Article

Hypoxia-Induced Energy = Stress=20 Regulates mRNA Translation and Cell Growth

Liping Liu,^¹,²=20Timothy=20 P. Cash,^²=20Russell=20 G. Jones,^²=20Brian Keith,^²=20Craig=20 B. Thompson,^²=20and M.=20 Celeste Simon^¹,²,

1 Howard Hughes=20 Medical Institute, University of Pennsylvania School of Medicine, = 421=20 Curie Boulevard, Philadelphia, Pennsylvania 191042 Abramson Family=20 Cancer Research Institute, University of Pennsylvania School of = Medicine,=20 421 Curie Boulevard, Philadelphia, Pennsylvania 19104

Correspondence:
M. Celeste Simon
Ph: (215) = 746-5532;=20 Fax: (215) 746-5511
celeste2@mail.med.upenn.edu

Article Options
= Table=20 of Contents Summary PDF=20 (679K) Supplemen= tal=20 Data
Search Medline = for
Liping= =20 Liu M. = Celeste Simon
Related = Articles
TOR=20 Signaling in Growth and Metabolism... TSC2=20 Mediates Cellular Energy Response to Control ... Phosphorylation=20 and Functional Inactivation of TSC... More=20 related=20 articles...

Summary

Oxygen (O₂) deprivation, = or=20 hypoxia, has profound effects on cell metabolism and growth. Cells = can=20 adapt to low O₂ in part through activation of = hypoxia-inducible=20 factor (HIF). We report here that hypoxia inhibits mRNA = translation by=20 suppressing multiple key regulators, including eIF2=CE=B1, eEF2, = and the=20 mammalian target of rapamycin (mTOR) effectors 4EBP1, = p70^S6K,=20 and rpS6, independent of HIF. Hypoxia results in energy starvation = and=20 activation of the AMPK/TSC2/Rheb/mTOR pathway. Hypoxic = AMP-activated=20 protein kinase (AMPK) activation also leads to eEF2 inhibition. = Moreover,=20 hypoxic effects on cellular bioenergetics and mTOR inhibition = increase=20 over time. Mutation of the TSC2 tumor suppressor gene confers a = growth=20 advantage to cells by repressing hypoxic mTOR inhibition and=20 hypoxia-induced G₁ arrest. Together, eIF2=CE=B1, eEF2, = and mTOR=20 inhibition represent important HIF-independent mechanisms of = energy=20 conservation that promote survival under low O₂=20 conditions.

Introduction

O₂=20 is essential for the viability and function of most metazoan = organisms and=20 closely monitored at both the organismal and cellular levels. To = survive=20 O₂ deficiency, cells activate a variety of adaptive = mechanisms=20 (Hochachka et al., 1996), including inhibition of energy-costly mRNA = translation=20 (Arsham et al., 2003; Koumenis et al., 2002; Wouters et al., 2005). mRNA translation is a critical component of cell = growth=20 and proliferation that is primarily regulated at the initiation = stage,=20 namely the assembly of the m⁷-GTP cap binding = eIF4E/eIF4A/eIF4G=20 (eIF4F) and the 40S ribosome binding eIF2/GTP/met-tRNA ternary = complexes.=20 Formation of these complexes is closely modulated by the = phosphorylation=20 status of eIF4E binding protein (4EBP) and eIF2=CE=B1 and highly = influenced by=20 environmental stimuli, such as mitogens and nutrients, and = environmental=20 stresses. For example, amino acid starvation and UV irradiation = inhibit=20 protein synthesis by inducing eIF2=CE=B1 phosphorylation at Ser51, = leading to=20 an enhanced affinity for GDP and reduced formation of the=20 eIF2/GTP/met-tRNA complex (Clemens, 2001; Liu et al., 2004). Additional regulation also occurs at translation = elongation (Patel et al., 2002). Upon energy starvation, the AMPK phosphorylates = eEF2=20 kinase (eEF2K) on Ser398 and activates its kinase activity (Browne = et al.,=20 2004). eEF2K then phosphorylates elongation factor eEF2 = at Thr56,=20 resulting in the inhibition of peptide elongation (Patel et al., = 2002).

The = Ser/Thr=20 kinase mTOR integrates signals from growth factors and nutrients = to=20 regulate cell size, division, and metabolism through control of=20 cap-dependent translation (Abraham, 2004). Phosphorylation of downstream targets = p70^S6K=20 and rpS6 by mTOR results in increased ribosome biogenesis, via = enhanced=20 translation of =E2=80=9CTOP=E2=80=9D mRNAs encoding ribosomal = proteins and elongation=20 factors (Schmelzle et al., 2000). mTOR also phosphorylates 4EBPs, rendering them = unable to=20 bind to and inhibit eIF4E, therefore promoting cap-dependent = translation.=20 mTOR activity is regulated by the PI3K/Akt, AMPK, and ERK pathways = (Inoki=20 et al., 2002; Inoki et al., 2003; Ma et al., 2005), which converge at the tuberous sclerosis complex = (TSC)=20 that consists of TSC1 and TSC2 dimers (Kwiatkowski, 2003; Li et al., 2004; Manning et al., 2003). The TSC complex represses mTOR by acting as a=20 GTPase-activating protein (GAP) toward the small GTPase Rheb (Li = et al.,=20 2004). Whereas TSC2 phosphorylation by Akt and ERK = suppresses=20 TSC2 GAP activity (Kwiatkowski, 2003; Ma et al., 2005), TSC2 phosphorylation by AMPK enhances its = activity (Inoki=20 et al., 2002; Inoki et al., 2003). The stress-induced proteins REDD1 and REDD2 = potently=20 inhibit mTOR signaling, and REDD1 functions downstream of Akt and = upstream=20 of TSC2 (Corradetti et al., 2005; Schwarzer et al., 2005). Multiple upstream regulators of mTOR, including = PTEN,=20 TSC1, TSC2, and LKB1, are frequently mutated in human cancers, = suggesting=20 a possible role for deregulated protein synthesis during tumor = development=20 (Tee et al., 2005).

We = have=20 previously shown that modest hypoxia (1.5% O₂) inhibits = mRNA=20 translation via the rapid hypophosphorylation of mTOR targets = 4EBP1,=20 p70^S6K, and rpS6 (Arsham et al., 2003). More recent studies have revealed that hypoxic = mTOR=20 regulation requires the TSC1/TSC2 complex and de novo = transcription and=20 expression of REDD1 (Brugarolas et al., 2004). REDD1 and its Drosophila homolog=20 Scylla are HIF-target genes induced by low O₂=20 (Brugarolas et al., 2004; Reiling et al., 2004). Simultaneous loss of Scylla and the related = protein=20 Charybdis decreases the survival rate of flies under low = O₂=20 (Reiling et al., 2004).

Although the=20 HIF-dependent REDD1-mediated feedback inhibition of mTOR provides = insight=20 into hypoxic regulation of mTOR, the rapid mTOR inhibition (within = 15 min)=20 by hypoxia cannot be explained by transcriptional activation of = REDD1. In=20 addition, the cellular mechanism by which hypoxia activates TSC2 = has not=20 been established nor has a role for the TSC2 substrate Rheb in = this=20 regulation been determined. In this study, we have investigated = the=20 effects of modest hypoxia (1.5% and 0.5% O₂) on = different=20 regulatory steps of mRNA translation. We report that hypoxia = concomitantly=20 inhibits eIF2=CE=B1 and eEF2, as well as the mTOR targets 4EBP1,=20 p70^S6K, and rpS6. Our data support regulatory roles for = TSC2=20 and HIF-inducible REDD1 during hypoxic mTOR regulation. However, = we also=20 demonstrate that hypoxic mTOR inhibition involves additional = signaling=20 pathways and can occur independent of HIF. Hypoxia promotes = cellular=20 energy starvation and activates the AMPK/TSC2/Rheb pathway, = resulting in=20 mTOR inhibition. TSC2 mutation in tumor cells blocks hypoxic mTOR=20 inhibition and hypoxia-induced cell cycle arrest, thereby = conferring a=20 growth advantage under low O₂. In summary, we present = evidence=20 that hypoxia induces conditions of energy starvation, which lead = to=20 inhibition of protein synthesis. Our findings indicate that AMPK = and TSC2=20 play important roles in coupling hypoxic effects on cellular = energy levels=20 to control protein synthesis, cell growth, and = proliferation.

Results

Hypoxia Inhibits Multiple Signaling = Pathways=20 Regulating mRNA Translation

Multiple=20 reports describe hypoxic (1%=E2=80=931.5% O₂) and = anoxic (<0.02%=20 O₂) effects on mRNA translation (Arsham et al., 2003; Blais et al., 2004; Brugarolas et al., 2004; Horman et al., 2002; Koumenis et al., 2002). Modest hypoxia can occur in various = physiological and=20 pathological conditions (embryogenesis, tumor formation, and=20 inflammation). Therefore, we evaluated the effects of modest = hypoxia (1.5%=20 O₂) on multiple regulators of protein synthesis by = using serum=20 replete HEK293 cells and rhabdomyosarcoma Rh30 cells that have = been used=20 for analyzing mTOR signaling (Dudkin et al., 2001; Inoki et al., 2003). As expected, insulin enhanced the = phosphorylation of=20 p70^S6K, rpS6, and 4EBP1, whereas LY294002 and = 2-deoxy-D-glucose=20 (2-DG) resulted in hypophosphorylation of these mTOR effectors in = HEK293=20 cells (Figure 1A). Notably, hypoxia = resulted in=20 hypophosphorylation of p70^S6K, rpS6, and 4EBP1 within 6 = hr, and=20 hypophosphorylation was enhanced by extended exposure to low = O₂=20 (20 hr) in HEK293 (Figure 1A) and Rh30 cells (Figure S1A available in the Supplemental Data with this = article online).=20 Hypoxia (up to 20 hr) resulted in no change in total rpS6 (Figure 1A), 4EBP1, or = p70^S6K=20 protein levels (data not shown). These results indicate that low=20 O₂ alone can effectively inhibit mTOR activity. In = addition to=20 mTOR inhibition, hypoxia also caused the hyperphosphorylation and=20 inhibition of eIF2=CE=B1 (Ser51) and eEF2 (Thr56) in HEK293 (Figure 1B) and Rh30 cells (Figure S1B). These changes were = detected=20 within 2 hr and enhanced by extended hypoxia. Taken together, = these data=20 indicate that moderate hypoxia suppresses translation by = inhibiting the=20 activity of mTOR, eIF2, and eEF2.

Figure 1. Hypoxia = Inhibits=20 Protein Synthesis by Rapidly Suppressing Multiple Key = Regulators of=20 mRNA Translation

(A) HEK293 cells were = subjected to=20 normoxia (21% O₂) or=20 hypoxia (1.5% O₂)=20 for 0=E2=80=9320 hr or treated with insulin, LY294002 (LY), = or=20 2-deoxy-D-glucose (2-DG) for 0.5 hr. Whole-cell extracts = were=20 analyzed by Western blot for phosphorylation of the indicted = residues of 4EBP1 and rpS6.

(B) Cell extracts were = analyzed for=20 phosphorylation of eIF2=CE=B1 at Ser51 and eEF2 at Thr56. = Sample loading=20 was adjusted with total eIF2=CE=B1 protein.

(C and D) Serum replete = (Ser+) and = serum-depleted=20 (Ser^=E2=88=92) HEK293 cells=20 were exposed to 21% or 1.5% O2 for = 2=E2=80=9348 hr. Protein=20 synthesis was measured by 35S-Met = incorporation=20 followed by TCA precipitation. The same samples were also = analyzed=20 by gel electrophoresis (see Figures S1D and S1E). Results were = expressed=20 relative to cells grown under normoxia. Data for serum = replete cells=20 are shown in (D). Error bars represent the standard error of = the=20 mean (SEM).

(E) HEK293 cells were grown = at 0.3%=20 O₂ for = 0=E2=80=936 hr, and=20 phosphorylation of 4EBP1, rpS6, eIF2=CE=B1, and eEF2 was = examined. 4EBP1=20 hypophosphoryaltion is indicated by its shift from the=20 hyperphosphorylated =CE=B3 and =CE=B2 forms to the less = phosphorylated =CE=B1=20 form.

(F) Protein synthesis for = serum replete=20 HEK293 cells grown at 0.3% O2 is shown. = Error bars=20 represent the SEM.

(G) Hypoxic mTOR inhibition = can occur=20 independent of HIF activity. ARNT+/+ or ARNT^{=E2=88=92/=E2=88=92} MEFs = were treated=20 with 1.5% O₂ for=20 0=E2=80=9320 hr, and phosphorylation of p70S6K and rpS6 = was examined=20 by Western blot.

View larger version: [In=20 this window] [In new = window]

Both mTOR and PERK Participate in = Hypoxic=20 Inhibition of Protein Synthesis

Anoxia has been=20 shown to rapidly reduce protein synthesis by 60%=E2=80=9370% = within 1 hr=20 (Kraggerud et al., 1995) and decrease the amount of polysome-associated = mRNA within=20 4 hr (Wouters et al., 2005). To assess the effect of modest hypoxia on = protein=20 synthesis, we measured the levels of ³⁵S-methionine = (Met)=20 incorporation in HEK293 cells grown under 1.5% O₂. = Despite more=20 rapid changes in 4EBP1, eIF2=CE=B1, and eEF2 phosphorylation = status, hypoxia=20 alone (up to 24 hr) did not cause any significant decrease in = protein=20 synthesis in serum replete HEK293 cells (Figure 1C). However, =E2=88=BC25% = and 40% reductions=20 in protein synthesis were detected after 32 and 48 hr of hypoxia = (Figure 1D). A similar delay in = protein=20 synthesis was observed in Rh30 cells where up to 24 hr hypoxia did = not=20 result in any significant drop in protein synthesis (Figure S1C). Of note, serum = starvation=20 facilitated hypoxic reduction of protein synthesis as = 20%=E2=80=9325% decreases in=20 ³⁵S-Met incorporation were detected in HEK293 cells = after 16=20 and 24 hr of hypoxia (Figure 1C). Trichloroacetic acid = (TCA)=20 precipitation and gel electrophoresis of labeled proteins yielded=20 comparable results for these assays. Densitometric analysis showed = a=20 significant 25%=E2=80=9330% attenuation in protein synthesis in = serum-deprived=20 HEK293 cells after 16=E2=80=9324 hr of hypoxia, but not in serum = replete cells (Figures S1D and S1E). This delay = in protein=20 synthesis inhibition could be due to a requirement for threshold = changes=20 (at least 3- to 4-fold) in eIF4F, eIF2, and eEF2 activities that=20 eventually cause a significant drop in protein synthesis, = especially as=20 hypoxic mTOR inhibition becomes more pronounced as the treatment = extends.=20 Of note, HEK293 cells cultured under more stringent hypoxia (0.3%=20 O₂) exhibited a 25% reduction in protein synthesis by 6 = hr (Figure 1F). This accelerated drop = in=20 translation was accompanied by 3- to 4-fold changes in 4EBP1 and = rpS6 by 2=20 hr (Figure 1E), as opposed to changes = that=20 occurred at 6-20 hr at 1.5% O₂. Furthermore, a 3-fold = change in=20 eEF2 phosphorylation occurred at 1.5% O₂ at 2 hr, = whereas a 20-=20 to 30-fold change was observed at 0.5=E2=80=932 hr at 0.3% = O₂. In=20 contrast, we saw comparable changes in eIF2=CE=B1 phosphorylation = at 1.5% and=20 0.3% O₂, in agreement with published data (Koumenis et = al.,=20 2002). Serum replete HEK293T cells, which exhibited a = higher=20 metabolic rate than HEK293 cells (indicated by a higher rate of=20 ³⁵S-Met incorporation), were inhibited more readily by = culture=20 at 1.5% O₂. Approximately 20%=E2=80=9325% reductions = were detected as=20 early as 16 hr in HEK293T cells (Figure S1C). These results suggest = that=20 intracellular ATP levels could influence changes in protein = synthesis=20 rates in hypoxic cells. Additionally, hypoxic mTOR inhibition is = enhanced=20 by a further decrease in O₂ levels and growth factor=20 withdrawal.

The = endoplasmic=20 reticulum (ER)-resident kinase PERK, which is activated by ER = stress, has=20 been shown to inhibit mRNA translation by phosphorylating = eIF2=CE=B1 on Ser51=20 (Harding et al., 1999). Severe hypoxia activates PERK in a number of = cell types=20 (Koumenis et al., 2002). Loss of PERK activity in MEFs significantly = attenuated=20 hypoxic inhibition of mRNA translation, as indicated by only a 25% = decrease in PERK^{=E2=88=92/=E2=88=92} MEFs relative to a = 55% decrease observed in=20 PERK^+/+ MEFs after 48 hr exposure to 1.5% O₂ = (Figure S1F). This difference = strongly suggests=20 that both the mTOR and eIF2=CE=B1 pathways contribute to = hypoxia-induced=20 translation inhibition.

Hypoxic Inhibition of mTOR Can Occur = Independent of HIF

Brugarolas et=20 al., 2004 recently reported that TSC2 and the HIF target = gene REDD1=20 are necessary and sufficient for hypoxic mTOR inhibition. This is = in=20 contrast to our previous report that hypoxia downregulates mTOR = activity=20 within 15 min (Arsham et al., 2003). Therefore, we evaluated the role of HIF and = HIF-inducible=20 REDD1 in hypoxic mTOR inhibition by using wild-type (wt) MEFs or = those=20 carrying a null mutation for the aryl hydrocarbon nuclear = translocator=20 (ARNT), the protein that forms heterodimeric DNA binding complexes = with=20 HIF-1=CE=B1 or HIF-2=CE=B1. HIF activity, examined by a hypoxia = response element=20 (HRE)-luciferase reporter assay, was completely abolished in=20 ARNT^{=E2=88=92/=E2=88=92} MEFs (Figure S2A), confirming that = hypoxic induction=20 of HIF activity is absent in ARNT^{=E2=88=92/=E2=88=92} = MEFs. Hypoxia (0.5 hr)=20 resulted in marked hypophosphorylation of 4EBP1, = p70^S6K, and=20 rpS6 in ARNT^{=E2=88=92/=E2=88=92} MEFs (Figure 1G), clearly indicating = that HIF=20 activity and REDD1 induction are not necessary for the rapid = effect of=20 hypoxia on mTOR, especially as REDD1 is induced solely by HIF = under=20 hypoxia (Schwarzer et al., 2005). Actinomycin D, a Pol II transcription inhibitor, = effectively blocked hypoxic mTOR inhibition (Figure S2B) (Brugarolas et al., = 2004). However, treatment with actinomycin D also = resulted in=20 hyperphosphorylation of p70^S6K, rpS6, and 4EBP1 under = both=20 normoxic and reoxygenation conditions (Figure S2B). The latter = observation is=20 consistent with a report that actinomycin D has a stimulatory = effect on=20 p70^S6K and rpS6 (Loreni et al., 2000). Thus, the effects of actinomycin D on hypoxic = mTOR=20 inhibition could in part result from mTOR stimulation by = actinomycin=20 D.

Interestingly,=20 whereas p70^S6K and rpS6 were hypophosphorylated in both = ARNT^+/+ and ARNT^{=E2=88=92/=E2=88=92} MEFs = throughout the 20 hr of=20 hypoxic treatment, hypophosphorylation was partially reversed = within 2 hr=20 in ARNT^{=E2=88=92/=E2=88=92} MEFs. This reversal was not = observed until 10 hr in=20 ARNT^+/+ MEFs (Figure 1G). We hypothesize that = the reduced=20 mTOR inhibition detected in ARNT^{=E2=88=92/=E2=88=92} = MEFs (2=E2=80=9310 hr) could be due=20 to the loss of REDD1-mediated mTOR inhibition. The biphasic = hypoxic mTOR=20 inhibition described here is likely caused by different = mechanisms.=20 Although HIF-induced REDD1 activity appears to contribute to = hypoxic mTOR=20 inhibition, our data indicate that HIF activity is not necessary = for=20 hypoxic mTOR regulation.

Hypoxia Affects Cellular = Bioenergetics and=20 Activates AMPK

Given that=20 O₂ levels regulate oxidative phosphorylation and that = AMPK=20 plays an important role in regulating cellular energy homeostasis = (Hardie=20 et al., 2003), we investigated the effect of hypoxia on = cellular energy=20 status and AMPK activity. AMPK is activated through binding of AMP = and=20 subsequent phosphorylation by upstream kinases at a threonine = residue=20 within the activation loop of the catalytic =CE=B1 subunit (Thr172 = in human=20 AMPK) (Carling, 2004). As shown in Figure 2A, exposure to 1.5% = O₂=20 (2=E2=80=9320 hr) resulted in a mild AMPK hyperphosphorylation at = Thr172 in serum=20 replete HEK293 cells. Hyperphosphorylation of AMPK targets = acetyl-coA=20 carboxylase (ACC), and eEF2 was more readily detected at all time = points=20 examined in both HEK293 (Figure 2A) and Rh30 cells (Figure S3B). Phosphorylation of = these=20 effectors in HEK293 cells (including AMPK) was further enhanced = when=20 O₂ tension was decreased to 0.5% (Figure S3A). These results = indicate that low=20 O₂ alone can activate AMPK and that hypoxic AMPK = stimulation is=20 affected by the severity of O₂ deprivation.

Figure 2. = Hypoxia-Induced Energy=20 Starvation and AMPK Activation Are Promoted by Growth Factor = Withdrawal

HEK293 cells were exposed to = 21% or=20 1.5% O₂ = for 0=E2=80=9320 hr=20 in the presence (A=E2=80=93C) or absence (D=E2=80=93F) of = 10% serum.

(A and D) Hypoxia activates = AMPK. Cell=20 extracts were analyzed for AMPK, ACC, eEF2, and 4EBP1=20 phosphorylation. Levels of total AMPK are shown to assess = sample=20 loading.

(B=E2=80=93F) Hypoxia = modulates cellular ATP=20 and ADP:ATP ratios. Levels of ATP (B and E) and ADP/ATP (C = and F) in=20 cell lysates. Total ATP was adjusted to levels of ATP at 0 = hr for=20 both normoxic and hypoxic conditions. Error bars represent = the=20 SEM.

View larger version: [In=20 this window] [In new = window]

Hypoxia, when=20 coupled with serum starvation, promoted AMPK, ACC, and eEF2=20 hyperphosphorylation within 30 min to levels comparable to those = induced=20 by 2-DG (Figure 2D). Furthermore, hypoxia = resulted in=20 more rapid and significant levels of mTOR inhibition in = serum-starved=20 cells when compared with serum replete HEK293 cells. = Hypophosphorylation=20 of 4EBP1 and rpS6 was readily detected as quickly as 30 min after = a 2 hr=20 serum depletion (Figure 2A and 2D and data not = shown). These=20 changes increased as hypoxic treatment extended. Serum deprivation = alone=20 did not cause any significant change in AMPK activity under = normoxia (0=E2=80=936=20 hr). However, we observed mild increases in AMPK and ACC = phosphorylation=20 after 20 hr of normoxia, possibly due to reduced glucose in the = culture=20 medium at this time point (Figure 2D). Hypoxic AMPK = activation did not=20 correlate with any significant increase in LKB1 = autophosphorylation on=20 Thr189, an AMPKK that has been shown to phosphorylate AMPK (data = not=20 shown) (Carling, 2004). Furthermore, Akt activity measured by = phosphorylation on=20 Ser473 was only slightly reduced by 6=E2=80=9320 hr hypoxia in = HEK293 cells in=20 both serum replete and serum-depleted conditions (data not shown). = The=20 latter suggests that the rapid hypoxic mTOR inhibition is unlikely = to be=20 caused by the PI3K/Akt cascade.

Although 1.5%=20 O₂ caused a modest hyperphosphorylation of AMPK, ACC, = and eEF2,=20 this level of O₂ did not correlate with any significant = decrease in cellular ATP over a 20 hr time course in serum replete = HEK293=20 cells (Figure 2B). In contrast, a = statistically=20 significant 18% drop occurred at 2 hr in serum-starved cells, = whereas a=20 50% drop in ATP levels was observed after 20 hr of hypoxia (Figure 2E). Moreover, 20 hr of = hypoxia=20 resulted in an 47% increase in the ADP:ATP ratio in serum replete = cells=20 (Figure 2C) and 30% and 55% = increases in=20 ADP/ATP after 2 hr or 20 hr of hypoxia in serum-depleted = conditions (Figure 2F). Hypoxia-induced = changes in ADP/ATP=20 were comparable to those caused by both 25 mM 2-DG and = staurosporine,=20 which inhibit intracellular bioenergetics (Figures S3C and S3D). Taken = together, our=20 results indicate that extended hypoxia regulates cellular = bioenergetics=20 and that AMPK is activated by hypoxia, possibly as a consequence = of=20 increases in the intracellular ADP:ATP ratio. A combination of = hypoxia and=20 growth factor withdrawal further promotes the changes in ATP = levels and=20 AMPK activity, rendering AMPK a better substrate for = LKB1.

AMPK Inhibits mTOR and eEF2 during=20 Hypoxia

To = investigate=20 the effect of AMPK activation on hypoxic mTOR modulation, = serum-depleted=20 HEK293 cells were treated with the AMPK inhibitor compound C (2 = =CE=BCM or 10=20 =CE=BCM) then exposed to 21% or 1.5% O₂. As shown in Figure 3, the AMPK inhibitor = effectively=20 blocked 4EBP1, p70^S6K, and rpS6 hypophosphorylation = caused by=20 2-DG (Figure 3A) and low O₂ = (Figure 3B), restoring 4EBP1,=20 p70^S6K, and rpS6 phosphorylation to levels similar to = those=20 observed under normoxic conditions. Compound C itself had a = negligible=20 effect on mTOR signaling under normoxia (Figure 3B). In addition, compound = C blocked=20 AMPK activation as evidenced by decreased AMPK, ACC, and eEF2=20 phosphorylation (Figure 3C), indicating that = hypoxic activation=20 of eEF2K by AMPK results in the phosphorylation and inhibition of=20 eEF2.

Figure 3. AMPK = Suppresses mTOR=20 Activity under Hypoxia

(A=E2=80=93C) Serum-starved = HEK293 cells were=20 pretreated with 2 or 10 =CE=BCM AMPK inhibitor compound C = for 30 min=20 before exposure to 2-DG (A) or 21% or 1.5% O2 for 1 or 3 hr = (B and=20 C).

(D) Serum replete HEK293 = cells were=20 preincubated with 10 mM methylpyruvate (MP) for 2 hr prior = to 2-DG=20 or 1.5% O₂ = (20 hr)=20 treatment. Phosphorylation of eEF2, 4EBP1, and rpS6 are=20 indicated.

(E and F) Kinase-inactive = AMPK =CE=B12=20 mutant suppresses hypoxic mTOR inhibition. HEK293 cells were = transfected with empty vector, wild-type AMPK-=CE=B12 = (AMPK-WT), or=20 kinase inactive K45R mutant (AMPK-DN). Clones expressing = similar=20 levels of AMPK protein were selected. (E) Levels of AMPK = protein in=20 selected clones. (F) Phosphorylation of p70S6K and rpS6 in = cells=20 exposed to 1.5% O₂=20 for 0=E2=80=9320 hr.

View larger version: [In=20 this window] [In new = window]

Methylpyruvate=20 (MP) has been shown to be efficiently metabolized by mitochondria, = resulting in increased ATP generation (Jijakli et al., 1996). HEK293 cells were preincubated with 10 mM MP = prior to 2-DG=20 or low O₂ treatment. MP effectively suppressed eEF2=20 hyperphosphorylation and 4EBP1 and rpS6 hypophosphorylation caused = by=20 hypoxia and 2-DG (Figure 3D). These results strongly = support a=20 role for energy depletion as the cause of AMPK activation in = hypoxic=20 cells. To further verify the role of AMPK in hypoxia-mediated = inhibition=20 of mTOR, we generated HEK293 clones stably expressing the wt = =CE=B12 subunit of=20 AMPK or a dominant-negative K45R mutant (Figure 3E) (Mu et al., 2001). Dominant-negative AMPK effectively blocked low=20 O₂-induced hypophosphorylation of p70^S6K and = rpS6,=20 whereas expression of wt AMPK affected neither basal mTOR activity = nor=20 hypoxic mTOR inhibition (Figure 3F). From these data, we = conclude that=20 AMPK activation by hypoxia promotes mTOR and eEF2 = inhibition.

TSC2-Dependent and TSC2-Independent = Effects of=20 Hypoxia on mTOR

TSC2 activity=20 has been shown to be regulated by cellular energy status = downstream of=20 AMPK. Upon activation by energy starvation, AMPK enhances TSC2 GAP = activity via phosphorylation on Ser1337, Ser1341, and Ser1345, = resulting=20 in mTOR inhibition (Inoki et al., 2003). To assess the role of TSC2 in hypoxic regulation = of mTOR,=20 we examined hypoxic effects on mTOR by using cells lacking TSC2. = Eker rats=20 are a genetic model for TSC2 and develop spontaneous tumors in = multiple=20 organs due to a germline insertional mutation in one allele of the = TSC2=20 gene and a subsequent loss of heterozygosity (Yeung, 2004). Therefore, we employed two = TSC2^{=E2=88=92/=E2=88=92} cells=20 derived from mutant rats (ERC15, a renal carcinoma cell line, and = ELT3, a=20 uterine smooth muscle cell line) and a TSC2^+/+ cell = line=20 (TRKE2, a kidney epithelial cell line) from wt rats. Expression of = the=20 functional full-length TSC2 protein is absent in ERC15 and ELT3 = cells but=20 readily detected in TRKE2 cells (Figure 5A).

Basal levels of=20 p70^S6K and rpS6 phosphorylation were lower in the=20 TSC2^+/+ TRKE2 cells compared to that of = TSC2^{=E2=88=92/=E2=88=92}=20 ERC15 and ELT3 cells (Figure 4A), confirming the loss of = TSC2=20 activity in mutant cells. p70^S6K and rpS6 were readily=20 hypophosphorylated in cells treated with the PI3K inhibitor = LY294002 (Figure 4A) and rapamycin (data not = shown) in=20 all three cell lines. Although p70^S6K and rpS6 = phosphorylation=20 was effectively reduced by 2-DG in TRKE2 cells, TSC2 mutations in = ERC15=20 and ELT3 cells effectively abrogated this change (Figure 4A). This result confirms = that mTOR=20 signaling is refractory to energy stress in TSC2 null cells.

Figure 4. TSC2 = Modulates Hypoxic=20 mTOR Inhibition

(A) TSC2 null mutation = abolishes=20 2-DG-induced mTOR inhibition. TRKE2 (TSC2+/+), ERC15 = (TSC2=E2=88=92/=E2=88=92), and = ELT3 (TSC2=E2=88=92/=E2=88=92) cells = were treated=20 with LY294002 or 2-DG for 30 min. Phosphorylation of p70^S6K and rpS6 = was assessed=20 by Western blot.

(B) Hypoxia activates AMPK in = TSC2=E2=88=92/=E2=88=92 ELT3 = cells.=20 Serum-starved ELT3 cells were exposed to 21% or 0.5% O2 for = 0.5=E2=80=9348 hr. Levels=20 of AMPK and ACC phosphorylation were examined.

(C=E2=80=93F) Serum-starved = TRKE2, ERC15, and=20 ELT3 cells were exposed to 21% or 0.5% O2 for = 0=E2=80=9320 hr. Cell=20 extracts from normoxic (Nor) and hypoxic (Hyp) conditions = were=20 examined for phosphorylation of 4EBP1 (C), p70S6K, and rpS6=20 (D=E2=80=93F).

View larger version: [In=20 this window] [In new = window]

Hypoxia (0.5%=20 O₂) caused a marked increase in AMPK and ACC = phosphorylation in=20 serum-starved TSC2^{=E2=88=92/=E2=88=92} ELT3 cells (Figure 4B), indicating a = regulatory effect of=20 hypoxia on cellular bioenergetics and AMPK activity. In addition, = 20 hr of=20 hypoxia resulted in hypophosphorylation of 4EBP1 in all three cell = types.=20 This is indicated by the shift of 4EBP1 from the more = hyperphosphorylated=20 =CE=B3 form to the less phosphorylated =CE=B1 and =CE=B2 forms (Figure 4C). We also observed an = enhanced=20 binding of eIF4E to 4EBP1 in all three cell types (data not = shown). These=20 results suggest that TSC2 is not absolutely required for hypoxic=20 inhibition of 4EBP1. Hypoxia resulted in p70^S6K and = rpS6=20 hypophosphorylation in serum-starved = TSC2^{=E2=88=92/=E2=88=92} ERC15 and ELT3=20 cells after 6=E2=80=9320 hr, confirming that mTOR inhibition by = hypoxia could=20 proceed independently of TSC2. However, p70^S6K and rpS6 = hypophosphorylation was greatly reduced in = TSC2^{=E2=88=92/=E2=88=92} cells=20 relative to TSC2^+/+ TRKE2 cells, especially after = 0.5=E2=80=932 hr=20 hypoxia (Figure 4D=E2=80=934F). We also = observed a similar=20 reduction in hypoxic mTOR inhibition in serum-starved = TSC2^{=E2=88=92/=E2=88=92}=20 MEFs in comparison with TSC2^+/+ MEFs (Figure S4B). Taken together, these = data=20 suggest that, although hypoxic regulation of mTOR could occur=20 independently of TSC2 under long-term low O₂, TSC2 is = critical=20 for the rapid hypoxic effects on mTOR.

TSC2 Contributes to Hypoxic = Inhibition of mRNA=20 Translation

To = assess the=20 role of TSC2 in hypoxic regulation of mTOR and protein synthesis, = we=20 established genetically matched ERC15 cell lines expressing either = wt TSC2=20 (clone T3 and T21) or N1643K TSC2 harboring a mutation in the GAP = domain=20 (clone DN4). As seen in Figure 5A, levels of wt or the = mutant protein=20 in ERC15 cells were similar to that of endogenous TSC2 in TRKE2 = cells (Figure 5A). Expression of wt TSC2 = effectively=20 promoted hypoxic mTOR inhibition in serum-starved = TSC2^{=E2=88=92/=E2=88=92} ERC15=20 cells, as indicated by rapid hypophosphorylation of = p70^S6K or=20 rpS6 (Figure 5B). In contrast, hypoxia = had little=20 effect on mTOR activity in cells expressing the TSC2 N1643K mutant = or=20 empty pcDNA3 vector (Figure 5B).

Figure 5. TSC2 and = Rheb Modulate=20 Hypoxia-Induced Inhibition of mTOR and Protein Synthesis

(A=E2=80=93C) TSC2=E2=88=92/=E2=88=92 ERC15 = cells were=20 stably transected with empty vector (Vec), wild-type (T3 and = T21),=20 or N1643K (DN4) TSC2. Clones expressing TSC2 protein at = levels=20 similar to TRKE2 cells were selected. The asterisk () indicates a nonspecific polypeptide = detected by the=20 anti-TSC2 antibody. (A) Expression of TSC2 protein in = selected ERC15=20 clones. (B) ERC15 clones were examined for phosphorylation = of=20 p70^S6K and = rpS6=20 after 0.5% O₂=20 treatment and serum deprivation. (C) ERC15 cells were = treated with=20 21% or 0.5% O₂ for=20 48 hr in the presence (Ser+) or absence = (Ser=E2=88=92) of = 10% serum. The=20 asterisk indicates statistically significant differences = between=20 Ser⁺ and = Ser=E2=88=92 = conditions; the diamond=20 (=E2=8B=84) indicates statistically significant differences = between normoxic=20 and hypoxic samples. In (C), error bars represent the = SEM.

(D) HEK293 cells were = transfected with=20 empty vector or Flag-tagged Rheb. Phosphorylation of 4EBP1 = and rpS6=20 and expression of Flag-Rheb were examined during O2 and serum = deprivation=20 by Western blot.

View larger version: [In=20 this window] [In new = window]

We = next=20 examined the effect of TSC2 on protein synthesis under hypoxic = conditions=20 by using these cell lines. As shown in Figure 5C, TSC2 null cells = (Vector) or cells=20 expressing the N1643K mutant (DN4) displayed a 15% and 30% = decrease in=20 protein synthesis after 48 hr in serum replete and serum-depleted=20 conditions, respectively. Of note, TSC2 expression resulted in=20 statistically significant changes in ³⁵S-Met = incorporation: 25%=20 and 47% under serum replete and serum-depleted conditions, = respectively=20 (Figure 5C). Taken together, these = data=20 indicate that TSC2 can regulate protein synthesis during hypoxia = and that=20 serum starvation exacerbates this response.

Rheb Regulates Hypoxic mTOR=20 Inhibition

Recent genetic=20 studies in yeast, Drosophila, and mammalian cells have = placed Rheb=20 downstream of TSC1/TSC2 in the control of mTOR activity. = Biochemical=20 analyses also show that Rheb is a physiological substrate for the = GAP=20 activity of TSC2 (Li et al., 2004). To investigate whether hypoxia-mediated mTOR = inhibition=20 involves Rheb, we examined the phosphorylation status of mTOR = effectors in=20 serum-starved HEK293 cells overexpressing Rheb under 1.5% = O₂.=20 As shown in Figure 5D, expression of Rheb in = HEK293 cells=20 induced a modest increase in basal rpS6 phosphorylation. = Importantly, it=20 also effectively blocked 4EBP1 and rpS6 hypophosphorylation during = 20 hr=20 of hypoxia, indicating that Rheb also mediates hypoxic mTOR=20 regulation.

Hypoxic Regulation of Cell Growth = and=20 Proliferation Requires TSC2

In = addition to=20 its regulatory effects on mRNA translation, the TSC1/TSC2 complex=20 negatively modulates cell growth and proliferation (Potter et al., = 2001; Tapon et al., 2001). Moreover, low O₂ alters cell = proliferation by=20 inducing cell cycle arrest or programmed cell death (Goda et al., = 2003; Schmaltz et al., 1998). TSC2 deletion results in abnormally high levels = of=20 proliferation under hypoxic conditions (Brugarolas et al., 2004), suggesting that TSC2 plays a key role in the = control of=20 cell proliferation under low O₂. To investigate the = role of=20 TSC2 in hypoxic control of cell survival, size, and proliferation, = we=20 examined the cell cycle profile of TSC2^+/+ TRKE2 cells = and=20 TSC2^{=E2=88=92/=E2=88=92} ERC15 cells by = bromodeoxyuridine (BrdU)/propidium=20 iodide (PI) staining. As shown in Figure 6A, 24 hr of hypoxia = significantly=20 reduced the proliferation rate of TSC2^+/+ TRKE2 cells = (Figure 6A), an effect that was = abrogated in=20 cells lacking TSC2 (Figure 6B). The percentages of = TSC2 null ERC15=20 cells (stably transfected with either vector or the N1643K mutant) = in S=20 phase were not changed by low O₂ relative to normoxia. = However,=20 expression of wt TSC2 (T3 and T21) effectively restored hypoxic = cell cycle=20 arrest (Figure 6B and 6C). As shown in Figure 6B, the reduction of cells = in S phase=20 corresponded to a =E2=88=BC22% increase in cells in the = G₁ phase of the=20 cell cycle. These data suggest that loss of TSC2 confers a = proliferative=20 advantage to cells under hypoxic conditions, allowing them to = overcome=20 hypoxia-induced G₁ arrest.

Figure 6. TSC2 = Mutation Enhances=20 Cell Proliferation under Low O2 by Abolishing = Hypoxia-Induced Cell Cycle Arrest

(A=E2=80=93C) Serum replete = TRKE2 and ERC15=20 cells were exposed to 21% or 0.5% O2 for 24 hr. = Cell cycle=20 profile was determined by BrdU/PI staining and analyzed by = flow=20 cytometry. (A) Hypoxia suppresses S phase entry in TSC2+/+ TRKE2 = cells. The=20 graph shows the percentages of BrdU+ TRKE2 cells = under=20 normoxic and hypoxic conditions. (B) Representative FACS = plots for=20 TSC2^{=E2=88=92/=E2=88=92} ERC15 = cells=20 transfected with vector (Vec) or wild-type TSC2 (T21). (C)=20 Quantification of BrdU+ cells for = ERC15 clones.=20 In (A) and (C), error bars represent the SEM.

(D=E2=80=93F) ERC15 cell = clones were grown in=20 the presence of 10% FBS under 21% or 0.5% O2 for 7 days. = (D) ERC15=20 colonies were stained with 0.2% crystal violet. (E) = Quantifications=20 of size of ERC15 cells subjected to normoxic (light bars) = and=20 hypoxic (dark bars) conditions. Cell size was determined by = forward=20 scatter (FSC) for cells in G1 phase of the = cell=20 cycle. (F) Representative FACS plots for cell size = measurements=20 described in (E). ERC15 cells (Vec and T3) treated under = normoxia=20 (a); representative plots for Vec (b) and T3 cells (c) under = normoxic and hypoxic conditions. In (E), error bars = represent the=20 SEM.

View larger version: [In=20 this window] [In new = window]

We = next=20 investigated the effect of hypoxia and TSC2 on cell survival and = cell=20 proliferation by using a colony formation assay. ERC15 cells = transfected=20 with empty vector, wt TSC2, or mutant TSC2 were grown in the = presence of=20 10% serum under 21% or 0.5% O₂ for 7 days. Similar = numbers of=20 cell colonies were observed among ERC15 cells treated under = normoxia or=20 hypoxia (Figure 6D), suggesting that = neither hypoxia=20 nor loss of TSC2 resulted in any significant cell death during = this 7 day=20 treatment. Interestingly, hypoxia caused decreases in overall = colony size=20 (Figure 6D). Expression of wt TSC2 = in=20 TSC2^{=E2=88=92/=E2=88=92} ERC15 cells promoted a further = reduction in colony=20 size. We then examined whether the difference in colony size was = the=20 result of changes in cell size. As shown in Figure 6E and 6F, TSC2 = inactivation resulted=20 in increased cell size. TSC2^{=E2=88=92/=E2=88=92} ERC15 = cells expressing control=20 vector (Vec) or the GAP mutant (DN4) were =E2=88=BC6% larger than = wt TSC2=20 expressing cells (T3 and T21). The 7 day hypoxia treatment = significantly=20 decreased cell size in all cell types. TSC2 did not show any = additive=20 effect with hypoxia on cell size; a 10% decrease in cell size was = observed=20 in all ERC15 cell lines regardless of TSC2 status. This result = suggests=20 that, although TSC2 plays an important role in mediating hypoxic = mTOR=20 inhibition, especially the early response, it has a less = significant=20 effect on cell mass accumulation under chronic hypoxic conditions. = Because=20 the differences in colony size cannot be explained by changes in = cell=20 size, our data demonstrate that tumor cells lacking TSC2 have a=20 significant proliferative advantage under low O₂, and = this=20 enhanced proliferation is likely due to escape from = hypoxia-induced G1=20 cell cycle arrest.

Discussion

We = report here=20 several mechanisms of HIF-independent hypoxic regulation of = protein=20 synthesis and cell growth: (1) hypoxia-induced cellular energy = depletion,=20 (2) mTOR inhibition via the AMPK/TSC2/Rheb pathway, (3) eEF2 = inhibition=20 mediated by AMPK, and (4) induction of ER stress that leads to = eIF2=CE=B1=20 inhibition (Figure 7). AMPK inhibition, TSC2 = null=20 mutation, and Rheb overexpression all suppress hypoxic mTOR = inhibition.=20 Mutations of TSC2 in tumor cells promote cell growth under low=20 O₂ by overcoming hypoxia-induced mTOR inhibition and=20 G₁ cell cycle arrest.

Figure 7. Schematic = Diagram for=20 the Signaling Pathways Modulated by Low O2 that Lead to=20 Translational Inhibition

Hypoxia suppresses signaling = for=20 translation initiation, elongation, and ribosome biogenesis = via=20 concomitant inhibition of eIF2=CE=B1, eEF2, and mTOR = downstream targets=20 4EBP1, p70^S6K, and=20 rpS6. The mTOR regulation involves the AMPK/TSC2/Rheb = pathway and is=20 activated by hypoxia-induced decreases in cellular = bioenergetics.=20 The eIF2=CE=B1 phosphorylation may be triggered by ER-stress = activated=20 PERK, and the eEF2 inhibition involves AMPK. HIF-inducible = REDD1 has=20 been implicated in the regulation of mTOR upstream of TSC2.=20 Mutations of TSC2 in tumor cells greatly impede = hypoxia-induced mTOR=20 inhibition and G₁=20 arrest. Thus, hypoxic translational control may ultimately = affect=20 cell growth and proliferation under low O2.

View larger version: [In=20 this window] [In new = window]

HIF = activity=20 and HIF-inducible REDD1 have been previously implicated as = necessary for=20 hypoxia-induced mTOR regulation (Brugarolas et al., 2004). However, although our data support a regulatory = role for=20 HIF in this hypoxic signaling, we demonstrate that hypoxic mTOR = inhibition=20 can occur independent of HIF. The latter is demonstrated by the = rapid mTOR=20 inhibition in both ARNT^+/+ and = ARNT^{=E2=88=92/=E2=88=92} MEFs. Our=20 data support a regulatory role for TSC2 during hypoxic mTOR = inhibition.=20 More importantly, we provide a molecular basis for TSC2 = regulation, namely=20 hypoxia-induced energy starvation and activation of the = AMPK/TSC2/Rheb=20 pathway. We propose that hypoxic mTOR regulation involves multiple = pathways, including energy-dependent control mediated by = AMPK/TSC2/Rheb,=20 inhibition by the HIF/REDD1/2 pathway, and TSC2-independent mTOR=20 inhibition (Figure 7).

Protein=20 synthesis consumes a great deal of ATP. Thus, a physiological = consequence=20 of hypoxia is the rapid change in signaling events controlling = protein=20 synthesis. We report such inhibitory signaling changes, as = indicated by=20 the concomitant inhibition of eIF2=CE=B1, eEF2, 4EBP1, = p70^S6K, and=20 rpS6, within 30 min. Whereas eIF2=CE=B1 inhibition is likely = induced by ER=20 stress via PERK, eEF2 and mTOR inhibition involves AMPK. In = contrast to=20 the rapid signaling changes, we observed a delay (between 24 and = 32 hr) in=20 the direct inhibition of ³⁵S-Met incorporation itself = by using=20 multiple cell lines. This delay suggests that signaling events = occur ahead=20 of a manifested biological change. The delay in protein synthesis=20 inhibition likely results from the requirement of threshold = amounts of=20 signaling changes and from compound effects of multiple stresses = such as=20 hypoxia and depletion of growth factors, amino acids, and glucose = in the=20 culture medium. This is supported by the acceleration of = hypoxia-induced=20 inhibition in protein synthesis by growth factor withdrawal or = more=20 stringent hypoxia (Figure 1C, 1F, and Figure 5C). We observed changes in = protein=20 synthesis when ADP:ATP achieved a ratio of 65%. It is noteworthy = that the=20 vast majority of solid tumors are characterized by insufficient=20 vasculature and heterogeneous oxygenation. This combination of = stresses=20 may mimic the environment in solid tumors where cells experience = chronic=20 depletion of O₂, growth factors and glucose. Therefore, = the=20 energy starvation and protein synthesis inhibition reported here = may=20 reflect the outcome of combined stresses experienced by cells = within solid=20 tumors.

Multiple=20 reports have shown that AMPK is critical for coordinating cell = growth,=20 proliferation, and survival under energy starvation conditions.=20 Phosphorylation of TSC2 by AMPK protects cells from energy=20 deprivation-induced apoptosis (Inoki et al., 2003). AMPK also coordinates cellular bioenergetics = with the=20 metabolic demands of cell proliferation by inducing a = p53-dependent cell=20 cycle checkpoint (Jones et al., 2005). Induction of this metabolic arrest promotes cell = survival.=20 We show here that AMPK acts as a hypoxic energy sensor and is = stimulated=20 by the changes in cellular energy status induced by low = O₂.=20 AMPK inhibits ATP-consuming anabolic processes during hypoxia such = as mRNA=20 translation and lipid biosynthesis, as reflected by = hyperphosphorylation=20 of eEF2 and ACC in our studies. It could also upregulate = ATP-generating=20 catabolic cellular processes such as lipid peroxidation and = glycolysis.=20 Thus, the AMPK-mediated mechanisms for energy maintenance could be = a=20 cellular mechanism for survival during hypoxia. Although it = remains to be=20 examined whether or not AMPK plays a role in hypoxic G₁ = arrest,=20 hypoxia-induced cell cycle arrest observed in our studies is = likely=20 another survival mechanism induced under this energy-deficient = state.=20 Tumor cells could take advantage of these mechanisms to survive = energy=20 deprivation, until hypoxia-stimulated adaptations such as = angiogenesis and=20 glycolysis allow for the reestablishment of energetic balance.

AMPK appears=20 not to be required for hypoxic mTOR regulation in some cases = (Arsham et=20 al., 2003; Brugarolas et al., 2004). It is worth noting that experimental conditions, = such as=20 the duration of hypoxic stress, the severity of hypoxia, and the=20 availability of growth factors all affect changes in AMPK activity = caused=20 by hypoxia. AMPK is a heterotrimeric complex comprised of a = catalytic =CE=B1=20 subunit and regulatory =CE=B2 and =CE=B3 units, and each subunit = exists in two to=20 three isoforms. The enzyme, when composed of different subunits in = different cell types, may respond with varying sensitivity to AMP = changes,=20 resulting in different levels of stimulation (Hardie et al., = 2003). Moreover, different cells may vary in their = relative=20 levels of AMP and ATP during hypoxia, therefore responding = differently to=20 energy starvation. Thus, although the AMPK/TSC2/Rheb/mTOR pathway=20 represents an important mechanism for mTOR regulation under = chronic=20 hypoxia, distinct cell types may be regulated differently by = hypoxia.

Our = data=20 clearly demonstrate that TSC2 is involved in hypoxic regulation of = protein=20 synthesis and cell proliferation. TSC2 mutation in tumor cells not = only=20 effectively suppresses AMPK-induced mTOR inhibition but also = overcomes=20 hypoxia-induced cell growth arrest. We further demonstrate that = TSC2=20 mutations confer a proliferative advantage to tumor cells by = abrogating=20 hypoxia-induced G₁ arrest. This change in the cell = cycle=20 profile cannot be explained by modulation of cyclin D1 or p27, as = hypoxia=20 reduces cyclin D1 and enhances p27 levels regardless of their TSC2 = status=20 (data not shown). More investigation is needed to determine the = mechanism=20 by which TSC2 mutation abrogates hypoxic cell cycle arrest. Our = data=20 suggest that the growth advantage of TSC2 null tumor cells under = low=20 O₂ does not result from increases in cell size. = Although=20 chronic hypoxia significantly reduces cell size, TSC2 mutation = does not=20 block this change. The lack of any significant effect of TSC2 on = hypoxic=20 cell size regulation could result from TSC2-independent mTOR = inhibition=20 and translational regulation under hypoxia. Alternatively, the = lack of=20 cell size regulation may simply be related to the enhanced = proliferation=20 of TSC2^{=E2=88=92/=E2=88=92} cells under low = O₂, especially because=20 cell size is coupled to both cell growth and proliferation.

In = summary, we=20 demonstrate that hypoxia controls protein synthesis by = concomitantly=20 inhibiting multiple key translational regulators independent of = HIF=20 activity. Inhibition of mTOR and eEF2 involves an AMPK-dependent = mechanism=20 triggered by hypoxic energy starvation. Although hypoxic mTOR = regulation=20 is mediated by multiple mechanisms, the AMPK/TSC2/Rheb pathway = represents=20 an important mechanism for HIF-independent control of translation, = especially during chronic hypoxia. TSC2 is also important for = hypoxic=20 regulation of cell growth, and loss of TSC2 function allows cells = to=20 escape hypoxia-induced mTOR inhibition and cell cycle arrest. = Targeting=20 components of these regulatory pathways may disrupt the growth and = proliferation of tumor cells and hamper their ability to adapt to = hypoxia=20 in vivo.

Experimental Procedures

Materials

Anti-TSC2=20 (C-20) and anti-actin were from Santa Cruz Biotechnology and = Chemicon=20 International. Anti-Flag was from Sigma. All other antibodies were = purchased from Cell Signaling technology. TRKE2 cells, ERC15 = cells, and=20 ELT3 cells were generously provided by Dr. C.L. Walker (University = Texas,=20 MD Anderson) (Howe et al., 1995). TSC2^{=E2=88=92/=E2=88=92}, = p53^{=E2=88=92/=E2=88=92}, TSC2^+/+,=20 and p53^{=E2=88=92/=E2=88=92} MEFs were gifts from Dr. = Kwiatkowski at Harvard=20 University. Rh30 cells were from Dr. P.J. Houghton (St. Jude = Children's=20 Research Hospital). Rat AMPK-=CE=B12 and K45R AMPK-=CE=B12 mutant = encoded in pcDNA3=20 vector were generously provided by Dr. M. Birnbaum (University of=20 Pennsylvania). Human wt TSC2 and N1643K mutant in pcDNA3-His = vector were=20 gifts from Dr. E. Henske (Fox Chase Cancer Institute). The Rheb = cDNA was=20 generated by RT-PCR and cloned into pcDNA3 vector with N-terminal = 2 =C3=97 Flag=20 tag. AMPK inhibitor compound C was obtained from Merck.

Cell Culture, Transfection, and = Clone=20 Selection

MEFs=20 established from ARNT^+/+ and = ARNT^{=E2=88=92/=E2=88=92} mice were=20 immortalized by SV40 large T-antigen. Both MEFs and HEK293 cells = were=20 maintained in DMEM containing 10% FBS and glutamine. TRKE2, ERC15, = and=20 ELT3 cells were cultured in DF8/F12 medium containing 10% FBS as = described=20 previously (Howe et al., 1995). For serum starvation, cells were cultured in = serum-free=20 media for 2 hr before normoxia, hypoxia, or pharmacological = treatment. For=20 serum replete conditions, cells were grown in fresh DMEM media = containing=20 10% FBS for 2 hr before any treatment. For pharmacological = treatments,=20 cells were exposed to insulin (100 =CE=BCM) (Sigma), LY294002 (10 = =CE=BCM) (Cayman=20 Chemical), or 2-DG (25 mM) (Sigma) for 30 min or 10 mM = methylpyruvate for=20 2 hr. The cells were plated at densities such that they reached = =E2=88=BC50%=E2=80=9370%=20 confluence at the time of study and were allowed to adhere = overnight prior=20 to any treatment.

HEK293 and=20 ERC15 cells were transfected with Lipofectin (Invitrogen). Stable=20 transfectants for AMPK=CE=B12 in HEK293 and TSC2 in ERC15 were = selected with=20 0.8 mg/ml or 0.4 mg/ml geneticin (Invitrogen), respectively. The = clones=20 were maintained in 0.5 mg/ml and 0.2 mg/ml geneticin.

Protein Synthesis by 35S-Met = Incorporation

HEK293 or ERC15=20 cells were exposed to normoxia or hypoxia (1.5% O₂) for = 2=E2=80=9348 hr=20 in serum replete (10% FBS) or serum-depleted (0% FBS) conditions. = Cells=20 were then switched to Met-free DMEM media supplemented with 10 = =CE=BCCi=20 ³⁵S-Met (Amersham Biosciences) and incubated for 1 hr. = The=20 cells were lysed in RIPA buffer containing complete protease = inhibitor=20 (Roche) and 1 mM PMSF. Radioactivity in the cell lysate was = examined by=20 TCA precipitation followed by Microbeta scintillation counting=20 (PerkinElmer) (Lang et al., 2002). Alternatively, cell lysate was resolved by 12.5% = SDS-PAGE=20 and examined by autoradiography. The amount of radiolabeling was = adjusted=20 with the protein content of each sample.

ATP and ATP/ADP = Measurements

Levels of ATP=20 or ADP:ATP ratio in HEK293 cell extracts were determined by ATP=20 bioluminescence assay kit CLS II (Roche Applied Science) and = ApoGlow assay=20 kit (Cambrex), respectively.

BrdU Labeling

ERC15 cells=20 were labeled with 10 =CE=BCM bromodeoxyuridine (BrdU) for the last = 45 min of=20 treatment after exposure to normoxia (21% O₂) or = hypoxia (0.5%=20 O₂) for 24 hr. Cells were fixed with ice cold 70% = ethanol and=20 stained with propidium iodide (10 =CE=BCg/ml) and Alexa Fluor-488 = conjugated=20 anti-BrdU antibody (Molecular Probes) and analyzed by flow = cytometry=20 (Becton Dickenson).

Statistical Analysis

Results are=20 average =C2=B1 standard error of mean (SEM) of six to eight = samples from two=20 independent studies. Statistic analyses were done by two-tailed = Student's=20 t test. Error bars represent SEM for Figure 1; Figure 2; Figure 5; Figure 6. Statistical significance = was defined=20 as and =E2=8B=84p < 0.05, p < 0.01.

Acknowledgments

We = thank Doug=20 Lin for technical assistance. Special thanks to John Gordan, Bryan = Barnhart, Regina Young, and Chengjun Hu for thoughtful discussions = and=20 reading of the manuscript. This research was supported by National = Institutes of Health Grant PO1 CA 104838 and the Abramson Family = Cancer=20 Research Institute. M.C.S. is an investigator of the Howard Hughes = Medical=20 Institute.

Supplemental=20 Data

Document S1. Four Figures

References

Abraham,=20 R.T. (2004). mTOR as a positive = regulator=20 of tumor cell responses to hypoxia. Curr. Top. Microbiol. = Immunol. 279, 299-319. [Medline]=20

Arsham, = A.M.,=20 Howell, J.J., and Simon, = M.C.=20 (2003). A novel hypoxia-inducible=20 factor-independent hypoxic response regulating mammalian target of = rapamycin and its targets. J.=20 Biol. Chem. 278, 29655-29660. [Medline]=20

Blais, = J.D.,=20 Filipenko, V., Bi, M., Harding, H.P., Ron, D., Koumenis, C., = Wouters,=20 B.G., and Bell, J.C. (2004). Activating=20 transcription factor 4 is translationally regulated by hypoxic=20 stress. Mol. Cell. Biol. 24, = 7469-7482. [Medline]=20

Browne, = G.J., Finn,=20 S.G., and Proud, C.G. (2004). Stimulation=20 of the AMP-activated protein kinase leads to activation of = eukaryotic=20 elongation factor 2 kinase and to its phosphorylation at a novel = site,=20 serine 398. J.=20 Biol. Chem. 279, 12220-12231. [Medline]=20

Brugarolas, J.,=20 Lei, K., Hurley, R.L., Manning, B.D., Reiling, J.H., Hafen, E., = Witters,=20 L.A., Ellisen, L.W., and Kaelin, W.G. (2004). Regulation of mTOR function in response to = hypoxia by=20 REDD1 and the TSC1/TSC2 tumor suppressor complex. = Genes Dev. 18,=20 2893-2904. [Medline]=20

Carling, D.=20 (2004). The AMP-activated protein = kinase=20 cascade=E2=80=93a unifying system for energy control. = Trends Biochem. Sci. 29, 18-24. [Medline]=20

Clemens,=20 M.J. (2001). Initiation factor = eIF2 alpha=20 phosphorylation in stress responses and apoptosis. = Prog. Mol. Subcell. Biol. = 27, 57-89. [Medline]=20

Corradetti, M.N.,=20 Inoki, K., and Guan, K.L. (2005). The=20 stress-inducted proteins RTP801 and RTP801L are negative = regulators of the=20 mammalian target of rapamycin pathway. J. Biol. Chem. 280,=20 9769-9772. [Medline]=20

Dudkin, = L.,=20 Dilling, M.B., Cheshire, P.J., Harwood, F.C., Hollingshead, M., = Arbuck,=20 S.G., Travis, R., Sausville, E.A., and Houghton, P.J. = (2001). Biochemical correlates of mTOR inhibition by = the=20 rapamycin ester CCI-779 and tumor growth inhibition. = Clin. Cancer Res. 7,=20 1758-1764. [Medline]=20

Goda, = N., Ryan,=20 H.E., Khadivi, B., McNulty, W., Rickert, R.C., and Johnson, = R.S.=20 (2003). Hypoxia-inducible factor = 1alpha is=20 essential for cell cycle arrest during hypoxia. = Mol. Cell. Biol. 23,=20 359-369. [Medline]=20

Hardie, = D.G.,=20 Scott, J.W., Pan, D.A., and Hudson, E.R. (2003). Management of cellular energy by the = AMP-activated=20 protein kinase system. FEBS Lett. 546, 113-120. = [Medline]=20

Harding, H.P.,=20 Zhang, Y., and Ron, D. (1999). Protein=20 translation and folding are coupled by an = endoplasmic-reticulum-resident=20 kinase. Nature 397, 271-274. [Medline]=20

Hochachka, P.W.,=20 Buck, L.T., Doll, C.J., and Land, S.C. (1996). Unifying theory of hypoxia tolerance:=20 molecular/metabolic defense and rescue mechanisms for surviving = oxygen=20 lack. Proc. Natl. Acad. Sci. USA 93, = 9493-9498.=20 [Medline]=20

Horman, = S., Browne,=20 G., Krause, U., Patel, J., Vertommen, D., Bertrand, L., Lavoinne, = A., Hue,=20 L., Proud, C., and Rider, M. (2002). Activation of AMP-activated protein kinase = leads to=20 the phosphorylation of elongation factor 2 and an inhibition of = protein=20 synthesis. Curr. Biol. 12, 1419-1423. = [Medline]=20 [Summary]=20 [Full=20 Text]

Howe, = S.R.,=20 Gottardis, M.M., Everitt, J.I., Goldsworthy, T.L., Wolf, D.C., and = Walker,=20 C. (1995). Rodent model of = reproductive=20 tract leiomyomata. Establishment and characterization of = tumor-derived=20 cell lines. Am. J. Pathol. 146, 1568-1579. = [Medline]=20

Inoki, = K., Li, Y.,=20 Zhu, T., Wu, J., and Guan, K.L. (2002). TSC2 is phosphorylated and inhibited by Akt = and=20 suppresses mTOR signalling. Nat. Cell Biol. 4, 648-657. = [Medline]=20

Inoki, = K., Zhu, T.,=20 and Guan, K.L. (2003). TSC2 = mediates=20 cellular energy response to control cell growth and = survival.=20 Cell 115, 577-590. [Medline]=20 [Summary]=20 [Full=20 Text]

Jijakli, H., Nadi,=20 A.B., Cook, L., Best, L., Sener, A., and Malaisse, W.J. = (1996).=20 Insulinotropic action of methyl = pyruvate:=20 enzymatic and metabolic aspects. Arch. Biochem. Biophys. 335, = 245-257.=20 [Medline]=20

Jones, = R.G., Plas,=20 D.R., Kubek, S., Buzzai, M., Mu, J., Xu, Y., Birnbaum, M.J., and = Thompson,=20 C.B. (2005). AMP-activated = protein kinase=20 induces a p53-dependent metabolic checkpoint. Mol. Cell 18, 283-293.=20 [Medline]=20 [Summary]=20 [Full=20 Text]

Koumenis, C.,=20 Naczki, C., Koritzinsky, M., Rastani, S., Diehl, A., Sonenberg, = N.,=20 Koromilas, A., and Wouters, B.G. (2002). Regulation of protein synthesis by hypoxia = via=20 activation of the endoplasmic reticulum kinase PERK and = phosphorylation of=20 the translation initiation factor eIF2alpha. Mol. Cell. Biol. 22,=20 7405-7416. [Medline]=20

Kraggerud, S.M.,=20 Sandvik, J.A., and Pettersen, E.O. (1995). Regulation of protein synthesis in human = cells exposed=20 to extreme hypoxia. Anticancer Res. 15, 683-686. = [Medline]=20

Kwiatkowski,=20 D.J. (2003). Tuberous = sclerosis: from=20 tubers to mTOR. Ann. Hum. Genet. 67, 87-96. = [Medline]=20

Lang, = K.J., Kappel,=20 A., and Goodall, G.J. (2002). Hypoxia-inducible factor-1alpha mRNA = contains an=20 internal ribosome entry site that allows efficient translation = during=20 normoxia and hypoxia. Mol. Biol. Cell 13, 1792-1801. = [Medline]=20

Li, Y., = Corradetti,=20 M.N., Inoki, K., and Guan, K.L. (2004). TSC2: filling the GAP in the mTOR signaling=20 pathway. Trends Biochem. Sci. 29, = 32-38. [Medline]=20

Li, Y., = Inoki, K.,=20 and Guan, K.L. (2004). Biochemical and=20 functional characterizations of small GTPase Rheb and TSC2 GAP=20 activity. Mol. Cell. Biol. 24, = 7965-7975. [Medline]=20

Liu, L. = and Simon, M.C. (2004). Regulation of transcription and translation = by=20 hypoxia. Cancer Biol. Ther. 3, 492-497. = [Medline]=20

Loreni, = F., Thomas,=20 G., and Amaldi, F. (2000). Transcription=20 inhibitors stimulate translation of 5=E2=80=B2 TOP mRNAs through = activation of S6=20 kinase and the mTOR/FRAP signalling pathway. Eur. J. Biochem. 267,=20 6594-6601. [Medline]=20

Ma, L., = Chen, Z.,=20 Erdjument-Bromage, H., Tempst, P., and Pandolfi, P.P. = (2005). Phosphorylation and functional inactivation = of TSC2 by=20 Erk implications for tuberous sclerosis and cancer = pathogenesis.=20 Cell 121, 179-193. [Medline]=20 [Summary]=20 [Full=20 Text]

Manning, B.D. and=20 Cantley, L.C. (2003). Rheb = fills a GAP=20 between TSC and TOR. Trends Biochem. Sci. 28, = 573-576.=20 [Medline]=20

Mu, J., = Brozinick,=20 J.T., Valladares, O., Bucan, M., and Birnbaum, M.J. (2001). = A role for AMP-activated protein kinase in=20 contraction- and hypoxia-regulated glucose transport in skeletal=20 muscle. Mol. Cell 7, 1085-1094. = [Medline]=20 [Summary]=20 [Full=20 Text]

Patel, = J., McLeod,=20 L.E., Vries, R.G., Flynn, A., Wang, X., and Proud, C.G. = (2002).=20 Cellular stresses profoundly inhibit = protein=20 synthesis and modulate the states of phosphorylation of multiple=20 translation factors. Eur. J. Biochem. 269, = 3076-3085. [Medline]=20

Potter, = C.J.,=20 Huang, H., and Xu, T. (2001). Drosophila=20 Tsc1 functions with Tsc2 to antagonize insulin signaling in = regulating=20 cell growth, cell proliferation, and organ size. = Cell 105, 357-368.=20 [Medline]=20 [Summary]=20 [Full=20 Text]

Reiling, J.H. and=20 Hafen, E. (2004). The = hypoxia-induced=20 paralogs Scylla and Charybdis inhibit growth by down-regulating = S6K=20 activity upstream of TSC in Drosophila. Genes Dev. 18,=20 2879-2892. [Medline]=20

Schmaltz, C.,=20 Hardenbergh, P.H., Wells, A., and Fisher, D.E. (1998). = Regulation of proliferation-survival = decisions during=20 tumor cell hypoxia. Mol. Cell. Biol. 18, = 2845-2854. [Medline]=20

Schmelzle, T. and=20 Hall, M.N. (2000). TOR, a = central=20 controller of cell growth. Cell 103, 253-262. [Medline]=20 [Full=20 Text]

Schwarzer, R.,=20 Tondera, D., Arnold, W., Giese, K., Klippel, A., and Kaufmann, = J.=20 (2005). REDD1 integrates = hypoxia-mediated=20 survival signaling downstream of phosphatidylinositol = 3-kinase.=20 Oncogene 24, 1138-1149. = [Medline]=20

Tapon, = N., Ito, N.,=20 Dickson, B.J., Treisman, J.E., and Hariharan, I.K. (2001). = The Drosophila tuberous sclerosis complex = gene=20 homologs restrict cell growth and cell proliferation. = Cell 105, 345-355.=20 [Medline]=20 [Summary]=20 [Full=20 Text]

Tee, = A.R. and=20 Blenis, J. (2005). mTOR, = translational=20 control and human disease. Semin. Cell Dev. Biol. 16, = 29-37.=20 [Medline]=20

Wouters, B.G., van=20 den Beucken, T., Magagnin, M.G., Koritzinsky, M., Fels, D., and = Koumenis,=20 C. (2005). Control of the = hypoxic=20 response through regulation of mRNA translation. = Semin. Cell Dev. Biol. 16, 487-501. [Medline]=20

Yeung, = R.S.=20 (2004). Lessons from the Eker rat = model: from=20 cage to bedside. Curr. Mol. Med. 4, 799-806. = [Medline]=20

Article = History

Received: July 28, = 2005
Revised:=20 November 1, 2005
Accepted: January 6, 2006
Published: = February 16,=20 2006

TOR=20 Signaling in Growth and Metabolism...
TSC2=20 Mediates Cellular Energy Response to Control ...
Phosphorylation=20 and Functional Inactivation of TSC...

More=20 related articles...

Article Options

= Table=20 of Contents
Summary
PDF=20 (679K)
Supplemen= tal=20 Data

Search = Medline=20 for

Liping= =20 Liu
M. = Celeste Simon

Cell		Immunity		Neuron		Molecular Cell		Structure
Current=20 Biology		Developmental = Cell		Cancer = Cell		Chemistry &=20 Biology		Cell=20 Metabolism

Article

Hypoxia-Induced Energy = Stress=20 Regulates mRNA Translation and Cell Growth

Liping Liu,1,2=20 Timothy=20 P. Cash,2=20 Russell=20 G. Jones,2=20 Brian Keith,2=20 Craig=20 B. Thompson,2=20 and M.=20 Celeste Simon1,2,

Hypoxia Inhibits Multiple Signaling = Pathways=20 Regulating mRNA Translation

Both mTOR and PERK Participate in = Hypoxic=20 Inhibition of Protein Synthesis

Hypoxic Inhibition of mTOR Can Occur = Independent of HIF

Hypoxia Affects Cellular = Bioenergetics and=20 Activates AMPK

AMPK Inhibits mTOR and eEF2 during=20 Hypoxia

TSC2-Dependent and TSC2-Independent = Effects of=20 Hypoxia on mTOR

TSC2 Contributes to Hypoxic = Inhibition of mRNA=20 Translation

Rheb Regulates Hypoxic mTOR=20 Inhibition

Hypoxic Regulation of Cell Growth = and=20 Proliferation Requires TSC2

Materials

Cell Culture, Transfection, and = Clone=20 Selection

Protein Synthesis by 35S-Met = Incorporation

ATP and ATP/ADP = Measurements

BrdU Labeling

Statistical Analysis

Liping Liu,^¹,²=20Timothy=20 P. Cash,^²=20Russell=20 G. Jones,^²=20Brian Keith,^²=20Craig=20 B. Thompson,^²=20and M.=20 Celeste Simon^¹,²,