Autophagy is a natural process of ‘self-eating’ that occurs within cells and can be either pro-survival or can cause cell death. As a pro-survival mechanism, autophagy obtains energy by recycling cellular components such as macromolecules or organelles. In response to nutrient deprivation, e.g. depletion of amino acids or serum, autophagy is induced and most of these signals converge on the kinase mTOR (mammalian target of rapamycin). It is commonly accepted that glucose inhibits autophagy, since its deprivation from cells cultured in full medium induces autophagy by a mechanism involving AMPK (AMP-activated protein kinase), mTOR and Ulk1. However, we show in the present study that under starvation conditions addition of glucose produces the opposite effect. Specifically, the results of the present study demonstrate that the presence of glucose induces an increase in the levels of LC3 (microtubule-associated protein 1 light chain)-II, in the number and volume density of autophagic vacuoles and in protein degradation by autophagy. Addition of glucose also increases intracellular ATP, which is in turn necessary for the induction of autophagy because the glycolysis inhibitor oxamate inhibits it, and there is also a good correlation between LC3-II and ATP levels. Moreover, we also show that, surprisingly, the induction of autophagy by glucose is independent of AMPK and mTOR and mainly relies on p38 MAPK (mitogen-activated protein kinase).
Equilibrium between anabolic and catabolic processes preserves the homoeostasis necessary to ensure normal cell growth and development. Upon nutrient deprivation or cellular stress, catabolic processes prevail in order to supply the energy required to maintain cell functions. One of these catabolic processes is macroautophagy, hereafter referred to as autophagy (reviewed in ). Autophagy is a highly conserved process by which the cells degrade their cytoplasmic material, including organelles and protein aggregates, in the lysosomes. In the last decades, more than 30 different ATGs (autophagy-related genes) have been discovered to play a role in this process . Autophagy starts with the formation of a double-membrane structure, called the phagophore, which sequesters large portions of the cytoplasm. This structure closes to become an autophagosome, which will ultimately fuse with late endosomes/lysosomes to form the autolysosome, where cargo is degraded by lysosomal hydrolases.
Autophagy has been implicated in many processes, including differentiation, growth, development and survival of cells, as well as in innate and adaptive immunity. Malfunction of autophagy is associated with aging and with several pathological conditions, including cancer, neurodegeneration and microbial infections, among others (reviewed in ). Depending on the cellular context, autophagy can promote either cell survival or death. Thus a full understanding of the signalling pathways that regulate autophagy will allow the development of new therapies to treat diseases in which this process is implicated.
Regulation of autophagy has been extensively studied, but there are still many unknowns . Under basal conditions, autophagy occurs at low levels in almost all cells, but it is strongly induced under stress conditions such as starvation. One of the main regulators of autophagy is mTOR (mammalian target of rapamycin), which blocks the function of Ulk1 (the mammalian homologue of Atg1), thus inhibiting the formation of the Ulk1–Atg13–FIP200 complex needed to initiate autophagosome formation . Many signalling pathways converge on mTOR, and an important regulator of this kinase is AMPK (AMP-activated protein kinase). AMPK is activated by an increased AMP/ATP ratio, leading to inactivation of mTOR via the tuberous sclerosis complex protein 2 and the GTP-binding protein Rheb and, subsequently, to the induction of autophagy in order to restore the energy levels . It has been established that AMPK competes with mTOR to phosphorylate Ulk1 [5,7]. AMPK phosphorylates Ulk1 at Ser317 and Ser777, forming an active complex with AMPK, Atg13 and FIP200, thus inducing autophagy . However, when nutrients are available, mTOR phosphorylates Ulk1 at Ser757 preventing the formation of this complex and, therefore, autophagy. In addition, there are many other signalling pathways that also regulate autophagy, and some of them may even act independently of mTOR [3,8,9].
Autophagy is highly sensitive to the levels of nutrients, hormones, and growth factors and cytokines, among which amino acids and insulin have been the most extensively studied (see e.g. ). The role of glucose has also been investigated and, although there is still controversy, it is generally assumed that glucose inhibits autophagy in mammalian cells and activates it in yeast (reviewed in ). Even though the availability of nutrients usually inhibits autophagy, an activation of autophagy by glucose also appears logical, because it is the main source of ATP production and many autophagic processes, such as cytoplasmic sequestration, lysosomal acidification and vacuolar H+-ATPase assembly, require ATP [12,13]. Therefore, in the present study, we have re-evaluated the regulation of autophagy by glucose in mammalian cells under starvation conditions. In contrast with reports indicating that autophagy is inhibited by glucose, we found that glucose induces autophagy. Moreover, we describe that this induction is regulated by p38 MAPK (mitogen-activated protein kinase) independently of AMPK and mTOR.
Reagents and antibodies
DMEM (Dulbecco's modified Eagle's medium), MEM (minimum essential medium), FBS (fetal bovine serum), penicillin and streptomycin were purchased from Gibco (Invitrogen). The inhibitors SB203580, PD98059, MK (MAPK-activated protein kinase) 2 inhibitor, KT5720, rapamycin and LY294002 were from Calbiochem. Ammonium chloride, bafilomycin A1, sodium oxamate and 3-methyladenine were obtained from Sigma–Aldrich. Lysotracker Red (Molecular Probes), leupeptin (Peptide Institute), BIRB796 (Selleck Chemicals), DAPI (4′,6-diamidino-2-phenylindole) nucleic acid stain (Gibco, Invitrogen) and D(+)-glucose (Merck) were purchased from the indicated sources. Antibodies against P (phosphorylated)-p38 MAPK (Thr180/Tyr182), P-AMPK (Thr172), AMPK, P-ACC (acetyl-CoA carboxylase) (Ser79), ACC, P-mTOR (Ser2448), mTOR, P-4E-BP1 (eukaryotic initiation factor 4E-binding protein 1) (Thr37/46), 4E-BP1, P-p70S6K (p70 S6 kinase) (Thr389), p70S6K, P-MKK (MAPK kinase) 3/6 (Ser189/207), MKK3/6, P-MKK4 (Ser257/Thr261), MKK4, P-CREB (cAMP-response-element-binding protein) (Ser133), CREB, P-ATF-2 (activating transcription factor-2) (Thr71), ATF-2, P-Ulk1 (Ser757) and Ulk1 were obtained from Cell Signaling Technology. Antibodies against p38 MAPK, LC3 (microtubule-associated protein 1 light chain 3), tubulin and actin were purchased from Santa Cruz Biotechnology, NanoTools, Abcam and Sigma–Aldrich respectively. The horseradish-peroxidase- and Alexa Fluor® 488-conjugated secondary antibodies were from Sigma–Aldrich and Gibco (Invitrogen) respectively.
Cell culture and RNAi (RNA interference)
The NIH 3T3 cell line was obtained from ECACC (European Collection of Animal Cell Cultures). MEFs (mouse embryonic fibroblasts) with p38 MAPK (p38−/−) knocked out and wild-type controls (p38+/+) were kindly provided by Dr A. Nebreda (IRB Barcelona, Barcelona, Spain) and have been described previously . HeLa and HEK (human embryonic kidney)-293 cell lines were also obtained from ECACC. Cells were cultured in DMEM (NIH 3T3 cells and MEFs) or MEM (HeLa and HEK-293 cells), supplemented with 10% FBS, 100 units/ml penicillin and 100 mg/ml streptomycin, at 37°C in a humidified 5% CO2 atmosphere. To evaluate the effects of glucose, cells were washed with PBS (pH 7.4) and incubated for 30 min in KH (Krebs–Henseleit) medium without glucose [118.4 mM NaCl, 4.75 mM KCl, 1.19 mM KH2PO4, 2.54 mM MgSO4, 2.44 mM CaCl2·2H2O and 18.5 mM NaHCO3, containing 10 mM Hepes (pH 7.4)]. Then, D(+)-glucose was added as indicated.
For inhibition of the expression of p38α MAPK, NIH 3T3 cells were plated in medium without antibiotics and 1 day later they were transfected with siRNA (small interfering RNA) duplexes targeting p38α MAPK or with non-targeting siRNAs, using DharmaFECT1 reagent (all from Dharmacon Thermo Fisher Scientific) according to the manufacturer's instructions. All siRNAs were tested and used at a final concentration of 100 nM. After 48 h, cells were replated and experiments were performed the following day.
Measurement of intracellular protein degradation
Intracellular protein degradation was measured as described previously . Briefly, NIH 3T3, or p38−/− and p38+/+ MEFs were incubated for 24 h in fresh full medium with 2 μCi/ml [3H]valine. After washing, the cells were incubated for 24 h in fresh full medium containing 10 mM L-valine to eliminate short-lived proteins and to prevent reutilization of [3H]valine. Cells were then washed with PBS and incubated for 30 min in KH medium without glucose, containing 10 mM L-valine. Then, 10 mM D(+)-glucose was added and the cells were incubated for an additional period of 2 h. Protein degradation was calculated by measuring the net release of trichloroacetic-acid-soluble radioactivity into the culture medium for 4 h. The contribution of macroautophagy was calculated with 10 mM 3-methyladenine as described previously .
Measurement of intracellular ATP
ATP was measured using the ATP bioluminescence assay kit HS II (Roche Applied Science). Briefly, cells were incubated in KH medium with different concentrations of glucose (1–10 mM) and sodium oxamate (5–100 mM) as indicated. After 2 h, intracellular ATP content was measured following the manufacturer's instructions, using a Spectra Max M5 microplate reader (Molecular Devices). Luminescence intensity was normalized to the protein present in each sample.
Cells plated on coverslips were incubated in KH medium in the presence or absence of 10 mM D(+)-glucose for 4 h. Cells were then rinsed twice with PBS, fixed in 4% paraformaldehyde/PBS for 20 min at room temperature (24°C), washed three times with PBS and quenched with 75 mM ammonium chloride and 20 mM glycine for 10 min. After cell permeabilization with 0.2% Triton X-100, cells were blocked overnight at 4°C in 1% BSA/0.05% Triton X-100. Cells were then incubated with an antibody against P-p38 (1:200) at 37°C for 2 h. Coverslips were washed four times with PBS and incubated with Alexa Fluor® 488-conjugated anti-mouse antibody (1:400 dilution) for 1 h at room temperature. After extensively washing, the coverslips were incubated with DAPI (1 μg/ml) for 5 min at room temperature, washed again and mounted using Fluor-Save reagent (Calbiochem). Images were obtained with an Apotome-equipped Axio Observer Z1 (Carl Zeiss).
Measurement of lysosomal content
NIH 3T3 cells were seeded in 24-well plates and incubated in fresh full medium 24 h before the experiment. Cells were then incubated for 2 h in KH medium, with or without 10 mM D(+)-glucose, and 50 nM Lysotracker Red was added for the last 20 min. Next, cells were washed and fixed with 4% paraformaldehyde/PBS. After extensive washing, Lysotracker Red signal was measured (excitation wavelength 544 nm, emission wavelength 590 nm) using a Spectra Max M5 microplate reader (Molecular Devices). Then, cells were incubated with DAPI (300 nM) for 5 min, washed and the DAPI signal was measured (excitation wavelength 358 nm, emission wavelength 461 nm). Lysosomal content is represented as Lysotracker Red fluorescence referred to DAPI fluorescence.
Electron microscopy and morphometric analyses
NIH 3T3 cells were incubated in KH medium in the presence or absence of 10 mM glucose and prepared for electron microscopy as described previously . Immature (Avi) and mature (Avd) autophagic vacuoles were identified in 200 randomly selected electron micrographs from four different experiments and their mean number per cell was calculated. Also, their fractional volume (volume density) was estimated using a multipurpose test system as described previously .
For immunoblotting, cells were collected and lysed in mammalian lysis buffer [150 mM NaCl, 1% Nonidet P40, 50 mM Tris/HCl, 0.5% sodium deoxycholate, 0.1% SDS, 15% glycerol, 1 mM PMSF, 100 μM leupeptin, 2 mM sodium orthovanadate, 100 mM NaF and 20 mM Na4P2O7 (pH 8.0)] or in RIPA buffer [150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris/HCl (pH 8.0), containing 1 mM PMSF and 100 μM leupeptin] at 4°C. Cellular suspensions were frozen–thawed at least eight times and centrifuged at 10000 g for 10 min at 4°C. Supernatants were collected and protein concentration was determined using the BCA (bicinchoninic acid) Protein Assay Kit (Thermo Scientific) according to the manufacturer's instructions. Samples were analysed by SDS/PAGE (10–16% gels) and proteins were transferred on to PVDF membranes (Millipore). Membranes were blocked with 5% (w/v) non-fat dried skimmed milk and incubated overnight with the corresponding primary antibodies. Then membranes were washed with TBS [Tris-buffered saline; 10 mM Tris/HCl and 150 mM NaCl (pH 7.4)] and probed with secondary antibodies for 1 h. Signals were visualized using Lumi-Light Western Blotting Substrate (Roche Applied Science) or ECL (enhanced chemiluminescence) Prime Western Blotting Detection Reagent (GE Healthcare) on Medical X-Ray Film (Kodak). Comparisons between different conditions, after calculating mean and S.D. values, were by Student's t test. Differences were considered significant at *P<0.05, **P<0.01 and ***P<0.005.
Glucose induces autophagosome maturation and proteolysis under starvation conditions
To investigate the regulation of autophagy by glucose, we incubated NIH 3T3 cells in KH medium with or without glucose. We first monitored the lipidation of LC3 (LC3-II) as a marker of autophagosome formation and, as shown in Figure 1(A), lanes 1 and 2, glucose increases LC3-II levels. This is compatible with an enhanced autophagy, although it could also reflect a reduced turnover of autophagosomes . To rule out this possibility, we performed the same experiment in the presence of the lysosomal inhibitors ammonium chloride and leupeptin. Accumulation of LC3-II confirms an increased autophagic flux in the presence of glucose. Thus, in the presence of lysosomal inhibitors, a 2 h glucose (10 mM) treatment increases by 260±25% (means±S.D. for three different experiments) the LC3-II levels (Figure 1A, lanes 3 and 4). Similar results were obtained with 500 nM bafilomycin A1 (results not shown). These results, however, are in contrast with previous studies that showed an induction of autophagy under deprivation of glucose [8,18–20].
Glucose induces autophagy in NIH 3T3 cells under starvation conditions
To investigate whether these effects could also be observed under other culture conditions, we also performed the same experiments using DMEM, which contains vitamins and amino acids. As shown in Supplementary Figure S1(A) (at http://www.biochemj.org/bj/449/bj4490497add.htm), addition of glucose increases autophagy when NIH 3T3 cells are incubated in DMEM, as well as when they are incubated in KH medium. This result was also obtained with other cell lines: LC3-II levels increased in HEK-293 and HeLa cells, and in MEFs in response to the addition of glucose, when incubated in KH medium or in DMEM (Supplementary Figures S1B–S1D). Taken together, these data confirm an inducing effect of glucose on autophagy under various starvation conditions.
An increased autophagic flux from autophagosomes to autolysosomes should increase the lysosomal mass in the cell. To verify this we used the fluorescent probe Lysotracker Red that selectively accumulates in organelles with low internal pH. When NIH 3T3 cells were incubated in KH medium with 10 mM glucose, the fluorescence produced by this probe increased almost 2-fold (Figure 1B). To confirm that the degradation of the autophagic cargo was taking place under these conditions, we measured, in pulse–chase experiments, the degradation of long-lived proteins (total degradation and degradation by macroautophagy) (Figure 1C). Glucose produced an almost 2-fold increase in both total and autophagic proteolysis. These results were also confirmed by electron microscopy. In the presence of 10 mM glucose, both immature (autophagosomes, not yet fused with lysosomes; Avi) and mature (autolysosomes, resulting from the fusion of autophagosomes with lysosomes; Avd) autophagic vacuoles increased (Figure 1D). Quantification of both the number of Avi and Avd per cell and the volume density of each, depicted respectively in Figures 1(E) and 1(F), showed again an approximately 2-fold increase induced by glucose, in good agreement with the data shown previously (Figures 1A–1C). Taken together, all of these results demonstrate that glucose induces autophagy. It should be noted that the increase in total proteolysis detected after the addition of glucose (see Figure 1C) is not only explained by an increase in autophagy, as we also observed an augmented ubiquitination and chymotrypsin-like proteasomal activity (Supplementary Figures S2A and S2B respectively at http://www.biochemj.org/bj/449/bj4490497add.htm). However, these changes appear to be less important than the autophagy increase.
Induction of autophagy by glucose depends on the ATP generated by its catabolism
Autophagy requires ATP, at least in the sequestration step and for the operation of the lysosomal proton pump [12,13]. To explore whether the induction of autophagy by glucose depends on the ATP produced by its catabolism, we first compared ATP and LC3-II levels in cells incubated in KH medium with increasing concentrations of glucose. As expected, ATP levels increased after the addition of glucose in a dose-dependent manner (Figure 2A). With regard to LC3-II levels, there was a clear rise between 0.5 and 1 mM of glucose, without further increase at higher concentrations up to 10 mM (Figure 2B).
Glucose-dependent autophagosome formation depends on ATP availability generated by glucose catabolism
To confirm that production of ATP from glucose is indeed responsible for the observed induction of autophagy, we analysed the levels of LC3-II in the presence of glucose and sodium oxamate, an inhibitor of lactate dehydrogenase. Blocking lactate dehydrogenase prevents the conversion of NADH into NAD+, the limiting factor in glycolysis. Since oxamate is an inhibitor of aerobic and anaerobic glycolysis, it decreased ATP levels (Figure 2C) and this correlated with a drop in autophagosome formation (Figure 2D). All of these data taken together indicate that induction of autophagy by glucose derives from the ATP produced by its catabolism.
Glucose induces autophagy independently of AMPK and mTOR
Glucose deprivation increases the AMP/ATP ratio, which leads to AMPK activation . This protein, once activated, increases autophagy by inhibiting mTOR through tuberous sclerosis complex protein 2 or by competition with mTOR [5,7]. Because glucose is a nutrient and an immediate source of ATP, we anticipated that AMPK and mTOR could play a role in the induction of autophagy. As expected, addition of glucose decreased the phosphorylation of AMPK and, consequently, of its substrate ACC (Figure 3A, left-hand panels). Consistent with this, we observed an increase in the phosphorylation of mTOR and in one of its targets, 4E-BP1, suggesting that, after the addition of glucose, this kinase is active (Figure 3A, middle panels). However, no phosphorylation could be detected in KH medium with or without glucose of other two downstream targets of mTOR, p70S6K and Ulk1 (Figure 3A, right-hand panels), indicating that this kinase may not be fully active to suppress autophagy. Therefore, and to clarify whether mTOR participates in the induction of autophagy produced by glucose, NIH 3T3 cells were treated for 2 h with rapamycin, a specific inhibitor of mTOR, in the presence or absence of glucose, and the levels of LC3-II were determined by Western blot analysis. Figure 3(B) shows that inhibition of mTOR does not affect LC3-II levels in the presence of glucose. Therefore mTOR or AMPK activities do not explain the induction of autophagy mediated by glucose, suggesting that another mechanism(s) might be involved.
Induction of autophagy by glucose is mTOR-independent
Glucose activates p38 MAPK
To understand the mechanism by which glucose promotes autophagy, we analysed the possible role of other kinases known to have an effect in this process. To this end, we used chemical inhibitors targeting protein kinase A (KT5720), ERK1/2 (extracellular-signal-regulated kinase 1/2) (PD98059) and p38 MAPK (SB203580). The inhibition of PI3K (phosphoinositide 3-kinase) class III, a protein necessary for autophagy to occur, was used as a positive control (LY294002). Inhibition of protein kinase A did not change LC3-II levels when compared with vehicle alone. We observed a slight decrease in autophagy in the presence of the ERK1/2 inhibitor PD98059, although the glucose-dependent increase in LC3-II levels (3.9-fold) is not affected. In contrast, p38 MAPK and, as expected, PI3K inhibition, produced a decrease in LC3-II accumulation (1.9- and 1.3-fold respectively) when glucose was added, suggesting that p38 MAPK could have an important role in the induction of autophagy by glucose (Figure 3C).
To further address this hypothesis, we investigated whether glucose regulates p38 MAPK activity. We first monitored the phosphorylation of p38 MAPK and its downstream targets. As observed in Figure 4(A), addition of glucose promotes the phosphorylation of p38 MAPK. Once activated, p38 MAPK can directly phosphorylate ATF-2 and, via other kinases [e.g. MK2/3 and MSK (mitogen- and stress-activated kinase) 1 and 2], CREB . Figure 4(A) shows that addition of glucose increases the phosphorylation of CREB and ATF-2, reflecting p38 MAPK activation. However, we do not observe phosphorylation of the kinases MKK3/6 or MKK4 that phosphorylate p38 MAPK, suggesting that other kinases play this role.
Induction of autophagy by glucose is p38 MAPK-dependent
It has been described that, under some conditions such as UV or X-ray exposure, p38 MAPK translocates to the nucleus upon its activation . We analysed whether this was also the case in our experimental model. Figure 4(B) shows representative images of NIH 3T3 cells incubated in KH medium in the absence (top panels) or presence (bottom panels) of glucose (10 mM). From these fluorescence experiments it appears clear that glucose promotes the translocation of p38 MAPK to the nucleus. Therefore taking into consideration all of the experiments described above, it can be concluded that glucose induces autophagy and also produces the activation of p38 MAPK.
Induction of autophagy by glucose is mediated by p38 MAPK
There are many reports on the effect of p38 MAPK in autophagy. However, depending on the stimulus and the cell type, p38 MAPK has been described to act either as a positive or as a negative regulator of this process [23–26]. In the present study, we have observed that p38 MAPK phosphorylation is induced by glucose under starvation conditions (Figure 4A) and that inhibition of p38 MAPK by SB203580 prevents the accumulation of LC3-II (Figure 3C), suggesting that p38 MAPK is necessary for the activation of autophagy by glucose. However, it has been recently established that SB203580 and the structurally related SB202190 may induce autophagic vacuole formation due to off-target effects [27,28]. Therefore we tested BIRB796, an inhibitor structurally unrelated to SB203580 that blocks the ATP-binding site of p38 MAPK, and an inhibitor of the kinase MK2, which is a downstream kinase of p38 MAPK. NIH 3T3 cells were pretreated with the inhibitors and then incubated under starvation conditions with or without 10 mM glucose for 2 h. Lipidation of LC3 was analysed by Western blotting. As shown in Figure 4(C), SB203580 (left-hand panels), BIRB796 (middle panels) and the MK2 inhibitor (right-hand panels) repressed the accumulation of LC3-II induced by glucose. These data confirm the results obtained previously, indicating that p38 MAPK has a stimulatory role in the induction of autophagy by glucose.
To further establish this role for p38 MAPK, we analysed the lipidation of LC3 in knockdown experiments. Knockdown of p38 MAPK produced a 2.8-fold decrease in the formation of autophagosomes induced by the addition of glucose (Figure 5A). This was further corroborated by the use of p38−/− MEFs, in which the addition of glucose under starvation conditions did not produce any accumulation of LC3-II compared with p38+/+ MEFs (Figure 5B). All of these data confirm that p38 MAPK is involved in the formation of autophagosomes.
p38 MAPK positively regulates autophagosome formation
Finally, and to further ascertain that p38 MAPK is involved in the induction of autophagy mediated by glucose, we analysed the degradation of long-lived proteins in pulse–chase experiments. After labelling of the proteins, wild-type and p38−/− MEFs were incubated in KH medium with or without 10 mM glucose for 4 h. Figure 5(C) shows that the increase in protein degradation induced by glucose is completely abolished in p38−/− MEFs.
In conclusion, the results of the present study strongly support that, under starvation conditions, the addition of glucose induces autophagy and that this effect is mediated by p38 MAPK.
Understanding the mechanisms involved in the regulation of autophagy is a central question in cellular biology. Regulation by insulin and amino acids has been extensively studied, but, comparatively, less attention has been paid to other factors such as glucose. Two decades ago, studies in yeast concluded that glucose induces protein degradation (see, for example, ). Conversely, studies in mammalian cells demonstrated that glucose deprivation induces autophagy, thus suggesting an inhibitory role for glucose [8,19]; however, there is still controversy about the role of glucose on autophagy. Therefore we decided to reinvestigate the effect of glucose on autophagy under starvation conditions. The results of the present study clearly demonstrate that in mammalian cells glucose induces autophagy.
How can this discrepancy with previous reports be explained? Most experiments showing that deprivation of glucose induces autophagy and, therefore, that glucose inhibits autophagy, were carried out in cells incubated in culture medium in the presence of serum, and/or using low concentrations of glucose or hyperglycaemic conditions. Serum, vitamins and amino acids found in the media affect autophagy themselves . Thus, to study the specific role of glucose, it is important to suppress any other factor and, therefore, we performed our experiments incubating the cells in a serum-free medium with or without glucose, using the nutrient-free medium KH or DMEM which contains vitamins and amino acids. To avoid side effects, all measurements were obtained between 2 and 4 h after deprivation of glucose.
We observed that addition of glucose quickly induced a significant increase in the formation of autophagosomes and autolysosomes as well as an increase in protein degradation by autophagy. In agreement with the results of the present study, glucose promotes cytoplasmic sequestration in yeast, as well as the assembly of the vacuolar-ATPase that is necessary for the acidification of the vacuole and, therefore, for the activity of the hydrolases . In addition, Ravikumar et al.  showed in mammalian cells that a high glucose concentration reduced aggregate accumulation and enhanced the clearance of mutant huntingtin by autophagy, probably following a signalling pathway mediated by the inhibition of mTOR. Moreover, trehalose, a disaccharide formed by two α-glucose units, also induces autophagy, leading to an enhanced clearance of mutant huntingtin and α-synuclein aggregates [33,34].
As autophagy consumes ATP , we investigated whether ATP production through glycolysis could be behind the observed increase in autophagy. The results of the present study indicate that autophagy requires a minimum amount of ATP, provided by 0.5–1 mM glucose, and that a further increase in ATP did not additionally increase autophagy. A reduction in ATP content in the cell is sensed by AMPK, which becomes activated and inhibits the mTOR kinase, leading to an induction of autophagy . In our experiments, we observe a correlation between the addition of glucose with the subsequent increase in ATP levels and inactivation of AMPK. Although we observe phosphorylation of the mTOR target 4E-BP1, which would suggest that glucose produces a decrease in autophagy, we do not observe phosphorylation of the targets p70S6K and Ulk1. It is important to note that these experiments were performed in KH medium, in which mTOR is normally inhibited. In fact, rapamycin does not alter the effect of glucose deprivation. Therefore other mechanisms might be involved. In this regard, we have identified p38 MAPK as an important effector in the induction of autophagy by glucose. In addition, ERK1/2 could also be a possible candidate in this regulation, but to a lesser extent (see Figure 3C). The p38 MAPK as well as its substrates ATF-2 and CREB are phosphorylated after the addition of glucose. Since p38 MAPK also translocates to the nucleus after glucose treatment, it is tempting to hypothesize that this activates some transcription factors, e.g. ATF-2 and CREB, that will drive the expression of genes implicated in autophagy induction. We do not observe activation of the upstream kinases MKK3, MKK6 or MKK4. However, it has been shown that p38 MAPK can be phosphorylated by other kinases and, for example, it has been described in rat aortic smooth muscle cells that PKCδ (protein kinase Cδ) phosphorylates p38 MAPK after glucose treatment .
There exist conflicting data about the role of p38 MAPK in autophagy. Depending on the stimulus, it can be considered as either an inducer or an inhibitor of autophagy. For example, it has been described that LPS (lipopolysaccharide) and interferon-γ activate p38 MAPK inducing autophagy [24,26]. In contrast, Webber and Tooze  showed that in full medium p38 MAPK is phosphorylated and inhibits autophagy by sequestering p38IP (p38-interacting protein) and, therefore, inhibiting the formation of the p38IP–mAtg9 complex needed for the redistribution of mAtg9 to the autophagosomal membrane.
There are other data that support an inhibitory role in autopagy for p38 MAPK. Comes et al.  showed that inhibition of p38 MAPK leads to autophagic cell death in colorectal cancer cells. In the same way, inhibition of p38 MAPK has been shown to induce autophagy by increasing the expression of Beclin1 and LC3 . However, these and other studies [23,36,37] concluding that p38 MAPK inhibits autophagy were based on the use of the chemical inhibitors SB203580 or SB202190, which may produce off-target effects on other proteins such as PI3K/Akt or AMPK [27,28]. Therefore the role of p38 MAPK as an autophagy inhibitor must be revised. In the present study, by using unrelated inhibitors, we have shown that inhibition of p38 MAPK prevented the increase in LC3-II levels by glucose. In addition, knockdown of p38 MAPK inhibited the increase in LC3 lipidation. Furthermore, there is also no increase in p38−/− MEFs compared with p38+/+ MEFs, and this is accompanied by a decrease in total protein degradation. Therefore we conclude that glucose is an activator of autophagy under starvation conditions and that this activation is mediated by p38 MAPK.
activating transcription factor-2
AMP-activated protein kinase
Dulbecco's modified Eagle's medium
eukaryotic initiation factor 4E-binding protein 1
European Collection of Animal Cell Cultures
extracellular-signal-regulated kinase 1/2
fetal bovine serum
human embryonic kidney
microtubule-associated protein 1 light chain
mitogen-activated protein kinase
mouse embryonic fibroblast
minimum essential medium
MAPK-activated protein kinase
mammalian target of rapamycin
p70 S6 kinase
small interfering RNA
Jose Félix Moruno-Manchón, Eva Pérez-Jiménez and Erwin Knecht conceived and designed the experiments. Jose Félix Moruno-Manchón and Eva Pérez-Jiménez performed the experiments. Jose Félix Moruno-Manchón, Eva Pérez-Jiménez and Erwin Knecht analysed the data and wrote the paper.
We thank Asunción Montaner for technical assistance and Ángel Nebreda (IRB Barcelona, Barcelona, Spain) for providing the p38-deficient and wild-type MEFs.
This work was supported by the Ministerio de Educación y Ciencia [grant number BFU2011-22630], Generalitat Valenciana [grant number PROMETEO/2012/061] and Fundació Marató TV3 [grant number 100130].
These authors are joint senior authors on the paper.