Secondary Logo

Journal Logo

Basic Science Aspects

Adenosine 5’-Monophosphate Protects from Hypoxia by Lowering Mitochondrial Metabolism and Oxygen Demand

Kondo, Yutaka; Sueyoshi, Koichiro; Zhang, Jingping; Bao, Yi; Li, Xiaoou∗,†; Fakhari, Mahtab; Slubowski, Christian J.; Bahrami, Soheyl; Ledderose, Carola; Junger, Wolfgang G.∗,‡

Author Information
doi: 10.1097/SHK.0000000000001440



Ischemia and reperfusion injury are unresolved problems that cause major clinical complications, for example in cardiac arrest, stroke, and trauma patients (1). Permanent brain damage is a frequent consequence of these and similar conditions that deprive the brain of oxygen. Current treatment strategies to reduce brain damage in these patients are limited. One attractive approach is therapeutic hypothermia that slows metabolism and thereby reduces nerve cell damage (2). However, the equipment needed to induce hypothermia is complex and the rate of cooling is slow, which limits the utility of this method in critical care settings (3).

Recent studies have shown that injection of adenosine 5’-monophosphate (AMP) in mice can cause a rapid drop in core body temperature, resulting in a hypometabolic state similar to that seen with therapeutic hypothermia and in animals undergoing hibernation (4, 5). Importantly, AMP injection may also reduce ischemic brain injury in experimental mouse models (6). However, the underlying mechanism by which AMP exerts these remarkable physiological effects is not clear and the clinical utility of AMP, for example for the treatment of critical care patients, has yet to be determined. Oxygen deprivation damages primarily those organs that depend heavily on mitochondrial energy production. Mitochondria produce ATP through the tricarboxylic acid cycle (TCA) and the electron transfer chain (ETC) that depend on oxygen and carbon sources such as glucose, lipids, or amino acids to drive the oxidative phosphorylation process that takes place within mitochondria. In the absence of oxygen, mitochondrial ATP production stalls and mitochondria can become irreversibly damaged (7). Hypoxia can cause reverse electron transport and the buildup of TCA intermediates that severely damage mitochondrial proteins when oxygen supply is restored (8).

ATP is not only the main energy carrier that fuels intracellular signaling processes needed for all biological functions but ATP also acts as an extracellular messenger molecule for intercellular communication in paracrine or autocrine fashions (9, 10). AMP is a breakdown product of ATP that can be further converted to adenosine. ATP and its breakdown products ADP and adenosine can stimulate a number of different purinergic receptors that are present on the surface of virtually all mammalian cell types (9). The nineteen known purinergic receptor subtypes are divided into three groups, the P1 (adenosine), the P2Y, and the P2X receptor families. The P1 and P2Y receptors belong to the G protein-coupled receptor (GPCR) superfamily. All four P1 receptors recognize adenosine, while the eight P2Y receptor members can recognize ATP, ADP, UTP, and other nucleotides (11, 12). While all seven known P2X receptors are ATP-gated ion channels that facilitate Ca2+ influx (13), P1 and P2Y receptors can elicit a range of different downstream signal transduction events such as cAMP, Ca2+, and MAPKs signaling. Through these diverse downstream signaling pathways, purinergic receptors fine-tune mammalian cell functions (9). Mitochondria are a main source of the ATP that stimulates these purinergic receptors and their downstream signaling mechanisms (14, 15).

In the current study, we demonstrate that AMP can interfere with these purinergic feedback mechanisms. We found that AMP rapidly diminishes mitochondrial respiration in mouse and human neurons, resulting in a hypometabolic state that transiently bypasses the need for oxygen and thereby prolongs survival of mice under hypoxic conditions. Our findings suggest that the mechanisms by which AMP slows mitochondrial respiration are potential therapeutic targets to limit brain damage in critical care patients.



JC-1, Fluo4-AM, and SytoxGreen were purchased from Molecular Probes (Thermo Fisher Scientific, Waltham, Mass). 5-Aminoimidazole-4-carboxamide ribonucleotide (AICAR), CGS21680, carbenoxolone (CBX), and suramin were obtained from Tocris (R&D Systems, Minneapolis, Minn). Antibodies against human phospho-mTOR (Ser2448, D9C2 XP rabbit mAB), phospho-P70S6 kinase (Thr389, 108D2, rabbit mAB), phospho-AMPKα (Thr172, 40H9, rabbit mAB), and total p38α MAPK (polyclonal rabbit AB) were purchased from Cell Signaling Technology (Danvers, Mass). Horseradish peroxidase-conjugated secondary anti-rabbit IgG antibodies, adenosine, AMP, ATP, UTP, and all other reagents were from Sigma-Aldrich (St. Louis, Mo) unless otherwise stated.

Mice and AMP treatment

All animal experiments were approved by the Institutional Animal Care and Use Committee of the Beth Israel Deaconess Medical Center and performed in accordance with the guidelines of the National Institutes of Health. C57BL/6J mice were obtained from Charles River Laboratories (Wilmington, Mass). These animals were housed in groups of three to five mice per cage under standard conditions with 12-h light and dark cycles at an ambient temperature of 20°C to 25°C and unrestricted access to water and standard laboratory mouse diet. The mice were allowed to adjust to the new housing environment for at least 1 week. Adult male and female mice (8–10 weeks old) were used for all experiments. Animals were randomly assigned to control or treatment (AMP) groups. The experiments were performed in an unblinded manner. Mice received an intraperitoneal (i.p.) injection of AMP (0.5 mg/g body weight, corresponding to 14.4 μL/g body weight of a 100 mM AMP stock solution dissolved in normal saline) or an equivalent volume of normal saline. This dose of AMP was previously shown to cause a reliable drop in metabolic parameters in mice (4, 5, 16). The core body temperature was measured with a rectal temperature probe and a digital thermometer (Greisinger Electronic, Regenstauf, Germany). The heart rate was determined under general anesthesia with 1% to 2% isoflurane (Precision Vaporizer V-1 Table Top System, Vet Equip, Pleasanton, Calif) by recording ECGs with a Dual Bio Amplifier instrument (ADInstruments, Colorado Springs, Colo). Respiratory rate, heart rate, and core body temperature were assessed in different cohorts of mice. Some mice that received an AMP injection were sacrificed to collect blood samples in order to assess nucleotide levels at the indicated time points (1–180 min). Other animals were allowed to recover. These animals were returned to their cages and monitored for at least 24 h. Spontaneous activity, grooming behavior, body posture, gait, and reflexes were evaluated after 24 h to examine animals for possible signs of behavioral or neurological impairments following AMP treatment.

To study the effect of AMP on hypoxic tolerance, mice were anesthetized with ketamine (100 g/kg i.p.) and xylazine (10 mg/kg i.p.; Patterson Veterinary, Devens, Mass). AMP (4 mg/g) or an equivalent volume of normal saline was administered intraperitoneally 2 min before mice were placed into a glass chamber filled with a controlled gas mixture of nitrogen and 6% or <1% oxygen (Live Cell Instrument, Seoul, Korea). Mice were carefully observed and the experiment was terminated when respiration stopped for more than 6 s.

Cell culture

Mouse neuronal progenitors were isolated from the cortex of E13.5-E14 mouse embryos as previously described (17). Briefly, the embryo cortices were microscopically dissected, dissociated with 0.25% trypsin, and the cells were cultured in KnockOut DMEM/F-12 medium supplemented with StemPro NSC SFM, EGF, and FGF according to the manufacturer's specifications (Thermo Fisher Scientific). Neuronal differentiation was induced by culturing the cells for 7 days in Neurobasal medium supplemented with B-27 and GlutaMAX as recommended by the manufacturer (Thermo Fisher Scientific). SH-SY5Y cells were purchased from the American Type Culture Collection (ATCC, Manassas, Va) and cultured in a mixture of 1 part DMEM and 1 part F12 media supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin (Thermo Fisher Scientific). Cells were maintained at 37°C in a humidified atmosphere containing 5% CO2.

Mitochondrial and cytosolic Ca2+ imaging

Mitochondrial Ca2+ uptake was assessed in SH-SY5Y cells expressing the mitochondrial Ca2+ biosensor mito-CAR-GECO1 (plasmid 46022, Addgene, Cambridge, Mass) (18). Twenty-four hours prior to transfection, SH-SY5Y cells (4 × 104/well) were seeded into 8-well chambered coverglass dishes (Nunc LabTek, Thermo Fisher Scientific) coated with 40 μg/mL human fibronectin. Transfection was carried out in cell culture medium without antibiotics using 0.4 μg plasmid DNA per well, Opti-MEM I reduced serum media (Thermo Fisher Scientific), and Lipofectamine 2000 transfection reagent as per the manufacturer's (Thermo Fisher Scientific) instructions. The medium was replaced with fresh culture medium after 6 h and cells were cultured for another 24 h. For cytosolic Ca2+ imaging, SH-SY5Y cells or differentiated mouse neurons were allowed to attach overnight to fibronectin-coated glass-bottom dishes. Then cells were stained with Fluo-4 AM (4 μM) for 20 min. Fluorescence imaging was done in phenol-red free DMEM medium (Thermo Fisher Scientific) using a Leica DMI6000B microscope (Leica Microsystems, Wetzlar, Germany) equipped with a temperature-controlled (37°C) stage incubator (Ibidi, Fichburg, Wis) and a Leica DFC365 FX camera. Fluorescence images were captured at a frame rate of one frame per second through a ×40 objective (numerical aperture 0.75) using fluorescein isothiocyanate (FITC; Fluo-4) or tetramethylrhodamine (TRITC; mitochondrial Ca2+) filter sets and Leica LAS X microscope imaging software. Images were analyzed with ImageJ software (National Institutes of Health).

Mitochondrial membrane potential

SH-SY5Y cells or primary mouse neurons were allowed to attach to fibronectin-coated coverglass dishes and then stained with the mitochondrial membrane potential probe JC-1 (2.5 μg/mL) for 20 min at 37°C. Imaging was done in phenol-red free DMEM medium. Cells were treated with AMP (1–20 mM), carbonyl cyanide m-chlorophenyl hydrazone (CCCP) (10 μM), or adenosine (0.1–5 mM) as indicated and JC-1 red fluorescence was recorded at 1-s intervals with the fluorescence microscope system described above. Adenosine was used at lower maximum concentrations than AMP because AMP breakdown in vivo or in cell cultures does not result in equimolar adenosine levels.

Cell viability

SH-SY5Y cells were cultured in fibronectin-coated chambered coverglass dishes for 16 h in the presence or absence of AMP (10 mM) or H2O2 (2 mM). Then, cells were stained with SytoxGreen (1 μM) for 15 min at 37°C and the percentage of dead cells was determined by fluorescence and bright field microscopy.

Resazurin assay

Metabolic activity of SH-SY5Y cells was estimated by measuring the reduction of resazurin to resorufin. This reaction is considered to depend primarily on the activity of mitochondrial enzymes (19). SH-SY5Y cells (1 × 104 cells per well of a 96-well cell culture plate) were treated with different concentrations of AMP in the presence of 1 mM resazurin. The fluorescent conversion product resorufin was measured after 3 h with a SpectraMax M5 microplate reader (Molecular Devices, Silicon Valley, Calif) using 544 nm as excitation and 594 nm as emission wavelengths.

HPLC analysis of adenine compounds

ATP, ADP, AMP, and adenosine plasma concentrations were determined by high performance liquid chromatography (HPLC) using a Waters HPLC system (Waters, Milford, Mass). Mice under general isoflurane anesthesia were injected with AMP (0.5 mg/g i.p.) and blood was collected by cardiac puncture into chilled heparinized tubes at the indicated time ranging from 0 to 180 min after AMP injection. Sample preparation, etheno-derivatization of adenine compounds, and HPLC measurements were performed as previously described (15). To determine the effect of exogenous AMP on intracellular adenine compound levels, SH-SY5Y cells were cultured overnight in fibronectin-coated 24-well culture plates (5 × 104 cells per well) using complete culture medium. Thirty minutes before treatment with AMP, the medium was replaced with prewarmed DMEM. Reactions were stopped at the indicated time points after the addition of AMP by chilling samples on ice. Cell culture supernatants were discarded. The cells were washed three times with ice-cold HBSS, lysed on ice by sonication in 0.4 M perchloric acid, and ATP, ADP, AMP, and adenosine concentrations in cell lysates were determined by HPLC.

AMPK and mTOR activation

SH-SY5Y cells were cultured overnight in fibronectin-coated 96-well cell culture plates (7 × 104 cells per well) and then treated with AMP (10 mM). The reactions were stopped at the indicated times by placing cells on ice. Cells were washed with ice-cold HBSS, lysed in RIPA buffer (25 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS) containing protease (Sigma) and phosphatase (Thermo Fisher Scientific) inhibitors and sonicated on ice. Proteins were separated by SDS-PAGE using 12% trisglycine polyacrylamide gels (Thermo Fisher Scientific) and transferred to PVDF membranes (Millipore, Bedford, Mass). Immunoblotting was performed using phosphospecific antibodies to assess mTORC1, AMPK, and P70S6 kinase activation and antibodies for total p38α MAPK as a loading control.


Data are expressed as mean ± standard deviation (SD) if not indicated otherwise. Differences between two groups were tested for statistical significance with two-tailed unpaired Student t test. One-way ANOVA followed by Holm–Sidak test was used for multiple group comparisons. Kruskal–Wallis test followed by post-hoc Dunn test was used for not-normally distributed data. Survival was analyzed with the Kaplan–Meier method and the log rank test. Differences were considered statistically significant at P < 0.05.


AMP treatment transiently downregulates physiological functions in mice

Previous reports have shown that AMP accumulates in animals during torpor and that injection of AMP can induce a transient, torpor-like state of hypothermia in mice (4, 5). Consistent with these reports, we found that intraperitoneal injection of mice with AMP (0.5 mg/g body weight) resulted in severe hypothermia (Fig. 1A). AMP caused a profound drop in core body temperature that reached near ambient levels about 1 h after AMP injection. In addition, AMP injection decreased the respiratory rate and the heart rate by half in less than 10 min after AMP injection (Fig. 1, C and D). AMP also induced a state of decreased alertness that dramatically reduced the responsiveness of these animals to physical arousal (Supplemental Digital Content 1-Video 1, All mice recovered spontaneously within 3 h with body temperature, heart rate, and respiratory rate restored to baseline levels. In accordance with previous reports by others (4), we did not find any signs of behavioral impairments when we evaluated the mice clinically and neurologically 24 h after recovery from AMP injection, suggesting the absence of cerebral damage in AMP-treated mice.

Fig. 1
Fig. 1:
AMP treatment of mice slows physiological functions.

AMP rapidly lowers mitochondrial metabolism of neurons

Although the human brain represents only 2% of the total body mass, it accounts for about 20% of the total oxygen demand at rest (20). Therefore, the dramatic drop in respiratory rate following AMP injection suggests that AMP reduces oxygen demand and mitochondrial metabolism of neurons in the brain. However, the fact that mice recovered from AMP injection without any apparent signs of neurological impairment indicates that AMP reduces oxygen demand without negative consequences for the brain (4). In order to study the mechanisms by which AMP affects neurons, we exposed the human neuronal cell line SH-SY5Y to different concentrations of AMP and assessed the metabolic rate with the resazurin conversion assay (19). We found that treatment with AMP dose-dependently reduced the metabolic rate of these cells (Fig. 2A). However, AMP did not affect the viability of SH-SY5Y cells, even when used at high concentrations (10 mM) and for prolonged times (16 h; Supplemental Digital Content 2-Supplemental Fig. 1, Neurons, more than any other cell type, depend on mitochondria for ATP production (21). Therefore, we tested how AMP treatment affects mitochondrial activity of SH-SY5Y cells. Because of the rapid in vivo effect of AMP (Fig. 1), we focused in the following experiments on assessing the early responses of cells to AMP. We found that AMP treatment caused a rapid and dose-dependent decrease in the mitochondrial membrane potential (Δψm) of SH-SY5Y cells, reaching levels similar to those obtained with CCCP, which is a chemical inhibitor that uncouples oxidative phosphorylation (Fig. 2, B–D; Supplemental Digital Content 3-Video 2, AMP had a similarly profound effect on Δψm of primary neurons that were obtained by the differentiation of neural stem cells from mice (Fig. 2D). These results are consistent with the notion that AMP treatment can rapidly reduce oxidative phosphorylation in the mitochondria of neurons.

Fig. 2
Fig. 2:
AMP rapidly decreases mitochondrial metabolism of neurons.

P2 receptors and AMP have opposing effects on mitochondrial activity

The findings above demonstrate that exposure of neurons to AMP interferes with signaling processes that regulate mitochondrial metabolism. Mitochondrial respiration depends on Ca2+ uptake by mitochondria (22). Changes in mitochondrial Ca2+ uptake regulate Δψm, mitochondrial energy metabolism, and neuronal activity (23). We studied whether and how AMP treatment of neurons affects mitochondrial Ca2+ uptake in SH-SY5Y cells. For that purpose, we engineered SH-SY5Y cells that express a genetically encoded mitochondrial Ca2+ biosensor (18). Treatment of these cells with AMP caused a rapid and dose-dependent drop in mitochondrial Ca2+ levels that occurred within 5 s after AMP addition (Fig. 3, A and B; Supplemental Digital Content 4-Video 3, In contrast to AMP, addition of ATP dose-dependently increased Ca2+ uptake of mitochondria in these cells. These findings suggest that stimulation of P2 receptors can dose-dependently increase mitochondrial respiration and that AMP and ATP have opposing effects on mitochondrial respiration in SH-SY5Y cells. Mitochondrial Ca2+ levels are tightly coupled to cytosolic Ca2+ signaling (23). Therefore, we examined next how AMP and ATP influence cytosolic Ca2+ homeostasis in SH-SY5Y cells. Using Fluo4-AM-loaded SH-SY5Y cells, we found that AMP and ATP had opposing effects on cytosolic Ca2+ levels (Fig. 3, C and D; Supplemental Digital Content 5-Video 4, Like with the human SH-SY5Y cell line, AMP and ATP had also opposing effects on Ca2+ signaling of primary mouse hippocampus neurons (Fig. 3, E and F). Taken together, these findings suggest that AMP and ATP can both influence mitochondrial Ca2+ uptake and cytosolic Ca2+ signaling of neurons and that AMP and ATP exert opposing effects on mitochondrial metabolism in these cells.

Fig. 3
Fig. 3:
ATP and adenosine have opposing effects on mitochondrial Ca2+ uptake and cytosolic Ca2+ levels of neurons.

AMP regulates mitochondria via AMPK and mTOR signaling

The results above demonstrate that AMP has profound inhibitory effects on mitochondrial activity and Ca2+ homeostasis of neurons. Next, we investigated the underlying mechanisms. Cells can take up extracellular AMP, which can alter intracellular signaling processes that regulate cell metabolism (24, 25). We used HPLC analysis to assess the effect of AMP treatment on intracellular levels of AMP, ADP, AMP, and adenosine. We found that AMP treatment of SH-SY5Y cells dose-dependently increased intracellular AMP levels (Fig. 4A). Intracellular AMP levels reached peak values that were 25 times higher than the baseline AMP levels in untreated cells (Fig. 4B). These peak AMP levels were reached within 1 min after AMP exposure. Intracellular ATP levels also rose briefly but returned to baseline levels within 1 min. Increased intracellular AMP concentrations can activate the AMP-activated protein kinase (AMPK), which is a signaling molecule that switches cell metabolism to an energy conservation mode. AMPK can phosphorylate raptor, which is a binding partner of the mammalian target of rapamycin complex 1 (mTORC1) that promotes mitochondrial respiration (26). Phosphorylation of raptor by AMPK can inhibit mTORC1 signaling and thereby reduce mitochondrial respiration (27). We tested whether AMP induces AMPK activation in SH-SY5Y cells. Using immunoblotting assays, we found that AMP treatment of these cells induced rapid and prolonged phosphorylation of the Thr172 residue on the α-subunit of AMPK (Fig. 4C). Phosphorylation of this AMPK subunit is indicative of AMPK activation. This activation step was paralleled by a decrease in the phosphorylation of mTOR on Ser2448, which reflects downregulation of mTORC1. Consistent with its inhibitory effect on mTORC1 signaling, AMP treatment of SH-SY5Y cells also decreased the phosphorylation of P70S6K, which is a downstream target of the mTORC1 signaling pathway (Fig. 4C). These findings suggest that exposure of SH-SY5Y cells to AMP activates AMPK signaling, which attenuates the mTOR pathway and thereby blocks mitochondrial activity in AMP-treated cells. In support of this notion, we found that the AMPK agonist AICAR could also block mitochondrial Ca2+ uptake in SH-SY5Y cells (Fig. 4D). However, AICAR was less effective in suppressing mitochondrial Ca2+ uptake when compared with AMP, which indicates that AMP may elicit additional mechanisms that downregulate mitochondrial metabolism of neurons.

Fig. 4
Fig. 4:
AMP inhibits mitochondria via AMPK signaling.

Adenosine is not responsible for downregulated mitochondrial metabolism of SH-SY5Y neurons

Ectonucleotidases such as CD73 convert AMP in the extracellular space to adenosine. Adenosine has been shown to cause hypothermia in mice by activating adenosine receptors (28). Moreover, adenosine contributes to the processes that regulate sleep, torpor, and hibernation (29, 30). Therefore, we wondered whether AMP treatment suppresses mitochondria indirectly through adenosine-mediated pathways. We found that adenosine treatment of SH-SY5Y cells had little effect on the mitochondrial membrane potential when compared to AMP or CCCP (Fig. 5A). Even at millimolar concentrations, adenosine treatment failed to significantly reduce the mitochondrial membrane potential of SH-SY5Y cells (Fig. 5B). A2a adenosine receptors are highly expressed in neurons and in SH-SY5Y cells (31, 32). However, treatment of SH-SY5Y cells with the selective A2a receptor agonist CGS21680 also failed to alter mitochondrial Ca2+ signaling in these cells (Fig. 5C). Taken together, these findings suggest that adenosine does not contribute to the AMP-induced downregulation of mitochondrial metabolism in SH-SY5Y neurons.

Fig. 5
Fig. 5:
AMP, not adenosine, intercepts the stimulatory effect of ATP on mitochondrial metabolism.

AMP prevents the stimulatory effect of ATP on mitochondrial metabolism

We have previously reported that mitochondrial ATP production, ATP release, and autocrine signaling through P2 receptors represent a feedforward amplification mechanism that maintains mitochondrial metabolism in human leukocytes (33, 34). Therefore, we hypothesized that similar mechanisms may regulate mitochondrial metabolism in neurons. In order to test this hypothesis, we examined whether inhibition of ATP release or antagonists of P2 receptors can affect mitochondrial Ca2+ uptake in SH-SY5Y neurons. Interestingly, we found that treatment of neurons with CBX, an inhibitor of connexin- and pannexin-induced ATP release, reduced mitochondrial Ca2+ signaling in SH-SY5Y cells (Fig. 5C). Inhibition of P2 receptors with suramin had a similar suppressive effect, which suggests that ATP release and autocrine stimulation of P2 receptors are involved in maintaining mitochondrial metabolism of SH-SY5Y cells. P2Y2 receptors are widely expressed in the nervous system where they regulate cell viability and neuroinflammatory responses (35). UTP and ATP are equipotent natural ligands of P2Y2 receptors, which are Gαi-coupled GPCRs that increase intracellular Ca2+ signaling (11). Simulation of SH-SY5Y cells with ATP or UTP induced mitochondrial Ca2+ uptake, which indicates that P2Y2 receptors act as positive regulators of mitochondrial metabolism in these neuronal cells (Fig. 5, D and E). Interestingly, we also found that AMP treatment prevented Ca2+ influx in response to both P2Y2 receptor agonists (Fig. 5, D–F). These findings suggest that AMP interferes with P2Y2 receptor stimulation, either directly by interfering with receptor binding or indirectly by down-regulating mitochondrial metabolism. Taken together, we conclude from these data that AMP exerts its metabolic effects by impairing mitochondrial respiration, which prevents P2 receptor-inducted metabolic stimulation and culminates in prolonged shutdown of cell metabolism in AMP-treated mice.

AMP treatment prolongs survival under hypoxic conditions

The findings described above demonstrate that AMP can rapidly disengage mitochondrial metabolism in neuronal cells. Therefore, AMP treatment may reduce oxygen demand of neurons and protect the brain under conditions of impaired oxygen supply. We tested this concept by studying whether AMP injection alters the tolerance of mice to hypoxia. Mice were anesthetized with ketamine/xylazine, injected with AMP, and then exposed to an atmosphere containing 6% oxygen. It was previously reported that AMP acts dose-dependently over a wide therapeutic range without apparent adverse effects on mice (4). Therefore, we chose a high dose of AMP (4 mg/g i.p.) designed to maximize its protective hypometabolic effects. We found that pretreatment with AMP significantly prolonged the survival of mice under hypoxic conditions. Animals treated with AMP survived nearly twice as long as control mice that did not receive AMP (Fig. 6A). AMP had similarly beneficial effects on mice subjected to anoxic conditions (<1% O2; Fig. 6B). AMP plasma concentrations measured after AMP injection were almost 400-fold elevated and remained 20 times higher than baseline levels for at least 60 min (Fig. 6C). Taken together, these findings indicate that AMP has important therapeutic properties that could be exploited to induce a transient hypometabolic state with reduced oxygen demand that protects the brain from neurologic damage by ischemia and reperfusion.

Fig. 6
Fig. 6:
AMP prolongs survival under hypoxic conditions.


In this study, we found that AMP treatment downregulates mitochondrial metabolism in neuronal cells, which may be responsible for the reduced consciousness and overall decrease in metabolism seen in mice following AMP injection. The resulting hypometabolic state resembles that of hibernation or torpor, which are characterized by transiently reduced oxygen demand and energy consumption without lasting negative impact on cell viability (36). The hypometabolic state induced by AMP may have therapeutic potential in clinical settings. For example, AMP may help stabilize patients after trauma, hemorrhage, stroke, or cardiopulmonary arrest until oxygen supply to vital organs, particularly the brain, can be successfully restored.

Therapeutic hypothermia has been used to reduce the core body temperature to as low as 32°C, which improves neurological outcome and survival of patients with global ischemia following cardiac arrest (37). Current guidelines recommend mild hypothermia as standard of care for comatose survivors of cardiac arrest (38). The therapeutic utility of hypothermia is, however, restricted by the rate-limiting constraint of body mass that inversely correlates with the time needed to achieve protective levels of hypothermia (3, 39). Cooling a fully-grown human at least 1 h even if invasive methods such as endovascular cooling are employed (2). This cooling rate may be too slow to fully attenuate neuronal damage that occurs within minutes after the loss of oxygen supply to the brain. Therefore, other strategies have been considered to lower mitochondrial metabolism and oxygen demand. Studies with mice have shown that hydrogen sulfide (H2S) can be used to inhibit complex IV of the ETC in mitochondria, which causes pharmacologically induced hypothermia that may prevent tissue damage (40, 41). However, the toxicity of H2S and other injectable sulfide compounds is of concern and has hampered translation of this approach into clinical practice (42).

In contrast to sulfide compounds, AMP is a ubiquitous physiological metabolite that is comparatively easy to administer. Our data show that AMP causes an immediate drop in mitochondrial metabolism of mouse and human neurons. Moreover, our findings and the results of several other groups have shown that the suppression of physiological functions by AMP does not seem to have lasting negative side effects (4, 6, 16). The kinetics of the metabolic shutdown caused by AMP injection are superior when compared to those of forced therapeutic hypothermia. A rapid reduction of oxygen demand following AMP injection may be able to reduce or prevent the reverse electron transport and subsequent deleterious effects that are caused by loss of oxygen supply in ischemic conditions (8). This is supported by a previous study that has reported beneficial effects of AMP in a mouse model of ischemic brain injury (6).

The mechanisms by which AMP reduces metabolism are controversial and have not been fully established (4, 16, 43). Our previous work has shown that mitochondrial metabolism and purinergic signaling via P2 receptors work hand in hand to fine-tune the metabolism and function of immune cells (15, 33, 34). In human T cells, we found that panx1-induced ATP release and autocrine stimulation of P2X1 receptors represent a feed-forward signaling loop that maintains basal mitochondrial metabolism in resting cells (33). Our current findings suggest that panx1 channels and P2 receptors have similar roles in the regulation of mitochondrial metabolism in neurons. Furthermore, we found that AMP has opposing effects on mitochondrial activity that counterbalance the stimulatory effects of ATP. While ATP increased mitochondrial activity in a P2 receptor-mediated fashion, AMP inhibited mitochondria in an AMPK-dependent manner. Our findings suggest that downregulation of mitochondrial metabolism by AMP reduces ATP production, ATP release, autocrine P2 receptor stimulation, and mitochondrial cell metabolism (Fig. 7). Adenosine, a metabolite of ATP and AMP, can induce A2 adenosine receptor-induced inhibitory cAMP signaling mechanisms that have been shown to contribute to hypothermia, sleep, and torpor in mice (28–30). We did, however, not find a direct inhibitory effect of adenosine on mitochondrial metabolism in SH-SY5Y cells. It is possible that adenosine contributes to the hypometabolic effects of AMP through mechanisms that are independent of mitochondrial metabolism. It is also likely that different cell types differ in their responses to adenosine and other purinergic signaling molecules. To answer these questions will require further studies. Similarly, the roles of ATP breakdown and endogenously generated AMP in the regulation of cell metabolism will have to be explored in future studies.

Fig. 7
Fig. 7:
Proposed model of AMP-induced hypometabolism.

Despite the prolonged survival time of AMP-treated mice under hypoxic conditions, AMP was not able to prevent mortality for extended periods. A possible explanation is that AMP is rapidly removed from the extracellular space, which is confirmed by the finding that plasma AMP concentrations decreased within minutes after AMP by more than 50%. Combining AMP with other drugs that prolong the bioavailability of AMP could therefore be a promising approach to further enhance the hypometabolic and protective effects of AMP and the therapeutic window that is available to treat critical care patients. Our preliminary data suggest that the nucleoside transport inhibitor dipyridamole might be a possible candidate for such a combination treatment, which should be explored in more detail. In preliminary experiments, we found that the combination of dipyridamole with AMP hastened the rate of the decline in metabolism of AMP-treated mice, which seemed to further increase the survival time under hypoxic conditions (Supplemental Digital Content 2-Supplemental Fig. 2,

However, many questions remain to be answered to fully evaluate the potential efficacy of AMP as treatment to prevent brain damage in clinical settings. Preclinical studies should evaluate whether AMP can provoke hypometabolic effects not only in rodents like mice, rats, and hamsters as previously shown (4, 6, 44, 45), but also in larger animals. Encouraging preliminary data from our group suggest that AMP may be effective in lowering metabolism when administered to pigs (unpublished observations). In mice, we have shown that pretreatment with AMP is protective under hypoxic conditions. In addition, it will be important to investigate whether and to what extent AMP is able to improve outcome when given after an ischemic event, e.g., in settings that recapitulate clinical scenarios encountered in cardiac arrest patients. More work is also needed to further explore the molecular mechanisms by which AMP and adenosine modulate mitochondrial function. This could reveal additional therapeutic targets to prevent mitochondrial damage in response to ischemia and reperfusion and to translate the exciting concept of drug-induced hypometabolism into clinical practice.


1. Eltzschig HK, Eckle T. Ischemia and reperfusion—from mechanism to translation. Nat Med 17 (11):1391–1401, 2011.
2. Song SS, Lyden PD. Overview of therapeutic hypothermia. Curr Treat Options Neurol 14 (6):541–548, 2012.
3. Diller KR, Zhu L. Hypothermia therapy for brain injury. Annu Rev Biomed Eng 11:135–162, 2009.
4. Zhang J, Kaasik K, Blackburn MR, Lee CC. Constant darkness is a circadian metabolic signal in mammals. Nature 439 (7074):340–343, 2006.
5. Daniels IS, O Brien WG 3rd, Nath V, Zhao Z, Lee CC. AMP deaminase 3 deficiency enhanced 5’-AMP induction of hypometabolism. PLoS One 8 (9):e75418, 2013.
6. Muzzi M, Blasi F, Chiarugi A. AMP-dependent hypothermia affords protection from ischemic brain injury. J Cereb Blood Flow Metab 3 (2):171–174, 2013.
7. Solaini G, Baracca A, Lenaz G, Sgarbi G. Hypoxia and mitochondrial oxidative metabolism. Biochim Biophys Acta 1797 (6–7):1171–1177, 2010.
8. Chouchani ET, Pell VR, James AM, Work LM, Saeb-Parsy K, Frezza C, Krieg T, Murphy MP. A unifying mechanism for mitochondrial superoxide production during ischemia-reperfusion injury. Cell Metab 23 (2):254–263, 2016.
9. Burnstock G, Fredholm BB, North RA, Verkhratsky A. The birth and postnatal development of purinergic signalling. Acta Physiol (Oxf) 199 (2):93–147, 2010.
10. Junger WG. Immune cell regulation by autocrine purinergic signalling. Nat Rev Immunol 11 (3):201–212, 2011.
11. Abbracchio MP, Burnstock G, Boeynaems JM, Barnard EA, Boyer JL, Kennedy C, Knight GE, Fumagalli M, Gachet C, Jacobson KA, et al. International Union of Pharmacology LVIII: update on the P2Y G protein-coupled nucleotide receptors: from molecular mechanisms and pathophysiology to therapy. Pharmacol Rev 58 (3):281–341, 2006.
12. Fredholm BB, IJzerman AP, Jacobson KA, Linden J, Müller CE. International Union of Basic and Clinical Pharmacology. LXXXI. Nomenclature and classification of adenosine receptors: an update. Pharmacol Rev 63 (1):1–34, 2011.
13. Coddou C, Yan Z, Obsil T, Huidobro-Toro JP, Stojilkovic SS. Activation and regulation of purinergic P2X receptor channels. Pharmacol Rev 63 (3):641–683, 2011.
14. Bao Y, Ledderose C, Seier T, Graf AF, Brix B, Chong E, Junger WG. Mitochondria regulate neutrophil activation by generating ATP for autocrine purinergic signaling. J Biol Chem 289 (39):26794–26803, 2014.
15. Ledderose C, Bao Y, Lidicky M, Zipperle J, Li L, Strasser K, Shapiro NI, Junger WG. Mitochondria are gate-keepers of T cell function by producing the ATP that drives purinergic signaling. J Biol Chem 289 (37):25936–25945, 2014.
16. Swoap SJ, Rathvon M, Gutilla M. AMP does not induce torpor. Am J Physiol Regul Integr Comp Physiol 293 (1):R468–473, 2007.
17. Zhang J, Neal J, Lian G, Hu J, Lu J, Sheen V. Filamin A regulates neuronal migration through brefeldin A-inhibited guanine exchange factor 2-dependent Arf1 activation. J Neurosci 33 (40):15735–15746, 2013.
18. Wu J, Liu L, Matsuda T, Zhao Y, Rebane A, Drobizhev M, Chang YF, Araki S, Arai Y, March K, et al. Improved orange and red Ca2 ± indicators and photophysical considerations for optogenetic applications. ACS Chem Neurosci 4 (6):963–972, 2013.
19. Nociari MM, Shalev A, Benias P, Russo C. A novel one-step, highly sensitive fluorometric assay to evaluate cell-mediated cytotoxicity. J Immunol Methods 213 (2):157–167, 1998.
20. Clarke DD, Sokoloff L. Siegel G, Agrano BV, Albers RW, Molino PV. Circulation and energy metabolism of the brain. Basic Neurochemistry. New York: Raven Press; 1989; 565–590.
21. Attwell D, Laughlin SB. An energy budget for signaling in the grey matter of the brain. J Cereb Blood Flow Metab 21 (10):1133–1145, 2001.
22. Glancy B, Balaban RS. Role of mitochondrial Ca2+ in the regulation of cellular energetics. Biochemistry 51 (14):2959–2973, 2012.
23. Kann O, Kovács R. Mitochondria and neuronal activity. Am J Physiol Cell Physiol 292:C641–657, 2007.
24. Yegutkin GG. Nucleotide- and nucleoside-converting ectoenzymes: Important modulators of purinergic signalling cascade. Biochim Biophys Acta 1783:673–694, 2008.
25. Zhang F, Wang S, Luo Y, Ji X, Nemoto EM, Chen J. When hypothermia meets hypotension and hyperglycemia: the diverse effects of adenosine 5’-monophosphate on cerebral ischemia in rats. J Cereb Blood Flow Metab 29:1022–1034, 2009.
26. Ramanathan A, Schreiber SL. Direct control of mitochondrial function by mTOR. Proc Natl Acad Sci U S A 106 (52):22229–22232, 2009.
27. Gwinn DM, Shackelford DB, Egan DF, Mihaylova MM, Mery A, Vasquez DS, Turk BE, Shaw RJ. AMPK phosphorylation of raptor mediates a metabolic checkpoint. Mol Cell 30 (2):214–226, 2008.
28. Fredholm BB, Johansson S, Wang YQ. Adenosine and the regulation of metabolism and body temperature. Adv Pharmacol 61:77–94, 2011.
29. Bjorness TE, Dale N, Mettlach G, Sonneborn A, Sahin B, Fienberg AA, Yanagisawa M, Bibb JA, Greene RW. An adenosine-mediated glial-neuronal circuit for homeostatic sleep. J Neurosci 36 (13):3709–3721, 2016.
30. Iliff BW, Swoap SJ. Central adenosine receptor signaling is necessary for daily torpor in mice. Am J Physiol Regul Integr Comp Physiol 303 (5):R477–484, 2012.
31. Sebastião AM, Ribeiro JA. Adenosine receptors and the central nervous system. Handb Exp Pharmacol 193:471–534, 2009.
32. Canals M, Angulo E, Casadó V, Canela EI, Mallol J, Viñals F, Staines W, Tinner B, Hillion J, Agnati L, et al. Molecular mechanisms involved in the adenosine A1 and A2A receptor-induced neuronal differentiation in neuroblastoma cells and striatal primary cultures. J Neurochem 92 (2):337–348, 2005.
33. Ledderose C, Bao Y, Ledderose S, Woehrle T, Heinisch M, Yip L, Zhang J, Robson SC, Shapiro NI, Junger WG, et al. depleted purinergic signaling, and defective T cell vigilance and immune defense. J Infect Dis 213 (3):456–464, 2016.
34. Bao Y, Ledderose C, Graf AF, Brix B, Birsak T, Lee A, Zhang J, Junger WG. mTOR and differential activation of mitochondria orchestrate neutrophil chemotaxis. J Cell Biol 210 (7):1153–1164, 2015.
35. Peterson TS, Camden JM, Wang Y, Seye CI, Wood WG, Sun GY, Erb L, Petris MJ, Weisman GA. P2Y2 nucleotide receptor-mediated responses in brain cells. Mol Neurobiol 41 (2–3):356–366, 2010.
36. Carey HV, Andrews MT, Martin SL. Mammalian hibernation: cellular and molecular responses to depressed metabolism and low temperature. Physiol Rev 83 (4):1153–1181, 2003.
37. Arrich J, Holzer M, Havel C, Müllner M, Herkner H. Hypothermia for neuroprotection in adults after cardiopulmonary resuscitation. Cochrane Database Syst Rev 2:CD004128, 2016.
38. Donnino MW, Andersen LW, Berg KM, Reynolds JC, Nolan JP, Morley PT, Lang E, Cocchi MN, Xanthos T, Callaway CW, et al. Temperature management after cardiac arrest: an advisory statement by the Advanced Life Support Task Force of the International Liaison Committee on Resuscitation and the American Heart Association Emergency Cardiovascular Care Committee and the Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation. Circulation 132 (25):2448–2456, 2015.
39. Testori C, Sterz F, Behringer W, Spiel A, Firbas C, Jilma B. Surface cooling for induction of mild hypothermia in conscious healthy volunteers—a feasibility trial. Crit Care 15:R248, 2011.
40. Blackstone E, Morrison M, Roth MB. H2S induces a suspended animation-like state in mice. Science 308 (5721):518, 2005.
41. Szabo C, Ransy C, Módis K, Andriamihaja M, Murghes B, Coletta C, Olah G, Yanagi K, Bouillaud F. Regulation of mitochondrial bioenergetic function by hydrogen sulfide. Part I. Biochemical and physiological mechanisms. Br J Pharmacol 171 (8):2099–20122, 2014.
42. Asfar P, Calzia E, Radermacher P. Is pharmacological, H2S-induced 'suspended animation’ feasible in the ICU? Crit Care 18 (2):215, 2014.
43. Carlin JL, Jain S, Gizewski E, Wan TC, Tosh DK, Xiao C, Auchampach JA, Jacobson KA, Gavrilova O, Reitman ML. Hypothermia in mouse is caused by adenosine A(1) and A(3) receptor agonists and AMP via three distinct mechanisms. Neuropharmacology 114:101–113, 2017.
44. Miao YF, Wu H, Yang SF, Dai J, Qiu YM, Tao ZY, Zhang XH. 5’-adenosine monophosphate-induced hypothermia attenuates brain ischemia/reperfusion injury in a rat model by inhibiting the inflammatory response. Mediators Inflamm 2015:520745, 2015.
45. Dugbartey GJ, Bouma HR, Strijkstra AM, Boerema AS, Henning RH. Induction of a torpor-like state by 5’-AMP does not depend on H2S production. PLoS One 10 (8):e0136113, 2015.

AMP; hypometabolism; hypothermia; mitochondria; purinergic signaling

Supplemental Digital Content

Copyright © 2020 by the Shock Society