Effects of Metformin in Experimental Stroke

Abstract

Background and Purpose

AMP-activated protein kinase (AMPK) is an important sensor of energy balance. Stroke-induced AMPK activation is deleterious as both pharmacological inhibition and genetic deletion of AMPK are neuroprotective. Metformin is a known AMPK activator, but reduces stroke incidence in clinical populations. We investigated the effect of acute and chronic Metformin treatment on infarct volume and AMPK activation in experimental stroke.

Methods

Male mice were subjected to middle cerebral artery occlusion (MCAO) after acute (3 days) or chronic (3 weeks) administration of Metformin. Infarct volumes, AMPK activation, lactate accumulation, and behavioral outcomes were assessed. The role of neuronal Nitric Oxide Synthase (nNOS) and AMPK were examined using mice with targeted deletion of AMPK or nNOS.

Results

Acute Metformin exacerbated stroke damage, enhanced AMPK activation, and led to metabolic dysfunction. This effect was lost in AMPK and nNOS knockout mice. In contrast, chronic Metformin given pre-stroke was neuroprotective, improved stroke-induced lactate generation, and ameliorated stroke-induced activation of AMPK. Similarly the neuroprotective effect of chronic pre-stroke Metformin was lost in nNOS knockout mice.

Conclusions

AMPK is an important potential target for stroke treatment and prevention. These studies show that the timing, duration and amount of AMPK activation are key factors in determining the ultimate downstream effects of AMPK on the ischemic brain.

AMP-activated protein kinase (AMPK) is a known sensor of peripheral energy balance. AMPK is activated when cellular energy supply is low, as signaled by increasing intracellular AMP and declining ATP levels ¹. In the periphery, AMPK acutely regulates cellular metabolism and chronically regulates gene expression, reducing energy storage (fatty acid, lipid and protein biosynthesis) and increasing energy utilization (fatty acid oxidation). AMPK is highly expressed in neurons and is rapidly activated in the brain during energy deprived states such as fasting and ischemia ^{2, 3}. We have previously shown that stroke-induced AMPK activation is deleterious after an induced focal stroke as both pharmacological inhibition and genetic deletion of AMPK are neuroprotective ^{3, 4}.

The downstream mediators of AMPK’s deleterious effects in the ischemic brain are not yet clear but may involve enhancement of lactic acidosis. Neurons, unlike astrocytes, have minimal activity of the key glycolytic enzyme PFK-2⁵, have no glycogen stores ⁶, and are therefore exquisitely sensitive to hypoxia and hypoglycemia. Within a short period of energy deficiency, AMPK activation enhances astrocytic glycolysis and ketosis to provide energy to ischemic neurons ⁷. However, prolonged anaerobic glycolysis in astrocytes leads to progressive acidosis and inhibits the ability of neurons to use lactate as an energy source ^{7, 8} contributing to neuronal death. In cerebral ischemia, when energy depletion is severe, lactate acidosis enhanced by AMPK activation could lead to an exacerbation of stroke injury.

Metformin is a drug widely used for the treatment of type 2 diabetes mellitus (DM2) ⁹. It has been well documented that acute Metformin treatment activates AMPK both in vivo and in vitro ^10–12. AMPK activation is responsible for a number of the actions of Metformin related to its glucose lowering effects, including a decrease in glucose production in hepatocytes and increase in glucose uptake in skeletal muscle¹². The activation of AMPK by Metformin requires nitric oxide (NO) as Metformin no longer activates AMPK when NO is directly inhibited in bovine endothelial cells ¹³ or in mice lacking the endothelial form of Nitric Oxide Synthase (eNOS−/−) ¹². This suggests that activation of AMPK by Metformin is NO dependent, at least in the vasculature. Currently, there is no data on the effect of Metformin on AMPK in the brain. Due to its AMPK activating effects, treatment may exacerbate damage during acute ischemia. However, in clinical populations, chronic Metformin treatment is associated with a lower risk of stroke, reducing cardiovascular mortality by 26%. This protection is independent of its glucose-lowering effect^{14, 26}. The effects of chronic activation of AMPK in experimental stroke have not been previously investigated. In this study we examined the effect of Metformin in experimental stroke to determine effects on lactic acidosis, AMPK, and infarct size.

Materials and Methods

Focal cerebral ischemia model and physiology

The present study was conducted in accordance with NIH guidelines for the care and use of animals in research and under protocols approved by the Center for Laboratory Animal Care at UCHC. Focal transient cerebral ischemia was induced in male mice (20–25g) by right middle cerebral artery occlusion (MCAO) followed by reperfusion as described previously⁴. At the end of ischemia (90 minutes MCAO), the animal was briefly re-anesthetized, and reperfusion was initiated by filament withdrawal. In separate cohorts, blood glucose, physiological measurements, femoral arterial blood pressure, and cortical perfusion (Laser Doppler Flowmetry; LDF) were evaluated throughout MCAO and early reperfusion as described previously⁴. Wild type (WT) mice were purchased from Charles River. AMPK α2 KO mice (C57BL6 background) were obtained from Dr. Benoit Viollet in France and re-derived in house ⁴. nNOS KO mice (C57BL6 background) and corresponding littermates were obtained from Jackson labs.

Behavioral measurements

Neurological deficits (NDS) were scored in the intra-ischemic period, 24 or 72 hours post-stroke. The scoring system was as follows: 0, no deficit; 1, forelimb weakness and torso turning to the ipsilateral side when held by tail; 2, circling to affected side; 3, unable to bear weight on affected side; and 4, no spontaneous locomotor activity or barrel rolling³.

Infarct analysis

At 24 or 72 hours after stroke, the brain was removed and cut into 5 2-mm slices and stained with 1.5% 2, 3, 5 triphenyltetrazolium (TTC) for 30 min at 3°C. Slices were formalin fixed (4%) then digitalized and infarct volumes analyzed (Sigma Scan Pro) as previously described ⁴. The final infarct volumes are presented as both direct volume (in mm3) and indirect volume (percentage of contralateral structures with correction for edema³). In animals assessed chronically post-stroke, mice were sacrificed at 3 weeks via pentobarbital overdose and perfused transcardially with cold PBS followed by 4% paraformaldehyde; the brain was postfixed for 18 h and placed in cyroprotectant (30% sucrose). The brain tissue was cut into 40-μm free-floating sections on a freezing microtome and every eighth slice was stained by cresyl violet staining for evaluation of ischemic cell damage. Sections were digitized by a CCD camera and analyzed with photoimaging software (Jandel Scientific). Due to the chronic nature of this study, cerebral atrophy was used as an indirect measure of cell loss. The volume of tissue atrophy was determined by measuring both hemispheres and lateral ventricles and transformed to mm³. Percent atrophy was computed by dividing the ischemic (right) hemisphere from the intact (left) hemisphere, then multiplying by 100 as in ³².

Metformin treatment

Metformin was dissolved in saline (vehicle). For acute treatment, Metformin was injected daily for 3 days (50 or 100 mg/kg Metformin or vehicle, i.p.) prior to stroke. For chronic pre-stroke Metformin treatment, WT and nNOS KO mice were injected for 3 weeks (50 mg/kg or vehicle) daily prior to stroke. Dose volume was 0.2ml/20g body weight. In an additional study, Metformin (i.p. 50 mg/kg per day) was administered 24 hours after the onset of MCAO for three weeks.

Western Blots

Western blots were done as described previously⁴. Four hours after the onset of cerebral ischemia, mice were sacrificed, brains were homogenized using lysis buffer and protein was loaded on a 4% to 15% gradient SDS–polyacrylamide gel and transferred to a polyvinylidene difluoride membrane. p-AMPK was probed with antibody from Cell Signaling (1:1000). Beta-actin (1:5000; Sigma) was used as the loading control. Blots were incubated overnight in primary antibody at 4°C in TBS containing 4% bovine serum albumin and 0.1% Tween20. Secondary antibodies (goat anti-rabbit IgG 1:5,000 for p-AMPK, goat anti-mouse IgG 1:5000 for Beta-actin; Chemicon) were diluted and ECL (pico) detection kit (ThermoScientific) was used for signal detection ⁴.

Brain lactate measurement

Four hours after the onset of stroke, mice were euthanized and brains were homogenized with perchloric acid (6%) and spun, the supernatants were used for lactate measurements. Lactate measurement was done by colorimetric assay using a lactate assay kit following the manufacture’s instruction (Abcam, Cambridge, MA).

Statistics

Data were expressed as mean±sem except for NDS which was presented as Median (Interquartile range). Statistics were performed either with Student t-test, one-way analysis of variance with Tukey post hoc test for multiple comparisons (for p-AMPK) or by Mann– Whitney U test (neurologic deficit scores). A p value < 0.05 was considered to be statistically significant. Investigators performing MCAO, behavioral and infarct size analysis were blinded to treatment group.

Results

Acute Metformin Treatment Increased Infarct 24 hours after stroke

24 hours following MCAO, Metformin (100 mg/kg) significantly increased total (Metformin 67.0±6.8% vs. vehicle 42.8±4.2%; p<0.05), cortical (Metformin 66.8±5.8% vs. vehicle 42.2±4.6%; p<0.05) as well as striatal (Metformin 77.4±4.5% vs. vehicle 57.6±3.3%; p<0.05) infarct volume (direct infarct volumes mm³; Cortex: Met: 67.8±7.1 vs. Vehicle: 38.5 ±6.0 p<0.05, Striatum: Met: 33.8±0.92 vs.21.9±2.3 p<0.05, Total: 121±9.8 vs. 70.4±7.5, p<0.05). At 50 mg/kg, Metformin increased both total (Met 61.2±4.7% vs. vehicle 42.8±4.2%; p<0.05) and striatal (Metformin 75.1±5.6% vs. vehicle 57.6±3.3%; p<0.05) infarction volume but had no significant effect on cortical infarction (Metformin 56.9±7.8 versus vehicle 42.2±4.6%) compared to vehicle-treated mice (Figure 1A) (direct infarct volumes mm³; Cortex: Met: 56.8±7.5 vs. Vehicle: 38.5 ±6.0, Striatum: Met: 32.2±2.0 vs.21.9±2.3 p<0.05, Total: Met 113±9.3 vs. Vehicle 70.4±7.5, p<0.05). The detrimental effect of Metformin was also reflected in the NDS (Metformin 100mg/kg; 3.0 (0) vs. vehicle 2.0(1), p<0.05) (Table 1, on line). No mortality was seen in any of the groups.

Fig. 1.

Acute Metformin treatment exacerbated stroke outcome in mice subjected to 90 minute MCAO with 24 hours (A & B) or 72 hours (C) survival. A & B. Met: Metformin; Metformin was given at 100 mg/kg or 50 mg/kg, i.p. daily for 3 days prior to stroke n=7 Metformin p/g; n=8 vehicle. C. Metformin was given at 100mg/kg daily for 3 days prior to stroke. n=10 vehicle, n=11 metformin. Cortical, striatal and total hemisphere infarcts were calculated (corrected for edema, percentage of non-ischemic hemisphere). * P<0.05 versus control (One-way ANOVA with Tukey post-hoc test); data are expressed as Mean±sem.

Acute Metformin Treatment Increased Infarct 72 hours after stroke

In the 72-hour survival group, Metformin (100 mg kg⁻¹) also significantly increased total (Metformin 66.0±2.2% vs. vehicle 49.6±5.1%; p<0.05), striatal (Metformin 74.4±2.3% vs. vehicle 61.0±4.5%, p<0.05) as well as cortical (Metformin 58.9±1.9% vs. vehicle 43.1±7.2%, p<0.05) infarct volume (contralateral hemisphere) in comparison to the saline-treated group (Figure 1B ) (direct infarct volumes mm³; Cortex: Met: 51.2±1.8 vs Vehicle: 41.2 ±5.1, Striatum: Met: 39.4±3.9 vs.27.3±1.4 p<0.05, Total: 104±4.8 vs. 79.2±8.2, p<0.05). Neurological deficits were also exacerbated by Metformin treatment (Metformin 3.0(1) vs. vehicle 2.0(0), p<0.05; Table 1).

There were no differences in physiological measurements between the acute or chronic Metformin and vehicle-treated groups (Table 2, on line). LDF was equivalently reduced during ischemia and was restored equally in early reperfusion (Table 2). The mortality rates were equivalent between groups (1 out of 11 in vehicle and 2 out of 13 in Metformin).

Acute Metformin Increased pAMPK levels

As expected, stroke significantly induced pAMPK levels 4 hours after the onset of stroke (stroke 1.63±0.19 versus sham 0.87±0.079, p<0.05, n=5). Metformin administration (100 mg/kg) significantly enhanced stroke-activated pAMPK levels compared to vehicle-treated MCAO mice (Metformin 3.02±0.36 versus vehicle 1.63 ±0.19, p<0.05) (Figure 2A & B).

Fig. 2.

Acute Metformin treatment activated AMPK after stroke. Brains were collected 4 hours after stroke. Vehicle/Metformin (100 mg/kg) was administered through i.p. for 3 days prior to MCAO. n= 5/pg. * p<0.05 stroke versus sham, #p<0.05, Metformin versus vehicle (One-way ANOVA with Tukey post-hoc test); data are expressed as Mean±Sem.

Deletion of AMPK Abolished the Deleterious Effects of Acute Metformin

To investigate if Metformin produces its detrimental effects specifically via AMPK activation, we examined the effect of acute Metformin in AMPKα2 knockout mice. The deleterious effects of acute Metformin treatment were ameliorated in AMPK α2−/− mice (Cortex: vehicle 38.5±3.4% vs. Metformin 35.6±5.8%; Striatum: vehicle 52.6±5.9% vs. Metformin 60.0±7.6%; Total: vehicle 33.3±3.2% vs. Metformin 33.4±4.1%) (Figure 3A) (direct infarct volumes mm³; Cortex: vehicle: 34.5±2.1 vs. Met 32.0 ±5.5, Striatum: Vehicle 18.0±3.4 vs. Met 19.3±2.7, Total: Vehicle 55.2±3.9 vs. Met 60.7±7.9). No mortality was seen in the groups.

Fig. 3.

The effect of acute Metformin treatment in stroke is mediated by AMPK. A. Metformin lost its effect in AMPK-2 null mice. n=7 saline, n=6 Metformin. B. Metformin had no effect in nNOS null mice. Metformin was given to mice at 100 mg/kg i.p. daily for 3 days prior to MCAO. Controls were given saline. n=10 saline, n=11 Metformin. Cortical, striatal and total hemisphere infarcts were calculated (corrected for edema, percentage of non-ischemic hemisphere). * p<0.05 Metformin versus vehicle (Student t-test). Data were expressed as Mean±sem.

AMPK mediates its effects through NO

It has been shown that activation of AMPK by Metformin is dependent on NO in the vasculature. The detrimental effects of acute Metformin treatment were lost in male nNOS−/− mice (Cortex: vehicle 31.4±5.0% vs. Metformin 30.5±4.8%, Striatum: vehicle 41.2± 6.5% vs. 44.8± 7.9%; Total: vehicle 34.4±5.2% vs. Metformin 32.3±5.0%) (Figure 3B) (direct volume mm³; Cortex 27.4±5.0 vs. 32.2 ±4.9, Striatum 18.3 ±3.7 vs. 19.9±3.4, Total 54.1±9.1 vs.54.3±7.9). The mortality rates were not significantly different between groups (0 out of 10 in vehicle and 1 out of 12 in Metformin).

Stroke-Induced Lactate Levels Were Elevated in Mice Acutely Treated with Metformin

Brain lactate levels were increased 4 hours after stroke (vehicle stroke 15.1 ± 0.36 vs. vehicle sham 9.62 ± 0.19, p<0.05, n=4/pg). Acute Metformin treatment significantly exacerbated the stroke-induced increase in lactate levels (Metformin stroke 20.6± 1.5 versus vehicle stroke 15.1 ± 0.36 p<0.05, n=4 p/g) (Figure 4).

Fig. 4.

Acute metformin treatment exacerbated mice brain lactate accumulation after stroke. Brains were removed 4 hours after stroke and homogenized in perchloric acid (6%). Lactate was immediately measured using Lactate Assay Fluorometric Kit (Abcam). n=4/group. Statistics were performed with one-way ANOVA with Tukey post-hoc test, *p<0.05 versus vehicle sham; #p<0.05 versus vehicle stroke; data are expressed as Mean±sem.

Chronic Pre-stroke Treatment with Metformin Was Neuroprotective

Chronic treatment with Metformin reduced infarct size 24 hours after stroke both in the striatum and total hemisphere (Cortex: 48.3±2.1% vs. 32.1±6.6%; Striatum: 53.6±4.0% vs. 45.3±2.9% p<0.05, Total: 45.3±2.9% vs. 29.3±4.9%, p<0.05) (Figure 5) (direct volumes mm³; Cortex: 46.9±2.6 vs. 29.7 ±6.9, Striatum: 25.6 ±2.6 vs. 12.4±1.9, Total: 82.8±6.1 vs.49.9±9.2, p<0.05). The protective effect of chronic Metformin treatment was reflected in the NDS (Metformin 1.5(1) versus vehicle 2.0(1), p<0.05) (Table 1). The mortality rates were equivalent between groups (1 out of 7 in vehicle and 1 out of 9 in Metformin).

Fig. 5.

Chronic treatment of metformin reduced infarct volumes in mice subjected to 90 minute MCAO with 24 hours reperfusion. Metformin was given to wild type mice at 50 mg/kg daily for 3 weeks prior to stroke. TTC was used to analyze infarct volumes. n= 6 for control and n=8 for metformin treated group. Cortical, striatal and total hemisphere infarcts were calculated (corrected for edema, as percentage of non-ischemic hemisphere). * P<0.05 versus control (two-tailed Student t-test); data are expressed as Mean±sem.

The effect of Chronic Pre-stroke Metformin treatment was Mediated by NO

Since NO may play a role in Metformin’s effect in stroke, we tested the effect of chronic Metformin treatment in nNOS KO mice. Chronic Metformin treatment (pre-stroke) lost its neuroprotective effect in nNOS KO mice suggesting that this effect is mediated by nNOS (Cortex: Met 38.4 ± 4.6% vs. Vehicle 32.5±3.4%, Striatum: Met 45.6±3.4% vs. 49.0±5.1%, Total: Met 35.8±2.0% vs. 32.0±2.7%) (Figure 6) (direct volume mm³; Cortex 29.8±2.2 vs. 26.9 ±3.4, Striatum 18.6 ±2.3 vs. 16.2±2.8, Total 53.8±4.1 vs.52.8±6.1). Mortality rate was 1 out of 7 in each group.

Fig. 6.

The effect of chronic Metformin treatment was mediated by nNOS. Metformin was given to nNOS KO mice at 50 mg/kg daily for 3 weeks prior to stroke. TTC was used to analyze infarct volumes. n= 6 for vehicle and n=6 for Metformin treated group. Cortical, striatal and total hemisphere infarcts were calculated (corrected for edema, as percentage of non-ischemic hemisphere). * P<0.05 versus control (two-tailed Student t-test); data are expressed as Mean±sem.

Chronic Pre-stroke treatment with Metformin down-regulated pAMPK and Ameliorated Lactate Accumulation

Chronic treatment with Metformin prior to stroke reduced pAMPK level in stroke when compared to sham Metformin-treated mice (Figure 7A& B). As expected, lactate levels after stroke were elevated in animals treated with vehicle. However chronic administration of Metformin led to an amelioration of stroke-induced lactate generation as levels were not significantly different from Metformin-treated sham animals (Figure 7C).

Fig. 7.

Effect of chronic Metformin onAMPK phosphorylation and lactate levels after stroke. Metformin was given to wild type mice at 50 mg/kg daily for 3 weeks prior to stroke. A &B, Metformin treated mice had reduced pAMPK levels after stroke when compared to sham mice treated with Metformin (n=2 in sham and n=3 stroke). Brains were removed 4 hours after stroke and homogenates were used for Western blots. Statistics were performed with Student t-test. *p<0.05 versus Metformin sham; C. Metformin chronic treatment did not enhance lactate accumulation in stroke. Brains were removed 4 hours after stroke and homogenized in perchloric acid (6%). Lactate was immediately measured using Lactate Assay Fluorometric Kit (Abcam). n=4/group. Statistics were performed with one-way ANOVA with Tukey post-hoc test, *p<0.05 versus vehicle sham; data are expressed as Mean±sem.

Chronic Metformin Treatment Initiated after Stroke was Neuroprotective

Chronic Metformin (50mg/kg i.p. per day for three weeks initiated 24 hours after stroke onset) reduced infarct volumes (measured as % tissue atrophy) three weeks after MCAO (Vehicle 18.84 ± 4.07% vs. Metformin 6.72 ±1.07%, n=7 per group, p <.01). Mortality was 2 out of 9 in each group.

Discussion

This paper demonstrates several important new findings regarding the role of AMPK in stroke. Firstly, acute Metformin activates AMPK in the brain in vivo as documented by elevations in pAMPK in brain homogenates. Secondly, Metformin-induced AMPK activation specifically exacerbates stroke damage as the detrimental effects of acute Metformin are abolished in AMPK α2 mice; the catalytic isoform that is responsible for the detrimental effect of AMPK activation in ischemic brain ⁴. Thirdly, the signaling pathway involves NO. Both acute Metformin-induced exacerbation of ischemic damage and the neuroprotection induced by chronic Metformin treatment were lost in nNOS knockout mice. Interestingly, the protective effect of chronic Metformin had a long therapeutic window as a neuroprotective effect was seen even if treatment was not initiated until 24 hours after the onset of stroke. Acute Metformin administration significantly increased brain lactate levels after stroke, which may represent increased anaerobic glycolysis secondary to AMPK activation. Finally, chronic pre-stroke Metformin treatment was neuroprotective, was associated with less stroke-induced AMPK activation, and ameliorated the detrimental metabolic changes seen with acute Metformin administration. This suggests that the duration and degree of AMPK activation are critical factors in determining the downstream physiological response in the ischemic brain.

The role of AMPK in stroke has been a subject of considerable debate¹⁵. Several studies, both in vivo and in vitro have demonstrated that acute AMPK activation is detrimental whereas others have suggested that it may play a neuroprotective role. Stroke represents a state of severe energy deficiency, as there is no available substrate for the energy consumptive pathways activated by AMPK. We have previously shown that the AMPK activator AICAR exacerbated stroke injury, however AICAR has numerous “off-target” effects that could contribute to enhanced injury, i.e., vasodilatation via activation of adenosine³. As Metformin is also known to activate AMPK, yet has clear beneficial effects on stroke incidence in clinical populations, we wanted to determine the effect of AMPK in an acute stroke model, recognizing that stroke incidence and acute neuroprotective effects may be mediated by different mechanisms. Surprisingly the effect of Metformin in the ischemic brain appears to be dependent on duration of treatment, and is linked to activation of AMPK.

There are several possible mechanisms by which Metformin can activate AMPK. Metformin can produce a decline in the free ATP/ADP and ATP/AMP ratios serving as a stimulus for AMPK phosphorylation and activation¹⁰. Alternatively, Metformin can directly activate an upstream AMPK kinase (AMPKK) or promote AMPK phosphorylation by binding to AMPK, making it a better substrate for an AMPKK¹⁶. Xie et al. has found that LKB1, an AMPKK, is required for Metformin-induced AMPK activation¹⁷. Metformin increased the phosphorylation and nuclear export of LKB1 into the cytosol in endothelial cells, leading to enhanced association of AMPK with LKB1. Inhibition of LKB1 abolished the stimulatory effect of Metformin on AMPK, and conversely, LKB1 over-expression enhanced Metformin induced AMPK phosphorylation ¹⁷. There is little data on possible interactions between LKB1 and NO in the brain. However, it is known that NO reacts with superoxide anions to form the potent oxidant peroxynitrite (ONOO), which activates AMPK in a LKB1 dependent manner in cultured endothelial cells ¹⁸. It is possible that in cerebral ischemia, where oxidative stress is high, Metformin activates NO, forming ONOO, leading to activation of AMPK. However, in this study we found no differences in LKB1 phosphorylation or translocation in Metformin treated animals (data not shown).

Our data demonstrate that the AMPK activation induced by Metformin requires NO, as acute Metformin treatment has no effect in male nNOS KO mice after stroke. It is well known that among the three major NOS isoforms in the brain, nNOS (neuronal NOS) is the primary isoform in neurons ¹⁹. Our studies suggest that the detrimental effects of acute Metformin are mediated in part by activation of nNOS. However, it is well known that AMPK, like NOS, is expressed in neurons, astrocytes and endothelium ^{1, 3, 20}. Our studies evaluated stroke outcome in an in vivo model, where all cell types interact “in-situ” and assessed AMPK activation in whole brain homogenates, therefore we cannot definitively exclude effects of Metformin in the endothelium or other cell types. Although it is likely that the activation of AMPK by acute Metformin in the neurons produced the detrimental effect of this drug in ischemia, further confirmation of this in vivo will require examination of mice with neuron-specific deletion of AMPK, which are currently under development.

The mechanisms through which acute AMPK activation exacerbates stroke injury are not yet clear. Studies have suggested exacerbated lactate accumulation, autophagy and increased glucose due to unregulated glucose-transporters in the reperfusion phase may contribute to stroke damage (For review see 15). AMPK can increase glycolysis by activating PFK-2, increasing fatty acid oxidation, activating glucose transport (via glucose transporter 4), and inhibiting glycogen synthesis to enhance available energy to the deprived cell ²¹. However, neurons are known to have minimal anaerobic glycolytic capacity ⁵. Astrocytes can perform glycolysis, however, over-activation of astrocytic glycolysis can be detrimental. Following ischemia in the brain, astrocytic PFK-2 activation is increased stimulating PFK-1, a glycolytic regulatory enzyme, subsequently leading to the production of lactate ⁶. Lactate can also be produced directly from glycogen, which astrocytes store in large amount ⁶. These processes result in progressive acidosis. Lactic acid is more cytotoxic than inorganic acids such as hydrochloric acid as it generates further intracellular acidosis ^{22, 23} and induces cellular edema ²⁴. In clinical settings, it has been well described that, although rare, Metformin can cause lactic acidosis in the periphery and the underlying etiology for this is not clear ²⁵.

As an initial step to investigate the mechanism by which acute Metformin exacerbates injury after MCAO, we examined the effect of acute Metformin on cerebral lactate levels. Acute Metformin significantly enhanced stroke-induced lactate accumulation, suggesting that lactate may contribute to the detrimental effect of acute Metformin. However, it is possible that the increased lactate may simply reflect the larger infarct volume. As the accumulation occurs quite early after ischemia it is likely that the increased lactate is in part responsible for the detrimental effects of acute Metformin. It is noteworthy that lactate levels in the chronic Metformin treated sham group are slightly higher than the vehicle sham group, suggesting long term activation of AMPK may enhance lactate in the intact, non-injured, brain. Metformin may serve as a preconditioning stimulus, making the brain less vulnerable to subsequent injury by exposing it to low levels of metabolic stress.

Chronic treatment with Metformin reduced infarct volume, reduced stroke-induced lactate generation and decreased stroke-induced AMPK activation. As we have shown that acute AMPK activation is detrimental after MCAO, chronic activation could lead to a down-regulation of AMPK and hence neuroprotection. Our data demonstrated that the neuroprotective effect of chronic Metformin was lost in nNOS KO mice. This suggests that neuronal NO plays an important role in both the deleterious effects of acute Metformin and the beneficial effects of chronic Metformin. Importantly, chronic metformin treatment reduced tissue atrophy measured three weeks after stroke, even when treatment was delayed for 24 hours. This suggests that continued cell death occurs after MCAO, and that this penumbra is potentially salvageable by manipulations of energy sensing mechanisms. The mechanism for this effect is unclear but may involve reductions in post-stroke inflammation, vascular dysfunction, or metabolic derangements, or alternatively, enhancement of post-stroke growth factors or post-conditioning, many of which are intricately tied to alterations in AMPK (see ¹⁵ for review), and will be investigated in future studies. Our finding of a neuroprotection with chronic Metformin administration is in agreement to observations from clinical studies. Intensive blood glucose control with Metformin reduced stroke ²⁶ and CV mortality by 26% relative to other antidiabetic agents or placebo ¹⁴. AMPK activation also reduces vascular inflammation, protects the endothelium, and enhances endothelial function via its actions on eNOS ^27–31 in peripheral blood vessels. Whether activation of AMPK in the cerebral vasculature has similar effects remains to be investigated.

Collectively, our data show that AMPK is a potentially important target for stroke treatment and prevention. These studies show that the timing, duration and amount of AMPK activation are key factors in determining the ultimate downstream effects of AMPK on the brain. The finding that acute activation of AMPK increased ischemic damage further confirms that acute AMPK activation is detrimental in stroke, consistent with previous findings from other pharmacological ³ and genetic studies ⁴. The detrimental effect of acute AMPK activation may be mediated, at least in part, by enhancement of lactic acidosis. Chronic Metformin treatment may lead to sublethal metabolic stress and down regulate AMPK protecting the brain from subsequent injury.