|Home | About | Journals | Submit | Contact Us | Français|
Oxygen–glucose deprivation (OGD) initiates a cascade of intracellular responses that culminates in cell death in sensitive species. Neurons from Arctic ground squirrels (AGS), a hibernating species, tolerate OGD in vitro and global ischemia in vivo independent of temperature or torpor. Regulation of energy stores and activation of mitogen-activated protein kinase (MAPK) signaling pathways can regulate neuronal survival. We used acute hippocampal slices to investigate the role of ATP stores and extracellular signal-regulated kinase (ERK)1/2 and Jun NH2-terminal kinase (JNK) MAPKs in promoting survival. Acute hippocampal slices from AGS tolerated 30 mins of OGD and showed a small but significant increase in cell death with 2 h OGD at 37°C. This tolerance is independent of hibernation state or season. Neurons from AGS survive OGD despite rapid ATP depletion by 3 mins in interbout euthermic AGS and 10 mins in hibernating AGS. Oxygen–glucose deprivation does not induce JNK activation in AGS and baseline ERK1/2 and JNK activation is maintained even after drastic depletion of ATP. Surprisingly, inhibition of ERK1/2 or JNK during OGD had no effect on survival, whereas inhibition of JNK increased cell death during normoxia. Thus, protective mechanisms promoting tolerance to OGD by AGS are downstream from ATP loss and are independent of hibernation state or season.
Ischemic injury to the brain can result from various traumas, including cardiac arrest, stroke, and traumatic brain injury, and is a primary cause of adult disability. Despite numerous therapeutic strategies found to reduce ischemia-induced damage in experimental models, only tissue plasminogen activator has been clinically approved; yet only 8% of stroke patients are eligible for treatment (Kleindorfer et al, 2004).
Hibernation is a unique physiologic adaptation characterized by periods of reduced metabolic activity and body temperature (Tb) that are interrupted at regular intervals by brief periods of arousal where metabolism and Tb return to euthermic levels (i.e., Tb of 37°C; interbout euthermic). Cerebral blood flow waxes and wanes up to 10-fold in parallel with body temperature throughout the 8- to 9-month hibernation season without neurologic damage (Frerichs et al, 1994; Lust et al, 1989; Ma et al, 2005). Tolerance to hypoxic and ischemic events in hibernating species may have evolved as a means to endure transitions into and out of torpor (Drew et al, 2004). Hibernating species tolerate hypoxia better than nonhibernating species (D’Alecy et al, 1990; Drew et al, 2004) and experience hypoxemia regularly during emergence from hibernation (Ma et al, 2005). In vitro acute hippocampal slices from hibernating 13-lined ground squirrels tolerate oxygen –glucose deprivation (OGD) better than euthermic squirrels or rat, and as temperature is decreased protection from OGD is also seen in tissue from euthermic animals (Frerichs and Hallenbeck, 1998). In contrast, in short-term culture at 37°C, hippocampal slices from hibernating and euthermic Arctic ground squirrels (AGS, Spermophilus parryii) tolerated modeled ischemia equally well (Ross et al, 2006). In support of this profound in vitro tolerance, euthermic AGS also tolerate 8 mins of global cerebral ischemia in vivo without loss of CA1 neurons (Dave et al, 2006). These results suggest that tolerance to ischemic-like conditions is not wholly dependent on the hibernating state and is, in part, an intrinsic characteristic of the tissue (Frerichs and Hallenbeck, 1998; Ross et al, 2006). This tolerance to in vivo and in vitro hypoxia and ischemia makes AGS an ideal model to identify endogenous molecular mechanisms that contribute to neuroprotection independent of Tb and hibernation state (Ross et al, 2006; Zhou et al, 2001).
Ischemia and subsequent reperfusion initiate a series of molecular responses that culminate in cell death. In hypoxia-sensitive species, loss of ATP occurs within minutes after the onset of ischemia and initiates a cascade of deleterious events including deterioration of ion homeostasis, neuronal depolarization, and excitotoxicity through activation of N-methyl-D-aspartate receptors resulting in excess intracellular Ca2+ (Drew et al, 2004; Lipton, 1999). Subsequent activation of intracellular factors including generation of reactive oxygen species and activation of proteins such as mitogen-activated protein kinases (MAPKs) results in necrotic and apoptotic cell death (Drew et al, 2004).
It has been shown previously in tissue culture models (Herdegen et al, 1997; Xia et al, 1995) and whole animal models (Hirt et al, 2004; Jones and Bergeron, 2004) that extracellular signal-regulated kinase (ERK) 1/2 and Jun NH2-terminal kinase (JNK) can regulate neuronal cell survival in homeothermic mammals. Inhibition of ERK1/2 and concurrent activation of JNK and p38 after nerve growth factor withdrawal are necessary for the induction of cell death (Xia et al, 1995). JNK is activated in cerebral ischemia and mediates excitotoxic and ischemic neuronal death (Hirt et al, 2004; Saporito et al, 2002). However, in addition to its role in promoting cell death, JNK has also been shown to regulate regeneration in response to injury by regulation of neuronal migration, axonal guidance, and cytoskeletal rearrangement (Waetzig et al, 2006).
Currently, it is unclear if hibernating species tolerate ischemic conditions by reducing metabolically demanding processes such as transcription and translation (Frerichs et al, 1998; Morin and Storey, 2006) to avoid energy deficits, or if mechanisms downstream to energy deficits interrupt the cytotoxic cascade observed in ischemia-vulnerable species. Here we begin to address this question by monitoring ATP levels and activation of MAPKs in response to modeled cerebral ischemia. We show that in an acute hippocampal slice model, CA1 neurons from both interbout euthermic AGS (ibeAGS) and hibernating AGS (hAGS) tolerate prolonged periods of OGD even though cellular ATP is more rapidly depleted in ibeAGS. In response to OGD, activation of ERK1/2 occurs only in hAGS whereas baseline JNK activation is maintained in both states even after ATP loss. However, inhibition of JNK increased cell death during normoxia, whereas neither ERK1/2 nor JNK inhibition influenced OGD-induced cell death. Therefore, ERK1/2 and JNK are not essential mediators of survival after OGD, but maintenance of baseline activation may be indicative of improved cellular homeostasis in AGS downstream from ATP loss.
The Institutional Animal Care and Use Committee of the University of Alaska Fairbanks (UAF) approved all procedures. Arctic ground squirrels were trapped on the northern slope of the Brooks Range, Alaska, approximately 20 miles south of the Toolik Field Station of UAF (68°38′ N, 149°38′ W; elevation 809 m) in July 2004 and 2005 under permit from Alaska Department of Fish and Game. On arrival at UAF, AGS were screened for salmonella and quarantined for at least 14 days. All AGS were housed individually in cages at approximately 18°C, fed approximately 40 g of Mazuri Rodent Chow per day, and kept on natural lighting for 64° latitude where the light:dark cycle changes over days from 20:4 to 16:8 h. While housed in environmental chambers, AGS were fed rodent chow ad libitum although animals do not eat when hibernating.
In early fall, AGS were fed 10 to 15 sunflower seeds each day for 2 weeks before being moved to environmental chambers where they were housed at approximately 2°C with a 4:20 h light:dark cycle. All AGS used were trapped between 2 and 3 months of age. Female ibeAGS (total of 8 ibeAGS) and female hAGS (total of 11 hAGS) were seasonally matched, cold adapted, experienced prolonged torpor bouts of regular frequency and comparable duration, and were of similar age and weight. Summer euthermic female AGS (a total of three) had experienced one hibernation season in captivity with regular hibernation bouts. These AGS were moved to 18°C with a 12:12 h light:dark cycle in mid-May and tissue was sampled in July and early August.
Animals were habituated to handling to avoid arousal from hibernation in response to movement from the chambers to the laboratory before euthanasia. Briefly, habituation trials were begun after AGS had been through at least three regular hibernation bouts lasting a minimum of 3 days each and three periods of interbout euthermy lasting approximately 24 h each. Arctic ground squirrels were habituated to handling by being handled in a 2°C chamber and progressing through trials A through C as described below. As respiration rate increases before heart rate or body temperature during arousal (Toien et al, 2001), it was used as a sensitive measure of responsiveness and recorded before and 5 mins and 1 h after handling. The trials were performed as follows: trial A, pick up briefly and return to cage; trial B, move AGS with their cotton nest to sanitized cooler for 5 mins; trial C, move AGS with nest to sanitized cooler and push cooler on a cart for 5 mins. Arctic ground squirrels were considered habituated once in trial C and if the change in respiration rate did not exceed 1 respiration per minute at 5 mins and 1 h after handling for 3 consecutive days.
The ‘shavings added’ method was used to monitor the state of AGS. Shavings were placed on the back of the AGS and checked every 24 h. An AGS was considered ‘torpid’ (hibernating) if the shavings remained on its back or considered to have been through an arousal if the shavings were disturbed or missing. This ‘shavings added’ or ‘sawdust method’ is a reliable indicator of prolonged torpor (Pengelley and Fisher, 1966). Animals in the hAGS group were euthanized after a minimum of 2 days in torpor. Animals in the ibeAGS group were aroused by transfer to 18°C to 21°C 14 to 16 h before euthanasia, after having spent a minimum of 24 h in the current torpor bout.
Hippocampal slices were prepared from juvenile (5- to 11- month-old) ibeAGS and hAGS and from 13- to 14-month-old summer euthermic AGS.
Euthermic animals were anesthetized using 5% halothane (maintained at 1% to 3%) with oxygen at a constant flow rate of 1.5 L/min while rectal body temperature and weight were recorded, and then animals were euthanized by decapitation. Hibernating AGS were euthanized without anesthetization as it is difficult to anesthetize hibernating animals to the same degree as euthermic animals because of the profound differences in respiration rate, pharmacokinetics, and response to anesthetics. Using a sterile technique, the whole brain was removed and placed in ice-cold oxygenated aCSF (120 mmol/L NaCl, 25 mmol/L NaHCO3, 10 mmol/L glucose, 3.3 mmol/L KCl, 1.2 mmol/L NaH2PO4, 2.4 mmol/L MgSO4, 1.8 mmol/L CaCl2 (pH 7.33 ± 0.01)) for 40 secs. Hippocampi were dissected and placed immediately into ice-cold oxygenated aCSF for 20 secs. Coronal hippocampal slices were simultaneously cut at a thickness of 400 μm using an MX-TS tissue slicer (SD Instruments, Grants Pass, OR, USA) and gently transferred to 20mL glass vials using wide-bore pipettes containing fresh aCSF to recover for 1 h at room temperature with constant bubbling with 95% O2/5% CO2.
Hippocampal slices were exposed to OGD by transferring the slices to deoxygenated aCSF (pH 7.32 ± 0.01) lacking glucose that had been previously bubbled with 95% N2/5% CO2 for the time indicated at 37°C. Control slices were transferred to another vial containing oxygenated aCSF and subjected to all subsequent transfers in parallel. The partial pressure of oxygen was monitored continually throughout the OGD insult using a miniature Clark-style electrode (Instech Laboratories, Plymouth Meeting, PA, USA) inserted in the vial. pO2 during OGD was 7.7 ± 2.1mmHg for ibeAGS over 119 ± 1.5 mins and 13.1 ± 2.4mmHg for hAGS over 116 ± 1.8 mins (P>0.05). Alternatively, slices were incubated in 300 μmol/L sodium cyanide (NaCN) and 2 mmol/L iodoacetate (IAA) in deoxygenated aCSF for 2 h at 37°C. For activation of ERK1/2, 200 nmol/L of phorbol dibutyrate (PdBU; BIOMOL, Plymouth Meeting, PA, USA) was added to oxygenated aCSF at 37°C. For inhibition of MAPK activation, 10 μmol/L of the MEK1 inhibitor U0126 (BIOMOL) or 20 μmol/L of the JNK inhibitor SP600125 (Calbiochem, San Diego, CA, USA) was added to the slices during the 1 h slice recovery period after slice preparation and throughout OGD or normoxia. Slices were then incubated for 1 h at 37°C in aCSF constantly bubbled with 95% O2/5% CO2 (pO2 was at least 760mmHg) to mimic reperfusion for cell death analysis or as indicated. Slices were randomly distributed for cell death analysis, homogenized for MAPK analysis, or sonicated for ATP analysis as indicated. As a positive control for detection of cell death by propidium iodide (PI) staining, slices were treated with 300 μmol/L NaCN, 2 mmol/L IAA, and 0.1% Triton X-100 in phosphate-buffered saline (pH 7.6) for 20 mins immediately after slice preparation.
For cell death analysis, triplicate slices were incubated immediately after 1 h recovery in aCSF at 37°C for 20 mins with 2.5 μg/mL Hoechst 33342 (Sigma, St Louis, MO, USA) to identify all cells and with 5 μg/mL PI (Molecular Probes, Eugene, OR, USA) to identify dead and dying cells with permeabilized membranes. Slices were then transferred to fresh aCSF to wash out excess dye. Images were acquired on a Nikon Eclipse TE2000-U inverted microscope (Nikon, Melville, NY, USA) equipped with MetaMorph software (Universal Imaging Corporation, Downingtown, PA, USA). Slices were illuminated with a halogen bulb to obtain bright-field images for identification of CA1 region and with a mercury lamp and a Texas Red filter to obtain PI fluorescent images and a UV filter for Hoescht epifluorescence. Images were acquired for 20 ms. Metamorph software was used to set the threshold intensity to identify PI- and Hoescht-positive cells. The number of PI-positive cells and the number of Hoescht-positive cells were quantified by automated cell counting using Metamorph. Percent cell death was calculated by dividing the number of cells that were positive for Hoescht and PI by the number of cells positive for Hoescht. This method eliminates from the calculation any artifactual staining of cells that were PI-positive but not Hoescht-positive.
Triplicate slices were combined with 200 μL of ice-cold 5% trichloroacetic acid. A probe sonicator was used to disrupt the tissue using 3 to 4 bursts of 3 secs each. The solution was then centrifuged for 1 min at 16,000 g and the supernatant was assayed for ATP activity using the ENLITEN ATP assay system (Promega, Madison, WI, USA) by comparison with a concurrently generated standard curve. The pellet was resuspended in 200 μL of 1 mol/L NaOH and protein concentration was determined using the Bio-Rad Rc protein assay kit (Bio-Rad, Hercules, CA, USA).
Blood ketone and glucose levels were assessed from core blood from the carotid artery immediately after decapitation. Blood was sampled directly onto ketone or glucose test strips and levels were determined using the Precision Xtra Blood Glucose and Ketone Monitoring System (Abbott Laboratories, Abbott Park, IL, USA) according to the manufacturer’s instructions (n = 3 ibeAGS, 4 hAGS).
Three hippocampal slices were homogenized in ice-cold RIPA lysis buffer (50 mmol/L Tris-HCl (pH 7.6), 0.02% sodium azide, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40, 150 mmol/L NaCl, 1 mmol/L phenylmethylsulfonyl fluoride, 1 μg/mL aprotinin, 1 μg/mL antipain, 10 μg/mL leupeptin, 1 mmol/L orthovanadate) using a hand-held homogenizer and left on ice for 20 mins. Insoluble material was removed by centrifugation and protein concentrations were determined using the Bio-Rad Rc protein assay kit. Proteins (10 μg for ERK or 20 μg for JNK and p38) were separated on 10% SDS–polyacrylamide gel electrophoresis gels and transferred to nitrocellulose membranes. The membranes were blocked for 1 to 2 h with 5% (w/v) skim milk powder in TBST (10 mmol/L Tris-HCl, 150 mmol/L NaCl, 0.1% Tween 20) and incubated overnight at 4°C with the appropriate primary antibody. Phosphorylated proteins were detected using antibodies recognizing ERK phosphorylated on Thr202/Tyr204, JNK phosphorylated on Thr183/Tyr185, and p38 phosphorylated on Thr180/Tyr182 (Cell Signaling Technology Inc., Beverly, MA, USA), which have been used previously on AGS protein extracts (Zhu et al, 2005). All antibodies were diluted in TBST. The membranes were washed with TBST and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody (Bio-Rad) for 1 h at room temperature. Immunoreactive bands were visualized using enhanced chemiluminescence (Pierce, Rockford, IL, USA). To reprobe the membranes, bound antibodies were eluted by incubation with 10 mmol/L Tris-HCl (pH 2) and 150 mmol/L NaCl for 30 mins. Equal loading was then confirmed by reprobing the membranes with antibodies recognizing total ERK, JNK, or p38 (Cell Signaling Technology Inc.). To quantitate results, scans of ECL exposures were saved as TIFF files and analyzed using ImageQuant 5.2 software (Molecular Dynamics, Sunnyvale, CA, USA).
Antibody reactivity and specificity to AGS proteins is shown by the observation that protein bands from AGS were of equivalent (and expected) molecular weight as protein bands detected simultaneously from rat protein. In addition, the detected phospho-proteins were of the same molecular weight as the corresponding total protein band.
Data were analyzed with one- or two-way analysis of variance (ANOVA). Significant main effects or interactions were subjected to post hoc analysis (Tukey or t-test where indicated) (SAS software, SAS Institute Inc., Cary, NC, USA). P < 0.05 was considered statistically significant. Data are expressed as mean ± s.e.m.
Previous work in hippocampal cultures from AGS showed approximately 50% cell death of CA1 neurons in normoxic conditions as quantified by total PI intensity (Ross et al, 2006). Therefore, the first goal of this work was to reduce baseline cell death so that high levels of baseline cell death would not obscure further manipulations. As AGS do not hibernate until they are 3 to 5 months of age, we moved to an acute hippocampal slice model that allows analysis of tissue from adult animals (Zhan et al, 2006). Using this acute model, we found baseline cell death to be 21.5% ± 3.3% for ibeAGS and 17.4% ± 2.6% for hAGS (Figure 1A and Supplementary Figure 1, normoxia), as quantified by the number of cells with permeabilized membranes detected by PI staining, an indicator of apoptotic and necrotic cell death.
We found that, in comparison to normoxia treatment, cell death was not significantly increased in response to 30 mins OGD in either ibeAGS or hAGS (Figure 1A). Statistical analysis indicated that there was no interaction between hibernation state and level of treatment but that there was a significant increase in cell death with 2 h OGD when ibeAGS and hAGS groups were combined (Figure 1B). Regression analysis showed no correlation between the number of days spent in torpor immediately before euthanasia and sensitivity to 2 h OGD in hAGS (R2 = 0.099, P = 0.13).
The tolerance of ibeAGS to OGD may be because of the preconditioning effects of hypothermia experienced during previous hibernation bouts (Nishio et al, 2000). Hypothermic preconditioning has been reported to last less than 7 days in rats (Nishio et al, 2000); therefore, we tested hippocampal tissue collected from AGS during the summer months (July–August) that had not been torpid for at least 4 months and should no longer be affected by hypothermic preconditioning. We found that baseline cell death increased to 50.9% ± 5.8% (P < 0.001 versus ibeAGS or hAGS), but similar to the winter animals, there was no increase in cell death in response to 30 mins OGD (Figure 1C), suggesting that tolerance to OGD is not because of preconditioning effects of recent low body temperatures. Body weights in this group of summer animals indicated that preparation had begun for the hibernation season (1012 ± 72 g for summer euthermic AGS versus 541 ± 18 g for winter AGS, P < 0.01). Interestingly, preparation for hibernation was not sufficient to protect the slices from trauma associated with slice preparation, evident from the high baseline cell death in this group of animals.
We next proposed that the CA1 hippocampus was using intracellular glycogen or making more efficient use of oxygen in oxidative phosphorylation to survive OGD. Therefore, NaCN, an inhibitor of complex IV of the electron transport chain, and IAA, an inhibitor of the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase, were added to slices during the 2 h OGD to chemically block glycolysis and oxidative phosphorylation. Addition of these chemicals did not affect the response of CA1 neurons to OGD, indicating that these cells were not relying on glycogen or oxidative phosphorylation for survival (Figures 1A and 1B; P < 0.05 versus normoxia). A separate group of slices was treated with the detergent Triton X-100 in a phosphate-buffered saline solution containing NaCN and IAA. Triton X-100 permeabilized the membranes and served as a positive control for PI fluorescence.
The brain preferentially uses glucose as its energy source, but in the absence of glucose, ketone bodies are metabolized from fats to provide energy. We found that hAGS have significantly higher levels of total blood ketones than ibeAGS (1.8 ± 0.5 versus 0.3 ± 0.09 mmol/L, n = 3 ibeAGS, 4 hAGS, P < 0.05), whereas ibeAGS have slightly higher levels of blood glucose than hAGS (124 ± 4.2 versus 84 ± 22 mg/dL, n = 3 ibeAGS, 4 hAGS, P = 0.09). The difference in blood glucose reported here is consistent with prior reports from this laboratory of a statistically significant decrease in blood glucose from 180 to 91 mg/dL on entrance into torpor (Osborne et al, 1999).
As the first effect of ischemia is loss of energy stores, we suggested that AGS might maintain their energy balance as a mechanism to improve survival. We analyzed levels of ATP in slices as a measure of cellular energy and found that ATP declined without a significant increase in cell death in response to OGD. Slices from hAGS lost ATP more slowly than slices from ibeAGS, consistent with metabolic suppression in torpor. ATP declined in slices from ibeAGS within 3 mins of OGD, whereas the decrease in ATP in slices from hAGS was not significant until after 10 mins of OGD (Figure 2). Addition of NaCN and IAA did not induce a greater decrease in ATP compared to 2 h OGD, arguing against a role for glycogen or oxidative phosphorylation in maintaining ATP and suggesting that OGD alone had a maximal effect on reduction of ATP levels.
Activation of prosurvival proteins and inhibition of prodeath proteins might be a mechanism used by AGS to improve survival in response to energetic challenge. To determine if ibeAGS and hAGS use differential regulation of MAPKs in response to OGD to improve survival, we first determined the levels of the activated ERK1/2, JNK, and p38 MAPKs immediately after slice preparation (Figure 3). Because phosphorylation of MAPKs is required for activation, we monitored phosphorylation on the Thr-Xxx-Tyr activation site, commonly used as a marker of MAPK activation (Nozaki et al, 2001; Zhu et al, 2005). We found that the levels of phosphorylated ERK1/2 were variable. Therefore, although the mean level of phospho-ERK1/2 was higher in hAGS, it was not statistically significant (Figures 3A and 3D). In contrast, the levels of phospho-JNK were significantly (P < 0.05) decreased in hAGS compared with ibeAGS (Figures 3B and 3D), consistent with previously published reports (Zhu et al, 2005). The levels of phospho-p38 were the same in ibeAGS and hAGS (Figures 3C and 3D) and as they did not change with OGD (data not shown), p38 was not examined further. The relative levels of these proteins were the same when the CA1 region was microdissected and analyzed immediately after slice preparation (data not shown); however, as this method did not allow for accurate time-course data, it was not used in further experiments.
To examine the role of ERK1/2 in tolerance to OGD, we analyzed the activation of ERK1/2 in response to OGD. Levels of phospho-ERK1/2 after OGD treatment were compared to those in normoxic control slices that had been exposed to normoxic conditions for the entire treatment time, where phospho-ERK1/2 levels had stabilized (between-animal variance immediately after slice preparation versus normoxia controls declined from 0.6 to 0.1 in hAGS and from 2.0 to 0.1 in ibeAGS). We found that in ibeAGS, phospho-ERK1/2 levels were maintained with 3 mins of OGD and then decreased markedly by 10 mins of OGD and remained suppressed (Figures 4A and 4C). Surprisingly, the potent activator of ERK1/2, PdBU, did not increase ERK1/2 phosphorylation in ibeAGS, suggesting that this classic mechanism for ERK1/2 activation may not be functional in ibeAGS. In contrast, phospho-ERK1/2 levels in hAGS increased with 3 mins of OGD and did not decrease below normoxic levels until 30 mins of OGD (Figures 4B and 4C). Hibernating AGS showed significant increases in phospho-ERK1/2 levels with PdBU treatment, indicating that this ERK1/2 activator can function as expected in AGS. Phosphorylated ERK1/2 was not upregulated in response to modeled reperfusion after prolonged OGD in hAGS or ibeAGS. Thus, ERK1/2 is upregulated in response to PdBU and OGD and remains active for longer after OGD in hAGS compared to ibeAGS.
To determine if regulation of JNK could promote cell survival, activation of JNK in response to OGD was analyzed. We found that both ibeAGS and hAGS sustained baseline levels of activated JNK through 3 and 10 mins of OGD (Figure 5, State P>0.05, Treatment P < 0.05). Arctic ground squirrels did not recover baseline levels of phospho-JNK after prolonged OGD or after 2 h OGD plus 1 h modeled reperfusion. Thus, maintenance of baseline JNK levels in AGS without rapid upregulation of this potentially prodeath signal at the onset of OGD may promote cell survival in either state.
As the activation states of both ERK1/2 and JNK are altered in response to OGD, we were interested in determining whether the presence of activated ERK1/2 or JNK activity is necessary for the intrinsic neuroprotection exhibited by AGS. Therefore, immediately after slice preparation and throughout OGD, we exposed hippocampal slices to chemical inhibitors that blocked activation of ERK1/2 and JNK before OGD. We found that treatment of slices with the MEK inhibitor U0126, which markedly inhibited ERK phosphorylation (Figure 6A), had no effect on survival of CA1 neurons in either ibeAGS or hAGS (Figure 6B). In contrast, treatment with the JNK inhibitor SP600125 increased cell death in the normoxic condition, but had no effect on cell survival in response to OGD regardless of the state of the animals (main effect of treatment, P < 0.05; Figure 7). As the JNK inhibitor SP60125 directly inhibits the kinase activity of all three JNK isoforms but not phosphorylation of JNK itself (Bennett et al, 2001), we were unable to confirm the efficacy of this inhibitor using analysis of JNK phosphorylation, and analysis of JNK substrates, ATF-2 and c-Jun, led to inconclusive results likely because of nonreactivity of the blotting antibodies to AGS protein (data not shown). Analysis of other JNK substrates, such as ELK-1 or Bcl-2, is difficult, as these can be phosphorylated by other kinases. However, we can confirm that SP60125 treatment did not block phosphorylation of other MAPKs and the unrelated kinase, Akt (data not shown). Nevertheless, these data do suggest that JNK normally promotes cell survival in CA1 neurons in AGS after trauma associated with slice preparation.
Using an acute hippocampal slice method, we found that tissue recently obtained from both ibeAGS and hAGS is remarkably resistant to OGD. We found that CA1 neurons from AGS, independent of hibernation state, show a significant increase in cell death with prolonged (2 h) but not shorter duration (30 mins) OGD as quantified by PI-positive neurons. This tolerance to OGD is maintained in spite of the depletion of ATP stores. We found that, although MAPK proteins are regulated in a manner consistent with promoting survival in AGS, particularly in hAGS, neither ERK1/2 nor JNK MAPKs are necessary for protection from OGD. In contrast, baseline JNK activation is necessary for survival after trauma of slice preparation. Rather than essential mediators of survival after OGD, maintenance of ERK1/2 and JNK activation may therefore be signatures of improved cellular homeostasis.
Tolerance to OGD in slices from both stages of the hibernation cycle (hibernating and interbout euthermic) at 37°C is consistent with previous observations made on the first day of culture in an organotypic-like hippocampal slice preparation (Ross et al, 2006). The partial pressure of oxygen was slightly higher during OGD in slices from hAGS than from ibeAGS and may have contributed to the tendency of slices from hAGS to tolerate OGD better than slices from ibeAGS, but this difference was eliminated in slices exposed to OGD combined with NaCN and IAA. Induced arousals, such as experienced by ibeAGS in this study, have an accelerated increase in metabolic rate and oxygen consumption compared to natural arousals (Tahti and Soivio, 1978) even though the Tb, heart rate, and respiratory rates have similar final values. The tolerance observed here is consistent with the previous study where ibeAGS had experienced natural arousals (Ross et al, 2006), suggesting that the conditions experienced during an induced arousal do not alter neuroprotective adaptations. Recent, repeated bouts of torpor were not necessary for OGD tolerance in summer euthermic AGS, suggesting that low body temperatures experienced during torpor do not augment OGD tolerance through hypoxic preconditioning. In other tissues, a seasonal effect of resistance to ischemic damage has been shown where summer animals show increased vulnerability to insult (Kurtz et al, 2006; Lindell et al, 2005). Further analysis of AGS sampled earlier in the summer season will be necessary to determine if seasonal factors associated with preparation for hibernation promote tolerance to OGD. Increased vulnerability of slices from summer animals to the trauma of slice preparation suggests that preparation for the hibernation season was not associated with tolerance to mechanical trauma. Moreover, the ability of the JNK inhibitor to increase baseline cell death in control slices suggests that different mechanisms contribute to tolerance to trauma and OGD and that JNK activation is involved in the former, but not the latter.
The long-term hippocampal slice model previously used in our lab produced high baseline cell death. The acute hippocampal slice model used here decreased baseline cell death in the CA1 region to approximately 20%, similar to what has been published previously for neonatal and adult rat tissue using PI staining (Xia et al, 1995; Zhan et al, 2006). This allowed us to manipulate extracellular conditions with the confidence that high baseline cell death would not obscure our findings. One caveat is that as staining with PI detects increased membrane permeability, we may not be able to detect cells that are in very early stages of cell death. Nevertheless, we found that, similar to the long-term slice model, both hAGS and ibeAGS show similar tolerance to hypoxia and aglycemia at 37°C. This is in contrast to the 13-lined ground squirrel where tolerance to OGD at 36°C was observed only in the hibernating state (Frerichs and Hallenbeck, 1998), suggesting AGS maintain their protective phenotype irrespective of state.
Many organisms tolerant of low oxygen levels possess large stores of glycogen and pH buffering mechanisms that fuel and protect against pH shifts resulting from anaerobic glycolysis (Jackson, 2004; Lutz and Milton, 2004). Intrinsic ischemia tolerance cannot be explained by enhanced peripheral stores of glycogen because glucose derived from glycogen will not reach ischemia tissue when blood flow is diminished. Nonetheless, astrocytes, and in some cases neurons, can store significant amounts of glycogen. Moreover, an increase in the brain glycogen content during hibernation may occur as observed in hamsters (Lust et al, 1989). Alternatively, use of more efficient fuels such as lipids and ketone bodies or more efficient coupling of oxidative phosphorylation to ATP synthesis might provide the energy needed for survival. To address these hypotheses, we treated slices with IAA and NaCN, inhibitors of glycolysis and oxidative phosphorylation, respectively. Resistance of slices to treatment with IAA indicates that slices are not using hippocampal glycogen to provide energy. In addition, resistance to NaCN treatment indicates that lipids, ketones, or glucose are not contributing to energy production by oxidative phosphorylation.
Although hippocampal slices in vitro did not appear to use glucose for energy, we were interested in determining which energy source is available in the torpid animal. Consistent with previous hibernation studies (Krilowicz, 1985; Rauch and Behrisch, 1981), we found that hAGS have significantly higher levels of blood ketones than ibeAGS. These data suggest that there is an increase in fatty acid oxidation rates during torpor bouts and the resulting build-up of ketone bodies is then rapidly processed during the first 24 h of an arousal bout. The increase in ketone bodies suggests that ketosis is an important energy source during torpor and arousal throughout the hibernation season (Muleme et al, 2006). The contribution of ketones as an energy source for hippocampal slices in vitro has not been determined in AGS.
Loss of ATP indicates that AGS hippocampus tolerates and does not avoid disruption in metabolic supply and demand. Even though ATP levels are maintained slightly longer in hAGS, 10 mins of OGD results in a significant decrease in ATP levels in both ibeAGS and hAGS; however, this drastic loss of ATP does not result in cell death. Loss of ATP is an initiator of cell death and subsequent inhibition of the Na+K+ ATPase results in loss of membrane potential and influx of Ca2+, resulting in cell death (Lipton, 1999). As cell death was measured in vitro at 37°C, the resistance to ATP loss is not because of to temperature effects of torpor.
Slower decline of ATP in slices from hAGS suggests that slices retain some degree of metabolic suppression characteristic of torpor even at 37°C. Previous evidence has suggested that AGS slices suppress ATP-consuming processes in vitro such as the Na+K+ ATPase pump (Ross et al, 2006). Alternatively, hibernating brain may contain higher levels of high-energy phosphates, such as phosphocreatine, which has been shown to promote neuronal survival in response to ischemia (Balestrino et al, 2002; Lensman et al, 2006). Brains of hibernating hamsters contain higher concentrations of phosphocreatine than euthermic hamsters (Lust et al, 1989), but the level of phosphocreatine in AGS brain is currently unknown.
In response to OGD, hAGS had increased followed by sustained baseline ERK1/2 activation in contrast to ibeAGS. This suggests that hAGS may be able to activate and maintain activation of prosurvival targets of ERK1/2 for longer duration than ibeAGS. Surprisingly, ibeAGS were unable to increase the levels of activated ERK1/2 in response to OGD or in response to the positive control, PdBU. Phorbol esters, such as PdBU, cause activation of the ERK1/2 pathway by activation of the Ras activator RasGRP. It will be important to determine if ibeAGS neurons uncouple ERK1/2 activation from OGD and PdBU treatment by downregulation of RasGRP, as shown in other cell types (Han et al, 2006). Alternatively, ERK1/2 may be maximally activated in ibeAGS in response to trauma, so OGD or PdBU treatment would not increase phosphorylation further.
To determine if ERK1/2 activation or sustained baseline activation of ERK1/2 promoted cell survival in AGS, we treated slices with the ERK1/2 inhibitor U0126. We found that inhibition of the ERK1/2 pathway did not affect OGD-induced cell death, indicating that ERK1/2 activation is not necessary for the observed tolerance to OGD. Thus, although ERK1/2 does not regulate cell death directly, it may be indicative of overall cellular homeostasis and signaling capability.
Baseline levels of JNK were sustained in both hAGS and ibeAGS in spite of ATP loss. Surprisingly, we found that inhibition of the JNK pathway did not affect OGD-induced cell death but that a baseline level of JNK activation was necessary for survival during normoxia after trauma from slice preparation, indicating that JNK is functioning in a prosurvival manner. This may be because of the physiologic role of JNK in regeneration and cytoskeletal integrity (Waetzig et al, 2006). Generation of hippocampal slices necessitates cutting coronal sections of hippocampus, resulting in physical damage to neurons that extend along the length. Therefore, sustained activation of JNK, necessary for the regeneration of the cut axons (Herdegen et al, 1998; Kenney and Kocsis, 1998; Lindwall and Kanje, 2005), is likely promoting cell survival in response to physical injury.
In summary, we have shown that CA1 neurons from both ibeAGS and hAGS are tolerant to prolonged periods of OGD. This tolerance persists despite drastic declines in ATP levels after OGD. We found that although inhibition of ERK1/2 or JNK activity during OGD did not affect neuronal survival, baseline JNK activation was necessary for survival in response to the trauma of slice preparation. Therefore, tolerance to OGD occurs downstream of ATP depletion and potentially involves improved maintenance of cellular homeostasis, as suggested by prolonged baseline levels of ERK1/2 and JNK after OGD.
This work was supported by NIH-NS41069, funded in part by NINDS, NIMH, NCRR, and NCMHD, by US Army Med. Res. and Materiel Command 05178001, by US Army Research Office #W911NF-05-1-0280, and by a post-doctoral fellowship to SLC from the Natural Sciences and Engineering Research Council of Canada (NSERC)
The authors thank Drs Barbara Taylor and Michael Harris for helpful discussion and technical assistance and Ms Jeanette Moore for technical assistance. The authors declare no conflict of interest.