Effects of Hyperoxia on Brain Tissue Oxygen Tension in Non-Sedated, Non- Anesthetized Arctic Ground Squirrels: An Animal Model of Hyperoxic Stress

Arctic Ground Squirrels (AGS) are classic hibernators known for their tolerance to hypoxia. AGS have been studied as a model of hypoxia with potential as a medical research model. Problem statement: Their unique resistance to the stressors of low oxygen led us to hypothesize that AGS might also be adaptable to hyperoxia. Approach: This study examined the physiological pattern associated with hyperoxia in response to brain tissue oxygen partial pressure (PtO2), brain temperature (Tbrain), global oxygen consumption (VO2) and respiratory frequency (fR) using non-sedated and nonanesthetized Arctic Ground Squirrels (AGS) and rats. Results: We found that 1) 100% inspired oxygen (FiO2) increased the baseline values of brain PtO2 significantly in both summer euthermic AGS (24.4 ± 3.6-87.3 ± 3.6 mmHg, n=6) and in rats (18.2 ± 5.2-73.3 ± 5.2 mmHg, n = 3); PtO2 was significantly higher in AGS than in rats during hyperoxic exposure; 2) hyperoxic exposure had no effect on brain temperature in either AGS or rats, with the brain temperatures maintaining constancy before, during and after 100% O2 exposure; 3) systemic metabolic rates increased significantly during hyperoxic exposure in both euthermic AGS and rats; moreover, VO2 were significantly lower in AGS than in rats during hyperoxic exposure; 4) the respiratory rates for rats were maintained before, during and after 100% O2 exposure, while the respiratory responding patterns to hyperoxic exposure changed after exposure in AGS. AGS fR was significantly lower after hyperoxic exposure than before the exposure. Conclusion: These results suggest that hyperoxic ventilation induced PtO2 and VO2 differences between AGS and rats and led to altered respiratory patterns between these species. AGS and the rat serves as an excellent comparative model for hypoxic and hyperoxic stress studies of the brain.


INTRODUCTION
One of the unique polar mammals encountered in an extreme environment is the Arctic ground squirrel (Buck and Barnes, 2000). Its unique metabolism during hibernation is a topic that is just beginning to be understood and has implications for neuroscience (Frerichs, 1999;Ma and Wu, 2008) and disease processes. Hypoxia in the AGS has been the focus of most studies because of its potential to answer questions related to reperfusion injury and neurodegeneration in Alzheimer's Disease, Parkinson's Disease and brain tumors. Hypoxia occurs in relation to certain tumors, promoting angiogenesis, metastasis and resistance to therapy. Pathways involved in hypoxia have profound implications in cancer prognosis and treatment. Hypoxic cells temporarily arrest in the cell cycle, reduce energy consumption and secrete survival and pro-angiogenic factors. On the other hand, hyperoxia has not been studied in this classic polar hibernator.
The inhalation of pure oxygen is often used in medical treatment and activities such as scuba diving and high-altitude flight. However, inhaling pure O 2 even at normal barometric pressure (~760 mmHg) creates unusual stresses within an organism. It is well known that oxygen can have a toxic impact on tissues under conditions of hyperoxia.
The Arctic Ground Squirrel (AGS), a polar hibernating species, is able to endure extreme changes in endogenous O 2 under different natural physiological states. During hibernation, AGS blood is well oxygenated with similar blood oxygen partial pressure (P a O 2 ) as compared to rat brains (Ma et al., 2005), but AGS brain tissue oxygen partial pressure (P t O 2 ) is hypoxic (Ma and Wu, 2008). When aroused from hibernation, AGS experiences extreme blood hypoxia (~ 9 mm Hg), but there is no evidence of cellular stress, inflammatory response, neuronal pathology, or oxidative modification in brain following the period of high metabolic demand necessary for arousal (Ma et al., 2005). While in the euthermic state, AGS have low P a O 2 and low hemoglobin oxygen saturation (sO 2 ) in contrast to non-hibernating rats, indicating a natural mild, chronic hypoxia in AGS blood (Ma et al., 2005). However, the P t O 2 in the euthermic state is higher than in the hibernating state (Ma and Wu, 2008;Ma et al., 2009).
Hypoxic blood gases and normal P t O 2 in euthermic AGS may precondition this species, help them cope with environmental hypoxia and maintain normal brain function.
When euthermic AGS are exposed to 8% O 2 inhalation under non-sedated, non-anesthetized natural conditions, P t O 2 declines quickly from normal values to very low levels (5 mmHg). P t O 2 returns to the normal level after hypoxic exposure ends and ambient air inhalation resumes, indicating that AGS brain tolerates hypoxic conditions (Ma et al., 2009). When both euthermic AGS and rats are exposed to 100% O 2 inhalation, sO 2 in rats is not affected by 100% O 2 inhalation; however, sO 2 is significantly increased in AGS blood, but remains at a lower level than in rats (Ma et al., 2005).
To our knowledge, there have been no investigations demonstrating the response and tolerance differences in P t O 2 to hyperoxia between hibernating and non-hibernating species. We hypothesize that AGS have a greater tolerance to hyperoxia in the brain than other terrestrial mammals such as the lab rat. This study was designed to examine the physiological mechanism associated with hyperoxia in response to brain tissue oxygen partial pressure (P t O 2 ), brain temperature (T brain ), systemic oxygen consumption (V O2 ) and respiratory frequency (f R ) using real time in-vivo measurements of brain tissue oxygenation with "Clarktype" O 2 electrodes in non-sedated and nonanesthetized, free-moving AGS and rats.

MATERIALS AND METHODS
Animals: All procedures were performed in accordance with and approved by University of Alaska Fairbanks Institutional Animal Care and Use (IAUC) guidelines. AGS and Sprague Dawley rats were used for these experiments. Adult AGS of both sexes were trapped during mid-July in the northern foothills of the Brooks Range, Alaska, approximately 40 miles south of the Toolik Field Station (68°38'N, 149°38'W; elevation 809 m) of the University of Alaska Fairbanks (UAF) and transported to Fairbanks under permit from the Alaska Department of Fish and Game. AGS were housed individually at 16-18°C and fed rodent chow, sunflower seeds, fresh carrots and apples ad libitum until mid-September when they were moved to a cold chamber set to an ambient temperature (Ta) of 2°C and 4:20 h light: dark cycle. Sprague Dawley rats (6-7 months of age 5 at time of experiment) were purchased from Simonsen Laboratories (Gilroy, CA) and transported by air to UAF. Rats were housed at 20-21°C, 12:12 h light: dark cycle and fed rodent chow ad libitum.
Surgery for brain guide cannula implantation and abdominal cannulation: The animals used in this research were both male and female. A total of 6 summer euthermic AGS (body mass: 889.2 ± 30.2 g) and 3 rats (body mass: 479.5 ± 20.7 g) were used in stereotaxic surgery for brain guide cannula implantation. The surgeries for AGS and rats were conducted between March and April. Animals fasted at least 12 hours before the surgery. Surgery was performed at room temperature under general anesthesia with isoflurane (Halocarbon Laboratories, River Edge, NJ), induced at 5% and maintained at 1.5-3% mixed with 100% medical grade O 2 at a flow rate of 1.5 L /min. During surgery, the animals' heads were put in a stereotaxic apparatus and their body temperatures were maintained at 35-37°C with a servo-controlled fluid-filled heating pad (Omni Medical Equipment Inc., Cincinnati, O H ).
Under strict aseptic conditions, a skin incision was made midline on the head and the working area was exposed with a retractor. The soft tissues were scraped with a surgical blade and cleaned with sterilized cottontipped applicators. After tissue cleaning on the skull, two holes (1.8-mm diameter) to accommodate the guide cannulae were made with a sterilized trephine operated by a battery-driven drill at the coordinates (AP, 13.5 or 14 mm; L, ±3.25 mm; D, -4.0 mm) for AGS and another four holes were made with a steel burr (0.5-mm diameter) near the cannulae holes (two in front of and two in back of the guide cannula holes) on the skull for the foundation of dental cement. After completing the drilling process, four anchor stainless steel bone screws (BAS, West Lafayette, IN) were driven into the skull through the screw holes. Two guide cannulae (CMA11, Acton, MA) were stereotaxically positioned above the right and left striatum (AP, 13.5 or 14 mm; L, ±3.25 mm; D, -4.0 mm), as previously described in detail (Osborne et al., 1999;Zou et al., 2002;Ma and Wu, 2008) and were slowly lowered 4 mm from the cortical surface. The surgical procedures for rats were similar to those for AGS. Guide cannulae for rats were stereotaxically positioned above the right and left striatum (A, +0.6 from bregma; L, ±3.5; v, 5.5) with the incisor bar set at -2.4 mm as previously described (Drew et al., 1990). Cannulae were secured to the screws with dental cement.
Antibiotics (Baytril; Bayer Corp., Shawnee Mission, KS, dose: 5 mg kg −1 ) were given twice a day, 1 day before surgery and 2 days after surgery, by subcutaneous injection in the back of the neck. After surgery, animals were allowed to recover completely at room temperature and daily cleanings were performed with 3% betadine for 10-15 days. Following recovery, AGS were housed individually in a cool chamber at 2°C under a light regime of 12:12 hours light: dark; rats were maintained in the same condition as during postoperative recovery.
Calibration for PO 2 electrode and thermocouple and temperature correction for P t O 2 : Calibration of PO 2 electrode and thermocouple. For the measurement of P t O 2 , a PO 2 microelectrode (Model: IPS-020; 200 μm; Inter Medical Co. Ltd, Japan) was calibrated in artificial cerebral spinal fluid (ACSF in mM: NaCl:124; Cl: 2.7; CaCl 2 :1.2; MgCl 2 : 0.85; D-Glucose:1.4; NaHCO 3 :24, pH 7.4) at 37 o C equilibrated with air (21% O 2 ) and N 2 (0% O 2 ), respectively, before and after each experiment to ensure no obvious change (less than 10%). ACSF solution in a water-circulating double layer glass beaker (15 mL) was thermostatically maintained at a temperature of 37° C by circulating an anti-freezing solution through plastic tubing connected to a circulating bath, which could change the temperature of the circulating fluids to any given temperature from -2 to 40°C. Compressed air containing 21% O 2 was used to set up a standard PO 2 value at a reading of 145-150 mm Hg, depending on the value of atmospheric pressure during calibration and pure compressed nitrogen gas was used to set up a standard PO 2 value as 0 mm Hg on the PO 2 monitor (IMP-201, Inter Medical Co. Ltd, Japan). The needle thermocouples were calibrated at 0 and 40°C with an accuracy of ± 0.1°C in a water circulating bath before gas sterilization.
Temperature correction for PO 2 measurement: Because brain temperature varies at any given time and can also change by exposure to different inhalation gases, P t O 2 measurement had to be corrected to obtain the real physiological effect. The details of the principle and the procedures of temperature correction for P t O 2 measurement was described in Ma and Wu, 2008. Briefly, to perform such a temperature correction, the following steps were taken: • The PO 2 meter was initially calibrated to a PO 2 value of 144-150 mm Hg (the value depended on the local atmospheric pressure during calibration) when the electrode was exposed to ACSF equilibrated with compressed air (21% O 2 ) at 37°C and to a PO 2 value of zero when ACSF was equilibrated with compressed nitrogen gas (0% O 2 ) at 37°C • Under bubbling conditions of 21% O 2 in ACSF solutions ranging in temperature from 0-40°C , the signal outputs of the PO 2 electrode (P O2 , meas .) were measured.
• The actual values of PO 2 (P O2 , cal .), independent of temperature effects, were calculated at each given temperature according to the equation: • where the values of water vapor pressure (P H2O ) (Weast, 1972), were appropriate for a given temperature • These data-P O2, meas ., P O2, cal . and temperature in Celsius (t)-were used to derive a temperature correction factor (f = A e -k t, where the values of A and k are determined by each electrode calibration) to calculate the actual PO 2 of tissues measured at various temperatures: f = P O2 , cal ./ P O2, meas . =A e -k t "P O2, cal ./ P O2, meas ." were plotted against t at temperatures of 0 to 40° C and the curve function fitted with an Excel program to find the values of A and k. After A and k values were defined, the function, P O2, cal . = (A e -k t )(P O2, meas .), was programmed into the data acquisition program (PowerLab) for online temperature correction of PO 2 measurements.
For each electrode and experiment, the calibration process was repeated as above.

Real-time measurement of P t O 2 and T brain in the striatum of AGS in conjunction with V O2 and f R :
After animals recovered from stereotaxic surgery for brain guide cannulae implantation, the insertions of the reference electrode, the PO 2 electrode and the microthermocouples were conducted under anesthesia with 2-3% isoflurane at room temperature. A sterilized silversilver chloride electrode (RC1, WPI, Sarasota, FL) was implanted to serve as a reference electrode under dorsal skin on the back through a soft silicone tube; a general anesthetic procedure with isoflurane gas was performed before insertion of the PO 2 electrode and microthermocouple. After implantation of the reference electrode, the calibrated and sterilized PO 2 electrode and needle micro-thermocouple for T brain were inserted through the pre-implanted guide cannulae into the left and right striatum, respectively and secured by medical tape. The tips of the PO 2 electrode and needle microthermocouple extended about 2 mm beyond the guide cannulae tubes. Buprenophine (0.03 mg kg −1 ), a pain reliever, was given by subcutaneous injection on the 8 back, with one single dose after insertions of the PO 2 electrode and thermocouple for both AGS and rats. Then for real-time in vivo recording of measurements, the animals were transferred into an experimental chamber with a normal swivel set-up for active animals. The wires connected from PO 2 electrode and reference electrode to PO 2 meter and from thermocouple to thermocouple meter went through a sprint tube and came out the experimental chamber in the middle of the top.
The systemic metabolic rate (oxygen consumption, V O2 ) of AGS and rats was recorded by indirect calorimetry in a vertical cylindrical-shaped Plexiglas metabolic chamber (I.D. 29 cm, height 32 cm) previously described (Ma and Wu, 2008). Air was drawn into a distribution tube through a vertical cylindrical-shaped Plexiglas metabolic chamber at a flow rate of 3 L/min by a membrane pump and measured by a mass flow meter (model AFSC-10 K, Teldyne Hastings-Raydist, Hampton, VA). A separate pump sampled gas from the distribution tube through a canister, which was filled with molecular sieves and connected in a series with a Nafion drying column. O 2 extraction was measured with an O 2 analyzer (Model: FoxBox 2.0, Sable System International, Inc, Las Vegas, NV), which was calibrated with air. O 2 was calculated with an Excel program on the basis of the O 2 fraction differences between fresh air (before being used by an animal) and expired air (after being used by an animal), flow rate of sampling air, animal body weight and a known respiratory quotient value from previous experiments with AGS (Tøien et al, 2001;Ma et al, 2005), according to Withers principles of equation (Withers, 1977). The respiratory quotient value of 0.74 was used for calculation of Sprague Dawley rat RQ cited from the literature (Strohl et al., 1997).
The P t O 2 , T brain and V O2 were recorded simultaneously 20 minutes after buprenophine injection by a data acquisition program matched with the interface (Power Lab/8 sp, AD Instruments, Inc., Colorado Springs, CO) at a sample rate of 1/min and without further signal filtering. A PO 2 monitor and a thermocouple meter (Sable Systems, TC-1000 thermometer, Henderson, NV) were used as preamplifiers and the outputs of those two meters were connected to the interface of the data acquisition program. A Sable interface (model: U12, Henderson, NV) was used as A/D converter for the V O2 measurement system. The output signals were then input into the interface of the data acquisition program. All experiments for P t O 2 , T brain and V O2 measurement were conducted in the experimental chamber at 18°C. Data in P O2, meas . recorded originally by the PO 2 electrode in one channel were calibrated online into corresponding values of P O2, cal. without the effects of temperature in another channel.
Respiratory frequency (f R ) was counted manually by direct observation during the entire experimental period with 30-minute intervals beginning 60 min before 100% O 2 exposure and ending 120 minutes after hyperoxic exposure. Counting for f R was blindly performed by a technician.
Hyperoxic exposure: When the recorded values in P t O 2 , T brain and V O2 were stable for at least 1 hour in fresh air, the air was replaced with 100% O 2 . Hyperoxic gas was delivered into the experimental chamber for 3 hours, followed by a return to ambient air. P t O 2 , T brain and V O2 were continuously recorded for an additional 2 hours after the return to air. f R was measured throughout the experimental period. The O 2 percentages in the experimental chamber were monitored with an O 2 analyzer before, during and after 100% O 2 delivery. The mean values of air O 2 fraction inspired by animals before and after hyperoxic exposure in the experimental chamber were 20.91 ± 0.06 (n=9) and 21.20 ± 0.08 (n = 9) mmHg, respectively. The animal inspired O 2 fraction during hyperoxic exposure in the chamber was 100.01± 0.31 mmHg (n = 9). After P t O 2 , T brain, f R and V O2 recordings were completed, the animals were euthanized with isoflurane under deep anesthetic condition. The brain tissues were immediately frozen in 2-methylbutane at -40°C and kept in a freezer at -80°C. Histologic analysis of the brains using the Cresyl Violet method to identify the PO 2 electrode tip location was performed for each animal, as previously described in detail (Ma and Wu, 2008).

Data analysis:
Values in this study are given as mean ± SEM. Data shown in Fig. 1-4 for P t O 2 , V O2 , T brain and f R before, during and after 100% O 2 exposure were averaged to represent the value in each of those three experimental periods. Data for P t O 2 , V O2 , T brain and f R shown in the bar graphs of Fig. 1-3 were analyzed with two-way ANOVA (SigmaStat, version 3.5, Systat Software, Inc. Chicago, IL) using two factors: animal groups (AGS and rats) and 10 experimental conditions (before, during and after 100% O 2 exposure). Significant differences (p<0.05) were determined by means of All Pairwise Multiple Comparison Procedure with Holm-Sidak's method. P values <0.05 were taken to represent significant differences. The horizontal lines above the tops of the bars in Fig. 1-4 represent significant differences either between the two groups or between the two experimental treatments.

RESULTS
The response of P t O 2 to 100% O 2 in the brain: Figure 1, panel A shows the real-time changes in the mean value of P t O 2 from six AGS and three rats before, during and after 100% O 2 exposure; panel B shows the mean values of P t O 2 of AGS and rats in these periods. The baseline value of P t O 2 (24.38 ± 1.91 mm Hg; n = 6) in summer euthermic AGS did not significantly differ from rats (18.19 ± 0.60 mm Hg; n = 3) before 100% O 2 exposure. The P t O 2 in both AGS and rats increased significantly, (AGS: 87.25 ± 6.17 mm Hg, n = 6; p< 0.001) and (Rats: 73.30 ± 2.24 mm Hg, n = 3; p< 0.001), during 100% O 2 exposure. P t O 2 increase in AGS was greater than in rats during 100% O 2 exposure (p = 0.038). The P t O 2 recovered to the levels of (26.91 ± 3.53, n = 6) in AGS and (25.64 ± 1.82 mmHg, n = 3) in rats. For both AGS and rats, there were no significant differences in the groups' P t O 2 before and after 100% O 2 exposure.

Response of systemic oxygen consumption to 100%
O 2 : Figure 2, Panel A shows mean real-time changes in V O2 before, during and after 100% O 2 exposure. In Panel A, V O2 data are missing for the first 30-minute interval after the beginning of 100% O 2 exposure and for the 30 min interval immediately after the end of exposure. Precise V O2 values could not be determined during these two transition periods because of the unavailability of a stable O 2 standard in the respiratory system for online calculation of V O2 . Panel B shows the averaged values in V O2 during the periods before, during and after 100% O 2 exposure in AGS and rats. Before 100% O 2 exposure the baseline value of V O2 of 11 AGS (1.43 ± 0.09 mL O 2 /g hr −1 , n = 6) was not different from that in rats (1.64 ± 0.36 mL O 2 /g hr −1 , n = 3). During hyperoxic exposure, V O2 increased significantly in both AGS and rats (P= <0.001), as compared with values before hyperoxic exposure. During the hyperoxic period V O2 in AGS (2.64 ± 0.25 mLO 2 /g hr −1 ) was significantly lower than in rats (3.40 ± 0.32 mL O 2 /g hr −1 ) (P = <0.001). During the period after 100% O 2 exposure, V O2 recovered in AGS (1.59 ± 0.15 mL O 2 /g hr −1 ) and in rats (1.36 ± 0.39 mLO 2 /g hr −1 ). For both AGS and rats, V O2 after 100% O 2 exposure was not different from baseline values before the 100% O 2 exposure. Fig. 3, demonstrates the real-time changes in brain temperature measured from non-sedated and nonanesthetized euthermic AGS and rats before, during and after exposure to 100% O 2 . Panel B shows the average values in brain temperatures before (AGS, 37.06± 0.42°C; Rat, 32.73± 2.03°C), during (AGS, 37.27± 0.38°C; Rat, 33.23± 1.67°C) and after exposure (AGS, 36.80± 0.49°C; Rat, 32.70± 1.76°C) to 100% O 2 in these two species. The brain temperatures maintained constant before, during and after 100% O 2 exposure. Hyperoxic exposure had no effect on brain temperature in either AGS or rats. However, brain temperature is lower in rats than in AGS before, during and after 100% O 2 exposure (p = <0.001).

Response of respiratory frequency to 100% O 2 :
Panel A in Fig. 4, demonstrates the real-time changes in the respiratory frequency (f R ) measured from nonsedated and non-anesthetized euthermic AGS and rats before, during and after exposure to 100% O 2 . Panel B shows the average values in f R before, during and after exposure to 100% O 2 in these two species. The baseline values of f R of AGS (102.00 ±10.86 bpm) and rats (113.00 ± 2.33 bpm) showed no significant difference in the present study. The exposure of 100% O 2 did not significantly change f R in either species during hyperoxic exposure; however, f R in AGS (71.20 ± 7.86 bpm) was significantly lower than in rats (120.50 ± 8.50 bpm) after 100% O 2 exposure (p = 0.018). In AGS, f R was also significantly decreased after 100% O 2 exposure (71.20 ± 7.86 bpm) than before the exposure (102.00 ± 9.408 bpm) (p = 0.032). Respiratory frequency of AGS declined continuously for two hours after their return to normal 12 air inhalation. Rats maintained a relatively stable level in f R after cessation of hyperoxic exposure and returning to air.

Hyperoxic response and brain tissue oxygenation:
High-dose O 2 is routinely used in clinical and nonclinical settings to prevent or treat hypoxemia and tissue hypoxia, as well as controversial treatment of other miscellaneous disorders (Balentine, 1982;Tibbles and Edelsberg, 1996). Little information is available about the precise P t O 2 response during 100% O 2 exposure under real time in-vivo natural state without anesthetic or sedated effect or whether AGS has a greater P t O 2 tolerance to hyperoxic exposure than nonhibernating species. The absence of information about the precise P t O 2 change pattern during 100% O 2 exposure in both hibernating and non-hibernating animals is either caused by anesthetic effect on P t O 2 or by the lack of an ideal methodology for P t O 2 measurement of free moving animals or conscious humans. Our study is the first to examine hyperoxic effects on directly measured P t O 2 in non-sedated and non-anesthetized AGS and rats under 100% O 2 .
Our results demonstrate that P t O 2 in both AGS and rats instantly increase to significantly higher levels during 100% O 2 . Similar changes in brain tissue oxygenation during hyperoxic exposures have been reported in non-hibernating animals (Shin et al., 2007;Liu et al., 2006;Rossi et al., 2000;Shinozuka et al., 1989) and humans (Hlatky et al., 2008;Longhi et al., 2002;Macey et al., 2007;Tolias et al., 2004;Menzel et al., 1999) under anesthetic conditions using different O 2 measurement methods. Our P t O 2 results demonstrate that hibernating species share some of the same response mechanisms as non-hibernating species, like the rat, during hyperoxic exposure. Intriguingly, we found that the response level of P t O 2 during 100% O 2 inhalation was significantly higher in euthermic AGS than in rats. This result in the response difference of P t O 2 under hyperoxia offers solid support to our hypothesis that AGS have a greater tolerance to hyperoxia in the brain than other terrestrial mammals such as the lab rat.
When animals or humans are exposed to hyperoxic inhalation, the functions of peripheral chemoreceptors in carotid body are eliminated and those peripheral chemoreceptors become silent (Dean et al., 2004;Watt et al., 1943;Lahiri et al., 2006). Therefore, hyperoxic gas mixtures are routinely used for chemical denervation of peripheral O 2 receptors in in-vivo studies of respiratory control. The response mechanisms under hyperoxic inhalation conditions are commonly considered to be regulated by oxygen-chemosensitive neurons of the central nerve system, which are distributed throughout the brain stem from the thalamus to the medulla (Neubauer and Sunderram, 2004;Dean et al., 2003Dean et al., , 2004Mulkey et al., 2001). Therefore, the higher response level of P t O 2 during 100% O 2 inhalation in AGS compared to rat is accounted by the regulation mechanism controlled by oxygen-chemosensitive neurons of the central nerve system. P t O 2 under hyperoxia is maintained by two primary factors, cerebral blood flow (CBF) and P a O 2 (Hlatky et al., 2008). Hyperoxia increases P a O 2 , but decreases CBF in humans (Diringer et al., 2007;Hlatky et al., 2008). Therefore, the net oxygen delivery to brain during hyperoxia depends on the combined effects of these two factors and other O 2 regulating mechanisms.
Hibernating species have a unique ability to undergo and tolerate dramatic changes in CBF during arousal Hashimoto, 2003, Ma et al., 2005). This special CBF regulation mechanism may contribute to a larger increase in P t O 2 through CBF change during hyperoxic ventilation. As reported in a previous study (Ma et al, 2005), hemoglobin-O 2 saturation (sO 2 ) increases significantly from 85-95% during100% O 2 inhalation in euthermic AGS. This contrasts with rats who maintain a constant sO 2 (~97%) during 100% O 2 exposure. These results demonstrate that euthermic AGS with unsaturated blood hemoglobin-O 2 have a greater capacity than the rats to buffer the hyperoxic exposure. The increase in sO 2 during 100% O 2 inhalation elevates the oxygen content in blood, which further induces a higher brain P t O 2 in euthermic AGS. Lower systemic oxygen consumption in euthermic AGS compared to rat during 100% O 2 exposure may be another cause for higher P t O 2 of AGS as indicated in Fig. 2. AGS brain may also have special mechanisms to increase tissue oxygen extraction, as reported for non-hibernating species (Shin et al., 2007) and humans (Murthy, 2006).
Despite the benefit of hyperoxia to increase tissue oxygen delivery to brain (Shin et al., 2007), hyperoxic ventilation can accentuate the effects of ischemia (Macey et al., 2007) and lead to oxidative stress and free radical damage (Liu et al., 2006). Paradoxically, hyperoxia results in increased ventilation, leading to hypocapnia, diminished cerebral blood flow and hindered oxygen delivery. Hyperoxic delivery could also induce other systemic changes, including increased plasma insulin and glucagon levels and reduced myocardial contractility and relaxation (Macey et al., 2007). To evaluate these side effects of hyperoxic ventilation on P t O 2 , it is critical to measure P t O 2 recovery after cessation of hyperoxic ventilation. Our results demonstrated that P t O 2 in both AGS and rats recovered to the baseline values immediately after returning to ambient air inhalation. However, P t O 2 of AGS recovered from a significantly higher level to the baseline value after cessation of hyperoxic exposure and return to ambient air. This new finding suggests that AGS brains have a greater tolerance to hyperoxic stress. Their greater ability to recover suggests a more tolerant protection mechanism to overcome the side effects of hyperoxic ventilation. Further histological evidence of the absence of cellular stress and tissue damage is necessary to confirm this physiological result.

Hyperoxic response and metabolic rate for O 2 :
Under normoxic ventilation (21% O 2 inhalation), changes in the systemic metabolic rate for O 2 (V O2 ) are temperature dependent in both rats (Dotta and Mortola et al., 1992a) and AGS (Buck and Barnes et al., 2000). When AGS inhale ambient air, V O2 decreases as ambient temperature increase from -16-0°C and remains relatively constant between 0 and 16° C. V O2 increases again from 16°C to room temperature (Buck and Barnes, 2000).
With hyperoxic ventilation, the systemic metabolic rate for O 2 increases in non-hibernating animals (Stanek et al., 1979;Stock et al., 1985;Dotta and Mortola, 1992) and humans (Wilson et al., 1975;Mortola et al., 1992aMortola et al., , 1992b. In hyperoxia the average values of V O2 are above the normoxic values at all ambient temperatures in newborn rats (Dotta and Mortola, 1992). At 20°C, close to the temperature of 18°C used in the present study, hyperoxic values of V O2 in newborn rats increased by 73% compared to the values of V O2 in normoxia (Dotta and Mortola, 1992). Diringer et al. (2007) found that hyperoxia had no impact on the brain's consumption of oxygen in humans despite their increased P a O 2 and blood oxygen content. This result suggests that the brain regional metabolic rate does not contribute to increased V O2 during hyperoxia in 15 non-hibernating species and humans. However, brain regional metabolic rate in AGS may change during hyperoxia. The brain metabolic response to hyperoxia in hibernating species has not yet been studied.
In this study, we found that V O2 during hyperoxic exposure increased and V O2 was significantly lower in AGS than in rats. This difference in V O2 during hyperoxia was not caused by body mass scaling as discussed in our previous study (Ma et al, 2009), where we found that V O2 during 8% O 2 exposure was significantly lower in euthermic AGS than in rats. Similar results in V O2 changes under opposite inhalation conditions (100 O 2 and 8% O 2 ) suggest that AGS may have a general mechanism protecting their brains from either hyperoxic or hypoxic stress.
Hyperoxic response and brain temperature: In contrast to the rapid drop of brain temperature during hypoxic inhalation (Ma et al., 2009), the brain temperatures in both AGS and rats maintain constancy before, during and after the 100% O 2 inhalation. The ability to maintain brain temperature during normobaric hyperoxic inhalation could account for increased cerebral blood flow (Shin et al., 2007), improved cerebral metabolism indicated by biochemical markers (Tolias et al., 2004) or maintained cerebral metabolic rate during hyperoxia (Diringer et al., 2007). The ability to maintain brain temperature during hyperoxia by AGS and rats demonstrates that these two different species share a protective thermogenic mechanism against hyperoxia.
We previously reported that the baseline value of brain temperature in rats is similar to euthermic AGS. However, in the current study, we observed that rat baseline brain temperature is lower than AGS. In our experience, the core body temperature of euthermic AGS varies between 30-39°C, while rats have a more constant core body temperature between 36-37°C. This greater variation in AGS core body temperature could account for the difference of T brain between the two species. Another contributing factor to the T brain difference could be attributed to depth of the temperature probe in the brain (Zhu et al, 2006).
Hyperoxic ventilatory response and f R : Studies in non-hibernating animals and humans suggest that breathing very high concentrations of oxygen can lead to an increase in ventilation (Becker et al, 1996;Dean et al., 2004;Forkner et al, 2007). Unlike nonhibernating animals or humans, AGS have a very specific ventilation pattern under hypoxic inhalation conditions (Ma et al, 2009). AGS f R decreases with a reduction in ventilation during hypoxic inhalation (8% O 2 ); rats, however, increase f R with an elevated ventilation during 8% O 2 exposure. The f R in rats gradually increases during hyperoxic exposure with 100% O 2 and maintains this increase after exposure. The f R in AGS show a slight decrease during hyperoxic exposure, followed by a significant decline after hyperoxic exposure ends. The mean value of f R after 100% O 2 exposure is significantly lower in AGS (71.20 bpm) than in rats (120.50 bpm). The respiratory control system for non hibernators and humans is particularly sensitive to an increase in inspired O 2 . Dean et al, 2004 demonstrate that the ventilation response to hyperoxia is a biphasic response. The first few breaths of hyperoxia (a short lived breath for 1-2 minutes) typically decrease expired minute ventilation (V E ) by inhibiting the peripheral chemoreceptors. The secondary response is a hyperoxic hyperventilation (a dose-dependent increase in tidal volume V T ), which may or may not be accompanied by an increase in f R . Following "chemical denervation" or "physiological chemodenervation" of the peripheral chemoreceptors, hyperoxic hyperventilation is of central origin. Oxygen has a stimulating effect, which tends to increase respiration and evoke a long lasting facilitation of ventilation. Hyperoxia may alter tonic output for central respiratory drive. The stimulating effect of hyperoxia on respiratory pattern may account for the gradual increase of f R in rats, the gradually decrease of f R in AGS and no immediate recovery after hyperoxia in both AGS and rats. Based on this knowledge, we believe that the response difference of f R between AGS and rat following hyperoxic inhalation may relate to a long lasting facilitation of V E or changes in tonic output for central respiratory drive.

Animal model for hypoxic and hyperoxic stress:
Oxygen pressure in mammalian CNS needs to be maintained at a level which is sufficiently high to ensure undisturbed function of brain cells and sufficiently low to minimize generation of free radicals (Erecinska andSilver, 2001, Al-Hashem, 2010). Excessive oxygen is toxic to mammalian CNS due to over production and accumulation of reactive oxygen species (ROS). Hyperoxia is not a natural existing condition in mammals, but hyperoxia has been a popular model of oxidative stress (Dean et al., 2004).
Laboratory reared rats have been used as a model for humanlike ventilatory adaptation to chronic hypoxia (Olson and Dempsey, 1978). Rat is a recognized control animal for hibernating species such as frog (St-Pierre et al., 2000), hamster (Fawcett and Lyman, 1954;Nikmanesh et al., 1996 ), bat (Mehrani and Storey, 1997), bear (Fustera et al., 2007) and ground squirrel (Fawcett and Lyman, 1954;Popov et al., 2007;Ma et al., 2005Ma et al., , 2009. Rat is the ususal mammalian representative for human physiology and serves as a proper control for hibernating species. AGS is a specific hibernating mammalian species that tolerates extreme hypoxic conditions with low cerebral blood flow, low P a O 2 and unsaturated sO 2 during arousal (Ma et al., 2005(Ma et al., , 2009. AGS additionally tolerate brain tissue penetration damage without inflammation (Zhou et al, 2002). In the current study, we demonstrate that AGS tolerates hyperoxic stress in brain tissue. Because of these tolerances, AGS is an excellent animal model for both hypoxic and hyperoxic stress.
Data from this exploratory methodological study demonstrates the effects of hyperoxic exposure on P t O 2 in the striatum recorded from non-sedated, nonanesthetized, freely moving hibernating species. Hyperoxic ventilation induces the response differences in P t O 2 and V O2 between AGS and rats and alters the respiratory pattern in AGS following exposure. Since AGS brains have a greater tolerance to hyperoxia than rats, AGS serves as an excellent comparative model for hypoxic and hyperoxic stress.

CONCLUSION
Hyperoxic ventilation induced P t O 2 and V O2 differences between AGS and rats and led to altered respiratory patterns between these species. AGS and the rat serves as an excellent comparative model for hypoxic and hyperoxic stress studies of the brain.