- Email: [email protected]
Assessment of cerebral pO2 by EPR oximetry in rodents: effects of anesthesia, ischemia, and breathing gas
BRAIN RESEARCH Brain Research 685 (1995) 91-98
ELSEVIER
Research report
Assessment of cerebral pO e by EPR oximetry in rodents: effects of anesthesia, ischemia, and breathing gas Ke Jian Liu a, Goran Bacic a, p. Jack Hoopes b, Jinjie Jiang a, Hongkai Du c, Lo Chang Ou c, Jeff F. Dunn a, Harold M. Swartz a, * a Department of Radiology, Dartmouth Medical School, Hanover, NH 03755, USA b DepartmeniofRadiation Oncology, Dartmouth Medical School, Hanover, NH 03755, USA c Department of Physiology, Dartmouth Medical School, Hanover, NH 03755, USA Accepted 14 March 1995
Abstract This report describes experiments designed to assess and illustrate the effectiveness of a new method for the measurement of cerebral interstitial pO 2 in conscious rodents. It is based on the use of low frequency electron paramagnetic resonance (EPR) spectroscopy with lithium phthalocyanine as the oxygen sensitive probe. Magnetic resonance imaging was used to document placement of the probe in the brain, and to assess potential cerebral changes associated with the placement. The technique provided accurate and reproducible measurements of localized pO 2 in the brains of conscious rodents under a variety of physiological conditions and for time periods of at least 2 weeks. Using this approach we quantitated the depressing effects on cerebral pO 2 of three representative anesthetics, isoflurane, ketamine/xylazine, and sodium pentobarbital. The effects of changing the content of oxygen in the breathing gas was investigated and found to change the cerebral pO 2. In experiments with gerbils, crystals of lithium phthalocyanine were implanted in each side of the brain and using a one-dimensional magnetic field gradient, simultaneous measurement of pO 2 values from normal and ischemic (ischemia induced by unilateral ligation of a carotid artery) hemispheres of the brain were obtained. These results demonstrate that EPR oximetry with lithium phthalocyanine is a versatile and useful method in the measurement of cerebral pO 2 under various physiological and pathophysiological conditions.
Keywords: Cerebral pO2; EPR oximetry; lithium phthalocyanine; anesthesia; ischemia
I. Introduction The brain has a high metabolic rate and depends primarily upon aerobic metabolism. Because of these characteristics, brain tissue is very sensitive to changes in the availability of oxygen and pathological changes can occur if there are significant changes in the delivery a n d / o r utilization of oxygen in local regions of the brain. Complex and effective physiological controls have evolved to enable the brain to maintain an adequate level of oxygen within the tissues. Knowledge of the control mechanisms for maintaining local partial pressure of oxygen (pO 2) and the conditions which lead to their failure to prevent patho-
* Corresponding author. Fax: (1) (603) 650-1935. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 0 4 1 3 - 0
logical changes is incomplete, because of the difficulty of making measurements of local pO 2 under various physiological and pathophysiological conditions. Such knowledge seems essential for devising optimal strategies for diagnosis and therapy for the many processes which can lead to hypoxic damage to the brain, especially injury from ischemia-reperfusion. The most common way to measure brain pO 2 has been to use oxygen sensitive electrodes, by insertion of small microelectrodes [12] or by surface multi-electrode arrays [9]. The former may cause mechanical injury to the tissue and repeated measurements at the same site are not possible. The latter method cannot make accurate measurements of local pO 2 deep in brain tissue. A non-invasive optical method has been developed recently, using the oxygen quenching of phosphorescence [24], but it requires intravenous injection of porphyrins and measurements are lim-
92
K.J. Liu et aL / Brain Research 685 (1995) 91-98
ited to a depth of 1 mm from the surface. Near infrared spectroscopy also has been used, but it measures cerebral blood flow, not the interstitial pO 2 [8]. Recently there has been significant progress in the use of electron paramagnetic resonance (EPR, or completely equivalently, electron spin resonance or ESR) spectroscopy for measurements of oxygen in vivo in conscious animals [19]. This has resulted from the development of low frequency EPR instruments (1 GHz or lower) suitable for use with living subjects [6,13] and the development of new stable paramagnetic particles (fusinite, India ink, lithium phthalocyanine) with oxygen sensitive EPR spectra [10,18,23]. These particles combine several desirable properties as probes of the pO 2 including high sensitivity to the pO 2, resistance to chemical reactions, and a high degree of inertness in biological systems. The exact location of the paramagnetic particles in tissues can be determined by magnetic resonance imaging (MRI) [2]. These developments have established EPR oximetry as a versatile tool for sensitive, non-invasive, and reproducible measurements of pO 2 in a variety of biological systems. In this paper we report the use of EPR oximetry with lithium phthalocyanine (LiPc) particles for the measurement of cerebral interstitial pO 2 in conscious rodents. The aims of the study were to assess the effectiveness of EPR oximetry in the measurement of cerebral pO 2, and to demonstrate the usefulness of the technique for in vivo brain studies by monitoring the effect of anesthesia, ischemia, and different breathing gases on cerebral pO 2. LiPc was selected for pO 2 measurements in brain because of its small size and high sensitivity which makes it particularly suitable for oximetry of neural tissues. MRI was used to document the location of the LiPc and to assess potential cerebral changes associated with its presence. Assessment of the post-implant tissue effects was made by microscopic examination of histologic sections of gerbil brains taken in the region where the LiPc was placed and in the same plane as the MR images.
2. Materials and methods 2.1. Materials
Lithium phthalocyanine crystals were obtained from Dr. M. Moussavi (LETI, Grenoble, France). Details concerning synthesis, characterization and calibration of LiPc for pO 2 measurements have been reported [10,21,22]. The line width of the EPR signal is a linear function of pO 2, and is independent of the type of tissue being measured, local metabolic processes, the presence of other paramagnetic species, or pH. The high density of unpaired spins (8 × 1019 spins/cm 3) combined with a narrow intrinsic line width of LiPc allows measurements of pO 2 in the brain using one or more crystals with a diameter of ~ 200 /zm.
2.2. A n i m a l preparation
Male Wistar rats weighing 200-250 g were obtained from Charles River Laboratories (Wilmington, MA). The male Mongolian gerbil (100 gram, Tumblebrosk Farm, PA) was chosen as the ischemia model because of the reportedly high percentage of animals which have an incomplete Circle of Willis [3,5]. An incomplete Circle of Willis allows for the production of unilateral cerebral ischemia via unilateral carotid occlusion. Twenty-four hours prior to EPR measurements the animals were anesthetized with ketamine/xylazine (100/10 m g / k g , i.m.), and LiPc crystals implanted in the cerebral cortex. The LiPc crystals were positioned by drilling a small hole (500 /zm in diameter) in the cranium followed by insertion of a needle containing the crystals. The crystals were then deposited by pushing a stylus through the barrel of the needle to a specified depth (4 + 1 mm). In the case of gerbils, LiPc crystals were similarly introduced into both left and right hemispheres through two holes (6 mm apart) in the skull. Prior to EPR measurements the spatial localization of the crystals was verified by MRI. Following final pO 2 measurements, the brains were removed and cut in the same plane as the MR image. The brains were fixed in 4% neutral-buffered formalin, imbedded in paraffin, stained with hematoxylin and eosin (H and E), and examined microscopically. 2.3. Electron paramagnetic resonance
The spectra of LiPc were obtained using an EPR spectrometer constructed in our laboratory with a low frequency (1.2 GHz, L-band) microwave bridge [13]. An extended loop resonator was positioned over the brain of the animal. Typical settings for the spectrometer were: incident microwave power, 10 mW; magnetic field center, 425 gauss; scan range, 1 gauss; modulation frequency, 27 kHz. Modulation amplitude was set at less than one-third of the EPR line width, typically around 20-50 mG. The linear magnetic field gradient, used to separate the signals originating from crystals in the two hemispheres of the brains of gerbils, was produced by a pair of coils in an anti-Helmholtz configuration capable of generating gradients along the B 0 up to 0.5 m T / c m . The magnitude of the gradient was calibrated by measuring the peak-to-peak splitting of two capillaries containing LiPc under nitrogen. Both unanesthetized and anesthetized animals were studied. The unanesthetized animals were restrained in a specially constructed device. To reduce tuning problems associated with movements of the unanesthetized animal, we constructed an external loop resonator tuned by a varactor diode and altered the automatic frequency control (AFC) system [15]. In this new system, the AFC error signal tunes the resonator to a fixed frequency. By this means, a shift of resonant frequency in the resonator due to
K.J. Liu et al. / Brain Research 685 (1995) 91-98
93
70
i
i
I
i 10
t 20
J 30
3.6
60 m 5O
3.3 O
4o
/ -+-
3o
20
10
10
I
I
I
I
I
I
20
30
40
50
60
70
3.0
/ o
2.7
0
80
T i m e (rain)
Inspired o x y g e n (~*)
Fig. 1. The relationship of the cerebral p O 2 of rats with oxygen content in the breathing gas. The unanesthetized animals were restraincd in a homc-made device, and wcre conscious throughout thc cxperiment. W e waited at least 20 rain after change of oxygcn content to allow p O 2 to rcach a new equilibrium. The inscrt shows a typical E P R spectrum of LiPc recorded from the brain. Ccrcbral p O 2 valucs arc cxpressed as mean + S E M (n = 3 animals).
animal motion can be compensated, keeping the frequency constant even if the animal moves. Scan time was 1 min. In some experiments such as kinetic studies, the data points from single scans were used directly. For steady state experiments, 5 scans usually were accumulated. The total EPR scan time for any measurement was less than 10 min, including positioning of the animal, adjusting the spectrometer, and accumulation of 5 scans. The EPR line widths were converted to pO 2 using a calibration curve determined for each batch of crystals.
2.4. Measurement of cerebral p O 2
Fig. 3. The samc data as in Fig. 2 to illustratefirst order kinctics of the
changes of brain pO 2 after pcrturbation. Data wcre transformed by taking logarithm of pO 2 values. Time zero is the starting time of switching inspired oxygen from 21% to 10.5% (O), or from 10.5% to 21% (Q).
allowed regulation of the breathing gases. Cerebral pO 2 was measured at various inhaled pO 2 and the breathing gas was monitored using an Oxygen Analyzer (Sensor Medics, Model OM-11, Anaheim, CA). To follow the time course of changes in cerebral pO 2 after the oxygen content in the breathing gas was switched from 21% to 10.5%, and back to 21%, the EPR spectrum was recorded repetitively at 60 s intervals. For the study of the effect of anesthesia on pOz, measurements were made before and after the peritoneal injection of a ketamine/xylazine mixture, or sodium pentobarbital, or inhalation of the gas anesthetic, isoflurane, which was delivered to the animals through the nose-cone of the restraining device. Occlusion of the common carotid artery was accom-
Unanesthetized rats were restrained in the specially constructed device with a nose-cone at one end, which 40 50
3o 4O
~
o°~ o
3O
isof
oo eJ
***
ket/xyl
•
~•~
2o
~
pen
~d
20
% o
L)
o:~oe
lo
10 •
21~, 02
10.5~ 02
o
I
I
I
I
20
40
60
80
100
Time (rain) Fig. 2. Time course of the change in cerebral pO 2 after perturbation in the oxygen content in breathing gas. The breathing gas was switched between 21% oxygen ( O ) and 10.5% (©). EPR spectra were recorded every minute.
0
- -
- -
- -
Fig. 4. Effect of anesthetics on the cerebral pO 2. Ketamine/xylazine mixture (100/10 mg/kg) and sodium pentobarbital (50 mg/kg) were injected into the rats intraperitnneally. Isoflurane (1%, flow rate 3 ml/min.) was delivered to the animal through a nose cone in restraining device. The values are expressed as mean +_S.E.M. ( n = 3 animals). • P < 0.003, * * P < 0.001, * * * P < 0.0001 vs. control.
94
K.J. Liu et al. / Brain Research 685 (1995) 91-98
50
left common carotid artery isolated. Following carotid isolation, blood flow was occluded by tying off the artery with elastic suture material (occlusion knots with this type of suture could be released without damage to the artery). The skin incision was then sutured closed and the gerbil positioned in ventral recumbency for EPR p02 measurement.
40
30
~eu
2o
2.5. M a g n e t i c r e s o n a n c e i m a g i n g
10
0 Day 1
Day 3
Day 7
Fig. 5. Measurement of cerebral pO2 over time after implantation of LiPc. EPR spectra were collected at day 1, 3 and 7 after implantation from unanesthetizedrats while the animals breathed 21% oxygen (11) or 10.5% oxygen ([]). The values are expressed as mean±S.E.M. (n = 3 animals). plished by anesthetizing the gerbils (ketamine/xylazine) and positioning them in dorsal recumbency. A 1-cm longitudinal skin incision was made in the ventral neck. Underlying subcutaneous tissues were bluntly dissected and the
MR images of the brain of anesthetized animals were obtained using a 7.0 T imager with a horizontal bore Magnex magnet and a SMIS console. Visualization of the LiPc in the brain was achieved using a bird-cage RF coil (d = 6 cm) and gradient-recalled echo (GRE) sequence ( 2 0 0 / 1 0 / 4 5 °) with FOV = 4 cm; matrix size = 128 × 128; slice t h i c k n e s s = 1 mm; and 4 acquisitions. T2weighted spin-echo images ( 2 0 0 0 / 6 0 ) were obtained in the same plane as the GRE images.
3. Results Fig. 1 shows that the cerebral pO 2 increased when the oxygen content in the breathing gas was increased. Fig. 2
Fig. 6. Axial MR images of rat brain with implanted LiPc crystals (arrows). A,C: gradient-recalledecho (200/10/45° ) and B,D: spin-echo(1800/50) images were obtained 1 day (top row) and 7 days (bottom row) after implantationof the LiPc.
K.J. Liu et a l . / Brain Research 685 (1995) 91-98
C
mmHg
95
shows the time course of changes in cerebral pO 2 when the breathing gas changed from 21% oxygen to 10.5%, and back to 21%. When the logarithm of the changes of p O 2 was plotted against time, straight lines were obtained, indicating that the changes of cerebral p O 2 followed first order kinetics (Fig. 3). The calculated half-life for reaching a new equilibrium was 23 min when the oxygen in the mixture was decreased from 21% to 10.5%, and 15 min in the reverse direction. The effect of anesthesia on the cerebral p O E w a s studied using three representative anesthetics, isoflurane, ketamine/xylazine, and sodium pentobarbital. As shown in Fig. 4, all three anesthetics substantially reduced the cerebral p O E. Sodium pentobarbital had the largest effect decreasing the pO 2 approximately 50%. To determine the feasibility of measuring cerebral p O 2 over extended periods of time with LiPc, we measured the EPR spectra at 1, 3 and 7 days after implantation of the crystals. Fig. 5 shows that p O 2 remained constant over this period of time, and when the breathing gas was changed from air to 10.5% 0 2, the cerebral p O 2 decreased reproducibly. Fig. 6 demonstrates the ability of MRI to locate LiPc crystals within the brain. The GRE technique greatly increases the apparent size of the crystals on the image ( ~ 1 mm). Considering the in-plane resolution of 0.3 mm and the slice thickness of 1 mm it should be very difficult to visualize an object the size of the LiPc crystal because of insufficient in-plane resolution and volume averaging. The ability to visualize LiPc crystals on MR images under such conditions is based on bulk magnetic susceptibility differences between the tissue and the crystal, which produces a susceptibility-based signal attenuation around the crystal which is much larger than the actual crystal [2]. This T 2 effect can be easily visualized on gradient-echo sequences (Fig. 6A, C), but is considerably smaller on spin-echo sequences (Fig. 6B,D) even when relatively heavily T2weighting is applied. This is illustrated in Fig. 6 where MRI was used to determine that there were no gross changes (e.g. edema) observed on images either acutely (Fig. 6B) or one week (Fig. 6D) following implantation of LiPc. Histologic assessment of the brains showed that MRI provides an accurate determination of the location of the LiPc, and that under the conditions of this experiment there is little tissue reaction to the presence of LiPc, or to the implantation procedure. Simultaneous measurements of p O 2 in both cerebral
\ Ischeinic p 0 2 = 11.4 rnrnHg I
399.8
i
I
,
Magnetic
I
400.2
400.0
field
(G)
Fig. 7. (A, B) Two contiguous axial MR images (GRE-100/7/45 ° ) of gerbil brain showing two LiPc crystals implanted in the cerebral cortex. (C) EPR spectrum of the LiPc crystals in gerbil brain after unilateral carotid occlusion (anesthetized animal) demonstrating the simultaneous assessment of pO 2 in the ischemic (right) and non-ischemic (left) cerebral hemispheres. A magnetic field gradient of 0.05 mT/cm was used.
96
K.J. Liu et al. / B r a i n Research 685 (1995) 91-98
hemispheres via bilateral LiPc implantation in both sides of brains (Fig. 7A, B) was successfully accomplished in the gerbil. Using this technique, we were able to study the effects of unilateral carotid occlusion, and measure the pO 2 in the ischemic and non-ischemic brain in the same animal simultaneously. Ten minutes after unilateral carotid occlusion, pO 2 was measured in both hemispheres using a magnetic field gradient of 0.05 m T / c m to separate the spectra from the two sites. Carotid occlusion resulted in a significant reduction of pO2 in the corresponding side of the brain; the pO 2 decreased to 11.4 mmHg compared with 24.1 mmHg in the non-occluded side (Fig. 7C). Because the gerbil was anesthetized with ketamine/xylazine, the observed pO 2 in the non-ischemic side of the brain was lower than in conscious animals and was consistent with the value obtained in the studies of the effects of anesthetics (Fig. 4).
4. Discussion These studies demonstrate that EPR oximetry is capable of assessing cerebral pO 2 in vivo with several favorable features, which are summarized in the following paragraphs. As illustrated in Fig. 5, pO 2 measurements can be made repeatedly on the same animal and at the same location without anesthesia or other invasive procedures, whereas in the use of electrodes, repeated insertions over time are likely to be perturbing and difficult to place at exactly the same location. The location of the site where the pO 2 is being measured by oxygen sensitive materials, such as LiPc, fusinite, and India ink can be verified histologically a n d / o r by MRI [2], since the paramagnetic material themselves are markers for the location where the pO e measurements are obtained. EPR oximetry can provide chronological studies of pO 2 over intervals ranging from short term variation (minutes) to long-term (days or weeks) follow-up studies of changes of pO 2 following occlusion or any other treatment (Fig. 2). Because the response time of LiPc to changes in pO 2 is immediate [10], EPR oximetry with LiPc measures the actual kinetics of the processes. Measurements of pO 2 by LiPc are highly localized. The pO 2 measured is the average pO 2 in the tissue that are adjacent to the paramagnetic particles, i.e., tissues in immediate equilibrium with the surface of the LiPc crystals. These materials can be as small as 200 /xm in diameter, therefore the pO 2 in any particular region of the brain can be selected for measurement of pO 2. Because of this capability to have localized measurements based on the placement of the particles, multisite (two or more) simultaneous measurements in different regions of the brain are possible, as shown in Fig. 7. Spatial resolution in the range of 2 mm easily can be achieved without significant distortion of the EPR signal.
This is a particularly useful feature of EPR oximetry because the measurement from one region can be used as an internal control while the signals from other regions are being monitored as indicators of the experimental procedures. EPR oximetry is non-invasive after the initial placement of the crystals and the measurements of pO 2 can be made repetitively at will without anesthesia since the LiPc is inert and remains responsive to changes of oxygen. The observed cerebral pO 2 value of 34.1 ___3.2 mmHg under normal conditions is consistent with the value of 36 mmHg reported using electrodes [12]. The cerebral tissue pO 2 measured under room air and 10.5% oxygen conditions are in good agreement with tissue pO 2 previously estimated under these conditions [20]. Thus it appears that EPR oximetry provides results similar to those obtained with microelectrodes but it offers several favorable features as described above, and in most situations, yields more precise data [17]. The EPR oximetry measurements indicated that it took longer to reach a new equilibrium when the oxygen in the breathing gas was decreased from 21% to 10.5% than in the reverse direction. Since the response time of LiPc to changes of pO 2 is virtually instantaneous [10], and the response time of blood pO 2 usually takes seconds to minutes, the observation of a slow decrease in cerebral pO 2 after an acute drop in inspired oxygen may reflect the operation of a physiological control mechanism which attempts to maintain the pO 2 in the brain in the presence of decreased availability of oxygen. The drop in pO 2 after administration of the anesthetics agents is likely related to depressed respiration induced by the anesthetics [14]. The results indicate the difficulties that may arise when studying oxygen dependent processes (including processes affected by oxidative metabolism) in anesthetized animals. In addition to changes in the pO 2 in various organs, anesthetics also are likely to alter the complex physiological control mechanisms that strive to maintain tissue pO 2. Therefore experimental manipulations which restore the pO 2 of the blood to normal levels may not compensate fully for the pO 2 changes in all tissues a n d / o r the responses of the systems to further experimental manipulation. We believe that EPR oximetry may provide a good means to investigate these problems in depth. Multisite spectral oximetry ideally should be performed in such manner that the applied magnetic field gradient is sufficient to separate signals from different sites, but not large enough to cause the distortion of the local crystal line-width which could give erroneous values of local pO 2. The degree of distortion depends on the relationship between the size of the crystal, applied gradient and intrinsic line width (i.e. linewidth in the absence of the gradient). The gradient strength used here (0.05 m T / c m ) allows good separation of signals from the two hemispheres of the rodent brain. Selection of the gradient strength was based
K.J. Liu et al. /Brain Research 685 (1995) 91-98
on knowledge of the distance between crystals (MRI) and value of linewidth (0.025 mT) under normal oxygenation of the brain. If necessary, line distortion induced by a gradient can be corrected, provided that the above parameters are known. It was not necessary in these studies, however, since it has been shown both theoretically and experimentally that a gradient of 0.05 m T / c m does not cause any significant distortion of the EPR line of LiPc [16]. The combination of EPR oximetry with MRI techniques can be quite useful. GRE MRI was used here to determine the site of the LiPc and could be repeated as necessary to confirm the localization during long-term EPR studies of oxygen changes in the brain. Spin echo MRI was used to show that there were no detectable changes in the MR image in the vicinity of the LiPc, indicating that there was not extensive tissue damage associated with the presence of the LiPc. It is important to note that susceptibility effects used for locating the paramagnetic materials do not cause severe effects on spin-echo images so that implantation of the LiPc does not preclude application of MRI techniques for assessment of physiological changes associated with ischemia such as diffusion weighted imaging [4,11]. Likewise, the presence of LiPc should not affect assessment of physiological changes by localized MR spectroscopy [1], even if the volume of interest (VOI) has to be placed directly over the position of the crystal, since the typical VOI for 31p or I H spectroscopy is around 100 mm 3 while the volume affected by LiPc is less than 1 mm 3. We emphasize this feature because both 31p spectroscopy and DWI have been advocated for non-invasive assessment of ischemia, but both techniques measure oxygen dependent changes such as ATP depletion and redistribution of the water as a consequence of changes in metabolism, but not pO 2 itself. Presently it is not known how these quantities are related, but by combining these techniques with in vivo EPR oximetry in the same animal model it should be possible to assess the correlation and 'calibrate' M R I / M R S findings [7]. In conclusion, these studies indicate that EPR oximetry can make noninvasive, accurate, and repeated measurement of cerebral p O E at multiple sites simultaneously in conscious animals. Our results indicate that cerebral p O E is altered by different classes of anesthetics and the amount of oxygen in the breathing gas. The results also demonstrate that EPR oximetry can assess alterations in cerebral pO 2 associated with experimentally produced cerebral ischemia.
Acknowledgements This research was supported by NIH Grant GM 34250 and used the facilities of the IERC at Dartmouth supported by NIH Grant RR-01811.
97
References [1] Allen, K.L., Busza, A.L., Proctor, E., King, M.D., Williams, S.R., Crockard, H.A. and Gadian, D.G., Controllable graded cerebral ischemia in the gerbil: studies of cerebral blood flow and energy metabolism by hydrogen clearance and 31p NMR spectroscopy, NMR in Biomed. 6 (1993) 181-186. [2] Bacic, G., Liu, K.J., O'Hara, J.A., Harris, R.D., Szybinski, K., Goda, F. and Swartz, H.M., Oxygen tension in a murine tumor: a combined EPR and MRI study, Magn. Reson. Med., 30 (1993) 568-572. [3] Bothe, H.W., Bosma, H.J., Hofer, H., Hossmann, K.A. and Angermeier, W.F., Selective vulnerability of hippocampus and disturbances of memory storage after mild unilateral ischemia of gerbil brain, Stroke 17 (1986) 1160-63. [4] Busza, A.L., Allen, K.L., King, M.D., van Bruggen, N., Williams, S.R. and Gadian, D.G., Diffusion-weighted imaging studies of cerebral ischemia in gerbils. Potential relevance to energy failure, Stroke, 23 (1992) 1602-1610. [5] Clifton, G.L., Taft, W.C., Blair, R.E., Choi, S.C. and DeLorenzo, R.J., Conditions for pharmacologic evaluation in the gerbil model of forebrain ischemia, Stroke, 20 (1989) 1545-1552. [6] Colacicchi, S., Alecci, M., Gualtieri, G., Quaresima, V., Ursini, C.L., Ferrari, M. and Sotgiu, A., New experimental procedures for in vivo L-band and radio frequency EPR spectroscopy/imaging, J. Chem. Soc. Perkin Trans., 2 (1993) 2077-2082. [7] Dunn, J.F., Ding, S., Hoopes, J.P., Liu, K.J. and Swartz, H.M., Relating the apparent diffusion constant to brain interstitial pO2, Proc. Soc. Magn. Res., 2 (1994) 1006. [8] Elwell, C.E., Cope, M., Edwards, A.D., Wyatt, J.S., Reynolds, E.O.R. and Deply, D.T., Measurements of cerebral blood flow in adult humans using near infrared spectroscopy-methodology and possible errors, Adv. Exp. Med. Biol., 317 (1992) 235-245. [9] Grota, J., Zimmer, K. and Schubert, R., Tissue oxygenation in normal and edematous brain cortex during arterial hypocapnia, Adv. Exp. Med. Biol., 180 (1984) 179-184. [10] Liu, K.J., Gast, P., Moussavi, M., Norby, S.W., Vahidi, N., Walczak, T., Wu, M. and Swartz, H.M., Lithium phthalocyanine: a probe for electron paramagnetic resonance oximetry in viable biological systems, Proc. Natl. Acad. Sci. USA, 90 (1993) 5438-5442. [11] Mintorovich, J., Moseley, M.E., Chileuitt, L., Shimizu, H., Cohen, Y. and Weinstein, P.R., Comparison of diffusion-and T2-weighted MRI for the early detection of cerebral ischemia and reperfusion in rats, Magn. Reson. Med., 18 (1991) 39-50. [12] Nair, P.K., Buerk, D.G. and Halsey, J.H., Comparison of oxygen metabolism and tissue pO E in cortex and hippocampus of gerbil brain, Stroke, 18 (1987) 616-662. [13] Nilges, M.J., Walczak, T. and Swartz, H.M., 1 GHz in vivo ESR spectrometer operating with a surface probe, Phys. Med., 5 (1989) 195-201. [14] Pavlin, E.G. and Hornbein, T.F., Anesthesia and the control of ventilation. In Handbook of Physiology, Section 3, Vol. lI, Control of Breathing, Part 1, pp. 793-813. [15] Smirnov, A.I., Norby, S.W., Walczak, T., Liu, K.J. and Swartz, H.M., Physical and instrumental considerations in the use of lithium phthalocyanine for measurements of the concentration of oxygen, J. Magn. Reson. B., 103 (1994) 95-102. [16] Smiruov, A.I., Norby, S.W., Clarkson, R.B., Walczak, T. and Swartz, H.M., Simultaneous multi-site EPR spectroscopy in vivo, Magn. Reson. Med., 30 (1993) 213-220. [17] Swartz, H.M., Goda, F., Rhodes, E.S., O'Hara, J.A., Liu, K.J. and Dunn, J.F., Measurement of pO z in tissue in vivo by EPR: principles, applications, and comparison with electrode techniques, Proc. Soc. Magn. Res., 2 (1994) 754. [18] Swartz, H.M., Liu, K.J., Goda, F. and Walczak, T., India ink: a potential clinically applicable EPR oximetry probe, Magn. Reson. Med., 31 (1994) 229-232.
98
K.J. Liu et al. / Brain Research 685 (1995) 91-98
[19] Swartz, H.M. and Walczak, T., In vivo EPR: prospects for the 90's, Phys. Med., 9 (1993) 41-48. [20] Tenney, S.M. and Ou, L.C., Physiological evidence for increased tissue capillarity in rats acclimatized to high altitude, Respirat. Physiol., 8 (1970) 137-150. [21] Turek, P. andre, J.J., Giraudeau, A. and Simon, J., Preparation and study of a lithium phthalocyanine radical: optical and magnetic properties, Chem. Phys. Lett., 134 (1986) 471-476. [22] Turek, P., Andre, J.J., Moussavi, M. and Fillion, G., Septet spin state in the lithium phthalocyanine radical compound, Mol. Cryst. Liq. Cryst., 176 (1989) 535-546.
[23] Vahidi, N., Clarkson, R.B., Liu, K.J., Norby, S.W., Wu, M. and Swartz, H.M., In vivo and in vitro EPR oximetry with fusinite: a new coal-derived, particulate EPR probe, Magn. Reson. Med., 31 (1994) 139-146. [24] Wilson, D.F., Gomi, S., Pastuszko, A. and Grindberg, J.H., Oxygenation of the cortex of the brain of cats during occlusion of the middle cerebral artery and reperfusion, Adv. Exp. Med. Biol., 317 (1992) 689-694.
ELSEVIER
Research report
Assessment of cerebral pO e by EPR oximetry in rodents: effects of anesthesia, ischemia, and breathing gas Ke Jian Liu a, Goran Bacic a, p. Jack Hoopes b, Jinjie Jiang a, Hongkai Du c, Lo Chang Ou c, Jeff F. Dunn a, Harold M. Swartz a, * a Department of Radiology, Dartmouth Medical School, Hanover, NH 03755, USA b DepartmeniofRadiation Oncology, Dartmouth Medical School, Hanover, NH 03755, USA c Department of Physiology, Dartmouth Medical School, Hanover, NH 03755, USA Accepted 14 March 1995
Abstract This report describes experiments designed to assess and illustrate the effectiveness of a new method for the measurement of cerebral interstitial pO 2 in conscious rodents. It is based on the use of low frequency electron paramagnetic resonance (EPR) spectroscopy with lithium phthalocyanine as the oxygen sensitive probe. Magnetic resonance imaging was used to document placement of the probe in the brain, and to assess potential cerebral changes associated with the placement. The technique provided accurate and reproducible measurements of localized pO 2 in the brains of conscious rodents under a variety of physiological conditions and for time periods of at least 2 weeks. Using this approach we quantitated the depressing effects on cerebral pO 2 of three representative anesthetics, isoflurane, ketamine/xylazine, and sodium pentobarbital. The effects of changing the content of oxygen in the breathing gas was investigated and found to change the cerebral pO 2. In experiments with gerbils, crystals of lithium phthalocyanine were implanted in each side of the brain and using a one-dimensional magnetic field gradient, simultaneous measurement of pO 2 values from normal and ischemic (ischemia induced by unilateral ligation of a carotid artery) hemispheres of the brain were obtained. These results demonstrate that EPR oximetry with lithium phthalocyanine is a versatile and useful method in the measurement of cerebral pO 2 under various physiological and pathophysiological conditions.
Keywords: Cerebral pO2; EPR oximetry; lithium phthalocyanine; anesthesia; ischemia
I. Introduction The brain has a high metabolic rate and depends primarily upon aerobic metabolism. Because of these characteristics, brain tissue is very sensitive to changes in the availability of oxygen and pathological changes can occur if there are significant changes in the delivery a n d / o r utilization of oxygen in local regions of the brain. Complex and effective physiological controls have evolved to enable the brain to maintain an adequate level of oxygen within the tissues. Knowledge of the control mechanisms for maintaining local partial pressure of oxygen (pO 2) and the conditions which lead to their failure to prevent patho-
* Corresponding author. Fax: (1) (603) 650-1935. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 0 4 1 3 - 0
logical changes is incomplete, because of the difficulty of making measurements of local pO 2 under various physiological and pathophysiological conditions. Such knowledge seems essential for devising optimal strategies for diagnosis and therapy for the many processes which can lead to hypoxic damage to the brain, especially injury from ischemia-reperfusion. The most common way to measure brain pO 2 has been to use oxygen sensitive electrodes, by insertion of small microelectrodes [12] or by surface multi-electrode arrays [9]. The former may cause mechanical injury to the tissue and repeated measurements at the same site are not possible. The latter method cannot make accurate measurements of local pO 2 deep in brain tissue. A non-invasive optical method has been developed recently, using the oxygen quenching of phosphorescence [24], but it requires intravenous injection of porphyrins and measurements are lim-
92
K.J. Liu et aL / Brain Research 685 (1995) 91-98
ited to a depth of 1 mm from the surface. Near infrared spectroscopy also has been used, but it measures cerebral blood flow, not the interstitial pO 2 [8]. Recently there has been significant progress in the use of electron paramagnetic resonance (EPR, or completely equivalently, electron spin resonance or ESR) spectroscopy for measurements of oxygen in vivo in conscious animals [19]. This has resulted from the development of low frequency EPR instruments (1 GHz or lower) suitable for use with living subjects [6,13] and the development of new stable paramagnetic particles (fusinite, India ink, lithium phthalocyanine) with oxygen sensitive EPR spectra [10,18,23]. These particles combine several desirable properties as probes of the pO 2 including high sensitivity to the pO 2, resistance to chemical reactions, and a high degree of inertness in biological systems. The exact location of the paramagnetic particles in tissues can be determined by magnetic resonance imaging (MRI) [2]. These developments have established EPR oximetry as a versatile tool for sensitive, non-invasive, and reproducible measurements of pO 2 in a variety of biological systems. In this paper we report the use of EPR oximetry with lithium phthalocyanine (LiPc) particles for the measurement of cerebral interstitial pO 2 in conscious rodents. The aims of the study were to assess the effectiveness of EPR oximetry in the measurement of cerebral pO 2, and to demonstrate the usefulness of the technique for in vivo brain studies by monitoring the effect of anesthesia, ischemia, and different breathing gases on cerebral pO 2. LiPc was selected for pO 2 measurements in brain because of its small size and high sensitivity which makes it particularly suitable for oximetry of neural tissues. MRI was used to document the location of the LiPc and to assess potential cerebral changes associated with its presence. Assessment of the post-implant tissue effects was made by microscopic examination of histologic sections of gerbil brains taken in the region where the LiPc was placed and in the same plane as the MR images.
2. Materials and methods 2.1. Materials
Lithium phthalocyanine crystals were obtained from Dr. M. Moussavi (LETI, Grenoble, France). Details concerning synthesis, characterization and calibration of LiPc for pO 2 measurements have been reported [10,21,22]. The line width of the EPR signal is a linear function of pO 2, and is independent of the type of tissue being measured, local metabolic processes, the presence of other paramagnetic species, or pH. The high density of unpaired spins (8 × 1019 spins/cm 3) combined with a narrow intrinsic line width of LiPc allows measurements of pO 2 in the brain using one or more crystals with a diameter of ~ 200 /zm.
2.2. A n i m a l preparation
Male Wistar rats weighing 200-250 g were obtained from Charles River Laboratories (Wilmington, MA). The male Mongolian gerbil (100 gram, Tumblebrosk Farm, PA) was chosen as the ischemia model because of the reportedly high percentage of animals which have an incomplete Circle of Willis [3,5]. An incomplete Circle of Willis allows for the production of unilateral cerebral ischemia via unilateral carotid occlusion. Twenty-four hours prior to EPR measurements the animals were anesthetized with ketamine/xylazine (100/10 m g / k g , i.m.), and LiPc crystals implanted in the cerebral cortex. The LiPc crystals were positioned by drilling a small hole (500 /zm in diameter) in the cranium followed by insertion of a needle containing the crystals. The crystals were then deposited by pushing a stylus through the barrel of the needle to a specified depth (4 + 1 mm). In the case of gerbils, LiPc crystals were similarly introduced into both left and right hemispheres through two holes (6 mm apart) in the skull. Prior to EPR measurements the spatial localization of the crystals was verified by MRI. Following final pO 2 measurements, the brains were removed and cut in the same plane as the MR image. The brains were fixed in 4% neutral-buffered formalin, imbedded in paraffin, stained with hematoxylin and eosin (H and E), and examined microscopically. 2.3. Electron paramagnetic resonance
The spectra of LiPc were obtained using an EPR spectrometer constructed in our laboratory with a low frequency (1.2 GHz, L-band) microwave bridge [13]. An extended loop resonator was positioned over the brain of the animal. Typical settings for the spectrometer were: incident microwave power, 10 mW; magnetic field center, 425 gauss; scan range, 1 gauss; modulation frequency, 27 kHz. Modulation amplitude was set at less than one-third of the EPR line width, typically around 20-50 mG. The linear magnetic field gradient, used to separate the signals originating from crystals in the two hemispheres of the brains of gerbils, was produced by a pair of coils in an anti-Helmholtz configuration capable of generating gradients along the B 0 up to 0.5 m T / c m . The magnitude of the gradient was calibrated by measuring the peak-to-peak splitting of two capillaries containing LiPc under nitrogen. Both unanesthetized and anesthetized animals were studied. The unanesthetized animals were restrained in a specially constructed device. To reduce tuning problems associated with movements of the unanesthetized animal, we constructed an external loop resonator tuned by a varactor diode and altered the automatic frequency control (AFC) system [15]. In this new system, the AFC error signal tunes the resonator to a fixed frequency. By this means, a shift of resonant frequency in the resonator due to
K.J. Liu et al. / Brain Research 685 (1995) 91-98
93
70
i
i
I
i 10
t 20
J 30
3.6
60 m 5O
3.3 O
4o
/ -+-
3o
20
10
10
I
I
I
I
I
I
20
30
40
50
60
70
3.0
/ o
2.7
0
80
T i m e (rain)
Inspired o x y g e n (~*)
Fig. 1. The relationship of the cerebral p O 2 of rats with oxygen content in the breathing gas. The unanesthetized animals were restraincd in a homc-made device, and wcre conscious throughout thc cxperiment. W e waited at least 20 rain after change of oxygcn content to allow p O 2 to rcach a new equilibrium. The inscrt shows a typical E P R spectrum of LiPc recorded from the brain. Ccrcbral p O 2 valucs arc cxpressed as mean + S E M (n = 3 animals).
animal motion can be compensated, keeping the frequency constant even if the animal moves. Scan time was 1 min. In some experiments such as kinetic studies, the data points from single scans were used directly. For steady state experiments, 5 scans usually were accumulated. The total EPR scan time for any measurement was less than 10 min, including positioning of the animal, adjusting the spectrometer, and accumulation of 5 scans. The EPR line widths were converted to pO 2 using a calibration curve determined for each batch of crystals.
2.4. Measurement of cerebral p O 2
Fig. 3. The samc data as in Fig. 2 to illustratefirst order kinctics of the
changes of brain pO 2 after pcrturbation. Data wcre transformed by taking logarithm of pO 2 values. Time zero is the starting time of switching inspired oxygen from 21% to 10.5% (O), or from 10.5% to 21% (Q).
allowed regulation of the breathing gases. Cerebral pO 2 was measured at various inhaled pO 2 and the breathing gas was monitored using an Oxygen Analyzer (Sensor Medics, Model OM-11, Anaheim, CA). To follow the time course of changes in cerebral pO 2 after the oxygen content in the breathing gas was switched from 21% to 10.5%, and back to 21%, the EPR spectrum was recorded repetitively at 60 s intervals. For the study of the effect of anesthesia on pOz, measurements were made before and after the peritoneal injection of a ketamine/xylazine mixture, or sodium pentobarbital, or inhalation of the gas anesthetic, isoflurane, which was delivered to the animals through the nose-cone of the restraining device. Occlusion of the common carotid artery was accom-
Unanesthetized rats were restrained in the specially constructed device with a nose-cone at one end, which 40 50
3o 4O
~
o°~ o
3O
isof
oo eJ
***
ket/xyl
•
~•~
2o
~
pen
~d
20
% o
L)
o:~oe
lo
10 •
21~, 02
10.5~ 02
o
I
I
I
I
20
40
60
80
100
Time (rain) Fig. 2. Time course of the change in cerebral pO 2 after perturbation in the oxygen content in breathing gas. The breathing gas was switched between 21% oxygen ( O ) and 10.5% (©). EPR spectra were recorded every minute.
0
- -
- -
- -
Fig. 4. Effect of anesthetics on the cerebral pO 2. Ketamine/xylazine mixture (100/10 mg/kg) and sodium pentobarbital (50 mg/kg) were injected into the rats intraperitnneally. Isoflurane (1%, flow rate 3 ml/min.) was delivered to the animal through a nose cone in restraining device. The values are expressed as mean +_S.E.M. ( n = 3 animals). • P < 0.003, * * P < 0.001, * * * P < 0.0001 vs. control.
94
K.J. Liu et al. / Brain Research 685 (1995) 91-98
50
left common carotid artery isolated. Following carotid isolation, blood flow was occluded by tying off the artery with elastic suture material (occlusion knots with this type of suture could be released without damage to the artery). The skin incision was then sutured closed and the gerbil positioned in ventral recumbency for EPR p02 measurement.
40
30
~eu
2o
2.5. M a g n e t i c r e s o n a n c e i m a g i n g
10
0 Day 1
Day 3
Day 7
Fig. 5. Measurement of cerebral pO2 over time after implantation of LiPc. EPR spectra were collected at day 1, 3 and 7 after implantation from unanesthetizedrats while the animals breathed 21% oxygen (11) or 10.5% oxygen ([]). The values are expressed as mean±S.E.M. (n = 3 animals). plished by anesthetizing the gerbils (ketamine/xylazine) and positioning them in dorsal recumbency. A 1-cm longitudinal skin incision was made in the ventral neck. Underlying subcutaneous tissues were bluntly dissected and the
MR images of the brain of anesthetized animals were obtained using a 7.0 T imager with a horizontal bore Magnex magnet and a SMIS console. Visualization of the LiPc in the brain was achieved using a bird-cage RF coil (d = 6 cm) and gradient-recalled echo (GRE) sequence ( 2 0 0 / 1 0 / 4 5 °) with FOV = 4 cm; matrix size = 128 × 128; slice t h i c k n e s s = 1 mm; and 4 acquisitions. T2weighted spin-echo images ( 2 0 0 0 / 6 0 ) were obtained in the same plane as the GRE images.
3. Results Fig. 1 shows that the cerebral pO 2 increased when the oxygen content in the breathing gas was increased. Fig. 2
Fig. 6. Axial MR images of rat brain with implanted LiPc crystals (arrows). A,C: gradient-recalledecho (200/10/45° ) and B,D: spin-echo(1800/50) images were obtained 1 day (top row) and 7 days (bottom row) after implantationof the LiPc.
K.J. Liu et a l . / Brain Research 685 (1995) 91-98
C
mmHg
95
shows the time course of changes in cerebral pO 2 when the breathing gas changed from 21% oxygen to 10.5%, and back to 21%. When the logarithm of the changes of p O 2 was plotted against time, straight lines were obtained, indicating that the changes of cerebral p O 2 followed first order kinetics (Fig. 3). The calculated half-life for reaching a new equilibrium was 23 min when the oxygen in the mixture was decreased from 21% to 10.5%, and 15 min in the reverse direction. The effect of anesthesia on the cerebral p O E w a s studied using three representative anesthetics, isoflurane, ketamine/xylazine, and sodium pentobarbital. As shown in Fig. 4, all three anesthetics substantially reduced the cerebral p O E. Sodium pentobarbital had the largest effect decreasing the pO 2 approximately 50%. To determine the feasibility of measuring cerebral p O 2 over extended periods of time with LiPc, we measured the EPR spectra at 1, 3 and 7 days after implantation of the crystals. Fig. 5 shows that p O 2 remained constant over this period of time, and when the breathing gas was changed from air to 10.5% 0 2, the cerebral p O 2 decreased reproducibly. Fig. 6 demonstrates the ability of MRI to locate LiPc crystals within the brain. The GRE technique greatly increases the apparent size of the crystals on the image ( ~ 1 mm). Considering the in-plane resolution of 0.3 mm and the slice thickness of 1 mm it should be very difficult to visualize an object the size of the LiPc crystal because of insufficient in-plane resolution and volume averaging. The ability to visualize LiPc crystals on MR images under such conditions is based on bulk magnetic susceptibility differences between the tissue and the crystal, which produces a susceptibility-based signal attenuation around the crystal which is much larger than the actual crystal [2]. This T 2 effect can be easily visualized on gradient-echo sequences (Fig. 6A, C), but is considerably smaller on spin-echo sequences (Fig. 6B,D) even when relatively heavily T2weighting is applied. This is illustrated in Fig. 6 where MRI was used to determine that there were no gross changes (e.g. edema) observed on images either acutely (Fig. 6B) or one week (Fig. 6D) following implantation of LiPc. Histologic assessment of the brains showed that MRI provides an accurate determination of the location of the LiPc, and that under the conditions of this experiment there is little tissue reaction to the presence of LiPc, or to the implantation procedure. Simultaneous measurements of p O 2 in both cerebral
\ Ischeinic p 0 2 = 11.4 rnrnHg I
399.8
i
I
,
Magnetic
I
400.2
400.0
field
(G)
Fig. 7. (A, B) Two contiguous axial MR images (GRE-100/7/45 ° ) of gerbil brain showing two LiPc crystals implanted in the cerebral cortex. (C) EPR spectrum of the LiPc crystals in gerbil brain after unilateral carotid occlusion (anesthetized animal) demonstrating the simultaneous assessment of pO 2 in the ischemic (right) and non-ischemic (left) cerebral hemispheres. A magnetic field gradient of 0.05 mT/cm was used.
96
K.J. Liu et al. / B r a i n Research 685 (1995) 91-98
hemispheres via bilateral LiPc implantation in both sides of brains (Fig. 7A, B) was successfully accomplished in the gerbil. Using this technique, we were able to study the effects of unilateral carotid occlusion, and measure the pO 2 in the ischemic and non-ischemic brain in the same animal simultaneously. Ten minutes after unilateral carotid occlusion, pO 2 was measured in both hemispheres using a magnetic field gradient of 0.05 m T / c m to separate the spectra from the two sites. Carotid occlusion resulted in a significant reduction of pO2 in the corresponding side of the brain; the pO 2 decreased to 11.4 mmHg compared with 24.1 mmHg in the non-occluded side (Fig. 7C). Because the gerbil was anesthetized with ketamine/xylazine, the observed pO 2 in the non-ischemic side of the brain was lower than in conscious animals and was consistent with the value obtained in the studies of the effects of anesthetics (Fig. 4).
4. Discussion These studies demonstrate that EPR oximetry is capable of assessing cerebral pO 2 in vivo with several favorable features, which are summarized in the following paragraphs. As illustrated in Fig. 5, pO 2 measurements can be made repeatedly on the same animal and at the same location without anesthesia or other invasive procedures, whereas in the use of electrodes, repeated insertions over time are likely to be perturbing and difficult to place at exactly the same location. The location of the site where the pO 2 is being measured by oxygen sensitive materials, such as LiPc, fusinite, and India ink can be verified histologically a n d / o r by MRI [2], since the paramagnetic material themselves are markers for the location where the pO e measurements are obtained. EPR oximetry can provide chronological studies of pO 2 over intervals ranging from short term variation (minutes) to long-term (days or weeks) follow-up studies of changes of pO 2 following occlusion or any other treatment (Fig. 2). Because the response time of LiPc to changes in pO 2 is immediate [10], EPR oximetry with LiPc measures the actual kinetics of the processes. Measurements of pO 2 by LiPc are highly localized. The pO 2 measured is the average pO 2 in the tissue that are adjacent to the paramagnetic particles, i.e., tissues in immediate equilibrium with the surface of the LiPc crystals. These materials can be as small as 200 /xm in diameter, therefore the pO 2 in any particular region of the brain can be selected for measurement of pO 2. Because of this capability to have localized measurements based on the placement of the particles, multisite (two or more) simultaneous measurements in different regions of the brain are possible, as shown in Fig. 7. Spatial resolution in the range of 2 mm easily can be achieved without significant distortion of the EPR signal.
This is a particularly useful feature of EPR oximetry because the measurement from one region can be used as an internal control while the signals from other regions are being monitored as indicators of the experimental procedures. EPR oximetry is non-invasive after the initial placement of the crystals and the measurements of pO 2 can be made repetitively at will without anesthesia since the LiPc is inert and remains responsive to changes of oxygen. The observed cerebral pO 2 value of 34.1 ___3.2 mmHg under normal conditions is consistent with the value of 36 mmHg reported using electrodes [12]. The cerebral tissue pO 2 measured under room air and 10.5% oxygen conditions are in good agreement with tissue pO 2 previously estimated under these conditions [20]. Thus it appears that EPR oximetry provides results similar to those obtained with microelectrodes but it offers several favorable features as described above, and in most situations, yields more precise data [17]. The EPR oximetry measurements indicated that it took longer to reach a new equilibrium when the oxygen in the breathing gas was decreased from 21% to 10.5% than in the reverse direction. Since the response time of LiPc to changes of pO 2 is virtually instantaneous [10], and the response time of blood pO 2 usually takes seconds to minutes, the observation of a slow decrease in cerebral pO 2 after an acute drop in inspired oxygen may reflect the operation of a physiological control mechanism which attempts to maintain the pO 2 in the brain in the presence of decreased availability of oxygen. The drop in pO 2 after administration of the anesthetics agents is likely related to depressed respiration induced by the anesthetics [14]. The results indicate the difficulties that may arise when studying oxygen dependent processes (including processes affected by oxidative metabolism) in anesthetized animals. In addition to changes in the pO 2 in various organs, anesthetics also are likely to alter the complex physiological control mechanisms that strive to maintain tissue pO 2. Therefore experimental manipulations which restore the pO 2 of the blood to normal levels may not compensate fully for the pO 2 changes in all tissues a n d / o r the responses of the systems to further experimental manipulation. We believe that EPR oximetry may provide a good means to investigate these problems in depth. Multisite spectral oximetry ideally should be performed in such manner that the applied magnetic field gradient is sufficient to separate signals from different sites, but not large enough to cause the distortion of the local crystal line-width which could give erroneous values of local pO 2. The degree of distortion depends on the relationship between the size of the crystal, applied gradient and intrinsic line width (i.e. linewidth in the absence of the gradient). The gradient strength used here (0.05 m T / c m ) allows good separation of signals from the two hemispheres of the rodent brain. Selection of the gradient strength was based
K.J. Liu et al. /Brain Research 685 (1995) 91-98
on knowledge of the distance between crystals (MRI) and value of linewidth (0.025 mT) under normal oxygenation of the brain. If necessary, line distortion induced by a gradient can be corrected, provided that the above parameters are known. It was not necessary in these studies, however, since it has been shown both theoretically and experimentally that a gradient of 0.05 m T / c m does not cause any significant distortion of the EPR line of LiPc [16]. The combination of EPR oximetry with MRI techniques can be quite useful. GRE MRI was used here to determine the site of the LiPc and could be repeated as necessary to confirm the localization during long-term EPR studies of oxygen changes in the brain. Spin echo MRI was used to show that there were no detectable changes in the MR image in the vicinity of the LiPc, indicating that there was not extensive tissue damage associated with the presence of the LiPc. It is important to note that susceptibility effects used for locating the paramagnetic materials do not cause severe effects on spin-echo images so that implantation of the LiPc does not preclude application of MRI techniques for assessment of physiological changes associated with ischemia such as diffusion weighted imaging [4,11]. Likewise, the presence of LiPc should not affect assessment of physiological changes by localized MR spectroscopy [1], even if the volume of interest (VOI) has to be placed directly over the position of the crystal, since the typical VOI for 31p or I H spectroscopy is around 100 mm 3 while the volume affected by LiPc is less than 1 mm 3. We emphasize this feature because both 31p spectroscopy and DWI have been advocated for non-invasive assessment of ischemia, but both techniques measure oxygen dependent changes such as ATP depletion and redistribution of the water as a consequence of changes in metabolism, but not pO 2 itself. Presently it is not known how these quantities are related, but by combining these techniques with in vivo EPR oximetry in the same animal model it should be possible to assess the correlation and 'calibrate' M R I / M R S findings [7]. In conclusion, these studies indicate that EPR oximetry can make noninvasive, accurate, and repeated measurement of cerebral p O E at multiple sites simultaneously in conscious animals. Our results indicate that cerebral p O E is altered by different classes of anesthetics and the amount of oxygen in the breathing gas. The results also demonstrate that EPR oximetry can assess alterations in cerebral pO 2 associated with experimentally produced cerebral ischemia.
Acknowledgements This research was supported by NIH Grant GM 34250 and used the facilities of the IERC at Dartmouth supported by NIH Grant RR-01811.
97
References [1] Allen, K.L., Busza, A.L., Proctor, E., King, M.D., Williams, S.R., Crockard, H.A. and Gadian, D.G., Controllable graded cerebral ischemia in the gerbil: studies of cerebral blood flow and energy metabolism by hydrogen clearance and 31p NMR spectroscopy, NMR in Biomed. 6 (1993) 181-186. [2] Bacic, G., Liu, K.J., O'Hara, J.A., Harris, R.D., Szybinski, K., Goda, F. and Swartz, H.M., Oxygen tension in a murine tumor: a combined EPR and MRI study, Magn. Reson. Med., 30 (1993) 568-572. [3] Bothe, H.W., Bosma, H.J., Hofer, H., Hossmann, K.A. and Angermeier, W.F., Selective vulnerability of hippocampus and disturbances of memory storage after mild unilateral ischemia of gerbil brain, Stroke 17 (1986) 1160-63. [4] Busza, A.L., Allen, K.L., King, M.D., van Bruggen, N., Williams, S.R. and Gadian, D.G., Diffusion-weighted imaging studies of cerebral ischemia in gerbils. Potential relevance to energy failure, Stroke, 23 (1992) 1602-1610. [5] Clifton, G.L., Taft, W.C., Blair, R.E., Choi, S.C. and DeLorenzo, R.J., Conditions for pharmacologic evaluation in the gerbil model of forebrain ischemia, Stroke, 20 (1989) 1545-1552. [6] Colacicchi, S., Alecci, M., Gualtieri, G., Quaresima, V., Ursini, C.L., Ferrari, M. and Sotgiu, A., New experimental procedures for in vivo L-band and radio frequency EPR spectroscopy/imaging, J. Chem. Soc. Perkin Trans., 2 (1993) 2077-2082. [7] Dunn, J.F., Ding, S., Hoopes, J.P., Liu, K.J. and Swartz, H.M., Relating the apparent diffusion constant to brain interstitial pO2, Proc. Soc. Magn. Res., 2 (1994) 1006. [8] Elwell, C.E., Cope, M., Edwards, A.D., Wyatt, J.S., Reynolds, E.O.R. and Deply, D.T., Measurements of cerebral blood flow in adult humans using near infrared spectroscopy-methodology and possible errors, Adv. Exp. Med. Biol., 317 (1992) 235-245. [9] Grota, J., Zimmer, K. and Schubert, R., Tissue oxygenation in normal and edematous brain cortex during arterial hypocapnia, Adv. Exp. Med. Biol., 180 (1984) 179-184. [10] Liu, K.J., Gast, P., Moussavi, M., Norby, S.W., Vahidi, N., Walczak, T., Wu, M. and Swartz, H.M., Lithium phthalocyanine: a probe for electron paramagnetic resonance oximetry in viable biological systems, Proc. Natl. Acad. Sci. USA, 90 (1993) 5438-5442. [11] Mintorovich, J., Moseley, M.E., Chileuitt, L., Shimizu, H., Cohen, Y. and Weinstein, P.R., Comparison of diffusion-and T2-weighted MRI for the early detection of cerebral ischemia and reperfusion in rats, Magn. Reson. Med., 18 (1991) 39-50. [12] Nair, P.K., Buerk, D.G. and Halsey, J.H., Comparison of oxygen metabolism and tissue pO E in cortex and hippocampus of gerbil brain, Stroke, 18 (1987) 616-662. [13] Nilges, M.J., Walczak, T. and Swartz, H.M., 1 GHz in vivo ESR spectrometer operating with a surface probe, Phys. Med., 5 (1989) 195-201. [14] Pavlin, E.G. and Hornbein, T.F., Anesthesia and the control of ventilation. In Handbook of Physiology, Section 3, Vol. lI, Control of Breathing, Part 1, pp. 793-813. [15] Smirnov, A.I., Norby, S.W., Walczak, T., Liu, K.J. and Swartz, H.M., Physical and instrumental considerations in the use of lithium phthalocyanine for measurements of the concentration of oxygen, J. Magn. Reson. B., 103 (1994) 95-102. [16] Smiruov, A.I., Norby, S.W., Clarkson, R.B., Walczak, T. and Swartz, H.M., Simultaneous multi-site EPR spectroscopy in vivo, Magn. Reson. Med., 30 (1993) 213-220. [17] Swartz, H.M., Goda, F., Rhodes, E.S., O'Hara, J.A., Liu, K.J. and Dunn, J.F., Measurement of pO z in tissue in vivo by EPR: principles, applications, and comparison with electrode techniques, Proc. Soc. Magn. Res., 2 (1994) 754. [18] Swartz, H.M., Liu, K.J., Goda, F. and Walczak, T., India ink: a potential clinically applicable EPR oximetry probe, Magn. Reson. Med., 31 (1994) 229-232.
98
K.J. Liu et al. / Brain Research 685 (1995) 91-98
[19] Swartz, H.M. and Walczak, T., In vivo EPR: prospects for the 90's, Phys. Med., 9 (1993) 41-48. [20] Tenney, S.M. and Ou, L.C., Physiological evidence for increased tissue capillarity in rats acclimatized to high altitude, Respirat. Physiol., 8 (1970) 137-150. [21] Turek, P. andre, J.J., Giraudeau, A. and Simon, J., Preparation and study of a lithium phthalocyanine radical: optical and magnetic properties, Chem. Phys. Lett., 134 (1986) 471-476. [22] Turek, P., Andre, J.J., Moussavi, M. and Fillion, G., Septet spin state in the lithium phthalocyanine radical compound, Mol. Cryst. Liq. Cryst., 176 (1989) 535-546.
[23] Vahidi, N., Clarkson, R.B., Liu, K.J., Norby, S.W., Wu, M. and Swartz, H.M., In vivo and in vitro EPR oximetry with fusinite: a new coal-derived, particulate EPR probe, Magn. Reson. Med., 31 (1994) 139-146. [24] Wilson, D.F., Gomi, S., Pastuszko, A. and Grindberg, J.H., Oxygenation of the cortex of the brain of cats during occlusion of the middle cerebral artery and reperfusion, Adv. Exp. Med. Biol., 317 (1992) 689-694.