NMR IN BIOMEDICINE, VOL. 5 , 145-153 (1992)
31PNMR Spectroscopy of Perinatal Hypoxic-ischemic Brain Damage: a Model to Evaluate Neuroprotective Drugs in Immature
Gerald D. Williams,$ Charles Palmer,$ Rebecca L. Roberts,§ Daniel F. Heitjantt and Michael B. Smith+$$* Departments of $Radiology (Center for NMR Research), PPediatrics (Division of Newborn Medicine), P~Cellularand Molecular Physiology, and Biological Chemistry and ttCenter for Biostatistics and Epidemiology, The Pennsylvania State University College of Medicine, The Milton S. Hershey Medical Center, Hershey, PA 17033, USA
Cerebral energy metabolism can be measured non-invasively in unanesthetized neonatal rats with 3'P NMR spectroscopy. Using this technique, serial changes in high energy phosphates were determined from the right cerebral hemispheres of 7 day postnatal rat pups during a hypoxic-ischemic insult known to produce focal brain injury. During 3 h of hypoxia-ischemia the concentration of ATP dropped to 33 If:8% of prehypoxic (baseline) levels, phosphocreatine (PCr)/Pi decreased from 1.5 & 0.51 to 0.16 f 0.06, while pH decreased nominally by 0.2 units. After 2.5 h of recovery in air, ATP returned to 75 k 10% of baseline levels, PCr/Pi rose to 1.1+- 0.28, and pH returned to its normal value of 7.16f0.06. This model was used to test the efficacy of the adenosine deaminase inhibitor, 2-deoxycoformycin (DCF) as a potential neuroprotective drug. The data for the drug- and saline-treated populations were analyzed by integrating ATP and Pi/PCr levels over specific time intervals, expressing it relative to baseline levels, and modeling it with cubic splines. Pretreatment with 500 pg/kg DCF shows a small, but statistically significant, preservation of both ATP and phosphorylation potential during hypoxia and initial recovery. Brain water content (edema) at 42 h recovery was apparently associated with both mean ATP and mean Pi/PCr in the last 2 h of hypoxia-ischemia. When ATP fell below 70% of baseline, brain edema was evident at 42 h of recovery. This methodology is suitable for extension to human infants.
INTRODUCTION
Perinatal hypoxia-ischemia (asphyxia) is a major cause of mortality and morbidity in the human newborn.' In this study we used a well established animal model of hypoxic-ischemic (HI) brain damage."' Cerebral metabolism was measured serially with 31PNMR spectroscopy during and following the combined HI insult to 7 day postnatal rat pups. The rat pup model consists of right common carotid ligation followed by 3 h of hypoxia in 8% oxygen. It produces perturbations in tissue energy metabolism and brain damage in the posterolateral region of the right cerebral hemisphere. Blood flow to the middle cerebral artery supply area is reduced to 15-35% of control values during hypoxiaischemia and is accompanied by increased glucose utilization.'.". Numerous studies of brain energy metabolism have been performed by in uiuo "'P NMR on anesthetized adult rats.s-" However, no study has yet been published using "P spectroscopy during cerebral hypoxiaischemia in unanesthetized neonatal rats. Extrapolation of results from adult rats to neonates is not valid. as cerebral blood flow and metabolism are
'
Author to whom correspondence should he addressed.
t Preliminary accounts of this work were presented at the 8th Annual Meeting of the Society for Magnetic Resonance in Medicine, Amsterdam. The Netherlands. 1989.' Abbreviations used: HI, hypoxic-ischemic: DCF. 2-deoxycoformycin: PME. phosphomonoester; PCr. phosphocreatine 0952-3480/92/030145-09 $09.SO
0 1992 by John Wiley gL Sons. Ltd.
slower, but pathologic reactions are faster in immature animals.' The rat pup, content in a warm, dark restricted space, is ideal for the study of unanesthetized cerebral metabolism by high resolution NMR. As the brain of the 7 day old rat is still immature, it provides a model for studying the effects of hypoxia-ischemia on the developing nervous system of the newborn, and the opportunity to monitor the potential benefits of neuroprotective drugs. Accordingly, we aimed to determine if inhibition of adenosine deaminase with 2-deoxy~oformycin'~~'~ (DCF; pentostatin, ParkeDavis Pharmaceuticals Inc., Ann Arbor, MI, USA, personal communication) could preserve energy metabolism during cerebral hypoxia-ischemia. Pretreatment with DCF increases adenosine levels as well as cerebral blood flow during hypoxia due to the rise in adenosine concentration." Adenosine has other potentially neuroprotective mechanisms. ". 'I If DCF increases the concentration of adenine nucleotide precursors and prevents catabolism of ATP, elevated levels of high energy phosphates should be detected by 3'P NMR during hypoxia-ischemia and early recovery. Our purpose was to use "P NMR to (i) measure and develop methods to quantitate the changes in high energy phosphates using a neonatal rat model of cerebral hypoxia-ischemia which reliably causes brain injury, (ii) determine if pretreatment with a potential neuroprotective agent, specifically DCF, will alter energy metabolism during the insult or early recovery and (iii) correlate these changes with the development of cerebral edema. Received 4 September 1991 Accepted (revised) 8 October 199I
I46
G . D. WILLIAMS ET AL.
MATERIALS AND METHODS Animal preparation and analysis Unsexed 7 day old Wistar (Charles River) rat pups weighing between 12 g and 18 g were subjected to right carotid artery ligation as described by Palmer et After surgery, the rat pups were returned to their dams for 2.5 h. Immediately before placement in the NMR probe, which was 60 min prior to the onset of hypoxia, pups were injected S.C. with physiological saline or DCF (Parke-Davis Pharmaceuticals Inc.). The DCF concentration was 5(H) or 50 pg/kg (DCF-500 or DCF-50). The choice of saline vs drug administration was randomized to minimize litter bias. Each rat pup had its front legs secured with a thin strip of surgical tape and was wrapped in a 4 x 4" gauze pad to restrict movement. Following NMR spectroscopy, animals were returned to their dams, and hemispheric water content was determined at 42 h of recovery when brain swelling is advanced. Hemispheric water content was computed as a percentage according to the formula: [(total wt -dry wt)/total wt] x 100. Parallel experiments were performed for determination of the effect of DCF on mortality. The pups were placed in 500 mL airtight jars which were partially submerged in a water bath at 36.5 "C. Air temperature in the center of the jar was 32-33°C. The pups were exposed to an 8% oxygen mixture for 3 h. Then the jars were opened to room air and the survivors returned to their dams.
Probe design and animal placement All studies were performed on a Bruker AM400 widebore spectrometer at 162.0MHz for 31P using a specially constructed NMR probe. The probe served as both an NMR RF receiver and a gas-tight chamber for control of oxygen and temperature. It was modified from a standard Bruker wide-bore probe body such that the top portion consisted of a gas-tight plexiglass environmental chamber. The 70 mm diameter cylindrical chamber contained a removable panel for easy access and positioning of the animal with respect to the 7 x 10 mm elliptical single-turn surface coil. The R F circuit was completed by adjustable tuning and matching capacitors (Jackson Brothers Ltd., London, U K ) located in close proximity to the coil at the top of the chamber. The chamber temperature was regulated by heated water from a temperature bath (Haake, Berlin, Germany), carried through a continuous polyethylene tubing which was wound and fixed into a coil at the back side of the animal chamber. The temperature adjacent to this heating pad was monitored with a standard Bruker thermocouple arrangement to be 36 f 0.5 "C during the entire experiment. A lower ambient temperature was recorded in close proximity to the skin, and was also constant during the experiment. Oral temperature was determined to be 36 "C in selected pups. In addition to an entrance at the bottom of the probe for heated water, an entrance for tubing to carry
a gas mixture was provided. The top open end of the chamber was sealed with plastic wrap, allowing maintenance of the calibrated gas mixture. A thin soft bandage (SpencoTM, Spenco Medical Corp., Waco, TX, USA) secured the rat's head to the surface coil with minimal discomfort. The NMR experiment was performed without anesthesia, with the rat in a vertical orientation inside the magnet. Careful positioning of the region of injury in the right hemisphere with respect to the coil assured minimal contamination from the left (contralateral) hemisphere (Fig. 1). Additional gauze material and a foam rubber casing were added prior to sealing the animal access door in order to further restrict movement and maintain body temperature. A standard Bruker aluminum alloy probe cover was used to provide structural integrity and RF shielding while the probe was in the magnet.
N M R experiment Following tuning to the "P frequency, the probe was placed in the magnet for optimization of field homogeneity using the 'H signal from the water protons of the same region of the hemisphere as used in the phosphorus spectroscopy. 'H shimming was accomplished with a double tuned circuit located external to the magnet. ?'P spectra were acquired with a single-pulse sequence using a pulse width of 11 ps, a repetition time of 2 s, acquisition time of 0.133s and a sweep width of 7.7 kHz. A typical resulting "P spectrum is shown in Fig. 1. Twenty-two spectra were acquired in 7.5 or 15 min blocks beginning with a 23min prehypoxic period during which three spectra were accumulated to obtain baseline levels. Eight per cent O2was supplied for the 3 h hypoxic period followed by a 2.5 h recovery in medical air. Gas delivery was maintained at constant levels at all times with a flow meter (Matheson Gas Products, East Rutherford, NJ, USA) located in series with the air tanks and the NMR probe.
Data analysis
A Lorentzian line broadening of 40 Hz and a 500 Hz profile correction function" to remove the broad lipid/ bone component were applied prior to Fourier transformation to 4 K data points. Spectra were analyzed by peak areas relative to an appropriate baseline. Slight overlap of P, with phosphomonoester (PME) was corrected by using the Bruker spectral simulation program 'Linesim'. Independently determined saturation factors to correct for incomplete relaxation were applied to the NMR peak areas. The pH was determined from the frequency of P,, calculated from the chemical shift difference between PI and phosphocreatine (PCr). The difference was compared with a standard 37°C pH calibration curve which uses a PK ,= ~ 6.75, and acid and base endpoints of d,,= 3.27 and db= 5.69.'' The time-series data were analyzed by two approaches. In the first method, the average ATP, P,/PCr and pH were computed within animals in both the last 2 h of hypoxia-ischemia and the initial recovery
I47
"P NMR OF HYPOXIC-ISCHEMIC BRAIN INJURY IN RATS
ATP
PME
Figure 1. Illustration of the head of a 7 day old rat pup. The broken line shows the region of the right hemisphere of the brain, while the solid line within that region shows the position of the elliptical surface coil, where injury is maximal. An 8 min 31Pspectrum (2 s repetition time) from this region is shown. Line broadening of 20 Hz was applied and the broad component was not filtered.
period. Intra-animal means were computed by the formula:
c h
Pub=
a
h
Pj Atj/N
Atj U
where P represents the parameter of interest (ATP, Pi/PCr or pH), j is the spectrum number, At, is the accumulation time for spectrum j , a and b are the lower and upper limits of the time range of interest and N is the baseline value of the parameter obtained during prehypoxia. Thus each animal yielded an intra-animal mean, normalized to baseline levels, and group averages of these mean values were compared by a pooledvariance two-sample t-test. This method is valid since data from different pups are statistically independent, but it leads to conservative p values because it uses the data inefficiently. In a second approach, cubic spline models were fit to the data using the method described by Heitjad3 and applied in English et aLZ4Briefly, the underlying mean curve is approximated by a cubic spline with knots at the times of initiation and terminaton of hypoxia." This error model accommodates both random animal effects and time-series a u t o ~ o r r e l a t i o n .The ~ ~ . model ~~ was fit by maximum likelihood, and parameter estimates were used to compute mean relative ATP, log(Pi/PCr) and pH in given time ranges. Estimated means in different treatment groups were compared by Wald tests.*' This method is both valid and efficient because it models the intra-animal error structures, thus making better use of the available data; it also avoids multiple comparisons problems associated with multiple t-testing.
RESULTS
Figure 2 shows the typical time course of metabolites during HI and recovery for a saline-treated rat pup. Eight out of the usual 22 spectra are represented. In this example, PCr/P, decreased from 1.7 to 0.25 during 3 h of hypoxia and ATP decreased to Since DCF did not reduce edema, it follows that the drug is unlikely to be effectual at reducing other measures of neuropathological brain damage. Mortality during hypoxia increased markedly with DCF dosage. With DCF-500 we were near the onset of toxicity (TDL,,= 500 pg/kg) reported in adult mouse'" but well below the half lethal dosage (LD,,, = 128 mg/kg; Parke-Davis Pharmaceuticals Inc., personal communication). The significant number of deaths in this study should be attributable to the added stress on the neonate from hypoxia-ischemia. The high mortality with DCF-500 must be remembered when evaluating the preservation of adenine nucleotides in those rats that survived. Although adenosine levels should increase significantly after DCF administration,lh ATP and phosphorylation potential levels were preserved in only selected animals. In summary, we have demonstrated that it7 uiuo serial assessment of brain energy metabolism can be obtained with this model using "'P NMR. A simple method of quantitation using mean values of ATP and P,/PCr, taken from hypoxia and initial recovery, has shown that there is a small preservation of high energy metabolites with DCF at 500pg/kg in animals that survive the 3 h o f hypoxia-ischemia. Analysis of the time-series data by cubic spline models" gave even smaller p-values. There was no significant reduction in subsequent cerebral edema with DCF-500. However, a threshold was found for predicting subsequent edema using either ATP or P,/PCr levels, integrated during the last 2 h of hypoxia. DCF was shown using both statistical methods of analysis to have n o effect on metabolite levels at a lower dosage of 50 pg/kg. We are currently in the process of evaluating other neuroprotective agents which both preserve metabolite levels and reduce brain damage.5' The methods ot analysis used in this study should be suitable for extension to human infants, and Cady and Azzopardi have demonstrated absolute quantitation of brain spectra from
152
G. D. WILLIAMS E T A L .
normal and birth asphyxiated infants5' In future experiments we would like to administer drug therapy following a period of H I insult in order to closely model the brain damage and treatment of the asphyxiated newborn human.
Acknowledgements 2-DCF was a gift from the Parke-Davis Pharmaceutical Research Division, a Warner-Lambert Co. We thank Scott Babe for assistance with the NMR analysis and Professor Robert C. Vannucci for helpful discussions
REFERENCES 1. Smith, M. B., Palmer, C. and Williams, G. D. 31P NMR spectroscopy of unanesthetized hypoxic-ischemic neonatal rats: a model t o evaluate neuroprotective drugs. 8th Annual Meeting of the Society of Magnetic Resonance in Medicine 1, 466 (1989). 2. Vannucci, R. C. Acute perinatal brain injury: hypoxiaischemia, in Management of Labor, ed. by W. R. Cohen, D. B. Archer and E. A. Friedman, pp. 138-244. Aspen Publishers, Inc., Rockville (1989). 3. Rice 111, J. E., Vannucci, R. C. and Brierley, J. B. The influence of immaturity on hypoxic-ischemic brain damage in the rat. Ann. Neurol. 9, 131-141 (1981). 4. Palmer, C., Brucklacher, R. M., Christensen, M. A. and Vannucci, R. C. Carbohydrate and energy metabolism during the evolution of hypoxic-ischemic brain damage in the immature rat. J. Cereb. Blood Flow Metabol. 10, 227235 (1990). 5. Palmer, C., Vannucci, R. C. and Towfighi, J. Reduction of perinatal hypoxic-ischemic brain damage with allopurinol. fediatr. Res. 27, 332-336 (1990). 6. Vannucci, R. C., Lyons, D. T. and Vasta, F. Regional cerebral blood flow during hypoxia-ischemia in immature rats. Stroke 19, 245-250 (1988). 7. Vannucci, R. C., Christensen, M. A. and Stein, D. T. Regional cerebral glucose utilization in the immature rat: effect of hypoxia-ischemia. fediatr. Res. 26, 208-214 (1989). 8. Horikawa, Y., Naruse, S., Hirakawa, K., Tanaka, C., Nishikawa, H. and Watari, H. !n vivo studies of energy metabolism in experimental cerebral ischemia using topical magnetic resonance: changes i n 31PNMR spectra compared with electroencelphalograms and regional cerebral blood flow. J. Cereb. Blood Flow Metabol. 5, 235-240 (1985). 9. Behar, K. L., Rothman, D. L., Fitzpatrick, S. M.,Hetherington, H. P. and Shulman, R. G. Combined 'H and 31PNMR studies of the rat brain in vivo: effects of altered intracellular pH on metabolism. Ann. NY Acad. Sci. 508, 81-88 (1987). 10. Ross, B. D.,Higgins, R. J., Boggan, J. E., Knittel, 6. and Garwood, M. 31PNMR spectroscopy of the in vivo metabolism of an intracerebral glioma in the rat. Magn. Reson. Med. 6, 403-417 (1988). 11. Younkin, D., Medoff-Cooper, B., Guillet, R., Sinwell, T., Chance, B. and Delivoria-Papadopoulos, M. In vivo 31P nuclear magnetic resonance measurement of chronic changes in cerebral metabolites following neonatal intraventricular hemorrhage. Pediatrics 82, 331-336 (1988). 12. Saunders, J. K., Smith, I. C. P., MacTa Ash, J. C., Rydzy, M., Peeling, J., Sutherland, E., Lesiuk, H. and Sutherland, G. R. Forebrain ischemia studied using magnetic resonance imaging and spectroscopy. NMRBiorned. 2,312-316 (1989). 13. Chang, L.-H., Shirane, R., Weinstein, P. R. and James, T. L. Cerebral metabolite dynamics during temporary complete ischemia in rats monitored by time-shared 'H and 31PNMR spectroscopy. Magn. Reson. Med. 13, 6-13 (1990). 14. Chopp, M., Chen, H., Vande Linde, A. M. Q., Brown, E. and Welch, K. M. A. Time course of postischemic intracellular alkalosis reflects the duration of ischemia. J. Cereb. Blood Flow Metabol. 10, 860-865 (1990). 15. Chopp, M., Vande Linde, A. M. 0.. Chen, H., Knight, R., Helpern, J. A. and Welch, K. M. A. Chronic cerebral intracellular alkalosis following forebrain ischemic insult in rats. Stroke 21,463-466 (1990). 16. Phillis, J. W., Delong, R. E. and Towner J. K. Adenosine deaminase inhibitors enhance cerebral anoxic hyperemia in the rat. J. Cereb. Blood Now Metabol. 5, 295-299 (1985). 17. Phillis, J. W., O'Regan, M. H. and Walter, G. A. Effects of deoxycoformycin on adenosine, inosine, hypoxanthine, xanthine and uric acid release from the hypoxemic rat
cerebral cortex. J. Cereb. Blood Flow Metabol. 8 , 733-741 (1988). 18. Budawari, S. (ed.) The 77th Mercklndex, pp. 1131. Merck & Co. Inc., Rahway (1989). 19. Hagberg, H., Lehmann, A., Sandberg, M., Nystrom, B., Jacobson, I. and Hamberger, A. Ischemia-induced shift of inhibitory and excitatory amino acids from intra- to extracelM a r compartments. J. Cereb. Blood Flow Metabol. 5, 413419 (1985). 20. Phillis, J. W. and O'Regan, M. H. Deoxycoformycin prevents ischemia-induced locomotor hyperactivity in the unanesthetized gerbil. Med. Sci. Res. 16, 897-898 (1988). 21. Gordon, R. E., Hanley, P. E. and Shaw, D. frogr. NMR Spectrosc. 15, 1-47 (1982). 22. Taylor, D. J., Bore, P. J. Styles, P., Gadian, D. G. and Radda, G. K. Bioenergetics of intact human muscles: a 31P NMR study. Mol. Biol. Med. 1, 77-94 (1983). 23. Heitjan, D. F. Generalized Norton-Simon models of tumour growth. Stat. Med. 10, 1075-1088 (1991). 24. English, H. F., Heitjan, D. F., Lancaster, S. and Santen, R. J. Beneficial effects of androgen-primed chemotherapy in the Dunning R3327 G model of prostatic cancer. Cancer Res. 51, 1760-1765 (1991). 25. Durrleman, S. and Simon, R. Flexible regression models with cubic splines. Stat. Med. 8, 551-561 (1989). 26. Chi, E. M. and Reinsel, G. C. Models for longitudinal data with random effects and AR(1 1 errors. J. Am. Stat. Assoc. 84, 452-459 (1989). 27. Diggle, P. J. An approach t o the analysis of repeated measurements. Biometrics 44, 959-971 (1988). 28. Silvey, S . D. Statistical Inference. Chapman & Hall, London (1975). 29. Chance, B., Leigh, J. S., Smith, D. and Nioka, S. 31PNMR as a criterion of critical states of brain oxidative metabolism. 4th Annual Meeting of the Society of Magnetic Resonance in Medicine 1, 250 (1985). 30. Chance, B., Leigh, J. S., Clark, 6.J., Maris, J., Kent, J., Nioka, S. and Smith, D. Control of oxidative metabolism and oxygen delivery in human skeletal muscle: a steady state analysis of the work/energy cost transfer function. froc. Natl Acad. Sci. USA 82, 8384-8388 (1985). 31. Palmer, C. and Smith, M. B. Assessing the risk of kernicterus using nuclear magnetic resonance, in Clinics in ferinatology: Neonatal Jaundice, ed. by M.J. Maisels, Vol. 17, pp. 307-329. W. B. Saunders Co., Philadelphia (1990). 32. Palmer, C., Sallavanti, R. A. and Roberts, R. L. Adenosine deaminase inhibition during cerebral hypoxia-ischemia (HI): effect on brain edema and mortality. fediatr. Res. 27, 347A (1990). 33. Tofts, P. and Wray, S. Changes in brain phosphorus metabolites during the post-natal development of the rat. J. fhysiol. 359, 417-429 (1985). 34. Reynolds, E. 0. R., Wyatt, J. S . , Azzopardi, D., Delpy, D. T., Cady, E. €3.. Cope, M. and Wray, S. New non-invasive methods for assessing brain oxygenation and haemodynamics. Br. Med. Bull. 44, 1052-1075 (1988). 35. Barany, M., Chang, Y. C. and Arus, C. Effect of halothane on the'natural abundance 13CNMR spectra of excised rat brain. Biochemistry 24, 7911-7917 (1985). 36. Barrere, B., Meric, P., Borredon, J., Beloeil, J.-C. and Seylaz, J. Cerebral intracellular pH regulation during hypercapnia in unanesthetized rats: a 31PNMR spectroscopy study. Brain Res. 516, 215-221 (1990). 37. Vannucci, R. C., Vasta, F. and Vannucci, S. J. Cerebral metabolic responses of hyperglycemic immature rats t o hypoxia-ischemia, Pediatr. Res. 21, 524-529 (1987). 38. Nioka. S., Smith, D., Chance, B. and Lockard, S. Metabolic
"P NMR OF HYPOXIC-ISCHEMIC BRAIN INJURY IN R A T S
39.
40.
41.
42.
43.
44.
45.
heterogeneity in brain tissue during incomplete ischemia and reperfusion. NMR Biomed. 3, 239-247 (1990). Levine, S. R., Welch, K. M. A,, Helpern, J. A., Chopp, M., Bruce, R., Selwa, J. and Smith, M. B. Prolonged deterioration of ischemic brain energy metabolism and acidosis associated with hyperglycemia: human cerebral infarction studied by serial 31P NMR spectroscopy. Ann. Neurol. 23, 416-418 (1988). Nakada, T., Houkin, K., Hida, K. and Kwee, I. L. Rebound alkalosis and persistent lactate: multinuclear ('H, 13C, 3'P) NMR spectroscopic studies in rats. Magn. Reson. Med. 18, 9-14 (1991). Lawson, B., Anday, E., Guillet, R., Wagerle, L. C., Chance, B. and Delivoria-Papadopoulos, M. Brain oxidative phosphorylation following alteration in head position in preterm and term neonates. Pediatr. Res. 22, 302-305 (1987). Azzopardi, D., Wyatt, J. S., Cady, E. B., Hamilton, P. A,, Delpy, D. T., Hope, P. L., Stewart, A. L. and Reynolds, E. 0. R. Prognosis of infants with hypoxic-ischaemic brain injury assessed by magnetic resonance spectroscopy. Pediatr. Res. 22, 220 (1987). Hamilton, P. A., Hope, P. L., Cady, E. B., Delpy, D. T., Wyatt, J. S. and Reynolds, E. 0. R. Impaired energy metabolism in brains of newborn infants with increased cerebral echo densities. Lancet i, 1242-1246 (1986). Geiger, J. D., Lewis, J. L., Maclntyre, C. J. and Nagy J. I. Pharmacokinetics of 2'-deoxycoformycin. an inhibitor of adenosine deaminase, in the rat. Neuropharmacology 26, 1383-1387 (1987). Morii, S.,Ngai, A. C., KO, K. R. and Winn, H. R. Role of
46. 47.
48.
49.
50. 51.
52.
I53
adenosine in regulation of cerebral blood flow: effects of theophylline during normoxia and hypoxia. Am. J. fhysiol. 253, H165-H175 (1987). Rothman, S. M. Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death. J. Neurosci. 4, 1884-1 891 (1984). Dolphin, A. C. and Archer, E. R. An adenosine agonist inhibits and a cyclic AMP analogue enhances the release of glutamate but not GABA from slices of rat dentate gyrus. Neurosci. Lett. 43,49-54 (1983). Fastbom, J. and Fredholm, B. B. Inhibition of 13H]-glutamate release from rat hippocampal slices by Lphenylisopropyladenosine. Acta fhysiol. Scand. 125, 121123 (1985). Stromski, M. E., van-Waarde, A., Avison, M. J., Thulin, G., Gaudio, K. M., Kashgarian, M., Shulman, R. G. and Siegel, N. J. Metabolic and functional consequences of inhibiting adenosine deaminase during renal ischemia in rats. J. Clin. lnvest. 82, 1694-1699 (1988). Sweet, D. V. (ed.) Registry of Toxic Effects of Chemical Substances, Vol. 3, p. 2772. US Government Printing Office, Washington, DC (1987). Williams, G. D., Palmer, C., Roberts, R. L. and Smith, M. B. 31PNMR spectroscopy of cerebral hypoxic-ischemic injury: a neonatal rat model to evaluate neuroprotective drugs. 9th Annual Meeting of the Society of Magnetic Resonance in Medicine 2, 1029 (1990). Cady, E. B. and Azzopardi, D. Absolute quantitation of neonatal brain spectra acquired with surface coil localization. NMR Biomed. 2, 305-311 (1989).
31PNMR Spectroscopy of Perinatal Hypoxic-ischemic Brain Damage: a Model to Evaluate Neuroprotective Drugs in Immature
Gerald D. Williams,$ Charles Palmer,$ Rebecca L. Roberts,§ Daniel F. Heitjantt and Michael B. Smith+$$* Departments of $Radiology (Center for NMR Research), PPediatrics (Division of Newborn Medicine), P~Cellularand Molecular Physiology, and Biological Chemistry and ttCenter for Biostatistics and Epidemiology, The Pennsylvania State University College of Medicine, The Milton S. Hershey Medical Center, Hershey, PA 17033, USA
Cerebral energy metabolism can be measured non-invasively in unanesthetized neonatal rats with 3'P NMR spectroscopy. Using this technique, serial changes in high energy phosphates were determined from the right cerebral hemispheres of 7 day postnatal rat pups during a hypoxic-ischemic insult known to produce focal brain injury. During 3 h of hypoxia-ischemia the concentration of ATP dropped to 33 If:8% of prehypoxic (baseline) levels, phosphocreatine (PCr)/Pi decreased from 1.5 & 0.51 to 0.16 f 0.06, while pH decreased nominally by 0.2 units. After 2.5 h of recovery in air, ATP returned to 75 k 10% of baseline levels, PCr/Pi rose to 1.1+- 0.28, and pH returned to its normal value of 7.16f0.06. This model was used to test the efficacy of the adenosine deaminase inhibitor, 2-deoxycoformycin (DCF) as a potential neuroprotective drug. The data for the drug- and saline-treated populations were analyzed by integrating ATP and Pi/PCr levels over specific time intervals, expressing it relative to baseline levels, and modeling it with cubic splines. Pretreatment with 500 pg/kg DCF shows a small, but statistically significant, preservation of both ATP and phosphorylation potential during hypoxia and initial recovery. Brain water content (edema) at 42 h recovery was apparently associated with both mean ATP and mean Pi/PCr in the last 2 h of hypoxia-ischemia. When ATP fell below 70% of baseline, brain edema was evident at 42 h of recovery. This methodology is suitable for extension to human infants.
INTRODUCTION
Perinatal hypoxia-ischemia (asphyxia) is a major cause of mortality and morbidity in the human newborn.' In this study we used a well established animal model of hypoxic-ischemic (HI) brain damage."' Cerebral metabolism was measured serially with 31PNMR spectroscopy during and following the combined HI insult to 7 day postnatal rat pups. The rat pup model consists of right common carotid ligation followed by 3 h of hypoxia in 8% oxygen. It produces perturbations in tissue energy metabolism and brain damage in the posterolateral region of the right cerebral hemisphere. Blood flow to the middle cerebral artery supply area is reduced to 15-35% of control values during hypoxiaischemia and is accompanied by increased glucose utilization.'.". Numerous studies of brain energy metabolism have been performed by in uiuo "'P NMR on anesthetized adult rats.s-" However, no study has yet been published using "P spectroscopy during cerebral hypoxiaischemia in unanesthetized neonatal rats. Extrapolation of results from adult rats to neonates is not valid. as cerebral blood flow and metabolism are
'
Author to whom correspondence should he addressed.
t Preliminary accounts of this work were presented at the 8th Annual Meeting of the Society for Magnetic Resonance in Medicine, Amsterdam. The Netherlands. 1989.' Abbreviations used: HI, hypoxic-ischemic: DCF. 2-deoxycoformycin: PME. phosphomonoester; PCr. phosphocreatine 0952-3480/92/030145-09 $09.SO
0 1992 by John Wiley gL Sons. Ltd.
slower, but pathologic reactions are faster in immature animals.' The rat pup, content in a warm, dark restricted space, is ideal for the study of unanesthetized cerebral metabolism by high resolution NMR. As the brain of the 7 day old rat is still immature, it provides a model for studying the effects of hypoxia-ischemia on the developing nervous system of the newborn, and the opportunity to monitor the potential benefits of neuroprotective drugs. Accordingly, we aimed to determine if inhibition of adenosine deaminase with 2-deoxy~oformycin'~~'~ (DCF; pentostatin, ParkeDavis Pharmaceuticals Inc., Ann Arbor, MI, USA, personal communication) could preserve energy metabolism during cerebral hypoxia-ischemia. Pretreatment with DCF increases adenosine levels as well as cerebral blood flow during hypoxia due to the rise in adenosine concentration." Adenosine has other potentially neuroprotective mechanisms. ". 'I If DCF increases the concentration of adenine nucleotide precursors and prevents catabolism of ATP, elevated levels of high energy phosphates should be detected by 3'P NMR during hypoxia-ischemia and early recovery. Our purpose was to use "P NMR to (i) measure and develop methods to quantitate the changes in high energy phosphates using a neonatal rat model of cerebral hypoxia-ischemia which reliably causes brain injury, (ii) determine if pretreatment with a potential neuroprotective agent, specifically DCF, will alter energy metabolism during the insult or early recovery and (iii) correlate these changes with the development of cerebral edema. Received 4 September 1991 Accepted (revised) 8 October 199I
I46
G . D. WILLIAMS ET AL.
MATERIALS AND METHODS Animal preparation and analysis Unsexed 7 day old Wistar (Charles River) rat pups weighing between 12 g and 18 g were subjected to right carotid artery ligation as described by Palmer et After surgery, the rat pups were returned to their dams for 2.5 h. Immediately before placement in the NMR probe, which was 60 min prior to the onset of hypoxia, pups were injected S.C. with physiological saline or DCF (Parke-Davis Pharmaceuticals Inc.). The DCF concentration was 5(H) or 50 pg/kg (DCF-500 or DCF-50). The choice of saline vs drug administration was randomized to minimize litter bias. Each rat pup had its front legs secured with a thin strip of surgical tape and was wrapped in a 4 x 4" gauze pad to restrict movement. Following NMR spectroscopy, animals were returned to their dams, and hemispheric water content was determined at 42 h of recovery when brain swelling is advanced. Hemispheric water content was computed as a percentage according to the formula: [(total wt -dry wt)/total wt] x 100. Parallel experiments were performed for determination of the effect of DCF on mortality. The pups were placed in 500 mL airtight jars which were partially submerged in a water bath at 36.5 "C. Air temperature in the center of the jar was 32-33°C. The pups were exposed to an 8% oxygen mixture for 3 h. Then the jars were opened to room air and the survivors returned to their dams.
Probe design and animal placement All studies were performed on a Bruker AM400 widebore spectrometer at 162.0MHz for 31P using a specially constructed NMR probe. The probe served as both an NMR RF receiver and a gas-tight chamber for control of oxygen and temperature. It was modified from a standard Bruker wide-bore probe body such that the top portion consisted of a gas-tight plexiglass environmental chamber. The 70 mm diameter cylindrical chamber contained a removable panel for easy access and positioning of the animal with respect to the 7 x 10 mm elliptical single-turn surface coil. The R F circuit was completed by adjustable tuning and matching capacitors (Jackson Brothers Ltd., London, U K ) located in close proximity to the coil at the top of the chamber. The chamber temperature was regulated by heated water from a temperature bath (Haake, Berlin, Germany), carried through a continuous polyethylene tubing which was wound and fixed into a coil at the back side of the animal chamber. The temperature adjacent to this heating pad was monitored with a standard Bruker thermocouple arrangement to be 36 f 0.5 "C during the entire experiment. A lower ambient temperature was recorded in close proximity to the skin, and was also constant during the experiment. Oral temperature was determined to be 36 "C in selected pups. In addition to an entrance at the bottom of the probe for heated water, an entrance for tubing to carry
a gas mixture was provided. The top open end of the chamber was sealed with plastic wrap, allowing maintenance of the calibrated gas mixture. A thin soft bandage (SpencoTM, Spenco Medical Corp., Waco, TX, USA) secured the rat's head to the surface coil with minimal discomfort. The NMR experiment was performed without anesthesia, with the rat in a vertical orientation inside the magnet. Careful positioning of the region of injury in the right hemisphere with respect to the coil assured minimal contamination from the left (contralateral) hemisphere (Fig. 1). Additional gauze material and a foam rubber casing were added prior to sealing the animal access door in order to further restrict movement and maintain body temperature. A standard Bruker aluminum alloy probe cover was used to provide structural integrity and RF shielding while the probe was in the magnet.
N M R experiment Following tuning to the "P frequency, the probe was placed in the magnet for optimization of field homogeneity using the 'H signal from the water protons of the same region of the hemisphere as used in the phosphorus spectroscopy. 'H shimming was accomplished with a double tuned circuit located external to the magnet. ?'P spectra were acquired with a single-pulse sequence using a pulse width of 11 ps, a repetition time of 2 s, acquisition time of 0.133s and a sweep width of 7.7 kHz. A typical resulting "P spectrum is shown in Fig. 1. Twenty-two spectra were acquired in 7.5 or 15 min blocks beginning with a 23min prehypoxic period during which three spectra were accumulated to obtain baseline levels. Eight per cent O2was supplied for the 3 h hypoxic period followed by a 2.5 h recovery in medical air. Gas delivery was maintained at constant levels at all times with a flow meter (Matheson Gas Products, East Rutherford, NJ, USA) located in series with the air tanks and the NMR probe.
Data analysis
A Lorentzian line broadening of 40 Hz and a 500 Hz profile correction function" to remove the broad lipid/ bone component were applied prior to Fourier transformation to 4 K data points. Spectra were analyzed by peak areas relative to an appropriate baseline. Slight overlap of P, with phosphomonoester (PME) was corrected by using the Bruker spectral simulation program 'Linesim'. Independently determined saturation factors to correct for incomplete relaxation were applied to the NMR peak areas. The pH was determined from the frequency of P,, calculated from the chemical shift difference between PI and phosphocreatine (PCr). The difference was compared with a standard 37°C pH calibration curve which uses a PK ,= ~ 6.75, and acid and base endpoints of d,,= 3.27 and db= 5.69.'' The time-series data were analyzed by two approaches. In the first method, the average ATP, P,/PCr and pH were computed within animals in both the last 2 h of hypoxia-ischemia and the initial recovery
I47
"P NMR OF HYPOXIC-ISCHEMIC BRAIN INJURY IN RATS
ATP
PME
Figure 1. Illustration of the head of a 7 day old rat pup. The broken line shows the region of the right hemisphere of the brain, while the solid line within that region shows the position of the elliptical surface coil, where injury is maximal. An 8 min 31Pspectrum (2 s repetition time) from this region is shown. Line broadening of 20 Hz was applied and the broad component was not filtered.
period. Intra-animal means were computed by the formula:
c h
Pub=
a
h
Pj Atj/N
Atj U
where P represents the parameter of interest (ATP, Pi/PCr or pH), j is the spectrum number, At, is the accumulation time for spectrum j , a and b are the lower and upper limits of the time range of interest and N is the baseline value of the parameter obtained during prehypoxia. Thus each animal yielded an intra-animal mean, normalized to baseline levels, and group averages of these mean values were compared by a pooledvariance two-sample t-test. This method is valid since data from different pups are statistically independent, but it leads to conservative p values because it uses the data inefficiently. In a second approach, cubic spline models were fit to the data using the method described by Heitjad3 and applied in English et aLZ4Briefly, the underlying mean curve is approximated by a cubic spline with knots at the times of initiation and terminaton of hypoxia." This error model accommodates both random animal effects and time-series a u t o ~ o r r e l a t i o n .The ~ ~ . model ~~ was fit by maximum likelihood, and parameter estimates were used to compute mean relative ATP, log(Pi/PCr) and pH in given time ranges. Estimated means in different treatment groups were compared by Wald tests.*' This method is both valid and efficient because it models the intra-animal error structures, thus making better use of the available data; it also avoids multiple comparisons problems associated with multiple t-testing.
RESULTS
Figure 2 shows the typical time course of metabolites during HI and recovery for a saline-treated rat pup. Eight out of the usual 22 spectra are represented. In this example, PCr/P, decreased from 1.7 to 0.25 during 3 h of hypoxia and ATP decreased to Since DCF did not reduce edema, it follows that the drug is unlikely to be effectual at reducing other measures of neuropathological brain damage. Mortality during hypoxia increased markedly with DCF dosage. With DCF-500 we were near the onset of toxicity (TDL,,= 500 pg/kg) reported in adult mouse'" but well below the half lethal dosage (LD,,, = 128 mg/kg; Parke-Davis Pharmaceuticals Inc., personal communication). The significant number of deaths in this study should be attributable to the added stress on the neonate from hypoxia-ischemia. The high mortality with DCF-500 must be remembered when evaluating the preservation of adenine nucleotides in those rats that survived. Although adenosine levels should increase significantly after DCF administration,lh ATP and phosphorylation potential levels were preserved in only selected animals. In summary, we have demonstrated that it7 uiuo serial assessment of brain energy metabolism can be obtained with this model using "'P NMR. A simple method of quantitation using mean values of ATP and P,/PCr, taken from hypoxia and initial recovery, has shown that there is a small preservation of high energy metabolites with DCF at 500pg/kg in animals that survive the 3 h o f hypoxia-ischemia. Analysis of the time-series data by cubic spline models" gave even smaller p-values. There was no significant reduction in subsequent cerebral edema with DCF-500. However, a threshold was found for predicting subsequent edema using either ATP or P,/PCr levels, integrated during the last 2 h of hypoxia. DCF was shown using both statistical methods of analysis to have n o effect on metabolite levels at a lower dosage of 50 pg/kg. We are currently in the process of evaluating other neuroprotective agents which both preserve metabolite levels and reduce brain damage.5' The methods ot analysis used in this study should be suitable for extension to human infants, and Cady and Azzopardi have demonstrated absolute quantitation of brain spectra from
152
G. D. WILLIAMS E T A L .
normal and birth asphyxiated infants5' In future experiments we would like to administer drug therapy following a period of H I insult in order to closely model the brain damage and treatment of the asphyxiated newborn human.
Acknowledgements 2-DCF was a gift from the Parke-Davis Pharmaceutical Research Division, a Warner-Lambert Co. We thank Scott Babe for assistance with the NMR analysis and Professor Robert C. Vannucci for helpful discussions
REFERENCES 1. Smith, M. B., Palmer, C. and Williams, G. D. 31P NMR spectroscopy of unanesthetized hypoxic-ischemic neonatal rats: a model t o evaluate neuroprotective drugs. 8th Annual Meeting of the Society of Magnetic Resonance in Medicine 1, 466 (1989). 2. Vannucci, R. C. Acute perinatal brain injury: hypoxiaischemia, in Management of Labor, ed. by W. R. Cohen, D. B. Archer and E. A. Friedman, pp. 138-244. Aspen Publishers, Inc., Rockville (1989). 3. Rice 111, J. E., Vannucci, R. C. and Brierley, J. B. The influence of immaturity on hypoxic-ischemic brain damage in the rat. Ann. Neurol. 9, 131-141 (1981). 4. Palmer, C., Brucklacher, R. M., Christensen, M. A. and Vannucci, R. C. Carbohydrate and energy metabolism during the evolution of hypoxic-ischemic brain damage in the immature rat. J. Cereb. Blood Flow Metabol. 10, 227235 (1990). 5. Palmer, C., Vannucci, R. C. and Towfighi, J. Reduction of perinatal hypoxic-ischemic brain damage with allopurinol. fediatr. Res. 27, 332-336 (1990). 6. Vannucci, R. C., Lyons, D. T. and Vasta, F. Regional cerebral blood flow during hypoxia-ischemia in immature rats. Stroke 19, 245-250 (1988). 7. Vannucci, R. C., Christensen, M. A. and Stein, D. T. Regional cerebral glucose utilization in the immature rat: effect of hypoxia-ischemia. fediatr. Res. 26, 208-214 (1989). 8. Horikawa, Y., Naruse, S., Hirakawa, K., Tanaka, C., Nishikawa, H. and Watari, H. !n vivo studies of energy metabolism in experimental cerebral ischemia using topical magnetic resonance: changes i n 31PNMR spectra compared with electroencelphalograms and regional cerebral blood flow. J. Cereb. Blood Flow Metabol. 5, 235-240 (1985). 9. Behar, K. L., Rothman, D. L., Fitzpatrick, S. M.,Hetherington, H. P. and Shulman, R. G. Combined 'H and 31PNMR studies of the rat brain in vivo: effects of altered intracellular pH on metabolism. Ann. NY Acad. Sci. 508, 81-88 (1987). 10. Ross, B. D.,Higgins, R. J., Boggan, J. E., Knittel, 6. and Garwood, M. 31PNMR spectroscopy of the in vivo metabolism of an intracerebral glioma in the rat. Magn. Reson. Med. 6, 403-417 (1988). 11. Younkin, D., Medoff-Cooper, B., Guillet, R., Sinwell, T., Chance, B. and Delivoria-Papadopoulos, M. In vivo 31P nuclear magnetic resonance measurement of chronic changes in cerebral metabolites following neonatal intraventricular hemorrhage. Pediatrics 82, 331-336 (1988). 12. Saunders, J. K., Smith, I. C. P., MacTa Ash, J. C., Rydzy, M., Peeling, J., Sutherland, E., Lesiuk, H. and Sutherland, G. R. Forebrain ischemia studied using magnetic resonance imaging and spectroscopy. NMRBiorned. 2,312-316 (1989). 13. Chang, L.-H., Shirane, R., Weinstein, P. R. and James, T. L. Cerebral metabolite dynamics during temporary complete ischemia in rats monitored by time-shared 'H and 31PNMR spectroscopy. Magn. Reson. Med. 13, 6-13 (1990). 14. Chopp, M., Chen, H., Vande Linde, A. M. Q., Brown, E. and Welch, K. M. A. Time course of postischemic intracellular alkalosis reflects the duration of ischemia. J. Cereb. Blood Flow Metabol. 10, 860-865 (1990). 15. Chopp, M., Vande Linde, A. M. 0.. Chen, H., Knight, R., Helpern, J. A. and Welch, K. M. A. Chronic cerebral intracellular alkalosis following forebrain ischemic insult in rats. Stroke 21,463-466 (1990). 16. Phillis, J. W., Delong, R. E. and Towner J. K. Adenosine deaminase inhibitors enhance cerebral anoxic hyperemia in the rat. J. Cereb. Blood Now Metabol. 5, 295-299 (1985). 17. Phillis, J. W., O'Regan, M. H. and Walter, G. A. Effects of deoxycoformycin on adenosine, inosine, hypoxanthine, xanthine and uric acid release from the hypoxemic rat
cerebral cortex. J. Cereb. Blood Flow Metabol. 8 , 733-741 (1988). 18. Budawari, S. (ed.) The 77th Mercklndex, pp. 1131. Merck & Co. Inc., Rahway (1989). 19. Hagberg, H., Lehmann, A., Sandberg, M., Nystrom, B., Jacobson, I. and Hamberger, A. Ischemia-induced shift of inhibitory and excitatory amino acids from intra- to extracelM a r compartments. J. Cereb. Blood Flow Metabol. 5, 413419 (1985). 20. Phillis, J. W. and O'Regan, M. H. Deoxycoformycin prevents ischemia-induced locomotor hyperactivity in the unanesthetized gerbil. Med. Sci. Res. 16, 897-898 (1988). 21. Gordon, R. E., Hanley, P. E. and Shaw, D. frogr. NMR Spectrosc. 15, 1-47 (1982). 22. Taylor, D. J., Bore, P. J. Styles, P., Gadian, D. G. and Radda, G. K. Bioenergetics of intact human muscles: a 31P NMR study. Mol. Biol. Med. 1, 77-94 (1983). 23. Heitjan, D. F. Generalized Norton-Simon models of tumour growth. Stat. Med. 10, 1075-1088 (1991). 24. English, H. F., Heitjan, D. F., Lancaster, S. and Santen, R. J. Beneficial effects of androgen-primed chemotherapy in the Dunning R3327 G model of prostatic cancer. Cancer Res. 51, 1760-1765 (1991). 25. Durrleman, S. and Simon, R. Flexible regression models with cubic splines. Stat. Med. 8, 551-561 (1989). 26. Chi, E. M. and Reinsel, G. C. Models for longitudinal data with random effects and AR(1 1 errors. J. Am. Stat. Assoc. 84, 452-459 (1989). 27. Diggle, P. J. An approach t o the analysis of repeated measurements. Biometrics 44, 959-971 (1988). 28. Silvey, S . D. Statistical Inference. Chapman & Hall, London (1975). 29. Chance, B., Leigh, J. S., Smith, D. and Nioka, S. 31PNMR as a criterion of critical states of brain oxidative metabolism. 4th Annual Meeting of the Society of Magnetic Resonance in Medicine 1, 250 (1985). 30. Chance, B., Leigh, J. S., Clark, 6.J., Maris, J., Kent, J., Nioka, S. and Smith, D. Control of oxidative metabolism and oxygen delivery in human skeletal muscle: a steady state analysis of the work/energy cost transfer function. froc. Natl Acad. Sci. USA 82, 8384-8388 (1985). 31. Palmer, C. and Smith, M. B. Assessing the risk of kernicterus using nuclear magnetic resonance, in Clinics in ferinatology: Neonatal Jaundice, ed. by M.J. Maisels, Vol. 17, pp. 307-329. W. B. Saunders Co., Philadelphia (1990). 32. Palmer, C., Sallavanti, R. A. and Roberts, R. L. Adenosine deaminase inhibition during cerebral hypoxia-ischemia (HI): effect on brain edema and mortality. fediatr. Res. 27, 347A (1990). 33. Tofts, P. and Wray, S. Changes in brain phosphorus metabolites during the post-natal development of the rat. J. fhysiol. 359, 417-429 (1985). 34. Reynolds, E. 0. R., Wyatt, J. S . , Azzopardi, D., Delpy, D. T., Cady, E. €3.. Cope, M. and Wray, S. New non-invasive methods for assessing brain oxygenation and haemodynamics. Br. Med. Bull. 44, 1052-1075 (1988). 35. Barany, M., Chang, Y. C. and Arus, C. Effect of halothane on the'natural abundance 13CNMR spectra of excised rat brain. Biochemistry 24, 7911-7917 (1985). 36. Barrere, B., Meric, P., Borredon, J., Beloeil, J.-C. and Seylaz, J. Cerebral intracellular pH regulation during hypercapnia in unanesthetized rats: a 31PNMR spectroscopy study. Brain Res. 516, 215-221 (1990). 37. Vannucci, R. C., Vasta, F. and Vannucci, S. J. Cerebral metabolic responses of hyperglycemic immature rats t o hypoxia-ischemia, Pediatr. Res. 21, 524-529 (1987). 38. Nioka. S., Smith, D., Chance, B. and Lockard, S. Metabolic
"P NMR OF HYPOXIC-ISCHEMIC BRAIN INJURY IN R A T S
39.
40.
41.
42.
43.
44.
45.
heterogeneity in brain tissue during incomplete ischemia and reperfusion. NMR Biomed. 3, 239-247 (1990). Levine, S. R., Welch, K. M. A,, Helpern, J. A., Chopp, M., Bruce, R., Selwa, J. and Smith, M. B. Prolonged deterioration of ischemic brain energy metabolism and acidosis associated with hyperglycemia: human cerebral infarction studied by serial 31P NMR spectroscopy. Ann. Neurol. 23, 416-418 (1988). Nakada, T., Houkin, K., Hida, K. and Kwee, I. L. Rebound alkalosis and persistent lactate: multinuclear ('H, 13C, 3'P) NMR spectroscopic studies in rats. Magn. Reson. Med. 18, 9-14 (1991). Lawson, B., Anday, E., Guillet, R., Wagerle, L. C., Chance, B. and Delivoria-Papadopoulos, M. Brain oxidative phosphorylation following alteration in head position in preterm and term neonates. Pediatr. Res. 22, 302-305 (1987). Azzopardi, D., Wyatt, J. S., Cady, E. B., Hamilton, P. A,, Delpy, D. T., Hope, P. L., Stewart, A. L. and Reynolds, E. 0. R. Prognosis of infants with hypoxic-ischaemic brain injury assessed by magnetic resonance spectroscopy. Pediatr. Res. 22, 220 (1987). Hamilton, P. A., Hope, P. L., Cady, E. B., Delpy, D. T., Wyatt, J. S. and Reynolds, E. 0. R. Impaired energy metabolism in brains of newborn infants with increased cerebral echo densities. Lancet i, 1242-1246 (1986). Geiger, J. D., Lewis, J. L., Maclntyre, C. J. and Nagy J. I. Pharmacokinetics of 2'-deoxycoformycin. an inhibitor of adenosine deaminase, in the rat. Neuropharmacology 26, 1383-1387 (1987). Morii, S.,Ngai, A. C., KO, K. R. and Winn, H. R. Role of
46. 47.
48.
49.
50. 51.
52.
I53
adenosine in regulation of cerebral blood flow: effects of theophylline during normoxia and hypoxia. Am. J. fhysiol. 253, H165-H175 (1987). Rothman, S. M. Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death. J. Neurosci. 4, 1884-1 891 (1984). Dolphin, A. C. and Archer, E. R. An adenosine agonist inhibits and a cyclic AMP analogue enhances the release of glutamate but not GABA from slices of rat dentate gyrus. Neurosci. Lett. 43,49-54 (1983). Fastbom, J. and Fredholm, B. B. Inhibition of 13H]-glutamate release from rat hippocampal slices by Lphenylisopropyladenosine. Acta fhysiol. Scand. 125, 121123 (1985). Stromski, M. E., van-Waarde, A., Avison, M. J., Thulin, G., Gaudio, K. M., Kashgarian, M., Shulman, R. G. and Siegel, N. J. Metabolic and functional consequences of inhibiting adenosine deaminase during renal ischemia in rats. J. Clin. lnvest. 82, 1694-1699 (1988). Sweet, D. V. (ed.) Registry of Toxic Effects of Chemical Substances, Vol. 3, p. 2772. US Government Printing Office, Washington, DC (1987). Williams, G. D., Palmer, C., Roberts, R. L. and Smith, M. B. 31PNMR spectroscopy of cerebral hypoxic-ischemic injury: a neonatal rat model to evaluate neuroprotective drugs. 9th Annual Meeting of the Society of Magnetic Resonance in Medicine 2, 1029 (1990). Cady, E. B. and Azzopardi, D. Absolute quantitation of neonatal brain spectra acquired with surface coil localization. NMR Biomed. 2, 305-311 (1989).
