0306-4522/91 s3.00 + 0.00 PergamonPress plc

Nertroscience Vol. 42, No. I, pp. 171-182, 1991

Printed in Great Britain

0 1991IBRO

THE EFFECTS OF CAFFEINE ON ISCHEMIC NEURONAL INJURY AS DETERMINED BY MAGNETIC RESONANCE IMAGING AND HISTOPATHOLOGY G.

R. SurnaRLAND,*t J. PEELING,$§ H. J. LESIUK,? R. M. BROWNSTONE,j’ J. IL SAUNDERS~~and J. D. GEIGER*T/

M. RYDZY,II

Departments of *Pharmacology and Therapeutics, TSurgery (Neurosurgery), and SRadiology, University of Manitoba Faculty of Medicine, 770 Bannatyne Avenue, Winnipeg, Manitoba, Canada R3E OW3 $Department of Chemistry, University of Winnipeg, Manitoba, Canada I/Division of Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada effects of caffeine on ischemic neuronal injury were determined in rats subjected to forebrain &hernia induced by bilateral carotid occlusion and controlled h~tension (50mmHg for 10 min). High resolution (100 pm) multi-slice, multi-echo magnetic resonance images were obtained daily for three consecutive days post-operatively in sham-operated rats and in rats that received either saline vehicle (controls), a single i.v. injection of iOmg/kg caffeine 30min prior to an ischemic insult (acute caffeine group), or up to 9Omg/kg per day of caffeine for three consecutive weeks prior to an ischemic insult (chronic caffeine group). Rats in the control group exhibited enhanced magnetic resonance image intensity in the striatum 24 h after ischemia which increased in the striatum and also appeared in the hippocampus after 48 h, and which began to resolve in both regions by 72 h post&hernia. Histopathological analysis of each rat following the final magnetic resonance examination showed that ischemic neuronal injury was strictly confined to the brain regions showing magnetic resonance image changes. Acute caffeine rats showed accelerated changes in the magnetic resonance images, with increased hip~mpal intensity appearing at 24 h post-ischemia. Although there was magnetic resonance evidence of accelerated injury, quantitative analysis of the histopatbological data at 72 h showed no significant difference in the extent of neuronal injury in any brain region between control-ischemic and acute caBeine rats. Nine out of 11 rats in the chronic caffeine group showed no magnetic resonance image changes over the three study days. Chronic caffeine rats had significantly less neuronal damage in ah vulnerable brain regions than either of the other groups of ischemic rats. The accelerated ischemic injury in rats treated with an acute dose of caffeine may occur secondary to, an~gonism of adenosine receptors, whereas protection from ischemic injury following chronic admimstration of caffeine may be mediated by up-regulation of adenosine receptors. Alx&act-The

Caffeine, ingested at the rate of approximately l~,~ tons/year, is the most widely consumed psychotropic drug in the world. In North America, for

example, the average daily intake per person is 200 mg; about 10% of the population consumes more than 500 mg/day. The pharmacological effects of caffeine and other commonly used methylxanthines such as theophylline are consistent with, although not exclusively restricted to, their ability to block adenosine receptors?8*4’ This capacity of methylxanthines to interact with a specific receptor system is relevant to the study of ischemic neuronal injury and stroke, as an increasingly large body of evidence suggests that adenosine functions as a neuroprotective substance in the CNS.‘O Experimental evidence that supports the notion that adenosine may be involved in protecting against ischemic maturation, a process in which neuronal injury in discrete areas of the CNS progressively

‘jiTo whom correspondence should be addressed. [‘H]CHA, [3~yclohexyladenosine; MR, magnetic resonance; NMDA, N-methyl-D-aspartate.

Abbr~iff~j~~:

worsens as the re-perfusion period following a transient ischemic insult is lengthen~,z3,~,~ includes the following. Adenosine is ubiquitously present and accumulates in ischemic tissues.‘* Adenosine receptors are concentrated in those brain regions selectively vulnerable to ischemic injury.z8~3’ These receptors are significantly and rapidly decreased in number in gerbils and rats subjected to brief periods of cerebral ischemia.2’*2s~M-3’ Post-ischemic neuronal hyperexcitability has been implicated as a cause of ischemic neuronal injury,4 and adenosine is known to decrease neuronal excitability.” Tt also inhibits the release of excitatory amino acid neurotransmitte& which have neurotoxic properties at high concentrations. Furthermore, adenosine receptor agonists protect against ischemic hippocampal cell loss in the rat” and gerbil,& while theophylline, possibly through competitive antagonism of adenosine receptors, significantly exacerbates ischemic cell damage in the gerbiL3’ Since adenosine has mainly depressive actions on the nervous system, the rapid down-regulation of adenosine receptors following ischemia242s could explain the known delayed post-ischemic increase 171

172

in the spontaneous rate of action potentials arising from hippocampal CA1 pyramidal neurons.” The resulting increased metabolic demand may in part contribute to ultimate irreversible neuronal injury.23,40 It is difficult to study the neurological effects of adenosine compounds directly as they have limited permeability through the blood-brain barrier. Therefore, the link between adenosine and ischemic neuronal damage in the CNS may best be investigated using competitive antagonists of adenosine receptors such as caffeine” which are known to have free access into the brain. The hypothesis that acute admi~stra~on of caffeine prior to ischemia may lead to potentiation of ischemic neuronal injury while chronic ingestion of caffeine may protect against ischemic neuronal injury as long-term use results in up-regulation of adenosine receptors are examined here.5~7.‘5J7~34*4**49 In addition magnetic resonance (MR) imaging is introduced’ as a method of non-invasively evaluating the regional temporal development of ischemic brain injury.

Ten control rats received an equivalent volume of normal saline. Eleven rats (chronic caffeine group) received caffeine by gavage three times daily on a schedule of 20 mg/kg per dose for the first week and 30 mg/kg per dose for the second and third weeks. Caffeine was withdrawn 24 h prior to ischemia. Control rats (n = 9) for the chronic caffeine group received saline vehicle by gavage, with equivalent volumes over a three-week period. Magnetic resonance microimaging

The remaining 43 rats were examined using MR imaging at 24, 48, and 72 h post-operatively (sham-operation or induced ischemia). During the time the animals were in the spectrometer, they were immobilized with isoflurane (1% in air) administered via a nose cone tlxed to the animal cradle. Each animal was maint~n~ no~o~e~c by means of a circulating water bath incorporated into the cradle. All MR examinations were performed on a Bruker Biospec 4.7/30 spectrometer with modified gradient coils and a two-coil imaging probe as described earlier.M Two interleaved sets of images (l-mm-thick slices spaced 4mm apart) were acquired using a multi-slice (3), multi-echo (4) sequence (echo time = 34 ms, repetition time = 1.2 s, six acoumulations), with the resulting six coronal slices spanning the forebrain. Each image was acquired using 256 x 256 data points with a field of view of 2.5 x 2.5 cm, giving a pixel resolution of about 100 pm.

EXPERIMENTAL PROCEDURES

induction of reversible forebrain iscchemia

Fifty-five male Sprague-Dawley rats (National Research Council, Canada) weighing 250-300g were used. Eleven control rats received sham operations while the remaining 44 rats were subjected to transient forebrain ischemia. As previously described,‘L42~43 each rat was fasted for 1 h, pre-treated with 0.5 mg/kg atropine, and then anesthetized with 30 mg/kg sodium pentobarbital. The rats were mechanically ventilated and maintained at 375°C on a heated water blanket placed under both the torso and head. A temperature electrode inserted into the temporalis muscle monitored head temperature prior to, during, and following the ischemic insult. Catheters were inserted into the tail artery and jugular vein of each rat for blood pressure monitoring and infusion of normal saline or caffeine. Both carotid arteries were exposed through a neck incision. After 2Omin of stabili~tion, forebrain ischemia was induced through bilateral carotid occlusion coincident with a reduction in systemic blood pressure to a mean of 50 mmHg through aspiration of heparinized blood. After IOmin, blood flow through the carotid arteries was restored and the aspirated blood was re-infused. Blood gas analyses and hematocrit determinations were obtained both prior to and following the ischemic insult. Ventilatory support was continued until the animal was breathing well and moving its extremities.

Following the tinal MR imaging study, each rat was nerfusion-tixed with I 1 of 10% buffered fo~~dehvde (nH 5.25). After fixation, the brain was placed in the s&me fixative for two weeks prior to sectioning. The brains were cut coronally into 1.5-mm slices, dehydrated in graded concentrations of ethanol, and embedded in paraffin. Serial sections (8 pm thick) were cut and stained with hematoxylin and eosin. All sections were examined to determine the qualitative and topographic extent of brain damage. For quantification of ischemic neuronal injury, standardized sections of the cerebral cortex, hippocampus, and caudate were used. Pre-determined regions of the hippocampal and frontal sections were photographed and ischemic neuronal injury was quantitatively determined by direct visual counting of all neurons.” The frequency of ischemic neurons was calculated by dividing the number of acidophilic and/or pyknotic neurons by the total number of neurons. Striatal damage was graded as follows < 10% necrotic neurons = 1, IQ-50% necrotic neurons = 2, 5*100% necrotic neurons = 3. The examiner was unaware of the identity of the sections examined. Statistics

Comparisons between groups were made using analysis of variance followed by Duncan’s intergroup comparison test.

Adenosine receplors

[3H]Cyclohexyladenosine ([‘H]CHA)-binding to hippocampal adenosine receptors was determined in 12 rats (six sham-operated, six ischemic) five days post-surgery using a radioligand binding assay containing 0.4-34 nM [3H]CHA (Moravek B&hem. Inc., Brea, CA), between 50 and 100 gg of membrane protein, and in a separate set of duplicate tubes 10 PM R-~6-phenylisopropyl~enosine (Boehringer Mannheim) for determination of non-specific displaceable binding.16 Incubations were for 120 min at 25°C. Membrane protein was determined using bovine serum albumin as a standard. Drug administration

Thirty minutes prior to the induction of ischemia, eight rats received lOmg/kg caffeine (acute caffeine group) dissolved in 0.5 ml normal saline through the jugular cannula.

RESULTS

Mean blood pressures were increased following the ischemic insult from 135 f 3 to 152 t 2 mmHg (P < 0.01). No intergroup differences were observed. The method of provoking the ischemic insult did not produce a significant systemic acidosis. Two minutes following re-perfusion, blood gas analyses yielded values for pH of 7.36 _+0.07, PaCOr of 35.2 + 4.3 mmHg, HCO,- of 21.1 + 2.6 mmHg, and a base excess of -3.1 + 3.2 meq/l compared with control values for pH of 7.41 f 0.03, PaCO, of 35.6 f 4.8 mmHg, HCO,- of 22.2 f 3.2 mmHg, and a base excess of - I .6 + 2.6 meq/l. By maintaining the rat

173

Caffeine and ischemic neuronal injury

Fig. 1. MR image (100 pm resolution) of a coronal section through a rat hippocampus and thalamus. Intmcranial structures am clearly seen, including the ~~rnp~, thalamus, middle cerebral artery branches (arrow), and possibly thalamic perforating vessels (arrowheads).

on a water blanket, minimal (< 0.25”C) fluctuation was observed during carotid occlusion and controlled hypotension. All experimental groups, both saline and caffeine treated, showed the same minor fluctuations in head temperature with no intergroup differences. This model consistently produces rapid energy failure, shown by a decrease in the intensity of peaks due to high energy phosphate compounds and a rise in inorganic phosphate peak in the in uivo 3iP MR spectrum.39 A typical coronal MR image (1OO~m resolution) through the hippocampus and thalamus of a sham-operated rat is shown in Fig. 1. The intracranial structures that are clearly identifiable include the hippocampus, thalamus, neocortex and the flow voids produced by the intracranial carotids, middle cerebral arteries and possibly small thaiamo perforating vessels. All five non-ischemic, sham-operated rats showed the same anatomical features, and the images showed no apparent changes over time. Non-caffeine treated rats given an ischemic insult showed increased MR image intensity, particularly in later echoes, in the region of the striatum 24 h post-ischemia (Table 1 and Fig. 2). The en-

in head temperature

hanced intensity became more pronounced and was accompanied by similar changes in the hippocampus at 48 h post-ischemia. By 72 h, although still present, the changes in the images were beginning to resolve. Eight of the 19 rats showed increased image intensity within the neocortex and five of these eight rats demonstrated unilateral changes only. Photomicrographs obtained from the striatum, dorsal lateral ~udoputamen, and the CA1 sector of the hippocampus showing normal cellular morphology are presented in Fig. 3. Similar photomicrographs showing ischemic neuronal injury in the Table 1. Post-ischemic magnetic resonance image changesserially observed rats that received normal saline (n = 19) Time 24h 48h 72h

Striatum 13/19 16119 16/19

Brain region Hip~pus 5119 16119 16/19

Neocortex 7/19 S/l9 8119*

*Five rats had his~~tholo~c evidence of neocortical injury that did not correlate with the MR image. In all other regions the histopathologic changes paralleled MR image changes.

Fig. 2. from a treated striatal

Third echo images obtained through the striatum (upper figure) and hipplw rat before ischemia (A) and serially at 24 (B), 48 (C). and 72 h (D) post- scl rat. At 24 h. the damage (enhanced intensity) was mainly confined to the lII damage was intensified, and hippocampal injury was now apparent. At 72 h J; in both regions, but was partially resolved when compared to imagt c z 174

I I ; il

I I I :

115

Caffeine and ischemic neuronal injury Table 2. Post-ischemic magnetic resonance image changesacute and chronic caffeine treated rats*

Time

Acute caffeine (n = 8) Striatum Hippo~mpus

Chronic caffeine (n = 11) Striatum Hipp~mpus

2111 O/l1 718 6/S z/11 z/11 718 8/S 2/11 2111 718 S/8 *Histo~tholo~c evidence of isehemic neuronal injury correlated precisely with the MR image changes.

24h 48 h 12h

striatum and the CA1 sector of the hippocampus at 72 h post-ischemia are presented in Fig. 4. Ischemic neurons in the neocortex were essentially restricted to layers III-VI. Ischemic neurons were characterized by retraction of the cell body, eosinophilia of the cytoplasm, disappearance of Nissl bodies, and pyknosis and hyperchromasia of the nucleus. In the hippocampus and striatum, MR image changes correlated precisely with cytological evidence of irreversible neuronal injury. In three rats MR neocortical changes did not correlate with histopathology. This probably arises from the use of a surface coil as the receiver, resulting in an artifactual enhancement of image intensity close to the coil, so that MR intensity changes in the neocortex were less well visualized. In the acute caffeine group, the increased image intensity evident in both the striatum and hippocampus at 24 h post-ischemia (Table 2) in six of the eight rats is indicative of accelerated ischemicinduced tissue injury. The histopathology obtained from two rats at 24 h post-&hernia confirmed that extensive neuronal injury had already’ occurred by this time in both the striatum and hippocampus. By 48 h post&hernia, the increased MR image intensity remained evident iir the hippocampus and striatum of all rats in the acute caffeine group. As was the case for the control animals, the enhanced intensity was beginning to resolve by 72 h post-ischemia. In three of the eight rats in this group, neocortical injury was more extensive and appeared earlier compared to the control animals. Histopathology obtained at 72 h post-&hernia showed ischemic neuronal injury

confined to selectively vulnerable brain regions (striaturn, hippocampus, and deep neocortical layers), correlating extremely well with the MR imaging results. Only two of the 11 rats in the chronic caffeine group showed changes in the MR images obtained post-ischemia (Table 2). In both of these animals, the changes were confined to the striatum at 24 h, involved both the striatum and ~p~ampus at 48 h, and were decreased by 72 h, as was seen in the control group. The images of the other rats in this group were indistinguishable from those of the sham-operated (non-ischemic) group of animals. In agreement with the MR imaging data, histopathology of the brains of nine of the 11 animals in the chronic caffeine group showed no evidence of ischemic neuronal injury, while the distribution of ischemic neurons in the other two rats corresponded to the regions showing increased intensity in the MR images. The ~st~bution of ischemic changes was similar in all groups, varying only in severity. Neocortical changes were primarily restricted to cortical layers III-VI, and were most severe in the watershed zone between the territory of supply of the anterior and middle cerebral arteries. Injury was asymmetrical in those animals that also had as~rnet~~l MR imaging changes. Within the striatum, neuronal injury involved the small- and medium-sized neurons located in the dorsal lateral portion of the caudoputamen while the large neurons were spared. In the ~pp~ampus, neuronal damage was greatest in the CA1 sector followed by the CA3 and CA4 sectors. In all regions, the glia and vasculature were unremarkable. The total number of neurons in the regions that were quantitated was not different between groups. Neuronal injury was signifi~ntly decreased in the hippocampus, striatum, and neocortex (Table 3) of the chronic caffeine group compared to acute caffeine or control groups. The apparent maximal number of adenosine receptors (B_) as labelled by [3H]CHA in hippocampus was signif&intly reduced (P < 0.05) by 34% from 806 + 59 in control to 536 f 59 in

Table 3. Ischemic neuronal injury in rats subjected to forebrain &hernia Treatment groups Acute Region Subiculum CAl/CAZ CA3 CA4 Frontal cortex Striatum

Control (n = 10) 0.49+0.11 0.83 & 0.06 0.37 It 0.15 0.36ztO.15 0.44 rt 0.12 1.9 &OS

Chronic

Caffeine

(n = 8) 0.75kO.14 0.85 + 0.08 0.48 k 0.21 0.47f0.22 0.36 f 0.10 1.8 f 0.5

Control (n = 8) 0.42&0.06 0.80 f 0.04 0.26 + 0.20 0.31 +0.08 0.38 IO.10 1.7kO.3

Caffeine (n = 11) 0.21 JrO.08 0.37 rt 0.1 I* 0.02 It O.Ol* 0.03 fO.Ol* 0.02 + O.OI* 0.09 * 0.09*

Data are mean + S.E.M. Ischemic neuronal injury = Ischemic neurons/ total neurons for subiculum, CAi/CAZ, CA3 and CA4 grade of injury < 10% = 1, l@-50% = 2, 50-100% = 3 for striatum. *P < 0.01 different from acute caffeine or control groups. Analysis of variance followed by Duncan’s test for multiple comparisons.

176

G.R.SUTHERLAND

et al.

Caffeine and ischemic neuronal injury

177

178

G. R. SUTHERLAND et al.

Caffeine and ischemic neuronal injury

179

G. R. SIJ~~EXLAND et al.

180

&hernia rats; the & values were unchanged. A separate group of rats chronically fed caffeine for three weeks according to the same schedule used in the present study exhibited an approximate 20% increase in rH]CHA binding to cerebral cortical membranes (unpublished observations). DISCUSSION

The ischemic model used in this study results in a predictable injury confined principally to selectively vulnerable neuronal populations. These include the small- and medium-sized neurons of the dorsal-lateral caudoputamen, hipp~ampal pyramidal neurons and neurons localized to layers III-VI of the neocortex.23+33*40,42 Both the physiological and histopatholo~~l manifestations of this injury appear as the duration following the ischemic insult increases.23*33,40pM The striatum manifests neuronal injury as early as 12 h post-ischemia, while hippocampal injury is usually delayed until 48 h post&hernia. 35,40 Both the regionally selective vulnerability and the variable ischemic maturation have been observed in this study using MR imaging where serial observations were made non-invasively in the same experimental animal. Both the ease with which temporal sequences of effects can be observed and the reduction of biological variability intrinsic to intergroup comparisons combine to make MR imaging a powerful approach to the study of stroke. The enhanced intensity observed post-ischemia in the MR images, particularly in the later echo-images, likely reflects increased free water from cytotoxi~/vaso$enic edema within selectively vulnerable brain regions.= The excitatory amino acid glutamate and the inhibitory neuromodulator adenosine have been implicated both separately and in an interrelated fashion in the development of ischemic neuronal injury. Excitatory amino acid nemotransmitters and adenosine” accumulate in isehemic tissues, and their receptors are heterogeneously distributed within the CNS, being heavily concentrated in those brain regions selectively vulnerable to ischemic neuronal injury. 24-26Following ischemia, adenosine A, sites are selectively and rapidly lost, showing a 20% decrease within 2 h following an ischemic insult.25 Administration of adenosine agonist compounds into the brain protects against ischemic neuronal and adenosine receptor injury.1.‘246 Down-regulation of ~-methyl-~-as~~ate (NMDA) receptors, however, is delayed, corresponding temporally to the onset of neuronal degeneration.47 The interrelated actions of these two classes of compounds include findings that adenosine decreases the retease of excitatory amino acids6 and excitatory amino acids provoke the release of adenosine from neural tissues.ig Furthermore, the adenosine-induced post-synaptic hyperpolarizatior? may lead to a voltage block of NMDA receptors.49 Finally, the rapid down-regulation of adenosine

receptors by 2 h post-ischemia and the delayed loss of NMDA receptors in response to an ischemic insult could account for the physiological observations of regional post-ischemic hyperexcitability,~ which is first observed 2 h post-is~hemia. If adenosine functions alone or in concert with other neuroactive substances as an endogenous neuroprotectant, then caffeine through its ability to accumulate in brain tissue and competitively block adenosine receptors may be expected to lead to a worsening of ischemic neuronal damage. Indeed, theophylline at doses of 30 mg/kg37 and 32 mg/kgg causes increased damage to hip~c~pal neurons in the CA1 area of Mongolian gerbils, and in the present study acute caffeine (10 mg/kg) accelerated damage in the hippocampus and worsened damage in the neocortex. By blocking the adenosine receptor the ~st-i~hemic hy~rexcited state might be accentuated and the increased metabolic demand might hinder the ability of the injured neurons to recover from the metabolic perturbation induced by ischemia, thus contributing to irreversible injury. If these data can be extrapolated to humans, people prone to strokes might be advised to refrain from acute ingestion of large amounts of methylxanthine-containing food and drink.” Most people, however, chroni~lly ingest large amounts of caffeine and consequently it is relevant to the possible involvement of adenosine in stroke to consider the effects of chronic caffeine on adenosine receptor systems and ischemic neuronal damage. Chronic caffeine exposure leads to increases in the number of adenosine receptors in ratsS~7,‘5,17,W,48,4g and mice.3t1,27,MAdenosine receptor numbers are similarly increased in rats2g,X,45and mice50 chronically administered theophylline. Thus, it is now well established that at least the numbers of recognition sites for adenosine are up-regulated following chronic trea~ents with methylxanthines. These increases are sustained for up to 30 days following methylxantbine withdrawa13*“*” Importantly, functional consequences of increased receptor numbers have been reported. Chronic caffeine ingestion by rats leads to increased sensitivity of the high affinity agonist receptor state to guanine nucleotides and enhanced responses of adenylate cyclase activity to the inhibitory effects of Rphenylisopropyladenosine, an A, receptor agonistfindings consistent with an increased coupling of the receptor to the Gi protein-adenylate cyclase comp1ex.i’~” Up-regulated adenosine receptors are probably responsible for the tolerance to the stimulatory action of caffeine on locomotor activity and other behavioral measurements,‘3~14~20the reduced sensitivity of rats to various chemical convulsants,38~4’ and firing rates of m~n~phalic reticular neurons5 The neuroprotection observed in the present study can similarly be explained by an up-regulation of adenosine receptors. Similar results have recently been observed in a study conducted on Mongolian gerbils.3s*36

Caffeine and ischernic neuronal injury

181

Although adenosine receptors are most likely involved in the acute and chronic effects of caffeine, the possibility must be entertained that there are additional parti~pating factors. For example, acute caffeine may increase intracellular calcium there-

maturation, while chronic caffeine protects against ischemic neuronal injury.

by accentuating neuronal damage while chronic caffeine may deplete intracellular calcium stores and lead to protection from ischemic neuronal damage. Regardless of the underlying mechanism(s), acute caffeine accelerates the process of ischemic

ante and Drs David Wilkins, Keith Butler and Jim MacTavish for their expertise in procuring the MR images. This work was supported by grants from the Canadian Heart Foundation (GS, JP) the Upjohn Company of Canada (GS, JP) and the Medical Research Council of Canada (JG) of which JG is a Scholar.

~~~~~~~~~~~~~~~_~he authors thank pat Mijjes, Maureen Donnelly and Emi Okamoto for technical assist-

REFERENCES 1. Arvin B., Neville L. F. and Roberts P. J. (1988) Adenosine analogues protect against kainate-induced toxicity in rat striatum and hippocampus. Sot. Neurosci. Abstr. 14, 746. 2. Eenveniste H., Drejer J., Schousboe A. and Diemer D. H. (1984) Elevation of the extracellular concentrations of glutamate and aspartate in rat hip~mpus during transient cerebral ischemia monitored by intracerebral microdiaiysis. J. Neurochem. 43, 1369-1374. 3. Boulenger J. P. and Marangos P. J. (1989) Caffeine withdrawal affects central adenosine receptors but not benzodiazepine receptors. J. Neural Transm. 78, 9-15. 4. Boulenger J. P., Pate1 J., Post R. M., Parma A. M. and Marangos P. J. (1983) Chronic caffeine consumption increases the number of brain adenosine receptors. Lrp Sci. 32, 1135-1142. 5. Chou D. T., Khan S., Forde J. and Hirsh K. R. (1985) Caffeine tolerance: behavioral, electrophysiological and neur~hemical evidence. Life Sci. 36, 2347-2358. 6. Corradetti R., Loconte G., Moroni F., Passani M. B. and Pepeu G. (1984) Adenosine decreases aspartate and glutamate release from rat hippocampal slices. Eur. J. pharmac. 104; 19-26.. I. Daval J. L.. Deckert J.. Weiss S. R. B.. Post R. M. and Marannos P. J. (1989) Uureaulation of adenosine A, recenters and forskolin binding sites following chronic treatment with cal%ine or carbarnazepiie: a quantitative autoradiographic study. Epilepsia 30, 2633. 8. Deckert J. and Jorgenson M. B. (1988) Evidence for pre- and post-synaptic localization of adenosine A, receptors in the CA1 region of rat hippocampus: a quantitative autoradiographic study. Brain Res. 446, 161-164. 9. DeLeo J., Schubert P. and Kreutzberg G. W. (1988) ~o~ntofylline (HWA 285) protects hipp~~p~ neurons of Mongolian gerbils against ischemic damage in the presence of an adenosine antagonist. Neurosci. Len. 84, 307-3 11.

10. Dragunow M. and Faull R. L. M. (1988) Neuroprotective effects of adenosine. Tren& pharmac. Sci. 9, 193-194. 11. Dunwiddie T. V. (1985) The physiological role of adenosine in the central nervous system. Int. Rev. Neurobiol. 27, 63-139.

12. Evans M. C., Swan J. H. and Meldrum B. S. (1987) An adenosine analogue, 2-chloroadenosine, protects against long term development of ischemic cell loss in the rat hippocampus. Neurosci. Lett. 113, 287-292. 13. File S. E., Baldwin H. A., Johnston A. L. and Wilks L. J. (1988) Behavioral effects of acute and chronic administmtion of caffeine in the rat. pharmuc. Riochem. Behau. 30, 809-815. 14. Finn I. B. and Holtxman S. G. (1988) Tolerance and cross-tolerance to theophylline-induced stimulation of locomotor activity in rats. Life Sci. 42, 2475-2482. 15. Fredhohn B. B. (1982) Adenosine actions and adenosine receptors after one week treatment with caffeine. Acfa physiol. stand.

115, 283-286.

16. Geiger J. D. (1986) Localization of [3Hlcyclohexyladenosine and pH]nitrobenzylthioinosine binding sites in rat striatum and superior colliculus. Brain Res. 363,404408. 17. Green R. M. and Stiles G. L. (1986) Chronic caffeine ingestion sensitizes the Al adenosine receptor-adenylate cyclase system in rat cerebral cortex. J. ciin. Invest. 7?, 222-227. 18. Hagberg H., Andersson P., Lacarewicz J., Jacobson I., Butcher S. and Sandberg M. (1987) Extracellular adenosine, inosine, hypoxanthine, and xanthine in relation to tissue nucleotides and purines in rat striatum during transient &hernia. J. Neurochem. 49, 227-231. 19. Hoehn K. and White J. D. (1989) Evoked release of endogenous adenosine from rat cortical slices by K+ and glutamate. Brain Res. 478, 49-51. 20. Holtzman S. G. and Finn I. B. (1988) Tolerance to behavioral effects of caffeine in rats. pharmac. ffiochem. Behau. 29, 411418.

21. Jarvis M. F., Murphy D. E., Williams M., Gerhardt S. C. and Boast C. A. (1987) A novel N-methyl-D-aspartate antagonist, CGS 19755, prevents ischemia-induced reductions in adenosine A-l receptors in gerbil brain: a quantitative autoradiographic study. Sot. Neurosci. Abstr. 13, 497. 22. Kato H., Kogure K., Ohtomo H., Izumiyama M., Tobita M., Matsui S., Yamamoto E., Kohno H. J., Ikebe Y. and Watanabe T. (1986) Characterization of experimental ischemic brain e&ma utilizing proton nuclear magnetic resonance imaging. J. cerebr. Hood Flow Metab. 6, 212-221. 23. Kirino T. (1982) Delayed neuronal death in the gerbil hip~p~ following &hernia. Brain Res. 239, 57-69. 24. Lee K. S., Schubert P., Reddington M. and Kreutzberg G. W. (1983) Adenosine receptor density and the depression of evoked neuronal activity in the rat hippocampus in vitro. Neurosci. Letr. 37, 81-85. 25. Lee K. S., Tetxlaff W. and Kreutzberg G. W. (1986) Rapid down-regulation of hippocampal adenosine receptors following brief anoxia. Brain Res. 380, 155-158. 26. Lupica C. R., Berman R., Williams M. and Jarvis M. F. (1990) Autoradiographic analysis of brain adenosine Al and A2 receptors following chronic theophylline exposure to rats. In Purines in Cellular Signalling: Targets for New Drugs (ads Jacobson K., Daly J. and Man~niello V.), pp. 393-394, Springer. New York.

182

G.R.

SUTHERLAND

et al.

27, Marangos P. J., Boulenger J. P. and Pate1 J. (1984) Effects of chronic caffeine on brain adenosine receptors: regional and ontogenetic studies: Life Sci. 34, 899-907. 28. Mathew R. J. and Wilson W. H. (19851 Caffeine consumotion. withdrawal and cerebral blood flow. Headache 25. , 1 I

305-309. 29. Murray T. F. (1982) Up-regulation of rat cortical adenosine receptors following chronic administration of theophylline. Eur. J. Pharmac. 82, 113-114. 30. Onodera H., Sato G. and Kogure K. (1987) Quantitative autoradiographic analysis of muscarinic cholinergic and adenosine Al binding sites after transient forebrain &hernia in the gerbil. Brain Res. 415, 309-322. 31. Onodera H. and Kogure K. (1985) Autoradiographic visualization of adenosine Al receptors in the gerbil hippocampus: changes in the receptor density after transient ischemia. Bruin Res. 505, 406-408. 32. Peeling J., Wong D. and Sutherland G. R. (1989) Nuclear magnetic resonance study of regional metabolism after forebrain ischemia in rats. Stroke 20, 633640. 33. Pulsinelli W. A., Brierley J. B. and Plum F. (1982) Temporal profile of neuronal damage in a model of transient forebrain ischemia. Ann. Neural. 11, 491-498. 34. Ramkumar V., Bumgarner J. R., Jacobson K. A. and Stiles G. L. (1988) Multiple components of the Al adenosine receptor-adenylate cyclase system are regulated in rat cerebral cortex by chronic caffeine ingestion. J. c&n. Invest. 82, 242-247.

35. Rudolphi K. A., Keil M., Fastbom J., Grome J. J. and Fredholm B. B. (1990) Contrary effects of acute versus chronic caffeine administration on ischemically induced hippocampal damage in the gerbil-involvement of adenosine receptors? In Purines in Cellular SignaIling: Targets for New Drugs (eds Jacobson K., Daly J. and Mananiello V.), pp. 411412. Springer, New York. 36. Rudolphi K. A., Keil M., Fastbow J. add Freholm B, 8. (1989) Ischemic damage in gerbil hippocampus is reduced following upregulation of adenosine (Al) receptors by caffeine treatment. Neuroci. Lett. 103, 275-280. 37. Rudolphi K. A., Keil M. and Hinze H.-J. (1987) Effect of theophylline on ischemically induced hippocampal damage in Mongolian gerbils: a behavioral and histopathological study. J. cerebr. Blood Flow Metab. 7, 74-81. 38. Sanders R. C. and Murray T. F. (1988) Chronic theophylline exposure increases agonist and antagonist binding to AI adenosine receptors in rat brain. Nearopharmaco~og~i 27, 757-760. 39. Saunders J. K., Smith I. C. P., MacTavish J. G., Rydzy M., Peeling J., Sutherland E., Lesiuk H. and Sutherland G. R. (1989) Forebrain ischemia studied using magnetic resonance imaging and spectroscopy. NMR Biomed. 2, 312-316. 40. Smith M.-L., Auer R. N. and Siesjo B. K. (1984) The density and distribution of ischemic brain injury in the rat following 2-10 min of forebrain ischemia. Rcta neuropath., Berlin 64, 319-332. 41. Snyder S. H., Katims J. J., Annau Z., Braun R. F. and Daly J. W. (1981) Adenosine receptors and behavioral action of methylxanthines. Proc. natn. Acad. Sci. U.S.A. 78, 3260-3264. 42. Sutherland G. R., Lesiuk H., Bose R. and Sima A. A. F. (1988) Effect of mannitol, nimodipine, and indomethacin singly or in combination on cerebral ischemia in rats. Stroke 19, 571-578. 43. Sutherland G. R., Ong B. Y., Louw D. and Sima A. A. F. (1989) Effect of lidocaine on forebrain ischemia in rats. Stroke 20, 119-122. 44. Suzuki R., Yamaguchi T., Inaba Y; and Wagner H. G. (1983) The effects of 5 min ischemia in Mongolian gerbils. II. Changes of spontaneous neuronal activity in cerebral cortex and CA1 sector of hippocampus. Acta neuropath., Berlin 60, 217. upregulation of A,-adenosine receptors associated 45. Szot P., Sanders R. C. and Murray T. F. (1987) Theophylline-induct with reduced sensitivity to convulsants. ~~rophurmacofogy 26, 1173-l 180. 46. Von Lubitz D. K. E. J., Dambrosia J. M. and Redmond D. J. (1989) Protective effect of cyclohexyladenosine in treatment of cerebral ischemia in gerbils. Neuroscience 30, 451462. 41. Westerbera E.. Monaehan D. T.. Kalimo H.. Cotman C. W. and Wieloch T. W. (1989) Dvnamic changes of excitatory amino a&? receptors-in the rat hippocamp& following transient cerebral ischemia. j. ieuro& 9, %8-805. . 48. Wu P. H. and Coffin V. L. (1984) Up-regulation of brain 13H]~a~parn binding sites in chronic ~ffeine-treated rats. Brain Res. 294, 186-189.

49. Wu P. H. and Phillis J. W. (1986) Up-regulation of brain [‘Hldiazepam binding sites in chronic caffeine-treated rats. Gen. Pharmac. 17, 501-503. 50. Zielke C. L. and Zielke H. R. (1987) Chronic exposure to subcutaneously implanted methylxanthines. Differential elevation of A,-adenosine receptors in mouse cerebellar and cerebra1 cortical membranes. B&hem. Pharmac. 36, 2.533-2538. (Accepted 5 September 1990)