Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The tight coupling between neuronal activity, rCBF and rCMR has been reported under physiological conditions. It is assumed that both glucose and oxygen demand increase the rCBF of the activated brain region. This provides the basis for activation studies using PET and functional magnetic resonance imaging, in which neuronal activation can be monitored as an increased rCBF response (for review, see Roland, 1993). On the other hand, the rCBF response to stimulation is very rapid (Lindauer et al., 1993), and the rCMR of oxygen response is much less than the rCBF response during physiological stimulation (Fox and Raichle, 1986; Fox et al., 1988b). These results suggest the involvement of neuronal mechanisms for the coupling of neuronal activity and rCBF during stimulation. Several animal studies demonstrated that the cortical neuronal activity was influenced by cholinergic projections from the basal forebrain (Kiyosawa et al., 1989; Kurosawa et al., 1989; Ouchi et al., 1996; Sato and Sato, 1992). Also, cholinergic muscarinic agonists were reported to increase rCBF in the cerebral cortex or pial vessel dilation with no increase in rCMR (Scremin et al., 1982). Recently, it was reported that the systemic administration of scopolamine, a muscarinic cholinergic receptor antagonist, abolished the rCBF response to somatosensory stimulation (Ogawa et al., 1994; Tsukada et al., 1997). The abolishment of the rCBF response by scopolamine was recovered by administration of physostigmine, a cholinesterase inhibitor (Tsukada et al., 1997), whereas the rCMR response was not affected by scopolamine (Ogawa et al., 1994) or physostigmine (Tsukada et al., 1997) administration. It was reported that this reactive rCBF increase was regulated by the intrinsic cholinergic neurons, which might have an important role in mediating the neuronal activity level in the blood vessels and dilating vessels locally to adjust glucose and oxygen supply (Fukuyama et al., 1996). Taken together, these results strongly suggested cholinergic regulation of rCBF responses to physiological stimulation.

In experimental animals, alterations of cholinergic systems by pharmacological blockade or lesions of the cholinergic projection neurons are accompanied by cognitive deficits. The muscarinic antagonist scopolamine is the most frequently used drug in studies of cognitive dysfunction. Administration of scopolamine produces transient cognitive impairment in various learning paradigms in both animals and humans, and several cognitive enhancers have been reported to prevent the scopolamine-induced disruption of memory function (Honer et al., 1987; Sitaram et al., 1978; Summers et al., 1986). Many compounds have been developed as centrally acting acetylcholine enhancers to improve cognitive deficits in subpopulations of patients with AD. In particular, cholinesterase inhibitors such as physostigmine and tacrine (tetrahydroaminoacridine) were reported to improve learning and memory (Davis and Mohs, 1982; Summers et al., 1986). Long-term administration of tacrine induced neurochemical changes such as increased nicotinic receptor binding and increased glucose metabolism, as measured by PET, in the temporal and frontal cortices of patients with AD, in whom neurochemical changes were paralleled by improvements in neuropsychological tests (Nordberg et al., 1992). Furthermore, long-term treatment with tacrine might not only improve symptoms but also delay disease progression (Nordberg et al., 1992). In young and aged rats, metabolic studies showed that physostigmine and tacrine increased rCMR in several regions of the brain, which topographically overlapped with the distribution of M₂ muscarinic receptors and cholinesterase activity (Bassant et al., 1993, 1995). The mechanisms of the effects of cognitive enhancers are still unclear. However, they might show their therapeutic effects, at least in part, by affecting cholinergic transmission (for review, see Freeman and Dawson, 1991).

In the present study, the influences of scopolamine and cholinesterase inhibitors on the rCBF response to somatosensory stimulation were evaluated in the somatosensory cortex of unanesthetized monkeys using [¹⁵O]H₂O and a high-resolution animal PET scanner. The drugs tested here were physostigmine, E2020 (Iimura et al., 1989; Sugimoto et al., 1992) and tacrine, all of which have been reported to improve learning and memory in experimental animals and also to show cholinesterase inhibitory activity, facilitating their actions on the cholinergic system.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and drugs. Four young-adult male rhesus monkeys (Macaca mulatta) weighing 5 to 8 kg were used for the PET measurements. Monkeys were maintained and handled in accordance with the recommendations of the United States National Institutes of Health and the guidelines of the Central Research Laboratory, Hamamatsu Photonics. They were trained to sit in a chair twice each week for >3 months. At least 1 month before the PET study, an acrylic plate, with which the monkey was fixed to a monkey chair, was attached to the head of each monkey, as described previously (Onoe et al., 1994).

Physiological monitoring. During PET scanning, heart rate and body temperature were continuously monitored using a life monitoring system (Nihon Kohden, Tokyo, Japan). The results shown were obtained 30 min after each drug treatment. Because of the technical limitations with the use of unanesthetized monkeys, PaCO₂, PaO₂, pH and blood pressure of arterial blood were measured under pentobarbital anesthesia. Arterial blood samples were obtained 30 min after each drug administration, via a cannula in a femoral artery, and monitored with a STAT PROFILE blood gas analyzer (NOVA Biomedical, Waltham, MA).

PET experiment. Data were collected by using a high-resolution PET scanner (SHR-2400; Hamamatsu Photonics, Hamamatsu, Japan), with transaxial resolution of 3.0-mm full width at half-maximum in the center of the scan field and a center-to-center distance of 6.5 mm (Watanabe et al., 1992). Animals were lightly sedated with i.m. injections of ketamine (5 mg/kg) and were seated in the monkey chair. During the sedation period, the head of the monkey was fixed to the chair, stereotaxically aligned parallel to the orbitomeatal line. Four hours after ketamine injection, at which time the monkey showed complete recovery from the effects of ketamine, the study was started under dim light.

Data analysis. Data analysis of PET experiments was performed in accordance with the method reported previously (Tsukada et al., 1997). PET images were reconstructed using data integrated throughout the scans for 1 min, determined from all slices from reconstructed images in each scan, and these values were used to normalize all data from the experiments of that day (Fox and Mintun, 1989; Fox et al., 1988a; Takechi et al., 1994). The images were then smoothed with a median filter with four pixels next to each pixel. Stimulation and control studies performed within 30 min were paired. Three sets of stimulation minus resting subtracted images were averaged and smoothed with a core r = 2 median filter. The averaged pixel values and the S.D. were calculated from all pixel values in each slice of subtracted images. The pixels exceeding a statistical criterion of P < .05 (two-tailed), corresponding to a Z score of 1.96 or greater, were superimposed on magnetic resonance images of the same subject, which were obtained with a Toshiba MRT-50A/II (0.5-T) scanner. The stereotaxic coordinates for PET and magnetic resonance imaging were adjusted based on the orbitomeatal line, with a specially designed head-holder (Takechi et al., 1994). Because three subtracted image pairs were obtained under each condition, pixel-by-pixel statistical analysis was not performed (Friston et al., 1990; Worsley et al., 1992). To statistically compare the data between conditions, the ROI in the contralateral somatosensory cortex corresponding to a Z score of 1.96 or greater was obtained from the subtracted image obtained after saline administration. The actual sizes of the ROI ranged from 70 to 80 mm². ROI were placed symmetrically and bilaterally in the contralateral and ipsilateral somatosensory cortical regions, and average counts associated with [¹⁵O]H₂O were determined in bilateral regions. The same ROI were used for different conditions in each monkey. The ratios (contralateral/ipsilateral) of radioactivities in the regions were analyzed statistically (Ogawa et al., 1994; Tsukada et al., 1997). Comparison between treatments was carried out using the unpaired two-tailed t test, and a probability level of <5% (P < .05) was considered to be statistically significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Administration of scopolamine (50 µg/kg) alone or of scopolamine plus physostigmine (10 µg/kg), E2020 (100 µg/kg) or tacrine (1000 µg/kg) did not produce any significant changes in physiological indices, as shown in table 1. The heart rate and body temperature were monitored 30 min after each drug injection, during PET scanning. In the present study, PaCO₂, PaO₂, pH and blood pressure of arterial blood were not monitored during PET scanning, because of the technical limitations of using conscious monkeys. When monitored in anesthetized animals with the same protocol as used in PET studies, PaCO₂, PaO₂, pH and blood pressure of arterial blood were not altered significantly under each condition (table 1).

After saline injection, vibrotactile stimulation significantly increased the rCBF response, represented as the ratio of radioactivities of contralateral and ipsilateral cortices, to 134% of the resting condition (figs. 1- 4). Systemic administration of scopolamine at a dose of 50 µg/kg induced complete abolishment of the rCBF response in the somatosensory cortex (figs. 1, 2, 3, 4). The rCBF response abolished by scopolamine (50 µg/kg) was recovered by administration of physostigmine, in a dose-dependent manner (123 and 146% of each resting condition at 1 and 10 µg/kg, respectively) (figs. 1 and 2). These findings were partially consistent with our previous results (Tsukada et al., 1997). Administration of E2020 at doses of 10 and 100 µg/kg induced the dose-dependent recovery of the rCBF response abolished by pretreatment of scopolamine (118 and 132% of each resting condition, respectively) (figs. 1 and 3). Much higher doses of tacrine were required (120 and 124% of each resting condition at 100 and 1000 µg/kg, respectively) for recovery of the abolished rCBF response, compared with physostigmine and E2020 (figs. 1 and 4). Administration of physostigmine (10 µg/kg), E2020 (100 µg/kg) or tacrine (1000 µg/kg) alone after saline treatment did not significantly affect the rCBF response induced by vibrotactile stimulation (figs. 1, 2, 3, 4). The ratios (contralateral/ipsilateral) of rCBF were approximately 1.0 under all conditions without vibrotactile stimulation (figs. 2, 3, 4).

View larger version (K): [in this window] [in a new window]   Fig. 1.   Typical images showing the rCBF response affected by scopolamine and cholinesterase inhibitors. Saline, scopolamine (50 µg/kg), physostigmine (1 or 10 µg/kg), E2020 (10 or 100 µg/kg) or tacrine (100 or 1000 µg/kg) was administered i.v. 30 min before PET scanning. Three stimulation-resting pairs of scans were averaged for each condition, and the averaged PET images were superimposed on magnetic resonance images of the same animals. The range of Z scores for PET data is color-coded.

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Effects of scopolamine and physostigmine on rCBF responses to vibrotactile stimulation in monkey brain. Saline, scopolamine (50 µg/kg) or physostigmine (1 or 10 µg/kg) was administered i.v. 30 min before PET scanning. When the stimulation was given, the ROI in the contralateral somatosensory cortex corresponding to a Z score of 1.96 or greater was obtained from subtracted images after saline administration. Without (resting) and with vibration, ROI were placed symmetrically and bilaterally in the contralateral and ipsilateral somatosensory cortical regions, and average counts associated with [15O]H2O were determined in bilateral regions. The same ROI were examined under different conditions in each monkey. The ratios (contralateral/ipsilateral) of rCBF in the regions were analyzed statistically. Values are means ± S.D. for four monkeys. *P < .05 vs. respective resting condition.

View larger version (45K): [in this window] [in a new window]   Fig. 3.   Effects of scopolamine and E2020 on rCBF responses to vibrotactile stimulation in monkey brain. Saline, scopolamine (50 µg/kg) or E2020 (10 or 100 µg/kg) was administered i.v. 30 min before PET scanning. The ratios (contralateral/ipsilateral) of rCBF for each condition were analyzed statistically as described for figure 2. Values are means ± S.D. for four monkeys. *P < .05 vs. respective resting condition.

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Effects of scopolamine and tacrine on rCBF responses to vibrotactile stimulation in monkey brain. Saline, scopolamine (50 µg/kg) or tacrine (100 or 1000 µg/kg) was administered i.v. 30 min before PET scanning. The ratios (contralateral/ipsilateral) of rCBF for each condition were analyzed statistically as described for figure 2. Values are means ± S.D. for four monkeys. *P < .05 vs. respective resting condition.

The recovery effect of physostigmine (10 µg/kg) on the rCBF response abolished by scopolamine showed a peak with the largest magnitude (140% of resting condition) among three inhibitors at 1 hr, and the magnitude returned almost to the basal level (contralateral/ipsilateral = 1) from 2 hr after administration and thereafter (fig. 5). Tacrine (1000 µg/kg) also showed a peak (123% of resting condition) at 1 hr, followed by a gradual reduction, and the recovery effect was not observed 4 hr after the administration (fig. 5). The recovery effect of E2020 (100 µg/kg) was comparable (136% of resting condition) to that of physostigmine at 1 hr, and this level was maintained up to 4 hr after administration (fig. 5). The abolished rCBF response induced by scopolamine (50 µg/kg) was observed during PET measurement up to 4 hr after systemic administration (fig. 5).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Duration of the effects of cholinesterase inhibitors on the rCBF response to the vibrotactile stimulation abolished by scopolamine in monkey brain. Scopolamine (50 µg/kg), physostigmine (10 µg/kg), E2020 (100 µg/kg) or tacrine (1000 µg/kg) was administered i.v., and PET measurements were performed 1, 2 and 4 hr after administration. Data at time 0 were obtained just before administration of cholinesterase inhibitors. The ratios (contralateral/ipsilateral) of rCBF for each condition were analyzed statistically as described for figure 2. Values are means ± S.D. of four monkeys. , scopolamine alone (50 µg/kg); , scopolamine plus physostigmine (10 µg/kg); , scopolamine plus E2020 (100 µg/kg); , scopolamine plus tacrine (1000 µg/kg). *P < .05 vs. time 0 value for each drug treatment; #P < .05 vs. scopolamine-alone condition at the respective time.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The rCBF response to vibrotactile stimulation that was abolished by systemic administration of scopolamine was recovered by administration of cholinesterase inhibitors (physostigmine, E2020 and tacrine) in the somatosensory cortex of unanesthetized monkeys. It is of interest that, when evaluated within 1 hr after administration of cholinesterase inhibitors, the order of dose ranges needed for the recovery of functional rCBF responses correlated well with that of cholinesterase inhibitory activity measured in vitro; that is, E2020 was 15 times more potent than tacrine and 10 times less potent than physostigmine when tested in vitro (Yamanishi et al., 1990). Physostigmine, a classical cholinesterase inhibitor with the highest inhibitory activity among compounds examined, has been shown to improve memory performance in patients with AD (Davis and Mohs, 1982). However, peripheral side effects of physostigmine have limited its use for treatment of AD. Tacrine was reported to exhibit hepatotoxicity. These side effects are partially due to the low selectivity of the effects of these drugs for AChE and BuChE. The brain contains exclusively AChE, whereas peripheral tissues contain BuChE rather than AChE. Physostigmine and tacrine inhibit AChE in the brain as well as BuChE in peripheral tissues, whereas E2020 inhibits AChE in the brain (Sugimoto et al., 1992; Yamanishi et al., 1990). Other reasons why the use of physostigmine for patients with AD is limited are its short duration of action and low bioavailability. The recovery effect of physostigmine on the rCBF response abolished by scopolamine reached a peak 1 hr after administration, and the effect disappeared after 2 hr. Tacrine also showed a short-term action, with no effect observed 4 hr after injection. In contrast, the recovery effect of E2020 lasted during PET measurements up to 4 hr after administration. These results seem to reflect the different durations of cholinesterase inhibition by physostigmine and E2020 measured in vitro; that is, E2020 (10 mg/kg) significantly inhibited cholinesterase in rat brain by 80% of control up to 8 hr after administration, whereas the inhibitory effect of physostigmine (10 mg/kg) on cholinesterase diminished within 2 hr (Sugimoto et al., 1992). The differences in durations of cholinesterase inhibition among these three compounds might, at least in part, be attributable to their differences in stability in vivo.

In addition to cholinesterase inhibition, other effects of these drugs in the central nervous system could not be excluded. Tacrine has been reported to modulate the release of acetylcholine (Nilsson et al., 1987), to block monoamine release and uptake (Drukarch et al., 1988) and to inhibit monoamine oxidase (Adem et al., 1989). Also, it has been demonstrated that tacrine directly blocks ion channels (Rogawski, 1987; Shaw et al., 1985). E2020 binds to  receptors (Eisai company report), and  receptor-binding agonists have been reported to increase acetylcholine release in the frontal cortex of rats (Matsuno et al., 1993). Some of these effects might be involved in the activation of neuron activity. However, the higher activity of cholinesterase inhibition provided a stronger recovery effect on the scopolamine-abolished rCBF response to stimulation, suggesting that the modulatory effect might be attributable, at least in part, to activation of the cholinergic neuronal system induced by cholinesterase inhibition.

Our previous reports demonstrated that the rCBF response abolished by scopolamine was recovered by administration of physostigmine (Tsukada et al., 1997), whereas the rCMR response, as measured with [¹⁸F]-2-fluoro-2-deoxy-D-glucose, was not affected by administration of scopolamine (Ogawa et al., 1994) or physostigmine (Tsukada et al., 1997). These previous results indicated that the recovery effects observed with the other two cholinesterase inhibitors, E2020 and tacrine, were also due to normalization of the coupling mechanism between neuronal activity and the rCBF response by enhancement of acetylcholine transmission and not to activation of the somatosensory neuronal response itself.

It has been reported that the rCBF response to stimulation might be influenced by changes in global cerebral blood flow induced by changes in arterial PaCO₂ (Shimosegawa et al., 1995; Wilder, 1953). Human studies indicated that scopolamine administration resulted in decreased cortical blood flow (Honer et al., 1987) and glucose metabolism (Blin et al., 1995). In the present study, we did not monitor PaCO₂ during PET scanning under unanesthetized conditions. It was impossible for us to obtain arterial blood samples during PET scans, because of the difficulty of placing and maintaining the cannula in femoral arteries of monkeys under unanesthetized conditions. When monitored under anesthesia with the same protocol, however, PaCO₂ levels were not significantly affected by scopolamine, physostigmine, E2020 or tacrine. The physiological indices monitored during PET scanning suggested that scopolamine and the cholinesterase inhibitors did not alter the heart rate or body temperature of monkeys. Our previous study indicated that PaCO₂ was not significantly affected by scopolamine or physostigmine, and the present results of the rCBF response obtained with scopolamine and physostigmine were also consistent with our previous observations (Tsukada et al., 1997). It was also reported that scopolamine butylbromide (100 µg/kg), which does not penetrate the blood-brain barrier, did not abolish the rCBF response to stimulation (Tsukada et al., 1997). Taken together, the observed changes, induced by scopolamine and cholinesterase inhibitors, in the rCBF response to vibrotactile stimulation can be attributed to modulation of the central cholinergic system acting as a cerebrovascular dilator and not to simple changes in the global cerebral blood flow caused by changes in physiological conditions.

The roles of cholinergic neuronal systems in cognition, learning and memory have been studied in experimental animals (mainly rats and mice), to explain the pathogenesis of human disorders associated with dementia. However, it is difficult to obtain clear answers relevant to human disorders from results obtained in rodents, because of the species differences and because most animal studies have been performed with anesthesia. The effects of cholinergic modulators on rCMR in rats were sometimes similar to and sometimes different from those in humans, reflecting species differences in the proportions of cholinergic receptor subtypes or in their affinities for each modulator (Blin et al., 1995). Different kinds of anesthesia induced different effects on cerebral functions such as rCBF and rCMR (Gjedde et al., 1980; Sokoloff, 1981), functional hemodynamic-metabolic coupling (Crosby et al., 1983) and neurotransmission, as measured by ligand-receptor binding in vivo (Kobayashi et al., 1995; Onoe et al., 1994). In the present study, we used unanesthetized monkeys as an experimental model to minimize the species differences and the effects of anesthesia. The combined use of conscious monkeys and PET, a method for noninvasive determination of cerebral biochemistry and physiology in vivo, can offer a unique bridge between animal and human experimental protocols, especially to predict therapeutic effects in humans from the results of animal studies.

In conclusion, the present study suggested that compounds with cholinesterase inhibitory activity can reverse the scopolamine-abolished coupling between neuronal activity and rCBF responses to somatosensory stimulation via enhancement of cholinergic neurotransmission. In addition, this study strongly supported the hypothesis that cholinergic mechanisms might be involved in regulation of the coupling mechanism in functional activation. This experimental procedure may provide the easiest method to evaluate compounds designed as cognitive enhancers, before clinical trials for treatment of patients with dementia. Among the three cholinesterase inhibitors examined here, the longer acting and less harmful compound E2020 may be ideal for therapeutic use, if applied with care.