Vascular Action as the Primary Mechanism of Cognitive Effects of Cholinergic,CNS-Acting Drugs,a Rat phMRI BOLD Study

Abstract

Concordant results of functional magnetic resonance imaging (fMRI) and behavioral tests prove that some non-blood-brain barrier-penetrating drugs produce robust central nervous system (CNS) effects. The anticholinergic scopolamine interferes with learning when tested in rats, which coincides with a negative blood-oxygen-level-dependent (BOLD) change in the prefrontal cortex (PFC) as demonstrated by fMRI. The peripherally acting butylscopolamine also evokes a learning deficit in a water-labyrinth test and provokes a negative BOLD signal in the PFC. Donepezil—a highly CNS-penetrating Cholinesterase inhibitor—prevents the negative BOLD and cognitive deficits regardless whether the provoking agent is scopolamine or butylscopolamine. Interestingly, the non-BBB-penetrating Cholinesterase inhibitor neostigmine also prevents or substantially inhibits those cognitive and fMRI changes. Intact cerebral blood flow and optimal metabolism are crucial for the normal functioning of neurons and other cells in the brain. Drugs that are not BBB penetrating yet act on the CNS highlight the importance of unimpaired circulation, and point to the cerebral vasculature as a primary target for drug action in diseases where impaired circulation and consequently suboptimal energy metabolism are followed by upstream pathologic events.

Keywords

animal models blood-brain barrier BOLD contrast cognitive impairment functional MRI

INTRODUCTION

The brain is the most energy-demanding organ in the body. The high energy requirement is met exclusively by the oxidative breakdown of glucose. The function of the brain depends on the availability of glucose and oxygen, and therefore, it is highly susceptible to minor changes in the blood supply, which is the carrier of nutrients.

A deficiency in energy supply is inevitably followed by a deficiency of function. It is generally recognized that ischemic deprivation (oxygen and glucose starvation) has detrimental effects on cognitive functions. Even a small fluctuation in cerebral oxygen delivery may have an impact on cognitive performance. With age, the cortical blood supply is reduced by up to 30%. A greater reduction in regional blood flow was demonstrated in patients with memory impairment. Wang et al¹ concluded from a multicenter clinical study that a decrease in cerebral blood flow is the first step towards Alzheimer's disease.

Most importantly, improving blood circulation, especially in regions associated with cognitive processes, may reverse the cognitive impairment caused by insufficient cellular energy production. This may happen without the direct modulation of neurotransmitters or their receptors. The restoration of blood supply will restore cognitive functions, and also prevents secondary damage. The role of vascular and endothelial pathologic processes in dementias is increasingly being addressed in the literature. Decreased blood flow in the brain can trigger the onset of dementia in Alzheimer's disease, conclude Palmer and Love² in their review.

It is generally recognized that a central nervous system (CNS) action requires a drug to be present in the CNS. The group of therapeutic Cholinesterase inhibitors is an excellent example for demonstrating the structural relationship of brain penetrability. Quaternary nitrogen-containing inhibitors such as neostigmine exert their effects only peripherally, as they are unable to penetrate into the CNS.³ The quaternary nitrogen-containing and therefore peripherally acting neostigmine has gained a therapeutic application in peripheral indications such as myasthenia gravis.

In contrast to neostigmine, donepezil penetrates into the CNS readily, resulting in a brain concentration of up to 10 times higher than plasma.⁴

There is ample early data in the literature to prove that neostigmine is free of CNS side effects in humans. However, in an early study by Prohovnik et al,⁵ when the effects of physostigmine and the non-penetrating neostigmine were compared in humans, the authors could not find a plausible explanation for the CNS effect produced by neostigmine. As the amelioration of cognitive deficit by neostigmine requires intracerebral action, they concluded that either the explanation could be an experimental error because of the small sample size, or neostigmine must penetrate into the brain to a greater extent than is generally believed. The unmentioned third possibility might be that the local circulatory change in the brain vasculature evoked peripherally by neostigmine has a cognitive effect because of the bidirectional neurovascular coupling. To test the idea, a pharmacological magnetic resonance imaging study and memory tests were conducted to compare the effect of the memory-disturbing agent scopolamine⁶ with its derivative, butylscopolamine, which does not cross the blood-brain barrier. Donepezil, a clinically well-established, highly brain-penetrating procognitive agent, and neostigmine, a peripherally acting Cholinesterase inhibitor, were tested for their ability to reverse the effect of memory-disturbing agents.^6–8 The similarities in the extent of memory and local circulatory impairment (measured by functional magnetic resonance imaging (fMRI)) and their reversal by brain-penetrating or non-penetrating drugs suggests that cognitive performance is greatly determined by vascular factors. Therefore, a target for new cognitive enhancer drugs does not necessarily have to be located beyond the blood-brain barrier. This is especially true under conditions where insufficient blood supply limits neuronal functions, and their direct stimulation may do more harm than good.

MATERIALS AND METHODS

The present study has been approved by Gedeon Richter Pic.

Functional Magnetic Resonance Imaging Study

Male Wistar rats were used for the experiments. For the fMRI studies, animals weighing 240 to 260 g were purchased from Harlan. The animals were kept in polycarbonate cages in a thermostatically controlled room at 21 ± 1°C The room was artificially illuminated from 6 a.m. to 6 p.m. The rats were fed with conventional laboratory rat food (ssniff R/M + H Spezieldiäten D-59494 Soest). All procedures conformed to the 1986/609 EU directive and were approved by the Ethical Committee of Gedeon Richter Pic.

Functional magnetic resonance imaging experiments were performed in a 9.4T ASZ Varian MRI system with a free bore diameter of 210 mm, fitted with a 120 mm inner size gradient coil (180 μs rise time). For excitation, an actively RF-decoupled 2-channel volume coil system with an internal size of 72 mm was used, and a fix-tuned receive-only phased array rat brain coil located directly above the dorsal surface of the rat's head to maximize the signal-to-noise ratio.

Scout pictures were obtained in planes of coronal and sagittal, to set the anatomic and functional images. Anatomic scans were acquired using a gradient echo multislice sequence with a field of view of 35 × 35 mm, slice thickness of 1 mm, and gap of 0.2 mm. Nine slices were received in an interleaved order; the scanner's default coronal orientation was changed slightly to obtain a standard anatomically coronal plane on the basis of the Rat Brain Atlas of Paxinos & Watson. Echo time, TE = 3.83 milliseconds, repetition time, TR = 200 milliseconds, flip angle 45°, averages 3, dummy scans 4, data matrix 192 × 192, total scan time 2 minutes. An interleaved triple-shot gradient-echo echo planar imaging sequence with compressed segments was used for the T2^*-weighted MR images. TE = 10 milliseconds, TR = 3,000 milliseconds, flip angle 90°, averages 1, dummy scans 4, data matrix 64 × 64, 1,000 repetitions. Field of view and slice parameters were the same as in the anatomic setup.

The rats were anaesthetized with isoflurane (5% starting concentration for ~5 minutes, and then 1% to 1.5% during scanning, depending on the breath number as a marker of depth of anesthesia) administered in compressed air. For intravenous drug administration, a cannula line was inserted and used during the scanning. The anesthetized rat was placed into the magnet. Body temperature was monitored using a rectal probe and maintained at 37 ± 1°C with thermostatically controlled air, circulated around the rat. The ventilation of the animal was also controlled. The experiments lasted for 50 minutes. After 16 minutes and 40 seconds (1,000 seconds) baseline period scopolamine, butylscopolamine, or saline was administered intravenously at a dose of 1 mg/kg. The automated drug administration was performed by an infusion pump controlled via optical signals. In the fMRI study, Cholinesterase inhibitors were applied intraperitoneally to the conscious animals as a pretreatment 1 hour before the scopolamine or butylscopolamine. Only one measurement was performed on each animal.

The results of each measurement were stored in the scanner's own file format (fdf-files). These files were converted to the widely used nifti-format (Neuroimaging Informatics Technology Initiative) using a proprietary Matlab (The Mathworks, Natick, MA, USA) script. The analysis and the visualization of the data were also performed using a proprietary Matlab script. This script uses one-way within-subjects one-factor analysis of variance (ANOVA) with two levels (with and without drug administration) for each scan. In this case, ANOVA is a voxel-wise paired t-test with samples from the pre-injection baseline and post-injection period, respectively.⁹

The t-test was corrected using the Benjamini and Hochberg false discovery rate method.⁹ A random effect analysis was not considered necessary, because the animals formed a homogeneous group (males only, same species, same weight, and other circumstances were also standard—see above), and the onset of blood-oxygen-level-dependent (BOLD) change coincided with the application of test drug and the change was compared with the pre-drug level. Movement was checked, and measurements with a higher than one voxel movement were rejected. To create t-maps in the pharmacological magnetic resonance imaging experiments, a paired t-test was performed on each voxel's two-time intervals (pre-injection baseline and post-injection) to determine the significant differences between the baseline signal and post-injection signal. Thus a t-value was assigned to each voxel, and voxels over the t-value limit were highlighted. Region of interest analysis was also performed using a proprietary Matlab script. The ROIs were determined on the basis of the Rat Brain Atlas of Paxinos & Watson.

The statistical significance of the drug effect was calculated using ANOVA and a post hoc Fisher test.

Water-Labyrinth Learning Performance

Male Wistar rats obtained from Toxicoop, Hungary, weighing 180 to 200 g were used in the water-labyrinth experiments. The water-labyrinth spatial learning test has already been described in detail.¹⁰ Briefly, the animals had to maneuver through three choice points of a labyrinth system to reach a platform that allowed them to escape from the water. The water tank (1 m long, 60 cm wide, and 60 cm deep) was filled with water at 24 ± 2°C to a depth of 30 cm. The labyrinth system was constructed with removable vertical metal plates. The escape platform (10 × 7 cm), with its top surface raised 0.5 cm above the water level, was placed in the corner of the tank farthest from the start point. The procedure for testing the water-labyrinth acquisition process was carried out on 4 consecutive days.

Adaptation (day 1): In order to get the animals to adapt to the test environment, the metal plates constituting the labyrinth system were removed from the water and the rats were conditioned three times to swim in the tank from the start point to the platform. The animals were left on the platform for 20 seconds and they were allowed to rest in their cage for 20 minutes between each swim.

Training procedure (day 2 to 4): One daily session consisting of three trials was performed. On the training days, the labyrinth system was in place. The rats were placed into the water at the start point of the labyrinth system and had to reach the escape platform. The animals spent 20 seconds on the platform, and then they were placed into a shared cage for ~20-minute rest periods between trials. The number of directional turning errors was measured as a variable reflecting learning performance. An error was defined as swimming through a choice point in the direction that did not allow the animal to reach the platform (blind alley). When the rat made an error, it was allowed to swim back to the choice point and try again. However, once a rat swam over a choice point in the correct direction (leading out of the labyrinth), the way back was manually closed by a metal plate. The swimming time (the time interval from the entry into the labyrinth until the exit from the water) could not exceed 5 minutes for any trial. If the rat did not find the platform during this period, it was assisted to the end of the labyrinth by the experimenter, and the number of errors was recorded as 12.

Thirty minutes before the start of the first daily trial, 3 mg/kg (2 mL/kg) scopolamine hydrobromide or 1 and 3 mg/kg butylscopolamine was injected intraperitoneally as amnestic agents. Donepezil (0.25, 0.5 mg/kg oral) and neostigmine (0.3 mg/kg intraperitoneal) were given 30 minutes before the first daily swim. Each study included a non-impaired solvent control, a memory-impaired group (induced by 3 mg/kg dose of scopolamine or butylscopolamine), and impaired groups treated with donepezil or neostigmine. There were 10 animals in each experimental group. The effect of neostigmine against butylscopolamine impairment was investigated in two repeated studies, and the results were summarized.

Data Analysis

The individual values (number of errors) were averaged across the three daily trials. Then the group means were calculated from these daily individual data. The results are given as mean error ± s.e.m. of groups for each daily session. Statistical comparisons between parameters of each group were made by ANOVA (two-way repeated-measures ANOVA) using ‘groups’ as the independent between-groups factor and ‘days’ as the repeated-measures factor. Post hoc comparisons (Duncan test) were performed in the event of a significant between-groups effect or a significant interaction between the independent and repeated-measures factors.

The percentage rate of reversal of the amnesia by the compound was calculated from the group means of pooled errors for all the trials in the training days, using the formula:

% = \frac{N_{err} of amnestic - N_{err} of compound}{N_{err} of amnestic - N_{err} of control} \times 100

RESULTS

Scopolamine, at 1 mg/kg intravenous dose caused a marked negative BOLD effect in the prefrontal cortex (PFC), while other brain areas remained essentially unchanged (Figure 1A). Butylsco-polamine's effect was highly similar; it showed an identical inhibitory pattern and BOLD reduction of similar extent in the brain (Figure 1B).

Figure 1

Effect of scopolamine (A) and butylscopolamine (B) on blood-oxygen-level-dependent (BOLD) responses in the rat brain after vehicle pretreatment. Clear negative BOLD effect can be seen in the area of prefrontal cortex. (C) Lack of negative BOLD effect by scopolamine after neostigmine pretreatment. Data are from individual experiments.

The effect of scopolamine was fully prevented by pretreatment with donepezil at a dose of 4 mg/kg, intraperitoneal. Pretreatment with the peripherally acting Cholinesterase inhibitor neostigmine (0.1 mg/kg intraperitoneal) had a similar, somewhat lower yet statistically significant effect on the BOLD response evoked by scopolamine in the PFC (Figure 1C). At the doses applied, neither donepezil nor neostigmine evoked any significant changes in the BOLD responses when administered alone (not shown).

The negative BOLD effect of butylscopolamine (buscopan) was fully reversed by both Cholinesterase inhibitors, donepezil and neostigmine (Figure 2).

Figure 2

Effect of saline, scopolamine, and butylscopolamine (buscopan) on BOLD responses in the rat brain after saline pretreatment, and prevention of the effect of scopolamine (scop) and butylscopolamine (busc) by neostigmine (neost) or donepezil (donep) pretreatment. The square marks the mean of group data. The box diagram shows the median of data in 25th and 75th percentiles. The whiskers are determined by the 5th and 95th percentiles, which is practically same as minimum and maximum value. ^*P<0.05, ^***P<0.005 from saline; ^#P<0.05, ^###P<0.005 from scopolamine; ^{^^^}P<0.005 from the butylscopolamine. The arrow means significant difference between the reversal effect of neostigmine and donepezil against scopolamine P<0.05 (analysis of variance and Fisher post hoc test).

In the water-labyrinth test, scopolamine (3 mg/kg) had a significant inhibitory effect on the learning performance of the animals. Donepezil at doses of 0.25 and 0.5 mg/kg significantly improved the learning deficit induced by scopolamine (F(3.36) = 12.13, P<0.0000), producing a 50% and 63.5% reversal, respectively (Figure 3A). Neostigmine also caused a highly significant (F(2.26) = 17.94, P<0.0000) and considerable (76.8%) reversal (Figure 3B).

Figure 3

Effect of donepezil (A) at a dose of 0.25 (donep/0.25) and 0.5 mg/kg (donep/0.5) and neostigmine (B) at 0.3 mg/kg (neost/03) on the learning deficit induced by 3 mg/kg scopolamine (scop/3) in the water-labyrinth test.⁺P<0.05, ⁺⁺P<0.01, ⁺⁺⁺P<0.001 versus control group ^*P<0.05, ^**P<0.01, ^***P<0.001 versus scopolamine-treated group.

Butylscopolamine showed a dose-dependent effect; it caused a significant impairment in learning performance at a 3 mg/kg dose (F(2.27) = 3.34, P<0.050) (Figure 4A), which was reversed by 0.3 mg/kg of neostigmine (F(2.52) = 4.19, P<0.020) (Figure 4B).

Figure 4

Effect of butylscopolamine at a dose of 1 (busc/1) and 3 mg/kg (busc/3) on the learning performance of rats (A) and effect of neostigmine at 0.3 mg/kg (neost/03) on the learning deficit induced by 3 mg/kg butylscopolamine (B) in the water-labyrinth test. ^*P<0.1, ⁺⁺P<0.01, ⁺⁺⁺P<0.001 versus control group, ^**P<0.01 versus butylscopolamine-treated group.

DISCUSSION

Both scopolamine and its non-brain-penetrating analog, butylscopolamine, strongly reduced the BOLD responses in the PFC. Interestingly, in other brain areas, e.g., the hippocampus, no significant BOLD changes could be detected. The PFC has been implicated in working memory, attention, response initiation, and emotion. In humans, dysfunction of the prefrontal cortical areas is related to psychopathologies such as schizophrenia, sociopathy, obsessive–compulsive disorder, depression, and drug abuse.¹¹ The medial PFC is one of the main parts of the so called default mode network, which has a key role in mental processes.¹² Our conclusions on the importance of cholinergic vascular regulation are in good agreement with the study by Zhuo et al¹³ on the cholinergic modulation of working memory.

The pretreatment of animals with the clinically well-established cognitive enhancer drug, donepezil, fully prevented the effect of scopolamine in the cognitive test, as well as in the fMRI test. Interestingly, the brain non-penetrating scopolamine analog, butylscopolamine, produced a BOLD effect that is highly similar to that of scopolamine. Furthermore, the peripherally acting Cholinesterase inhibitor neostigmine was also capable of preventing the effect of both scopolamine and butylscopolamine. These BOLD results suggest that the memory-disturbing effect of scopolamine is mediated, to a considerable extent, by a vascular mechanism. The water-labyrinth test confirmed the conclusions drawn from the fMRI BOLD study. Scopolamine as well as butylscopolamine caused deterioration in the learning performance of rats. The learning disability evoked by either scopolamine or by butylscopolamine was inhibited by the non-penetrating neostigmine.

The impairment of cognition by anticholinergic drugs probably happens via a vascular mechanism. The primary mechanism of procognitive Cholinesterase inhibitors is probably also vascular.

The cross-checking of brain-penetrating or non-penetrating anticholinergic drugs and Cholinesterase inhibitors proves that, at least in the prefrontal and frontal cortex, intact circulation is a determining factor for normal cognition.

Neurovascular coupling works in both directions. Increased neuronal activity induces circulation, which is reflected by a positive BOLD answer. Cholinergic inhibition will reduce circulation of prefrontal vasculature which in turn reduces neuronal activity reflected by impaired learning and a negative BOLD, while cholinergic stimulation will restore normal circulation and neuronal functioning.

Our conclusion is supported by several findings in the literature. Sato and Sato^14–16 demonstrated that electric stimulation of the cholinergic nucleus basalis of Meynert produced a metabolism-independent vascular dilatory effect in the frontal-prefrontal area. Hoff et al¹¹ showed that the non-selective muscarinergic agonist pilocarpine increased blood flow in the cortex, thalamus, and hippocampus. They detected this effect using the sensitive cerebral blood volume method, but not with BOLD fMRI. Among the studied regions, the cortical areas revealed the highest cholinergic sensitivity.

In an early report,¹⁸ scopolamine attenuated the memory task-induced increases in cerebral blood flow in the left and right PFC and the right anterior cingulate region, suggesting a role of vascular muscarinic receptors in the neurovascular coupling, activated during cognitive tasks. It has been reported by Andrews et al⁹ that the non-brain-penetrating scopolamine analog, methyl scopolamine, disturbed the performance of rats in a PFC-specific paradigm, the delayed matching to position procedure.

These literature data support the view that a vascular pathomechanism, with implications that reach far beyond the immediate impact on cognition reported here, may be a major contributing factor in dementias such as Alzheimer's disease (AD), among others. Several clinical studies have been conducted to address the vascular aspect of AD. A study by Marchant et al²⁰ suggests that in ischemic cerebrovascular disease, cognition is more influenced by cerebral blood flow, and Aβ showed no association with cognition in this case. In the review by Tong and Hamel²¹ on the role of cholinergic system in blood supply, its possible role is discussed in the development of Alzheimer's disease. Indeed, in mild AD subjects, the regional blood flow was significantly increased after donepezil treatment in the middle cingulate cortex and posterior cingulate cortex,²² which are the neural substrates of the medial cholinergic pathway. In both brain regions, the baseline blood flow and its alteration on donepezil treatment were significantly correlated with the behavioral changes in ADAS-cog scores. Some procognitive agents directly increase cerebral blood flow, for example, alpha 7 nicotinergic agonists,²³ or even donepezil. The improved circulation is not only a marker of improved cognition, but also a prerequisite for it.

In conclusion, the present study and others from the literature demonstrate that cholinergic compounds that do not cross the blood-brain barrier can exert a remarkable effect on delicate CNS functions such as cognition, and they also highlight the role of vascular components in the action of scopolamine on brain function. Curiously, the peripherally acting butylscopolamine used as premedication for gastrointestinal endoscopy reportedly caused anterograde amnesia in patients.²⁴ The cognitive amelioration of scopolamin's effect by neostigmine in humans also points to the human relevance of cholinergic vascular regulation in cognitive processes.⁵

Footnotes

The authors declare no conflict of interest.

ACKNOWLEDGMENTS

The authors thank Katalin Tóth and Petra Schreiber for their excellent technical assistance, and Attila Csáki for his help in the mathematical analysis.

References

Wang

Das

Xie

Arnold

Detre

Wolk

. Arterial spin labeled MRI in prodromal Alzheimer's disease: a multi-site study. NeuroImage Clin 2013; 2: 630–636.

Palmer

Love

. Endothelin receptor antagonists: potential in Alzheimer's disease. Pharmacol Res 2011; 63: 525–531.

Goodman Gilman

Goodman

Rail

Murad

. The Pharmacological Basis of Therapeutics. 7th edn. (Macmillan Publishing Company: New York 1985, pp 116.

Matsui

Mishima

Nagai

Yuzuriha

Yoshimura

. Absorption, distribution, metabolism, and excretion of donepezil (Aricept) after a single oral administration to rat. Drug Metab Dispos 1999; 27: 1406–1414.

Prohovnik

Arnold

Smith

Lucas

. Physostigmine reversal of scopolamine-induced hyperfrontality. J Cereb Blood Flow Metab 1997; 17: 220–228.

Snyder

Bednar

Cromer

Maruff

. Reversal of scopolamine-induced deficits with a single dose of donepezil, an acetylcholinesterase inhibitor. Alzheimers Dement 2005; 1: 126–135.

Lenz

Baker

Locke

Rueter

Mohler

Wesnes

. The scopolamine model as a pharmacodynamic marker in early drug development. Psychopharmacology 2012; 220: 97–107.

Riedel

Kang

Choi

Platt

. Scopolamine-induced deficits in social memory in mice: reversal by donepezil. Behav Brain Res 2009; 204: 217–225.

Friston

Ashburner

Kiebel

Nichols

Penny

. Statistical Parametric Mapping: The Analysis of Functional Brain Images, 1st edn. Academic Press: New York, NJ, UK 2007.

10.

Laszy

Laszlovszky

Gyertyan

. Dopamine D3 receptor antagonists improve the learning performance in memory-impaired rats. Psychopharmacology (Berl) 2005; 179: 567–575.

11.

Heidbreder

Groenewegen

. The medial prefrontal cortex in the rat: evidence for a dorsoventral distinction based upon functional and anatomical characteristics. Neurosci Biobehav Rev 2003; 27: 555–579.

12.

Broyd

Demanuele

Debener

Helps

James

Sonuga-Barke

. Default-mode brain dysfunction in mental disorders: a systematic review. Neurosci Biobehav Rev 2009; 33: 279–296.

13.

Zhou

X-L

Douglas

Palaninathan

Kang

Buccafusco

. Cholinergic modulation of working memory activity in primate prefrontal cortex. J Neurophysiol 2011; 106: 2180–2188.

14.

Sato

. Cholinergic neural regulation of regional cerebral blood flow. Alzheimer Dis Assoc Disord 1995; 9: 28–38.

15.

Sato

Uchida

. Regulation of regional cerebral blood flow by cholinergic fibers originating in the basal forebrain. Int J Devl Neurosci 2001; 19: 327–337.

16.

Sato

Uchida

. Activation of the intracerebral cholinergic nerve fibers originating in the basal forebrain increases regional cerebral blood flow in the rat's cortex and hippocampus. Neurosci Lett 2004; 361: 90–93.

17.

Hoff

Oostenbrugge

Otte

van der Marel

Steinbusch

HWM

Dijkhuizen

. Pharmacological magnetic resonance imaging of muscarinic acetylcholine receptor activation in rat brain. Neuropharmacology 2010; 58: 1252–1257.

18.

Grasby

Frith

Paulesu

Friston

Frackowiak

RSJ

Dolan

. The effect of the muscarinic antagonist scopolamine on regional cerebral blood flow during the performance of a memory task. Exp Brain Res 1995; 104: 337–348.

19.

Andrews

Jansen

JHM

Linders

Princen

. Effects of disrupting the cholinergic system on short-term spatial memory in rats. Psychopharmacology 1994; 115: 485–494.

20.

Marchant

Reed

DeCarli

Madison

Weiner

Chui

. Cerebrovascular disease, beta-amyloid, and cognition in aging. 2012 Neurobiol Aging 33: 1006.e25–1006.e36.

21.

Tong

Hamel

. Regional cholinergic denervation of cortical microvessels and nitric-oxid synthase-containing neurons in Alzheimer's disease. Neuroscience 1999; 92: 163–175.

22.

Antuono

Xie

Chen

Jones

Ward

. Changes in regional cerebral blood flow and functional connectivity in the cholinergic pathway associated with cognitive performance in subjects with mild Alzheimer's disease after 12-week donepezil treatment. NeuroImage 2012; 60: 1083–1091.

23.

Lee

TJF

. α7-Nicotinic acetylcholine receptors on cerebral perivascular sympathetic nerves mediate choline-induced nitrergic neurogenic vasodilation. Circ Res 2002; 91: 62–69.

24.

Lee

Son

Rhee

Kim

Rhee

. Unexpected anterograde amnesia associated with Buscopan used as a predmedication for endocscopy. World J Gastroenterol 2007; 13: 28.