HOW ANTICHOLINERGIC DRUGS MIGHT PROMOTE ALZHEIMER’S DEMENTIA: MORE ABETA AND LESS PHOSPHATIDYLCHOLINE
Richard J. Wurtman, MD, Massachusetts Institute of Technology, 77 Mass Ave., Cambridge, MA 02139 USA
Corresponding author: Richard J. Wurtman
77 Mass Ave., Bldg. 46-5009, Cambridge, MA 02139
Phone: 617-253-6731; FAX 617-253-6992
Emal:
ABSTRACT
Drugs that block muscarinic cholinergic neurotransmission in the brain can, as a consequence, increase the formation of Abeta, and decrease brain levels of phosphatidylcholine (by slowing its synthesis and accelerating its turnover). Both of these effects might cause a decrease in brain synapses, as characterizes and probably underlies the memory disorder of early Alzheimer’s disease.
Key words: acetylcholine, choline, amyloid, neurotransmission, synaptic membrane, synapses
INTRODUCTION
A decrease in cortical and hippocampal synapses is an early and robust finding in AD brains (1), and generates memory impairments by weakening the connections among brain regions that underlie memory (2). This decrease is widespread, however it is most pronounced among cholinergic synapses - numbers of which may decline by 90% (3,4) - and is ultimately accompanied by the loss of cholinergic parikarya in the basal forebrain and septum (5).
The decrease in synaptic number could result from either of two processes, or from a combination of both: an acceleration in their breakdown, perhaps caused by an endogenous neurotoxin like Abeta42 polymers (6), or a slowing in their formation (which, in normal brains, occurs continuously, replacing synapses that disappear after their short life-spans of days to weeks [7]). The decrease in synapses in AD brain is associated with a parallel decline in the numbers of dendritic spines - the essential anatomic precursors for excitatory brain synapses; this has been observed both in animal models of AD (8) and in brains of AD patients (9).
A recent epidemiologic study (10) has shown that the high cumulative use of anticholinergic drugs may promote the development of AD and other dementing diseases. It had been known that single doses of muscarinic receptor blockers, like scopolamine, could acutely impair short-term memory, and that the impairment would subside if subjects were also given an acetylcholinesterase inhibitor to raise intrasynaptic acetylcholine levels (11). However the possibility that anticholinergics could also produce long-term damage was generally unrecognized. This neurotoxic effect apparently is independent of an anticholinergic agent's chemical family or its other pharmacological properties, occurring with drugs as diverse as antihistamines, antispasmodics, antiemetics, and antipsychotics: Apparently all that is required is that the drug block cholinergic receptors (10). Hence the repeated blockade of central cholinergic neurotransmission, like AD itself, probably decreases cortical and hippocampal synapses, probably also by accelerating their breakdown or slowing their production or both.
How might the chronic drug-induced blockade of central cholinergic transmission promote the development of AD? At least two established biochemical mechanisms, both affecting synaptic number, could underlie such an effect. One would involve increasing the production of potentially-toxic Abeta peptides and their oligomers (12) and decreasing the ACh-stimulated hydrolysis of the amyloid precursor protein (APP) (13-16) which forms non-amyloidogenic intermediates. The other mechanism, operating principally in cholinergic brain neurons, would impair synaptogenesis by lowering the levels of phosphatidylcholine (17,18) and other, related constituents of synaptic membrane. As discussed below, the activation of m1 and m3 muscarinic receptors (or other receptors that generate diacylglycerol (DAG) from phosphatidylinositol [PI]) is known to inhibit the cleavage of APP by beta- and gamma-secretase enzymes which would form Abeta and its oliogomers, and to enhance the activity of alpha-secretase enzymes which cleave APP to form non-amyloidogenic metabolites (13-16). Thus, as might be expected, the blockade of m1 and m3 receptors by muscarinic antagonists enhances Abeta synthesis and suppresses alpha-secretase activity (12).
The depolarization of cholinergic neurons also affects, indirectly, the breakdown of the PC in neuronal membranes (17 18) and slows its synthesis. PC breakdown liberates free choline, most of which is diverted for ACh synthesis and away from regenerating PC (17). It thereby reduces levels of PC and other membrane phosphatides, thus diminishing the quantity of neural membrane available for synaptogenesis. In AD the major loss of cortico-hippocampal cholinergic synapses very likely increases the firing frequencies of the surviving terminals; moreover their firing is further increased if a muscarinic antagonist has blocked postsynaptic responses to the ACh. The higher firing frequency enhances PC breakdown and conversion of the liberated choline to ACh, not PC (17). And if, in addition, a muscarinic receptor antagonist has been administered, the resulting blockade of presynaptic receptors will further increase the loss of ACh and choline from the terminal, and shunt free choline into forming ACh and not PC (17,20,21). Hence PC levels fall, - as has been observed in frontal and parietal regions of AD brain [19 ]. The enhanced turnover of membrane PC to liberate choline which is then used for ACh synthesis reflects a kind of "autocannibalism" [18- 21])
This review briefly describes these two mechanisms, and suggests ways of ameliorating their consequences in AD patients.
CHOLINE, CHOLINERGIC TRANSMISSION, AND THE PC IN SYNAPTIC MEMBRANE
The enzymes choline acetyltransferase (CAT), which acetylates choline to form ACh, and choline kinase (CK), which phosphorylates choline to initiate PC synthesis, both have unusually low affinities for their choline substrate; their Km's are 540 uM and 2.6 mM respectively, while the brain's choline concentration is only about 35uM (22,23). Hence both enzymes are highly unsaturated in vivo, and even small changes in brain choline levels can have important effects on the rates at which it is acetylated or phosphorylated. As discussed below, the conversion of choline to PC involves a three-step process: first CK converts the choline to phosphocholine; that compound then combines with cytidylyltriphosphate (CTP) to form CDP-choline; and the CDP-choline then combines with particular diacylglycerol (DAG) molecules which contain an omega-3 acid (i.e., DHA or EPA) to yield the PC. The CTP and the DHA-containing DAG species needed for this pathway are formed in brain from circulating uridine and omega-3 fatty acids, by enzymes that are also unsaturated with these substrates (24,25).
Circulating choline derives both from dietary sources (e.g., egg yolks, in the form of PC) and, probably to a greater extent, from the hepatic synthesis and hydrolysis of PC (25), formed by the sequential, B-vitamin-dependent methylation of phosphatidylethanolamine [PE] ) (26), The recommended dietary intake of choline by normal adults is about half a gram per day, however many women and older people of both sexes fail to attain this intake. Moreover basal choline requirements may be significantly greater among individuals who, unknowingly, have common variations in gene alleles for some of the hepatic enzymes involved in PE methylation or folic acid metabolism (27). An increased need for dietary choline - e.g., for people carrying such alleles, or in AD patients who might benefit from accelerating synaptogenesis, as described below - can, to some extent, be satisfied by providing oral choline. However doses in excess of 500 mg cause many users to develop an unacceptable "fishy body odor" caused by choline's intra-intestinal breakdown to trimethylamine, Thus the liver’s contribution to plasma choline may also need to be enhanced, by administering the B-vitamins that promote the methylation of PE (26).
Circulating choline enters the brain principally via a concentration-dependent, facilitated diffusion mechanism, mediated by a transport protein in endothelial cells that line the brain's capillaries. PC and lysoPC are not substrates for a transport protein, and enter the brain only poorly. Once in the brain's extracellular fluid choline enters all cells by a low-affinity transport system, or is taken up into terminals of ACh neurons by a high-affinity uptake mechanism. As discussed below, much of the choline used for ACh synthesis in rapidly-firing neurons, or following administration of an anticholinergic drug, derives from the local breakdown of membrane PC (17).
The initial experimental evidence that choline used for ACh synthesis can derive from the breakdown of membrane PC came from studies on electrically-stimulated, superfused slices of rat corpus striatum that were exposed to various concentrations of free choline (17, 18). Stimulation of slices in a choline-free medium for four or six 20-minute periods caused 4-fold increases in ACh release, with no reduction in tissue ACh levels. The total amount of ACh released during the three-hour superfusion period was about three times the initial amount in the tissues, and the amount of free choline released was even greater, i.e. 45 times the initial or final amounts in the tissues.
Stimulation also caused significant 14% reductions in tissue PC levels, and comparable reductions in levels of protein (10.5%) and in the other principal membrane phosphatides (PE and PS). This indicated that the fall in PC represented the loss of both major membrane constituents, phosphatides and proteins (17). The reductions in phosphatide levels and the increase in ACh release were all blocked by addition of tetrodotoxin to the medium, indicating that they were mediated by neuronal depolarization. Moreover slices of cerebellum - a tissue with, unlike striatum, minimal cholinergic innervation - neither released significant amounts of ACh during stimulation nor exhibited post-stimulation reductions in phosphatide or protein contents. This suggested that most of the reductions in membrane constituents in stimulated striatal tissue (and most of the choline used for ACh synthesis) had originated in synaptic membranes of cholinergic neurons.
Addition of choline (10-40 uM) to the media partially or completely blocked the stimulation-induced reductions in striatal phophatides and protein, the highest choline concentrations actually increasing tissue levels of all of the phosphatides. This suggested that, during stimulations, the tissue retained the capacity to produce neural membranes, but that, without supplemental choline, most of the newly-synthesized membrane was broken down to provide choline for ACh synthesis.
It can be hypothesized that chronic treatment with anticholinergic drugs has effects on PC breakdown and choline utilization that are similar to those observed after electrical stimulation of striatal slices, since blockade of postsynaptic muscarinic receptors in vivo would cause a reflex increase in the firing frequency of the presynaptic cholinergic neurons, and blockade of the presynaptic receptors on those neurons would very much increase their release of ACh and consequent loss of choline.
CHOLINERGIC TRANSMISSION AND ABETA FORMATION
The activation of muscarinic receptors that are coupled to phosphatidylinositol (PI) breakdown, DAG formation, and protein kinase C (PKC) activation enhances the activity of alpha-secretase enzymes that convert APP to non-amyloidogenic peptides, and suppresses the formation by beta- and gamma-secretases of the amyloidogenic Abeta peptides (6). Thus administration of carbachol (13), a general muscarinic agonist, or drugs like AF 102B which specifically target M1 (28) or M3 (29) receptors, decreases Abeta levels and increases non-amyloidogenic peptides in brain. While the general muscarinic antagonist scopolamine (12) or the selective M1 antagonist dicyclomine (30) promote amyloidogenic APP processing (31).
It can thus be understood that chronic frequent administration of muscarinic antagonists might cause prolonged increases in brain Abeta formation, and thereby, as Gray et al have shown (10) constitute a risk factor for Alzheimer's disease and possibly other dementias.
SOME IMPLICATIONS OF THESE RELATIONSHIPS
Clearly, though AD brains ultimately show major reductions in most varieties of synapses and neurons, the decrease in cholinergic synapses may be of special importance, both because muscarinic cholinergic neurotransmission is essential for normal memory and because, when impaired, this has effects on phosphatide availability and Abeta formation which may extend the synaptic loss. And if cholinergic transmission is further compromised by chronically administering anticholinergic drugs, the likelihood of developing dementia is thereby enhanced (10). In contrast, a cholinergic agonist might be expected to have opposite and beneficial effects: It might act postsynaptically to decrease Abeta production and, by slowing the neuron’s firing frequency, increase the levels of phosphatides that can be used to generate synaptic membranes. And acting presynaptically the agonist might further reduce the loss of choline to ACh production, thus allowing more to be used to form PC.
This hypothesis might be tested by determining whether the chronic administration of presently-utilized acetylcholinesterase (AcChE) inhibitors - which are presumed to increase intrasynaptic ACh – actually elevate ACh release into cerebrospinal fluid and, if they do, whether the dose-response relationship of the ACh increase parallels that for reducing CSF Abeta levels. Higher drug doses might be required for these effects than the doses presently used to treat patients with early AD, and even those relatively low current doses are sometimes associated with unacceptable peripheral side-effects.
Another implication of the ability of cholinergic antagonists to promote dementing diseases is that sufficient choline needs to be provided so that adequate PC production can continue. This might most effectively be done by administering the choline as a component of a nutrient mixture that contains all three of the PC precursors which limit PC’s rate of synthesis – uridine as its monophosphate; DHA or EPA; and choline (24). This mixture has been shown to diminish Abeta formation in experimental animals (32), and could have the added benefit of partially restoring the deficient brain synapses.
1. Terry R, Masliah E, Salmon D, Butters N, DeTeresa R, Hill R, et al (1991) Physical basis of cognitive alterations in Alzheimer's disease: synapse loss is the major correlate of cognitive impairment. Ann Neurol 30, 572-580.
2. DeWaal J, Stam CJ, Lansbergen MM, Wieggers RL, Kamphuis PJGH, Scheltens P, Maestu F, Van Straaten ECU (2014) The effect of Souvenaid on functional brain network organization in patients with mild Alzheimer’s disease: A randomized control study. PloS One 9 (1) e86558.
3. Davies P, Maloney AJF (1976) Selective loss of central cholinergic neurons in Alzheimer’s disease. The Lancet 2(8000), 1403.
4. Bowen DM, Smith CB, White P, Davison AN (1976) Neurotransmitter-related enzymes and indices of hypoxia in senile dementia and other abiotrophies. Brain 99, 459-496.
5. Coyle JT, Price DL, DeLong MR (1983) Alzheimer’s disease: a disorder of central cholinergic innervation. Science 11(219), 1184-1190.
6. Hardy J, Selkoe DJ (2002) The amyloid hypothesis of Alzheimer’s disease: Progress and problems on the road to therapeutics. Science 297, 353-356.
7. Lardi-Studler B, Fritschy JM (2007) Matching of pre- and post-synaptic specializations during synaptogenesis. Neuroscientist 12, 115-126.