Pharmacotherapy in the treatment of Alzheimer's disease: an update

Abstract

This review summarises the pharmacological properties of the main classes of drugs in current use for the symptomatic treatment of Alzheimer's disease. These may be divided into two major groups: those enhancing cholinergic function which has been shown to be defective in the disease, and those which either directly or indirectly reduce free radical/inflammatory processes in the brain. To date, none of the drugs available has been shown to reverse the pathological changes associated with the disease. However, a number of drugs are in development which are designed to block the neurotoxic action of amyloid beta peptide and thereby reverse the underlying pathological processes. These include the gamma secretase inhibitors and vaccines against amyloid beta peptide. The limitations of these novel approaches are discussed.

For the past 20 years, an increased understanding of the pathology of Alzheimer's disease (AD) has led to the development of numerous drugs for the treatment of the disorder. At the present time, there are at least 60 drugs estimated to be in development for the symptomatic treatment of AD, some of which may ultimately be expected to affect the development of the disease. The drugs in current use can be broadly divided into those that are designed to enhance cholinergic function, those that reduce the synthesis of free radicals, the anti-inflammatory agents, the oestrogens, and a miscellaneous group of natural products which include the Ginkgo biloba alkaloids. In addition, some drugs are in development which are aimed at counteracting the possible causes of neuronal cell loss by blocking the neurotoxic effects of amyloid beta peptide (Ab). These include the inhibitors of gamma secretase and vaccines against Ab. Some of these drugs will now be considered.

DRUGS ENHANCING CHOLINERGIC FUNCTION

The cholinergic hypothesis of AD is based on the loss of histochemical markers of forebrain cholinergic neurons that correlates with diminished cognitive function and with the degree of accumulation of neuritic plaques and neurofibrillary tangles. Assuming that AD bore some resemblance to Parkinson's disease, in which dopaminergic agonists correct the endogenous deficiency of striatal dopamine, it was speculated that directly and indirectly acting cholinergic agonists should correct the symptoms of the disorder. In the past decade drug development has therefore largely focused on centrally acting anticholinesterases and, to a lesser extent, muscarinic agonists and acetylcholine releasing agents. Other approaches have included the administration of high doses of acetylcholine precursors (such as lecithin and choline), which have not been shown to be therapeutically effective, and more recently galanin receptor antagonists.

Because of the progressive neuronal loss that occurs in AD, drugs that enhance the endogenous cholinergic system are inevitably limited in their duration of action. However, at postmortem the M1 and M4 type of cholinergic receptors appear to remain intact in patients with AD, which has strengthened an interest in drugs which have direct cholinomimetic effects (1).

Anticholinesterases

Tacrine, donepezil, rivastigmine and galantamine are cholinesterase inhibitors which preserve endogenous acetylcholine following its synthesis. The inhibition of the cholinesterase may be either reversible, irreversible or pseudoirreversible. In addition, the inhibitor may be either competitive or non-competitive for true (acetyl) cholinesterase, pseudo (butyryl) cholinesterase or for both types. Some anticholinesterases also have a weak affinity for the nicotinic cholinergic receptors. These drugs also differ in their pharmacokinetic properties (for example, protein binding, elimination half-life) and in their interactions with other drugs.

Tacrine is a non-competitive, irreversible inhibitor of both acetyl and butyryl cholinesterase, with a greater potency for the latter enzyme. Based on the outcome of placebo controlled, double-blind studies, tacrine was the first anticholinesterase to be licenced for the symptomatic treatment of AD in the United States (2, 3). The main disadvantage of tacrine lies in its hepatotoxicity (approximately 50% of patients were found to develop elevated liver transaminases, which reversed on discontinuation of the drug). Because of such side effects and limited efficacy, tacrine is no longer widely prescribed.

Donepezil is primarily a reversible inhibitor of acetylcholinesterase with a long elimination half-life. It lacks the hepatotoxicity of tacrine but frequently causes nausea, vomiting and diarrhoea (4). These side effects, together with occasional bradycardia, syncope and changes in the sleep architecture, are directly associated with a central and peripheral enhancement of cholinergic function (5). At the present time, donepezil is the most widely prescribed anticholinesterase in the United States and Europe.

Rivastigmine is a pseudoirreversible inhibitor of both acetyl and butyryl cholinesterases. Although the drug initially blocks the enzymes, it is metabolized by them, so that its half-life is relatively short (6). The top dose is often necessary to achieve therapeutic efficacy, at which dose the central and peripheral cholinergic side effects become apparent.

Galantamine, unlike the other anticholinesterases in clinical use, is derived from the alkaloids from the daffodil and snow drop family. It is a reversible, competitive inhibitor of acetylcholinesterase with some inhibitory action on butyryl cholinesterase. It is also an agonist at nicotinic receptor sites. Although a clinically effective drug, galantamine frequently causes gastrointestinal side effects (7).

Other anticholinesterases in development include metrifonate. This is an irreversible organophosphorus inhibitor of acetylcholinesterase and is a pro-drug for dichlorvos (8). The development of metrifonate has been delayed because some patients developed muscle weakness, and a delayed neurotoxicity has been described for compounds that are chemically related to the drug (9).

To complete the list of anticholinesterases, extended release physostigmine has been shown to have some therapeutic efficacy, but has been restricted in its development because of the high frequency of nausea and vomiting (10).

In addition to their ability to increase the endogenous concentrations of acetylcholine, anticholinesterases have also been found in in vitro studies to increase the synthesis of non-amyloidogenic amyloid precursor protein (APP) and to decrease the neurotoxicity of Ab (11). Regarding the effect of some anticholinesterases on nicotinic receptors, there is also evidence that neurodegeneration is delayed, thereby suggesting that such drugs may be neuroprotective. This view has been supported by epidemiological studies in which the incidence of Parkinson's disease has been shown to be lower than expected in cigarette smokers (12).

Muscarinic receptor agonists

The first generation of cholinomimetics, such as arecoline, bethanecol and pilocarpine, were not designed for the treatment of AD and the results of the early clinical trials were consistently disappointing. In addition to their poor bioavailability and short duration of action, any therapeutic benefits were limited by their cholinergic side effects. The second generation of muscarinic agonists were therefore developed to specifically treat AD. These drugs, exemplified by milameline and xanomeline, have improved pharmacokinetic profiles relative to the first generation drugs. In controlled clinical trials, xanomeline has shown moderate clinical efficacy but, despite in vitro data showing that it was selective for M1 and M3 receptors, it still caused mild to moderate parasympathomimetic side effects (13). Milameline has equal affinity in vitro for all five muscarinic receptor subtypes; peripheral cholinergic side effects were clearly evident in the initial clinical trials (14). Neither of these cholinomimetics has been marketed at the present time.

Despite the theoretical interest in specific nicotinic agonists for the treatment of AD, to date none has reached clinical development.

Glutamate receptor antagonists

Glutamate is a major excitatory neurotransmitter in the brain, estimated to be involved in the regulation of 70% of excitatory synapses. Of the two major types of glutamate receptor - the ionotropic N-methyl-D-aspartate (NMDA) and the metabotropic once - the former plays a crucial role in neuroplasticity and memory formation. It is postulated that excessive activation of the NMDA receptors by glutamate plays an important role in the neurodegenerative changes found in AD (15); Ab has been shown to increase the release of glutamate (16). Several experimental studies have shown that NMDA antagonists prevent glutamate induced neurotoxicity (17).

Of the various NMDA antagonists which have undergone clinical investigation, memantine has been shown to be the least toxic in vivo and also to have neuroprotective properties. Memantine has been the subject of a double-blind, placebo controlled trial and proven to be superior to placebo in improving cognitive dysfunction in patients with mild Alzheimer's disease (18). It should be added that this drug has been available for clinical use in Germany for the treatment of AD for the past 10 years. Thus, if the placebo controlled studies are replicated, it is likely to provide a useful treatment for the mild to moderate form of the disease.

Anti-inflammatory drugs

Inflammatory processes are well known to be associated with AD. Elevation in circulating pro-inflammatory cytokines, acute phase proteins and complement and the presence of activated microglia have been described in patients with AD (19-22). It has also been shown that the complement cascade can be activated by Ab and result in neurotoxic changes (23, 24). The seminal studies of McGeer and coworkers laid the basis to the preventative strategy for the treatment of AD (25). These investigators showed that the prolonged use of non-steroidal anti-inflammatory drugs for the treatment of arthritis and related conditions was associated with a significant decrease in the incidence of AD. Studies of siblings who had a differential exposure to anti-inflammatory drugs also showed that the incidence of AD was significantly reduced in those to whom such drugs were administered (26, 27).

Despite the clinical evidence implicating the involvement of inflammatory processes in the pathology of AD, the mechanisms behind the accumulation of inflammatory mediators are complex. Nevertheless, it would appear that cyclooxygenase 2 (COX2) plays a crucial role. It is known that COX2 activity is elevated in the brain of patients with AD (28) and that there is an increased expression of COX2 mRNA in the frontal cortex of such patients. Furthermore, the severity of the symptoms correlates with both the COX2 activity and the increased expression of Ab (28).

A number of anti-inflammatory drugs have now been tested for their therapeutic efficacy in AD. For example, the steroid prednisolone, which is lipophilic, has been administered to patients with AD for up to a year, but the results were disappointing (29). The potent non-steroidal anti-inflammatories diclofenac and indomethacin have also been tested, but shown to have minimal benefit with a high frequency of side effects (30). Perhaps these results are not surprising, since the inhibition of COX is expected to have little beneficial effect on the symptoms of AD once neuronal death has occurred, as seems likely in the clinical studies in which the patients were in the advanced stage of AD. Another problem arising in the interpretation of the data concerns the effects of the selective COX2 inhibitors such as celecoxib and rofecoxib which, in in vitro studies, have been shown to enhance the formation of the highly neurotoxic form of Ab, Ab42 (31). By contrast, the non-selective COX inhibitors ibruprofen and suldinac were shown to reduce Ab to its less neurotoxic form Ab38. With regard to the mechanism of action of the non-steroidal drugs, it has been speculated that the beneficial effects might be linked to a reduction in the activity of gamma secretase, the enzyme assumed to be responsible for the cleavage of APP to its neurotoxic product. In addition, it is known that these drugs also act as free radical scavengers, which is unconnected with their inhibitory actions on COX.

The most positive result for the action of an anti-inflammatory has been obtained from the studies of propentofyline. This drug inhibits the action of microglia which act as macrophages within the brain and release inflammatory cytokines. However, while there was evidence of some therapeutic benefit, the results were modest and insufficient for the drug to be marketed either in Europe or North America (32, 33). Thus the jury is still out regarding the potential therapeutic efficacy of the non-steroidal drugs, at least in the treatment of patients with established AD.

Antioxidants

Free radicals have been considered to play an important role in initiating neuronal death. Neurotoxic processes arising from an overactivity of the glutamatergic system, from ischaemia, and from the direct action of Ab lead to increased oxidative stress and free radical synthesis (15). It would therefore be anticipated that free radical scavengers and antioxidants could delay or prevent the progression of neuronal degeneration due to such causes (34). Of the compounds tested in a clinical situation, vitamin E has been the most extensively studied. The results of a 2-year placebo controlled trial (35) have shown that vitamin E, particularly when combined with the monoamine oxidase B inhibitor and free radical scavenger selegiline, had a significant beneficial effect, but the high doses of vitamin E which were necessary are also known to cause disorders of blood coagulation, so it seems unlikely that this will become a treatment of choice.

Chelating agents

It is known that metals such as zinc and copper become more concentrated in the brain with increasing age and that these metals can induce Ab aggregation, thereby enhancing the deposition of senile plaques. In addition, the presence of these metals with Ab initiates the formation of hydrogen peroxide, which causes oxidative damage to neurons. By using metal chelating agents such as cliquinol, it has been possible to reduce the zinc and copper concentrations in the brains of patients with AD, leading to a small, but significant improvement in cognitive function (36). The use of chelating agents may therefore be of some therapeutic benefit in the future.

Oestrogens

The notion of a potential therapeutic value of the oestrogens came from their protective effect against AD in post-menopausal women (37, 38). Experimental studies have shown that oestrogens protect hippocampal dendrites from damage and also augment the activity of choline acetyl transferase. Additional properties of oestrogens which may contribute to their neuroprotective effects include antioxidant properties and a facilitation of the processing of APP along a non-amyloidogenic pathway (39, 40).

Despite the promising experimental and epidemiological studies, the clinical trials of the oestrogens in AD have been disappointing. There is some evidence that women who had taken oestrogens for the treatment of post-menopausal symptoms had a better response in terms of an improvement in cognitive symptoms than those who had not taken oestrogens. It seems possible that selective oestrogen receptor modulators (SERMs), which are drugs that act as selective agonists for central oestrogen receptors, may eventually replace the non-selective steroidal oestrogens which have been used to date.

Secretase inhibitors

Amyloid deposition is now regarded as one of the earliest changes that initiate AD. It would appear that, regardless of the position of the mutation, whether this is in the APP gene, presenilin 1 or 2, the final outcome is the increase in neurotoxic Ab42 in the brain and plasma. A similar finding has been observed with the increased frequency of the apolipoprotein E4 allele (41). This has led to the hypothesis that the aggregated form of Ab is primarily responsible for the symptoms of AD and therefore it might be possible to develop appropriate drugs to prevent the neurotoxic damage by blocking the synthesis of Ab.

It is known that beta and gamma secretases are responsible for cleaving APP to Ab, so that by inhibiting these enzymes it might be possible to block the progression of the disease. Alternatively, enhancing the activity of alpha secretase, leading to the formation of a non-amyloidogenic end product, might also be beneficial. Another approach could involve the increase in the breakdown of Ab once it has been formed. All these possibilities are actively under consideration but so far no drugs have emerged.

Regarding the possibility of developing secretase inhibitors, it appears that the pancreas also contains beta secretase and it is presently unclear what the consequences could be if the pancreatic enzyme was inhibited in addition to the brain enzyme (42). With regard to gamma secretase, it is known to be closely associated with the presenilins, the enzymes that are critically involved in a number of metabolic pathways in addition to the formation of Ab (43). Thus, there may be unexpected toxicity problems which arise by inhibiting gamma secretase.

Despite some exciting experimental findings in mice that have been genetically programmed to develop the human form of Ab, so far there is no evidence to indicate that secretase inhibiting drugs have been developed that offer some prospects for treating AD (44). However, following the recent identification of a novel membrane bound aspartic protease (BAC1) as beta secretase, the possibility still remains that drugs targeting the beta - and gamma - secretases could be developed in the near future (45).

There is evidence that cholesterol contributes to AD by enhancing Ab synthesis (46). This provides a theoretical basis for the use of statins to lower the blood cholesterol concentration. There is also recent evidence that the statins have unexpected anti-inflammatory properties by reducing the adhesion and activation of leucocytes, which may contribute to the moderate improvement in the cognition scores which have been observed in a placebo controlled trial (47).

VACCINES

Transgenic mice that overexpress Ab have been used to determine whether vaccines could be produced to reduce the concentration of the peptide in patients with AD (48). Experimental studies have shown that Ab peptide immunization reduces the cognitive impairments and the formation of plaques in rodent models of AD (44). This finding led to the development of vaccines for human use. While the phase 1 trials in the UK suggested that the vaccine was safe (49), more extensive studies in Europe led to the termination of the clinical trials because 5% of the patients developed meningoencephalitis (50). Further studies are presently underway to induce an immune response against Ab without initiating T-cell activation which underlies the inflammatory process in the brain. The possible mechanism whereby cerebral haemorrhage occurs following anti-Ab immunotherapy has been suggested recently by Pfeifer and coworkers (51).

CONCLUSIONS

Major progress has been made in the past decade to develop drugs to treat the symptoms of AD. To date, important advances have been made regarding the reversal of the disease process, in particular with respect to preventing the accumulation of Ab and preventing the central inflammatory response which appears to initiate the neurotoxic changes. Undoubtedly the following decade will see the development of vaccines and other strategies that will alter the course of the disease. Thus we can expect the therapeutic pessimism of the past to be replaced by therapeutic optimism in the future.