READING TEA LEAVES Okello found that tea might act on the same targets as more sophisticated Alzheimer's treatments.
	NORTH NEWS AND PICTURES

One class of medication that may reduce one's risk of getting Alzheimer's disease is familiar to anyone who takes drugs such as ibuprofen (Advil, for example). Unfortunately, some of these nonsteroidal anti-inflammatory drugs (NSAIDs), which are currently marketed as pain relievers, have recently come under suspicion of causing cardiovascular problems (C&EN, Jan. 3, page 7). The emergence of these possible side effects has brought clinical trials of NSAIDs as a potential Alzheimer's treatment to a halt. Nevertheless, valuable information can still be garnered from studies of the drugs.

NSAIDs may exert their beneficial effect against Alzheimer's by influencing the cleavage of amyloid precursor protein. APP is a large molecule of unknown function that pokes through the membrane of nerve cells. The molecule is metabolized by -, -, and -secretase enzymes. Depending on the secretases involved and the location of the cuts that they make in APP, cleavage of the molecule can produce either harmless fragments or amyloid-₄₂ (A₄₂) peptides that clump together into the insoluble toxic aggregates associated with Alzheimer's.

Researchers are developing a variety of possible explanations for how certain NSAIDs increase the likelihood that APP is cleaved at the more favorable location. For instance, these drugs may reshape the protein site at which cleavage occurs, according to Bradley T. Hyman and colleagues in the Alzheimer's Research Unit at Massachusetts General Hospital, Charlestown [Nat. Med., 10, 1065 (2004)].

First, Hyman's team studied the effect of NSAIDs on the interaction between -secretase and APP in cell culture. The researchers showed that NSAIDs reduce the production of A₄₂ relative to nontoxic amyloid- molecules but don't reduce the overall quantity of amyloid- molecules. "We interpret these data to suggest that NSAIDs alter APP--secretase interactions to change the site of the APP cleavage, rather than to change whether cleavage occurs," the researchers explain. "We reasoned that an alteration in the configuration of the APP--secretase interaction might lead to the observed outcome."

The researchers next surveyed the geography of -secretase's cleavage site. -Secretase is an enzyme complex that includes a section known as presenilin 1 (PS1), which is thought to be the actual cleavage site. Using an assay based on fluorescence resonance energy transfer, Hyman and his colleagues showed that NSAIDs including ibuprofen and flurbiprofen open up the PS1 site and alter the relative positions of APP and PS1. "This change in the conformation of PS1 shifts the cleavage of APP toward shorter species such as A₃₈," they note. Hyman's model is supported by results with aspirin and naproxen. These two NSAIDs--which have no effect on the relative amount of A₄₂ generation--leave APP-PS1 positioning unchanged.

LIKE NSAIDS, some statin drugs--which are widely used as cholesterol-lowering agents--appear to show efficacy against Alzheimer's disease. Also like NSAIDs, statins may achieve this effect, at least in part, by influencing APP cleavage.

The connection between statins and APP may lie along a pathway involving Rho, one of a family of molecules known as small guanosine triphosphate-binding proteins. These molecules, which are also referred to as small G proteins, controls many cellular functions, ranging from gene expression to cytoskeletal reorganization.

Samuel E. Gandy, a neurologist at Thomas Jefferson University, Philadelphia, recently explored the effect of statins on the Rho pathway and APP processing in cultured neural cells [PLoS Medicine, 2, e18 (2005)]. In particular, Gandy and his collaborators were interested in ROCK1, a protein that Rho activates to carry out its tasks. Once activated, ROCK1 phosphorylates other proteins. This phosphorylation is mediated by molecules called isoprenoids.

The researchers first confirmed that ROCK1 activation promotes "bad cleavage" of APP, thereby boosting A₄₂ production. They then determined that the statin drugs atorvastatin (Lipitor) and simvastatin (Zocor) inhibit ROCK1 activity and tip the balance in favor of "good cleavage."

The researchers believe that ROCK1 may exert its effect on APP cleavage via a-secretase. Gandy thinks that the interaction may be indirect, with ROCK1 phosphorylating an "accessory molecule" that then interacts with -secretase. Whatever the exact mechanism, the results "reveal an unsuspected pathway linking statins and amyloid metabolism," Gandy says. "This may help unravel statin action in Alzheimer's as well as point the way toward novel antiamyloid drugs," which could be designed to target ROCK1 directly.

Neuroscientists at the Alzheimer Research Laboratory at Case Western Reserve University proffer an additional explanation for the beneficial effects of statins. Gary Landreth and Andrew Cordle note that amyloid- deposition evokes an inflammatory response in the brain. In particular, amyloid- plaques prompt immune cells known as microglia and white blood cells called monocytes to release inflammation-promoting molecules such as interleukin-1, tumor necrosis factor , and reactive oxygen and nitrogen species. Exposure to these compounds contributes to the death of neurons.

When Landreth and Cordle dosed cultured microglial cells and monocytes with statins, they found that the drugs "robustly inhibit the amyloid--stimulated expression" of the proinflammatory compounds. Adding cholesterol to the culture can't reverse this suppression. The researchers interpret this evidence to mean that statins' "anti-inflammatory actions are distinct from their cholesterol-lowering actions" [J. Neurosci., 25, 299 (2005)].

However, both statin effects can be traced to the drugs' inhibition of 3-hydroxy-3-methylglutaryl coenzyme A reductase. This rate-limiting enzyme controls an early step in the biosynthesis of cholesterol. Inhibiting the enzyme not only reduces production of cholesterol but also reduces production of a whole pile of intermediates that precede cholesterol.

Those intermediates include the isoprenoids, which can branch off into another pathway unrelated to cholesterol production. In this alternative pathway, isoprenoids activate members of the Rho family of small G proteins that are involved in inflammatory signaling. By reducing isoprenoid production, statins limit activation of this inflammatory pathway.

"The Gandy paper suggests that statins act in neurons through mechanisms similar to those described by us in microglia," Landreth explains. Furthermore, "statins and NSAIDs share a common general mechanism to alter Alzheimer's disease pathogenesis in which APP processing is downregulated in neurons and inflammatory responses are suppressed in microglia."

In their work, Landreth and Cordle focused on the cholesterol-independent activities of statins. But that doesn't mean the drugs' effects on cholesterol are unimportant in the war against Alzheimer's. In fact, until the anti-inflammation connection was made, the reduced risk of Alzheimer's among statin users generally had been attributed solely to cholesterol reduction.

SO HOW DOES cholesterol fit into the equation? Several scientists think that the amount of cholesterol in a neural cell's membrane influences the activity of the secretases that snip APP.

Researchers are trying to get a clearer picture of these interactions by studying another class of drugs that affect cholesterol. Known as ACAT inhibitors, the drugs block acyl-coenzymeA:cholesterol acyltransferase (ACAT), an enzyme that regulates cholesterol distribution. Some of these inhibitors, such as avasimibe, are in clinical trials to treat cardiovascular disease. In addition, avasimibe and another ACAT inhibitor--CP-113,818, which is a fatty acid anilide derivative--are being studied in the lab as a therapy for Alzheimer's.

ACAT normally converts cholesterol located in cell membranes into a form that is stored in intracellular droplets for later use. When ACAT's active site is blocked, less cholesterol is stored.

In 2001, Dora M. Kovacs, a neurologist at Harvard Medical School, and colleagues showed that ACAT also regulates the formation of amyloid- in cell culture. In more recent work, she showed that inhibiting ACAT with CP-113,818 virtually eliminates amyloid pathology in a mouse model of Alzheimer's disease [Neuron, 44, 227 (2004)].

Furthermore, "we found that this way of reducing cholesterol levels in the brains of living animals improved learning," Kovacs adds. "As far as we know, this is the first study of cholesterol metabolism's impact on amyloid levels that included cognitive testing."

Analysis of brain tissue from the mice in the trial leads Kovacs to believe that the ACAT inhibitor lowers production of amyloid- rather than reducing its deposition. Kovacs and her team are now studying the effect of avasimibe on amyloid-.

The apparent connection between cholesterol and Alzheimer's disease is strengthened by findings related to apolipoprotein E (apoE), a molecule that helps transport and metabolize cholesterol and triglycerides.

The genes that code for apoE influence the risk of getting Alzheimer's disease. Most people carry the 3 version of the apoE gene. Those lucky enough to possess the 2 allele are much less likely than the 3 carriers to get the disease. Those with the 4 allele, which facilitates aggregation and deposition of amyloid- in the brain, are more likely than the average person to develop Alzheimer's.

Inder M. Verma, a geneticist at Salk Institute, La Jolla, Calif., and colleagues reasoned that inserting copies of the 2 allele into an organism's DNA could boost production of the protective form of apoE. They tested their hypothesis in a mouse model of Alzheimer's.

Verma's team injected viruses carrying apoE alleles into the hippocampus of mice. Once the new genes were taken up by the animals' DNA, the mice began producing the apoE version programmed by the particular allele they had received. The researchers found that expression of the 4 allele promoted fibril formation and deposition of amyloid-. Expression of the 2 allele, on the other hand, "resulted in a rather robust reduction in hippocampal amyloid- burden," they reported [Proc. Natl. Acad. Sci. USA, 102, 1211 (2005)]. The researchers suggest that a similar gene delivery technique might be suitable for preventing or treating Alzheimer's disease in humans.

Cholesterol and inflammation, and the drugs to control them, aren't the only possible answers to Alzheimer's disease. Antidepressants are also being eyed as potential treatments. For instance, Ottavio Arancio and Michael Shelanski, pathologists at Columbia University, and colleagues have found that rolipram can halt and even reverse deterioration of learning and memory in a mouse model of Alzheimer's [J. Clin. Invest., 114, 1624 (2004)].

Rolipram inhibits phosphodiesterase 4 (PDE4), an enzyme that breaks down cyclic adenosine monophosphate (cAMP). cAMP participates in a series of reactions known as the cAMP/PKA/CREB pathway. This pathway is thought to be involved in learning and memory functions, possibly through the formation and maintenance of synapses and growth of neurons. The Columbia researchers found that rolipram, by boosting cAMP levels, counteracts the tendency of A₄₂ to suppress this pathway.

Although rolipram has a half-life of just three hours and is rapidly cleared from the body, it has an enduring effect on cognitive function. Mice treated for just three weeks still show memory benefits at least two months later. Arancio and his colleagues, who plan to explore just how long the treatment is effective, note that the drug doesn't alter production or deposition of amyloid- peptide. Instead, they suspect that rolipram's activation of the cAMP/PKA/CREB pathway alters expression of the genes responsible for the architecture of synapses and dendrites.

The Columbia researchers believe their results are the first to show that "inhibition of phosphodiesterase activity might effectively counteract learning and memory defects in Alzheimer's disease. ... Rolipram and other PDE4 inhibitors represent a new approach to treatment that appears to make the synapse more robust and resistant to the effects of amyloid-." Rolipram has another appealing characteristic: It works even better in older mice than in young ones. As the researchers point out, "this widens the possible therapeutic window of this class of compounds, not limiting it to the initial phases of the disease."

ON ANOTHER FRONT, researchers are probing one of the mysteries of Alzheimer's disease--that the amount of insoluble amyloid- plaque deposited in the brain of patients may not correlate with the severity of their symptoms. Some researchers believe that that's because amyloid- fibrils, which are the main protein component of plaques, are not the primary toxic species. Instead, a precursor to the fibrils, such as amyloid- oligomers, may be the main problem for neurons.

Furthermore, according to researchers at the National Institutes of Health, not all fibrils are created equal: Fibrils come in at least two different forms, with different morphologies and different molecular structures, and these different forms may not be equally hazardous.

Chemical physicist Robert Tycko, neuroscientist Mark P. Mattson, and their NIH colleagues first set out to explain how the two different forms of amyloid fibrils come about. They found that the protofilaments that are the basic building blocks of amyloid- fibrils themselves possess two different three-dimensional structures. The in vitro study showed that subtle variations in the growth conditions under which amyloid- is allowed to aggregate determine the structure of the protofilaments. In turn, the structure of the protofilaments dictates whether a fibril adopts a configuration in which the protofilaments are parallel or one in which they are twisted. The researchers also found that these two fibril morphologies exhibit significantly different toxicities in neuronal cell cultures. The researchers think that the difference stems from the nature of the amino acids exposed on the fibril surfaces [Science, 307, 262 (2005)].

Many potential Alzheimer's treatments--and the insights they provide about the disease--are based on compounds that are synthesized in a lab. But such humble, low-tech provisions as tea and curry are also contributing to the body of knowledge about Alzheimer's.

Indeed, tea may be able to act on the same targets as more sophisticated Alzheimer's treatments, according to Edward J. Okello, a visiting lecturer at the University of Newcastle upon Tyne, in England, and colleagues [Phytother. Res., 18, 624 (2004)].

"The primary target of licensed drugs for the treatment of Alzheimer's disease is the inhibition of the enzyme acetylcholinesterase, although preventing -amyloidosis is a prime target for drugs in development," notes Okello, a tea drinker who conducts research at the university's Medicinal Plant Research Centre. Acetylcholinesterase breaks down acetylcholine, a neurotransmitter involved in memory and cognition that is in short supply in the brains of people with Alzheimer's disease.

Okello's team found that infusions of green or black tea inhibit acetylcholinesterase's activity in vitro. Tea also inhibits butyrylcholinesterase, another enzyme that can break down acetylcholine. And that's not all: Green tea inhibits the activity of -secretase, one of the enzymes that chops APP to release amyloid- peptides. Coffee is a less potent inhibitor of acetylcholinesterase and is inactive against butyrylcholinesterase and -secretase. The British researchers haven't yet isolated the active ingredients in tea, though they have already ruled out caffeine, theophylline, and theobromine. They hope to raise funds for clinical trials to further scrutinize tea's impact.

Tea is not the only food product that shows promise. A number of studies have indicated that curcumin, the yellow pigment in curry spice, may fight Alzheimer's disease. One study that explored how curcumin works comes from a team led by Gregory M. Cole, a professor in medicine and neurology at the University of California, Los Angeles [J. Biol. Chem., 280, 5892 (2005)].

The researchers note that curcumin's structure resembles that of Congo red, a dye known to bind to amyloid- oligomers and fibrils in vitro. Congo red, unfortunately, is both toxic and unable to cross the blood-brain barrier to reach amyloid- aggregates.

FIBRIL DISSIPATORS Some compounds that inhibit aggregation of amyloid- feature two charged or polar groups separated by a hydrophobic bridge.

Unlike the negatively charged Congo red, curcumin's hydrophobic nature might allow it to enter the brain, the researchers reasoned. In addition, curcumin has a long and benign history. It has been used in the form of turmeric as an antioxidant food preservative and as an anti-inflammatory turmeric extract in traditional Indian medicine. Cole thinks that curcumin may combat both the oxidative damage and inflammation associated with Alzheimer's disease.

Through in vitro tests, Cole's team has shown that curcumin inhibits the formation of amyloid- oligomers and fibrils. In addition, the researchers have found that curcumin injected into the bloodstream or fed to a mouse crosses over into the animal's brain and binds to amyloid-.

Better yet, adding curcumin to the diet of an aged mouse appears to break up amyloid- plaques in the animal's brain. "Our in vivo observations suggest that curcumin may be beneficial even after the disease has developed," Cole's team explains. That capability would be invaluable because the outward symptoms of Alzheimer's disease don't become apparent until decades after amyloid begins to accumulate in the brain. UCLA's Alzheimer's Disease Research Center has begun human clinical trials of curcumin in patients with mild to moderate cases.

With research proceeding on so many fronts--and with tangible results accruing--treatments for Alzheimer's are likely to emerge from the lab as well as the larder.