Mairin B. Brennan

C&EN Washington

"There is a wonderful kind of excitement in modern neuroscience, a romantic, moon-walk sense of exploration and setting out for new frontiers. The science is elegant, the scientists dismayingly young, and the pace of discovery absolutely staggering."—Kay R. Jamison, "An Unquiet Mind," New York: Alfred A. Knopf Inc., 1995

In the sweep of history, the 1990s will surely be regarded as the decade that opened an unprecedented window on the brain. During these 10 years, research in neuroscience pushed forward as never before, delivering a flood of advances. In their wake, a greater understanding of the mechanisms underlying neurodegenerative diseases emerged, insights into the biological underpinnings of addiction and behavior were gained, the molecular roots of chronic pain were postulated, the molecular basis of learning and memory began to unfold, and hints surfaced that spinal cord injury might be reversible.

At the same time, advances in imaging techniques, including magnetic resonance imaging (MRI) and positron emission tomography (PET), enabled researchers to capture snapshots of the brain responding to a task or reacting to intake of a drug of abuse. And animal models of a variety of neurological disorders proliferated. The "Decade of the Brain," proclaimed in 1990 by then-President George Bush, lived up to its title.

The spate of information that gushed through the 1990s laid the foundation for extraordinary contributions to the treatment of brain disease in the 21st century. But it also underscored the complexity of the brain and the challenges that lie ahead in mapping its underlying biology. "We're learning that it's very naive to think of schizophrenia as a 'dopamine' disease or depression as a 'serotonin' disorder, because all neurotransmitters interact," notes Stephen L. Dewey, a neuroscientist in the chemistry department at Brookhaven National Laboratory, Upton, N.Y. The dynamics of the brain are such that "if one system changes, there are going to be compensatory changes in others," he says.

Clues to how complex the circuitry of the human brain might be can be gleaned from the neurobiology of Caenorhabditis elegans, a nematode whose genome was sequenced last year [Science, 282, 2011 (1998)]. In this tiny worm, genes related to the nervous system encode, among other proteins, at least 80 potassium ion channels, 90 ion channels regulated by neurotransmitters, 50 peptide-activated receptors, and up to 1,000 "orphan" receptors presumed to respond to chemical stimuli.C. elegans' genome contains roughly 19,000 genes.

Sidebar: Memories are made like this

The Human Genome Project, now racing toward the finish line, is expected to deliver an estimated 140,000 genes, nearly half of which are believed to be expressed in the brain. Already, genes with sequences for ion channels and receptors and other proteins still unknown are on the table, adding layers of complexity to a scenario that even now is beset by a deficiency in understanding.

But that is what makes neuroscience an endless frontier. That is why a community of sciences must together solve this giant scientific jigsaw puzzle in the new millennium, piecing it together bit by elusive bit. Molecular pathways must be defined, interconnections among nerve cell circuits mapped, and discoveries converted into cures. In these endeavors, chemistry is poised to play an increasing role.

"For many years, molecular biology was moving forward at an incredible pace," notes Solomon H. Snyder, director of the department of neuroscience at Johns Hopkins University School of Medicine. "Now chemists are having their own revolution." It's a practical revolution, he says, a revolution not so much in theory but one that offers sophisticated strategies for addressing the molecular basis of psychiatric and neurological diseases. Stephen J. Lippard, chairman of the chemistry department at Massachusetts Institute of Technology, agrees, suggesting that "structural and mechanistic work is where chemists can make a big contribution to neuroscience, when they turn their attention to it and the tools are available."

It's not that chemists haven't always contributed to neuroscience—behind every drug for treating a brain disorder are the synthetic organic chemists who made it possible. Indeed, Snyder points out, it was organic chemists who launched the modern pharmaceutical industry in the early 1900s by transferring their expertise in modifying chemical ring structures from dyes to drugs, an evolution that, in 1952, led to the first drug (chlorpromazine) for treating schizophrenia. PET, too, is indebted to chemistry. The technique that can image a tissue's biochemistry and physiology in vivo owes much of its capabilities to the synthetic organic chemists who incorporate short-lived tracer isotopes into the biological molecules targeted for imaging.

Now a new era is in the wings. Soon, probes for the entire human genome will be available on DNA chips, providing scientists with the opportunity to track genes turning on and off in response to stimuli. "That will be a major contribution to neuroscience," Lippard says. "The question is, 'How are you going to interpret the data?' We don't know what the vast majority of the proteins encoded in human genomes are doing. It will be a lot like having the Egyptian hieroglyphics without the Rosetta stone, so chemists have got to provide the Rosetta stone." Chemists must figure out the biochemical pathways controlled by these genes, he explains. "We are going to need armies of chemists to attack this problem."

A few chemists already are focusing their research on the biology of the brain, drawn by the challenges it presents and the pressing need to develop drugs that could arrest or reverse the degenerative process in Alzheimer's and Parkinson's diseases. But the expertise of chemists also is needed in helping to unravel the biochemical pathways that preordain nerve cells to an untimely death in these and other neurodegenerative diseases and in stroke and to pinpoint the biology of mental illnesses.

Despite the advances that came thick and fast in the 1990s, drugs that can stop the progression of Alzheimer's or Parkinson's diseases have not advanced much beyond the drawing board. And other than genetic testing for rare familial forms, ways to detect these diseases early on have not yet been developed. Meanwhile, the incidence of both diseases is demographically programmed to soar: The Baby Boom generation—the bulge of babies born in the aftermath of World War II—will soon be growing old.

The prevalence of Alzheimer's disease doubles every five years after age 65, according to the National Institute on Aging (NIA), and the disease currently afflicts more than 4 million people in the U.S. Parkinson's disease, too, is largely a disorder of the elderly, afflicting an estimated 500,000 people in the U.S., according to the National Institute of Neurological Disorders & Stroke.

Genes underlying Alzheimer's, Parkinson's, and several other neurodegenerative diseases have now been discovered and the proteins they encode identified. A common theme has emerged: Each of the diseases is associated with the overproduction of a protein that either misfolds or polymerizes into aggregates within or around nerve cells.

In Alzheimer's, for example, the culprit is the amyloid -peptide (A), which forms fibrillar deposits in amyloid plaque that invades the brain's seat of memory and cognition before spreading to other areas. In Parkinson's, protein aggregrates called Lewy bodies form in an area of the brain that controls movement, their fibrillar component derived mainly from a protein known as -synuclein. In Huntington's disease, an inherited movement disorder and dementia that afflicts roughly 30,000 people in the U.S., a protein called huntingtin accumulates in fibrillar form within neurons. But the picture is more complicated: Alzheimers' also produces tangles of a protein called tau, and Huntington's packs ubiquitin—a protein that normally tags proteins for destruction—into aggregates of huntingtin.

The precise mechanism of cell death in these and other neurodegenerative diseases is still far from clear. But an emerging theory suggests that protein misfolding is a key player—the diseases, although initiated by different proteins, may proceed along parallel routes to destroy different regions in the brain. "The idea that misfolded proteins could be involved in diseases was pretty far out just a few years ago," says protein researcher Michael H. Hecht, an associate professor of chemistry at Princeton University. "Now many people are working in this area." Molecules that interfere with protein aggregation might be one way to halt the progression of a neurodegenerative disorder or somehow reverse it, he suggests.

PET images of glucose uptake—the yardstick for measuring brain activity via glucose metabolism—in the brain of a healthy older person (left) and patients with moderate (center) and severe (right) Alzheimer's disease reflect neurodegeneration that accompanies the disease. White and red colors indicate the highest uptake of glucose; dark blue and darker colors, the lowest. [National Institute on Aging]

"What's interesting to me is trying to bring together a bunch of diseases that had previously been pulled apart by the clinical world," admits organic chemist Peter T. Lansbury, an associate professor of neurology at Brigham & Women's Hospital, Boston. "From a chemist's point of view, we're not concerned with the fact that Alzheimer's disease and Parkinson's disease and Huntington's disease and Lou Gehrig's disease are all different chapters in a neurology textbook and different subspecialties within neurology. Instead we see the biochemical and chemical similarities."

Lansbury's group exploits solid-state nuclear magnetic resonance (NMR) and atomic force microscopy (AFM), in addition to using more traditional techniques, to study the fibrillar proteins associated with Alzheimer's and Parkinson's. The group collaborates with chemistry professors Robert G. Griffin at MIT and Charles M. Lieber at Harvard, respectively, on NMR and AFM approaches. Solid-state NMR is particularly useful in structure studies of the amyloid fibril, Lansbury notes, because the fibril is neither soluble nor crystalline—and thus not amenable to X-ray crystallography. The only way to solve its structure is to look for novel methods, he insists. "We've made progress, but we're not grinding out structures yet. People have been spoiled into believing they can learn about biology from looking at structures determined by crystallography and solution NMR. That works only for a minority of proteins."

Chemists and biophysicists in Lansbury's lab collaborate with molecular biologists in carrying out protein fibrillation experiments in mice. Genetics, chemistry, cell biology, biophysical chemistry, pathology, and clinical medicine "are equal partners that offer different ways to converge on the truth," Lansbury argues. "The medical community is interested in solving problems. If you can show them that you can do it, they will beat a path to your door."

On his wish list for the new millennium are imaging techniques that would detect Alzheimer's and Parkinson's at an early stage—before symptoms of the diseases signal damage already wrought. "There's no information at the molecular level about the pathway. You just know the end product, the pathology," he says. "It blows my mind that people are not jumping to [develop imaging tools]. But it requires a huge interdisciplinary collaboration among physicists, synthetic organic chemists, and physicians."

Other Alzheimer's researchers are looking for ways to circumvent amyloid formation by preventing production of A. Among them is chemist Michael S. Wolfe, an associate professor of neurology at Brigham & Women's Hospital and Harvard Medical School. A is clipped from a large transmembrane protein called amyloid precursor protein (APP) by enzymes dubbed secretases. -Secretase cleaves APP at a site outside the membrane, while -secretase cuts it in a region that traverses the membrane. Molecules that would inhibit either of these secretases could potentially stop A production.

-Secretase "caught my imagination," Wolfe says, because protein cleavage wasn't known to occur within the membrane. Wolfe and his colleagues synthesized small-molecule, substrate-based -secretase inhibitors that led them to discover relationships between -secretase and presenilins, the products of genes linked to familial forms of Alzheimer's. They have postulated that presenilins may, in fact, be -secretases—self-activated enzymes that carry out their work within the membrane [Nature, 398, 513 (1999)].

Sidebar: MRI offers diagram of brain wiring

As Alzheimer's disease progresses, cholinergic neurons waste away. The couple of drugs currently used to treat the disease are designed to improve memory and cognition only in the early stages. In effect, they delay the onset of more severe symptoms by increasing the concentration of acetylcholine—a neurotransmitter crucial to nerve-cell "firing"—by inhibiting cholinesterase, the enzyme that hydrolyzes it. But as neurons die off, the supply of acetylcholine shrinks, and the drugs become increasingly less effective.

In stroke, a clot in the cerebral artery can affect nerve cells highly dependent on blood supplied by the artery (dark purple area) as well as surrounding nerve cells that receive some blood from other arteries (light purple area). [© Carol Donner/Originally published in Scientific American]

Because the drugs delay the progression of Alzheimer's, their usefulness in stalling onset of the disease itself is now being explored. So is the usefulness of antioxidants, since destructive oxidative species have been implicated in promoting nerve-cell death in Alzheimer's. Earlier this year, NIA launched a study that will evaluate the effectiveness of both the anticholinesterase agent donepezil and vitamin E, which has antioxidant properties, in delaying onset of Alzheimer's in an elderly population presumed to be at risk (http://www.memory study.org). A study sponsored by NIA in 1997 showed that vitamin E could slow the progression of Alzheimer's in those already afflicted with it.

The exact relationship between A aggregation and the oxidative process that destroys nerve cells in Alzheimer's disease is still something of a mystery. Among those seeking to find the connection is D. Allan Butterfield, a professor of chemistry at the University of Kentucky, Lexington, and director of the university's Center of Membrane Sciences, who has shown that vitamin E prevents A-induced oxidation and cell death in neurons growing in culture <:url:http://pubs.acs.org/cgi-bin/jcen?crtoec/10/i05/html/tx960130e.html>[Chem. Res. Toxicol., 10, 495 (1997)]. More recently, the group has suggested that a methionine residue in A may trigger the oxidative process [Neurobiol. Aging, 20, 325 (1999)]. The Kentucky team also is exploring the source and course of oxidative damage in both Parkinson's and Huntington's, using rodent models for these diseases. Both Parkinson's, which destroys dopamine-producing neurons, and Huntington's, which attacks neurons that house -aminobutyric acid, are thought to produce free radicals that impair mitochondrial function.

Current findings in Butterfield's lab suggest that oxidative stress precedes the onset of neuronal loss in a rat model of Huntington's, a discovery that potentially could lead to therapies based on antioxidants. In other work, his group is investigating ways to increase the concentration of one of the brain's natural antioxidants, glutathione, by administering a precursor molecule whose metabolite would cross the blood-brain barrier and be converted to glutathione.

At the University of Chicago, assistant professor of chemistry Ka Yee C. Lee is studying the interaction of A with model membranes to figure out how the peptide aggregates. An electrical engineer with a Ph.D. in applied physics, Lee held postdoctoral fellowships in both chemistry and chemical engineering before joining Chicago. "What I have found in going through the different disciplines is that people in different disciplines are trained differently, and so they approach the problem differently," she says. "This is an interesting thing. Because neuroscience is so complicated, I think coming from a different angle . . . would help us understand the whole problem."

Lee believes one can "borrow" knowledge from the physical sciences, such as principles of ordering and self-assembly, and apply it to problems in neuroscience. The chemistry of colloids and complex fluids and the high-resolution imaging used by physicists to study ordered crystal packing can be brought to bear on problems in neurobiology, she says. "I think you will see the hard border between the physical sciences and the biological sciences soften," as people in different disciplines learn from each other.

Devastating as Alzheimer's and Parkinson's diseases are, they don't claim nearly as many victims as stroke. Indeed, stroke is the third leading cause of death in the U.S. and "may be the largest unmet need in drug development," Hopkins' Snyder suggests. Drugs that prevent nerve-cell death in stroke are still not available, although tissue plasminogen activator, a clot buster used in heart attack, can help if administered within a few hours. Some of the major steps in stroke's destructive pathway have been identified: Nerve cells cut off from a blood supply cannot maintain a resting potential. As a result, they release excitatory amino acids, including glutamate, which binds to ion-channel receptors on neighboring neurons, flooding these cells with Ca²⁺. In the metabolic overdrive that ensues, the cells burn out and die, propagating the degenerative process.

"We've learned a lot in the past five to 10 years," Dennis W. Choi, a professor of neurology at Washington University, St. Louis, acknowledges. "We now have identified specific players in the injury process—for example, glutamate and nitric oxide—that allow us to do some rational targeting" for drug development. In fact, a dozen or more drug candidates are at various stages of clinical trials, some intended to prevent glutamate from binding to its receptors and others meant to block voltage-gated ion channels. Choi is cautiously optimistic, while pointing out that past trials "have been a bit of a disappointment," because lead candidates were found to induce psychosis or other severe side effects.

"We would be in a different place if we had a complete map of the injury mechanisms," Choi says. "We need to understand more of the underlying biology. There's been a great deal of information learned, but some of the information has told us just how complicated things are. It's not as if there's a single process going on that we can attack. There are a multiplicity of processes, some of which are influenced in opposite directions by the same manipulations."

Understanding the chemistry of ion channels could provide more precise targets for drugs to treat stroke and perhaps other disorders. The challenge is to define the structure-function relationship of these membrane proteins and the properties that enable them to select the ions they allow to pass through.

In a major step forward, the first crystal structures of ion channels were obtained last year. Biophysicist Roderick MacKinnon and coworkers at Rockefeller University, New York City, characterized a potassium ion channel (C&EN, April 6, 1998, page 12), and chemistry professor Douglas C. Rees and colleagues at California Institute of Technology described a mechanosensitive ion channel [Science, 282, 2220 (1998)]. Although both are bacterial ion channels, they've provided a glimpse into what the architecture of mammalian ion channels might be, observes Dennis A. Dougherty, a professor of chemistry at Caltech. Characterizing them "was a heroic effort in each case," Dougherty says, because, as membrane proteins, ion channels are not good candidates for X-ray crystallography.

"Ion channels are very dynamic structures—the action is in motion and movement, opening and closing, ion flux, ligand binding, and ligand release," Dougherty notes. Understanding how all of this happens at the molecular level would require creative approaches even if crystal structures were available, he says. In collaboration with Caltech biologist Henry A. Lester, Dougherty is expressing ion channels in frog oocytes, systematically incorporating unnatural amino acids at targeted sites. Using photoreactive groups and other designed reporter molecules at these sites, he can follow changes in structure induced by ligand binding or current flow. "Some of the electrophysiological tools we use are foreign to chemists," he says—for example, the patch clamp technique that measures current flow through single cells. "But once you get a handle on them, they are tremendously powerful."

Much of Dougherty's work focuses on the molecular basis for nerve-cell signaling, using the nicotine receptor—an acetylcholine receptor—as a model system. Some acetylcholine receptors are more sensitive to nicotine than others, he notes, and thus may be implicated in nicotine addiction. "There's got to be a chemical basis for [addiction]," he argues. "Some binding interaction or gating interaction may be subtly different from one receptor to another. That's what we're addressing right now. It's a great chemistry problem."

And it promises to be a complex problem, given that the C. elegans' genome has 40 subunits for the acetylcholine receptor. This type of diversity in the human genome could potentially produce a tremendous variety of nicotine receptors, Dougherty notes. Because the human receptor is a pentamer, any five subunits could, in principle, assemble to form it. Learning how to target specific subunits and develop drugs that would home in on only one class of receptor is "a fabulous chemical challenge," he suggests. But it's also scary, he says, because it implies "one could chemically dial in a neurological state."

MIT's Lippard believes neuroscience represents "a spectacular frontier for bioinorganic chemistry," offering opportunities to study the role of copper in prion diseases, for example, or metal-ion transport through membranes, a key event in nerve-cell signaling. His group is developing models for the potassium ion channel's selectivity filter to uncover the mechanism by which it allows the potassium ion to pass through while excluding the smaller sodium ion. On another front, the group is synthesizing sensors for zinc, with the goal of producing an intracellular sensor that can detect free zinc in nerve cells. Such sensors are available for calcium, he notes, and a fluorescent sensor for nitric oxide is under construction.

"There are a lot of transition metals in the brain whose functions aren't known," Lippard says. "The hippocampus, for example, has mossy fibers that contain synaptic vesicles loaded with zinc, and nobody knows for sure what the zinc is doing there."

Space-filling model of mechanosensitive bacterial ion channel whose crystal structure was determined by Rees and colleagues at California Institute of Technology offers cutaway side view (left) and surface view (right). Purple regions are more basic, red more acidic. [Adapted from Science]

Some tools needed for high-resolution structural information have yet to be developed, Lippard points out. For example, the ability to image at the cellular or subcellular level will require better spatial resolution than MRI currently can deliver. To achieve this goal, physicists and bioengineers will need to design more powerful machines and chemists will have to develop better contrast agents.

PET imaging will present its own challenges, especially for synthetic organic chemists. Researchers studying addiction, behavior, and mental illness increasingly are calling for radioligands that can image individual receptor subtypes and pinpoint the neuronal circuit each feeds into. Such information would enable them to zero in on the subtype that modulates a specific pathway and, as a result, to design better therapies for mental illnesses, for example.

"People are trying to tease apart the dopamine system in schizophrenia," Dewey says, because there's disagreement over which of several dopamine receptors is involved in this mental disorder. Dewey, who studies mental illness and addiction, stresses the crucial need in these fields of research for synthetic organic chemists who can devise rapid synthetic routes to ligands labeled with ¹¹C, an isotope with a 20-minute half-life. "It's not going to help if it takes four hours to make the ligand," he points out.

¹¹C-labeled ligands allow researchers to follow an individual patient's response to a drug, he explains. "We can inject the tracer and image the brain. Two hours later, the isotope is 'gone,' and we can pretreat the patient with a targeted drug, inject the tracer, and image again. That way, each patient serves as his or her own control. And that gives us enormous power."

A major challenge in radioligand design is understanding how chemical compounds interact with living systems, Joanna S. Fowler, director of Brookhaven National Lab's PET program, points out. The ligand must cross the blood-brain barrier, and it must withstand all the metabolic and binding processes in the body and be delivered to the brain. Chemists need to know which enzymes might break it down or which proteins might bind to it and prevent it from reaching its target, she says. Another challenge is the requirement for chiral synthesis when only one form of a drug is active in the brain, Dewey notes.

"It's chemists who can label the [PET] probes that allow us to look at the mechanisms," affirms psychiatrist Nora D. Volkow, associate director for life sciences at Brookhaven. However, she says, more sensitive instruments are needed, and that will involve physicists.

Dewey predicts that the next decade of brain research will make "enormous inroads" into understanding the genetic underpinnings and neurochemistry of addiction and behavior. Already some insights are emerging: Earlier this year Volkow postulated that people with low levels of a particular dopamine receptor (the D2 receptor) may be at higher risk for abusing drugs. In previous studies, she had noticed that drug addicts have low levels of this receptor, so she wanted to find out whether one's quota of dopamine receptors might affect how much one "likes" a drug.

PET images by Yu-Shin Ding at Brookhaven National Laboratory of human brain after administration of¹¹ C-d-threo-methylphenidate (top) and ¹¹C-l-threo-methylphenidate (bottom) show the d-enantiomer concentrated in the basal ganglia—the area of the brain that houses the "reward circuit"—where it binds to the dopamine transporter. The l-enantiomer does not bind. Ritalin, a racemic drug used to treat attention deficit hyperactivity disorder, is a mix of the two enantiomers. The images indicate that the d-enantiomer is the active form[Proc. Natl. Acad. Sci. USA, 96, 11073 (1999)]. [Images © Journal of Psychopharmacology]

Volkow measured the levels of the D2 receptor in 23 men who were not drug abusers. The men were given Ritalin (methylphenidate), a mild stimulant used to treat children with attention deficit hyperactivity disorder. Ritalin increases dopamine levels in the brain, an effect associated with the "high" produced by drugs of abuse. The 12 men who had low levels of the receptor described the drug's effects as "pleasant," while those with high levels of the receptor said the effects were "adverse." Men with very high levels of D2 receptors "will say the drug makes them lose control, or become anxious, or suspicious, or even paranoid," Volkow says. "This is telling us that how one responds to a drug is not motivated by the drug itself, but by the biochemical characteristics of one's brain." She cautions, however, that addiction involves more than a head count of dopamine receptors, because the men who had fewer receptors were not abusing drugs. "It's not just genes—it's genes and the environment" that underlie addiction, Volkow says, and perhaps risk taking and aggressive or violent behaviors as well.

"The nervous system is tremendously complex, and therefore it's unlikely there will be simple answers," suggests Clifford Woolf, a professor of anesthesia research at Massachusetts General Hospital and Harvard Medical School. Woolf studies mechanisms of pain and nerve regeneration. Earlier this year, he and his colleagues induced severed nerves to grow across the site of an injury in a rat spinal cord, even though the environment of the central nervous system traditionally forbids regeneration of adult neurons. The role of that environment still holds, Woolf says, so "a large part of the equation becomes how to switch adult nerve cells into an actively growing state. And that involves identifying signal molecules that control whole programs of gene function."

Biotechnology and genomics will help to unravel the nervous system, he concludes, but a whole new mathematics will be needed to deal with changes in populations of genes and complicated sets of neurons. "That's going to be a big challenge," he says, "but I have a sense that we are on the brink of something big."

A new decade in neuroscience research is around the corner, a decade that will usher in a new millennium. The age where science and the mysteries of the mind converge is beginning, and chemists will be center stage.

[Feeding The World][Science & Security]

Chemical & Engineering News
Copyright © 1999 American Chemical Society