If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Analytical Chemistry

A Chemical Map Of The Mind

Targeted radiotracers help drugmakers navigate the neurological landscape by positron emission tomography

by Bethany Halford
September 8, 2008 | A version of this story appeared in Volume 86, Issue 36

Credit: Andrea Varrone/Karolinska Institute

There was a time when making a drug destined for the brain was like firing a missile without any telemetry technology: You aimed in the right direction and then had to wait for reports to trickle in to find out whether or not you'd hit your target.

As the average cost of bringing a drug to market edges past the billion-dollar mark, pharmaceutical companies have come to realize that they can no longer afford to fire blindly. But surveying the terrain behind the skull is no trivial task. Reconnaissance methods that work well for other parts of the body—biopsy and animal models, for example—aren't well suited to assessing the central nervous system (CNS). It's not as if you can routinely scoop out a sample of a patient's brain. And using animals to probe the treatments for diseases particular to the human condition, such as Alzheimer's, schizophrenia, and depression, provides limited intelligence at best.

To navigate the neurological landscape, drugmakers are increasingly turning to positron emission tomography, or PET. This highly sensitive imaging technology lets scientists create a chemical cartograph of the brain by using radiolabeled molecules. Drugmakers no longer have to wonder whether their compounds have hit the target; PET lets them see for themselves.

Without PET, a scientist trying to chart the course of a CNS drug in a living person would administer the compound and then draw blood to track its concentration in the body. This only tells you that the compound is in the body, not that it is, in fact, in the brain.

"You don't know how much of the drug really gets into the human brain," explains Joanna S. Fowler, a senior chemist and PET expert at Brookhaven National Laboratory. "You don't know how fast it gets in. You don't know if inhibition of the enzyme or the receptor persists for a long time or if it's rapidly reversible. You don't have any of that information. You're just guessing. With PET technology you can look at what the drug is doing directly in the human brain. That is an extremely powerful tool."

PET allows scientists to trace a treatment's path by imaging "reporter" molecules—known as radiotracers—that bear a short-lived, positron-emitting radioactive isotope. As these isotopes decay, they set off a radiation-generating series of events that a PET scanner uses to create a picture.

For example, researchers commonly turn to the radiotracer 2-[18F]fluoro-2-deoxy-d-glucose ([18F]FDG), which contains the isotope fluorine-18, to image glucose metabolism. Regions of the brain where glucose is being hungrily consumed will light up, while areas without ramped up glucose metabolism will appear dim. [18F]FDG patterns will look very different in a normal brain compared with a brain with advanced Alzheimer's. This radiotracer is also commonly used in oncology, because certain tumors tend to consume a lot of glucose.

Some examples of lead compounds and their corresponding radioligands.

"Thanks to advancement in radiochemistry, it is possible to label 70 to 80% of all molecules" in the syntheses which are especially time-sensitive because of the isotopes' short half-lives, says Lars Farde, chief scientist for CNS and pain research with AstraZeneca and a professor of psychiatry at Sweden's Karolinska Institute. Carbon-11 and 18F, with half-lives of 20 and 110 minutes, respectively, tend to be the most popular radiolabels.

PET can detect radiotracers in picomolar concentrations, so the amount of radiotracer that patients receive is vanishingly small. "I think the beauty of PET is that it has remarkably high sensitivity in terms of its molecular resolution, and it's able to image targets that are present in very low density," notes Richard Hargreaves, worldwide head of basic neuroscience at Merck & Co.

Other imaging technologies that physicians rely on, such as X-ray, ultrasound, and magnetic resonance imaging, can provide topographical maps of the brain, Hargreaves explains, but they don't give any quantitative information about the biochemistry that's going on. "PET allows you to see the molecular pathology of disease, and it allows you to see the molecular targets for drug therapies because of this supreme resolution that it has," he says.

Although PET has been around for decades, it's only been within the past few years that pharmaceutical companies have made the imaging technology an integral part of the drug development process. "I think that there's been a paradigm shift in the way early drug discovery has been done," Hargreaves says. "PET has had a particularly important role in neuroscience and CNS drug discovery."

Drugmakers can use PET in a couple of different ways. If they've got a lead compound that's amenable to synthesis with a radioactive isotope, then they can follow the molecule, along with any metabolites that might retain the radiolabel, as they travel throughout the body. Not only does this tell you whether the compound makes it to the brain, but it also lets you see what other organs sequester the drug.

The other way pharmaceutical firms take advantage of PET's molecular imaging power is by creating radioligands that bind to specific receptors, enzymes, or other molecular targets in the brain. Researchers can use such radiotracers, for example, in occupancy studies, where scientists see how much drug they have to administer to kick a radiolabel off of a receptor.

That kind of information can be critical when setting up a clinical trial, Hargreaves says. "In the old days when you did your drug discovery, the doses that you tested in your clinical trials just defined the tolerability of the drug. You had no idea whatsoever whether you'd actually delivered the drug to the target in sufficient quantity to have any chance of having a therapeutic benefit," he explains. PET has become popular because it yields data useful for determining likely therapeutic dosages, making for a much more rational drug discovery process.

With PET, Hargreaves says, it's simple to identify the loser molecules quickly because they don't engage the target sufficiently within their safety and tolerability levels. It's equally easy to pick out more promising compounds because they occupy the target to a high degree at low concentrations.

Of the 12,000 or so genes expressed in the brain, scientist think that perhaps 500 to 1,000 may serve as drug targets, according to AstraZeneca's Farde. The challenge for chemists is in finding good radioligands to map them. "I would say there are not more than 25 to 30 really good radioligands that serve the purpose of receptor occupancy measurement," Farde says. "It's almost as difficult to find a good radioligand as it is to find a good drug."

A good CNS radioligand needs to fulfill a specific set of criteria, explains Victor W. Pike, chief of PET radiopharmaceutical sciences at the Molecular Imaging Branch of the Bethesda, Md.-based National Institute of Mental Health. First, it is has to be able to breach the stronghold of the blood-brain barrier and evade any transporters that might pump it out. Then, once the radiotracer gets into the brain, it should bind only to the target protein. If it latches onto anything else, you won't get a clear picture of how effectively your drug candidate displaces the tracer from the target.

Along those lines, Pike notes, a good CNS radioligand needs to have the right lipophilic balance: greasy enough to slip past the blood-brain barrier but not so lipophilic that it sticks to the brain's hydrophobic white matter.

Credit: Chris Rowe
[18 F]-BAY94-9172 targets amyloid-β plaques (highest concentrations in red and yellow) that are characteristic of Alzheimer’s disease.

Kinetics and metabolism are important, too. Because its radioactive signal is steadily fading, the tracer needs to get on and off of the target fairly quickly. That's one reason drugs generally don't make good radiotracers. Finally, the radioligand's metabolic breakdown needs to be carefully choreographed so that radiolabeled metabolites aren't produced in the brain. After all, PET just images radiation. It doesn't tell you which molecule it's coming from.

"All of these properties have to be built into the radiotracer, and it has to be a compound that can be quite readily labeled with 11C or 18F," Pike adds. Although radiochemists have a number of synthetic tools for tacking on those labels, not every method works equally well.

For example, Pike and his colleagues recently redesigned a radiotracer for imaging the metabotropic glutamate subtype-5 receptor (mGluR5)—a potential therapeutic target for pain, anxiety, depression, addiction, schizophrenia, and fragile X syndrome. A few years ago, researchers at Merck reported [18F]MTEB, a promising radioligand for imaging mGluR5. Unfortunately, the reaction that adds the 18F label to the compound's aryl ring proceeds in only 2–5% yield.

Credit: Victor Pike

By moving the radiolabel from the aryl position to an aliphatic spot, Pike's group was able to boost the yield of their radiolabeling step to 87%. The new radiotracer, known as [18F]SP203, is as potent as the Merck compound and, thanks to its relatively easy radiosynthesis, [18F]SP203 could become an important tracer for evaluating drugs aimed at mGluR5.

Developing a new radioligand can take as long as developing a new drug. That's why, Hargreaves says, at Merck chemists seek out both candidate CNS therapeutics and their tracers in tandem. Quite often, molecules with kinetics that are better suited to tracers than drugs will shake out during the initial stages of candidate screening.

Hargreaves says PET played a pivotal role when the company was evaluating its drug Emend (aprepitant), which is used to prevent the nausea and vomiting brought on by chemotherapy and surgical procedures. Emend targets the brain's neurokinin 1 receptors, and PET studies with the radiotracer [18F]SPARQ showed that Emend was very effective at saturating these sites.

That was a critical piece of information to have when Merck was evaluating Emend as a potential antidepressant, Hargreaves says. When Emend proved to have no clinical efficacy as an antidepressant, the decision to drop that line of research was an easy one. "If we had not done the imaging, we wouldn't have known if we'd engaged the target sufficiently," Hargreaves says. "But we could see that we had absolutely saturated the target in the brain with the drug but it didn't have the effect that we hoped it had."

These radiotracers have been used to image amyloid-β plaques in the brain. The fluorinated compounds are currently in clinical trials.

"PET just gives so much more confidence in what you're doing," says M. Edward Pierson Jr., a psychiatry discovery project leader at AstraZeneca. At the American Chemical Society's national meeting in Philadelphia last month, Pierson presented work his group had done in collaboration with the Karolinska Institute to image 5-hydroxytryptamine 1B (5HT1B) receptors.

The Tracer Race

Chemists who want to make radiotracers need to be synthesis speed racers. Each moment they spend in the chemical triathlon of synthesizing, purifying, and formulating these molecules, the precious radiolabel is decaying, steadily dimming its usefulness as a beacon in positron emission tomography (PET).

n terms of time, a radiochemist has no more than two to three half-lives of the radionuclide to prepare a tracer for injection. That gives chemists working with fluorine-18—which has a half-life of 110 minutes—just a few hours to assemble their PET agents. But that's a leisurely pace compared with carbon-11-labeled compounds, which fizzle out much faster, thanks to the nuclide's 20-minute half-life.

"When you make a radiotracer, you can't reflux your reaction overnight and work it up the next morning," says Chester A. Mathis, a professor of radiology and director of the PET facility at the University of Pittsburgh. "If you did that, there'd be nothing left in the morning in terms of radioactivity."

With time being such a critical element, the radioactive nuclide has to be installed in the final steps of a radiotracer's synthesis. With 18F, there's enough time for three synthetic transformations once the radiotracer arrives from the cyclotron, where it's made by bombarding stable isotopes with protons and deuterons.

If you're working with 11C, there's usually only time for two steps. "In 11C chemistry we may be performing multistep radiotracer synthesis within an hour," notes Victor W. Pike, chief of PET radiopharmaceutical sciences in the Molecular Imaging Branch of the National Institute of Mental Health.

Chemists who make radiotracers tagged with 11C need to work where there's a cyclotron on-site, but 18F's longer half-life means the radionuclide can be shipped short distances. Steven M. Larson, chief of nuclear medicine at Memorial Sloan-Kettering Cancer Center, recalls how when he was chief of nuclear medicine at the National Institutes of Health in the 1980s, the closest cyclotron to his lab in Bethesda, Md., was 12 miles away at the Naval Research Laboratory. It wasn't uncommon back then, he says, for traffic around the District of Columbia Beltway to sabotage his tracer synthesis.

These days the distribution of 18F radiotracers is much more reliable, Larson says. Still, Mathis points out, "most organic chemists would feel an extreme amount of stress" when faced with the challenge of making a radiotracer.

And it's not just the time crunch that radiochemists have to contend with. Because they're working with radioactive materials, all the synthetic machinations and purifications are done by robotic manipulation inside a box with lead-lined walls—all to get less than a microgram of radiotracer. "It's totally different thinking when you put it behind lead glass, and you can't put your hands in," Mathis says.

The receptor has been implicated in several psychiatric disorders and is a promising target for depression therapies. Pierson and his colleagues developed the radiotracer designated as [11C]AZ10419369 to help guide their research on this receptor. At the meeting, Pierson showed how the radiotracer could be used to measure how well the compound AZ12320927 binds to 5HT1B, although he wouldn't comment on whether or not it was a potential drug.

PET's ability to scope out the brain's biochemical blueprint has also been helping scientists get a better picture of Alzheimer's disease, thanks to radiotracers that bind specifically to the amyloid-β plaques that characterize the disease. These targeted tracers can highlight plaques just as they are starting to form. This distinguishes them from [18F]FDG, which images glucose metabolism and can only readily detect Alzheimer's in the late stages of the disease when the brain is riddled with plaques and billions of neurons are dead.

"Today the name of the game is not detecting Alzheimer's disease. The name of the game is to detect Alzheimer's disease before it manifests clinically," says Chester A. Mathis, a professor of radiology at the University of Pittsburgh. Evidence suggests that plaques develop in the brain for years before any symptoms of Alzheimer's emerge. By the time a patient visits a neurologist, chances are that the damage already is extensive, Mathis explains. "The neurons are dead. You can't bring them back."

Although there's some debate about the precise role plaques play in Alzheimer's—postmortem examinations have found plaques in patients who had no symptoms of the disease—scientists are hoping that by finding these plaques as they emerge, they'll be able to fight the disease before it's ravaged the brain.

Three plaque tracers are currently in Phase II clinical trials, and if they prove to be successful, doctors could someday use PET scans to screen high-risk candidates for Alzheimer's like they use mammograms to screen for breast cancer.

Pittsburgh Compound B (PiB), developed by Mathis and University of Pittsburgh psychiatry professor William E. Klunk, has been the most popular amyloid-β imaging agent. Because it's labeled with short-lived 11C, which means it must be made close to where it is used, PiB is not practical for widespread clinical use. Right now, of the 2,000 PET centers in the U.S., only about 25 have access to PiB.

Hoping to make the molecule more amenable to commercialization, GE Healthcare has developed a longer-lived 18F-labeled version of PiB, known as [18F]AH110690. The compound is currently in Phase II clinical studies.

The other two amyloid-β imaging agents that have made it to Phase II clinical studies are [18F]BAY94-9172 and [18F]AV-45. Both compounds were developed by University of Pennsylvania radiology and pharmacology professor Hank F. Kung and Avid Radiopharmaceuticals, a company Kung cofounded. [18F]BAY94-9172 has subsequently been licensed by Bayer.

Kung points out that these amyloid-targeting tracers help pharmaceutical companies ensure subjects in clinical trials for amyloid-β-fighting compounds have the Alzheimer's plaques as opposed to brain features that are diagnostic of some other kind of dementia. "If a drug company wants to test a drug specifically designed to lower plaque, they want to test a patient population that has these plaques. If the target is not even there, how can you measure the effects of your drug?" he says.

Eli Lilly & Co. is currently using [18F]-AV-45 to track some patients participating in the Phase III clinical trials of its Alzheimer's therapeutic LY450139, a compound that targets γ-secretase enzymes associated with amyloid production. The radiotracer should show if LY450139 changes the amount of plaque in the brain and how such changes translate clinically.

"The innovation that's come along with the imaging scientists just dovetailed perfectly with our γ-secretase inhibitor work," says Eric R. Siemers, medical director of Lilly's Alzheimer's disease research team. "It's a very interesting time for the field."

But even as pharmaceutical companies embrace PET as a tool for drug development, many in the field see trouble ahead, particularly in basic research. A report released last year by the National Academy of Sciences concluded that there was a "loss of federal commitment" to nuclear medicine. The Department of Energy, which has supported the bulk of nuclear medicine research since the 1950s, cut back its funding for the field by 85% from 2005 to 2006, and funding levels haven't made any appreciable gains in the past two years.

PET researchers' other chief complaint has particular relevance to C&EN readers. "What we're missing in this field are chemists," says Gilles Tamagnan, laboratory R&D director at Molecular NeuroImaging and associate professor of psychiatry at Yale University. "There's plenty of nice chemistry to be done. There just aren't enough people doing it," he says.

PET researchers say more radiochemists are required to staff the rapidly growing number of PET centers (see page 68). They also note that only a handful of synthetic chemists are working on new methods to incorporate radionuclides into molecules. Medicinal chemists are also needed to expand the limited list of radiotracers currently available. "There are less than 20 teams developing new radiolabeled compounds for the brain," Tamagnan notes, "and the brain is very big."


This article has been sent to the following recipient:

Chemistry matters. Join us to get the news you need.