News this Week

Science  24 Mar 2006:
Vol. 311, Issue 5768, pp. 1688

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution


    Violent Reaction to Monoclonal Antibody Therapy Remains a Mystery

    1. Eliot Marshall*
    1. With reporting by Gretchen Vogel in Berlin.

    CAMBRIDGE, U.K.—It took only minutes to realize that something had gone seriously wrong. On 13 March, six healthy volunteers in a clinical trial were injected with a “superagonist,” a drug meant to boost a type of T cell in the immune system, and soon all of them became violently ill. According to relatives and friends last week, the six vomited, collapsed, and passed out; one became bloated “like the Elephant Man,” his girlfriend told the press. Two additional participants who had received a placebo showed no ill effects.

    The volunteers were paid to participate in the trial (according to one, about $3460), the first human tests of a drug aimed at treating leukemia and autoimmune diseases such as multiple sclerosis and rheumatoid arthritis. They were given a synthetic (monoclonal) antibody called TGN1412, designed by TeGenero in Würzburg, Germany, and manufactured by Boehringer Ingelheim. In animal tests, the molecule triggered the production of so-called regulatory T cells, which keep the immune system in check. But something went amiss. The worst-affected volunteers were kept alive with mechanical life support and large doses of steroids to reduce inflammation. After 5 days, a doctor at Northwick Park Hospital near London said four were improving and three had been removed from machines. But the two worst affected remained in critical condition early this week.

    Exactly what triggered the reaction is not known. It seemed at first that an error in drug dosing or manufacture may have been to blame, says Simon Gregor, spokesperson for the U.K. Medicines and Healthcare Regulatory Authority (MHRA), which approved the trial. Managers at Northwick Park were so surprised that they even called in the police to check for evidence of a crime. But as MHRA and other investigators analyze materials and swarm over the private, 36-bed ward where the test took place, no crime or technical error has come to light. Suspicion is focusing instead on TGN1412 itself.

    Strong medicine.

    The human T cell, a multitasking agent in the immune system, is normally activated only when two receptors are stimulated (left). But the “superagonist” used in a London clinical trial can activate T cells by stimulating a single receptor (right).


    Six healthy volunteers injected with a test drug had to be rushed into critical care at Northwick Park Hospital; two others injected with a placebo weren't affected.


    MHRA, TeGenero, and the company that managed the trial, Parexel in Boston, Massachusetts, say that their procedures still look watertight. The volunteers' reactions were unforeseeable, they maintain. TeGenero's chief scientific off icer Thomas Hanke expressed “shock” in a statement on 17 March: “Extensive preclinical tests showed no sign of any risk.”

    Hanke told Science that a rodent version of the molecule was tested extensively at high doses in rats and mice, with no ill effects; TGN1412 itself was given to 20 cynomolgus monkeys in an unpublished study—after it was shown that their T cells were activated in the same way as human cells—with no significant adverse effects other than a short-lived increase in lymph node size. MHRA's Gregor says, “We have gone back [to the files] this week, and there is nothing in the documentation that would cause us to think there is a concern here.”

    But some independent observers have suggested that the trial was moving too aggressively. It was “a mad concept” to give a potent drug never tested in humans to six people at once, says medicines policy expert Joe Collier of St. George's Hospital Medical School in London. It would have been better to do one test and pause, he says. Monoclonal cancer vaccine researcher Angus Dalgleish of St. George's agrees that the procedure looks “bizarre,” because the results of T cell activation are notoriously hard to predict.

    Hanke responds that the trial's approach was “fairly common,” reflecting “current practice in biopharmaceutical development.” He adds: “We did not have any evidence to suspect that this drug would be unsafe at the dosage we applied,” which was, at 0.1 milligram per kilogram of weight, one-500th that given as a safe dose in animals.

    Some also question TeGenero's decision to move into human testing without a better developed—or at least a more publicly documented—rationale for how TGN1412 could up the count of regulatory T cells without turning on other, destructive responses. TeGenero's co-founder and scientific adviser Thomas Hünig of Würzburg University says that research since 1997 has shown that TGN1412 and analogous antibodies bind to the CD28 receptor on T cells, triggering a powerful expansion of cells dominated by regulatory T cells. Even at “horrific” doses in rats and mice, he says, regulatory cells dominated, giving credence to the view that these cells' damping effect would swamp out the more harmful effects of conventional T cells, also activated by TGN1412. The monkey study supported this confidence, says Hanker: “We saw no drug-related adverse events.”

    Some experts in monoclonal antibodies—including Dalgleish and Arlene Sharpe and David Hafler of Harvard Medical School in Boston, as well as John Isaacs of the University of Newcastle, U.K.—say that without having seen the relevant monkey data or results from tests with human cells in vitro, it's difficult to evaluate the argument that TGN1412 would likely have the same selective, benign effect in humans as in test animals. But they say they would not be surprised to find that TGN1412 stimulates harmful as well as beneficial effects. “A lot of cells” carry the CD28 receptor and might be activated, Hafler notes, adding, however, that “I wouldn't have thought [an accident like this] could happen.”

    Johannes Löwer, president of the Paul Ehrlich Institute in Langen, Germany, says his center was also approached by TeGenero to assess the TGN1412 trial. “We reviewed it very carefully” and reached the same conclusion as the U.K. group: The trial was safe and should proceed. Löwer offers two lessons for the future. Research is needed to define better animal models of the human response to CD28 agonists, he says. And he recommends that extra precaution be taken when antibodies are used to stimulate rather than neutralize components of the immune system.


    Long-Awaited Data Sharpen Picture of Universe's Birth

    1. Adrian Cho

    If good things come to those who wait, astrophysicists and cosmologists are reaping a well-deserved reward. Last week—a year later than originally planned—researchers working with NASA's Wilkinson Microwave Anisotropy Probe (WMAP) satellite released their second batch of data. The new measurements pinpoint the emergence of the first stars and tighten the screws on theories of how the infant universe expanded from the size of a marble to billions of light-years across in 10−35 seconds.

    “The new WMAP results are extremely important,” says Andrew Lange, an astrophysicist at the California Institute of Technology in Pasadena. “They usher in a whole new phase of cosmological research.”

    Launched in 2001, WMAP detects light from the big bang, which has cooled and stretched to longer wavelengths, leaving a pervasive haze of microwaves with a temperature of 2.7 kelvin. Three years ago, WMAP researchers used data collected in the satellite's first year in space to chart the faint variations in the temperature of the microwaves across the sky (Science, 14 February 2003, p. 991).

    Scrutinizing the fluctuations, the team hammered down the cosmos's vital statistics to unprecedented precision. The universe is 13.7 billion years old; is “flat” (curved neither inward like a gigantic sphere nor outward like a gigantic potato chip); and consists of a smattering of ordinary matter, much more unseen dark matter, and a whopping amount of space-stretching “dark energy.”

    Now, WMAP researchers have analyzed data collected during the second and third years of the satellite's mission. The microwaves coming from different places in the sky can point in different directions, like wind-speed arrows on a weather map, and the new work tracks how that polarization varies across the sky. The data give researchers another window into the infant universe, team leader Charles Bennett, an astrophysicist at Johns Hopkins University in Baltimore, Maryland, said Friday in a telephone press conference at Princeton University.

    Lights on.

    New data from NASA's WMAP satellite show that stars emerged 400 million years after the big bang and limit theories of how the newborn universe inflated at faster-than-light speed.


    The polarization arose when photons in the big bang afterglow collided with free electrons whizzing through the youthful universe. And the electrons popped out of neutral atoms when the atoms were illuminated by the light of the first stars. So by studying the polarization, the researchers could tell that the first stars emerged about 400 million years after the big bang, says David Spergel, a theoretical astrophysicist at Princeton University and member of the WMAP team.

    Knowing when the stars turned on and the fog of electrons emerged, researchers refined their analysis of the temperature variations, which are also affected by the electrons. The results rule out certain models of inflation, the mind-boggling expansion that took place in the universe's first split second, Spergel says. “This is a powerful step toward winnowing the field of contenders of how inflation took place,” says Brian Greene, a theoretical physicist at Columbia University.

    The polarization fluctuations are only a hundredth as pronounced as the temperature variations, and checking and rechecking the analysis took longer than researchers expected, Bennett says. “It wasn't anything fundamental that was difficult, but a lot of little things that had to be done right,” he says.

    Next, researchers hope to detect tiny swirls in the microwave background, which would be a sign of gravity waves from the big bang itself. But those swirls should be fainter still and may fall to WMAP's successor, Europe's Planck satellite, scheduled for launch in 2007.


    Plastics Break the Speed Barrier

    1. Robert F. Service

    Conducting plastics have long been a bit of a tease. Depending on their makeup, they can carry a current freely like metals or switch on and off like semiconductors. But when it comes to making transistors and other electronic devices, semiconducting plastics have been slowpokes. Even amorphous silicon, the low-grade silicon used to make the transistor arrays that drive liquid crystal displays, whisks electrical charges along almost an order of magnitude faster than semiconducting plastics do. That's a major reason plastics haven't dethroned amorphous silicon in large-scale applications. But new work from researchers in the United Kingdom and the United States could change all that.

    In a paper published online this week by Nature Materials, researchers led by Iain McCulloch, a polymer chemist at Merck KGaA in Southampton, U.K., report making plastic-based transistors that ferry electrical charges at nearly the same speed as amorphous silicon, a sixfold improvement over the previous generation of materials. “That's a very significant improvement” and could give plastic electronics a major push into electronics applications, says Zhenan Bao, a plastic electronics expert at Stanford University in California, who helped develop a previous record holder.

    The speed boost could finally help plastic electronics take advantage of their other selling points. Chief among these is their ease of fabrication. Unlike amorphous silicon, which must be grown in a vacuum chamber, plastics can be laid down from solution. That opens the door to patterning huge arrays of devices with what amounts to an ink-jet printer or other simple and cheap technologies. Plastic electronics researchers have been doing just that for over a decade. But when researchers first started laying down plastic conductors and semiconductors from solution in the early 1990s, the molecules in their films formed a jumble, like straw strewn on the ground. That disorder made it hard for conducting electrical charges to hop from one molecule to another in their devices, slowing their speed to a crawl.

    Bao and her colleagues, then at Lucent Technologies' Bell Laboratories, improved matters in 1996 when they came up with new semiconducting polymers known as regioregular polyhexylthiophenes. These plastics contained a series of ring structures in the polymer backbone with hydrocarbon arms that hung off the sides. Because the rings on neighboring molecules preferred to lie flat atop one another, the polymers automatically stacked themselves into perfectly ordered 20-to-50-nanometer crystals. On surfaces, the crystals packed side by side like bricks in a patio, forming a continuous sheet. Thanks to that orderly pattern, electrical charges jogged through the films at 0.1 centimeters squared per volt per second (cm2/V-s)—a respectable speed, but still only 1/10 as fast as in the best amorphous silicon films. The charges, it turned out, still ran into speed bumps each time they tried to hop from one tiny crystal in the sheet to the next.

    Crystal power.

    New semiconducting plastics form large crystals that help whisk electrical charges at higher speeds than ever before.


    McCulloch and his colleagues at the Palo Alto Research Center in California and Stanford decided to remove some of those speed bumps by growing larger crystals. They started with regioregular polythiophenes and fused a pair of neighboring rings at regular intervals along the polymer backbone. The rings locked the arms into a flatter shape and made it energetically even easier for neighboring molecules to stack side by side. As a result, crystallites in their film grew larger, to about 200 nanometers across, and the speed of charges in their devices ultimately reached 0.6 cm2/V-s. As a bonus, McCulloch says, the fused rings also are harder for oxygen atoms to break apart, making the plastics more resistant to degradation when exposed to air or water, another common problem with conducting plastics.

    McCulloch says the new plastics still need improvements before they can replace amorphous silicon. For one, because the materials are still prone to degradation, researchers must find ways to hide them from air. If they succeed, perhaps plastic electronics will stop teasing and get down to business.


    Cancer Institute Director Tapped for FDA

    1. Jocelyn Kaiser,
    2. Jennifer Couzin

    The National Cancer Institute (NCI) will be getting a new leader following the nomination last week of its controversial director Andrew von Eschenbach to head the U.S. Food and Drug Administration (FDA). As Science went to press, federal officials said von Eschenbach would step down “soon.”

    The $4.8 billion NCI is the National Institutes of Health's (NIH's) largest institute and the only one whose director is appointed by the president. A urologic surgeon and three-time cancer survivor, von Eschenbach arrived 4 years ago from the University of Texas M. D. Anderson Cancer Center in Houston, where he became friends with former president George H. W. Bush and his family. Von Eschenbach's tenure at NCI has been rocky, not least of all because he set the contentious goal of eliminating suffering and death from cancer by 2015. He also launched initiatives in nanomedicine, proteomics, and other areas at the same time success rates have dropped for investigator-initiated grant proposals.

    Von Eschenbach was named acting FDA commissioner last fall after the surprise resignation of Lester Crawford (ScienceNOW, 26 September 2005, After lawmakers protested that the two jobs posed a conflict of interest and too heavy a workload, von Eschenbach turned over day-to-day operations at NCI to John Niederhuber, a former University of Wisconsin surgical oncologist who joined NCI as a deputy director last fall.

    Rolling along.

    NCI's Andrew von Eschenbach, shown promoting a drug company-sponsored bicycle tour to fight cancer, is headed to FDA.


    A fight over FDA's handling of Plan B, the “morning after” pill, will likely slow von Eschenbach's confirmation in the Senate. Hillary Clinton (D-NY) and Patty Murray (D-WA) want the agency first to rule on whether the drug, now sold by prescription, should be made available over the counter, as many scientists and FDA officials have advocated. In August, Crawford announced he was putting off a decision indefinitely.

    Von Eschenbach's nomination is also likely to face some opposition. Citing the 2015 goal, Sidney Wolfe of the Washington, D.C.-based consumer activist group Public Citizen says von Eschenbach “continues to exhibit extraordinarily bad judgment” at NCI and is “a very bad choice to head this critical agency [FDA].”

    Two days after von Eschenbach was nominated, Niederhuber notified NCI staff that he would continue overseeing day-to-day operations until an acting director is announced. Several prominent cancer researchers told Science they hope the White House will launch a national search for von Eschenbach's successor. “Let's look for the very best person in the United States who's willing to take this job on,” says John Mendelsohn, president of M. D. Anderson. Applicants may be scared away, however, by the prospects of a declining NCI budget, a lame-duck Administration, and tight new conflict-of-interest rules for NIH senior officials.


    Studies Suggest Why Few Humans Catch the H5N1 Virus

    1. Dennis Normile

    This week, two research groups are independently reporting results that help explain why the H5N1 avian influenza virus is so lethal to humans but so difficult to spread. Unlike human influenza viruses, the teams report, H5N1 preferentially infects cells in the lower respiratory tract. Residing deep in the airways, the virus is not easily expelled by coughing and sneezing, the usual route of spread. The results “explain a lot of the mysteries” surrounding H5N1, says K. Y. Yuen, a virologist at the University of Hong Kong.

    A better understanding of the virus couldn't be more timely. Endemic in much of Asia, H5N1 has recently spread through Europe and to Africa. It has killed 98 of the 177 humans it has infected. Flu experts worry that if the virus mutates into a form that could be easily passed among humans, it could spark a pandemic. The two reports, which used different strategies but reached the same conclusion, suggest just what sort of mutation would be needed.

    One team, led by Yoshihiro Kawaoka of the University of Wisconsin, Madison, tested various tissues of the human respiratory tract for receptors to which the virus can bind. Human flu viruses preferentially bind to what are known as α 2,6 galactose receptors, which populate the human respiratory tract from the nose to the lungs. Avian viruses prefer α 2,3 galactose receptors, which are common in birds but were thought to be nearly absent in humans. Using marker molecules that bind to one receptor or the other, the team found that humans also have α 2,3 galactose receptors, but only in and around the alveoli, structures deep in the lungs where oxygen is passed to the blood. They describe their findings in the 23 March issue of Nature.

    The second team, led by pathologist Thijs Kuiken of Erasmus University in Rotterdam, the Netherlands, used a more direct technique to show that H5N1 readily binds to alveoli but not to tissues higher up in the respiratory tract. Kuiken, whose team publishes its findings online this week in Science (, notes that this pattern is consistent with autopsies that have shown heavy damage to the lungs but little involvement of the upper respiratory tract. Among experimental animals, the team reports, cats and ferrets more closely match the human pattern of infection than do mice and macaques. “This is an important factor to consider when planning experiments” to understand the pathology of H5N1, says Yuen.

    Yuen notes that the findings also explain clinical anomalies such as why nasal swabs of H5N1 patients are less reliable than throat swabs in detecting the virus. And they suggest that clinicians need to exercise particular care when performing procedures, such as intubation, that might give the virus a route out of a patient's lungs.

    The risk of a pandemic would ratchet up substantially should the virus acquire the ability to bind to receptors in the upper respiratory tract, Kuiken warns. But just how difficult that mutation is to acquire “is something this research did not address,” he says.


    Free-Flowing Supersolid Confirmed, But Origins Remain Murky

    1. Adrian Cho

    BALTIMORE, MARYLAND—Two years ago, a team of physicists created a stir by reporting that solid helium could flow without resistance, like a liquid devoid of viscosity. Now, three other groups have reproduced the bizarre effect. But data presented here last week at the March meeting of the American Physical Society also suggest that the “supersolid” flow occurs only in crystals riddled with defects.

    “The more perfect the crystal, the less supersolid there is,” says Ann Sophie Rittner of Cornell University. That observation suggests the flow is not an intrinsic property of crystalline helium, as the most exotic explanation would have it.

    The first signs of supersolidity were spotted by Eunseong Kim and Moses Chan of Pennsylvania State University in State College in a tiny can filled with pressurized solid helium-4, the heavier isotope of helium (Science, 1 July 2005, p. 38). When they set the can twisting atop a thin shaft and cooled it below 0.2 kelvin, the frequency of the twisting suddenly shot up.


    New data suggest supersolid flow can occur in a defect-riddled helium crystal (top), but not in an orderly one.


    The jump indicated that about 1% of the helium had let go of the can and was standing still as the rest of the helium crystal continued to gyrate. And that implied the solid helium flowed freely through itself. Theorists disagree on how that might occur, however, or whether it's possible in a perfect crystal. And until now experimenters hadn't reproduced the spectacular results.

    New experiments also reveal the helium letting go, says Keiya Shirahama of Keio University in Yokohama, Japan. “We have confirmed the Kim and Chan observation,” Shirahama says. Minoru Kubota and colleagues at the University of Tokyo have obtained similar results, as have Rittner and John Reppy of Cornell. But Rittner and Reppy say the flow vanishes when they heat the crystal to just below its melting temperature and slowly cool it, a process called annealing that eliminates defects.

    Chan says he sees no such effect, but Rittner and Reppy's data suggest that supersolidity is not an inherent feature of crystalline helium-4. Other experiments seem to support that conclusion. Kim, now at the Korea Advanced Institute of Science and Technology in Daejeon, and Chan have found that the supersolid signal reaches its maximum when their helium-4 contains several parts per million of the lighter isotope helium-3. It shrinks steadily as the helium-3 concentration falls to the lowest achieved level of a part in a billion.

    Chan's group also sees no clear spike in the specific heat—the amount of heat needed to raise the temperature a fixed amount—which ought to accompany the onset of superflow in a crystal. And Yuki Aoki and Haruo Kojima of Rutgers University in Piscataway, New Jersey, have yet to spot a tell-tale type of sound caused by the free-flowing portion of the helium sloshing back and forth.

    The new results still leave physicists debating how the flow occurs. Some had argued that it might arise when many helium atoms crowd into a single quantum wave in a phenomenon called Bose-Einstein condensation. That weird effect enables liquid helium-4 to flow without resistance, but some theorists argue that it is impossible in a well-ordered crystal. The new data make Bose-Einstein condensation in the solid “very unlikely,” says David Ceperley, a theorist at the University of Illinois, Urbana-Champaign.

    Others aren't so sure. Boris Svistunov of the University of Massachusetts, Amherst, and colleagues argue that the helium may jumble together to form a kind of glass. The atoms in the disorderly solid might undergo Bose-Einstein condensation, Svistunov says. That would make it a glass more than half-full with new physics.


    Diabetes Studies Conflict on Power of Spleen Cells

    1. Jennifer Couzin

    Three separate attempts have failed to replicate promising results that electrified the diabetes community 2 years ago. The outcome severely weakens the theory that the spleen cradles stem cells with curative powers against mouse diabetes. But some researchers, including the leader of the original studies, haven't given up on the idea.

    “An experimental result is not a valid one until it's been repeated,” says Diane Mathis of the Joslin Diabetes Center in Boston, Massachusetts. That's one reason her team and two others set out to duplicate startling findings published more than 2 years ago. Both the original and the recent studies reversed diabetes in ailing mice. But the recent studies attribute that reversal to a therapy that's too toxic for humans, although it has previously been used to prevent, and occasionally treat, mouse diabetes.

    Inconsistent observation.

    A 2003 Science paper reported that donor spleen cells (labeled in pink) could morph into pancreatic islets in a diabetic mouse. But other scientists have seen no hint of donor cells in their experiments.

    CREDIT: S. KODAMA ET AL., SCIENCE 302, 1223 (2003)

    The story began in 2001, when Denise Faustman of Massachusetts General Hospital in Boston and her colleagues described in The Journal of Clinical Investigation a combination of standard and novel strategies for treating diabetic mice. Her team gave diabetic animals an immune-modulating agent along with donor spleen cells. As is fairly common in such studies, the mice also got a temporary transplant of new islets, which include insulin-producing beta cells; the islets weren't curative, but they normalized the animals' blood sugar to give the other treatments a chance to take effect.

    When the sugar-stabilizing islets were removed, Faustman reported, 78% of the mice maintained normal blood sugar. Publishing more detailed work in Science in 2003, the team obtained cure rates as high as 90%. Furthermore, Faustman and her colleagues detected donor spleen cells that had morphed into beta cells and shown up in the animals' pancreases.

    The work engendered special hope. Unlike traditional islets, which must be culled from cadaver pancreases in short supply, “here you would have the possibility of collecting spleens” that people “don't really need,” says Louis Philipson of the University of Chicago, Illinois. “That promise … demanded that there be some sort of replication.” Starting on page 1774, the three groups—Joslin's Mathis, Christophe Benoist, and their colleagues; Philipson, Anita Chong, and their colleagues at Chicago; and Emil Unanue, Anish Suri, and colleagues at Washington University in St. Louis, Missouri—describe efforts to do just that.

    Following the Faustman protocol, the three injected between 22 and 53 diabetic animals with the immune modulator, called complete Freund's adjuvant (CFA). Containing killed bacteria that cause tuberculosis, CFA provokes a massive immune response and protects the pancreas from attack. It has previously been found to prevent and, sometimes, cure mouse diabetes. The immune reaction it triggers, however, is considered too hazardous for humans.

    Along with a CFA shot, the mice received injections of spleen cells from donor animals. Functioning islets were also wedged under their kidney capsules for 120 days to keep animals healthy during treatment.

    “The [spleen] injections did not take,” says Unanue. No group detected donor spleen cells in the blood or lymph nodes of mice or as new beta cells in the pancreas, suggesting that the immune system had destroyed them.

    However, like Faustman and her colleagues, all the groups successfully treated a subset of animals, although their “cure” rates were lower, roughly 10% to 25%. These animals, as well as some others that weren't cured, were found to have some beta cells in their pancreases, suggesting that mice can retain or perhaps regenerate the cells after diabetes develops. This finding is consistent with earlier mouse work and is increasingly suspected to apply to humans with newly diagnosed diabetes. But because the three groups could not detect spleen-derived beta cells, and because treatment with CFA and islets alone yielded the same results as when spleen cells were added to the mix, the groups attribute these cures to CFA and temporary islets.

    “It's hard to know what to say,” admits Kevan Herold, an endocrinologist at Columbia University. The groups “went to extremes” to replicate the spleen cell findings and were unable to do so.

    Faustman disagrees with this interpretation on several levels. The important point, she says, is that in every experiment mice were cured, an outcome she calls “fantastic.” The spleen cells, she adds, were “a minor point” in her paper. Still, she says she is certain that spleen cells can become beta cells and suggests that the groups used techniques too weak to detect them.

    Neuroanatomist Éva Mezey of the National Institute of Dental and Craniofacial Research in Bethesda, Maryland, endorses the idea that in mice, spleen cells can be a powerful mediator of autoimmune disease. With help from the Faustman lab, she has successfully tested CFA and spleen cells against the autoimmune disease Sjögren's syndrome in a handful of mice; the work is not yet published. In diabetes, another autoimmune disease, Mezey believes that CFA can jump-start residual beta cells, and that spleen cells can create new ones if none remain naturally.

    But most diabetes experts consider the spleen question settled. The problem now, says George Eisenbarth, executive director of the Barbara Davis Center for Childhood Diabetes at the University of Colorado Health Sciences Center in Denver, is that using CFA to cure mice is probably not relevant to humans. CFA's effects on mice have been studied for years, and a related but less toxic substance, the tuberculosis vaccine BCG, has failed to counter human diabetes. Faustman is raising money through the Iacocca Foundation to test BCG again (Science, 27 August 2004, p. 1237). Diabetes experts disagree whether, in light of these new findings, additional experiments with BCG should be considered.


    Turmoil Threatens to Sink Canadian Journal

    1. Paul Webster*
    1. Paul Webster writes from Toronto, Canada.

    TORONTO—Canada's premier journal of medical science continues to implode in the wake of allegations of censorship by its publisher, the Canadian Medical Association (CMA). The 20 February firing of Editor John Hoey and Deputy Editor Anne Marie Todkill (Science, 3 March, p. 1226) has prompted the resignations of 14 of 19 board members and a succession of senior and intermediate CMA Journal editors, as well as calls for a boycott. Researchers fear the loss of an important outlet for Canadian science.

    “The CMAJ no longer exists as we knew it,” says Amir Attaran, an associate professor of health law at the University of Ottawa, who wants researchers and reviewers to boycott the journal unless CMA reinstates Hoey and Todkill or explains more fully why it dismissed them. “It would be sad if such action were to lead to the CMAJ's demise, but it would still be preferable to accepting anything less than a fully free journal.”

    CMA officials say Hoey and Todkill were fired without notice because a “fresh approach” was required. But most observers believe it was the final step in a series of clashes over articles and editorials Hoey published during his 10-year reign and his commissioning of an outside panel that last month issued a scathing report of CMA's behavior. Hoey has declined comment due to a confidentiality agreement he and other editors signed with the CMA last year.

    The Ottawa-based journal is now under its third acting editor, and only three of nine section editors remain on staff. Noni MacDonald, a pediatrician and medical professor in Halifax, Nova Scotia, says she stepped in as editor after the CMA promised to investigate how the journal is managed and adopt guidelines employed by the Journal of the American Medical Association to protect editorial independence. MacDonald was a member of the journal's Oversight Committee, which departing editorial board members say failed to preserve the journal's editorial independence.

    In the meantime, MacDonald warns, a boycott could be fatal. “What's the goal there?” she asks. “To kill the journal, so we have no voice for national research issues?”

    CMAJ's influence has been rising, say scientists, according to both quantifiable measures such as impact factor and anecdotal evidence. “Ten years ago, this journal was just another throwaway publication produced by the doctors' association,” says Jacques Pepin, a microbiologist at the University of Sherbrooke in Quebec. “Under Hoey and Todkill, it has become an interesting journal for clinicians.”

    In August 2004, for example, CMAJ published Pepin's analysis of an outbreak of Clostridium difficile bacterial infections that had killed 200 patients in Quebec hospitals. Pepin says the paper alerted authorities in other countries about the difficulty of controlling the infection, which subsequently showed up in the United States. “For 18 months or so, the CMAJ was the only international journal with new research [on the outbreaks],” he says.

    Retired Canadian Supreme Court judge Antonio Lamer is leading a review of governance issues at the journal. But Pepin and others are worried that their recommendations may not go far enough. “The CMA seems not to have carefully planned any of its actions,” Pepin says. “We are worried the CMAJ may be headed back to throwaway status.”


    Seoul National University Dismisses Hwang

    1. Yvette Wohn*
    1. Yvette Wohn is a reporter in Seoul.

    SEOUL—Seoul National University's (SNU's) disciplinary committee announced on 20 March that it would dismiss disgraced cloner Woo Suk Hwang, a professor at its Veterinary College, for his involvement in fabricating data. Six other professors and co-authors on the two papers on embryonic stem cell cloning, which were published in 2004 and 2005 and later retracted from Science, received lighter sentences.

    At a press conference, Chang Ku Byun, dean of academic affairs, said dismissal is the harshest punishment the committee could impose. Hwang will be banned from working in a public position for 5 years after his dismissal and will receive only half of his retirement money.

    According to Byun, Hwang said that he would take all responsibility for the fabrication because he was the leader of the cloning project. In particular, Hwang admitted to ordering a junior researcher to take photographs of two stem cell lines in the 2005 article so that it would look as if the team had created 11 customized stem cell lines.

    The committee also suspended four other professors and cut the wages of two. Shin Yong Moon and Sung Keun Kang were both suspended for 3 months; Byung Cheon Lee and Curie Ahn were suspended for 2 months. They all will receive one-third of their wages during the suspension period and are not eligible for promotions for an additional 18 months. Chang Gyu Lee and Sun Ha Baek will have their wages deducted by one-third for 1 month.


    ASNU's disciplinary committee has dismissed Hwang for fabricating data.


    “The professors fundamentally went back on the values of integrity and honesty that should have been kept as an academic and professor of a national university,” Byun said. He explained that the committee imposed comparatively harsher punishment on Moon and Kang because Moon was a co-author of the 2004 paper and Kang was working in the same lab as Hwang and was deeply involved in the data manipulation. Lee and Baek were listed as co-authors but did not make any contributions to the paper, Byun added.

    Meanwhile, the special investigative team of Seoul Central Prosecutors' Office said, also on 20 March, that the initial contamination of stem cells in January last year was not deliberate, as they had previously thought, but arose from “accidents” by the researchers. The prosecutors are still investigating whether stem cells from Hwang's project were intentionally switched with those of MizMedi Hospital, which collaborated with Hwang on the research. The prosecutors, who are also examining how Hwang spent state funds and private donations, hope to conclude their investigations early next month.


    How a Marine Bacterium Adapts to Multiple Environments

    1. Elizabeth Pennisi

    Ahab chased a giant sperm whale through the seas for years. Sallie Chisholm has spent much of her professional career tracking far smaller denizens of the ocean: photosynthetic bacteria called Prochlorococcus. A marine microbiologist at the Massachusetts Institute of Technology (MIT), Chisholm helped discover these microbes 20 years ago. Now, she and her colleagues have charted the distribution of Prochlorococcus across the Atlantic Ocean and found that various strains congregate at different depths and places.

    In a Research Article on page 1737 of this issue of Science, Chisholm's team describes how water temperature and other environmental factors determine which strains thrive at particular places. And in a Report on page 1768, she and her colleagues reveal newly discovered “guest” genes in two strains that may help each strain survive in particular ocean environments. The studies “give us a rare glimpse into the population genetics of a native population,” says Stephen Giovannoni, a marine microbiologist at Oregon State University in Corvallis.

    With 1700 genes, Prochlorococcus has the smallest genome known for a free-living, photosynthetic organism. Yet, in low-nutrient environments typical of deep oceans, it outnumbers all other marine microbes and can account for up to 48% of net primary production there. Moreover, although all Prochlorococcus strains are quite closely related—at least according to traditional comparisons of their ribosomal RNA gene—Chisholm and her colleagues found a diversity of physiologically distinct strains, or ecotypes. For example, some grow best in dim light, whereas others require four times as much sunshine as their low-light-adapted peers.

    Hot on the trail.

    Erik Zinser searches for the cyanobacterium Prochlorococcus (inset).


    In 2003, Zackary Johnson, Chisholm's postdoctoral fellow, surveyed the ocean distribution of these ecotypes during a research cruise from Europe to the tip of South America. He and another postdoc, Erik Zinser, used the polymerase chain reaction to identify which ecotypes lived where. They also measured the biomass of Prochlorococcus and the amount of photosynthesis at specific depths at each sampling site. “This is the first [extensive] study about Prochlorococcus distributions at the ecotype level,” says Jakob Pernthaler, a microbial ecologist at the University of Zurich, Switzerland. “It's an impressive data set.”

    On the trip, water temperatures ranged from 5°C to 29°C, and nitrate, phosphate, silicate, and other nutrients utilized by Prochlorococcus were more abundant in the southern reaches. At certain sampling sites, two of the light-loving strains populated the surface in about equal numbers. And at depths of 50 meters, a low-light-adapted strain typically appeared, its abundance increasing with depth until about 75 meters down. Deeper down, a second low-light-adapted strain would take over. However, these distributions varied by latitude. Close to the equator, one low-light-adapted strain was as common at the surface as a high-light-adapted one. Together, these data provide “an ocean snapshot of Prochlorococcus biogeography” that will aid the understanding of this organism's ecology and role in the environment, says Eugene Madsen, a microbial ecologist at Cornell University.

    Edward DeLong, a microbiologist at MIT, also recently found stratification of the Prochlorococcus strains, but through a slightly less direct method (Science, 27 January, p. 496). His team, which included Chisholm and her postdoc Matthew Sullivan, sampled all the DNA in a water column from 10 meters to 4000 meters down, at a monitoring station off Hawaii. They then used a sophisticated computer program to assemble the bits of DNA into sequences large enough to identify genes and to get a read on which organisms were present. For example, that work showed that genes from high-light-adapted and low-light-adapted Prochlorococcus strains had different abundances at various depths.

    DeLong's group also discovered high concentrations of bacterial viruses called phages at ocean depths where Prochlorococcus lives; the greatest concentrations were at a depth of 70 m. The second study by Chisholm's team indicates that such phages are key to the adaptability of Prochlorococcus. Chisholm says that the viruses have inserted “islands” of genes into this bacterium, helping to create the various ecotypes she and others have found. In each strain, the islands are located in about the same place on the bacteria's chromosome, but the genes they contain vary from strain to strain. “These are hot spots in the genomes that appear to be very dynamic,” says Chisholm.

    When Chisholm's graduate student Maureen Coleman looked at these islands in two high-light-adapted strains growing in the lab, she found genes important for photosynthesis and for proteins for coping with stress, such as phosphorus deficiencies. Some of the islands' genes were the same in all strains. But not always: Prochlorococcus collected from the Mediterranean had a gene likely to be important for nitrogen scavenging, but this gene was lacking in another strain. Thus, the island genes may enable various strains to thrive at particular depths, light intensities, nutrient content, and temperatures. “The case is very strong that the islands and their variations play a role in [Prochlorococcus] evolution,” says Madsen.

    Pathogenic bacteria often have phage-delivered islands of DNA that contain genes for increased virulence or drug resistance, so it's not unexpected that similar islands enhance Prochlorococcus's survival, says Xabier Irigoyen of AZTI-Tecnalia in Pasaia, Spain. Nonetheless, Pernthaler notes, “it's a great piece of science that will inspire follow-up for other marine bacteria.”


    A Worrying Trend of Less Ice, Higher Seas

    1. Richard A. Kerr

    Startling amounts of ice slipping into the sea have taken glaciologists by surprise; now they fear that this century's greenhouse emissions could be committing the world to a catastrophic sea-level rise


    Greenland glaciers like those dumping into Sondre Sermilik Fjord have sped up and retreated.


    HAVE AN URGE LATELY TO RUN FOR higher ground? That would be understandable, given all the talk about the world's ice melting into the sea. Kilimanjaro's ice cloak is soon to disappear, the summertime Arctic Ocean could be ice-free by century's end, 11,000-year-old ice shelves around Antarctica are breaking up over the course of weeks, and glaciers there and in Greenland have begun galloping into the sea. All true. And the speeding glaciers, at least, are surely driving up sea level and pushing shorelines inland.

    Scientists may not be heading for the hills just yet, but they're increasingly worried. Not about their beach houses being inundated anytime soon; they're worried about what they've missed. Some of the glaciers draining the great ice sheets of Antarctica and Greenland have sped up dramatically, driving up sea level and catching scientists unawares. They don't fully understand what is happening. And if they don't understand what a little warming is doing to the ice sheets today, they reason, what can they say about ice's fate and rising seas in the greenhouse world of the next century or two?

    That uncertainty is unsettling. Climatologists know that, as the world warmed in the past, “by some process, ice sheets got smaller,” says glaciologist Robert Bindschadler of NASA's Goddard Space Flight Center (GSFC) in Greenbelt, Maryland. But “we didn't know the process; I think we're seeing it now. And it's not gradual.” Adds geoscientist Michael Oppenheimer of Princeton University, “The time scale for future loss of most of an ice sheet may not be millennia,” as glacier models have suggested, “but centuries.”

    The apparent sensitivity of ice sheets to a warmer world could prove disastrous. The greenhouse gases that people are spewing into the atmosphere this century might guarantee enough warming to destroy the West Antarctic and Greenland ice sheets, says Oppenheimer, possibly as quickly as within several centuries. That would drive up sea level 5 to 10 meters at rates not seen since the end of the last ice age. New Orleans would flood, for good, as would most of South Florida and much of the Netherlands. Rising seas would push half a billion people inland. “This is not an experiment you get to run twice,” says Oppenheimer. “I find this all very disturbing.”

    A rush to the sea

    Much of the world's ice may be shrinking under the growing warmth of the past several decades, but some ice losses will have more dramatic effects on sea level than others. Glaciologists worried about rising sea level are keying on the glaciers draining the world's two dominant ice reservoirs, Greenland and Antarctica. Summertime Arctic Ocean ice may be on its way out, but its melting does nothing to increase the volume of ocean water; that ice is already floating in the ocean. The same goes for floating ice shelves around Antarctic. The meltwater from receding mountain glaciers and ice caps is certainly raising sea level, but not much.

    The truly disturbing ice news of late is word that some of the ice oozing from the 3-kilometer-thick pile on Greenland has doubled its speed in just the past few years. In the 17 February issue of Science, for example, radar scientists Eric Rignot of the Jet Propulsion Laboratory in Pasadena, California, and Pannir Kanagaratnam of the University of Kansas, Lawrence, analyzed observations made between 1996 and 2005 by four satellite-borne radars. These synthetic aperture radars measure the distance to the surface during successive passes over a glacier. The changing distance can then be extracted by letting successive observations form interference patterns. The changing distance, in turn, translates to a velocity of the ice toward the sea.

    In central east Greenland, Kangerdlugssuaq Glacier more than doubled its speed from 2000 to 2005, Rignot and Kanagaratnam found, from 6 kilometers per year to 13 kilometers per year. That made it the fastest in Greenland. To the south, Helheim Glacier accelerated 60%. And on the west of Greenland, Jakobshavn Isbrae almost doubled its speed between 1996 and 2005. The accelerations are “actually quite surprising,” says glaciologist Julian Dowdeswell of the University of Cambridge in the United Kingdom. Even at its slower speed, Jakobshavn had ranked as one of the fastest-flowing glaciers in the world, perhaps the fastest; now it's just one of the pack.

    As glaciers draining the Greenland Ice Sheet are picking up speed, researchers are realizing that nothing has made up for the increased loss of ice. Greenland's pile of ice is getting smaller. How much smaller is still being debated, if only because of the vast scope of an ice sheet. What goes out through glaciers is just one part of the equation: Ice sheets also lose mass by melting and gain it from snowfall. To gauge those gains and losses, Rignot and Kanagaratnam used previously published estimates of how the warming climate over Greenland has increased meltwater losses and slightly increased snowfall, making for a growing net loss in addition to the glacier flow. All told, the scientists find that the loss of mass from Greenland doubled from 1996 to 2005, reaching 224 ± 41 cubic kilometers per year. Los Angeles uses 1 cubic kilometer of water per year.

    In another approach to estimating mass balance, researchers sketch the changing shape and therefore volume of the ice sheet. In a paper just out in the Journal of Glaciology, glaciologist Jay Zwally of GSFC and colleagues use satellite radars to measure the height of the Greenland Ice Sheet's broad plateau and airborne laser altimeters to monitor the height of glaciers draining to the coast, which are too small for satellite radars to see reliably. “We have strong evidence the ice sheet was near balance [during] the last decade of the 20th century,” says Zwally. “Our measures show a slight positive gain of 11 [cubic kilometers] per year” between 1992 and 2002.

    Global warming contrarians have already taken up Zwally's result as evidence that nothing much is happening with the ice sheet, so there's nothing to worry about. Zwally disagrees. “There's no question there's been an acceleration of some of Greenland's glaciers over the last 5 years,” after his surveys were completed, he says. “I would say that right now the current loss is 30 to 40 [cubic kilometers] per year,” he says, based on his gut feeling about the most recent radar and laser observations.

    That's getting close to the mass loss reported last fall using a third approach: repeatedly weighing the ice sheet. Geophysicists Isabella Velicogna and John Wahr of the University of Colorado, Boulder, reported in Geophysical Research Letters how the two satellites of the Gravity Recovery and Climate Experiment (GRACE), flying in tandem, gauge the mass beneath them. They precisely measure the changing distance between them caused by the gravitational pull of the passing ice. Between 2002 and 2004, GRACE found a loss of about 82 cubic kilometers of ice per year.

    Going under?

    Global warming might trigger a 6-meter rise in sea level that would inundate coasts (red) worldwide. Southern Louisiana (left) and South Florida (lower right) would be hard hit.


    All things considered, it seems clear that “Greenland has been shifting to a negative mass balance the last few years,” says glaciologist Richard Alley of Pennsylvania State University in State College. The same can be said for the West Antarctic Ice Sheet. All recent surveys have the far more massive East Antarctic Ice Sheet slowly gaining mass from increased snowfall. But that gain falls far short of compensating for the loss from West Antarctica. There, Zwally's analysis has the ice shrinking by about 47 cubic kilometers per year. And Velicogna and Wahr, writing in this week's issue of Science (p. 1754), report a GRACE-estimated loss of about 148 cubic kilometers per year. In West Antarctica, as in Greenland, the culprit is the acceleration of outlet glaciers in recent years (Science, 24 September 2004, p. 1897).

    Why the rush?

    The recent proliferation of galloping glaciers caught researchers unawares. “None of the models [of glacier flow] predict there should be such rapid change,” says glaciologist Ian Joughin of the University of Washington, Seattle (see Perspective on p. 1719). “If you look at a textbook, you'll see an ice sheet response time of 1000 years or more.” That's because models “treat ice sheets as a big lump of ice,” he says. They melt, or they don't melt.

    In the case of West Antarctica, there is tentative agreement about what is triggering the acceleration of the glaciers. Around the Palmer Peninsula that juts northward from West Antarctica, the world's strongest regional warming of the past 50 years first puddled the surface of ice shelves with meltwater. The meltwater then drove into the ice along growing cracks, breaking up shelves over a few weeks. Without the shelves to hold them back, apparently, the glaciers feeding them sped up (Science, 30 August 2002, p. 1494). To the south, where it's still far too cold for surface melting, a third-of-a-degree warming of the ocean seems to have eaten away at the shelves jutting into the Amundsen Sea. That in turn sped up Pine Island Glacier and its neighbors.

    Off to sea.

    The acceleration of glaciers draining both the Greenland and Antarctic ice sheets has meant more icebergs and thus more sea-level rise around the world.


    Around Greenland, however, both surface melting and shelf-bottom melting seem to be happening to some extent. Surface melting around the ice sheet's periphery has increased in recent years. Some of the meltwater plunges into open crevasses, where Zwally has shown that it can lubricate the bottom of the ice and accelerate ice flow. But, as Bindschadler argues on page 1720 of this issue of Science, the accelerating Greenland glaciers all flow through deep troughs that expose the ice to any warming ocean water, and all lost their buttressing ice shelves before or during acceleration. So both mechanisms are plausible drivers of glacier acceleration, but glaciologists cannot agree on their relative importance.

    Whither the world's ice

    If the recent behavior of ice sheets is not fully understood, their future is largely a blank. “We don't actually understand what's driving these higher velocities,” says Dowdeswell, so “it's difficult to say whether that's going to continue,” or spread.

    At the moment, ice loss from Greenland and West Antarctica combined is contributing less than half of the ongoing 2-millimetersper-year rise in sea level; the rest comes from melting mountain glaciers and the simple thermal expansion of seawater. If the recent surge of ice to the sea continues, sea level might reach something like half a meter higher by 2100. That would be substantial but not catastrophic. To produce really scary rises really fast (say, a meter or more per century), the air and water will have to continue warming in the right—or wrong—places. The temperature rise will have to spread northward around Greenland and in the south around West Antarctica, reaching the big ice shelves where most of that ice sheet drains. And glacier accelerations triggered near the sea must propagate far inland to draw on the bulk of an ice sheet.

    Faced with uncertainty about the present, paleoclimatologists look to the past. About 130,000 years ago, between the last two ice ages, the poles may have warmed as much as they will with only a couple of degrees of global warming. But sea level was considerably higher then, something like 3 to 4 meters higher. In two articles in this issue of Science, paleoclimatologist Jonathan Overpeck of the University of Arizona, Tucson (p. 1747), paleoclimate modeler Bette Otto-Bliesner of the National Center for Atmospheric Research in Boulder (p. 1751), and their colleagues consider whether the greenhouse world of a century hence might be as warm— and thus as destructive of ice—as during the previous interglacial.

    Not just the heat.

    Greenland glaciers retreat (tan area) when warming increases melting, but they can also accelerate when warmer ocean water destroys their lower reaches or added meltwater lubricates their undersides.


    First they simulated the climate of 130,000 years ago. Back then, Earth was tilted slightly more on its axis, so more solar radiation hit the high northern latitudes, driving warming there. Because the model included that added radiation, it had Greenland warming by about 3°C in the interglacial period. When that warming was put into a model of the ice sheet, the ice melted away slowly (because the model lacked any acceleration mechanisms) until about half remained. That produced enough meltwater to raise sea level 2 to 3 meters. Overpeck and colleagues suggest that another couple of meters of sea level rise could have come from West Antarctica; it was not as warm there, but much of the ice sheet lies below sea level, making it inherently unstable.

    When the climate model simulates the next 140 years of rising greenhouse gases, Greenland warms as much by 2100 as it did in the previous interglacial and would thus—eventually—melt as much. “Ice sheets have contributed meters above modern sea level in response to modest warming,” Overpeck and his colleagues conclude, and “a threshold triggering many meters of sea-level rise could be crossed well before the end of this century.”

    The paleoclimate argument for large, imminent ice losses “is fascinating and scary at the same time,” says Oppenheimer. “Paleoclimate always has a large amount of uncertainty, [but] we should take this as a serious warning sign. You could lock in a dangerous warming during this century.”

    An icy conundrum

    The ice sheet problem today very much resembles the ozone problem of the early 1980s, before researchers recognized the Antarctic ozone hole, Oppenheimer and Alley have written. The stakes are high in both cases, and the uncertainties are large. Chemists had shown that chlorine gas would, in theory, destroy ozone, but no ozone destruction had yet been seen in the atmosphere. While the magnitude of the problem remained uncertain, only a few countries restricted the use of chlorofluorocarbons, mainly by banning their use in aerosol sprays.

    But then the ozone hole showed up, and scientists soon realized a second, far more powerful loss mechanism was operating in the stratosphere; the solid surfaces of ice cloud particles were accelerating the destruction of ozone by chlorine. Far more drastic measures than banning aerosols would be required to handle the problem.

    Now glaciologists have a second mechanism for the loss of ice: accelerated flow of the ice itself, not just its meltwater, to the sea. “In the end, ice dynamics is going to win out” over simple, slower melting, says Bindschadler. Is glacier acceleration the ozone hole of sea level rise? No one knows. No one knows whether the exceptionally strong warmings around the ice will continue apace, whether the ice accelerations of recent years will slow as the ice sheets adjust to the new warmth, or whether more glaciers will fall prey to the warmth. No one knows, yet.


    Along the Road From Kyoto

    1. Eli Kintisch,
    2. Kelly Buckheit

    A special two-page graphic spread in this issue of Science depicts the state of greenhouse gas emissions for 10 countries that are key to the 1997 Kyoto Treaty, and where the planet is headed. (Read more.)


    A Clearer View of Macular Degeneration

    1. Jean Marx

    Genes tied to age-related macular degeneration confirm the notion that inflammation helps destroy the central area of the retina in this vision disorder

    More than 10 million people in the United States and an estimated 50 million worldwide suffer from age-related macular degeneration (AMD). This deterioration of the retina wipes out their central vision, robbing them of the ability to perform essential activities such as reading or driving a car. Until recently, researchers had few clues to what causes AMD. But a series of recent gene discoveries has gone a long way toward solving the mystery.

    The work has shown that variations in two genes encoding proteins in the so-called complement cascade account for most of the risk of developing AMD. This complex molecular pathway is the body's first line of defense against invading bacteria, but if overactive, the pathway can produce tissue-damaging inflammation.

    Sign of trouble.

    The retinas of AMD eyes (left) develop abnormal deposits called drusen (yellow spots); as shown by the red staining (right), drusen contain complement proteins generated by inflammation.


    The new links between AMD and the complement genes suggest that excessive inflammation resulting from uncontrolled complement activity underlies the vision-destroying changes that particularly strike the macula, the central region of the retina. Indeed, such inflammation-promoting gene variants contribute to the development of perhaps as many as 75% of AMD cases. Several other genes have also been implicated recently, including one of as-yet-unknown function that may also make a substantial contribution to AMD risk.

    At long last, say researchers, they have the kind of information needed to find ways to prevent or treat AMD. “I think we're headed for a period of time when we're going to come up with possible therapies,” says AMD researcher Michael Gorin of the University of Pittsburgh School of Medicine in Pennsylvania. This might be accomplished, for example, by finding ways to inhibit complement activity in the eye.

    A tough disease to crack

    The causes of AMD have been hard to pin down partly because the disease develops late in life, usually after age 60. In addition, AMD is a complex disease, caused by an interaction between multiple genes and environmental factors such as diet and smoking. That's made it hard to do studies aimed at tying particular gene variants to the disease. “Until last year, we just didn't have a clue about the etiology [of AMD]. It's been very frustrating,” says Gregory Hageman of the University of Iowa in Iowa City, one of the field's pioneers.

    In their search for clues, researchers have looked at hereditary eye diseases that develop early in life and mimic some features of AMD pathology, such as the development of drusen: abnormal deposits of proteins and other materials in the retina. They identified the genes at fault in some of these early developing diseases and for a time hoped that the same genes might also be major contributors to AMD. “It turns out that for the most part that idea was wrong,” says Gorin.

    About 15 years ago, Hageman began pursuing a different tack, collecting donated eyes from both people afflicted with AMD and those who were not. His analyses of those eyes suggested that inflammation is a key player in the etiology of AMD.

    For example, working with Iowa colleague Robert Mullins and Don Anderson and Lincoln Johnson of the University of California, Santa Barbara, Hageman found that proteins associated with immune system activity are located in or near the drusen in eyes with AMD. These proteins included various activated components of the complement system, such as the membrane attack complex, which is the business end of the complement cascade. It destroys cells infected with bacteria or viruses by poking holes in their membranes, but it can also damage normal cells if not controlled.

    Based on these and other results, Hageman and his colleagues proposed about 5 years ago that drusen growth begins when some as-yet-unknown insult damages cells in the retina. The leftover cell debris provides the seed for drusen formation and triggers complement activation and local inflammation. Over time, the drusen grow as they accumulate inflammatory proteins and other materials, and the inflammation persists, causing additional damage to the retina and, in the worst cases, blindness.

    At the time, this idea “was not met with a lot of positivity,” Hageman wryly recalls. “He pushed the hypothesis for many years, and nobody believed him,” says Rando Allikmets of Columbia University, a recent collaborator. Allikmets notes that he, like many other AMD researchers, acknowledged that there could be an inflammatory component in AMD but “thought this was secondary or tertiary” to whatever was actually causing the retinal damage.

    A genetic link to inflammation

    The turning point in the inflammation story came just over a year ago, thanks mainly to new gene-hunting tools, including the human genome sequence as well as the growing library of human single-nucleotide polymorphisms (SNPs), subtle DNA sequence changes that can be used to pin down the gene variants at fault in a disease. Over the past few years, researchers have used these SNPs to identify several chromosomal regions likely to contain genes that influence the risk of getting AMD and then zeroed in on the genes themselves.

    Last March, three independent groups-led by Josephine Hoh of Yale University School of Medicine; Margaret Pericak-Vance of Duke University Medical Center in Durham, North Carolina; and Albert Edwards of the University of Texas Southwestern Medical Center in Dallas and Lindsay Farrer of Boston University School of Medicine-reported online in Science that they had uncovered a gene on chromosome 1 that greatly increases the risk of getting AMD. The gene encodes a protein called complement factor H that helps keep the complement system under tight control so that it doesn't attack the body's normal cells. The researchers found that people bearing a particular variant of the factor H gene were much more likely to get AMD than were people with other variants (Science, 15 April 2005, pp. 362, 385, 419, and 421). Their calculations showed that the high-risk variant could explain up to 50% of the cases, presumably because the protein product of that gene is less effective in inhibiting the complement pathway.

    Hageman describes this confirmation of complement involvement in AMD as “pretty wonderful”—even though the other teams got their work into print before his own. In the 17 May 2005 Proceedings of the National Academy of Sciences, Allikmets and Hageman, working with Bert Gold and Michael Dean of the National Cancer Institute in Bethesda, Maryland, and colleagues reported that they, too, had linked the same high-risk variant of the factor H gene to AMD and also identified other variants that appeared protective. In addition, they showed that factor H is made in the macula and is present in drusen. In other words, it's right where it should be to influence AMD development.

    Missing faces.

    As illustrated by the simulated picture at right, AMD robs its victims of their central vision and can leave them legally blind even though a rim of peripheral vision may remain.


    The case for a link between complement activity and AMD risk got another major boost earlier this month. In the March issue of Nature Genetics, Allikmets, Hageman, Gold, Dean, and their colleagues report linking the gene for a second component of the complement cascade to the eye disease. This gene, located on chromosome 6, produces a protein called factor B involved in complement activation; the researchers again found both high-risk and protective variants.

    The earlier factor H studies showed that some people with healthy eyes also carry the AMD-predisposing variant, although it is present in more AMD patients. Allikmets says that the existence of protective variants of the factor H and B genes can help explain why some people with the “bad” variants don't get AMD. By looking at the variants of both genes that people have, he says, “you get a much clearer picture. Seventy-four percent of the patients had no protective [gene variants], while 56% of the controls had at least one.”

    Still unclear, however, is the nature of the initial trigger that touches off complement activation. An infection is one possibility. About a year ago, Joan Miller of Harvard Medical School in Boston and her colleagues reported evidence indicating that the eyes of patients with the so-called wet form of AMD, characterized by blood vessel growth in the macula, had been infected by the bacterium Chlamydia pneumoniae. A causative link between Chlamydia infection and AMD needs to be confirmed, however.

    More AMD genes

    A third gene recently tied to AMD doesn't fall neatly into the complement story. In work published last summer in the American Journal of Human Genetics, Gorin and his colleagues followed up on previous studies placing an AMD gene on chromosome 10. Gorin says that his group's analysis homed in on two tightly linked genes (designated PLEKHA1 and LOC387715) but couldn't discern which is the culprit. However, in a paper that appeared a few months later in Human Molecular Genetics, a team led by Bernhard Weber of the University of Regensburg, Germany, reported that the strongest AMD risk seemed to be associated with a single amino acid change in the protein predicted to be encoded by LOC387715. In work published online early this month in the American Journal of Human Genetics (AJHG), Pericak-Vance and her colleagues confirmed that conclusion.

    The genetic analyses indicate that LOC387715's effect on AMD is independent of that of the factor H gene but is almost as strong, contributing to perhaps 40% of AMD cases. Still, the function of the gene's protein is a mystery. “Until one knows what it does, you can't really say it's the gene,” Gorin cautions.

    In their AJHG paper, Pericak-Vance and her colleagues also point out an intriguing connection between the high-risk variant of LOC387715 and cigarette smoking, one of the strongest environmental risk factors for AMD. The researchers found that the combined risk of smoking and carrying the AMD-promoting gene variant was more than the sum of the risk of the two individually. This indicates, Pericak-Vance says, that the two factors interact to foster AMD development.


    Despite early skepticism from his colleagues, Gregory Hageman's view that inflammation plays a causal role in AMD is proving correct.


    The factor H and B genes and LOC387715 are likely not the only ones that affect AMD risk. For example, in a study published in January in Investigative Ophthalmology and Visual Science, Pericak-Vance's team looked at eight candidate genes that were suspected of involvement in AMD. Their analysis implicated three of them, including the VEGF gene and two involved in lipid metabolism. The product of the VEGF gene stimulates blood vessel growth, suggesting it might be involved in wet AMD, which is the most severe form.

    But the complement genes and LOC387715 are certainly the major contributors to AMD risk, and establishing that, Hageman says, is good news for people who might develop AMD. Having to look at just a few genes could make it easier to identify high-risk individuals, who could then take preventive steps such as avoiding smoking, decreasing their fat intake, and increasing their intake of antioxidants and carotenoids. Many of these steps, suggested by epidemiology studies, are the same ones prescribed to reduce a person's risk of heart attack and stroke. “There is a very similar risk profile for cardiovascular disease and AMD,” says epidemiologist Johanna Seddon of Harvard Medical School.

    If just a few genes account for almost all of the risk of getting AMD, that should also help in devising therapies that can slow or prevent vision loss in people with the disease. “If each gene contributed just 4% or 5%, developing a therapy [based on those genes] would be pretty much impossible,” Hageman says. However, the new studies suggest a much brighter outlook for efforts to beat this devastating disease.