# News this Week

Science  16 Aug 2002:
Vol. 297, Issue 5584, pp. 1104
1. EVOLUTION

# Could Poor Nutrition Have Held Life Back?

1. Richard A. Kerr

Our planet's middle age was “the dullest time in Earth's history,” as one wag put it. From about 2 billion to 1 billion years ago, in the middle of the Proterozoic Eon, the green scum of cyanobacteria and their ilk reigned supreme, changing little from eon to eon. During that long hegemony, our cellular ancestors, the eukaryotes, went nowhere evolutionarily. Prospects looked dim for them despite their billion-plus-year history and rising oxygen levels.

Now a geochemist and a paleontologist propose an intentionally provocative hypothesis: that the eukaryotic algae, at least, were held back for a billion years by a nutritional deficiency brought on by rising oxygen levels and mediated by trace metals dissolved in the sea. The eukaryotic algae were not unleashed until further increases in atmospheric oxygen finally permeated the deep sea fewer than a billion years ago, according to the hypothesis, which is presented on page 1137 of this issue of Science by geochemist Ariel Anbar of the University of Rochester in New York state and paleontologist Andrew Knoll of Harvard University. The idea “is elegant,” says geochemist Timothy Lyons of the University of Missouri, Columbia. “It explains a lot, but we really need to have more data. This paper will define research for years to come.”

What might be called the Malnourished Earth hypothesis is founded on a 1998 proposal by geochemist Donald Canfield of Odense University in Denmark. With the exception of the near-surface, he suggested, the world ocean remained oxygen-free, or anoxic, after oxygen first appeared in the atmosphere about 2.2 billion years ago. That was when the great banded iron formations now mined for their metal ceased growing from the anoxic, iron-rich seawater that typified Earth's oceans since its beginnings. In the conventional view, oxygen had ended the age of banded iron formations by penetrating throughout the ocean and removing the iron in the form of insoluble iron oxides.

But Canfield, drawing on new measurements of the sulfur-isotope composition of Proterozoic ocean sediments, suggested that banded iron formation had been turned off by rising sulfide levels, not by oxygen. Below the uppermost waters the ocean remained anoxic, he argued, while the newly oxygenated atmosphere weathered large amounts of sulfur off the land and into the ocean. There it took the form of sulfides (including hydrogen sulfide). By forming insoluble compounds with the iron, these sulfides removed most of the iron from the sea. The ocean in the Mesoproterozoic, from 1800 million to 800 million years ago, would thus have been like neither Earth's primordial ocean—anoxic and iron-rich—nor the modern ocean—oxygen-rich and iron-poor. Chemically, it would have been like today's Black Sea.

In such an ocean, iron would be removed to low concentrations, as would other metals, including molybdenum, copper, zinc, vanadium, and cadmium. The eukaryotic algae, in particular, would have missed some of those metals sorely. Their enzymes for taking up essential nitrogen in the form of nitrate are built around an iron atom and a molybdenum atom. The less of each of these metals in seawater, the harder it is for eukaryotic algae to get their nitrogen from nitrates. They have no way to fix nitrogen—convert the nitrogen gas of the atmosphere into usable nitrogen—as cyanobacteria do. Even bacteria's nitrogen fixation requires iron-based enzymes, the most efficient of which requires molybdenum as well. And eukaryotes generally lack bacteria's ability to ingest nitrogen-containing particles.

All in all, the Mesoproterozoic could have meant nutritionally hard times for ocean life, Anbar and Knoll conclude. The dearth of fixed nitrogen induced by scarce iron and molybdenum could have caused the generally low ocean productivity of the Mesoproterozoic, they say. And eukaryotic algae would have been at an evolutionary disadvantage. Multicellular algae, in particular, compete best when there are high levels of nitrate, not just the bare minimum. That could explain the steady and persistently low diversity of relatively simple eukaryotic algae through the Mesoproterozoic, Anbar and Knoll say, until 2 billion years after their first appearance. Only after mountain-building wrenched North America, sending more weathered metals into the sea, and atmospheric oxygen levels increased further, converting sulfides to sulfates and freeing up the trace metals, could the eukaryotes end their “profoundly boring” period and diversify toward larger, multicellular plants.

“They've put together a nice working hypothesis,” says geochemist Louis Derry of Cornell University in Ithaca, New York. “It could be true, … [but] the literature on [Proterozoic] environments is full of interesting ideas. There's not as much data, so it is easy to carry out thought experiments. The trick to doing good science on these old materials is finding things you can test.”

Geochemist Yanan Shen of Harvard, Canfield, and Knoll have in fact recently tested the foundation of the working hypothesis using Mesoproterozoic rocks from two former ocean basins in northern Australia. To judge by indicators of anoxia such as iron pyrite preserved in the 1.7-billion-year-old sediments, the two ancient ocean basins “simply look like the deep Black Sea,” says Knoll. But these might have been restricted ocean basins such as the Baltic, not the open ocean, so Anbar is working on a molybdenum- isotope analysis that could gauge the oxidation state of the world ocean from a few samples. Even then, many geophysiological links would remain to be proven between ancient ocean chemistry and the rise of well-fed eukaryotes.

2. LANGUAGE EVOLUTION

# 'Speech Gene' Tied to Modern Humans

1. Michael Balter

The ability to communicate through spoken language is the trait that best sets humans apart from other animals, most human origins researchers say. Last year the community was abuzz over the identification of the first gene implicated in the ability to speak. This week, a research group shows that the human version of this so-called speech gene appears to date back no more than 200,000 years—about the time that anatomically modern humans emerged. The authors argue that their findings are consistent with previous speculations that the worldwide expansion of modern humans was driven by the emergence of full-blown language abilities.

“This is the best candidate yet for a gene that enabled us to become human,” says geneticist Mary-Claire King of the University of Washington, Seattle. But other researchers caution that uncertainties underlying the team's mathematical analysis, as well as debate about the gene's function, make dramatic conclusions premature. The case that the gene is closely linked with language ability “can only be said to be circumstantial,” comments geneticist David Goldstein of University College London.

The gene, called FOXP2, was identified last fall by geneticist Anthony Monaco's group at Oxford University, in collaboration with cognitive neuroscientist Faraneh Vargha-Khadem and colleagues at the Institute of Child Health in London (Science, 5 October 2001, p. 32). They showed that FOXP2 mutations cause a wide range of speech and language disabilities. Geneticist Svante Pääbo's group at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, in collaboration with Monaco's team, then set about tracing the gene's evolutionary history.

The Leipzig team, with graduate student Wolfgang Enard taking the lead, sequenced the FOXP2 genes of several primates—chimpanzee, gorilla, orangutan, and rhesus macaque—as well as that of the mouse and compared them with the human sequence. The gene encodes a protein with 715 amino acids; it resembles other members of a family of regulatory genes implicated in embryonic development. Since the last common ancestor of humans and mice, which lived some 70 million years ago, there have been only three changes in the protein's amino acid sequence, the team reported online in Nature on 14 August. And two of these changes have occurred in the human lineage since it split with that of chimps roughly 6 million years ago.

These amino acid changes might have given some evolutionary advantage to the hominids who harbored them, the researchers surmised. This hypothesis gained support from calculation of a parameter known as Tajima's D statistic, an estimate of how much selection pressure has been exerted on a particular gene over the course of evolution. In general, the more negative this D value, the more selection has occurred. FOXP2 had a highly negative D value—in fact, out of 313 well-characterized human genes recently analyzed, only one outscored FOXP2 (Science, 20 July 2001, p. 489).

The team estimated how recently the human version of FOXP2 became “fixed” in human populations—that is, when all humans harbored the last amino acid substitution. Although the date cannot be pinpointed, the team concluded that the fixation was 95% likely to have occurred no more than 120,000 years ago and was virtually certain to have occurred no earlier than 200,000 years ago.

Most of the researchers who spoke with Science agree that the authors make a strong argument that the human version of FOXP2 has been favored by natural selection. “Overall, [that] case has been made,” says Goldstein. But he and others were less eager to accept the dating of the gene: “Dating analyses [such as these] are fraught with uncertainty.”

Now that the human version of FOXP2 has been found to be advantageous to human evolution, the debate over the gene's role in language has become even more relevant. Some scientists caution against overstating the importance of FOXP2 in the evolution of language ability. “It would be foolish to talk about FOXP2 as the gene that evolved to permit the emergence of speech and language,” says Elizabeth Bates, a neuroscientist at the University of California, San Diego, although it is clearly “one of the genes” that did so. Indeed, Pääbo suggests that this gene, which may be implicated in the ability to make the mouth and facial movements essential to speech, might have been selected for precisely because it improved vocal communication once language had already evolved.

3. NUMBER THEORY

# Simple Recipe Creates Acid Test for Primes

1. Barry Cipra

Quick, now: Is 341 a prime number? That one's pretty easy to answer. How about 4,294,967,297? That's still a snap if you use a computer. But what if the number you're interested in has thousands of digits? Then things get murky, because the obvious way to settle the issue—systematically checking whether smaller numbers divide it—takes far too long. In recent decades, theorists have devised clever algorithms for telling whether a large number is prime, but none that could be proven to work quickly.

Until now.

Three computer scientists at the Indian Institute of Technology in Kanpur have found what researchers have long sought: a provably efficient algorithm for testing primes(whole numbers evenly divisible only by themselves and 1). Manindra Agrawal, a professor of computer science, and two students, Neeraj Kayal and Nitin Saxena, announced their result early this month, e-mailing copies of it to a number of experts in computational number theory.

“This is really fantastic, that we now know this to be true,” says Hendrik Lenstra, one such expert at the University of California, Berkeley. The new result tidies up one of the corners of modern cryptography, which relies on hard-to-factor composite (nonprime) numbers to encrypt information. Although the algorithm is not practical at present, just knowing that it exists “simplifies our picture of what is going on,” notes Andrew Odlyzko, a computational number theorist and director of the Digital Technology Center at the University of Minnesota, Twin Cities.

What especially intrigues number theorists is that the algorithm and the proof of its efficiency are both very simple. Lenstra says he printed out the paper and discussed it with colleagues over dinner. “Before we had coffee we knew it was right,” he says.

Put technically, the new result is a “polynomial time” algorithm for primality testing. That means it can take any N-digit number and return a yes-or-no answer in a run time bounded by a power of N—in this case, N12. In computer science, polynomial time is the gold standard of efficiency. Any algorithm whose run time increases more steeply than that (as, say, 2N does) quickly becomes time-consuming enough to bog down even the fastest computer.

Like other modern tests for primes, the new algorithm is based on a number-theoretic fact that Pierre de Fermat (of Last Theorem fame) discovered in the 17th century: If n is prime, then it evenly divides an — a for any number a. Fermat's test makes it possible to prove that a number n is not prime without finding any of its factors. For example, 29 − 2 = 510, which is not divisible evenly by 9. Hence, 9 cannot be a prime number.

Unfortunately, some composite numbers n also evenly divide an — a. To eliminate such “false positive” readings, the new algorithm runs a more elaborate but still elementary test, based on searching for pairs of numbers that fulfill a few simple conditions. If the search is successful, then n is declared composite; otherwise, it's prime. The key to the algorithm's efficiency is that the search can be restricted to a small range of numbers.

Constructing the primality test took a year of grueling work, Agrawal says. “There was nothing like an ‘in a flash everything started making sense’ feeling,” he says. He and his students whittled away at the problem from various angles until July, when Agrawal hit on a substitution that made the last roadblock melt away.

“It's a bit of a surprise that such an easy algorithm had been missed all these years,” says Carl Pomerance, a number theorist at Bell Labs in Murray Hill, New Jersey. “It's a delightful surprise—and perhaps also a bit of an embarrassment for those who have been working in the field, such as myself.”

Cryptographers might well have more mixed feelings. They rely on now-difficult number theoretic computations, such as factoring large numbers, to safeguard cryptosystems that have become mainstays of the computer security business. If primality can be vanquished so easily, who's to say that a polynomial-time algorithm for factoring isn't just around the corner? Or, as Pomerance puts it, “What else have we overlooked?”

4. ITALIAN REFORMS

# Planned Mergers Raise Hue and Cry

1. Alexander Hellemans*
1. Alexander Hellemans is a science writer in Naples.

NAPLES—Earlier this month Italian scientists were stunned to learn that their government is planning a major overhaul of the country's scientific establishment, including the axing of several institutes. Researchers are denouncing the plan, leaked to a prominent newspaper and denied by the government, as a ham-handed power grab drafted without their input.

According to a report in the 2 August edition of la Repubblica, the Ministry of University and Scientific Research intends to abolish eight publicly funded institutes and transfer the staff to the country's National Research Council (CNR). The institutes include the Anton Dohrn marine biology research station (Stazione Zoologica) in Naples, the National Institute of Applied Optics in Florence, and the National Institute for Experimental Oceanography and Geophysics in Trieste. Internal reforms would also be implemented at CNR, the Italian space agency, the National Institute of Geophysics and Volcanology, and the National Institute for Astrophysics.

“The Stazione Zoologica is a center of excellence,” says Antonio Miralto, the institute's former director. “To have researchers of the Stazione Zoologica become part of the CNR would be a great error on the part of the government.” William Speck, director of the Marine Biology Laboratory in Woods Hole, Massachusetts, says he was “shocked” by the news. “It is the oldest marine laboratory in the world.”]

CNR researchers are especially upset about a provision to choose institute directors through a top-down selection process, overseen by a president and an administrative council dominated by government appointees. Franco Pacini, an astronomer at the University of Florence and president of the International Astronomical Union, blasts this proposal as “obscene and unacceptable. It will basically put in the hands of politicians or administrators the choice for the best scientific leaders, even at the intermediate level.”

The government's draft plan is close to its final form, says Claudia Di Giorgio, the journalist who received a leaked copy of the so-called “decree” after promising not to distribute it. However, a ministry spokesperson told Science that research minister Letizia Moratti is still working on the decree, which will be distributed to political bodies and the scientific community at the end of next month.

CNR is already moving to consolidate its 300 institutes into 100 larger units, and several scientists say it's ironic that the decree comes just as CNR has moved toward a more open selection process for its scientific directors. In addition, they are extremely unhappy that the proposed changes were drafted without their input. “I was not consulted by the minister … we didn't receive any formal information,” says Lucio Bianco, CNR's president, adding that he first learned about the decree's contents in the newspaper. “If the description of the document is true, it cannot be acceptable to the researchers of the CNR.”

Bianco says that he hopes to discuss the reforms with Moratti “in an open way” after the government returns from vacation in September. In the meantime, scientists are planning to stage a public protest and will discuss the government's proposal at a meeting on 10 September in Rome organized by Italy's Association of Ph.D.s and other groups. Few scientists contacted by Science believe that the reform measures will survive in their current version.

5. ANIMAL BEHAVIOR

# Birds Spy on Neighbors To Choose Nest Sites

1. Jay Withgott

Information is power, even for birds. Faced with tough choices, animals that know how others have fared in comparable situations can make better decisions. On page 1168, researchers report that collared flycatchers decide where to nest and whether to return the next year based in part on knowledge of their neighbors' reproductive success. “How individuals collect this information is enigmatic,” says Tomas Pärt, an evolutionary ecologist at the Swedish University of Agricultural Sciences at Uppsala. “This result suggests that the cues used may be unexpectedly fine.”

Choosing a good breeding site may mean the difference between begetting many offspring or none at all. Previous work on group-nesting seabirds, such as cormorants and kittiwakes, had turned up observational evidence that birds monitor the success of their fellows in assessing breeding sites. A team led by Blandine Doligez, then at France's National Center for Scientific Research (CNRS) in Paris and Uppsala University, tested experimentally to what extent birds make use of information gleaned by watching their neighbors, which ecologists call “public information.”

Doligez, now at the University of Bern, Switzerland, worked at a long-term research site at Gotland, Sweden, where collared flycatchers sport identifying color-coded leg bands. Researchers there had noticed flycatchers peering into the nest boxes of other birds. “No one really focused on this behavior, [but] I thought, that's really a sign” they're gathering information, she says.

To manipulate such information, the researchers took nestlings from some nests and added them to others, creating some plots of woodland with supersized broods and others with measly numbers of young. The team then monitored these plots and two types of control areas for 3 years.

The manipulation had a marked effect. Outsiders preferentially moved to plots augmented with nestlings, apparently judging these plots to be productive. But having extra mouths to feed forced parents to spread food more thinly, so youngsters on these plots were smaller. Emigrants picked up on both cues—and viewed the cup as half-empty rather than half-full. They fled both treatment plots at equally high rates, responding negatively to lowered quantity or quality of young. Emigrating birds “know what's going on in their own area,” Doligez explains. Immigrants, however, are at a disadvantage and may be unable to pick up on relatively subtle clues, she says.

The birds also appear to be using information from their own breeding experience when making dispersal decisions. Parents of experimentally reduced nests were more likely to fly the coop than birds with unmanipulated nests, whereas the nestling recipients were more likely to stay put.

Such findings highlight the importance of animal behavior to population biology, notes conservation biologist J. Michael Reed of Tufts University in Medford, Massachusetts. “Dispersal is often treated as diffusion in population models even though for many species it is a result of a series of behavioral decisions,” Reed says. The work may also prove useful for conservation efforts to reintroduce species to new areas, he adds.

The new study beats to press studies on colonial seabirds that show a similar use of public information. Thierry Boulinier and Etienne Dauchin of CNRS and co-workers have found that kittiwakes whose young were removed by researchers were less likely to leave when surrounded by successful neighbors than when surrounded by failed ones. And Thomas Bregnballe and colleagues at the National Environmental Research Institute in Rønde, Denmark, have found comparable patterns with cormorants. Says Pärt: “I'm convinced that this is a widespread phenomenon.”

6. COMPUTER SECURITY

# Congress Expands Cyberfellows Program

1. Jeffrey Mervis

In his 14 years at the University of Tulsa, says computer scientist Sujeet Shenoi, “I never had a student go on to work for the [U.S.] government.” But this year some two dozen have promised to join the federal workforce to safeguard the nation's computing and communications infrastructure, with 30 more banging on the door. Shenoi, who sees a terrorist attack on the country's power or communications grid as a matter of “when, not if,” couldn't be more pleased with his students' sudden shift in career plans: “I want them to make a difference before they make a buck.”

The Oklahoma students are part of a growing network of scientists and technical experts of all ages being trained in various aspects of computer security. “It's time to be as smart about cybersecurity as we are about cyberspace,” says Joseph Bordogna, deputy director of the National Science Foundation (NSF), which runs the Scholarship for Service (SFS) program (www.ehr.nsf.gov/DUE/programs/sfs). A $29 billion supplemental spending bill signed into law 2 August gives NSF an additional$19.3 million for the program, which offers 2-year full scholarships for students to earn a bachelor's or master's degree in return for at least 2 years of government service.

The scholarships are aimed at filling a years-long shortage of scientists, engineers, and policy professionals in computer security and information assurance. NSF made the first SFS awards in May 2001, averaging $2.5 million over 4 years, to the University of Tulsa and five other institutions. Last fall's terrorist attacks convinced Congress to nearly double the current year's$11.3 million budget, which in May was distributed to five more institutions. The program has also awarded more than a dozen “capacity- building” grants to universities to train faculty members at institutions breaking into the field of cybersecurity.

The intent of the 2002 supplemental funding is “to produce more professionals as quickly as possible,” explains Norman Fortenberry, head of NSF's division of undergraduate education, which manages the SFS program. A partisan fight between Congress and the White House on broader homeland security issues delayed passage of the funding bill until nearly the end of the current fiscal year and the start of the new academic year. To save time, Fortenberry says NSF is likely to “ask existing grantees” if they could grow larger rather than staging a new competition. The foundation might also consider funding highly ranked proposals that didn't make the earlier cut.

Corey Schou of Idaho State University in Pocatello hopes that his school's proposal, submitted in the hope of a supplemental bill, falls in that category. “We didn't apply in 2001 because our program is already in pretty good shape,” says Schou, who also chairs a national organization of university programs on computer systems security. “But the government's need for trained professionals is real. We also have a shortage of faculty trained to teach this stuff.” Schou points to one student who left Idaho State this winter before completing his bachelor's degree to take a corporate job that pays $83,000 a year. “I couldn't in good conscience tell him to stick around,” he confesses. “This stuff” goes well beyond computer technology. Idaho State's program includes a heavy dollop of business management training along with technical expertise, for example. Carnegie Mellon University in Pittsburgh, another member of the 2001 class, offers public policy as one of its four tracks; it has asked NSF to support a new master's degree program in information security that will be run jointly by its computer science and public policy schools. Shenoi, who came to the U.S. from India 20 years ago for graduate training, says he's looking for students who “want to make a career of federal service” and who see themselves “not just as scientists but also as public-minded citizens.” The program, he says, is also “a way to pay back this wonderful country for everything it's done for me.” 7. PATIENT RECORDS # Researchers Welcome Revised Privacy Rules 1. Jocelyn Kaiser Greater protection of patient records doesn't have to come at the expense of research. That's the message intended for scientists in final rules announced last week by the Bush Administration (www.hhs.gov/ocr/hipaa) giving patients more control over how their records are used. The “privacy rule” is a response to growing concerns about access to medical records by health care providers. However, many researchers were upset by a December 2000 rule issued by the Department of Health and Human Services (HHS), especially a provision that applied to using “de-identified” records without prior permission. The language would have stripped the records of so much demographic information, including ZIP codes and birth dates, that the data would no longer be usable for research. The modified rule includes several changes suggested by researchers. One creates a new “limited data set” specifically for research, public health, and health care that retains more identifiers, including ZIP code and birthdate. Researchers must sign an agreement stating they will keep the information secure and use it only for specific purposes. In addition, the rule no longer requires separate forms for getting informed consent and authorization to use a patient's data, and it no longer sets an expiration date for using data for a particular study. “We're encouraged that they made many of the changes we proposed,” says Jennifer Kulynych of the Association of American Medical Colleges (AAMC), one of 160 scientific societies and universities that complained about the earlier version (Science, 7 December, p. 2070). At the same time, AAMC is worried that because HHS is requiring a very detailed, binding data use agreement for the limited data set, a health care provider may have to review each agreement and research could be delayed. “The intent was a streamlined alternative” to a waiver from an ethics review board, Kulynych says, adding that the result may not be any faster. But the department feels that the new rules should appease scientists who worried that health care providers wouldn't share data. “We worked very hard with the research community to take care of what they feared would be the most chilling aspects of the rule,” says an HHS official. The rule goes into effect April 2003. 8. ALTERNATIVE MEDICINE # NIH Trial to Test Chelation Therapy 1. Constance Holden The National Institutes of Health (NIH) is putting$30 million into a major clinical trial of a cardiovascular therapy that skeptics say has no scientific rationale. Even supporters of the trial acknowledge that they aren't sure how the therapy might work and that a successful trial will leave them no closer to understanding its mechanism.

Last week the National Heart, Lung, and Blood Institute (NHLBI) and the National Center for Complementary and Alternative Medicine (NCCAM) announced joint funding of a 5-year study into whether chelation therapy can help sufferers from heart disease. Chelation, whose active ingredient is a synthetic amino acid called EDTA, binds with minerals in the body and has long been an established treatment for heavy-metal poisoning. In the past 2 decades, it has also become a widely used “alternative” treatment for arterial plaque through its purported ability to draw off calcium.

Skeptics say that the method is scientifically implausible and that the only supporting evidence is anecdotal. “The people at NIH are doing it out of political fear,” says quackery battler Saul Green of New York City, a biochemist and former cancer researcher. In addition, critics believe that no amount of negative data will persuade supporters to abandon faith in the procedure. But the heart institute's Claude Lenfant says his agency is “enthusiastic” about the project. “Only a large clinical trial can definitively answer the question of whether chelation treatment is truly safe and effective,” he says.

NIH has been under continuing pressure from Congress to support such research. There have been several small controlled trials that detected no benefit from the treatment, and 3 years ago Lenfant told legislators that he'd love to support a big one if he received the right proposal. The study, headed by Gervasio Lamas, director of cardiovascular research and academic affairs at Mount Sinai Medical Center and Miami Heart Institute in Miami Beach, Florida, will enroll 2300 patients at 100 sites.

Supporters of chelation offer two major theories for its action. One has it sucking the calcium from plaques and facilitating their dissolution, whereas the other involves EDTA's role as a powerful antioxidant. Lamas says that the calcium idea is “probably not a reasonable hypothesis” because, for one thing, EDTA is water soluble and therefore could not penetrate fatty membranes and affect plaque. He prefers the notion that EDTA acts primarily by decreasing oxidation of plaque-forming cholesterol.

Green and Wallace Sampson, a retired Stanford oncologist and hematologist, ridicule both theories. Calcium plays little role in plaque, they say, and chelation may actually promote rather than reduce oxidation. “When you chelate iron, you increase its ability to produce free radicals,” says Green. Lamas says that it's “quite a stretch” to extrapolate from basic science and that a clinical trial offers the last word. But he doesn't promise patients a miracle: At best, he says, chelation may only stem further deterioration of blood vessels.

If the trial does show a benefit, researchers still won't know the active ingredients. The usual solution actually has 10 different components, including vitamins and minerals. But NCCAM director Stephen Straus says his primary goal is to test its efficacy. If the answer is yes, he says, “then there will be time to look at mechanisms.”

9. BIODEFENSE

# Peering Into the Shadows: Iraq's Bioweapons Program

1. Richard Stone

Four years after international weapons inspectors were thrown out of Iraq, they are preparing to return with more questions

CAMBRIDGE, U.K.—At the height of the foot-and-mouth disease outbreak in Great Britain last year, Iraq asked the United Nations (U.N.) for permission to renovate a lab complex that in the 1980s made vaccine against the dreaded livestock virus. From any other country, that would have been a reasonable request. But there was nothing reasonable about the foot-and-mouth disease facility in Daura, southeast of Baghdad.

In 1996, the U.N. had compelled Iraq to destroy fermenters and other equipment there after inspectors learned that the lab had been used during the Persian Gulf War to produce the most poisonous substance known: Clostridium botulinum toxin, a single gram of which could in theory kill more than 1 million people. In the 3 months before coalition air strikes began in January 1991, Daura had churned out 5400 liters of botulinum solution. Its researchers had also tinkered with potential viral weapons.

Not surprisingly, U.N. officials spurned Iraq's bid in March 2001 to tap revenue from the U.N.'s oil-for-food aid program to upgrade Daura. Iraq can freely import foot-and-mouth disease vaccine, notes a former U.N. weapons inspector, who argues that it is therefore not essential for the country to rebuild the facility. But according to an official at the U.K. Foreign Office in London, fresh satellite images and other intelligence suggest that Iraq is thumbing its nose at the U.N. “They're refurbishing Daura on their own,” the official claims, while leaving a Potemkin village of damaged buildings for press and dignitaries.

The buzz of activity at Daura, some weapons experts say, is one of many hints that Iraq may be attempting to replenish its biological arsenal after booting out U.N. inspectors in November 1998. Intelligence reports also suggest that Iraq is smuggling in ingredients for chemical weapons and is seeking to rejuvenate its nuclear weapons program. But the ease of concocting nasty brews and concealing them in ostensibly civilian settings such as Daura sets the biological risks apart. “Biological weapons are probably the greatest immediate threat,” says Kelly Motz, an analyst on Iraq at the Wisconsin Project on Nuclear Arms Control in Washington, D.C.

Weapons inspectors want to defuse this ticking bomb. With U.S. leaders talking up an invasion to oust Saddam Hussein, Iraqi officials last week invited the head of the U.N.'s weapons inspection team, Hans Blix, to Baghdad and bid the U.S. Congress to dispatch a fact-finding mission. Both invitations were rejected, largely on the grounds that the Iraqi government refuses to accept a U.N. Security Council resolution outlining the scope and duration of the next round of weapons inspections. Then earlier this week, Iraq's information minister told the Arabic satellite news station Al Jazeera that weapons inspections “have finished.” Still, some U.N. officials perceive in the mixed signals that Iraq is preparing, however grudgingly, to permit renewed inspections in the hopes of averting war and seeing U.N. sanctions lifted. (Iraq's Ministry of Foreign Affairs did not respond to requests for comment.)

Inspectors would face a daunting challenge. Iraqi bioweaponeers honed their deception skills during the 1990s, says Graham Pearson, former director-general of the U.K. Ministry of Defence's Chemical and Biological Defence Establishment in Porton Down. “Saddam Hussein is very good at hiding things in the desert,” he says. Several U.N. weapons inspectors agreed to speak with Science, on condition of anonymity, about the inconsistencies and riddles they hope to explore on a return trip.

## A chicken coup

The effort to root out Iraq's clandestine program began when a U.N. biological team arrived for the first time in Baghdad on 2 August 1991, the anniversary of Iraq's invasion of Kuwait. From intelligence reports, they knew that the Iraqi military had explored three tried-and-true biowarfare agents: Bacillus anthracis, or anthrax; botulinum; and Clostridium perfringens toxin, which causes gas gangrene. But the U.N. team was not prepared for Iraq's apparent show of openness. A senior U.N. inspector recalls stepping off the plane that day and, to his amazement, being met on the tarmac not by some bureaucrat but by Rihab Taha, known from intelligence as a senior figure in the biological weapons program.

During the first visit, Taha and others confirmed the existence of Iraq's military program on the three bacteria and handed over a sheaf of papers purporting to show that the program was purely defensive. Although inspectors learned later that names and references to weapons had been stripped out, the heavily censored documents threw up enough red flags for investigators to infer that the Iraqis had indeed pursued an offensive program. Weapons inspectors also learned that Iraq had tried—and failed—to import three 5000-liter pathogen fermenters in 1988, as well as machines to spray-dry bacteria for storage in 1989. And there were rumors that the military had dabbled in viruses.

Although the first few years of inspections across the battle-scarred country yielded no smoking guns, they left plenty of disturbing impressions. For example, a senior inspector on the first visit to Daura in October 1991 says that before going in, they had “no indication” that the facility was part of the bioweapons program. But one of their minders for that visit was Hazem Ali, identified by intelligence as the chief of Iraq's viral weapons research program. “There was a whole bevy of attractive young ladies working as technicians—and they all clearly knew Hazem Ali,” says the inspector. “It was eerie.” At the time, however, the team could conclude only that Daura had the potential to contribute to a weapons program.

After 4 years of probing, inspectors presented Taha and others with waybills showing that Iraq had imported 39 tons of bacterial growth media in 1988 alone. “Iraq had anticipated this and had forged documents” purporting to show that the agar was distributed to hospitals and clinics for diagnostic purposes, says the senior inspector. But the Iraqi figure “was way too high,” he says, compared to what Iran and Syria bought for medical use. Even with the fraudulent paper trail, a whopping 17 tons of growth media remained unaccounted for. On 1 July 1995, top Iraqi officials admitted that the program had cooked up vast quantities of anthrax, botulinum, and perfringens spores—but continued to deny that the agents had been turned into weapons.

The biggest break came a few weeks later, when Gen. Hussein Kamel Hassan, chief of Iraq's weapons of mass destruction programs until 1990, defected to Jordan. Iraqi officials labeled Hussein, Saddam's son-in-law, a traitor and led U.N. weapons investigators to a chicken coop stuffed with documents that Hussein had supposedly squirreled away. Most of the estimated 2 million pages of materials concerned Iraqi efforts on nuclear and chemical weapons and missile development. But one crate contained revelations depicting a wide-ranging weapons program that tested dozens of microbes and produced thousands of liters of anthrax spores, botulinum toxin, and aflatoxin. “We soon began wondering what the hell else they were doing that they hadn't revealed to us,” says the inspector.

## Plagued by uncertainties

U.N. investigators felt they were just starting to get a feel for the top-secret program's broad outlines when Iraq slammed the door on them in late 1998. “We are very much limited by what Iraq has chosen to declare,” explains Pearson, author of The UNSCOM Saga (St. Martin's Press, 1999).

Most intriguing to some experts is Iraq's decision to weaponize aflatoxin, a compound derived from Aspergillus molds that grow on peanuts and other crops. Iraq declared that it had produced 2390 liters of concentrated aflatoxin, filling roughly 70% of this amount into munitions. But aflatoxin is a curious choice of weapon, as it is best known for causing liver cancer—hardly a knockout punch on the battlefield.

Some observers speculate that Iraq may have developed aflatoxin as an ethnic weapon against Kurds or Shiites. The cancer might not show up for years, but the psychological effects could be devastating, possibly emptying contaminated villages. “That would make it a true terror weapon,” says one U.N. inspector. Or aflatoxin could simply have been what one scientist calls the “pet toxin” of an Iraqi specialist.

Another mystery is Iraqi statements about work on camelpox virus at the Daura lab. Viral research runs through other connections to Daura: the apparent involvement of Hazem Ali, the viral guru; and documents about studies at Daura on rotavirus, a microbe that causes severe diarrhea, and on enterovirus 70, which causes hemorrhaging in the eyes but is not fatal. Enterovirus 70 hadn't shown up on any lists of potential agents before Iraq's research was uncovered, says a weapons inspector. But it makes sense as an incapacitating agent, he says: “You certainly can't shoot a rifle or fly a plane with your eyes bleeding.”

Although the Iraqis never produced documents on camelpox studies, analysts say that the work could have served as a surrogate for its closest known kin, smallpox virus. “We have no idea when smallpox work started, or even if it has,” says the senior weapons inspector. But U.N. investigators do know that Iraqi military scientists obtained three nonclassified papers on smallpox as a weapon in the mid-1980s, and they made smallpox vaccine, testing its potency in rabbits in 1990. “They had a theoretical if not a practical interest,” the inspector notes.

View this table:

Speculation also swirls around one obvious candidate bioweapon noticeable for its absence in Iraqi logs: Yersinia pestis, the bacterium that causes plague. The Japanese army unleashed plague-infested fleas in China during World War II, to devastating effect. “It's impossible to believe that they didn't obtain plague when they tried to obtain damn near everything else,” says one inspector. Some experts fear that Iraq has managed to hide work on plague—and possibly stocks of it—as a bacteriological ace in the hole.

## Hidden agenda

But getting to the truth requires prolonged access to suspected facilities and key scientists. Satellite surveillance and other limited intelligence suggest that Iraq, in the past few years, has been restoring labs such as Daura, building facilities, and shuffling materials from site to site. A high priority will be to investigate ostensibly civilian labs. As long as outsiders lack hard evidence, notes a U.N. inspector, “Saddam knows it presents a moral dilemma” to destroy equipment at such facilities.

A lesser priority is to check out defector accounts. Last month, for instance, The Washington Post described a defector's report of a new lab devoted to the Ebola virus. Some are dubious. “In my experience, 95% of information from defectors has come to nothing,” says the senior weapons inspector. Another U.N. investigator, however, notes that there have been at least 10 instances in which Iraqi defectors produced “exactly correct, highly valuable information.” Even the small chance that Iraq is running an Ebola lab is a chilling prospect.

Iraq could have acquired new strains of pathogens since the Gulf War despite the embargo, say experts. At the International Congress of Bacteriology and Applied Microbiology in Paris last month, one scientist asked a weapons inspector about the danger of sharing biological samples with colleagues in Iraq. The question, says a prominent virologist who was present, made him wonder how many researchers might have been—and still are—naïvely sending pathogens to Iraqi collaborators.

If Iraq can obtain the right materials, experts agree, it retains the capacity to produce biological weapons. One U.N. inspector says he has a database of 600 names associated with the bioweapons program, including 30 key scientists. The vast majority, he believes, remain in Iraq. Iraq can also make its own spray driers and fermenters; “maybe these look like tin cans,” says the senior weapons inspector, “but you don't need high tech to produce anthrax.”

As analysts wait for a diplomatic breakthrough that might let them back into Iraq, they wonder how well the West has been able to read the traces left behind by Iraqi military scientists and technologists. “They have shown themselves to be masters at exploiting discontinuities in the inspection process,” says Pearson. And the longer the current discontinuity lasts, the harder it will be to penetrate Iraq's hidden capabilities.

10. PROFILE: CLAUDIE HAIGNERÉ

# France's Highflier Comes Down to Earth

1. Michael Balter

As an astronaut, Claudie Haigneré performed experiments in space; back here on Earth she's now working to advance the research agenda of an entire nation

PARIS—On a sunny August day 6 years ago, French astronaut Claudie Haigneré and two Russian cosmonauts sat squeezed into their narrow, padded seats aboard a Soyuz space capsule, perched atop a towering rocket at the Baikonur space center. Below them, the barren steppes of Kazakhstan stretched endlessly in every direction. Although Haigneré had been training as an astronaut for more than a decade, this launch—the 1996 Cassiopeia mission to the space station Mir—was her first venture into space. Yet, Haigneré says, she was too busy carrying out her long list of prelaunch tasks to be afraid. “Then came the countdown,” she recalls. “5, 4, 3, 2, 1, lift-off! At that moment, I had a feeling of total exaltation and liberation.”

Haigneré was instantly rocketed into stardom as one of France's most exalted celebrities. Her reputation was boosted yet again in October 2001, when she moored another Soyuz capsule to its docking port on the international space station (ISS), becoming the first European woman to visit that orbiting installation. But in June, Haigneré hung up her space suit. In the new conservative government of President Jacques Chirac and Prime Minister Jean-Pierre Raffarin, Haigneré is now minister of research and new technologies, charged with getting France's stagnating research effort off the ground. As an astronaut, she was trained to pilot a Soyuz spacecraft back to Earth in case the ISS crew had to make an emergency getaway. As science chief, she has already had to make her first emergency maneuver: steering Raffarin away from slashing France's research and development budget to help deliver on a promised tax cut (Science, 9 August, p. 917).

Many researchers have privately expressed surprise that Haigneré has entered politics by accepting a government post. Haigneré says that she herself was “surprised” when, immediately after June's legislative elections, Raffarin telephoned to ask her to take the research job. But she does not see her acceptance as a partisan choice. “For me, science is not something left-wing or right-wing. We are talking about a shared national goal and about the construction of European research.”

Haigneré, known as Claudie André-Deshays before her marriage to French astronaut Jean-Pierre Haigneré last year, was born in May 1957 in the Burgundy town of Le Creusot. Her passion for space was kindled on 20 July 1969. As a 12-year-old girl camping with her family on France's Mediterranean coast, Haigneré gazed with rapture at television images of astronaut Neil Armstrong stepping onto the moon. “I looked at the image on the TV screen and then at the moon in the sky,” Haigneré recalls. “Something inaccessible, that had been just a dream, suddenly became a reality.” After graduating from high school at age 15, she entered medical school, eventually qualifying in sports, aviation, and space medicine, as well as rheumatology. Later she earned a graduate degree in biomechanics and physiology of movement, and in 1992 she was awarded a doctorate in neuroscience.

It was in 1985, when France's National Center for Space Studies put out a call for astronaut applications, that Haigneré seized the chance to realize the dream she had had as a 12-year-old. She was one of seven candidates chosen from more than 1000 applicants and the only woman. She spent 11 years on the ground as a space scientist, coordinating numerous scientific experiments on joint French-Russian space missions. When she finally got her own chances to fly in 1996 and 2001, she carried out a large number of experiments herself, including a study of how frog embryos develop in weightless conditions and the way the human inner ear responds to gravity-free conditions.

Now, as France's research minister, she has the chance to shape the nation's entire science effort. So far, she has not launched any specific initiatives, most of which will have to wait until the proposed budget is approved by the Council of Ministers in September. But in a number of speeches and policy statements, she has set down her main priorities, including stemming France's “brain drain” by making research careers in France more attractive. “We must liberate the creativity of researchers,” she says, for example, by making it easier for them to go into industry or teaching while still retaining positions with basic research agencies such as CNRS. Haigneré also wants to ensure that France takes its rightful place in the European Research Area (Science, 29 June 2001, p. 2425). And although she says that fundamental research should always be at center stage in France's science effort, she wants to do much more to boost technological innovation.

Over the past few weeks, Haigneré has been forced to fight a rearguard action against proposed cuts of 7.6% in the $9 billion France spends on nonmilitary R&D each year. The proposal, first reported by the daily newspaper Libération, which had obtained an internal government working document, came as a major shock to French scientists. Just last month, they heard President Chirac repeat—in a speech at the Airbus factory in Toulouse—his campaign promise to boost French R&D spending from its current 2.17% of gross domestic product to 3% by 2010. Although Haigneré and finance ministry officials decline to confirm specific figures, she says she “battled a great deal to get the message across” to Raffarin that such a cut was unacceptable. “The message was understood,” Haigneré says. “Things are not as good as I had wished but much better than [they were].” She sees one positive result from the controversy over science budget cuts: “Now everyone is talking about research, which they weren't before!” If the current government serves out its full term, Haigneré could expect to be research minister for the next 5 years. Since her appointment, she says she has had little time to keep up with her normally rigorous daily exercise schedule. “Being a minister takes a lot of energy, but after a period of adjustment I hope to find a rhythm so that I can stay in [physical] shape.” And she has not given up the idea of returning to space one day: “I will never close that door.” 11. ASTROPHYSICS # Gravitational Wave Hunters Take Aim at the Sky 1. Robert Irion European and American scientists are eager to build LISA, a 5-million-kilometer triangle of spacecraft tuned to murmurs from the biggest black hole encounters of all UNIVERSITY PARK, PENNSYLVANIA—The idea sounds outlandish. Launch three spacecraft into orbit around the sun and spread them into a triangle 5 million kilometers on a side. Use 1-watt lasers and 30-centimeter telescopes to monitor changes in the relative distances of metal cubes that float quietly within each vessel. Gauge variations as small as 10 picometers—yes, just one-tenth of an angstrom. “Invariably, these numbers induce giggles or outright chuckles in the back of the room,” says optical engineer Eugene Waluschka of NASA's Goddard Space Flight Center in Greenbelt, Maryland. However, the next steps are spurring Waluschka and his colleagues past the snickers: Convert the data into details about passing gravitational ripples in the fabric of spacetime, yielding exquisite insights about merging black holes that astronomers can get in no other way. That's the promise of the Laser Interferometer Space Antenna (LISA), a project that has graduated from years of blue-sky dreams to a joint European-U.S. mission gaining momentum toward a planned launch in 2011. LISA will not supersede the Laser Interferometer Gravitational-Wave Observatory (LIGO) and similar devices on the ground, which are striving to suppress the noises of Earth (see sidebar, p. 1115). Instead, LISA will observe a different part of the gravity spectrum, much as a radio telescope sees aspects of the universe that an optical telescope cannot. Whereas LIGO and its brethren try to detect high-frequency bursts from sudden events, such as off-center supernova blasts or the collisions of two neutron stars or black holes with starlike masses, LISA will tune into deep gravitational murmurs that rumble for months or years. If LIGO listens for the squeaks of cosmic mice, LISA will record intricate whale songs. The vocalists, according to theory, include binary pairs of white dwarfs in our Milky Way so numerous that they generate a persistent gravitational buzz, like the static between AM radio stations. More compelling are small black holes swooping into big ones and the mergers of supermassive black holes at the cores of galaxies—the most energetic events the universe can spawn. “That's the gold-plated objective for LISA,” says U.S. project scientist Robin Stebbins of NASA Goddard. “Ten years ago, people didn't even think there were massive black holes in most galaxies. The science of LISA keeps getting better.” That consensus echoed through a recent meeting* of more than 100 researchers here at Pennsylvania State University. Daunting tasks loom, especially getting a firmer commitment from the U.S. government to fund its share of the project. The total cost may rise from today's estimate of$600 million to more than $1 billion by launch. However, design and engineering are under way for a European Space Agency (ESA) satellite flight in 2006 to validate crucial LISA technology in the microgravity of space. “People are ready to cut metal,” says astrophysicist L. Samuel Finn of Penn State, the meeting organizer. “It's no longer an academic exercise.” ## A distant whisper Plans call for LISA's trio of identical spacecraft to orbit the sun at the same distance as Earth, trailing 50 million kilometers behind our planet. Seen from above the solar system, the center of the LISA triangle will track Earth's orbital path, while the three vessels pinwheel around that center once a year (see figure below). Each spacecraft will look like a metallic doughnut enclosing a Y-shaped assembly of instruments. More critical than the vessels are their cargo: six gold-platinum cubes just 4.6 centimeters on a side, with one floating freely inside each upper arm of the Y. These are LISA's “proof masses,” the bodies drifting peacefully through the warp and woof of spacetime. For LISA to work, each 2-kilogram cube must trace an independent orbit around the sun. The spacecraft will isolate them within vacuum chambers and provide the necessary hardware—lasers and small telescopes—to monitor fluctuations in their relative motions. As the orbiters travel through space, each will track changes in its motion relative to the other two, rather than the far harder task of measuring absolute distances. The aim is to detect cyclical expansions and compressions of spacetime created by one massive body spiraling close to another, like waves from an eggbeater whirling in batter. Each spacecraft's optical system will combine the laser beams to create a fast-changing interference pattern, with about a million dark fringes riffling past each second. Spacetime waves will alternately enhance and retard that pattern with signatures that range from subtle to dramatic. Indeed, the expected strength of gravitational radiation at LISA's low frequencies—cycles lasting tens to thousands of seconds—is the main reason why astronomers are enthusiastic about LISA. “With LIGO [in its first generation], you're really at the margins of detectability, struggling to try to dig signals out of the noise,” says astrophysicist Neil Cornish of Montana State University, Bozeman. “People assume that's true for gravitational waves in general, but it is not at all the case for LISA. We have massive signal-to-noise, far better than for some optical instruments. We will see sources within the first hour of turning on.” Still, the demands for utterly quiet spacecraft surrounding the proof masses are steep. “It's a new concept in serenity,” says physicist Bonny Schumaker of NASA's Jet Propulsion Laboratory in Pasadena, California. “She's a very sensitive girl, this LISA. She demands the most pristine environment imaginable.” Schumaker calculates that a nudge equivalent to the air pressure from a human whisper 40 kilometers away would tip a cube out of whack. Today's technology suffices to cancel such disturbances, says Karsten Danzmann, director of the Albert Einstein Institute in Hannover, Germany, and ESA's mission scientist for LISA. “There are no breakthroughs required, no miracles,” he says. “Just sound engineering practices will be enough to bring it to the sensitivity we envision.” For example, new microthrusters will keep the spacecraft centered around the masses by spitting vanishingly thin jets of ions into space. To get within a factor of 10 of LISA's requirements, ESA and NASA physicists will fly two test packages of proof masses and hardware inside the same ESA satellite, called SMART-2, in 2006. “The only thing we cannot simulate on the ground is zero gravity,” says physicist Stefano Vitale of the University of Trento, Italy, architect for the ESA package on the test flight. His colleagues repeatedly teased him about one other tough issue that LISA will face: a distinct lack of graduate students in space to fix the instruments if something goes wrong. ## Bothrodesy: It's the pits Beyond some sure-fire sources, astrophysicists differ about LISA's most promising quarries. Calculations suggest that LISA will easily see emissions from several thousand close pairs of white dwarfs in our galaxy. These partners shed gravitational energy as they whirl more tightly, often speeding around each other in mere minutes (Science, 15 March, p. 1997). Waves from another 100 million such binaries in the Milky Way will blur into an underlying “confusion” noise. A few binaries involving neutron stars or stellar-mass black holes also should trigger LISA, although their numbers are poorly known. Far more tantalizing is the prospect of precisely measuring the fantastic gravitational fields near black holes, a venture dubbed bothrodesy by astrophysicist E. Sterl Phinney of the California Institute of Technology in Pasadena. (In ancient Greece, a bothros was a sacrificial pit or well.) Phinney thinks LISA will provide the best test of Einstein's general theory of relativity by monitoring the plunges of pedestrian black holes—with 5 to 10 times the mass of our sun—into gigantic ones at the hearts of galaxies. “The smaller hole will spend its last 10,000 to 100,000 orbits in a strongly relativistic regime,” Phinney says. “There will be hundreds of these at the same time, producing the richest and most complicated signals that LISA will see.” Unraveling the end stages of those signals, which theorists cannot begin to forecast today, may be “long, messy, and painful,” Phinney says—but that's the sort of pain astrophysicists crave. Another potential signal that defies prediction is the merger of two supermassive black holes. That event, anywhere in the universe, would warp spacetime between LISA's proof masses more severely than anything else. However, it's not clear whether LISA will be lucky enough to catch one during its operational lifetime of 3 to 10 years. A recent analysis of active radio galaxies concluded that such mergers could happen about once a year (Science, 2 August, p. 753). That's more optimistic than other studies, which suggest that pairs of giant black holes could get trapped in orbits a few light-years wide for billions of years. “We know galaxies are merging today,” says astronomer Douglas Richstone of the University of Michigan, Ann Arbor. “The rate has declined, but they continue. So at any given time there should be binary supermassive black holes in the pipeline. One of them somewhere will be just about to go down the chute.” Still, the uncertainties in event rates prompt this warning from physicist Peter Bender of JILA in Boulder, Colorado: “We can't guarantee any of the massive black hole science.” Even without that prize, LISA still may expose how the cores of galaxies grow. Thanks to the expanding universe, higher-pitched gravitational signals from more frequent mergers of distant, smaller black holes will stretch into LISA's low-frequency waveband. “There's a good chance that supermassive black holes build up from the mergers of black holes with a few hundred to a few thousand solar masses,” says Bender. “LISA could do a lot to unravel that process.” Finally, some sources might not fit any category. Montana State's Cornish says that searching for novel waves in the data will be similar to code-breaking: finding signs of a pattern with little or no knowledge of its template. “With luck and ingenuity, we may be able to figure out what those unexpected sources are,” Cornish says. “But it's going to be phenomenally hard.” ## Transatlantic glow The challenges and promise of LISA have forged tight bonds on both sides of the pond. If the camaraderie at Penn State was any indication, LISA will become a model space mission for international collaboration. LISA is a cornerstone mission for ESA, which has committed to proceeding with the full project. NASA has not yet done so, although its share of the 2006 test flight appears solid. LISA program scientist Michael Salamon of NASA Headquarters in Washington, D.C., told the audience that LISA is one of the two highest priority missions for the next 2 decades in a soon-to-be-released road map for the agency's Structure and Evolution of the Universe theme. The other, called Constellation-X, is a fleet of advanced x-ray telescopes that would work in unison. Many scientists, however, worry that future NASA budgets may belie Salamon's reassuring words. “As things stand now, if you look at the numbers, NASA can't do LISA and Constellation-X simultaneously,” observes Penn State's Finn. “ESA is much further along in its commitment to LISA, and that has already caused some problems. If NASA doesn't get caught up, those problems will get worse.” ESA representatives are happy with NASA's apparent wish for a true 50-50 partnership on LISA, but they concur with Finn's assessment. “It would be extremely negative if a lack of funds [from NASA] were to induce a delay or cancellation,” says Alberto Gianolio, LISA study manager at ESA in Noordwijk, the Netherlands. “That would be a major blow to our cooperation.” But as fireflies glowed rhythmically in the heavy calm of Pennsylvania summer evenings, few scientists wanted to dwell on such pitfalls. “I've been retraining myself as a gravitational wave astronomer,” said Cornish. “We're virtual astronomers at the moment, but it becomes more realistic with each passing month.” • *4th International LISA Symposium, 19–24 July. 12. ASTROPHYSICS # LIGO: The Shakedown Continues 1. Robert Irion Thirty-three months after its dedication, the$365 million Laser Interferometer Gravitational-Wave Observatory (LIGO) will begin its first “science run” 23 August. For 17 days, physicists intend to keep their beams of light on target at LIGO's sites in Louisiana and Washington state (Science, 21 April 2000, p. 420) for simultaneous data collection. “If we can keep running, we'll reach interesting scientific limits,” says deputy director Gary Sanders of the California Institute of Technology (Caltech) in Pasadena. Don't expect startling news, however: According to the latest estimates, LIGO's detectors are still 100 to 1000 times more jittery than their blueprints demand. Only a nearby supernova might rise above the background noise.

Not to worry, Sanders says: “It's what we expected during commissioning. We've made steady progress in reducing the noise since we turned on.” Adds Caltech colleague Ron Drever, a pioneer of the laser interferometry technique: “I think people are jolly well pleased.”

Many of LIGO's systems have surpassed expectations. The immense vacuum chambers, spanning 8 kilometers of steel tubing at each site, hold tight with nary a leak. The lasers are stable, snapping back to their “locked” focus through the optics after each engineering adjustment.

Still, figurative and literal bumps have jostled the team. The worst problems are local traffic and commercial logging in the woods around the Louisiana site. Damp soil channels the vibrations of a felled and dragged log at just the right frequency—about 2 hertz—to jiggle LIGO's nested sets of springs. “When a tree falls and there's nobody there, we hear it,” says physicist David Shoemaker of the Massachusetts Institute of Technology in Cambridge. So for now, most usable data come only at night.

To compensate, the team has accelerated testing of a new system that will sense and actively suppress vibrations. Plans call for engineers to install that fix after a second science run in December and early January has ended. By mid-2003, “we expect that we will be within a stone's throw of the target sensitivity for our first-generation detector,” Shoemaker says.

That timeline forces a tough decision. LIGO's leaders must apply to the National Science Foundation for “Advanced LIGO,” a $100 million upgrade for a stronger laser, better optics, quieter suspension, and more sophisticated protection from vibrations. A successful proposal this year would keep the project on track for an upgrade starting in 2006. In the minds of most observers, that's the only realistic chance for LIGO to see gravitational waves, as Advanced LIGO would search a volume of space 10,000 times larger than LIGO 1. However, the team is mulling whether to wait a year, pending progress in the first science run. “We're in this dilemma where we may propose the construction of an upgrade before we have really shown that we've fulfilled the promise of the first round of detectors,” says Sanders. LIGO hasn't yet won the hearts of astronomers, but fewer disparage the project now than early in its history. “I think astronomers are ignoring LIGO,” says Douglas Richstone of the University of Michigan, Ann Arbor. “They just don't think it will get sources. But if it does,” he adds with a smile, “it will inspire a great deal of interest.” 13. DRUG TARGETING # Breaking Down Barriers 1. Greg Miller The brain is guarded by a nearly impenetrable cellular barrier, but researchers are looking for ways to blast through this wall of defense—for the brain's own good Any attempt to develop a new therapy for brain disease must deal with a vexing impediment. Neuroscience textbooks bury it in the appendix, Ph.D. programs give it a cursory treatment, and pharmaceutical companies have tried to ignore it. But the blood-brain barrier is a stubbornly real obstacle for potential drugs against many disorders of the central nervous system (CNS). The barrier is built into the densely packed capillaries that feed the brain. The endothelial cells lining these tiny tubes are cemented together by junctions so tight that even ions have a hard time squeezing through. Transport proteins in the endothelial membrane act like bouncers, ushering sugars, amino acids, and other important players into the brain and making sure any undesirables that manage to slip in are promptly shown the way back out. Unlike the far leakier capillaries elsewhere in the body, brain capillaries form a superfine filter that protects the brain from any riffraff, such as toxins and viruses, circulating in the blood. But this barrier also keeps out many would-be therapeutic agents. Molecules can cross the barrier only if they are escorted by the choosy transport proteins or if they are small enough and lipid-friendly enough to pass through one side of the endothelial cells and out the other. As a general rule, that means that only lipophilic molecules smaller than about 500 daltons can cross from blood to brain. But many drugs that show promise in animal studies for treating CNS disorders are much bigger—commonly weighing in at tens or hundreds of kilodaltons. Now a small but growing community of predominantly academic researchers is studying the blood-brain barrier with an eye to improving drug delivery to the brain. Their research ranges from trying to better understand the basic biology to creating molecular Trojan horses that sneak anything from neural growth factors to therapeutic genes into the brain. “It's an area that's been underresearched, and its significance hasn't been recognized,” says David Begley, a neurophysiologist at King's College London. “There must be millions of good drugs sitting in pharmaceutical company stores simply because they can't be delivered,” he adds. Drug companies have long been reluctant to tinker with the blood-brain barrier, but a recent surge in academic research on the subject might show them the way. Traditionally, pharmaceutical companies have focused their efforts to develop CNS drugs on small molecules. It's worked for a few disorders, says William Pardridge, a blood-brain barrier researcher at the University of California, Los Angeles. He rattles them off: epilepsy, chronic pain, mood disorders, and schizophrenia. But a host of other disorders have proved far more recalcitrant, including Alzheimer's disease, Huntington's disease, stroke, and brain cancer. Pardridge believes that the small-molecule approach to CNS drug development is fundamentally flawed. “It's based on two misconceptions: There's the misconception that you can come up with a small molecule for any disease and the misconception that all small molecules cross the blood-brain barrier.” In an editorial this year in Drug Discovery Today, Pardridge argued that the industry's fixation on small molecules is the reason there are relatively few CNS drugs on the market. The global market for CNS drugs, at$33 billion in 1998, was roughly half that of the global market for cardiovascular drugs, even though in the United States nearly twice as many people suffer from CNS disorders as from cardiovascular disease. The reason for this lopsidedness, he claims, is that more than 98% of all potential CNS drugs don't cross the blood-brain barrier.

## Sugary solution

The first clue that a barrier separated the brain's blood supply from the body's came in the late 1800s. German bacteriologist Paul Ehrlich found that dyes injected into the body stained every organ except the brain. A student of his, Edwin Goldmann, later hit on the other side of the story: Dye injected into the cerebrospinal fluid stained the brain but not other organs. The anatomical basis of the barrier remained a mystery until the late 1960s, when electron microscope studies revealed the tight junctions between the endothelial cells lining the brain's capillaries.

If not for the blood-brain barrier, the capillaries—the branches of which meander more than 600 kilometers through the human brain—would provide a fantastic drug-delivery system, designed as they are to bring oxygen and nutrients virtually to the doorstep of every cell in the brain.

One of the earliest techniques to circumvent the barrier for therapeutic purposes—and the first to be used in humans, more than 20 years ago—was developed by neuroscientist Stanley Rapaport and neurosurgeon Edward Neuwelt, then at the National Institutes of Health in Bethesda, Maryland. The idea behind the approach is to break the barrier down—temporarily—by injecting a sugar solution into arteries in the neck. The resulting high sugar concentration in brain capillaries sucks water out of the endothelial cells, shrinking them and opening gaps between cells. In current practice, the effect lasts 20 to 30 minutes, during which time drugs that wouldn't normally cross the barrier have a free pass.

Neuwelt, now at Oregon Health & Science University in Portland, still uses the technique, called blood-brain barrier disruption (BBBD), primarily to deliver chemotherapy drugs to brain cancer patients. Animal studies suggest that the method increases drug delivery to the brain 10- to 100-fold over injection into the neck arteries without the sugar solution. BBBD has been used to treat hundreds of patients and is currently in phase I and II clinical trials at nine U.S. institutions, as well as one in Canada and one in Israel. The National Cancer Institute has twice turned down proposals for a controlled phase III trial of the technique, however. Designing a phase III BBBD trial is exceptionally tough, Neuwelt says, because appropriate brain cancer patients are relatively rare and because only a few institutes are qualified in the technique.

In the laboratory, Neuwelt's team is working on other applications for BBBD, including using it to deliver chemotherapy drugs attached to antibodies that home in on cancer cells. The researchers are also investigating using BBBD to shuttle therapeutic genes into the brain—either to make tumor cells more susceptible to chemotherapy or to replace defective genes in neurodegenerative disorders.

But others see limitations to the BBBD approach. For one, it is invasive and requires considerable expertise, says Raymond Bartus, senior vice president of life sciences research and development at Alkermes, a biotech company in Cambridge, Massachusetts. Bartus and colleagues pioneered an alternative method for opening the barrier using intravenous injections of a compound called RMP7, which binds to receptors on the surface of endothelial cells and kicks off a biochemical cascade that loosens the tight junctions.

Although the technique showed promise in early clinical trials, Bartus says Alkermes has more or less given up on the project. The method would be useful only for a small number of patients, such as those suffering from brain cancer, and thus wasn't worth continued investment, he says. BBBD has a similarly limited scope, he points out.

Both procedures are unacceptably risky, according to Pardridge. Disrupting the barrier—even for brief periods—leaves the brain vulnerable to infection and damage from toxins, he explains. Even some substances that circulate harmlessly through the peripheral bloodstream, such as the protein albumin, can have deleterious effects if they get into the brain, he says.

## Escort service

Pardridge and others are investigating another option. They hope to take advantage of the transporter proteins in the endothelial membrane. The number of known transporters has grown rapidly in recent years, says William Banks, who studies blood-brain barrier physiology at Saint Louis University in Missouri. And that has changed the way people view the barrier: “It's not only a wall but a gate.”

Pardridge and his team are looking for ways to fool the gatekeepers. Many of their initial studies have focused on a receptor that binds transferrin, a protein critical for transporting iron. When an iron-laden transferrin molecule on the blood side of a capillary binds to the receptor, a dimple forms in the endothelial cell membrane and the transferrin gets sucked into the cell inside a tiny bubble of membrane. This bubble makes its way to the other side of the pancake-shaped cell (a mere 300 nanometers away), where it fuses with the membrane and spits the transferrin out into the brain to deliver its iron.

In studies with rats, Pardridge's team has managed to trick transferrin receptors into ferrying an unusual cargo across the barrier: molecules called neurotrophins that protect neurons and encourage them to grow.

One day, says Pardridge, such a strategy could protect the brain from damage caused by strokes. Many of the few-dozen known neurotrophins have been shown to reduce stroke damage in animals, but the molecules are too big to pass the barrier on their own. The only way to deliver them into lab animals' brains has been to drill a hole in the skull and inject the compounds—a procedure widely judged too invasive and impractical for humans.

Pardridge and colleagues tested a new hybrid transport method: They attached a neurotrophin called brain-derived neurotrophic factor to antibodies that attach to the transferrin receptor and prompt endothelial cells to engulf them. The researchers simulated a stroke in rats by blocking an artery to the brain and then gave the rats a shot of the antibody-neurotrophin combo. The injection reduced the volume of cerebral cortex killed by the mock stroke by up to 70% when given 1 or 2 hours later, the team reported in June 2001 in Stroke. And in the May issue of The Journal of Pharmacology and Experimental Therapeutics, the team reported that delivery of another neurotrophin via the same antibody has an even more powerful protective effect.

The team hopes to use this method to ferry other cargo across the blood-brain barrier, including material for gene therapy. Earlier efforts, using viruses to deliver genes into the human brain, spurred potentially deadly immune responses. But the researchers can sneak genes across, sans virus, by enclosing genetic material in membrane bubbles coated with antibodies tuned to the transferrin receptor. The approach works with a single IV injection, and in the July issue of Molecular Therapy, Pardridge's team reported doubling the life-span of mice with brain tumors by using the technique. In that case, the team delivered so-called antisense messenger RNA to block expression of the gene for epidermal growth factor, which is often hyperactive in especially nasty brain tumors.

The transferrin receptor isn't the only baggage handler pulling molecules into the brain. Pardridge has also developed a drug-delivery system based on an insulin receptor, and tests in monkeys suggest it is 10 times more effective than the transferrin receptor system, he says. And he has begun a blood-brain barrier genomics project to identify more transporters. His team is isolating the capillary endothelium from rat and human brain tissue, identifying genes specific to the blood-brain barrier, and fishing through them for transporters. So far, about 15 transporters are known; Pardridge estimates there are at least 50, all told.

The advantage of the transporter approach, many researchers say, is its flexibility. “You could get almost anything across in principle,” says Tom Davis, a pharmacologist at the University of Arizona in Tucson. But while Davis agrees the approach is promising, he cautions that therapies based on it would probably be expensive, given the “exotic chemistry” involved.

Another stumbling block, argues Jörg Kreuter, a pharmaceutical scientist at Johann Wolfgang Goethe University in Frankfurt, Germany, is that only limited amounts of drug can be delivered this way. The traffic is limited by the number and carrying capacity of the transporters, as well as by the number of drug molecules that can attach to each antibody, he says. “You have a very big engine and very few passengers,” he says. That might be fine for neurotrophins, which are effective in low doses, says Begley, but it might not work as well in cases where large quantities of the drug are needed.

## Nanotricks

Looking for another entryway into the brain, Kreuter, Begley, and others have found that polymer nanoparticles can be used to sneak drugs across the barrier. The tangled balls of polymer have grooves and pockets that can be stuffed with drugs. Nanoparticles are typically hundreds of nanometers across—big enough to transport almost any molecule, even strips of DNA, Kreuter says. He has used the particles to deliver the chemotherapy drug doxorubicin, for instance, to rats with brain tumors. Nanoparticle drug injections cured up to 40% of the rats, which normally die in 10 to 20 days. Six months later, the tumors were gone—only a bit of swelling and scar tissue remained.

How nanoparticles cross the barrier is a matter of some debate. Pardridge argues that the technique is something like BBBD in disguise. Detergents added to keep the particles from clumping together have been shown to disrupt the barrier, he says, either by loosening the tight junctions between the endothelial cells or by dissolving their membranes. But Kreuter and others argue that they have good evidence that the particles are sucked up by the cells without disrupting the barrier, although they acknowledge the process isn't fully understood. “We don't know everything about the mechanism, but we can measure the pharmacological effects,” Kreuter says.

## Worth the investment?

Many researchers say blood-brain barrier research is just gathering momentum. A recent Gordon conference in Tilton, New Hampshire, attracted almost everyone working in the field—150 researchers from 11 countries. It wasn't a huge crowd, as far as biomedical meetings go, but was larger than similar gatherings in past years. “That's been the problem in blood-brain barrier research—the community is small but the problem is big,” says Tom Jacobs, scientific leader of the Neural Environment Cluster, an advisory group at the National Institute of Neurological Disorders and Stroke (NINDS) in Bethesda, Maryland.

Although enthusiasm runs high among those in the field, many researchers are frustrated that the pharmaceutical industry hasn't expressed much interest in their work—and isn't doing much of its own. “I'm aware of no company that has a focused R&D group for developing technology for getting drugs across the blood-brain barrier,” says Davis. Others, including some who work in industry, agree. “It's not only surprising, it's not cost effective,” Davis says. “You spend tremendous amounts of money developing drugs. You should also spend money on technology to deliver the drugs to the organ of choice.”

This neglect in part reflects a historic bias toward developing small-molecule therapeutics. “Industry has more expertise with small molecules,” says Andreas Reichel, a pharmacologist at Schering in Berlin. “People are more comfortable working with what they know.”

But Reichel says awareness of the blood-brain barrier is growing in industry. Companies now screen compounds earlier in the drug development process to see if they will cross the barrier. In vitro and in silico models of the barrier have helped avert many dead ends, he says.

Still, for the most part, industry has taken a wait-and-see attitude toward delivery systems for large molecules. “If there were a clear high-value target that was really valuable for patients, we'd make the investment,” says Jim Baxter, executive director of pharmacokinetics dynamics and metabolism at Pfizer's Groton, Connecticut, research facility. The cost of drug development is so high, he says, that putting additional money into sophisticated delivery systems for large molecules is too great a risk. “Until we saw that it's the only way, we'd rather try the more precedented [small-molecule] methods.”

Pardridge scoffs at this logic. “It's completely ludicrous,” he says: A neuroprotective drug for the treatment of stroke would be a “high-value drug.” For every disorder that hasn't responded to small-molecule drugs, he says, “they'd have a blockbuster drug.”

But even he acknowledges that more research is needed to get large-molecule drug-delivery systems into the clinic. For now, everyone agrees most of that work will have to come from the academic side. Although some researchers complain that getting funding for blood-brain barrier research has been tough, the situation may be improving, at least in the United States. An initiative to be announced this fall by NINDS will include money for blood-brain barrier research, says Jacobs. He says researchers in this field have made a compelling case that understanding the barrier is critically important for understanding and treating CNS disease.