# News this Week

Science  14 Nov 2003:
Vol. 302, Issue 5648, pp. 1126
1. NEW FACILITIES

# ITER Tops DOE's List of Next Big Science Projects

1. David Malakoff,
2. Adrian Cho*
1. With reporting by Robert F. Service. Adrian Cho is a freelance writer in Grosse Pointe Woods, Michigan.

Like the mythical phoenix, a once-dead fusion energy project this week rose from the ashes to become the U.S. Department of Energy's (DOE's) highest scientific priority. The $5 billion International Thermonuclear Experimental Reactor (ITER) topped a list of 28 major new research facilities that DOE hopes to complete over the next 20 years, ahead of projects ranging from a biomolecules factory to a new high-energy particle smasher. “These facilities and upgrades … will revolutionize science,” Energy Secretary Spencer Abraham predicted on 10 November as he unveiled the rankings (see table). View this table: Researchers and policy experts are applauding DOE for tackling the difficult task of setting priorities across disparate disciplines, saying that a clear road map could help inspire Congress to boost DOE's stagnating$3.3 billion science budget. And although budget watchers say the plan's funding assumptions may be a bit rosy, many scientists are just happy that projects they support made the list. “I take the attitude that all the boats will rise with the plan,” says physicist Frederick Gilman, who backs a $6 billion linear collider that finished 13th, to the disappointment of some supporters. Gilman is a professor at Carnegie Mellon University in Pittsburgh, Pennsylvania, and head of DOE's high-energy physics advisory panel. The list is the brainchild of DOE science chief Raymond Orbach, who joined DOE 2 years ago. Late last year he asked his seven division heads to identify facilities costing at least$50 million that the United States needs to build over the next 20 years to capture “world scientific leadership.” Their 46 nominations went to outside advisory panels, which ranked the entries and added seven projects to the mix. Orbach then made the final cuts, leaving 25 proposals off the list. Abraham says that DOE rejected a more traditional committee approach “because committees, despite their best efforts, are notorious for delivering compromise documents.”

Orbach assumed that Congress would follow a plan it is now considering to boost DOE's science budget by about 60% by 2008, followed by 4% annual increases until 2023. “No one believes we will get all 28 machines,” says a senior DOE official. “But this is a recipe [for Congress to follow] if there were additional funds available.” The official briefed reporters on the condition that he not be identified.

ITER's top ranking is vindication for fusion researchers, who had watched the United States pull out of the collaboration in 1998 due to cost concerns, only to rejoin it last year. “This is fabulous,” says Robert Goldston, a fusion physicist at Princeton University. “To see fusion energy science front and center is wonderful.” The U.S. is one of six partners in the project and is expected to foot 10% of the bill for the $5 billion experiment. Japan, France, and Spain have offered sites for the facility. Computer scientists are pleased that a supercomputing initiative captured the silver medal. Researchers hope to identify key classes of computationally intensive problems, such as modeling Earth's climate, and then use interdisciplinary teams to tailor-make supercomputers to solve them. Each such facility would cost upward of$100 million, says Margaret Wright, a computer scientist at New York University. The ranking is “a tribute to Orbach's vision to treat high-end computing as a facility,” says Wright, who chairs DOE's Advanced Scientific Computing Advisory Committee.

Four projects tied for third place on the list, including a proteomics facility that signals DOE's increasing emphasis on “big biology,” the senior official said. The Protein Production and Tags facility would mass-produce tens of thousands of proteins and matching chemical “tags” that are essential to unraveling chemical signals within cells, says Richard Smith, a proteomics researcher at Pacific Northwest National Laboratory in Richland, Washington. “It's a capability that's difficult to implement in small laboratories,” he says.

The $900 million Rare Isotope Accelerator, also ranked third, would generate beams of rare and radioactive atomic nuclei in a bid to reveal the intricacies of the strong force that binds protons and neutrons into nuclei. The$220 million Linac Coherent Light Source would use a high-energy particle beam to make an x-ray laser for studying matter at the Stanford Linear Accelerator Center in Menlo Park, California. A $1 billion satellite known as the Joint Dark Energy Mission, to be built in collaboration with NASA and launched in 2014, rounds out the bronze medalists. The news was less encouraging for particle physicists, as none of their experiments cracked the top 10. Gilman says that pending design issues left the massive collider in the middle of the pack. But Frank Wilczek, a theoretical physicist at the Massachusetts Institute of Technology, worries that it may rank too low to ever be built. “One's initial reaction is that lists like this never get down to number 13,” he says. The senior official said the list may not influence the Bush Administration's spending plans until next fall, when the White House starts preparing its 2006 budget request. (Congress was expected to approve DOE's 2004 budget this week, and its 2005 request is being vetted by budget officials.) Until then, researchers are likely to be working on behalf of their pet projects. 2. PRIMATE RESEARCH # U.K. Mulls Military 'Safe Haven' for Embattled Center 1. Richard Stone CAMBRIDGE, U.K.—Escalating costs appear to have doomed plans by the University of Cambridge (UC) to build a controversial primate research center on the edge of this academic city. But Science has learned that the U.K. government is weighing an innovative alternative: to put the$50 million lab at a tightly guarded military facility, where it would presumably be sheltered from protests by animal activists.

UC officials want a center that would bring all of the university's primate research under one roof and allow it to expand work on potential treatments for everything from schizophrenia to Alzheimer's disease. The proposal won support from the highest echelons: Prime Minister Tony Blair, for example, said it was in the national interest, and Britain's top biomedical officials last month issued a joint letter defending primate research. But local authorities have twice refused to grant planning approval on the grounds that protesters might snarl traffic and perhaps harm staff or themselves. A public hearing last December on these issues erupted into a debate on animal research (Science, 6 December 2002, p. 1862). The proposal now rests with the U.K.'s deputy prime minister, John Prescott, who has yet to issue a ruling.

In the meantime, support for the center has eroded within UC. Last month the university's oversight board complained that it had not been told early on that primates would be used for research at what has been billed as a neuroscience center. In a 29 October response, the University Council noted that it is not obliged to detail scholarly or experimental work in proposed buildings or facilities.

More telling, however, was the council's acknowledgement of a “significant shortfall” in construction funds and operating expenses. Over the past 5 years the center's price tag has risen 33%, in part from the cost of implementing tighter rules on lab animal care and safety regulations. The university also anticipates being millions of pounds in the red this year, with further deficits next year. UC's top official, Vice Chancellor Alison Richard, was not available for comment, but other officials have signaled that the university would need additional outside funds to proceed if Prescott were to approve the project. “The mounting costs of security and animal welfare … must be addressed at a national level,” says virologist Anthony Minson, UC's pro-vice-chancellor.

With a sword of Damocles hanging over Cambridge, one solution would be to whisk the center to a safe haven. The U.K. government is now mulling a proposal to build a national primate lab at the Defence Science and Technology Laboratory's Porton Down facility. Backers of the confidential proposal argue that the biodefense lab in Wiltshire would offer a secure environment for researchers from across the United Kingdom. Portland Down already has a macaque breeding facility for the academic community that opened last August.

UC has distanced itself from the proposal, although it could in principle operate the center. “We have no intention of siting the facility elsewhere,” says a spokesperson. Nevertheless, observers say, the new site would provide UC officials and Cambridgeshire authorities with a face-saving alternative and ensure that scientists have a place to conduct primate research.

3. # The Eyes of Texas Are Upon Him

LUBBOCK, TEXAS—The trial of Thomas Butler continued here this week, with prosecutors pressing their case in a court battle that could set legal precedent for life scientists operating under antibioterrorism laws. For daily online coverage, log on to sciencenow.sciencemag.org.

4. CONDENSED-MATTER PHYSICS

# Competing Research Teams Create Long-Sought State of Matter

1. Charles Seife*
1. With reporting by Daniel Clery.

After years of trying, scientists have coldly gone where no one has gone before. Two groups of physicists—one in Austria and one in the United States—have created a Bose-Einstein condensate (BEC) out of fermions rather than the usual bosons. The achievement opens up a whole new area of research that might soon help scientists understand the mystery of high-temperature superconductivity.

“It's a big step,” says Randall Hulet, a physicist at Rice University in Houston, Texas. “It's really very different and a major new direction.”

The excitement stems from the difference between bosons and fermions, based on a quantum-mechanical property known as “spin.” Spin comes in packets that physicists reckon in multiples of 1/2. Fermions are particles (such as the electron) that have a half-integer spin (1/2, 3/2, 5/2, and so on), whereas bosons are those (such as the photon) that have an integer spin (0, 1, 2, and so on).

The distinction gives bosons and fermions very different properties. No two fermions can have the same quantum state: They can't have the same properties and be in the same place at the same time. Bosons can. So, when, in 1995, physicists cooled a dollop of rubidium and sodium atoms—bosons, the lot—to a few billionths of a degree above absolute zero, many of the atoms sank down into the same low-energy quantum state, becoming, to all intents, a single giant ensemble atom: a BEC. With fermions such as potassium-40 or lithium-6, on the other hand, each particle must have a slightly different property, even at the very coldest temperature.

But physicists know an easy way to overcome that obstacle. They can turn fermions into bosons by pairing them up. Two half-integers make a whole integer, fermion pairs act as bosons, and all of a sudden the gas can condense into a BEC.

Now, for the first time, two groups have made this happen. The first, based at the University of Innsbruck in Austria, loaded lithium-6 atoms into an optical trap and cooled them carefully while applying a steady magnetic field to encourage the fermions to bond. “Somewhere in the final stages of cooling” the fermions pair up, says team leader Rudolf Grimm. The team describes its work in a paper published online by Science this week (www.sciencemag.org/cgi/content/abstract/1093280).

The second group, based at the Joint Institute for Laboratory Astrophysics (JILA) in Boulder, Colorado, used a slightly different technique. Led by physicist Deborah Jin, the JILA team first cooled and trapped potassium-40 atoms and then applied a magnetic field. By varying that field, Jin altered how strongly each atom attracted its neighbor—and induced them to form pairs and condense into a BEC, as the researchers report in a paper they posted on the arXiv preprint server (http://www.arxiv.org/) and have submitted to Nature.

“Nobody's seen, prior to this, a molecular BEC,” says Hulet, who adds that it could lead to a radical new way to study superconductivity (Science, 8 August, p. 750). A similar process to Bose-Einstein condensation happens in superconductors—where fermionic electrons form loosely bound bosonic “Cooper pairs,” as described by Bardeen-Cooper-Schrieffer (BCS) theory. Because the teams can change the binding strength of the Cooper pairs, they can, in theory, study the region in between loosely bound BCS pairs and the tightly bound bosons of BECs—the realm where high-temperature superconductors are thought to reside. “You can study the regime between the BEC and BCS pairing limits,” says Hulet. “You can go in between. It's a regime nobody's seen.”

5. SPACE SCIENCE

# Europe Forced to Turn Blind Eye to Far-Flung Planets

1. Daniel Clery

CAMBRIDGE, U.K.—Bowing to harsh budgetary realities, the European Space Agency (ESA) has aborted plans for the Eddington observatory, a $205 million effort to hunt for planets beyond our solar system. The agency has also canceled the first-ever visit to Mercury's surface by axing the lander from the upcoming BepiColombo mission. The cuts, imposed by ESA's Science Programme Committee (SPC), are the unhappy consequence of a string of recent unforeseen events. “The wheels fell off the wagon,” admits ESA science chief David Southwood. That's scant solace to aggrieved researchers involved in Eddington. “It's a bloody awful decision,” says Barrie Jones of the Interdisciplinary Centre for Astrobiology at the Open University in Milton Keynes, U.K. The Eddington mission would have looked for planets by detecting dips in the brightness of parent stars as planets pass in front. It also would have probed for subtle oscillations in a star's output that give clues to its interior. Two problems forced the crisis, Southwood says. First was the failure last December of an upgraded version of Europe's Ariane 5 launcher. Grounding all flights delayed the launch of two science missions, the Rosetta comet chaser and the SMART-1 moon mission, and boosted their costs by$90 million. Second was the failure of some ESAmember countries to pony up money promised for building instruments for upcoming missions, most notably the twin observatories Herschel and Planck. Altogether, the science program found itself facing a $115 million shortfall for 2003, and a loan from another ESA division gave only a temporary respite (Science, 23 May, p. 1213). The science program reassessed its priorities, which it then presented to ESA's Space Science Advisory Committee. The panel's key recommendation was to halt work on Eddington until the financial picture improves. Project researchers mounted a last-ditch effort last week to save the mission, calling it “essential.” But it was in vain. The program committee, which includes representatives from ESA's 15 member states, went beyond the advisory panel's suggestion and canceled it outright. Southwood explains that delaying Eddington would have been impractical. The observatory was to have been built right after Herschel and Planck, allowing ESA to use the same production team and similar components. Restarting at a later date would require rebuilding all that from scratch at a much higher cost. “I don't want the community to hold out false hope,” says Southwood. The decision has left the Eddington team fuming. Ian Roxburgh of London's Queen Mary College, the project's coordinating scientist, accuses ESA of adopting a last-in, first-out policy to exclude Eddington. “I don't respect an institution that makes decisions on this basis,” he says. Roxburgh and his colleagues are planning to lobby politicians and members of the ESA Council to reverse the decision. Mercury buffs also got bad news last week when the SPC downsized the BepiColombo mission to Mercury, last visited by Mariner 10 in the 1970s. Axing the mission's lander was a difficult decision, says Southwood. “It would have been a nice first” to land on Mercury, he says. But the lander's design costs would have been substantial, because “it's tough to get down on that planet.” ESA also decided to delay BepiColombo's launch from 2010 to 2012, hoping to avoid a budget crunch at the end of the decade by spreading out the costs. 6. TEACHING EVOLUTION # Texas Resolves War Over Biology Texts 1. Constance Holden Fending off attempts to weaken references to evolution, the Texas State Board of Education last week approved a list of biology textbooks that scientists believe do justice to Darwin's theories. The 11-4 vote allows schools to choose from among 14 high school and advanced placement biology books. Scientists had been on high alert in response to complaints by advocates of “intelligent design” that pro-Darwin “inaccuracies” appear in books under review in Texas, the nation's second-largest buyer of textbooks after California. Shortly before the school board vote, 550 Texas scientists and teachers urged the 15 members to maintain “high standards of science”—and they think the appeal got through. “The publishers have done a terrific job of holding the line and keeping the creationists at bay,” says Skip Evans of the National Center for Science Education in El Cerrito, California. Complaints about the books had come from the Discovery Institute (DI) and its Texas allies, Texans for Better Science Education. They were unhappy with descriptions of presentations such as the famous “peppered moth” experiments (changes in coloring offer a classic example of microevolution); illustrations indicating that early human and animal embryos look a lot alike; and the “bogus claim” that human embryos have “gill slits.” In response, some textbook publishers made minor changes, including replacing embryo drawings with photos and dropping the term “gill slits.” One also eliminated the assertion that Darwin's theory is the “essence of biology.” Now both sides seem to be happy. DI officials praised publishers for correcting “a number of embarrassing errors that overstate the evidence for evolutionary theory.” Scientists, for their part, say the changes pose no problems. The embryo drawings were old and inaccurate, says Evans. And whether or not books talk about gill slits is simply a matter of “terminology.” The settlement doesn't quite bring peace to Texas. The author of an environmental text is suing the school board over its rejection of his book after a conservative group called it “anti-American.” And the adoption of health textbooks next year is expected to trigger a battle over sex education. In the meantime, Darwin's defenders are getting a respite. 7. EARTH OBSERVATION # Loss of Japanese Satellite Narrows Window on World Climate 1. Dennis Normile TOKYO—Satellite oceanographer Timothy Liu always knew that space science was risky. But he wasn't prepared for such a string of bad luck. Japan's Advanced Earth Observing Satellite (ADEOS-I) was carrying a scatterometer developed by Liu and his colleagues at NASA's Jet Propulsion Laboratory (JPL) in Pasadena, California, to measure ocean wind speeds when the satellite stopped sending back data in 1997. Late last month the same thing happened to its$630 million successor, ADEOS-II, which was carrying another one of Liu's scatterometers. It's one of five instruments—three Japanese and one French—that are now lost to science.

The failure has raised questions about the reliability of Japanese spacecraft, the fate of several satellites in the pipeline, and the impact on climate change research. “The loss of any environmental observation capacity is an extremely serious matter,” says Taroh Matsuno, director-general of Japan's Frontier Research System for Global Change in Yokohama.

Designed to last 3 years, ADEOS-II (also called Midori-II) was launched in December 2002 and has spent the past 10 months gathering data on ocean winds, cloud cover, and the distribution of aerosols in the atmosphere for both weather forecasting and environmental studies. Satellite operators first noted problems with data transmissions on 25 October; by 31 October they had thrown in the towel.

The episode marks the first major crisis for the Japan Aerospace Exploration Agency (JAXA), recently formed through the merger of three space-related agencies. The stakes are high: JAXA has more than a dozen satellites planned for launch through 2007. Although JAXA's predecessors—the National Space Development Agency of Japan and the Institute of Space and Astronautical Science (ISAS)—performed much of the development work, the same outside contractors make components for both groups.

“If the [ADEOS] loss is traced to one of these common components, we would have to be very careful” about verifying its reliability, says JAXA's Takeo Kosugi, a solar physicist formerly at ISAS. Failing to find the cause could be even worse, he says, because it would cast a shadow on all the satellites in the pipeline. Yuriko Koike, Japan's environment minister, has already hinted that she might try to place future ministry instruments on foreign satellites. The ministry lost ozone layer-observing instruments on both ADEOS missions.

In the meantime, scientists are unhappily coping with the loss of their instruments and their reduced capacity to understand ocean-atmosphere interactions. “It's a big, big disappointment,” says Liu, a senior research scientist at JPL. He had just released the first 6 months of data from ADEOS-II and was enjoying the 6-hour monitoring cycles made possible by it and a scatterometer he cobbled together from spare parts for a 1999 launch of a U.S. mission after ADEOS-I went awry. The third instrument is already 2 years past its scheduled life span, and Liu doesn't know how long he'll be able to keep together his 40-member scientific team now that ADEOS-II is history.

Likewise, Matsuno's organization had hoped to use ADEOS-II data to refine a new high-resolution model of cumulus cloud formation and related storms in the tropics. The lack of detail of these ephemeral processes in existing models “is a bottleneck for both weather forecasting and climate modeling,” he notes. The work will go more slowly without help from ADEOS-II.

A JAXA public relations spokesperson says the agency is hoping “that a quick and convincing determination of the cause of the failure will start to restore public trust.” JAXA has upward of 100 engineers on the investigation, but reaching a definitive conclusion could take months.

8. NEUROSCIENCE

# Brain Model Puts Most Sophisticated Regions Front and Center

1. Laura Helmuth

The prefrontal cortex is the part of the brain that brainstorms. Brain imaging studies show that it's activated when someone solves problems, juggles competing hypotheses, or remembers to stick to a tight schedule. Given this variety of responsibilities, many cognitive neuroscientists have overheated their own prefrontal cortices trying to figure out how this area of the brain is organized. Now a study on page 1181 proposes that the prefrontal cortex has a hierarchical organization built around a set of nested general functions.

According to Etienne Koechlin of Pierre and Marie Curie University in Paris and colleagues, information flows from the front of the prefrontal cortex to the back. They identify way stations along this path, suggesting that each makes a particular type of decision and feeds information into the next subregion, which makes a different type of decision.

The first subregion, the researchers propose, keeps track of what they call the episode: the particular situation in which someone is trying to figure out what to do. The next handles context, or what a given stimulus means in that situation. Finally, the back-most subregion reacts according to what the stimulus is.

It all sounds very abstract, Koechlin admits, but it translates into a real-world example. When the phone rings—a stimulus—people normally pick it up. But if the phone rings at a friend's house—that is, in a different context—people let it ring. However, if the friend, before hopping into the shower, asks a visitor to pick up the phone if it happens to ring—this, now, is the episode—the visitor answers the phone.

The researchers modeled these factors by asking volunteers to perform a series of tasks in the lab while each one's prefrontal cortex was being monitored. For example, volunteers were asked to press a key if they saw a green square, mimicking a stimulus-only condition. Sometimes color provided context, telling them to determine whether a letter was a vowel or consonant. And depending on the block of trials, or episode, they either ignored or responded to blue. Under functional magnetic resonance imaging, successive portions of the prefrontal cortex lit up as each subject carried out these tasks.

The research “is getting at an interesting dimension” of prefrontal cortex organization, says John Duncan of the Medical Research Council's Cognition and Brain Sciences Unit in Cambridge, U.K. Researchers have wrestled with how to describe the prefrontal cortex's organization, he says. Some have proposed that different parts handle particular types of information, such as the shape of an object or where it's located. But this study, Duncan says, “suggests we should be looking for more abstract grounds” to tease apart the contributions of different subregions.

Koechlin claims that his model accounts for and explains results from other brain imaging studies. Michael Petrides, head of the Cognitive Neuroscience Unit at the Montreal Neurological Institute of McGill University, says that Koechlin's “context” subregion is also activated in Petrides's studies when monkeys have to learn conditional responses, such as “if you see a red dot, push the right lever.” And the “episode” subregion appears to be the same one that is activated when the monkeys have to keep track of which pictures they've already responded to and which ones they should choose in the future. If so, “I'm quite comfortable with the conclusions of the article,” Petrides says.

But Duncan cautions, “I don't think this is going to be the be-all and end-all” of prefrontal cortex organization, a point Petrides echoes. There are many other regions of the prefrontal cortex that don't play a role in Koechlin's new model, Petrides says, and in any case, “I think things are much more complex in real life.”

9. GEOCHEMISTRY

# Minerals Cooked Up in the Laboratory Call Ancient Microfossils Into Question

1. Richard A. Kerr

Looks can be deceptive, researchers are finding as they search for traces of ancient as well as alien life. The wormy-looking shapes discovered in a meteorite from Mars turned out to be purely mineralogical and never were alive. And last year some researchers claimed that the textbook examples of the earliest known life on Earth—the 3.5-billion-year-old “Warrawoona” microfossils from Australia—are nothing more than suggestively shaped geologic detritus, not fossils (Science, 8 March 2002, p. 1812). Now, a paper in this issue of Science (p. 1194) strikes another blow for geology over biology: A group of researchers details how to cook up minerals in the laboratory that bear a striking resemblance to the Warrawoona microfossils.

The lab products “look remarkably ‘lifelike,’” says organic geochemist George Cody of the Carnegie Institution of Washington's Geophysical Laboratory in Washington, D.C. That doesn't prove that paleontologists mistook inorganic minerals for fossils at Warrawoona, but it does remind everyone that “just because something looks familiar doesn't mean it's biogenic,” says Cody. “A lot more thought has to go into biomarkers for ancient life.”

The recipe for fossil-like minerals turns out to be rather simple: silica, carbonate, and barium in an alkaline medium with a dash of simple organics. Geologist and crystallographer Juan Manuel García-Ruiz of the CSIC-University of Granada, Spain, and his colleagues mixed the inorganic ingredients at near room temperature and, at certain dilute concentrations of ingredients, produced sheets adorned with filaments, all composed of barium carbonate crystals coated with silica. These synthetic filaments look enough like true microfossils to be mistaken as the products of life, the authors say. And the lab conditions are similar to those at the time of the Warrawoona fossils, they say, to judge by the abundant carbonate, silica, and barium sulfate found among neighboring rocks.

When García-Ruiz and his colleagues immersed the synthesized filaments in a brew of formaldehyde and phenol, then heated the mix, a brown coating of complex organic matter formed on the filaments. The Warrawoona fossils are also coated with dark, complex organic matter. García-Ruiz argues that heat-induced breakdown of iron carbonates, which could have happened at Warrawoona, is known to produce relatively simple organic materials. With time, he argues, these materials could combine into complex organics, much as happened in the lab.

“Our work affords a coherent, completely abiotic scenario for formation of the [Warrawoona] ‘microfossils,’” says co-author Stephen Hyde of Australian National University in Canberra. “This means simply that we must rethink the criteria for detection of life, both here on Earth and elsewhere. Morphology and organic chemistry are evidently not enough.”

Many researchers would agree. “Nine and a half out of 10 paleontologists would say that's a microfossil,” says sedimentologist and paleontologist John Grotzinger of the Massachusetts Institute of Technology. One who wouldn't is William Schopf of the University of California, Los Angeles, who was first to publish on the Warrawoona fossils in Science in 1993 (30 April 1993, p. 640). “It's very interesting and ingenious work,” he says. “At the same time, the resemblance is superficial.” He sees walled-off voids in the Warrawoona structures where cells would have resided, whereas the lab structures are solid rods or, after etching with acid, hollow tubes. Hyde responds that the adsorption of organics onto a hollow tube—something the group hasn't attempted yet because of the tubes' fragility—could well create a structure that looks like a cell wall.

It's not surprising that the details of putative microfossils from Earth's earliest days are being debated, says paleontologist Andrew Knoll of Harvard University. “I think the synthesized microstructures look only moderately like real microfossils,” he says. “Unfortunately, that is also the case for the Warrawoona structures.” But it all may work for the good, suggests Grotzinger. The lab creations, he says, are “just annoying enough that paleontologists will go back to the field,” in hopes of finding convincing examples of the real thing.

10. PHARMACOGENOMICS

# FDA Puts the Brakes on Roche's Gene Array Test

1. Jocelyn Kaiser

The U.S. government has blocked the sale of a new kind of DNA diagnostic test, putting up an unexpected barrier to the marketing of technology to distinguish genetic differences in how patients metabolize certain drugs.

The Geneva-based diagnostics firm Roche had billed its test, launched last June, as “the first wide-scale application of diagnostic microarray technology in the world,” making it a milestone in genomic medicine. But in a letter released last week, the Food and Drug Administration (FDA) told Roche to stop selling the test and instead submit more data to support its claims.

The AmpliChip CYP450, which Roche developed with Affymetrix, detects mutations in two genes that can cause people to have serious side effects or fail to respond to some drugs (Science, 24 October, p. 589). In selling the chip to clinical labs as a component for their own tests, Roche compared it to genetic testing reagents, which aren't directly regulated by FDA. But in a 29 October letter, FDA said the AmpliChip requires a higher level of review because it is “of substantial importance in preventing impairment of human health” and uses “sophisticated” technology.

Roche “was testing the waters” by selling AmpliChip as a reagent while advertising it as a diagnostic, says regulatory analyst Ron Eisenwinter of Boston Healthcare Associates-Expertech. But the company denies that it was trying to avoid more intensive review. “We have always planned to seek FDA approval for in vitro diagnostic use,” says Melinda Baker, spokesperson for Roche Molecular Diagnostics in Pleasanton, California.

The “big unknown” is how much data FDA will require, says Kathy Hudson, director of the Genetics and Public Policy Center in Washington, D.C. Under one scenario, gene chips might have to be clinically tested like drugs. But requiring quantitative data on every mutation tested for, notes Christopher Austin of the National Human Genome Research Institute, might not be “economically feasible.”

The agency itself is struggling to sort out how to regulate these new tests, observers say. FDA officials expect to get plenty of advice this week at a Washington, D.C., workshop on pharmacogenomics.

11. MARINE RESEARCH

# A Plan to Go Where Few Have Gone Before

1. David Malakoff

The United States should take the lead in launching a new era of global ocean exploration, argues a report* from a National Academy of Sciences committee. But with a hefty price tag—start-up costs alone could run $270 million—the proposal is likely to encounter rough seas. “The money will be difficult, … [but] we need to set an example and then seek collaboration,” says Representative James Greenwood (R-PA), a longtime proponent of ocean exploration, who spurred Congress to request the study. The international panel, led by physical oceanographer John Orcutt of the Scripps Institution of Oceanography in La Jolla, California, reported last week on the feasibility of a global initiative to explore little-studied regions of the world (Science, 24 May 2002, p. 1386). Current funding priorities favor narrow, hypothesis-driven studies best tested by sending scientists repeatedly to the same sites, say critics. But there's another way, says Shirley Pomponi, head of the Harbor Branch Oceanographic Institution in Fort Pierce, Florida, and a member of the panel: “Hypothesis-generating exploration has gotten short shrift,” she says. Discovery won't come cheap. The panel estimates that a “minimal” exploration program, using leased ships and submarines and targeting just a few topics, would cost$30 million a year. A “fully capitalized” program would cost $270 million to start, it estimates, and up to$110 million annually to operate. The full program would need a dedicated, $70 million exploration vessel and 18 submersibles, including piloted, tethered, and autonomous models. The panel concludes that international support for such an ambitious program isn't there yet, but that the United States could get the project started. It suggests that the National Oceanographic Partnership Program, a 15-agency coordinating body, should oversee a new exploration effort, and an independent nonprofit organization, similar to those that run the Hubble Space Telescope and ocean drilling programs, should operate it. Pomponi and others acknowledge that some scientists have derided the idea of a new vessel as a “ship of fools.” The criticism, she says, is based on the misperception that exploration “is just people cruising the high seas on a lark.” Other skeptics question whether marine science budgets—roughly$400 million annually in the United States—can grow fast enough to accommodate a new exploration program. Marine researchers are already seeking funds for a $500-million-a-year ocean observing system and a$200 million network of sea-floor observatories.

Greenwood hopes to overcome these obstacles by gaining the support of President George W. Bush. Bush could polish his environmental record with a “blue-green initiative,” says Greenwood: “If Kennedy could go to the moon, there is no reason Bush can't go to the bottom of the sea.”

• *Exploration of the Seas: Voyage Into the Unknown, National Research Council, November 2003.

12. OCEAN SCIENCE

# Report Says U.S. Needs Bigger Deep-Sea Fleet

1. David Malakoff

A National Academy of Sciences panel says the U.S. government should build a new human-operated submarine for deep-sea scientists—but only after it constructs an unpiloted explorer and gives scientists greater access to existing submersibles. “The scientific demand for deep-diving vehicles … is not being adequately met,” the panel concluded this week in a report* that attempts to resolve a sometimes feisty debate over how to replace the field's flagship, the 35-year-old Alvin (Science, 19 July 2002, p. 326).

Alvin's owners and funders—the National Science Foundation (NSF), the National Oceanic and Atmospheric Administration (NOAA), and the U.S. Navy—requested the study earlier this year after the Woods Hole Oceanographic Institution (WHOI) in Massachusetts, which operates the craft, began exploring replacement options. The three-passenger Alvin has made nearly 4000 dives to as deep as 4500 meters, and although engineers have upgraded or replaced nearly all of its parts, maintenance costs are growing. Some researchers have argued for a new state-of-the-art piloted submarine that would go deeper. But others questioned the price tag of up to $20 million, saying less expensive tethered vehicles or autonomous robots could do a better job. The panel, led by retired IBM executive John Armstrong of Amherst, Massachusetts, backed both sides. It said the government should build “a new, more capable” human-operated submarine, perhaps by as early as 2006. To save money, the panel suggested that engineers consider using Alvin's existing titanium sphere or a leftover hull from a Russian program. First, however, the government should spend about$5 million to build a remotely operated vehicle (ROV) that could dive to 7000 meters, similar to WHOI's existing Jason craft. The new ROV would be based in the Pacific to expand research options. The panel says that NSF and NOAA should also alter the present arrangement that makes it more expensive for scientists to use submersibles not operated by WHOI.

“We're very pleased with the report,” says James Yoder, head of NSF's marine science program, which provides the bulk of Alvin's operating costs. “It will be a big help in making decisions.”

• *Future Needs in Deep Submergence Science: Occupied and Unoccupied Vehicles in Basic Ocean Research, National Research Council, November 2003.

13. ESTROGEN RESEARCH

# The Great Estrogen Conundrum

1. Jennifer Couzin

In mice and monkeys it works wonders for the heart, but in a study of thousands of women it falls flat. Scientists are wrestling with hormone replacement therapy's mixed results

The study was supposed to prove what people thought they already knew: Hormones taken by millions of postmenopausal women protect against heart disease. Instead, in July 2002 the Women's Health Initiative (WHI) abruptly ended its flagship experiment of 16,600 participants, half of whom were taking a popular estrogen and progestin combination pill and half a placebo. Those on hormones were slightly more likely to be felled by a heart attack than those on dummy pills. The headlines sent a quake through doctors' offices.

An aftershock is now spreading through research labs devoted to the study of estrogen and the heart. Decades of research and hundreds of scientific papers had consistently shown that estrogen was the heart's guardian. Mice, rabbits, pigs, and monkeys displayed reduced signs of vascular damage after receiving the hormone, and genetically susceptible animals given estrogen never got heart disease. The hormone's effects on the brain and cognition appeared no less remarkable. These, too, were upended by a related WHI study (see following story, p. 1138).

When the WHI news surfaced, attention focused on the disparities between observational studies and the clinical trial. But basic scientists grappled with their own set of questions. Why didn't the effects seen in human beings match lab results? Had they been testing the wrong hormones or using inappropriate animal models all this time? Were the deficits in their own work, in the clinical studies, or in efforts to link the two?

Some researchers are fascinated by the stark disparity. “From a scientific perspective, [WHI] has made the field substantially more intriguing,” says David Herrington, a cardiologist at Wake Forest University School of Medicine in Winston-Salem, North Carolina. But not for doctors and patients, he says: The WHI report “has dampened enthusiasm considerably about widespread use of hormone replacement therapy.”

This new disconnect between patient and researcher interests has scientists jittery—nearly as jittery, in fact, as they are over the apparent contradictions between basic research and clinical data. Exploring the contrasting results could help clarify and reshape the role of lab research in medicine. But estrogen therapy today carries the faint stench of failure. Many people may prefer to leave the mystery in the dust and move on. Already, researchers fear, that's happening.

## Mismatched models?

The popularity of hormone replacement therapy (HRT) stems partly from a rock-solid, decades-old observation: Given the same risk factors, “women have about a 10-year delayed onset in the expression of coronary heart disease” compared to men, says Lori Mosca, director of preventive cardiology at New York-Presbyterian Hospital in New York City. But the gender gap gradually shrinks with age, as levels of estrogen and other hormones decline. “There is something there” benefiting women, says Mosca, and many in her field have pointed to estrogen.

But artificially boosting a postmenopausal woman's estrogen levels increases her risk of uterine cancer. To counteract this effect, modern hormone therapies pair estrogen with some version of progestin, a class of hormones that includes naturally produced progesterone. They interfere with estrogen's stimulation of uterine tissue. Both therapies can trigger breast cancer, however, and most researchers believe that combination hormones are more likely to do so than estrogen alone.

Women and men both produce three different kinds of estrogen molecules, all of which activate estrogen receptors that modulate gene expression in various tissues, including the breast, uterus, and heart. Estradiol predominates. In both humans and animals, it relaxes arteries and boosts blood flow; it also minimizes thickening of the carotid artery, an indicator of atherosclerosis.

The WHI study results, however, painted a bleaker picture: Women on estrogen and a synthetic progestin were 24% more likely to suffer serious heart problems than those on placebo. For every 10,000 women on hormones, 39 would be affected, compared with 33 on placebo. Another HRT study, the Heart and Estrogen/Progestin Replacement Study (HERS), reported in 1998 that estrogen and progestin in combination failed to slow progression of heart disease in more than 2700 women who were already sick. “This is a hypothesis that failed,” says Marcia Stefanick of Stanford University, a lead WHI investigator.

The WHI also includes an estrogen-only arm with 11,000 women who have had hysterectomies and are thus not at risk of uterine cancer. That study continues, although the hormone's effect on heart attacks, strokes, and blood clots appears to be heading “in the same direction” as in the ill-fated combination trial, Stefanick says.

Some basic researchers, however, reject attempts to apply WHI results broadly. “There has been overinterpretation and misinterpretation” of the data, argues Michael Mendelsohn, director of the Molecular Cardiology Research Institute at Tufts-New England Medical Center in Boston.

The mean age of WHI participants at enrollment was 63, older than most menopausal women who initiate hormone treatment. The study began in the early 1990s. Volunteers received a standard dose of Prempro, then a relatively new combination drug that quickly became a favorite in the United States. But Mendelsohn and other basic researchers contend that the dose was too high, a sudden boost for women who had lived with lower hormone levels for years. Furthermore, most animal studies had focused solely on estrogen rather than a combination of estrogen and progestin. “The question is, Have we [through the WHI] tested the hypothesis that the [animal] data supports?” he asks, “and the answer's no.”

But one could also ask a different question: Have the bulk of animal studies yielded a hypothesis that's testable in humans? Most experiments in mice involve “yanking out the ovaries when these animals are 5 to 6 weeks old” to induce the equivalent of menopause, says Carole Banka of La Jolla Institute for Molecular Medicine in San Diego, California. (Mice enter their version of menopause at around 14 months, and, says Banka, “no one wants to pay to house animals for 14 months.”)

There may be other disconnects. Nearly all mice in hormone studies receive estrogen by means of an implant under the skin, not orally. The subcutaneous delivery route bypasses the liver and may change how the steroid is metabolized. Many lab animals, particularly mice, are treated with estradiol. Prempro, however, contains not estradiol but different estrogens from horses. (Some studies on monkeys and rabbits have found protection from heart disease with equine estrogens.) Hormone experts can't agree on whether the different compounds have different cardiac effects.

There's much debate, too, about the role of progestin, the hormone added to avert the risk of uterine cancer. Until about 5 years ago, it was rarely studied in combination with estrogen in animals. Some work hints that progestin may blunt estrogen's positive effect on vascular cells, which could render the combination useless. But studies in monkeys suggest that “it doesn't seem to be a large issue,” says Thomas Clarkson of Wake Forest.

Modeling heart disease in animals is an imperfect science. To mimic early-stage disease, scientists can injure a major artery. They sometimes feed monkeys extremely high-fat diets. Or they rely on genetically altered mice deficient in the enzyme ApoE. These animals develop a form of atherosclerosis, but it's unclear how closely their disease hews to the human version.

Oddly enough, even mice with serious disease rarely die of it. “Most of these animals don't drop dead of heart attacks,” says Banka. So researchers rely on other measures, such as improvements in arterial lesions or carotid thickness, to assess estrogen's benefits. Still, lesions aren't a surefire way to predict death from atherosclerosis in humans, who normally succumb when an arterial plaque ruptures. “These are highly artificial systems,” says Jan-Åke Gustafsson, a molecular endocrinologist at the Karolinska Institute in Stockholm, Sweden.

## Back to the lab

In the 16 months since the WHI combination hormone trial was halted, the response of animal researchers has varied. Because the study focused on women over 60, some reject its applicability to women in their early 50s, who make up the bulk of those starting hormone therapy. Other scientists are more circumspect. “The failure of these trials is sending us a strong message,” says Michael Rosenfeld, a pathologist at the University of Washington, Seattle. Amid all the debate and defensiveness lies an uncomfortable fact: The design of many animal trials did not reflect the reality of hormone therapy in humans, he says. Going forward, Rosenfeld adds, the two need to run as parallel as possible.

Whether anyone will follow his recommendation isn't clear. Banka, who last month gave her mice combination hormones for the first time, says she's never seen a published mouse study documenting the effects of estrogen and progestin. And although researchers such as Clarkson have experimented with combination hormones in monkeys, Clarkson says it's too costly to do so in every study because that would require a third group of animals on estrogen alone.

But some researchers are already tackling one glaring contrast between the lab and clinic: They are studying animals as unhealthy as human subjects. Animals in traditional hormone experiments receive estrogen while their arteries are still clean or when they have fatty lesions but no arterial plaques. Roughly half of all middle-aged women, meanwhile, are thought to harbor subclinical atherosclerosis, says Jacques Rossouw of the National Heart, Lung, and Blood Institute in Bethesda, Maryland, who is project officer for the WHI.

Recent animal work suggests that arteries spotted with plaques may react very differently to estrogen than healthier ones do. In a 4-year study, Clarkson's team removed the ovaries of monkeys, then fed them what he calls an “imprudent” diet high in fats; some of those monkeys also received equine estrogens orally. After 2 years (the equivalent of 6 years in a woman), all the animals received hormones. Those given estrogens when the imprudent diet began—and, presumably, before their arteries were streaked with fat—had 70% less atherosclerosis than monkeys that didn't get them right away. Clarkson says this suggests that it was the timing of hormone initiation, and not the steroids themselves, that caused problems in the WHI. Rossouw agrees that it's possible that estrogen triggers “events” in those already suffering from atherosclerosis by perturbing blood clotting or causing inflammation. This is what some suspect happened in the WHI.

A nascent human trial—one of very few to launch since the WHI combination study ended—will test how hormones perform in women without arterial plaques. Last month, the Kronos Longevity Research Institute in Phoenix, Arizona, announced plans to fund a 5-year study of atherosclerosis progression in 675 women around ages 45 to 55. S. Mitchell Harman, director and president of the institute, anticipates divvying participants into three groups: one receiving progestin and an estradiol patch; another, progestin and equine estrogens; and a third, placebos. “The basic science gives us lots of reasons” to believe that younger women do better on hormones, says Harman. In contrast to Rossouw, he estimates that only 10% of menopausal women have subclinical heart disease.

Timing may matter for estrogen's other cardiovascular effects as well. Cardiologist Suzanne Oparil of the University of Alabama, Birmingham, is noticing that when estrogen is given to older rats who have spent time without it, their carotid artery doesn't respond as impressively. “I think the target degenerates,” she says, referring to receptors to which estrogen binds.

San Diego's Banka, with colleagues at the University of Arizona in Tucson, is trying to bridge another chasm between human and animal hormone studies: the way menopause arrives. “Having your ovaries removed and undergoing natural menopause are very different,” she says. Researchers need a way to trigger menopause in young animals to keep studies affordable, Banka says, so the team is experimenting with a chemical compound that can induce a mouse version of menopause.

WHI researcher Stefanick agrees that better animal models are sorely needed. But she's not certain they'll make any difference where HRT is concerned. “The bottom line on the clinical trials is that they are absolutely relevant to people,” she says. “My days of looking at hormone replacement are over.”

## Short of support

Reconciling clinical studies and basic research has gotten tougher as funding agencies and scientific journals have cooled to estrogen research. The chill is evident to Raghvendra Dubey, a biochemist at the University of Zürich, Switzerland, who submitted a grant proposal to the U.S. National Institutes of Health with Edwin Jackson, a pharmacologist at the University of Pittsburgh Medical Center. The pair have been investigating how metabolites of estrogen might be beneficial to the heart. They received reviewers' comments soon after the WHI news broke: “Hormone replacement therapy is not protective, so why do the study?” is the bottom line Dubey recalls. Richard Karas of the Molecular Cardiology Research Institute at Tufts says, “I personally have seen [similar comments] in both manuscript reviews and grant reviews.”

Mosca, at New York-Presbyterian Hospital, says that results from trials like WHI and HERS should nudge basic science in new directions, such as closer study of the two known estrogen receptors and their functions. Already under investigation is a class of compounds called selective estrogen receptor modulators, nonhormonal drugs that stimulate estrogen receptors in some tissues and block them in others. One, raloxifene, blocks receptors in the breast and uterus, which should steer clear of potentially cancerous effects. The Raloxifene Use for the Heart trial, which spans 26 countries and includes 10,000 women, is slated to continue for roughly 3 or 4 more years.

WHI and HERS have led some scientists to reexamine the source of their faith in estrogen: the premenopausal gender gap in heart disease. Perhaps, some now say, it's not that estrogen is a protector but rather that testosterone is an aggravator of heart disease. Another possible explanation for the gender gap may be differences in cholesterol levels. Still, estrogen remains a star for many. The promise of hormones “has just been scratched,” says Tufts's Mendelsohn. He and others fervently hope they'll have shovels to keep digging.

14. ESTROGEN RESEARCH

# Brain Researchers Try to Salvage Estrogen Treatments

1. Ingrid Wickelgren

Despite a clinical trial's result that hormone therapy increases the risk of dementia, some scientists say it's not time to give up on the strategy

In July 2002, investigators with the Women's Health Initiative (WHI) halted a clinical trial in part because hormone therapy was bad for the heart (see previous story). Ten months later, they announced that it might be bad for the brain as well: Women taking a widely used hormone pill were twice as likely to suffer signs of dementia as those taking a placebo.

The results roiled the small but burgeoning cadre of scientists who study estrogen's effects on the brain. “The WHI blew me away,” says James Simpkins, a neuroscientist at the University of North Texas Health Science Center in Fort Worth. Its results, Simpkins and others say, flew in the face of hundreds of studies conducted over a decade suggesting that estrogen could protect brain cells from damage and improve cognition in people and animals (Science, 2 May 1997, p. 675).

Far from ready to cast out that body of data, researchers fear that reports of the WHI results, and subsequent regulatory actions, may make the work appear obsolete. With many women convinced that hormone replacement therapy (HRT) is harmful, further clinical trials are unlikely to draw volunteers. Already, some funding for basic research may be in jeopardy. “There is an incredible backlash that has obscured much of the basic science,” says neuroscientist Bruce McEwen of Rockefeller University in New York City.

Researchers are fighting back. A dozen prominent scientists have formed a response team dubbed the Consortium for the Assessment of Research on Progestin and Estrogens. The consortium is charged with informing the public that the results of the 5000-woman WHI Memory Study (WHIMS) have been widely overinterpreted. They claim that the increase in dementia risk the study found cannot be extended to all forms of HRT, or even to the women most likely to start HRT. At the same time, scientists are scurrying back to their benches to try to reconcile the WHIMS results with laboratory data.

## Bad recipe?

Some researchers are critical of the formulation WHI tested: a mixture of horse estrogens and a synthetic progesterone. Although this mixture, dubbed Prempro, was by far the most widely prescribed form of HRT in the United States, researchers say that both its dosing and its type of progestin work against estrogen's ability to protect neurons. Prempro delivers a constant dose of hormones throughout the month. Before menopause, the female body gets only intermittent pulses of both. A constant supply of estrogen, researchers say, may dull the sensitivity of estrogen receptors.

Intermittent exposure to estrogen is mainly responsible for the positive effects observed in animals. For example, this year John Morrison, Peter Rapp, and their colleagues at Mount Sinai School of Medicine in New York City reported injecting monkeys with estradiol once every 3 weeks. The treatment enhanced the monkeys' cognitive performance throughout the month and also boosted the number of synapses in two brain areas associated with memory: the hippocampus and prefrontal cortex.

The progestin in Prempro, so-called medroxyprogesterone acetate (MPA), may be a poor choice relative to natural progesterone, at least for the brain. Roberta Brinton's team at the University of Southern California in Los Angeles recently demonstrated that MPA counteracts estrogen's ability to protect neurons against toxic insults such as exposure to the culprit peptide in Alzheimer's disease, amyloid.

By contrast, natural human progesterone seems to either enhance or moderate estrogen's effects. In Brinton's cell cultures, progesterone bolstered estrogen's protective effect. In monkeys, it worked synergistically with estrogen to restore synaptic connections in certain brain neurons involved in mood and memory.

Still, there is a paucity of laboratory data on estrogen-progesterone combination treatments—a major gap given that most women shouldn't take estrogen alone because it increases the risk of uterine cancer. Morrison and Rapp plan to compare the cognition and brain structure of monkeys given estrogen with those given estrogen followed by progesterone.

Aside from spotting flaws in Prempro's recipe, many researchers believe that the results of WHIMS are limited by the age of women tested. The study included only women age 65 and older to increase the odds of finding dementia, and because some doctors were giving HRT to older women to enhance cognitive and cardiovascular health. However, most women who begin HRT do so between 45 and 55. And the brains of 65-year-olds are more likely to harbor vascular problems or early Alzheimer's disease. HRT might exacerbate these existing conditions, some researchers say.

Timing could make a big difference in the effectiveness of HRT, neuroscientists say, if estrogen preserves healthy neurons rather than repairing sick ones. Estrogen does not appear to do much, if any, good if applied to brain cells or given to people after damage is done. Because of this, “I don't think the WHIMS trial is directly relevant to the question of whether estrogen protects against Alzheimer's disease,” says Yale University gynecologist Frederick Naftolin.

## Some explaining to do

Even given the critiques of Prempro, no one really understands why WHIMS found an increased risk of dementia among HRT users rather than simply no effect. WHIMS principal investigator Sally Shumaker of Wake Forest University School of Medicine in Winston-Salem, North Carolina, wonders if the drug somehow raised the risk of silent strokes. Her team, along with a group led by Susan Resnick of the National Institute on Aging (NIA) branch in Baltimore, Maryland, proposes to use magnetic resonance imaging to search the women's brains for the type of strokes that lead to dementia. Brinton's team is investigating another possibility. Her previous work hints that MPA might induce a chemical reaction in neurons that can promote neurofibrillary tangles, a hallmark of Alzheimer's disease.

No one can say whether using a different formulation of HRT or starting therapy earlier would be safe and beneficial for women over the long run. “The onus is in the other direction now,” says Shumaker. “If a woman suffers severe symptoms associated with menopause, she should look at hormone therapy, but at the lowest dose and for the shortest period of time.”

Supporting this cautious view, some researchers point out that earlier epidemiological studies linking HRT to a decreased risk of dementia may be seriously flawed. Most, if not all, tracked women who had decided on their own to take HRT. Such studies are prone to so-called healthy-user bias, in which women who choose a therapy may be different in other ways, such as being better educated or more health conscious, from those who do not. Such incidental differences, which might elude statistical techniques meant to control for them, could explain a resistance to dementia quite apart from any hormone treatments. WHIMS, in contrast, assigned women randomly to HRT or placebo. It was also the largest study ever done.

But few scientists are ready to close the book on estrogen. NIA is funding three ongoing human studies to determine whether estrogen helps prevent or ameliorate the symptoms of Alzheimer's disease or other cognitive impairment. The agency is eager for basic science proposals as well. “We want to continue to try to figure out the role of estrogen in the brain, the heart,” and other organs, says Susan Molchan, who directs NIA's program for Alzheimer's disease clinical trials.

Meanwhile, the race is on to develop a safer estrogen that preserves brain cells. Simpkins suspects that estrogen's neuroprotective talents do not require the receptors thought to be responsible for many of estrogen's side effects in the breast and uterus. He and his team have synthesized more than 70 estrogenlike compounds that do not bind to estrogen receptors but have a potent protective effect on neurons in culture and in rodent models of stroke and Alzheimer's disease. McEwen and his colleagues are exploring novel ways estrogen exerts its effects on the brain in hopes of pinpointing additional avenues for bolstering neurons without impacting other tissues.

The WHIMS results are “not a signal that we should stop research, but that we need to redouble our efforts,” says Simpkins. Adds Brinton: “It's a tough—but exciting—time for all of us in the field.”

15. VIROLOGY

# New Vaccine and Treatment Excite Ebola Researchers

1. Martin Enserink

Recent advances—including a second vaccine candidate and the first treatment shown to work in monkeys—offer hope for controlling a gruesome disease

The thrill is gone. Ebola, long the most frightening disease in the public's mind, has begun to lose its sex appeal. In 1995, the world was mesmerized when the virus erupted in Kikwit, Zaire, now the Democratic Republic of the Congo. But an outbreak in the neighboring Republic of the Congo earlier this year barely made headlines. Smallpox has become the world's favorite nightmare.

But the researchers who study Ebola and its close relative, Marburg, remain fascinated by the things that titillated readers and moviegoers a decade ago: the 50% to 90% mortality rate, the horror of widespread hemorrhage in some patients, and the ongoing hunt for the natural reservoir. And bit by bit, they're unraveling the mechanisms that make Ebola so devastating. At a meeting* last month at the National Institutes of Health's Vaccine Research Center in Bethesda, Maryland, researchers also unveiled data about a second promising vaccine candidate and presented a potential treatment for Ebola, an advance that virologist C. J. Peters of the University of Texas Medical Branch in Galveston called “extremely impressive,” especially given the obstacles in the field.

Although Ebola incidence seems to be on the rise, it remains rare—according to the World Health Organization (WHO), it has claimed just over 1000 human lives in 30 years—so only a handful of labs study it. Progress is also hampered by the lack of so-called biosafety level 4 labs with space or permission to house monkeys. (Many new ones are on the drawing boards or under construction, but most will take years to complete.)

It's unclear why epidemics occur almost yearly now, whereas almost 20 years elapsed between the first known outbreaks, in Zaire and Sudan in 1976, and the next, in Kikwit. Nor do researchers understand why Ebola is currently pushing some ape populations in central Africa to the brink of extinction (Science, 11 April, p. 232).

Finding a treatment for Ebola is still one of the top priorities. Although researchers have found several antivirals that worked in mice and guinea pigs, none was effective in monkeys. Other approaches have failed as well. For instance, a monoclonal antibody called KZ52 that was derived from the bone marrow of an Ebola survivor in Kikwit looked promising in guinea pigs. In a recent study, however, it failed to protect rhesus macaques, Dennis Burton of the Scripps Research Institute in La Jolla, California, reported at the meeting.

Another way to treat Ebola would be to try to block the physiological cascade that the virus triggers rather than target the virus itself. Many researchers are now focusing on this complex, and poorly understood, process. Researchers know that the bleeding in Ebola is caused by coagulopathy, a dysfunction of the blood-clotting system, but it's not quite clear where this starts or how it can be stopped. At the meeting, a team led by Tom Geisbert of the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID) in Fort Detrick, Maryland, presented evidence that macrophages, which are a type of white blood cell, play an important role. When infected by the virus, they start expressing a clotting protein called tissue factor (TF) on their surfaces. These molecules attract other clotting molecules from the blood, resulting in localized, harmful blood clots that leave the body vulnerable to bleeding elsewhere.

The USAMRIID group decided to treat rhesus monkeys with an anticoagulant called rNAPc2, which counteracts the effects of TF. (A company called Dendreon is also testing the anticoagulant as a treatment for acute coronary syndromes.) Given immediately after Ebola infection or within 24 hours and continuing for 8 days, the drug saved the lives of three out of nine monkeys and slowed death by several days in the remaining six. All three controls died.

Because infection in rhesus monkeys is almost universally lethal, a 30% reduction in mortality is “incredible,” says Pierre Formenty, a virologist at WHO in Geneva, Switzerland. Virologist Mike Bray of the U.S. National Institute of Allergy and Infectious Diseases in Bethesda, Maryland, says, “Tom has really done quite a remarkable thing: to show where the coagulopathy is arising, and that interfering with it actually has some benefit.” Peters notes that in humans, who have a fighting chance of their own, mortality rates might be reduced even more. But Vincent Deubel of the Pasteur Institute in Lyon, France, was less impressed. “I think we need to do better,” he says.

Meanwhile, several vaccine strategies appear to be paying off. Three years ago, Nancy Sullivan and colleagues at the Vaccine Research Center reported success in monkeys using a two-pronged approach: three shots of “naked DNA,” expressing Ebola's glycoprotein, followed by one shot with an adenovirus engineered to express glycoprotein as well (Science, 3 November 2000, p. 923). But that regimen takes 6 months to produce immunity; in September of this year, the same team, together with Geisbert's group, reported in Nature that even a single shot of the adenovirus alone was protective, in as little as 28 days. Such a fast-acting vaccine would be a major advantage during outbreaks, Sullivan says, when villagers would need to be protected on short order. The center is now doing safety trials for the DNA component of the vaccine in humans, and it plans to start doing the same for the adenovirus soon.

In Winnipeg, Canada, Heinz Feldmann, Steven Jones, and their colleagues at the National Microbiology Laboratory have developed a competing vaccine, based on an animal pathogen called vesicular stomatitis virus (VSV), which causes a disease whose symptoms resemble those of foot and mouth—although it's in a different virus family and not nearly as virulent—in cattle, horses, and pigs. Using techniques developed by Yale University's John Rose, who's pioneering the use of VSV as a vector for an HIV vaccine, the team replaced VSV's glycoprotein with that of Ebola. After experiments in mice and guinea pigs proved successful, the Canadian team had Geisbert test the vaccine at USAMRIID. To their surprise, they discovered that the VSV vaccine, too, could protect after a single shot and within a month. “It looks very promising,” Deubel says. “Three weeks ago, we had one promising vaccine,” says Bray. “Now we have two.”

But the new vaccine has potential downsides: VSV can occasionally cause illness in humans; what's more, animal health experts worry about using a livestock disease agent as a human vaccine. If the researchers can show that the engineered VSV is not pathogenic and vaccinees don't shed much of the virus, the vaccine might have a future, said Feldmann. Still, Deubel predicts that it will face an uphill regulatory battle.

With safety trials for the prime-boost vaccine under way, the next phase—a field test in Africa to demonstrate efficacy—is beginning to give researchers headaches. Clashes over issues such as how to bury deceased patients have often strained relations between local populations and the international teams trying to quash outbreaks. During last year's outbreak in Gabon, health care workers felt so threatened that they withdrew from an affected town. Conducting a trial under such circumstances “is going to be extremely difficult,” says Deubel.

The Vaccine Research Center is already discussing these issues with WHO and plans to talk to African governments about a trial with the adenovirus-alone variant of the vaccine, says Sullivan. Among the many thorny questions: Should there be a placebo group? Many think that with Ebola, it would be unethical and impossible to explain why some don't get a real vaccine; but dropping the placebo arm would make it harder to demonstrate efficacy.

WHO's Formenty thinks that a drug might have a better chance than a vaccine of being accepted by the community—and could help reduce Ebola's spread as well. Currently, patients often avoid hospitals because doctors don't have much to offer them; as a result, caregivers are more likely to become infected. If Geisbert's or another therapy could give patients a better shot at survival, they might come to the clinic, Formenty says, where they could be isolated: “That would be a great step forward.”

• *Symposium on Viral Hemorrhagic Fevers, 14 to 17 October.

16. SOCIETY OF VERTEBRATE PALEONTOLOGY MEETING

# Mammal Menagerie

1. Erik Stokstad

ST. PAUL, MINNESOTA—From 15 to 18 October, 1000 paleontologists flocked to the banks of the Mississippi for SVP's 63rd annual meeting. Early whales were a hot topic, as were furry predators—demonstrating that dinosaurs had no monopoly on violence.

Being top dog may not be so great after all. A new study of the fossil canids of North America shows that as dogs evolved to be bigger, they began to hunt larger prey—and tended to go extinct sooner. The findings, presented here last month, could have implications for conservation biology. “It's a very clever blend of modern ecology and paleontology,” says Christine Janis of Brown University in Providence, Rhode Island.

The study was inspired by a 1999 survey of living carnivores. That research noted that carnivores bigger than 21.5 kilograms tend to eat differently. Although smaller carnivores can subsist on invertebrates and prey less than half their own size, larger predators hunt prey that's as big as they are, or bigger. This thick-steak lifestyle, known as hypercarnivory, is likely rooted in an animal's energy budget: For a hungry lion, mice don't have enough calories to be worth pursuing. Blaire Van Valkenburgh of the University of California, Los Angeles, wondered if this pattern held true in the past and, if so, what long-term consequences might have resulted from a supersize diet.

Luckily for Van Valkenburgh, detailed studies of two extinct subfamilies of North American dogs had recently been published. As with many other ancient species, she found that dogs in both subfamilies tended to evolve to larger sizes. Then she looked for signs of hypercarnivory, such as a relatively deep jaw. Deeper jaws make for a stronger bite and correlate with larger prey size in modern dogs. It turned out that hypercarnivory was much more common among the bigger dogs, whereas smaller ones showed signs of being omnivorous.

This difference had a bearing on extinction rates. The hypercarnivore species went extinct after 6 million years or less. In contrast, species of dogs that took down smaller prey lasted for up to 12 million years. “It's the clearest demonstration that hypercarnivorous dog species [persist for] a shorter time,” says Lars Werdelin of the Swedish Museum of Natural History in Stockholm. One possible explanation, he says, is that competition may be stiffer for larger prey, which is less common fare than fruit, bugs, or mice.

With many hypercarnivores today threatened by extinction, Werdelin says, research like this can help explain why they're vulnerable. “At the very least, we need to know what aspects of extinction are natural and not caused by man,” he notes.

17. SOCIETY OF VERTEBRATE PALEONTOLOGY MEETING

# Early Whales Get All Wet

1. Erik Stokstad

ST. PAUL, MINNESOTA—From 15 to 18 October, 1000 paleontologists flocked to the banks of the Mississippi for SVP's 63rd annual meeting. Early whales were a hot topic, as were furry predators—demonstrating that dinosaurs had no monopoly on violence.

When the ancestors of whales headed into the seas, they evolved teeth to eat meat; modified their ears to listen underwater; and, of course, reshaped their bodies to become highly efficient swimmers. At the meeting, two presentations focused on the aquatic abilities of early whales. Skeletal proportions reveal how one such whale paddled, and bone tissue suggests that the most primitive whale ancestor known, originally thought to have run on land, must have been a wader. The studies “certainly broaden our understanding of the locomotor transition,” says Annalisa Berta of San Diego State University in California.

Paleontologists since Charles Darwin have surmised that whale ancestors might have swum like today's otters: paddling with their hind limbs and wiggling their bodies and flat, somewhat flukelike tails. Philip Gingerich of the University of Michigan, Ann Arbor, decided to test this idea by seeing how the bones of modern semiaquatic mammals compare with those of ancestral whales.

Gingerich picked 50 mammals that spend at least some time in water, from water mice to hippos. Then he measured 14 aspects of their limbs and torsos. By using a statistical technique called principal components analysis, Gingerich determined that most of the variation in their skeletal proportions was due to body size. But two other variables also played a role. The degree of aquatic agility explained 3.6% of the variation in body proportions; a seal, for example, has relatively longer toes and shorter leg bones than a tapir does. Another 0.7% depended on whether a mammal swims mainly with its feet, like a sea otter, or with its hands, like a platypus.

For the early whales, Gingerich chose two with exceptionally well-preserved skeletons. Like a modern whale, the 37-million-year-old Dorudon lacked hind limbs and had powerful tailbones for propulsion. But the big surprise came from the 47-million-year-old Rodhocetus, a four-legged whale from Pakistan with feet probably too delicate to travel far on land. Rather than resembling an otter's, its limb proportions were more like those of a Eurasian aquatic mole called a desman. That implies that Rodhocetus paddled with its feet alone, using its tail as a rudder.

Also at the meeting, Sandra Madar of Hiram College in Ohio discussed bones from the earliest known whale, Pakicetus. Some 400 bones of various individuals have been collected by Hans Thewissen of Northeastern Ohio Universities College of Medicine in Rootstown and colleagues. In 2001, he and his co-authors noted that Pakicetus bones are long and slender, similar to those of cursorial mammals today.

Madar says the bones tell a more complicated story. When she examined the internal tissue under a microscope, she found that the shafts were extremely dense. “Their bones would be so brittle that they couldn't bear the force of running,” Madar says. Although weight-bearing microstructures in its hip and femur bones suggest that Pakicetus could have walked on land, Madar and Thewissen think it spent most of its time wading in streams. Aquatic mammals such as sea cows, she notes, also have dense bones, which serve as ballast to keep their bodies from floating. Pakicetus, she says, may have enjoyed a similar lifestyle on the bottom.

18. SOCIETY OF VERTEBRATE PALEONTOLOGY MEETING

# Mastodon Gladiators

1. Erik Stokstad

ST. PAUL, MINNESOTA—From 15 to 18 October, 1000 paleontologists flocked to the banks of the Mississippi for SVP's 63rd annual meeting. Early whales were a hot topic, as were furry predators—demonstrating that dinosaurs had no monopoly on violence.

Until 10,000 years ago, a menagerie of big animals roamed North America, including fearsome predators such as saber-toothed cats and bone-crushing dogs. Not all of the killers were carnivores, though. Rival mastodons, stoked by surging hormones, occasionally battled to the death, according to research presented here at the meeting. “The injuries were severe,” comments paleontologist Jeffrey Saunders of the Illinois State Museum in Springfield. “It was mortal combat.”

Mastodons are extinct relatives of modern elephants. At 6 to 8 tons, mastodons were some 30% more massive than elephants, and their tusks were more curved. While studying mastodon remains from the Great Lakes region, paleontologist Dan Fisher of the University of Michigan, Ann Arbor, came across signs of injuries reminiscent of those found in dueling elephants. When male elephants are in heat, called musth, they often establish dominance by knocking heads or fencing with their tusks.

Fisher spotted hints of violent combat in five out of 10 mastodon specimens well enough preserved to show signs of injuries. The evidence includes broken ribs and tail vertebrae—places where elephants in musth get injured. Looking closely at the tusks, Fisher found strong ligaments near the outer edge of the tusk socket that could have absorbed shocks from dueling. “You wouldn't expect that [reinforcement] for other tusk uses,” he says. Moreover, the tips of the tusks on one mastodon were fractured so badly that they probably would have broken off if they had ever been used for normal activities again; this suggests that the animal died shortly after hitting something.

That something may have been another male. Several skulls had puncture wounds that could have been inflicted by a tusk. To double-check the feasibility of tusk warfare, Fisher created computer models of the skulls. The models showed that by jutting its head upward, one animal would have thrust its upcurved tusks into its opponent's cheek, where Fisher saw the damage. Such a deep wound could have been fatal, Fisher explains, because the tusk that made it would have punched through some 50 centimeters of muscles used for chewing. Even if the animal survived the trauma, it would likely have starved. “The adult males sometimes did fight to the death,” Fisher concludes.

Was it for love? The growth rings in the tusks show that the deaths occurred from the early spring to the late summer, not a normal time for mastodon mortality, Fisher says. The rings also suggest that the animals had been fasting before death. Elephants in musth, Fisher notes, are so fixated on mates and competitors that they forgo eating. Love or not, it certainly looks like war. “This was definitely conflict between two mastodons,” Saunders says.