# News this Week

Science  16 Mar 2001:
Vol. 291, Issue 5511, pp. 2060
1. FINANCIAL CONFLICT

# Universities Puncture Modest Regulatory Trial Balloon

1. Eliot Marshall

Even the blandest words can be incendiary when they're about money. U.S. officials have learned that lesson the hard way this winter after the chief lobbies for academic medicine kicked up a fuss about suggestions on how to deal with financial conflicts of interest. Their opposition is likely to shoot down a mildly worded “draft interim guidance on financial relationships in clinical research” issued by the Department of Health and Human Services (HHS) in January.

The HHS rule-writing effort was designed to tune up policies that were last examined in 1995. The current push began after a young man died in a university-based gene therapy experiment in 1999. The case drew attention because one of the clinicians in the project, and the academic institution itself—the University of Pennsylvania—had equity in a company that was hoping to benefit from the research (Science, 12 May 2000, p. 954). But the cry for clear and consistent new standards regarding money and medicine, led by former HHS Secretary Donna Shalala, so far has failed to win the attention of the new Bush Administration.

The HHS draft guidance (ohrp.osophs.dhhs.gov/nhrpac/mtg12-00/finguid.htm) reflects what officials saw as a consensus on how to deal with the increasing role of industry in academic medicine. Among other things, it suggests that researchers' potential conflicts be disclosed to the same Institutional Review Boards (IRBs) that now monitor other ethical issues, and possibly to patients as well. About a third of the publicly funded IRBs are considering whether to take a look at financial issues, according to HHS. The draft statement also encourages academics to become involved in reviewing “institutional” conflicts—the kind that occur when a university itself has a financial stake in the outcome of a clinical trial. Drawing on public comments from a meeting last August, the guidance sought to harmonize patchy federal and university policies. It was developed by the new Office for Human Research Protections (OHRP), a high-profile version of an outfit previously housed within the National Institutes of Health.

But even that gentle prodding was too much for academic leaders. The draft is “quite premature,” says David Korn, a former dean of medicine at Stanford University who now works on government issues at the Association of American Medical Colleges (AAMC) in Washington, D.C. “I think it is necessary to address these issues,” says Korn, but “I don't think the government has any great wisdom [to offer]. We don't even know how to define an institutional conflict of interest.”

On 2 March, four major education organizations wrote to OHRP director Greg Koski, asking him to “withdraw” the guidance and “reissue portions of it as points for consideration.” They argued that some of the HHS ideas—particularly on potential institutional conflicts—were based on anecdote rather than good evidence. On 8 March, the Federation of American Societies for Experimental Biology echoed those views in a separate letter.

It's not that the community is ignoring the issue. The AAMC, which signed the call for withdrawing the draft along with the Association of American Universities, the Council on Government Relations, and the National Association of State Universities and Land Grant Colleges, is setting up a new panel to formulate its own policy. AAMC president Jordan Cohen is hoping that its 125 member institutions will “agree voluntarily to abide by a common set of principles for managing those conflicts.” The panel is headed by William Danforth, former president of Washington University in St. Louis, who has yet to set a date for the first of the proposed twice-yearly meetings.

Koski says he was taken aback by the sharp and “misleading” tone of the response from academia. “We haven't issued any guidance yet,” he points out, “and you can't withdraw something that hasn't been issued.” HHS published the statement “to start a broad discussion,” he adds.

The government and the private sector can work in parallel, says Koski, adding that he hopes HHS can learn from the Danforth committee as it undertakes its review. And he doesn't think it will be too hard to clarify the rules and build public confidence in research: “This isn't rocket science.”

2. PARKINSON'S RESEARCH

# Fetal Cell Transplant Trial Draws Fire

1. Gretchen Vogel

In just 1 week, an experimental treatment for Parkinson's disease—fetal cell transplants—went from promising to perilous. At least, that's how much of the general media reported the publication of mixed results from the first double-blind study. But Parkinson's researchers caution that results from a single trial, especially one that was controversial from the start, should not be the final word on the technique.

On 8 March, neuroscientist Curt Freed of the University of Colorado School of Medicine in Denver, neurologist Stanley Fahn of Columbia University College of Physicians and Surgeons in New York City, and their colleagues reported in The New England Journal of Medicine that injecting fetal cells into the brains of Parkinson's patients resulted in a significant improvement in some recipients. Several patients, however, also experienced troubling side effects.

Parkinson's disease is marked by the mysterious death of brain cells that produce a chemical messenger called dopamine, which helps control motor function. The researchers had hoped to replace the lost cells by injecting dopamine-producing neurons from the brains of aborted fetuses into the affected areas of patients' brains. Several studies of this experimental technique had yielded encouraging results, but none included a control group.

The Colorado-Columbia study was the first one the National Institutes of Health (NIH) supported after President Bill Clinton lifted the ban on federal funding for research involving fetal tissue. From the outset, the 1993 award was controversial. Some researchers worried that the decision to fund a trial that used only one of several transplant techniques could harm the field if the results weren't positive (Science, 4 February 1994, p. 600; 11 February 1994, p. 737). Those fears seemed prescient last week as newspapers, magazines, and television news programs called the results disappointing and the technique a failure. “It's a bit of a setback,” says neuroscientist John Sladek of the Chicago Medical School, one of the early critics of the NIH decision. “But it should not be the end of research on cell therapy for Parkinson's.” Indeed, in 1995 NIH funded a second study using slightly different transplant techniques; results are expected early next year.

Freed's team randomly assigned 40 patients to two groups: Half had four holes drilled in their skulls through which fetal cells were injected, while the other half underwent “imitation” surgery, in which the same holes were drilled but no cells injected. (The design itself raised questions because of the risks to those in the control group.) One year later, the patients were asked to evaluate the overall severity of their disease on a scale of −3 to +3. By that measure, the two groups reported no significant difference.

However, on a standardized test in which physicians evaluated patients' symptoms while they were off their medicine, the data were more encouraging. After 12 months, those who had undergone the imitation surgery experienced no significant change, but transplant recipients improved by an average of 15%. After the evaluations, patients were told whether they had received cells, and those in the control group had a chance to receive transplants.

Three years after the operation, transplant recipients who were under 60 when they underwent surgery had improved by an average of 38% on the standardized test, and older patients by 14%. But by then, some troubling side effects had also appeared: Five recipients began to show jerky movements typical of Parkinson's patients who become oversensitive to dopaminergic drugs. The condition persisted after the patients reduced or stopped taking the drugs. Freed attributes these effects to a possible overgrowth of the transplanted cells or an oversensitization of dopamine-receiving cells in the region.

“Few in the field anticipated that too much dopamine would be an issue,” says neurosurgeon Thomas Freeman of the University of South Florida in Tampa, an investigator in the second NIH trial. Instead, he says, most researchers have concentrated on encouraging enough cells to survive to produce sufficient dopamine.

Patients in other ongoing studies in Europe and the United States have experienced similar side effects, although none as severe as those reported by Freed, says neurologist Olle Lindvall of the University of Lund, Sweden. In these studies, he says, researchers transplant fresh tissue rather than cultured cells and use different doses and surgical techniques. Lindvall does not think an overgrowth of dopamine-producing neurons caused the side effects. He notes that autopsy data from two transplant recipients in the Colorado-Columbia study who later died—one in a car accident, the other of a heart attack—found between 45,000 and 63,000 surviving cells per patient. Other studies have suggested that as many as 100,000 surviving cells are required for a functional graft, he says.

Freeman and his colleagues hope their study will answer some of these questions. That double-blind trial, which also uses imitation surgeries, includes 34 patients and tests different doses of cells. Patients are not told whether they received cells for 2 years. Last week's report “has put a huge burden on our trial,” Freeman says. “If our trial using different methodologies is negative as well, [continuing the research] certainly will be a bigger uphill battle.”

3. HUMAN CLONING

# Experts Assail Plan to Help Childless Couples

1. John Pickrell

ROME—A plan to create the first human clone announced here last week is drawing widespread condemnation from the scientific community. Unlike previous such pronouncements, however, experts worry that the three researchers who are intent on treading into this moral and political minefield may have the expertise to carry out their plan—with potentially disastrous consequences for both the mother and her offspring. “They want to use humans as guinea pigs, and this is absolutely preposterous,” says Rudolf Jaenisch of the Whitehead Institute for Biomedical Research in Cambridge, Massachusetts.

This is not the first time individuals have threatened to break what has become a taboo in many countries and religions. Three years ago, physicist Richard Seed aired plans to launch a human cloning clinic in Chicago (Science, 16 January 1998, p. 315), and has since vowed that he would clone his wife. In addition, a Canadian cult, the Raëlians, laid out its vision of human cloning (Science, 25 June 1999, p. 2083). Few people took either announcement seriously.

The latest pronouncement comes from a trio composed of Severino Antinori, a fertility expert at the Institute of Clinical Obstetrics and Gynaecology in Rome; Panos Zavos, a reproductive physiologist at the Andrology Institute of America in Lexington, Kentucky; and Avi Ben-Abraham, an American-Israeli biotechnologist whose current affiliation was not revealed. Speaking at a workshop at Antinori's institute, Ben-Abraham said that the team has “unlimited funding”—he declined to reveal the source—and plans to carry out the experiments in an undisclosed Mediterranean country. Ben-Abraham hinted that it could be Israel or an Arab nation, claiming that “the climate is more [receptive to human cloning research] within Judaism and Islam.”

The group wants to use cloning to help childless couples—particularly infertile men—start families. “Cloning may be the last frontier … in our attempts aimed at defeating male sterility,” says Antinori, who is no stranger to controversy: In 1994, he used in vitro fertilization to impregnate a 62-year-old woman. The trio would attempt cloning only for childless couples in which the men produce no sperm, Antinori says. He claims to have 600 such couples on a waiting list.

One of the few scientific details of the project revealed at the meeting was that the group plans to follow essentially the same approach that was used to produce the sheep Dolly: Transplant a nucleus from a somatic cell into an enucleated egg and kick-start the process with a jolt of electricity. Zavos claims that the group has many scientists on board, including animal cloning experts; he refused to reveal their names, citing “security” concerns.

Jaenisch and others have denounced the effort. “What these guys are suggesting is ridiculous,” he says, warning that the rare cloned mammals that survive from hundreds of fertilized eggs often suffer severe health problems. “Many die very soon after or have serious problems, such as kidney and brain abnormalities or no immune system,” he says. There's no reason, Jaenisch adds, to think that such problems—seen in all five mammalian species cloned so far—won't affect human clones. Dolly's creator, Ian Wilmut of the Roslin Institute in Edinburgh, U.K., adds: “We had a lamb born recently which looked perfectly formed, but it couldn't stop hyperventilating; in the end we decided it was kinder to kill it. It turned out that the muscles and arteries leading to its lungs were malformed. I would like to know what they propose to do with a human in a situation like this.”

The mother of a human clone might also be at risk. Mammalian clones are often extra-large, and pregnant mothers become dangerously swollen and frequently miscarry. Antinori's team claims that problems with embryo culture medium could be the cause of this syndrome, and that altering the medium's ingredients could avoid the complication. Wilmut acknowledges that's a possibility but says that until this problem is resolved, human surrogate mothers would be put at great risk. Jaenisch says epigenetic factors may affect a clone's health and account for the high rate of failure in bringing cloned embryos to term. “They cannot screen for epigenetic abnormalities in the same way they can screen for chromosomal aberrations,” he says.

Undaunted, Antinori revealed that the trio would meet in October in Monte Carlo, Monaco, to fine-tune its plan; the researchers hope to start implanting embryos within 2 years. Said Zavos, “The genie is out of the bottle.”

4. GENDER EQUITY

# NSF Program Targets Institutional Change

1. Jeffrey Mervis

Huddled around a campfire in the Colorado Rockies last fall, 30 women engineers plotted how to improve conditions for their academic colleagues. Out of that meeting, part of a 3-day workshop, came the idea for a Women in Engineering Leadership Institute (WELI). The campers' timing couldn't have been better: Last month, the National Science Foundation (NSF) unveiled plans for a new $20-million-a-year program aimed at improving career prospects for women scientists and engineers in academia, and organizers of the nascent institute are already working on a grant proposal. The competition may be fierce. WELI will be competing for one of five to 10 “institutional transformation” awards that NSF hopes to make by this fall as part of its new program, called ADVANCE. The program, which replaces NSF's earlier efforts to tackle the chronic problem of women being underrepresented in science, will also fund fellowships for women just starting or returning to the academic track, and provide leadership awards to enhance institutional efforts already under way (see table). View this table: The case for such a program is easy to make: Women comprise only 22% of the U.S. scientific and engineering workforce, and their growing share of undergraduate and graduate enrollments is not reflected in faculty hiring and movement up the tenure ladder (Science, 21 July 2000, p. 379). A recent study of the University of California system, for example, shows a drop over the last 2 years in the share of new faculty appointments going to women (Science, 2 February, p. 806). But improving those statistics is difficult. ADVANCE's predecessor, called Professional Opportunities for Women in Research and Education, gave out nearly 500 grants over 4 years to women who needed a boost on the road to an academic career. Although satisfied with individual success stories, NSF officials felt that the program wasn't doing enough to remove institutional barriers. They also worried that restricting participation to women could make the program vulnerable to attack by foes of affirmative action. The new program addresses both those concerns by targeting the place where academic women work. NSF hopes that its money will be a carrot for universities to reform their attitudes toward everything from dual-career couples to those needing time off the tenure track. “We're trying to look at the problem at an institutional level, to both help raise their consciousness and give them the tools to change their policies and procedures,” says Norman Bradburn, head of NSF's social and behavioral sciences directorate, which will manage the program. “And one nice aspect of this program is that it's not restricted to women.” Last May, President Clinton announced the program in a Rose Garden ceremony on pay equity for women. But it wasn't until 5 February that NSF finally spelled out the details and issued a call for proposals (NSF 01-69, www.nsf.gov/home/crssprgm/advance). There's an 8 May deadline for the institutional and leadership awards, while fellowship requests are due 21 to 24 August. Deb Niemeier, incoming chair of civil engineering at the University of California, Davis, welcomes the new program, and she sees it as a chance to get WELI off the ground. The nascent institute, a virtual structure that draws on faculty at 10 research universities around the country, plans to submit a proposal for$750,000 a year for 5 years to run workshops, provide mentoring for women academic engineers, and help universities trying to increase the number of women in administrative ranks. “I'm really excited. It's a wonderful opportunity to channel our activities,” says Niemeier, a WELI board member, who organized last fall's NSF-sponsored workshop.

NSF hopes to continue the program at its current level for at least 5 years, although Bradburn says its fortunes are tied to the overall NSF budget. That could mean tough sledding, as President Bush has requested only 1.3% more for the agency in 2002, a shock after NSF's 13.5% boost in the current year (Science, 9 March, p. 1882). At the same time, Bradburn notes that money isn't the real impediment to change. “None of this is a substitute for academic leadership,” he says. “Without a strong institutional commitment, nothing will happen.”

5. NEUROSCIENCE

# Dyslexia: Same Brains, Different Languages

1. Laura Helmuth

Pity the poor speakers of English. New research suggests that they may be especially prone to manifest dyslexia, the language disorder that makes reading and writing a struggle, simply because their language is so tricky.

The distinctive pattern of spelling and memory problems that characterizes dyslexia has a strong genetic basis, suggesting that some neurological oddity underlies the disorder. But there appears to be a cultural component to the disease as well, because dyslexia is more prevalent in some countries than others; for instance, about twice as many people fit the definition of dyslexic in the United States as in Italy. Researchers have suspected that certain languages expose the disorder while others allow dyslexics to compensate. Now a brain imaging study backs this theory up.

A multinational team of researchers used positron emission tomography (PET) scans to observe brain activity in British, French, and Italian adults while they read. Regardless of language, the team reports on page 2165, people with symptoms of dyslexia showed less neural activity in a part of the brain that's vital for reading.

“Neurologically, the disease looks very much the same” in people who speak different languages, says neurologist Eraldo Paulesu of the University of Milan Bicocca in Italy. “Therefore, the difference in prevalence of clinical manifestations [among different countries] must be attributed to something else.” The researchers blame language.

English consists of just 40 sounds, but these phonemes can be spelled, by one count, in 1120 different ways. French spelling is almost as maddening. Italian speakers, in contrast, must map 25 different speech sounds to just 33 combinations of letters. Not surprisingly, Italian schoolchildren read faster and more accurately than do those in Britain. And it's no surprise that people have a harder time overcoming reading disorders if their language, like English or French, has a very complex, arbitrary system for spelling. “English comes with a built-in deficit,” says education researcher Ken Spencer of the University of Hull in the United Kingdom.

Diagnosing a learning disability is notoriously subjective. Lack of access to good education and other social factors probably account for most reading disorders, says psychologist Richard Olson of the University of Colorado, Boulder. To avoid some of these issues, the researchers tested university students—people who have served plenty of time in classrooms and don't lack intelligence or willpower. The English and French dyslexic students have compensated for their disorder and are “very successful people,” says study co-author Ute Frith of University College London, even though they need more time when taking exams and make frequent spelling mistakes.

Finding dyslexic Italian subjects was trickier, because practically no university students have been diagnosed with the disorder, Frith says. The team tested 1200 students and identified 18 with a pattern of verbal memory problems (such as difficulty remembering telephone number-like strings of digits) and slowed reading typical of the diagnosed dyslexics in France and the U.K. The “dyslexic” Italian students weren't told how they scored, but some (aware that they were participating in a dyslexia study) volunteered that they'd had trouble learning to read as children, Paulesu says.

Even though the Italian subjects were unaware of and unhindered by having dyslexia-like reading skills, under the PET scan they looked just like British and French students who struggled with reading. Compared to normal readers, dyslexics from all three countries showed less activation in parts of the temporal lobe while reading. The underutilized areas are familiar to neurologists: Patients with strokes in this area often lose the ability to read and spell, even though they still speak fluently.

The researchers aren't sure why the dyslexics seem to access this brain area less than normal readers do. Other PET studies and scattered neuropathological reports have led to speculation that, in general, dyslexics have fewer neural connections among cells in this region. Although most researchers think that's plausible, little consensus exists on more detailed explanations of how or why dyslexics' brains are different from those of normal readers, Olson says. His research suggests that genetic factors account for about half of someone's risk of developing dyslexia, although no single gene is likely to be to blame.

This research doesn't supply ready solutions for how to help dyslexic students overcome their reading disability, Paulesu says, short of moving to Italy, Turkey, or Spain, where spelling is simple and straightforward. So sympathize when English- or French-speaking students complain about having to memorize arbitrarily spelled words; they're right to feel wronged.

6. CELL BIOLOGY

# How Bacterial Flagella Flip Their Switch

1. Dennis Normile

Bacteria move with the agility of a tailback in American football, first cutting left, then right—all propelled by their whirling flagellar tails. Exactly how the tiny creatures achieve their feats of broken-field running has long been a mystery. New results now provide a probable solution to the puzzle.

When bacteria “run,” the rotation of their flagella, long, whiplike filaments that project from the cells, is driven by motors located at the flagellar base in the bacterial membrane. The flagellar filaments are helical, and when they wind in the left-handed direction, they rotate counterclockwise. To change direction then, the bacterial cell momentarily reverses the direction of the motor, generating a torque on the flagellar filaments that flips them into right-handed helices. This causes the bacterial cell to “tumble,” or change its orientation for its next run.

The mystery concerns how flagellin, the protein that makes up the filaments, achieves this dramatic structural shift. An x-ray crystallographic study of the protein described in this week's issue of Nature by Keiichi Namba of Matsushita's Advanced Technology Research Laboratories in Kyoto, Japan, Fadel Samatey of the Protonic NanoMachine Project, also in Kyoto, and their colleagues provides the first good look at a possible answer. It suggests that a sharp change within a very small “switch” region of flagellin is all that it takes to flip the left-handed flagellar helix into the right-handed form.

Previous structural studies, using the electron microscope and other methods, had characterized the bacterial flagellar filament to a resolution of 10 angstroms. These showed that it is a tubular structure made up of 11 protofilaments, each formed by stacking together numerous molecules of flagellin. This protein comes in either left- or right-handed states, or L or R types, and all the molecules in a given protofilament are in the same state.

But the left- and right-handed protofilaments aren't just mirror images of one another. The repeat distance, which is the distance from any given point on one flagellin molecule to the corresponding point on the molecule above or below it, is less—by 0.8 angstroms—for the R-type protofilament. As a result, a flagellar filament containing both L and R protofilaments isn't straight but supercoils into a helix. For example, the Salmonella filament, which contains nine L-type and two R-type protofilaments, twists into a gentle left-handed helix. For tumbling, two of the L protofilaments are switched to the R type, and that's enough to give the filament a right-handed coil.

Researchers have long wanted to see this structural change, but that requires x-ray crystallography and they were stymied in their efforts because flagellin doesn't crystallize. (The proteins link together in polymers instead.) In the current work, Namba, Samatey, and their colleagues got around that problem by clipping off the ends of the protein, thus removing the regions that form links with other flagellin molecules.

But yet another hurdle appeared: Only the R type of the protein crystallized, making it impossible to directly compare the structures of the two forms. Namba suggests that the longer L structure exists only in the filament, produced by the interaction between the protofilaments. “I don't think we will ever crystallize the L type of the protein,” he says. That meant the researchers had to find another way to observe the change of the L form of flagellin to the R form.

So after determining the x-ray structure of the R-type protein, they used a computer simulation to stretch it, in 0.1-angstrom steps, into the longer L type. Up to a point, the stretching was accommodated by elastic strain throughout the structure. But then, a key hairpin fold in the protein snapped into a new position, pushing two regions of the protein farther apart and producing the L form.

While the simulation strongly suggests that Namba and his colleagues have spotted the flagellar switching mechanism, he cautions that his group has more to do to confirm that and to understand the interactions among the protofilaments. Even so, Hirokazu Hotani, a biophysicist at Nagoya University in Japan, describes the achievement as “very difficult to accomplish.” He adds that the finding is important not only because it gives a clearer picture of how bacteria control their movements, but also because flagellin is the only protein known to undergo such a large degree of conformational change. And there could be a practical payoff in using the mechanism as a switch in nanomachines, although just how it might be incorporated into practical devices remains to be seen.

7. ASTRONOMY

# Stars Rise From Ashes in Globular Cluster

1. Govert Schilling*
1. Govert Schilling is an astronomy writer in Utrecht, the Netherlands.

If you think cleanliness is next to godliness, avoid globular clusters. These huge, spherical concentrations of millions of old stars are among the filthiest places in the universe, according to a new study by Italian astronomers. Like houses in an industrial area, the stars in a globular cluster are polluted by the exhausts from nearby chemical plants.

In a paper to appear in Astronomy and Astrophysics, Raffaele Gratton of the Astronomical Observatory of Padua and his colleagues describe their discovery of polluted stars in a globular cluster known as NGC 6752, some 13,000 light-years from Earth in the southern constellation Pavo the Peacock. The cluster is about 100 light-years across and contains millions of stars, which are hundreds of times closer together than the stars in the solar neighborhood. From the Southern Hemisphere, it can easily be seen with a pair of binoculars.

Gratton and colleagues used a sensitive spectrograph on the European Southern Observatory's (ESO's) 8.2-meter Kueyen telescope in Chile (part of the Very Large Telescope) to study the chemical makeup of 18 dwarf stars—stars about the size of the sun—in NGC 6752. Because the stars in a globular cluster are believed to have formed simultaneously from the same cosmic ingredients, you would expect them to have a similar spectral fingerprint. Instead, the team found huge star-to-star variations in the composition of the stars' outer layers.

For giant stars, that wouldn't be too surprising. Their high internal temperatures churn up their insides vigorously enough to carry the “ashes” of a star's nuclear burning processes from its core to its surface. But that can't be happening in dwarf stars, says team member Luca Pasquini of ESO in Garching, Germany, because they're not hot enough. So what causes the anomalous abundances?

One clue comes from the observations that the stars in the cluster show “anticorrelations” between certain elements. Dwarf stars that are high in oxygen tend to be low in sodium, and vice versa. A similar relation holds for magnesium and aluminum. Sodium and aluminum form relatively late in a large star's life cycle, when oxygen and magnesium tend to be depleted.

The astronomers concluded that the dwarf stars had picked up their heavy elements from hot, massive, short-lived stars that perished billions of years ago, when the cluster was young. Mixing carried the processed material from the massive stars' cores to their outer layers. Later, the dying stars ejected those layers as planetary nebulae—vast, slowly expanding shells of gas that polluted the space between the stars in the cluster and contaminated nearby dwarf stars. Pasquini thinks the “dirty” stars may have become 10% to 30% more massive because of the stellar pollution.

The new observations challenge a scenario proposed 20 years ago by Gary Da Costa of the Australian National University in Canberra and Peter Cottrell of the University of Canterbury in Christchurch, New Zealand. Da Costa and Cottrell thought the giant stars had polluted the cluster gas even before the dwarf stars formed out of it. Pasquini says it's hard to see how dwarf stars in a globular cluster could have formed much later than giant stars, as that model requires. But he acknowledges that the evidence so far is inconclusive. “We really need a better statistical sample to distinguish between the two models,” he says. Da Costa agrees: “This is an important discovery. What needs to happen now is further work to understand how this process works [in detail].”

Stellar pollution has never been observed before, except in binary star systems. In the galaxy at large, astronomers say, stars are too far apart to intercept the ashes from old stars before they slowly disperse into space. But in massive globular clusters, stars are much closer together, and their combined gravity is strong enough to keep most of the “space pollution” in the neighborhood. Gratton's team also studied eight dwarf stars in a less massive cluster, NGC 6397, but found no pollution there, probably because the exhausts of dying stars have escaped the cluster altogether.

Pasquini says it's unclear how contaminants might be affecting the life cycles of the tainted stars: “We are just starting to investigate this. We'll have to run the [stellar evolution] models to see in detail what's going on.”

8. INNATE IMMUNITY

# Ancient System Gets New Respect

1. Trisha Gura*
1. Trisha Gura is a science writer in Cleveland, Ohio.

The antimicrobial peptides produced by the so-called innate immune system are not only widely effective but may provide a new source of antibiotics

In the deadly struggle between invading microbes and immune system defenses, antibodies and killer cells tend to get all the attention. But now appreciation is growing for another defense system that, until recently, has been largely ignored. Called innate immunity, the system is a kind of chemical barrier that organisms from insects to humans deploy to stop dangerous microbes at their first point of contact—the skin, mucous membranes, or other surfaces. This first line of defense has recently become a subject of intense study in several labs, because it represents a promising source of potential new antibiotic drugs. But the basic, primitive system is also turning out to be interesting in its own right.

Unlike the so-called acquired immune system, with its disease-fighting cells and antibodies, innate immunity depends on peptides and small proteins to fight off dangerous microbes. As genomic information has flooded out of sequencing labs within the last few years, researchers have identified hundreds of these peptides from a broad range of species. They are also now working out exactly what trips these defenses and how they specialize to fight off the pathogens that attack each kind of organism.

As expected, the sugars and proteins that coat certain microbes are good triggers, but so too are various molecules that the infecting pathogens produce. Innate immunity “is quite a well-designed surveillance system whose extent, depth, complexity, and power have not been fully appreciated,” says pediatrician and molecular biologist Michael Zasloff of Magainin Pharmaceuticals in Plymouth Meeting, Pennsylvania.

These defensive peptides tend to work very differently from conventional antibiotics, which generally block a crucial protein in an invading microbe. The peptides are less subtle killers: They punch holes in an invader's membranes or disrupt its internal signaling. Some even appear to pump up the host's own immune cell activity. As a result, most of the peptides are effective against a broad range of germs.

What's more, because the peptides home in on basic physical properties such as the overall charge on a microbial membrane, pathogens may be less likely to develop resistance. “It is hard for bacteria to change the physical properties of their membranes,” notes Robert Hancock, a microbiologist at the University of British Columbia in Vancouver. “To develop resistance would entail that the microbes change their cells in very fundamental ways.”

These attributes have caught the attention of the pharmaceutical industry. Clinical trials are already under way to see if the germ-killing peptides can combat everything from acne to catheter-related infections, the lung infections that plague cystic fibrosis patients, and the oral ulcers that often follow radiation and chemotherapy treatments for cancer. Early results look promising, although researchers caution that toxicity issues loom large, especially for molecules intended for internal use as opposed to topical application.

Although research on antimicrobial peptides has intensified in the past 5 years, its roots can be traced back to the 1980s, when Hans Boman and his colleagues at the University of Stockholm, Sweden, sighted an intriguing phenomenon in pupae of the silkworm moth, Hyalophora cecropia. The researchers knew that insects, which do not have immune systems like those of higher animals, have some sort of innate immunity, but they were hazy about the details. To get at those, Boman's group inoculated the worms with bacteria and watched as they unleashed a batch of defensive agents. From that array, the team isolated two novel peptides, dubbed cecropin A and B, after the name of the moth.

The compounds, made from a string of 35 to 39 amino acids, appeared similar to a potent peptide called melittin found in bee venom. But Boman's curiosity was piqued when he noted that cecropins only kill bacteria such as Escherichia coli, while leaving moth cells unharmed. In contrast, melittin kills both bacteria and the cells of higher organisms. Boman thought he might be on to something interesting.

Over the next 2 decades, his group and others showed that antimicrobial peptides are extremely widespread in nature. They turned up in insects and eventually in higher species such as frogs and mammals, suggesting that innate im-munity has deep evolutionary roots. Some, such as cecropin and attracin—another moth peptide identified by the Boman team—kill E. coli and certain other gram-negative bacteria (bacteria that fail to take up the so-called gram stain). Since then, researchers have identified other peptides that kill gram-positive bacteria. And still others, such as a peptide called drosocin, identified in the fruit fly by Jules Hoffmann's team at the Institute of Molecular and Cellular Biology at CNRS in Strasbourg, France, primarily attack fungi.

Hoffmann recalls that his group and others began studying fly defenses because even though the insects lack the cell-based immune systems seen in higher organisms, they are “extremely resistant to microorganisms.” The large number of antimicrobial peptides found in insects could go a long way toward explaining their resistance, as well as that of other organisms lacking advanced mammalian immune systems. Take frogs, for example.

In the late 1980s, Zasloff, then at the National Institutes of Health (NIH) in Bethesda, Maryland, found that frog skin harbors a potent defense system. When stimulated by injury or microbes, the animals sweat out large amounts of antibiotic peptides that Zasloff called “magainins” after the Hebrew word for “shield.” In 1988, he left the NIH to launch a company, also named Magainin, to commercialize his discovery.

But even animals with the most advanced immune systems produce defense peptides. In 1988, Boman's group purified cecropin-like peptides from the intestines of pigs. Robert Lehrer, along with Tomas Ganz and Michael Selsted of the University of California, Los Angeles, School of Medicine, discovered that other mammals, including rabbits, cows, and even humans, make similar antimicrobial peptides, which were named defensins. (Some insect and plant antimicrobial peptides are also called defensins, although their structures are different from those of the mammalian peptides.) Most recently, a team led by Yong-Lian Zhang at the Shanghai Institute for Biological Sciences in China identified a new defensin in the reproductive tract of male rats that may aid sperm development and help guard against infection (Science, 2 March, p. 1783).

In fact, in higher animals, cells in and around the linings of the intestines and of the respiratory and urogenital tracts commonly churn out the peptides. So do phagocytic cells called neutrophils, which use defensins to kill the microbes they engulf. The potent molecules are produced at low levels all the time. But Zasloff and others, such as Ganz, have found that injury or inflammatory signals such as interleukin-1β can cause defensin production to skyrocket in the alimentary tract of cows. “We have the basis of a shield that is not just sitting there passively but is capable of induction,” Zasloff says.

The peptides are proving to be vital for survival. For instance, in work reported on page 113 of the 1 October 1999 issue of Science, Carole Wilson of the Washington University School of Medicine in St. Louis and Andre Ouellette of the University of California, Irvine, genetically engineered mice to be defective in an enzyme needed to activate a defensin produced by cells located at the base of the tiny finger-like projections lining the small intestine. Researchers had known for decades that the cells maintain a “sterile” microenvironment, but they didn't know how. The knockout mice gave some clues. They seemed normal at birth, but when the animals were infected with disease-causing strains of either E. coli or Salmonella bacteria, they succumbed more readily to illness, even death. The same susceptibility to infection occurs in flies that have had their peptide defenses knocked out.

## How the peptides work

Even as the number of antimicrobial peptides has continued to swell, researchers have been trying to figure out how they work. Boman and Stockholm colleague Dan Hultmark showed early on that some, such as the cecropins and insect or mammalian defensins, seem to home in on and rupture the negatively charged membranes of bacteria. But because the peptides pick out some bacteria and not others with a similar charge, researchers are sure there is more to the story. They also note that the peptides don't damage the cells of the organisms that make them. “It's amazing,” says Zasloff. “Despite being so broad spectrum and working on membranes, these antimicrobial peptides don't harm the host.”

So far, the explanation for this remains a mystery. Some investigators have suggested that host cells may carry a lower negative charge than the microbes. But charge differences alone can't explain how some peptides can distinguish among members of the same class of germs. Another possibility is that rather than just recognizing net charge, the peptides home in on patterns of charge, say in the sugars or amino acid chains that sit on a microbe's surface.

In one special case, researchers do have a clear idea of how a peptide works. In 1994, Hoffmann and his colleagues identified an antimicrobial peptide in the European sap-sucking bug, Pyrrhocoris apterus, that acts very differently from previously discovered antimicrobial peptides. Those tend to kill quickly and to act either on gram-positive or gram-negative bacteria. But pyrrhocoricin takes anywhere from 6 to 12 hours to work and kills representatives from both classes of bacteria. That suggests that the peptide targets a protein common to different classes of bacteria. In October, Laszlo Otvos and his colleagues at the Wistar Institute in Philadelphia fingered that target: a protein called DnaK that normally works to preserve the three-dimensional shapes of many proteins in the face of heat. Interfering with DnaK destroys many enzymes involved in normal microbial housekeeping, thus killing the organisms.

Researchers are also getting a handle on just what triggers the production of the peptides. While investigating how fungi spur the production of peptides such as drosocin, Hoffmann and others found that elusive molecular signals produced by the microbes trigger the so-called “Toll” pathway to turn on the genes needed to synthesize the peptides (Science, 25 September 1998, p. 1942). The pathway is named after the Toll protein, a receptor that sits in the cell membrane, where it can pick up signals from the outside. Until then, Toll was known only for its role in regulating early fly embryo development.

Researchers have since learned that the fly carries at least eight Toll receptor variants, each of which seems to be triggered by a specific class of attacking microbe. One Toll family receptor can pick out, say, fungi, while another detects gram-positive bacteria. The activated receptor then signals the cell nucleus to make the appropriate peptide to combat that type of microbe.

Researchers, including Charles Janeway of Yale University in New Haven, Connecticut, and Shizuo Akira and colleagues at Osaka University in Japan, have shown that something similar happens in human cells. For example, Janeway demonstrated that endotoxin, a lipid- and sugar-laden toxin made by gram-negative bacteria such as E. coli and Salmonella typhimurium, prompts Toll-like receptors in human cells to trigger the production of mammalian defensins. To date, at least nine human Toll-like receptors (TLRs) have been identified, each of which may recognize a distinct pattern of molecules characteristic of a class of microbes. For example, TLR4 seems to trigger a defense cascade in response to endotoxin, while TLR2 recognizes components from the walls of gram-positive bacteria such as Listeria monocytogenes.

## Into the clinic

Whatever drives the production of so many different antimicrobial peptides, they are providing a bonanza of potential agents for treating human infections. Still, the researchers caution against overoptimism: If antimicrobial peptides are not specific enough, they can overzealously target human cell membranes or rupture red blood cells, leading to rapid toxicity, even death, for the host, as has been found in animals.

Investigators are now exploring several ways to circumvent these toxicity problems—and they are seeing some signs of success. One possibility is to screen for natural peptides that spare host tissues. The fruit fly peptide drosocin, for example, may have the type of specificity desired. “If you take peptides, which have been selected by nature for hundreds of millions of years, you're bound to find some that are not cytotoxic,” Hoffmann says.

To develop that idea—and possible candidates—the molecular biologist has spun off a company in Strasbourg called EntoMed. Headed by Hoffmann's former Strasbourg colleague Jean-Luc Demarcq, the company is now conducting studies in rodents to see whether drosocin can combat two fungi, Candida albicans and Aspergillus fumigatus, that cause often fatal infections in patients undergoing chemotherapy or organ transplants. Demarcq says the company hopes to move into human studies by 2002.

Meanwhile, Intrabiotics in the San Francisco Bay area of California is conducting clinical trials to see whether a peptide spawned by pig neutrophils can treat mouth ulcers in patients undergoing chemotherapy or radiotherapy for head and neck cancers. Those trials are in their final stages, and Intrabiotics has also begun testing whether the peptide in aerosol form can prevent the Pseudomonas infections that often kill cystic fibrosis patients and the pneumonia that afflicts patients on ventilators.

Other researchers, including Hancock in Vancouver and Lehrer in Los Angeles, are trying to produce safer peptides by modifying some of the more potent ones to dampen their activity. They either synthesize the variants from scratch or use recombinant DNA technology to produce modified peptides in bacteria. The researchers add the peptide candidates to test tubes filled with various strains of bacteria and sheep red blood cells, and look for those that kill the bacteria without damaging the blood cells.

Already, this strategy has produced three peptides that have moved into clinical trials. None are being injected systemically, however, because the researchers consider that route of administration too risky right now. Micrologix Biotech Inc. in Vancouver is currently conducting a late-stage clinical trial of one of these peptides, aimed at seeing whether it will prevent the catheter-related infections caused by many microbes, including the bacteria Pseudomonas aeruginosa and Staphylococcus aureus, a gram-positive superbug that now resists most common current antibiotics. Company researchers are also testing whether the peptide will cure severe acne and prevent acute S. aureus infections.

Zasloff's team at Magainin is taking another tack: trying to get the body to produce more of its own defensive peptides, which presumably won't damage the cells of their native host. The team began by looking for potential triggers in disease-causing bacteria such as Staphylococcus and Pseudomonas. Those pathogens did not boost β-defensin gene expression. But much to Zasloff's surprise, beneficial organisms, such as bakers' yeast and Lactobacillus, found in yogurt, did. The team then purified the active component from bakers' yeast, which turned out to be the amino acid isoleucine. The amino acid may work by binding to a Toll-like receptor in people, the researchers reported in the Proceedings of the National Academy of Sciences in November.

Although isoleucine is a necessary building block of the body's proteins, humans cannot make the amino acid on their own and must get it in their diets. Zasloff suggests that isoleucine is a good signal of an invading pathogen, because the microbes do make the amino acid and would drive isoleucine levels up during infection—either by secreting it or by breaking down host tissue. If so, then something as simple as an amino acid supplement might boost immunity. But it remains to be seen whether such a strategy will be successful in designing actual therapies.

Even if the peptides now being tested don't pan out, there are plenty of other candidates waiting in the wings. “It is becoming clear that the hard-wiring of an animal, even the mammal, can be fine-tuned to very carefully distinguish between different classes of microbe,” Zasloff says.

9. INNATE IMMUNITY

# Engineering Protection for Plants

1. Trisha Gura*
1. Trisha Gura is a science writer in Cleveland, Ohio.

Animals aren't the only creatures that need potent defenses against microbial pathogens. Plants need them as well. Now, researchers are trying to use antimicrobial peptides originally found in insects and other animals (see main text) to boost plants' ability to resist infectious bacteria and fungi. Current targets include the potato blight fungus, which caused the great Irish famine of the mid-1800s, and the bacterium that causes soft rot in stored potatoes.

In previous efforts, researchers tried to bolster plant defenses by rubbing antimicrobial peptides from insects on their leaves or by genetically engineering plants to make the peptides internally. But those approaches failed, because the insect peptides proved toxic to the plants. To get around this problem, plant geneticist Santosh Misra and microbiologist William Kay of the University of Victoria in British Columbia tried making less toxic peptides, using a “mix and match” strategy devised with the help of microbiologist Robert Hancock of the University of British Columbia in Vancouver.

In work reported in the November issue of Nature Biotechnology, Misra constructed hybrid DNAs that combined DNA sequences from moth and bee peptide genes. Test tube studies showed that some of the resulting peptides retained the ability to kill bacteria without harming potato cells. Misra's team then shuttled the DNA encoding the best of these into Russet Burbank and Desiree potatoes with the help of Agrobacterium tumefaciens, a bacterium that injects genetic material into plant cells.

Once the genetically altered potato plants reached 2 weeks of age, the British Columbia team exposed them to the blight-causing fungi. The microbes infected unaltered potato plants within 11 days, but the modified Desiree plants grew just fine, even through a cloud of fungi clustered about their roots. The genetically altered Russet Burbank plants didn't do quite as well, however. Although they proved more resistant to infection than controls, they grew yellow, curly leaves and smaller, branched tubers than did normal Russets. The researchers don't know what accounts for this difference.

In a second set of studies, the researchers exposed sliced potato tubers from both genetically modified and control plants to Erwinia carotova, the bacterium that causes soft rot. Exposed controls were quickly infected and lost 60% of their weight. In contrast, tubers from the altered potato plants appeared similar to uninfected controls. They could also withstand storage for at least 6 months, while controls succumbed to soft rot. What's more, when fed to mice, the peptide-producing potatoes were as benign to the rodents as control tubers.

In similar studies, Ai-Guo Gao's team at Monsanto in St. Louis reported in the December issue of Nature Biotechnology that Russet potatoes expressing modified alfalfa antifungal peptides showed enhanced resistance to the fungus that causes “early dying” disease. The Monsanto team also found that the modified plants retained their resistance to the pathogen in field trials.

The British Columbia team is now testing its defense-enhanced potatoes in field trials, and Misra estimates that the tubers could make it to supermarket shelves in 3 to 5 years—provided they aren't stalled by public reluctance to accept genetically modified foods. Indeed, such concerns have caused plant pathologist Herbert Aldwinckle at the Geneva campus of Cornell University to decide against commercializing a genetically modified apple strain his group produced. The trees carry a variant of the antimicrobial peptide attacin, which has worked well at protecting them from the bacterial disease fire blight. “I think these antimicrobials do have the potential to work in crops in the field,” Aldwinckle says. But even if the crops are safe, he says, “we are very concerned [that] anything we put into apple might reflect badly on the image of apple as a crop.”

10. BIOPHYSICAL SOCIETY MEETING

# Crossover Research Yields Scents and Sensitivity

1. R. John Davenport

BOSTON, MASSACHUSETTS—Over 5000 scientists gathered here last month for the 45th annual meeting of the Biophysical Society.* The meeting brought together physicists, chemists, biologists, and others to discuss how physics can be used to address fundamental biological problems in new ways.

## Watching a Virus Get Stuffed

For decades, scientists have used viruses that infect bacteria, known as phages, as models for viruses that infect humans. Yet despite intense scrutiny, how phages manufacture more phages remains mysterious. Now, for the first time, biophysicists from the University of California (UC), Berkeley, have caught a phage in a key act of self-assembly. Using microscopic beads and laser light, they watched the virus stuffing its genome into a protein shell, and they played tug-of-war with its DNA. The results reveal that the microscopic winch responsible for DNA packaging is the most powerful molecular motor ever measured.

“They've done a remarkable job,” says Dwight Anderson, a phage biologist at the University of Minnesota, Minneapolis. “For a long time, we've been using an integrated biochemical and genetic approach [to understand packaging], but what we really needed is the biophysics.”

The DNA packaging machinery of the phage, known as phi29, has been studied extensively. To stuff DNA into the virus's five-sided head, a phage protein called gp16 brings the DNA to a protein-RNA complex at the base of the head. The gp16 burns up adenosine triphosphate (ATP) fuel and, together with the protein connector, pulls the DNA into the head. Enlarged 600,000 times, the whole process would resemble stuffing a strand of spaghetti into a matchbox. But although its biochemistry is well known, the mechanical details of how the virus pulls the DNA into the head and organizes it in just a few minutes are sketchy.

To get a handle on the problem, Sander Tans of UC Berkeley literally took hold of the phage while it packaged. He and colleagues Doug Smith, Steve Smith, and Carlos Bustamante strung a phage head and DNA between two small plastic beads, attaching them with a protein glue. Using suction, the researchers held the bead with the phage head on the end of a micropipette. The DNA bead, meanwhile, was caught in a laser trap—a specially focused laser beam that holds the bead stationary. By pulling on the DNA-phage head assembly and measuring the deflection of the laser trap's light, the researchers could measure the force on the bead—and thus on the biological molecules attached to it. After assembling the complex, they added ATP and watched.

As the motor pulled the DNA into the phage head, the researchers saw the beads moving closer and closer together. The DNA packaging began smoothly, at a constant rate of 100 base pairs per second. But once the head was half full of the DNA, the pressure began to build up and the motor to slow down. Tans and his co-workers used the optical trap to pull against the packaging motor. They could bring the motor to a stop by resisting with a force of 57 piconewtons—the highest “stalling force” ever seen for a molecular motor. From that force and the volume inside a phage head, Tans calculates that the pressure inside a fully stuffed head is 15 megapascals. “That's the pressure inside an oxygen bottle,” he says. He conjectures that the high pressure may help the phage inject its DNA when it attaches to the outside of a cell it's going to infect.

A thorough understanding of packaging, Anderson says, could reveal a new target for novel antiviral drugs. Other scientists are also excited by the application of biophysical techniques to the problem. Without them, “we would never learn what a strong engine the virus packaging machinery has,” says Roman Tuma of the University of Helsinki in Finland.

## A Sniff in Time Scents Mines

Natural olfactory systems, such as the one inside the nose of a dog, can distinguish tens of thousands of different smells and can learn new ones. Artificial noses are much less versatile. Most of those developed to date are geared toward detecting a particular class of smell, such as spilled gasoline or rotting food. The reason is that a living nose contains perhaps 1000 different chemical receptors for identifying smells, a number no machine has come close to matching.

Drawing on their years of studying the olfactory system of the salamander, neuroscientists are now working around that numbers gap. John Kauer and Joel White of Tufts University School of Medicine in Boston have built a device that, Kauer says, employs “about 20 different attributes” of natural noses. Like many similar machines, the device uses an array of sensors, each of which consists of a fluorescent dye embedded in one of many subtly different polymers. When an odor molecule diffuses into the matrix, it changes the fluorescent properties of the dye. The resulting pattern of fluorescence on the entire array gives a signature of the molecules in the air sample.

Unlike other array-based noses, though, the Tufts Medical School nose “sniffs” air samples by periodically puffing air across its sensors. Monitoring how sensors change over time, Kauer says, enables it to discriminate among more odors with the same number of sensors. Other features drawn from nature include feedback mechanisms to keep the sensors from getting saturated and odor-analyzing algorithms based on olfactory nerve circuits.

The device can detect a panoply of odors, Kauer says, including one that is both faint and grim. With funding from the Defense Advanced Research Projects Agency and the Office of Naval Research, Kauer has investigated whether the artificial nose can be used to detect landmines in the field. Most of the estimated 100 million landmines buried throughout the world contain TNT (trinitrotoluene). Dogs can scent landmines, probably by cueing on the byproduct DNT (dinitrotoluene). But an artificial device wouldn't get fatigued like a four-legged nose and could be operated remotely (Science, 3 September 1999, p. 1476).

To see how the artificial schnoz would fare, Kauer tested it “nose-to-nose” against dogs in a special chamber. Kauer's nose proved about 10 to 15 times less sensitive than a dog's, which can detect 1 part per billion or less of DNT and related compounds. In a field test, the nose detected rough signatures of landmines but couldn't locate them precisely, Kauer says—probably because varying background odors cloak vapor emitted by landmines.

Locating landmines is a difficult problem, says chemist David Walt of Tufts University, Medford, and whether artificial noses will ever be up to the task remains to be seen. But, he adds, “there have been some incredible advances.” Kauer says that those advances could be applied to other, more tractable problems, including monitoring complex environments such as the inside of airplane cabins for toxic fumes.

## Old Movies Reveal New Moves

As a cell grows, moves, or divides, it repeatedly has to tear down and rebuild parts of its own structure called the cytoskeleton. Amid the confusion of a living cell, scientists can get only fleeting glimpses of that process of creative destruction. But a twist on an off-the-shelf image-processing technique may help them squeeze new kinds of data out of old cytoskeletal images.

The technique—a computer algorithm called cross-correlation—was developed to detect patterns in signals of many kinds. Zachary Perlman and colleagues at Harvard Medical School in Boston realized it could also be used to study mitotic cell division. During mitosis, protein filaments called microtubules form a network known as the mitotic spindle, which aligns paired chromosomes down the middle of the cell. The microtubules pull the chromosomes to opposite sides of the cell so accurately and efficiently that, when the cell splits in two, each of the resulting cells has a full complement of chromosomes. Scientists know that somehow the microtubules migrate from the center of the cell toward the poles in the process, but many important details of the process remain murky.

In 1998, cell biologists Ted Salmon and Clare Waterman-Storer of the University of North Carolina, Chapel Hill, discovered a new way to track microtubule movement. Using a standard technique, they assembled microtubules from subunits of the protein tubulin, some of which were labeled with fluorescent tags. Because the researchers used less tagged tubulin than usual, the microtubules wound up speckled with glowing spots. By carefully watching the speckles move in a mitotic spindle, the scientists could track the gyrations of the microtubules in their cellular dance.

Interpreting the motions was another matter. Data were hard to analyze; it could take weeks to convert just a couple of spots to hard numbers. Spindles would drift around the microscope slide. Speckles would move in and out of focus and then fade.

To improve on the method, Perlman and his colleagues Tarun Kapoor and Tim Mitchison set out to find a better way to analyze the images. Using cross-correlation, they took successive images of a moving spindle and rotated them into alignment. By comparing spots on the altered images, they could determine how fast and in which direction the spots were moving. “Cross-correlation is not anything that would shock someone in image processing,” Perlman says, but applying it to cell biology has allowed him to extract quantitative data from a previously qualitative experiment.

The technique is useful for more than mitosis, Perlman says. When eukaryotic cells move along a surface, polymerization of other cytoskeletal structures called actin filaments pushes the edge of the cell forward. Perlman has already used cross-correlation to analyze movies of cells with fluorescently labeled actin fibers.

Waterman-Storer, now at the Scripps Research Institute in La Jolla, California, suspects that cross-correlation may have trouble accurately comparing complicated patterns of speckles. But Salmon predicts that the technique will prove to be a powerful way to mine new information from speckle images. “There's a great need to have intelligent computer image-processing technology developed” to analyze such images, he says.

• * 17 to 21 February 2001, Hynes Convention Center, Boston, Massachusetts.

11. RESEARCH COLLABORATIONS

# NASA Lab Offers Land to Lure Research Partners

1. Andrew Lawler

Ames Research Center hopes to strengthen its programs by having universities and industry set up shop on its Silicon Valley campus

MOUNTAIN VIEW, CALIFORNIA—An aging federal research lab here on the outskirts of Silicon Valley hopes to use some of its prime real estate to lure universities and industrial companies to become its neighbors and partners in a unique research campus. If successful, the arrangement could help NASA's Ames Research Center over the next decade to become a major player in the hot new fields of information technologies, astrobiology, and nanotechnology. In return, its academic and business partners would get a window into the technology-rich region.

Built on the eve of World War II as an aircraft research laboratory, Ames, which sits on 800 hectares of land just west of San Francisco Bay, has spent the past decade struggling to gain a beachhead in high-tech areas such as information technologies and to attract the needed talent. Last year, officials at NASA headquarters shot down the lab's proposal for an industry-built lab dedicated to astrobiology. Now it is negotiating deals with aerospace giant Lockheed Martin and three universities—Carnegie Mellon, San Jose State, and the University of California, Santa Cruz—to build more than 1 million square meters of office and lab space over the next decade. Ames managers hope to break ground on the center, called the NASA Research Park, in early 2002.

The plan must still clear several hurdles, including potential environmental problems, a likely change in NASA leadership, and the current economic slump. But officials think the new facilities—and those who will work in them—will greatly enhance Ames's research capacity. “We get fresh blood and the chance to interest them in our mission,” says Nancy Bingham, special assistant to the Ames director. The prospective partners agree. “It's a win-win for everyone,” says Don Fulop, Lockheed's vice president of business development.

The park would consist of separate university, industry, historical, and museum areas. Each university would provide the funds to build and operate its own lab and office space. Lockheed would also build Ames a $40 million Laboratory for Advanced Science and Research (LASR) in exchange for use of a large parcel of land. The 10,000- square-meter lab will be designed to house a multidisciplinary array of researchers from NASA, academia, and industry. Carnegie Mellon has the firmest plans to date among the three universities. It hopes to raise$150 million to establish its presence in Silicon Valley, says James Morris, dean of the computer science school. The park is a better, cheaper alternative to developing its own campus in the region, says Morris, who adds that the university may eventually spend $400 million over the next decade to beef up its advanced computing and robotics research. San Jose State University wants to develop an$87 million technology center starting in the next couple of years, says associate vice president Nabil Ibrahim. Half of the money would come from a tax on current research programs and the rest from state and industrial grants. The University of California, Santa Cruz, hopes to begin development in 2004 of up to 16 hectares that will contain an undefined amount of lab and office space. “It's coming together,” says Chancellor M.R.C. Greenwood, who has a $2 million planning budget in 2000 and 2001 for facilities that will focus on biotechnology, information processing, nanotechnology, and astrobiology. The largest single investment would come from Lockheed Martin's space company based in Houston, which this month expects to sign a letter of intent with NASA. In exchange for its use of land, the company would sublease some of the 56,000-square-meter office and lab space to other science and technology firms. It also would build LASR over 18 months starting in late 2002 or early 2003 and hand it over to a nonprofit group to operate. Lockheed wants closer ties to information technology companies, and NASA documents indicate that Sun Microsystems, Oracle, Intel, Apple, and Raytheon are among those interested in partnering with Lockheed and NASA. A portion of rent increases may go into a research fund for LASR operations, says Fulop. “This has never been done, and it requires creativity,” he adds. In addition to all the new construction, the science park also hopes to rejuvenate existing structures. The core of an old Navy base, an area designated as a historic site, will be renovated by Invisible Studios, an affiliate of Dreamtime. Last year, the Mountain View-based company won a NASA contract to handle the agency's image archives. Renovations of four historic buildings are slated to begin this year. A 33,000-square-meter hangar will be converted into the California Air and Space Center to be operated by a nonprofit foundation with state and other funds, with a museum on the history of computer science as a neighbor. To meet environmental requirements, space will be set aside for a nature reserve. Despite the number of agreements already signed, however, the center still has a way to go. The development plans will be debated at public hearings this spring, and an environmental impact statement will not be ready before early 2002. The likely replacement of Dan Goldin, who has endorsed the park, as NASA administrator could also affect the pace and scope of the effort. Finally, the economic downturn may hamper university fund raisers. But participants remain enthusiastic. “There will be many twists and turns and surprises,” says Morris. “But this is a unique effort, and the basis of cooperation is there.” Ames officials are just as eager to get started. “It's a lot better than sitting here with a bunch of vacant buildings,” says Bingham. 12. DEMOGRAPHY # Sardinia's Mysterious Male Methuselahs 1. Robert Koenig More men live past 100 on this Italian island, proportionally, than anywhere else, it appears. Scientists are now trying to explain why SASSARI, ITALY—When Antonio Todde celebrated his 112th birthday on 22 January in the tiny village of Tiana in the Sardinian mountains, among the hundreds of pilgrims who journeyed to the frail shepherd's house to wish him akentannos—congratulations for living to a century—were a half-dozen Italian scientists and a Belgian demographer. The hairpin turns on the narrow road into Tiana may have given the longevity researchers gray hairs, but the harrowing trip was worth it. Nowhere else in the world, they believe, does a larger proportion of men survive to the ripe old age of 100 than on this sunny Mediterranean island. Todde—the world's oldest living man, according to The Guinness Book of World Records—and his long-lived Sardinian brethren could hold clues to the secrets of long life. In countries with reliable data, five women reach the century mark for every man who reaches this milestone. Two years ago, researchers here at the University of Sassari found that on Sardinia, the female-male centenarian ratio is only about two to one. And in Sardinia's mountainous interior, there are roughly equal numbers of 100-year-old men and women. “The proportion is remarkable,” says Claudio Franceschi, Italy's leading centenarian researcher and an immunologist at the University of Bologna—which, as the world's oldest university, is a fitting venue for longevity research. Because relatively few people have moved to Sardinia's villages, he says, the island's genetic isolation makes it “a perfect place to pursue research into complex traits such as longevity.” Whereas many past claims of extreme male longevity around the world have withered under scrutiny, the Sardinia data appear to be holding up. That's why researchers are now beating a path to Todde's door, and to many other homes in the Sardinian uplands, to take blood samples and quiz the centenarians about their genealogies and lifestyles. A clear explanation for what sets these men apart has not yet emerged, but there are hints that genetic factors—perhaps even inbreeding—may play a role. “Centenarians are the pioneers on the frontiers of survival,” says James W. Vaupel, director of the Max Planck Institute for Demographic Research in Rostock, Germany. “By studying these pioneers, we can learn a lot about why they reach such extreme age and perhaps [about] what can be done to help others reach that age.” ## A checkered past Although no one knows the exact number, as many as 100,000 centenarians worldwide may have witnessed the dawn of the new millennium. Vaupel and other demographers agree that the ranks of 100-year-olds are rising rapidly as nutrition and health care improve. Still, a minuscule fraction of the population attains supercentenarian status—someone like Todde who has lived past the age of 110. While no biblical Methuselah, said to have lived more than 900 years, the oldest person today is a 114-year-old Frenchwoman. But the numbers of supercentenarians are on the rise, according to Jean-Marie Robine of INSERM's demography institute in Montpelier, who helped validate the world's most long-lived person according to verifiable records, Jeanne Calment, who died in 1997 at the age of 122. At first, European demographers dismissed the Sardinian longevity numbers as too fantastic to be true. At a meeting in France in October 1999, Vaupel recalls, “I stood up and said: ‘There's no need to present these results, because they are almost certainly false.’” He had good reason to be suspicious. Similar claims of extreme longevity—among populations in the Caucasus, the Andes, western China, and on the island of Okinawa—all fell apart because they could not be documented. Even past U.S. censuses have exaggerated the number of American centenarians. The 1980 census, for example, listed more than twice what demographers believe to be the true number due to errors such as elderly people misreporting their ages or 4-year-olds reported as 104 because their birth years were marked wrongly. Several projects are gathering more robust data on centenarians (see table), and top centenarian researchers around the world have banded together to form the Alliance for Research on Longevity and Exceptional Survival. View this table: In light of the research area's checkered past, Vaupel cut a deal with Franceschi to split the cost of sending a crack historical demographer, Michel Poulain of the Catholic University of Louvain in Louvain-la-Neuve, Belgium, to vet the Sardinian numbers. After shuttling to the island last year to examine birth, marriage, and other records of about three dozen of Sardinia's known centenarians, Poulain announced last October that those age claims were, in fact, checking out. “This was a shock for people who were supposing that the data were false,” says Poulain, who adds that Sardinia's excellent records have allowed him so far to verify the ages of more than 50 of Sardinia's centenarians, including most of the oldest men. “This is really striking,” says Vaupel. “It's the first case I know of an alleged male-longevity region looking plausible to demographers.” Not everyone is fully convinced. “I'm still a bit skeptical, even though Poulain has done a fine job so far,” says Odense University's Bernard Jeune, who heads up Denmark's centenarian project. He, Poulain, Vaupel, and several Italian scientists are planning to join forces on a study to fully verify all the age claims and extend the demographic analysis to Sardinian men in their 80s and 90s. Whereas female mortality among Sardinian women older than 80 “is about the same as elsewhere,” Vaupel says, male mortality after age 80 “appears to be substantially lower than elsewhere.” If the numbers are correct, Jeune says, “the mortality figures for the men over age 80 are amazing.” Indeed, the exceptional survival of males over 80 accounts for the large number of male centenarians on Sardinia. ## Programmed for long life· Some experts assume that a healthy, low-stress, agrarian lifestyle is the main reason why Sardinian centenarians have outlived most of their peers. “I worry about focusing so much attention on centenarians,” says molecular geneticist Mario Pirastu of the Institute of Molecular Genetics in Alghero, who is also conducting genetic research in central Sardinia (see sidebar). “I suspect that most of the reasons for their longevity may turn out to be the style of life.” Several studies of elderly, noncentenarian twins in Scandinavia suggest that only about 25% of the variance in adult longevity is attributable to genetic differences, says Vaupel. The other 75% seems to be related to early life events and a person's current environment. But no one knows for sure whether that holds for centenarians. “It is possible that genetic factors are more important for them,” Vaupel says. And if lifestyle factors are the answer, it's not clear what they would be. After helping interview more than 100 Sardinian centenarians, Gianni Pes and Ciriaco Carru of the University of Sassari point out that some of the men—while leading otherwise low-stress, outdoor lifestyles—were once heavy smokers and weren't exactly risk adverse in their youth, having fought in one or both of the 20th century's world wars. Franceschi and his Sardinian colleagues suspect that genetics underlies the unusual male longevity, which appears to run in families. Take the case of the Brundu brothers—Pietro, 103, and Antonio, 101—who share a house in the village of Erula. Whereas Pietro's health has declined in recent years, the dapper Antonio—a confirmed bachelor—is a raconteur who regales visitors with stories about his exploits as a policeman. Asked the secret of his longevity, Antonio laughs: “I have no explanation, other than a perfectly normal life.” Todde attributes his longevity to a carefree attitude, but as Vaupel points out, “there are a lot of people with a good attitude toward life who drop dead at 50.” The Italian researchers are hoping to get more provocative answers by interrogating centenarian genes. Geneticists Giovanna De Benedictis of the University of Calabria and L. Luca Cavalli-Sforza of Stanford University are examining gene variations, or polymorphisms, of the male Y chromosome in Sardinian centenarians to see if these differ markedly from those of younger Sardinian men. In the general population, some Y chromosome polymorphisms are associated with a reduced likelihood of fathering children, says De Benedictis. Given the known trade-off between fertility and longevity, variations in Y chromosome genes, she says, could “in principle play a role in the high number of male centenarians found on the island.” De Benedictis has also teamed up with human geneticist Giuseppe Attardi of the California Institute of Technology in Pasadena—an expert on aging and mitochondrial DNA (mtDNA), which is inherited from the mother—to examine the mtDNA of centenarians. They're testing whether mtDNA gene variations protect cells against the ravages of aging. After examining mtDNA from 212 Italian centenarians and a control group, the team has found that the frequency of the J haplogroup—one of the groups of mtDNA types that population geneticists use to reconstruct human evolution lineages—”was notably higher in centenarians than in younger individuals,” De Benedictis says. In male centenarians, the frequency was about 20%, versus about 2% in younger Italian males. Researchers are now trying to explain how the J haplogroup could be linked to long life. One surprising indication from the research so far is that inbreeding, long known to increase the chances of inheriting recessive genes that can be detrimental to health, may actually help the Sardinian men live longer. In Sardinia's remote mountain villages, most residents are descendants of a few founding families, and preliminary data suggest that they do, indeed, have less genetic diversity than the general population. “At first glance, the finding seems to be counterintuitive,” says Franceschi. One possible explanation could lie in the fact that extreme longevity itself is not a logical consequence of evolution: If life's raison d'être is to reproduce, excessively old individuals—well past their prime reproductive years—offer little if any biological benefit to a population. “Longevity is a trait with some peculiarities because of the unnecessary nature of aging,” says Franceschi. Another interesting concept from the Italian research is that, in certain individuals, the immune system may adapt to aging. According to Franceschi, this “complex reshaping of the immune system” in extremely old persons tends to follow this pattern: Whereas the most sophisticated immune response, centered on T and B lymphocytes, tends to deteriorate with age, the body's innate immunity—in which macrophages gobble up foreign proteins and cells—appears to improve, like a fine wine. “This phenomenon starts at about 60 to 70 years of age,” says Franceschi, who believes that “those who live longer … are able to adapt continuously to the deteriorative changes occurring in the immune system with age.” Other immunologists, including Beatrix Grubeck-Loebenstein of the Institute for Biomedical Aging Research in Innsbruck, Austria, have noticed similar immune-system “shifts” in certain healthy patients over age 65. When such shifts occur, says Georg Wick, an Innsbruck colleague who edits the journal Experimental Gerontology, the body's innate immunity “is not only preserved, but even compensates for the deteriorating response of the specific immunity.” Years of work lie ahead to decipher any lessons encrypted in the centenarians' unusual genetics. “When I first started this work a dozen years ago, centenarians were considered a rare curiosity,” says Franceschi. “Now they are important subjects of research.” He and his colleagues know the clock is ticking. Besides drawing blood from the oldsters, a team led by Giovannella Baggio and Luca Deiana has been banking DNA samples and taking detailed medical, genealogical, and mental-health histories. Is the Fountain of Youth hidden in the verdant Sardinian mountains? Probably not, unless that vital liquid is the local red wine, which all the old men seem to enjoy. Asked whether he had learned any secrets of longevity from his 103-year-old father, 75-year-old Italo Brundu smiled and said: “I touch him every day, and I hope it rubs off.” 13. DEMOGRAPHY # An Island of 'Genetic Parks' 1. Robert Koenig NUORO, ITALY—Scientists are flocking to isolated communities such as those found in the rugged interior of the island of Sardinia, hoping to mine their genetic riches. The first major genetic campaign has been mounted in Talana, a remote village near the Gennargentu Mountains southeast of Nuoro. While longevity researchers have zeroed in on a few dozen unusual individuals scattered across the highlands (see main text), one molecular geneticist has turned the entire village of Talana into a “genetic park.” NUORO, ITALY—When the Romans invaded the island of Sardinia in the second century B.C., they called its rugged interior, where the native nuraghic tribes had taken refuge, the land of barbarians. Today's descendants of those warriors are welcoming, not fighting, the latest wave of invaders: scientists mining the genetic riches of these isolated communities. The first major genetic campaign has been mounted in Talana, a remote village near the Gennargentu Mountains southeast of Nuoro. While longevity researchers have zeroed in on a few dozen unusual individuals scattered across the highlands (see main text), molecular geneticist Mario Pirastu has turned the entire village of Talana into a “genetic park.” The goal is to probe for genes that might contribute to illnesses common in Sardinia, including diabetes, kidney stones, and asthma. Focusing on such Sardinian towns makes sense, says David Schlessinger of the U.S. National Institute on Aging's gerontology research center in Baltimore, as they often were settled by a few individuals. The small genetic variability of so-called founder populations, says Schlessinger, who has teamed up with a rival group of Sardinian scientists to look at other mountain villages, “makes the work of geneticists simpler.” Sardinia's genetically isolated populations, adds Pirastu, offer a more precise target for gene hunters than larger surveys such as one now under way in Iceland by deCODE Genetics Inc. (Science, 1 January 1999, p. 13). “In Iceland, they are examining the whole island,” says Pirastu. “But here in Sardinia, we think that we can get the same sorts of results using maybe 5000 people in a few villages. You don't have to start with a half-million people.” With the support of Talana's 1200 residents, Pirastu—who heads the Italian national research agency's Institute of Molecular Genetics in Alghero—has spent 6 years building a research clinic, tracing the entire town's family tree back 400 years, canvassing health information, and taking DNA samples from every adult inhabitant. Now collaborators are helping to map the frequency of genetic polymorphisms, or variations, that Pirastu hopes to link to diseases. “His project looks very promising,” says molecular geneticist James L. Weber, whose group at the Marshfield Medical Research Foundation in Wisconsin is one of two contracted to do the gene mapping. “We'll learn a lot about the population genetics of the village and maybe about Sardinia in general,” he says. Weber's group and a team in Edinburgh together will sequence 1600 markers for each adult Talana inhabitant. Some Italian critics contend that the Sardinians may not fully understand the project's implications and therefore might be exploited. Pirastu, however, says that villagers sign privacy waivers, and that Talana and other towns enrolled in the study get new clinics, free medical testing, and jobs connected with the project. The villagers, he notes, are aware of the study's goals: Pirastu says he gets written permission from each resident “every time we collect a new blood sample or clinical test.” The villagers, he claims, “are happy to collaborate, because they feel they are part of something important.” Pirastu's research has attracted the attention of Sardinia's wealthiest man—Internet entrepreneur Renato Soru—who last year put up most of the$5 million in capital for a start-up company, called Shar.DNA, that is now discussing how to commercialize any findings from the mountain villages. Italy's national research agency—which could use its share of any future profits to beef up the research—is a partner in the firm, as are Soru and Sardinia's biggest bank. Pirastu is Shar.DNA's scientific director, but he says he holds no shares in the company.

Pirastu already has expanded the “genetic park” experiment well beyond Talana. The town of Perdasdefogu has given him a \$1 million grant and lab space to start plumbing the genetics of its 2400 inhabitants, for example, and he has extended his studies to several other isolated villages with populations of between 1000 and 2000. His group will soon publish a paper linking a region of chromosome 10 with kidney stones. “Have we found a gene? Not yet,” says Pirastu. “But we think that we are very close. Our fingers are crossed.”

14. NEUROBIOLOGY

# Does Alcohol Damage Female Brains More?

1. Bernice Wuethrich*
1. Bernice Wuethrich is a writer in Washington, D.C.

As they begin to probe this question, researchers are finding surprising and sometimes contradictory evidence that gender matters

In the blockbuster film Traffic, newly appointed drug czar Judge Robert Wakefield and his wife Barbara argue about who drinks more alcohol. The judge swills a nightly Scotch and soda to take the edge off his boredom at home, while his wife quaffs at least three times as much. But their nasty spat over who drinks more begs a nagging scientific question: Does alcohol affect the brains of men and women differently, perhaps causing even more damage to the female than the male brain?

The answer has been surprisingly hard to pin down, despite anecdotal observations that women alcoholics suffer more severe motor problems and cognitive impairment than men do. Alcoholism has been traditionally thought of as a “male disease” because of its higher prevalence among men: There are about three times as many male as female alcoholics. In addition, researchers have often turned to the mostly male Veterans Administration hospitals for their research subjects; only recently have studies been published that compare men and women. Methodological problems are rife, too, especially controlling for alcohol consumption.

For these reasons, those sex differences that studies have turned up, such as increased damage to women's livers and hearts, have been largely attributed to the different ways in which men and women metabolize alcohol. Because women tend to be smaller and have more body fat than men—and because women may have less of a stomach enzyme that digests alcohol—women's blood alcohol levels (BALs) tend to be higher after imbibing the same amount as a man. Higher exposure simply translates into more severe effects, the argument goes.

Over the past few years, however, researchers have found tantalizing clues suggesting that more is at work than higher BALs. Using increasingly precise molecular and brain imaging techniques, researchers have been scrutinizing brains of both men and women, as well as male and female rats. From these studies are emerging new—although sometimes contradictory—evidence that when it comes to alcohol, one's sex matters. The findings reinforce the idea that, in addition to higher BALs, biological differences between male and female brains contribute to increased damage in women.

“It's still controversial, but more of us are starting to recognize the likelihood that the female brain is indeed more sensitive to the deleterious effects of alcohol,” says neuroscientist Mark Prendergast of the University of Kentucky, Lexington, who is examining the issue.

“Sex differences in alcohol-induced brain damage is certainly an area that has to be explored,” agrees David Lovinger, a neurophysiologist specializing in brain damage caused by alcohol at Vanderbilt University in Nashville, Tennessee.

## The shrinking brain

Magnetic resonance imaging (MRI) studies conducted in male alcoholics have consistently shown that excessive alcohol consumption shrinks male brains, particularly the white matter, and that there is a corresponding increase in the volume of cerebrospinal fluid (CSF). In spring 1999, psychiatrist Daniel Hommer of the National Institute on Alcohol Abuse and Alcoholism in Bethesda, Maryland, got a surprise when he began looking for similar effects in females' brains. Specifically, Hommer compared the entire brains of 43 alcoholic men and 36 alcoholic women to those of nonalcoholic men and women of the same ages. The contrast was startling. “After reviewing the first dozen people, we thought, ‘Wow, there's something here,’” Hommer recalls. Alcoholic women lost about 11.1% of their gray matter compared to healthy women, whereas alcoholic men lost only about 5.6% of their gray matter. Alcoholic women lost 8.2% of their white matter verses 5.3% for men. At the same time, the volume of space filled with CSF increased by 24.1% in alcoholic women over healthy women, and by 10.5% in alcoholic men relative to their controls. This latter difference was most striking in the intrahemispheric fissure, a gully that runs through the top center of the brain and separates its two hemispheres. Like a stream that erodes the landscape, the fissure widened and deepened in alcoholic women. In women as in men, brain shrinkage was evident by the time an alcoholic had reached her early 30s.

Hommer's MRI results, published in the February issue of the American Journal of Psychiatry, are the first to show that alcoholic women have reduced gray and white matter volumes and correspondingly greater CSF volumes than healthy women—and that their damage is greater than that of men. Already, the data are being challenged.

Indeed, a study published in the same journal found decided sex differences—this time, suggesting men are the hardest hit. In their MRI study, neuroscientists Edith Sullivan of Stanford University School of Medicine and Adolf Pfefferbaum of SRI International, a research and consulting firm in Menlo Park, California, compared 44 alcoholic men and 42 alcoholic women with healthy controls. The California team imaged a slab of brain, concentrating on the cortex, the brain's outermost layer and home to most of its gray matter. The images included the white matter just below and the lateral ventricles, the largest fluid-filled space in the middle of the brain.

They found a striking decrease in brain volume in men, especially in the frontal cortex—but not in women. The size of the ventricles, however, increased in both men and women to about the same extent, Sullivan says. This provided the only hint of brain shrinkage in women. The researchers also found that the women who had been abstinent for longer periods had larger amounts of white matter than women who were more recently sober, implying as other studies have that brain structure and volume can recover, at least partially.

“The new, contradictory results underscore the complexity of studying this problem,” explains clinical psychologist Terry Jernigan of the University of California, San Diego. The discrepancies could stem from differences in either methodology or in patient groups, she says. In terms of methodology, the studies used subtly different MRI resolutions, leaving both research groups prone to different kinds of error. In addition, Hommer's alcoholic subjects were all recruited from an inpatient program, whereas the men in Sullivan's study were inpatient military veterans while the women were outpatients, and inpatients tend to have more severe disease than outpatients. In addition, the participants in Hommer's study had been sober for just 3 weeks, while Sullivan's participants had been sober for an average of several months, allowing more time for brain recovery. Finally, both studies could have been more rigorous in controlling for alcohol consumption, Jernigan says.

The studies raise more questions than they answer, including possible mechanisms of damage in the female brain. Brain imaging studies cannot determine specific causes, only effects. Nor can human studies probe actual molecular changes in the brain. For that, researchers turn to rats.

## Molecular suspects

At the University of Kentucky, Prendergast and colleague John Littleton recently identified a molecular suspect in sex-specific brain damage: spermidine. Unrelated to its namesake, spermidine is a polyamine, a natural substance that plays a role in cell growth, differentiation, and death. Produced in cells throughout the body, polyamines can modulate the functions of RNA, DNA, and protein synthesis.

Studies by Peter Wilce of the University of Queensland in Brisbane, Australia, in 1998 had shown that spermidine is released in rat brain tissue during alcohol withdrawal. Wilce and colleagues also reported in the journal Alcoholism: Clinical and Experimental Research that spermidine's presence is accompanied by an increase in the seizure activity of neurons, in which they fire nonstop. Such seizures can kill nerve cells.

Subsequently, Prendergast and Littleton looked for possible sex differences in this effect, comparing slices of the hippocampus from male and female rats. They first dosed the brain slices with alcohol and then compared the amount of neuronal death during withdrawal, the brain's most sensitive time. They saw no difference between males and females—until they added spermidine to the slices. As reported in the December 2000 issue of the same journal, the presence of spermidine during alcohol withdrawal caused 15% to 20% more neuron death in female than male rats. Prendergast and Littleton speculate that, in living brains, nerve cells damaged by alcohol release spermidine as part of an intended repair process. But in females, the process goes haywire, and spermidine increases the seizure activity of neurons already agitated by alcohol.

Previous research established that spermidine alters the neurons' NMDA receptors in a way that enhances the activity of glutamate, an excitatory neurotransmitter. Upon binding to the NMDA receptor, glutamate normally opens a pore that allows a measured flow of calcium to enter the cell, where it contributes to an array of activities. But transformed by alcohol and, subsequently, by spermidine, the tap stays open, allowing more and more calcium to flood the neuron. This riptide of ions activates enzymes that destroy the cell. This basic process takes place during withdrawal, causing cell death; now, with Prendergast and Littleton's findings, there is evidence that spermidine exacerbates it in females.

Fulton Crews, a pharmacologist specializing in alcohol studies at the University of North Carolina, Chapel Hill, notes that the finding challenges another presumption about alcohol and the brain: that any differential damage would be due to hormonal differences in the male and female brains. “By putting the hippocampus in culture and taking it out of the brain's hormonal milieu, it implies that the difference is inherent to the system,” he says.

## Hormones with a new twist?

Throughout her lifetime, a woman's brain is exposed to a more variable hormonal milieu than a man's brain is. But researchers have had difficulty establishing a definitive link between those differences and the brain's response to alcohol. Now, one hint of such differences comes in the work of endocrinologist Catherine Rivier of the Salk Institute for Biological Studies in San Diego.

In a series of rat studies over the past 10 years, Rivier has found that alcohol overstimulates the hormonal cascade produced by the hypothalamic-pituitary-adrenal (h-p-a) axis far more in females than in males. One of the last hormones in the cascade is cortisol, and several studies have shown that the chronic release of cortisol can produce mild brain damage.

Although Rivier's studies strongly suggest that hormonal differences influence alcohol's effects on the brain, she cautions that they don't explain it entirely. When she removes female circulating sex steroids by removing the ovaries, alcohol still overstimulates the h-p-a axis, although to a much lesser degree.

Looking beyond the usual hormonal suspects, Leslie Devaud, a neuropharmacologist at Idaho State University in Pocatello, is exploring sexual dimorphism in brain circuitry and pathways. “Gender differences in the response to alcohol may arise from subtle differences in molecular responses,” she says.

Devaud has found initial indications of such molecular differences by studying minute changes in both NMDA and GABAA receptors in male and female rats. (γ-aminobutyric acid or GABA is the main chemical messenger that inhibits the firing of neurons, and the GABAA receptor plays an important role in alcohol dependence and tolerance.) Devaud found that chronic alcohol exposure predominantly wrought changes in brain GABAA receptors in male rats, whereas in female rats, the most pronounced changes were in the NMDA glutamate receptors. But the impact of these differences remains a mystery.

Regardless of sex, the brain responds to alcohol by trying to counteract its depressing effects on the central nervous system. Devaud and others are continuing to search for molecular responses that vary between males and females as the brain attempts to compensate for the changes wrought by alcohol use. Says Devaud: “Big questions remain, but if we can identify relevant gender differences, it could ultimately affect how we treat alcohol dependence in men and women.”