# News this Week

Science  30 May 2008:
Vol. 320, Issue 5880, pp. 1142
1. PHYSICS

# Energy Department Pulls Plug on Overbudget Fusion Experiment

Fusion may someday yield cheap power, but a troubled experimental reactor has proved too pricey for the U.S. Department of Energy (DOE). This week, DOE terminated the National Compact Stellarator Experiment (NCSX) at the Princeton Plasma Physics Laboratory (PPPL) in New Jersey. The not-yet-completed reactor would have been one of four large “magnetic confinement” reactors in the United States.

Donald Rej, a plasma physicist at PPPL and NCSX project manager since February, says he was disappointed to hear the news. “My colleagues have put a good fraction of their careers into this,” Rej says. “It's a technological tour de force.”

Like a supernova, the stellarator's budget had exploded. In 2001, it was proposed as a $58 million project to be completed in 5 years. It won approval in 2005 with a “baseline” budget of$102 million and a completion date of 2009. But in April, a DOE review showed that the cost had ballooned to $170 million and the machine could not be completed until 2013 at the earliest. The review suggested that even those estimates might not hold. “We were unable at this point to rebaseline, to formally say that we knew what it would take to finish,” says Raymond Fonck, associate director for fusion energy sciences in DOE's Office of Science. Had it been completed, the NCSX might have served as the prototype for the next great fusion experiment to come after ITER, the$12 billion machine that will be built in Cadarache, France. In a magnetic confinement reactor, scientists heat an ionized gas, or plasma, of light nuclei to 100 million degrees while trapping and squeezing it with magnetic fields. ITER and the three remaining machines in the United States are tokamaks, reactors built around a tubelike magnetic coil curled into a doughnut shape, or torus. The coil produces magnetic fields that go straight around in the horizontal torus. But to confine the plasma, the field must be modified so that it spirals around the doughnut like the stripes on a candy cane. To make that happen, the plasma must flow around the tokamak to produce an electrical current. The flow is generated by applying pulses of magnetic field or by other, more complicated means.

A stellarator, by contrast, uses bizarrely shaped coils to generate a spiraling field from scratch. Those coils are immensely complicated. NCSX researchers were able to fabricate them, but they needed more time and money to assemble the parts into a whole machine. The problems were far trickier than sticking tab A into slot B and required overcoming engineering challenges such as how to weld the massive asymmetrical pieces without deforming them, Rej says.

The cancellation of the NCSX strikes a body blow to the United States's domestic fusion program, says Stewart Prager, a plasma physicist at the University of Wisconsin, Madison, and chair of DOE's Fusion Energy Sciences Advisory Committee. “The loss of important scientific knowledge is very large,” Prager says. “NCSX would have tested a fascinating physics concept and advanced understanding of a very promising fusion configuration.”

The cut intensifies the uncertainty already facing plasma physicists. Over the past decade, DOE's budget for fusion research has stagnated at $300 million, and since the United States rejoined the ITER collaboration in 2003, researchers have fretted that money for smaller experiments at home might be siphoned off to pay for the nation's commitment overseas. (This year, however, the U.S. Congress zeroed out a scheduled$149 million contribution to ITER and bumped up the budget for running domestic facilities to $93.5 million,$6 million more than DOE had requested.)

Fonck says he would prefer to keep the $19.6 million requested for NCSX next year within the domestic fusion program. That money could be used simply to run the other U.S. fusion experiments longer. For example, PPPL's tokamak, the National Spherical Torus Experiment (NSTX), ran for only 13 of a possible 25 weeks in 2007 and will run for 15 weeks this year. “The plan for next year is down to 9 weeks,” says A. J. Stewart Smith, dean of research at Princeton University, which runs PPPL. “This was necessary to accommodate NCSX.” It's too early to know whether the cancellation of NCSX will lead to layoffs at PPPL, Fonck says. The lab has a staff of 420 and a budget of$77 million this year.

Some physicists argue that, in spite of the cost overruns, DOE should have stuck with the NCSX. “Given the energy problem we have, it doesn't make any sense to shut down projects like this,” says Miklos Porkolab of the Massachusetts Institute of Technology in Cambridge. He notes that China and South Korea, both members of ITER, completed their own large fusion experiments in 2006 and 2007, respectively. The United States hasn't completed a new machine since PPPL finished the NSTX in 1999.

2. PARKINSON'S DISEASE

# Streamlined Clinical Trials, From a Home Computer

1. Jennifer Couzin

A Parkinson's research and treatment center and the genetic testing company 23andMe, both in California's Silicon Valley, are experimenting with an unusual new approach to clinical trials: have participants assess themselves from their home computers potentially using everything from videos of tremors to a mouse that senses motor abilities. If it works—still a big if—the strategy could greatly reduce the need for doctors' visits and make trial participation vastly cheaper and possible from anywhere in the world.

Researchers have long considered how the Web might enhance clinical trials, in particular its ability to aggregate data from the thousands of people needed for studies tracking genetic and environmental factors linked to common diseases. But the new alliance comes with many uncertainties. Will there be a bias in who completes the assessment? Can the approach capture subtle changes in disease symptoms? Will participants understand what they're signing up for, and how will the data they supply be used?

“Can you really do valid clinical research via the Web?” asks Katie Hood, chief executive officer of the Michael J. Fox Foundation for Parkinson's Research in New York City, which is funding the project with a $625,000, 2-year grant. “That's what this will show.” The Parkinson's Institute and Clinical Center in Sunnyvale, California, will recruit 150 people, half with Parkinson's disease and half without, who recently participated in a traditional study examining occupational risk factors behind Parkinson's; through that study, researchers will have amassed a wealth of data on them. 23andMe will help design or adapt computer technologies and questionnaires for the 150 participants to see whether they accurately capture their health and health history. If they do, the tools could be expanded to much larger populations and to other research groups. Volunteers will also submit saliva samples to 23andMe, and those samples, like all supplied to the company, will be sequenced for more than 580,000 single-base variations. Eventually, their genetic information will be matched with information they provide online and, say the study leaders, stored securely. “We want to build this engine basically to be able to power genomewide association studies,” says Linda Avey, a co-founder of 23andMe. Such enormous studies compare the genomes of those with a particular disease to the genomes of those without. To accomplish this, 23andMe needs to collect large online cohorts of people with different ailments, including Parkinson's disease. Eventually, 23andMe might sell information on participants, with their consent, to pharmaceutical companies looking to recruit for clinical trials. Avey notes that the company is considering similar strategies in other diseases, although she declined to say which ones. “We're all trying to come up with new clinical approaches that can help us get large cohorts,” says Kenneth Marek, a neurologist at the Institute for Neurodegenerative Disorders in New Haven, Connecticut. Marek is experimenting with another in-home assessment for Parkinson's disease, independent of the 23andMe venture: a sniff test that includes 40 different odorants. One of the first symptoms of the disease is a diminished sense of smell. Marek and his colleagues have begun recruiting 10,000 people whose family members have Parkinson's to see whether the disease can be detected in its earliest stages. One concern, says Marek, is that relying on volunteers to run through such tests out of sight of clinicians may lead to bias. People who struggle with the tasks could “get frustrated and may be less willing to complete this at home than they would be in a clinic,” he says. “We have some sense that that might be happening,” but more information is needed. A question for 23andMe is whether participants in a disease-focused study will be ready for all the other genetic information the company offers when it surveys their genomes. Will someone in a Parkinson's trial want to know whether they—and by extension, their family members—are at increased risk for heart attacks or prostate cancer? The project leader, Parkinson's Institute scientific director and CEO J. William Langston, expects that the institutional review board considering the ethics of the venture, which could start later this year, will scrutinize it closely. As that process gets under way, Langston and 23andMe are bringing together technology companies to explore how far computers can be pushed to assess Parkinson's disease from a person's home. The Michael J. Fox Foundation, meanwhile, is gearing up to spend as much as$1 million more on other Web-based assessment tools: It put out a call for applications in March.

3. WENCHUAN EARTHQUAKE

# Damaged University Mourns Its Dead--and Plans Fast Recovery

1. Richard Stone*
1. With reporting by Chen Xi.

MIANYANG, CHINA—At 2:28 p.m. on 12 May, Gao Kun was relaxing in his fifth-floor dorm room at Southwest University of Science and Technology (SWUST) when the building began to shake violently. The 23-year-old computer science major scrambled down the stairwell. Outside, people were screaming. As the ground convulsed, some students jumped from dorm windows. Gao, ducking chunks of falling masonry, helped a young woman who had jumped and broken her leg to limp away from the building. He ran back and dragged a more severely hurt jumper to safety. Minutes later, several professors arrived to aid the injured and soothe the terrified students.

Mianyang, a sprawling science city with a population of 5.2 million, was one of the worst hit by the rupture of the Longmenshan Fault, 50 kilometers to the west. As Science went to press, the number of local dead had climbed above 20,000, and the overall official toll for the Wenchuan earthquake stood at 67,183. In addition to SWUST, Mianyang has several institutes, including the China Air Dynamics Research and Development Center, and a high-technology R&D park anchored by the Sichuan Changhong Electronics Group. Mianyang may be best known, however, for China's main nuclear weapons design laboratory, the Chinese Academy of Engineering Physics. Although damage assessments at “China's Los Alamos” and nuclear weapons-fabrication facilities northwest of Mianyang are not publicly available, there have been no reports of radiation leaks. In the meantime, a lake formed by a landslide in the mountains west of the city is rising (Science, 23 May, p. 996), prompting the evacuation of more than 70,000 Mianyang residents as Science went to press.

5. GENETICS

# Ancient DNA From Frozen Hair May Untangle Eskimo Roots

1. Michael Balter

Humans began to brave the frozen northern reaches of Alaska, Canada, and Greenland about 4500 years ago, according to archaeological evidence. Researchers have long pondered the family history of these so-called Paleo-Eskimos, who were skillful hunters. Did they descend from the same Asian peoples who 10 millennia earlier crossed the Bering Strait and headed south, giving rise to Native Americans? And were they the ancestors of modern Eskimos? Now DNA recovered from an ancient clump of hair suggests that the answer to both questions is no.

This week, researchers from Europe and Greenland report online in Science (www.sciencemag.org/cgi/content/abstract/1159750) the sequencing of mitochondrial DNA from a male Paleo-Eskimo who lived in western Greenland roughly 4000 years ago—the first near-complete ancient mtDNA genome ever published. The sequence is distinct from that of both Native Americans and modern Eskimos but closely resembles that of small populations living today in the Bering Sea area, implying that the earliest Eskimos derived from an independent wave of migration from this region that left no living descendants. “The methodology appears to be excellent and their conclusions are believable,” says Michael Crawford, a biological anthropologist at the University of Kansas, Lawrence.

For decades, archaeologists have attempted to trace the peopling of the far north. The earliest Paleo-Eskimos show up all across the Arctic region about 4500 years ago. But by about 1000 years ago, the Paleo-Eskimos were replaced by new migrants called the Neo-Eskimos, which researchers have concluded are the ancestors of modern Eskimo groups such as the Inuit.

Eske Willerslev, who specializes in ancient DNA at the University of Copenhagen in Denmark, learned recently that Bjarne Grønnow, an archaeologist at the National Museum of Denmark, had found human hair during his excavations 20 years ago at a Paleo-Eskimo site on the west coast of Greenland. The hair was “a huge clump,” says Willerslev, and its DNA was well-preserved by the permafrost.

The mtDNA genome derived from the hair bore a relatively rare genetic marker called D2a1, which is absent in modern Native Americans. To check possible links between the Paleo-Eskimo sample and Neo-Eskimos, Willerslev's team also sequenced the mtDNA genomes from 14 Greenlandic Inuits; none had the D2a1 marker. Nor does the marker show up in the handful of partial ancient DNA sequences from Neo-Eskimo skeletons that have been excavated and analyzed by others.

The genetic marker is also absent from Europeans, one reason Willerslev is confident that the mtDNA his team analyzed is not from the Danish-Inuit team that found the hair. Beth Shapiro, an expert on ancient DNA at Pennsylvania State University in State College, agrees that the mtDNA sequence is “not likely” to be the result of the contamination problems that often plague ancient DNA studies.

The D2a1 marker found in the Paleo-Eskimo mtDNA is closely related to a marker called D2a1a, which is found in present-day inhabitants of the Bering Sea area such as the Aleuts and a group of Siberian Eskimos. Scientists agree that this suggests that Paleo-Eskimos originated from this area. Still, notes Agnar Helgason, a biological anthropologist at deCODE Genetics in Reykjavík, Iceland, “this interpretation rests on a single sequence.” Grønnow and his colleagues found two smaller clumps of hair during the 1980s excavations, and Willerslev's team plans to repeat their ancient DNA analysis. “I think there is a good chance of getting one more individual,” Willerslev says.

6. EPIDEMIOLOGY

# Children's Study Needs Pilot Testing, Panel Finds

1. Jocelyn Kaiser

An expert panel last week urged that several changes be made in a controversial $2.7 billion health study of 100,000 U.S. children before the first pregnant mothers are enrolled in September. The National Institute of Child Health and Human Development says it is responding to some of the 24 suggestions, but others would be too expensive. Mandated by the U.S. Congress in 2000, the National Children's Study (NCS) will follow children from before birth to age 21 and explore how environmental factors, including pollution and television, influence disease and normal development. The study “has had a long and difficult gestation,” notes the report from the National Research Council and the Institute of Medicine. Congress appropriated$69 million for this project in 2007 and $111 million this year, although the White House and National Institutes of Health Director Elias Zerhouni opposed it (Science, 9 February 2007, p. 751). Last year, the child health institute produced a 700-page research plan describing the study's 28 broad hypotheses. It then requested a review. The 12-member review panel, chaired by Samuel Preston, a sociologist at the University of Pennsylvania, praises NCS's large size and its plan to visit randomly chosen households at 105 locations. The easy way would have been to find participants through health care providers, but a random sample means the results will be nationally representative and will provide “valuable” data. But the panel also points to “important weaknesses and shortcomings.” For one, it faults the absence of a built-in pilot stage for recruiting patients, collecting data, and managing the “huge databases.” The panel suggests a 6- to 12-month delay between enrollment at the vanguard centers later this year and at the next wave of centers. Extra time is needed because “it is such a complicated study with so many variables,” says Preston. View this table: The reviewers also saw a problem in follow-up: After the babies turn 1, they will be evaluated in person once every few years. In the meantime, researchers will rely on telephone interviews with mothers to find out about sicknesses and diagnoses. The report urges that NCS conduct more frequent home or clinic visits and gather medical records, school records, and other documents that could help confirm illnesses and problems such as child abuse. NCS study director Peter Scheidt says that “we agree absolutely” on the need for a pilot phase and proposed one a few years ago, but it wasn't funded. The vanguard centers will test the protocol, he says, but he admits that NCS will have less than ideal flexibility. Still, NCS is delaying full enrollment by 6 months to January 2010 to allow for revisions. Scheidt says NCS will follow several more recommendations, such as reconvening a working group on health disparities. But adding more in-person visits or collecting paper medical records, Scheidt says, would simply be too expensive. Pediatrician Philip Landrigan of the Mount Sinai School of Medicine in New York City, an early proponent of NCS and a leader of a vanguard center, says his team would have to hire two full-time nurses to visit 100 medical practices. He thinks interviews will suffice: “If it's a significant illness, like diabetes or a surgical procedure, parents remember,” he says. Preston notes that the panel was not asked to weigh in on whether the study should go forward or is worth the cost. “That's beyond our level of expertise,” he says, “but it's not an unimportant issue.” The study's annual cost rises to$192 million next year.

7. PARTICLE PHYSICS

# Does Fermilab Have a Future?

The United States's last particle physics lab finds itself in turmoil, with its current experiments soon to wind down and nothing under construction to replace them. Physicists wonder whether the lab--and particle physics in the United States--will survive.

The United States's last particle physics lab finds itself in turmoil, with its current experiments soon to wind down and nothing under construction to replace them. Physicists wonder whether the lab—and particle physics in the United States—will survive

BATAVIA, ILLINOIS—Like a magnet, particle physics drew David Mason when he was an undergraduate. “I was initially attracted by all the cool toys we play with,” says the postdoc here at Fermi National Accelerator Laboratory (Fermilab). “Basically, everything we use we have to construct for ourselves because it's never been thought of before.” Mason, 37, first worked in a lab as an undergrad at the University of Oregon, Eugene. In 1996, he came to Fermilab, whose bucolic 2750-hectare campus preserves a patch of quiet in the suburban sprawl 60 kilometers west of Chicago, as an Oregon graduate student to study particles called neutrinos. After finishing his doctorate 2 years ago, he signed on to collaborate on an experiment that will be done in Europe.

Now, Mason finds himself spending his savings to keep his young family afloat. Rocked by budget cuts late last year, Fermilab will soon lay off about 140 of 1950 staff members (Science, 16 May, p. 858). In February, the lab instituted a rolling furlough that, until year's end, requires employees like Mason to take 1 week every 2 months as unpaid leave. The 25% cut in every other paycheck hurts, says Mason, whose wife stays home with his 2-year-old son.

Perhaps more troubling, the budget crunch leaves the future of the 40-year-old lab, the United States's last dedicated particle physics lab, uncertain at best. Fermilab's current experiments will wind down early next decade, and the U.S. Congress cut funding for the projects meant to replace them (Science, 11 January, p. 142). The action is shifting to Europe and Japan, and Mason, who says moving abroad is probably out of the question, wonders how long he can stick with the field. “On one hand, this is what I've spent years of my life preparing for,” he says. “On the other hand, my family has to eat.”

American particle physics stands at a crossroads. Since the invention of the cyclotron in 1929, the United States has led the quest to bust matter into bits and see what the universe is made of. The science is more exciting than it has been in decades, researchers say. Fermilab's particle smasher, the 6.3-kilometer-long circular Tevatron collider, is cranking out copious data that could reveal the long-sought Higgs boson, the missing link in the “standard model” of the known particles. This summer, the European particle physics lab, CERN, near Geneva, Switzerland, will turn on its Large Hadron Collider (LHC), a 27-kilometer ring that could blast out scads of new particles and recreate conditions of the big bang. The United States has 1300 researchers working on the LHC, more than any other country.

But the United States's position in particle physics has been slipping, and this year the decline has snowballed into a crisis. In the past 3 months, U.S. researchers have shuttered colliders at Cornell University and the Stanford Linear Accelerator Center in Menlo Park, California. Only the 25-year-old Tevatron remains, and it will shut off in 2010. Fermilab's smaller experiments will end at about the same time. In this country, the cupboard is bare, and physicists have only unapproved plans with which to restock it.

The immediate cause of the turmoil at Fermilab is the last-minute budget-Congress passed in December. It trimmed Fermilab's budget to $320 million this year from$342 million in 2007, $52 million less than requested by the U.S. Department of Energy (DOE), which owns the lab. Congress zeroed out$36 million for a proposed neutrino experiment called NOνA; slashed $14 million for work on the proposed multibillion-dollar International Linear Collider (ILC), which American physicists hope someday to build at the lab; and clipped$18 million for research on superconducting accelerator technology. “You took all the things that the lab was working toward for a future facility and you lopped them off,” says Fermilab Director Piermaria Oddone. “When you do that, you're pointing the laboratory straight for the rocks.”

The roots of the problem reach further back. Knowing that the LHC would eclipse the Tevatron, many U.S. physicists and some DOE officials have pushed to start building the ILC at the lab as early as 2016. In their haste, they scrapped smaller projects that otherwise might have protected Fermilab from cost cutters looking for vulnerable research and development expenditures. Or so others say. “It's pretty clear that there was a plan at DOE to clear the decks for the ILC, and a number of us saw that this was incredibly risky,” says Sheldon Stone of Syracuse University in New York, who calls the current jam “predictable.”

Fermilab is not giving up. To secure their future through the next decade, researchers have proposed a relatively modest billion-dollar proton accelerator, dubbed Project X, to feed neutrino studies and other smaller scale experiments. The stakes are high: If all the accelerators are overseas, U.S. particle physics may simply die, researchers say. “Five years down the road, Congress may look at it and say, ‘If there's nothing here, why are we funding this at all?’ “says Robert Harr of Wayne State University in Detroit, Michigan. The odds may be long: DOE has completed just one major project at Fermilab in 9 years.

## The ILC: A gamble that didn't pay

On 19 February 2006, William Foster penned an open letter to his colleagues. Foster had worked at Fermilab for 22 years, the last five as head of a project that, he thought, would provide the lab with a decade of research to do. Dubbed the Proton Driver, the billion-dollar linear accelerator would have pumped out protons that would crash into targets to generate neutrinos and particles called muons, K mesons, and D mesons for experiments.

But weeks earlier, officials from DOE's Office of Science decided that they would not put the project up for the first of five “critical decision” reviews that any DOE project must pass as it wends its way from idea to facility. The reason for DOE's refusal, Foster wrote, was that the Proton Driver, the concept for which had been proposed by others as early as 1994, would interfere with efforts to get the ILC built as quickly as possible. “This position apparently applies not only to the Proton Driver, but to any intermediate-scale projects which might provide any alternate or interim future for U.S. [high-energy physics] at a cost significantly less than the [approximately] $10 billion estimated cost of the ILC,” Foster wrote. “I fear that this approach is likely to end very badly. …” Foster quit the project, the lab, and the field. “The Proton Driver represented my last best effort for Fermilab,” Foster says. “And when it was clear that it wouldn't go through, I wanted to try something else where I could be more successful.” That something else was politics: In March, Foster won a seat in the U.S. House of Representatives (Science, 14 March, p. 1470). Numerous factors put Fermilab in its riches-to-rags predicament. For more than a decade, the DOE's high-energy physics budget failed to keep pace with inflation; this year it fell 8.5% to$688 million, from $752 million in 2007. In 2001, after a major upgrade, the Tevatron performed so poorly that the lab had to throw all it had at the problem (Science, 8 February 2002, p. 942). Plainly put, in the past 2 decades, particle physics has produced few of the major discoveries that thrill the public and secure generous funding. However, many Fermilab researchers argue that the lab is in trouble because, following the lead of advisory panels stacked with ILC supporters, DOE officials have sacrificed modest experiments that the department can afford for a chance at a dream machine that may never come. “We ended up in an ILC-or-bust mode,” says Stephen Holmes, Fermilab's associate director for accelerators. View this table: In the past decade, physicists at Fermilab have proposed several hundred-million-dollar experiments that would have searched for new physics by studying the decays of familiar particles in great detail (see table, above). For example, the Tevatron feeds two particle detectors, named CDF and D0, that are searching for the Higgs boson and other new particles. The proposed BTeV detector would have studied well-known particles called B mesons, which the Tevatron produces in spades. The$200 million effort was canceled in 2005, just as physicists expected the go-ahead for construction. Since 1999, only the MINOS neutrino experiment, which shoots a beam of the elusive particles to a detector in the Soudan Mine in Minnesota, has made it to completion.

Some physicists say it's unfair to blame the ILC for the demise of smaller experiments. In tight budgets, those projects simply weren't worth the costs, they say. For example, BTeV would have required running the Tevatron into the middle of the next decade at a cost of $40 million per year. “I don't think that anyone can say that BTeV was canceled because of the ILC,” says Barry Barish, a physicist at the California Institute of Technology in Pasadena and leader of the ILC Global Design Effort. But Raymond Orbach, DOE undersecretary for science, who declined to be interviewed for this article, has indicated in the past that DOE was foregoing smaller projects in favor of the ILC. “There is a fear, and the fear is well-grounded, that we may be sacrificing a lot and that [the ILC] may not come to pass,” Orbach told Science in a June 2006 interview. “But if we don't take the risk, then we won't have the ILC on shore. … I want it here, and I want the United States to maintain its leadership in this area, and it's the only way I know how to do it.” Orbach struck a more cautious tone 8 months later. On 8 February 2007, physicists working on the ILC design released a cost estimate that indicated that if the machine were built in the United States, it would cost upward of$10 billion, of which the nation's share would be roughly $7.5 billion (Science, 9 February 2007, p. 746). Two weeks later, Orbach told researchers on DOE's High Energy Physics Advisory Panel that the ILC probably could not be built until the middle of the 2020s and asked for smaller projects that the United States could pursue in the meantime (Science, 2 March 2007, p. 1203). Orbach's warning suddenly presented Fermilab physicists with a gap of 15 years or more without an accelerator project. And with nothing beyond the planning stage, many physicists say, the lab became an easy target for congressional budget cutters who had to shear$22 billion from the 2008 budget to avoid a veto by President George W. Bush.

Now Fermilab researchers have come up with a plan to restore their future. To some measure, lab leaders hope to take up where they left off before this year's crisis. Congress did not cancel the NOνA neutrino experiment, notes Oddone, and physicists hope to resume work on it. Similarly, DOE has requested 35 million for ILC work in 2009. But most say that, if it is to survive, Fermilab needs a new accelerator project, and getting one may be difficult because the United States's particle physics community has painted itself into a corner, says Joel Butler, a 28-year veteran of Fermilab. “The things we can do are deemed not grandiose enough, and the things that are grandiose enough we can't afford,” he says. Fermilab hopes to solve that paradox with Project X, the conceptual son of the Proton Driver. Similar to the Proton Driver, Project X would consist of a superconducting linear accelerator measuring 700 meters long. Like Proton Driver, it would produce intense beams of protons that could be used to generate neutrinos and other familiar particles. But unlike the Proton Driver, the guts of Project X—the “cavities” through which particles surf on electromagnetic waves—would be more like those in the ILC, says Young-Kee Kim, deputy director at the lab. “The technology is aligned, so any progress we make with Project X will help us with our efforts to host the ILC,” she says. Fermilab hopes to have Project X up and running by 2016, but securing it is not a slam dunk. Many physicists question whether the menu of experiments it would support—precision studies of muons, K mesons, and neutrinos—is hearty enough to justify the expense and sustain the lab. “Thus far, I haven't seen anybody stand up and make the case that the physics that can be done with Project X is as important as building Project X” for the sake of the accelerator program, says Peter Cooper, a physicist at Fermilab. Kim says that everyone to whom she's presented the science case seems convinced. Project X also has competition. Japanese physicists will fire up their own proton source, the Japan Proton Accelerator Research Complex (J-PARC), this year. It will pursue much of the same physics as Project X. At least in its first incarnation, J-PARC will produce a beam only half as intense as Project X's. But researchers already plan to upgrade the facility, and Japanese scientists will enjoy a head start of at least 8 years over their Fermilab rivals. If Project X is going to help, then lab and DOE officials will have to hustle it along. The Tevatron will shut down in just over 2 years, and the accelerators that feed it and the current MINOS neutrino experiment probably won't run much longer. And if the lab goes too long without a working accelerator, it will likely lose the people it needs to build a new one, says Fermilab accelerator boss Holmes. “If you're not operating an accelerator, then you're not going to be able to design and construct a future facility,” he says. Fermilab is hoping to get DOE's preliminary okay in 2009 and start construction in 2012. If nothing is in the works by the time the Tevatron shuts down, then Fermilab will likely cut another 10% to 15% of its staff, Oddone says. Fermilab's Butler warns that, having shelved the Proton Driver once, physicists and DOE may have a tough time selling Congress on a similar project and the experiments it can do. “It's going to be difficult to say that the things that we said weren't that important are now the most important things,” Butler says. ## The no longer unthinkable If Project X does not come to fruition, Fermilab won't vanish as soon as the Tevatron shuts down. It will still be the national headquarters for the 630 physicists from the United States who are working on CMS, one of four gargantuan particle detectors at CERN that will be fed by the LHC. The lab is also broadening its mission into astrophysics and cosmology. For example, Fermilab is one of 25 institutions in the Sloan Digital Sky Survey, which since 1998 has used a 2.5-meter telescope on Apache Point, New Mexico, to map 1/5 of the sky. Fermilab also leads the proposed Dark Energy Survey, which would use the 4-meter Blanco Telescope at Cerro Tololo in Chile to probe the bizarre dark energy that is accelerating the expansion of the universe. But such efforts cannot sustain the lab at its present size. What's more, if Fermilab has no operating accelerator to anchor it, these other activities could be moved to other institutions, researchers worry. “The worst [possible outcome] is that we get shut down,” says Fermilab physicist Stephen Pordes, “and the question is, do we get shut down quickly or slowly.” Still, there are rays of hope for the lab, and Florencia Canelli is one of them. One of Fermilab's best and brightest, the 35-year-old Argentinean holds a Wilson Fellowship—the equivalent of a tenure-track professorship at a university—and has been working at the lab since 1997, when she was a grad student at the University of Rochester, New York. She and her husband, a postdoc from Ohio State University in Columbus, both work on the CDF particle detector, which is fed by the Tevatron. They have fielded offers of dual professorships from three different universities. But Canelli has just decided to stay, taking a joint position with the lab and the University of Chicago, and her husband is taking a staff position at the lab. The two want to devote themselves full-time to exploiting the Tevatron data and gearing up for the LHC, she says. Their decision marks a small victory not only for Fermilab but also for the U.S. program. Canelli has Italian citizenship, and her husband holds a passport from the United Kingdom; in principle, they could take off for Europe. But the United States offers opportunities that Europe does not, Canelli says, such as the chance for anyone with talent to climb to the top. “It's a good thing about the U.S.,” she says. “I haven't heard of a lot of non-Italians getting a position in Italy or a lot of non-French people getting positions in France.” Canelli says she remains optimistic that the United States won't drop out of the most fundamental physics. “What is matter made of? How does the universe work?” she says. “We will always have these questions. The only question is, is this country going to be involved in getting the answers.” Right now, the answer is a definite maybe. 8. PARTICLE PHYSICS # Whither the International Linear Collider? 1. Adrian Cho Efforts to develop the International Linear Collider (ILC), a 40-kilometer-long, straight-shot particle smasher, have taken some thumps in the past 16 months. But like a seasoned pugilist, the ILC has rolled with the blows, project leaders say. Efforts to develop the International Linear Collider (ILC), a 40-kilometer-long, straight-shot particle smasher, have taken some thumps in the past 16 months. But like a seasoned pugilist, the ILC has rolled with the blows, project leaders say. “We've been slowed down,” says Barry Barish, a physicist at the California Institute of Technology in Pasadena, who leads the ILC Global Design Effort (GDE). Still, he says, “in terms of the threat of it being turned off, I don't think there's much chance of that.” Physicists generally agree that the ILC or something like it represents the future of particle physics. This summer, the European lab, CERN, near Geneva, Switzerland, will turn on the Large Hadron Collider (LHC), which could cough up a slew of new particles and perhaps reveal new dimensions of space. The LHC will fire protons into protons, each of which is a knot of particles called quarks and gluons, so it will produce extremely messy collisions. The ILC would collide indivisible electrons and positrons and produce cleaner collisions, which should allow researchers to study in detail the new particles glimpsed by the LHC. The ILC's troubles began in February 2007, after the GDE released a cost estimate for the machine (Science, 9 February 2007, p. 746). It set the “value” of the ILC at6.7 billion, not including contingency or inflation during planning and construction. Adding those factors meant that, if the United States hosted the ILC and paid for half of it, its share would total $7.5 billion. Two weeks later, Raymond Orbach, undersecretary for science at the U.S. Department of Energy (DOE), warned that it could take until 2025 or later to get the go-ahead for the machine (Science, 2 March 2007, p. 1203). Then in December, the U.K.'s Science and Technology Facilities Council announced that Britain was pulling out of the project entirely, saying it could “not see a practicable path towards the realization of this facility” (Science, 21 December 2007, p. 1851). Two weeks later, a quarter of the way into the 2008 fiscal year, Congress cut funding for ILC research and development from a requested$60 million to $15 million, stopping work in the United States for the year (Science, 11 January, p. 142). Still, physicists in Europe and Asia continue to soldier on and make progress, Barish says. And Spain and India have recently joined the effort. What has really suffered, Barish says, are the chances that the machine will be built in the United States, at Fermi National Accelerator Laboratory (Fermilab) in Batavia, Illinois. “The most likely place that a machine like this will be built is CERN,” Barish says. “It's hard to see a scenario that would bring it to Fermilab at this point.” The United States's prospects for hosting the machine suffered not so much because of the cuts to ILC development but more because of cuts to the U.S. contribution to the international fusion experiment, ITER, which will be built in Cadarache, France. Congress zeroed out the$149 million that the United States was supposed to contribute this year, leaving the six other ITER partners in the lurch. “At the moment, the U.S. is not a reliable partner for long-term projects, with the obvious consequence that few people think the U.S. is a probable candidate” for hosting the ILC, says Albrecht Wagner, head of the German Electron Synchrotron Laboratory (DESY) in Hamburg and chair of the International Committee for Future Accelerators.

The United States is not giving up on hopes for landing the ILC, however. “There was never—never—a suggestion in my comments or my actions that we were somehow moving away from the ILC,” Orbach told Science in an interview in January. DOE has requested \$35 million for ILC research and development in 2009. However, observers say Congress is likely to continue with the present budget until a new president takes office in January 2009, at which point the United States's role in the ILC will lie in the hands of the next Administration.

9. ANTARCTICA

# Freeze-Dried Findings Support a Tale of Two Ancient Climates

1. Douglas Fox*
1. Douglas Fox is a science journalist based in northern California.

A surprising cache of ancient plant material adds evidence for divergent climate histories of the East and West Antarctic ice sheets over the past 14 million years.

A surprising cache of ancient plant material adds evidence for divergent climate histories of the East and West Antarctic ice sheets over the past 14 million years

MCMURDO STATION, ANTARCTICA—Setting out on foot from camp on a clear, cold November day, three graduate students picked their way around the boulder-strewn flank of Mount Boreas, a naked butte in the Olympus Range of the Dry Valleys in Antarctica. The students had wandered 20 kilometers in search of glacial deposits or volcanic ash that would help them reconstruct the geologic history of this region 120 km from McMurdo Station. As they climbed over a pile of boulders, they spotted a low drift of white powder concealed in a depression.

At first, it looked like volcanic ash. But when they dropped to their knees and dug in with their fingers, they discovered papery layers of brown and white that resembled the stacked pages of a buried dictionary. Along the edges, inch-long fibers fluttered in the wind. The soft organic material was out of place in this sterile, stone-paved landscape. Once the trio had returned to camp, their microscopes revealed the brown and white layers to be desiccated slabs of an ancient lakebed that had been pried up by powerful winds. The white powder on the surface turned out to be thick deposits of diatoms. And the fibers were tiny brown stems and shriveled leaves.

The 2000 discovery, only now coming to light through a series of recent meeting presentations, represents the last gasp of a tundra ecosystem before it plunged 14 million years ago into a winter from which it has never emerged. It's one of several findings that point to an alpinelike ecosystem millions of years ago that has remained in a freeze-dried state ever since. “For 14 million years, presumably, they've been close to the surface,” says paleoecologist Allan Ashworth of North Dakota State University (NDSU) in Fargo, who is coordinating analysis of the specimens. “If the climate had gotten warmer and wetter, microbes would have mined these deposits as carbon sources.”

Juxtaposed against these findings are results from two recent cores drilled into the seabed of McMurdo Sound, 120 kilometers from the Olympus Range. The analysis of the ANDRILL (ANtarctic geological DRILLing) cores, taken in the fall of 2006 and 2007, indicate huge fluctuations in temperature over the same period in that general region.

These findings appear to be contradictory at first glance, but in fact they buttress an evolving view among scientists that the two major features of the continent, the western and eastern ice sheets, have experienced vastly different climate histories. Data from the Dry Valleys reveals an East Antarctic Ice Sheet that is high, dry, cold, and stable, at least in its central area. And the ANDRILL cores suggest a more volatile West Antarctic Ice Sheet that is subject to the changing temperatures of the sea in which it wades. “It reaffirms the fragility of the West Antarctic Ice Sheet [WAIS] and the stability of the central part of the East Antarctic Ice Sheet,” says Peter Barrett, a sedimentologist at the Victoria University of Wellington (VUW) in New Zealand, who advised the ANDRILL project.

Those results also have major implications for what lies in store for the two ice sheets, which together cover more than 95% of the frozen land mass. Their fate represents a key element in any scenario about the long-term effects of global warming.

In fact, there's so much at stake, and the fossil findings were so unexpected, that David Marchant, a geomorphologist at Boston University (BU), initially suspected a practical joke when his students asked him to look through the microscope. “I thought someone had snuck in grass from New Zealand,” he says. “It looked so fresh. As soon as I knew it was a real find, I thought this was among the biggest discoveries in the Dry Valleys, and maybe all of continental Antarctica, in the last decade.”

## Freeze-dried samples

Adam Lewis, then pursuing his Ph.D. at BU under Marchant, had already spent several years studying glacial deposits in the Olympus Range. That work had given him a nodding acquaintance with dating them and allowed him to draw a general picture of past climates in the region. So although the withered roughage spotted by him and fellow grad students Jane Willenbring and Brett VandenHeuvel was novel, he immediately sensed its significance. “We knew it could be used to reconstruct temperature,” says Lewis. “We knew that we would be able to say, 15 million years ago, how warm it was in summer.”

It was some time before the graduate students could indulge their curiosity, however.

“We bagged it up, wrote down our notes, and made a little rock pile,” says Lewis, now on the faculty at NDSU. “Then we had to get back to work on glacial geology.” The group—Willenbring is now a postdoc at the University of Hanover, Germany, and VandenHeuvel is an environmental lawyer in Oregon—kept the find secret to prevent the site from being excavated before its stratigraphy was properly dated.

But Lewis returned 2 years later, and the story is now appearing in a series of recent presentations and upcoming publications. Lewis and Ashworth also star in a new documentary, Ice People, that describes their work in the Dry Valleys.

The shriveled fibers Lewis and his colleagues stumbled across were 14-million-year-old aquatic mosses whose leafy stems once undulated in shallow, lazy currents. When dunked in water, which Ashworth immediately tried with some of them, they reinflated and unfurled. “These things look like freeze-dried museum samples,” says Ashworth. The bodies of pea-sized freshwater seed shrimp, or ostracods, have also turned up with their soft mouth parts intact. A second lakebed, also in the Dry Valleys, offered up pollen, leaves, and twigs of southern beech—the twigs still composed of pliable, burnable wood. And lake sediments have coughed up megaspores of aquatic quillwort plants and fragments of insects, including a species of weevil.

These fossils paint a picture of an alpine lake dammed behind glacial moraines, surrounded by tundra and weather-beaten southern beech shrubs. The setting is similar to modern, above-tree-line communities in New Zealand or Patagonia. The wide range of diatom species show that the lake didn't dry up seasonally, says Alex Wolfe of the University of Alberta in Edmonton, Canada, who examined the diatoms: “It was a proper lake ecosystem.”

This spongy lawn of alpine tundra suggests an average summer temperature well above freezing, according to early results presented last summer by Marchant, Lewis, and Ashworth at the 10th International Symposium on Antarctic Earth Sciences in Santa Barbara, California. (More complete climate reconstructions are in press.) That's at least 20°C warmer than current conditions, in which only a few minuscule nematodes, springtails, and mites eke out an existence below the surface.

Isotope analysis of volcanic ash sprinkled over one lakebed dates the tundra at 14.1 million years. Last November, Marchant and Lewis published geomorphic surveys suggesting that the Olympus Range shifted within a few hundred thousand years from warm, seasonally melting glaciers that fed lakes to dry-based glaciers that sublimated rather than melted. The dry glaciers have advanced and evaporated since then, but cosmogenic nuclide dating, which estimates the times that rocks on the surface have been exposed to cosmic rays, suggests that this landscape—perhaps the oldest on Earth—hasn't tasted liquid water in the last 14 million years. “We'll move a boulder, and we'll think to ourselves, ‘Our ancestors were swinging in the trees, and that boulder was already there,’ “says Lewis.

## Cork in the bottle

While the portion of Antarctica east of the Transantarctic Mountains stayed locked in the freezer for the last 14 million years, West Antarctica, which constitutes about a fifth of the continent, was having a much more on-again, off-again relationship with its ice. The two 1200-meter cores from the ANDRILL project provide a 20-million-year climate record. The 2006 core was of particular interest because it came from seabed beneath the northwest corner of the Ross Ice Shelf, a Spain-size slab of floating ice that feeds out of the West Antarctic Ice Sheet.

Glaciologists regard the Ross Ice Shelf as a barometer of the ice sheet's health. It buttresses the flow of major glaciers flowing out of WAIS, providing what some call “the cork in the bottle.” Sea levels could rise by 5 meters were WAIS to melt. Satellite studies indicate that it is currently losing as much as 150 cubic kilometers of ice per year, and many observers consider it at risk for collapse because it slides on a bed that sits below sea level, where rising and warming seas could destabilize it. (All other ice sheets that survived the last deglaciation sit atop land.) The disappearance of the Ross Ice Shelf would lead to an acceleration of WAIS's glaciers and maybe major ice loss.

Alternating layers of diatoms and glacial debris from the ANDRILL cores revealed 60 cycles of advance and retreat of the Ross Ice Shelf over 14 million years, according to results reported last December. “The variation surprised us,” says Timothy Naish, a paleoclimatologist at VUW, who co-led the 2006 ANDRILL effort and presented in April at the assembly of the European Geosciences Union in Vienna, Austria. Most striking of all was a 90-meter layer of green, diatom-rich sediment that revealed 200,000 years of biologically productive, ice-free sea starting 4.2 million years ago. “So as well as seeing variability,” says Naish, “we're also getting this sense of extended periods when the West Antarctic Ice Sheet was very small, if not gone altogether.”

For Lewis, an ice-free McMurdo Sound teeming with life at various points in the last 14 million years is consistent with having the Olympus Range maintained in a freeze-dried limbo if one factors in the powerful winds screaming off the polar plateau. “There has to be a fairly big ice sheet perched right behind those mountains [in East Antarctica],” he says, “to blow all that cold air down the slope.”

That the Ross Ice Shelf underwent major collapses rather than minor fluctuations is supported by glacial drop stones found in the 2006 ANDRILL core. The stones come from 300 kilometers south, where Byrd Glacier pours through the Transantarctic Mountains. Rocks scooped up by Byrd were delivered to McMurdo Sound because the Ross Ice Shelf bent the glacier's flow. But in sections of core showing open sea, rocks come from local sources. To Naish, it means that the Ross Ice Shelf was absent, or least too small to bend Byrd Glacier toward McMurdo.

The ANDRILL cores support a volatile view of West Antarctica and the Ross Ice Shelf that was emerging from earlier studies. In 1995, glaciologists drilled 1000 meters to the base of WAIS and found marine diatoms in the subglacial sediments. These diatoms, dated between 120,000 and 1 million years old, indicate an open sea and, hence, a major ice-sheet collapse.

Reed Scherer, the paleoecologist at Northern Illinois University in DeKalb who dated the diatoms, sees corroboration for WAIS collapse in other records. Beaches, corals, and water lines suspended in sea cliffs high above current water levels reveal major sea level spikes on at least three occasions during the time when Scherer's diatoms might have grown: at 125,000 years, 400,000 years, and 1.07 million years ago. The million-year-old event corresponds with an episode of open sea in the 2006 ANDRILL core. It also corresponds to a spike in 16O/18O ratios in marine cores worldwide, a sign that ice sheets had injected fresh water into the oceans, because ice sheets preferentially incorporate water molecules containing 16O over 18O.

One potential payoff from knowing Antarctica's history is a better understanding of how its ice will respond to global warming. The 90-meter segment of diatom-rich core that Naish helped drill in 2006 points to an open Ross Sea 4 million years ago, a time when marine-core records from other parts of the world suggest that CO2 levels hovered around 400 parts per million, with global temperatures 3° to 4°C warmer than today. Naish notes that those conditions correspond to optimistic predictions in last year's Intergovernmental Panel on Climate Change report of CO2 levels in 2100. “If we can control our carbon emissions, we might be able to keep warming to 3°C above present and stabilize CO2 at 450 parts per million,” he says. “But even at those levels, we don't have a West Antarctic Ice Sheet,” implying that WAIS would disappear in the coming centuries.

Marchant and Lewis are already looking for fossils in other parts of the Dry Valleys. Ashworth advocates broadening the search to provide a more nuanced view of how Antarctica's two massive ice sheets have changed over time. For example, the East Antarctic Ice Sheet as a whole seems to have been stable over time. But some areas around its fringes sit on marine beds like WAIS does, and people would love to know more about the history of those areas. “All the way around the edge of Antarctica, you could have pockets of these fossils preserved,” says Ashworth. “I've got this gut feeling that [more of] these deposits are going to be turning up.” If he's right, what was once a rare discovery may someday become a standard tool for understanding the history of Antarctica's ice sheets and climate.

10. MEDICINE

# Nothing Rotten About Hydrogen Sulfide's Medical Promise

1. Mitch Leslie

Despite its toxicity and famously bad odor, hydrogen sulfide's ability to lower metabolism and create a hibernation-like state has scientists wondering whether the gas can help soldiers and other people withstand injuries or surgeries.

Despite its toxicity and famously bad odor, hydrogen sulfide's ability to lower metabolism and create a hibernation-like state has scientists wondering whether the gas can help soldiers and other people withstand injuries or surgeries

It's foul-smelling, corrosive, flammable, and deadly. It's the bane of oil fields, sewage treatment plants, and farms because at concentrations workers sometimes encounter, a single breath of it can kill. Hydrogen sulfide (H2S), the rotten egg gas, is not something you would think to pump into sick or injured people.

But that's exactly what some scientists plan to do, reflecting the gas's improving reputation over the past 2 decades. Long known for its distinctive smell and toxicity—it starves our cells by disabling an enzyme necessary for extracting energy from food—the molecule has proven to be an influential physiological signal, with effects on everything from blood flow to hormone secretion. Eager to capitalize on these newfound capabilities, scientists are trying to exploit hydrogen sulfide to tame the side effects of common painkillers, for example, and curb heart attack damage. After announcing last week that injecting low doses of hydrogen sulfide into healthy people produced no dangerous side effects, one company plans later this year to start testing the molecule as a treatment for several conditions, possibly including restricted blood flow to the liver and lung injuries. “We are right at the beginning of an expanding field that could have enormous clinical implications,” says cardiovascular physiologist David Lefer of Albert Einstein College of Medicine in New York City.

In the past few years, hydrogen sulfide research has also veered into science-fiction territory, as investigators have found that nonlethal doses of the molecule can send small animals into a hibernation-like state. Whether this unexpected effect can be reproduced in large animals, or people, remains a matter of debate, but that hasn't stopped some scientists—and the U.S. military—from investigating whether the gas could allow patients to better survive severe injuries or traumas, such as a stroke, by placing them in a form of suspended animation.

## Not just sewer gas anymore

Toxicologists know plenty about the downside of hydrogen sulfide. Even at 10 parts per million (ppm), the exposure limit set by the U.S. National Institute for Occupational Safety and Health, the gas can irritate the eyes. If you breathe 500 ppm, you can die within half an hour, and 1000 ppm knocks you out instantly and kills within a few minutes.

So researchers were surprised to discover that the human body naturally makes the potentially lethal molecule, although at much lower concentrations. This ability represents a legacy from some of the earliest microbes, says physiologist Rui Wang of Lakehead University in Thunder Bay, Canada. These organisms—like a few living today—relied on sulfur, not oxygen, to obtain energy from metabolism. Although oxygen took over this job for the most part, organisms still put hydrogen sulfide to use for numerous other functions.

Hydrogen sulfide isn't the only noxious gas to reveal a good side. Nitric oxide was the first so-called gasotransmitter that scientists identified. Nobel Prize-winning work that started in the mid-1980s demonstrated that nitric oxide relaxes blood vessels, quells inflammation, nudges the hypothalamus to release hormones, and even transmits signals between the brain's neurons. Another killer gas, carbon monoxide, later joined the gasotransmitter family (Science, 21 November 2003, p. 1320). Despite these precedents, hydrogen sulfide researchers say they still get grief for their work, particularly from toxicologists. “A week doesn't go by when I'm not answering questions like, ‘Isn't this a poison?'” says pharmacologist John Wallace of the University of Calgary in Canada.

Interest in hydrogen sulfide's possible benefits bubbled up about 20 years ago, when three papers described surprisingly high concentrations in brain samples from animals and humans. What was the molecule doing in healthy tissue, researchers wondered. It might help us to learn, replied neuroscientist Hideo Kimura and postdoc Kazuho Abe, then at the Salk Institute for Biological Studies in San Diego, California. In 1996, the pair found that in slices of rodent brain tissue, hydrogen sulfide spurs long-term potentiation, an increase in synapse sensitivity that can promote learning and memory. Unlike nitric oxide, hydrogen sulfide doesn't appear to transmit messages between neurons, says Kimura, who's now at the Institute of Neuroscience in Tokyo. Rather, it acts as a neuromodulator that adjusts the responsiveness of neural circuits.

The molecule's résumé includes other responsibilities, subsequent studies have shown. Many of its effects stem from the power to open the membrane channels that allow potassium to leave the cell, says Wang. For example, in 2001, his group showed that by prodding these channels, hydrogen sulfide relaxes smooth muscle cells in the walls of blood vessels, suggesting that the molecule regulates blood pressure. Researchers are still trying to resolve the sometimes contradictory reports on hydrogen sulfide's effects—for instance, studies conflict about whether it is pro- or anti-inflammatory—and nail down all of its functions. But physiologist David Kraus of the University of Alabama, Birmingham, predicts that hydrogen sulfide “will rival the prominence of nitric oxide.”

One apparent function of the molecule—soothing cells under stress—has attracted medical interest. Kimura and colleagues revealed 4 years ago that hydrogen sulfide can shield cultured neurons from oxidative damage, which often occurs after a stroke. Instead of combating oxidants directly, the molecule spurs cells to bump up the levels of glutathione, neurons' natural antioxidant. Researchers still need to test whether hydrogen sulfide curbs brain damage from a stroke in animals, Kimura says.

However, administration of hydrogen sulfide does seem to limit the damage from a heart attack, as Lefer and colleagues revealed last year. To simulate a clogged artery, the researchers temporarily tied off one of the vessels that delivers blood to the left ventricle in mice. An injection of hydrogen sulfide directly into the heart cut the amount of scarring and inflammation that resulted once blood flow resumed. Protection of mitochondria, the cell's energy-generating organelles, might explain that finding. In the untreated mice, the organelles' capacity to use oxygen plummeted. The mitochondria were swollen, and their complex interior structure appeared scrambled. This damage was absent in treated mice. Lefer and colleagues are already assessing hydrogen sulfide for other conditions, including heart failure, and plan to test sulfide-releasing pills that are now under development.

Pills that promote the creation of hydrogen sulfide might also protect the gut. Cells of the gastrointestinal (GI) tract naturally make the molecule, possibly to regulate blood flow or shield intestinal linings. In rodent studies, Wallace and his colleagues have found that intravenous administration of hydrogen sulfide fends off side effects, such as GI bleeding and ulcers, frequently caused by nonsteroidal anti-inflammatory drugs (NSAIDs), the drug class that includes aspirin and ibuprofen. Seeking easier-to-stomach NSAIDs, the researchers designed versions that emit small amounts of hydrogen sulfide. In tests on rats, the modification almost eliminated intestinal injury from the NSAID diclofenac, the team reported last year. These new NSAIDs “don't cause any GI damage at all and are as potent as the parent drugs,” Wallace says. He predicts that the drugs will reach clinical trials within 18 months. And a sulfide-delivering version of mesalamine, a treatment for inflammatory bowel disease that causes similar GI anguish, could be ready for testing later this year, he says.

## A big sleep

While researchers have been gradually unearthing the physiological roles of hydrogen sulfide, Mark Roth's heart-slowing experiments with the gas grabbed headlines 3 years ago. A physiologist at the Fred Hutchinson Cancer Research Center in Seattle, Washington, Roth was studying how animals drastically reduce their metabolism, as some do during hibernation. His lab found that small animals such as nematodes and zebrafish embryos could endure oxygen concentrations supposedly below the minimal level for survival. The animals remained barely alive—the embryonic fish's hearts often stopped beating—but revived when researchers cranked up the oxygen levels. The team then started looking for ways to induce the same effect by preventing animals from using available oxygen. Carbon monoxide worked for nematodes, but they thought it would be too risky for humans. Roth says he decided to try hydrogen sulfide instead after seeing a TV documentary that mentioned its dangers to cavers.

When Roth's team exposed mice to 80 ppm of the gas, which toxicology studies had established was safe for rodents, the animals passed out. Their core body temperature plunged more than 20°C, their oxygen consumption fell, and their carbon dioxide output—an indicator of metabolic rate—tumbled. Once the researchers shut off the gas and provided heat, the mice were up and gnawing within an hour. They passed a battery of behavioral tests, indicating that they incurred no brain damage during their down time. In a follow-up study published last year, Roth's team showed that hydrogen sulfide could enable mice to survive low oxygen concentrations that are otherwise lethal to rodents.

After being gassed, the mice slip into an altered state that differs from the unconsciousness of sleep, hibernation, and anesthesia. Their eyes are closed, but unlike anesthetized patients, they are not paralyzed and respond to pain, says Warren Zapol, chief of anesthesia at Massachusetts General Hospital in Boston, who has studied the effects of hydrogen sulfide on the rodents. “You couldn't take their appendix out.” he jokes. Pinch a tail, and they try to wriggle away, Zapol says. The animals aren't hibernating, either, as that involves much more gradual changes in everything from fat deposition to protein synthesis.

In an April paper in Anesthesiology, Zapol and colleagues provided the most detailed view yet of the cardiovascular changes induced by hydrogen sulfide. After breathing 80 ppm of gas, the animals' heart rate declined more than 50%, to about 250 beats per minute, the team reported. Typically, when a mouse's heart rate plunges, the organs can begin to run short of oxygen, Zapol notes. But the rodents' blood pressure remained steady, suggesting that the delivery of oxygen throughout the body didn't falter, he says.

Hydrogen sulfide appears to serve as a master metabolic regulator, Roth says. He notes that 18th century British chemist Joseph Priestly likened humans to burning candles, consuming oxygen to keep the flame going. “We have stumbled across the mechanism by which human beings are regulating the degree to which they burn their candle,” Roth says.

## Scent of survival

Roth now wants to lower the flame on people's candles. He envisions temporarily turning down metabolism and oxygen demand with a dose of hydrogen sulfide, buying time for patients who have suffered heart attacks, strokes, or wounds that produce drastic blood loss. In a study that The Journal of Trauma will publish in July, Roth and colleagues show that 66% of rats injected with a hydrogen sulfide solution can survive the loss of 60% of their blood, versus 14% of rats given a control solution. He's now testing for the same benefits in larger animals and has received a grant from the U.S. Defense Advanced Research Projects Agency to design injectable hydrogen sulfide kits that troops could carry in the field. Wounded soldiers could be “deanimated,” suggests Roth, until they can be evacuated to a hospital.

Roth acknowledges that using hydrogen sulfide in this way faces some public relations hurdles. People expect doctors and paramedics to resuscitate their loved ones, not put them into a near-death state, he says: “The notion that you are better off deanimated than animated is not something people want to think about.”

The bigger problem, say some critics, is that hydrogen sulfide won't elicit the same effect in people as it does in mice. Two studies on large animals bolster their skepticism. Last fall, for instance, a French team reported no metabolic decline in sheep that breathed 60 ppm of hydrogen sulfide gas.

Recent work on pigs by cardiologist Andrew Redington of the Hospital for Sick Children in Toronto, Canada, and colleagues suggests that nonlethal concentrations of the gas could exacerbate patients' problems. In their experiment, piglets that had been anesthetized much like a surgical patient breathed gradually increasing concentrations of hydrogen sulfide gas, from 20 ppm to 80 ppm. Instead of slowing the metabolism of the piglets, hydrogen sulfide revved it up, the team reported in January. Heart rate, blood pressure, and cardiac output were all higher in animals that breathed the gas than they were in controls. Many of the target patients for hydrogen sulfide treatment have weakened hearts, and forcing the organ to work harder could kill them, says Redington. His take is that in large animals, hydrogen sulfide “has no effect at best but possibly a detrimental effect.” Resuscitation researcher Samuel Tisherman of the University of Pittsburgh Medical Center in Pennsylvania isn't as harsh but concedes that “some of the promise seems to be slipping away.”

Roth counters that both large-animal studies share a crucial flaw: The concentration of hydrogen sulfide was too low. Although 80 ppm will knock out a mouse, a pig or sheep requires a bigger hit, he claims. He and his colleagues are conducting their own tests on large animals, but he says they're not ready to discuss the results.

Hydrogen sulfide doesn't need to produce dramatic effects to provide benefits, Roth notes. Ikaria, the company that he co-founded 3 years ago and for which Lefer serves as a consultant, has just announced the results of the first safety trial of an injectable form of hydrogen sulfide. Last week, the company revealed that 36 volunteers who received doses well below the level that causes unconsciousness in people showed no ill effects. Csaba Szabo, Ikaria's chief scientific officer, says that before the end of the year, the company plans to launch phase II trials. What types of patients will receive the drug isn't certain, but they could include people who have suffered heart attacks or who are undergoing operations such as a heart or lung bypass.

If suspended animation isn't futuristic enough for you, Roth published a paper last year hinting at another dramatic effect: increased longevity. His lab found that exposure to the gas stretches the life span of nematodes by up to 70%. Whether or not other organisms react the same way, it's clear that hydrogen sulfide has come a long way from just being a killer.