# News this Week

Science  02 May 2008:
Vol. 320, Issue 5876, pp. 594
1. ENDANGERED SPECIES

# Spotted Owl Recovery Plan Flawed, Review Panel Finds

A blue-ribbon panel of scientists has confirmed major flaws in the proposed recovery plan for the northern spotted owl, a threatened species that has driven forest policy in the northwestern United States for nearly 2 decades. As did earlier reviews, the final one, by the Sustainable Ecosystems Institute (SEI) in Portland, Oregon, concludes that the Fish and Wildlife Service's (FWS's) plan does not put enough emphasis on protecting the owl's habitat. It also says that massive thinning of dry forests is needed to prevent habitat from going up in smoke—a recommendation that makes some environmentalists nervous. All eyes are now on FWS, which intends to release a final version of the plan by the end of the month. The plan “could have a very profound effect,” says forest ecologist Jerry Franklin of the University of Washington, Seattle, who participated in the SEI review.

The northern spotted owl (Strix occidentalis caurina) lives mainly in old-growth forests, which range from British Columbia to northern California. After decades of logging and population decline, the species was put on the federal list of threatened species in 1990. FWS began to create a recovery plan to use in determining habitat critical for the species' survival. Although recovery plans lack the legal clout of regulations, they are a major influence on FWS decisions to permit logging and other activities that might harm a particular species. The recovery plan for the spotted owl was never finished, however, because it was superseded by the overarching Northwest Forest Plan, which guides forest policy across the region.

Many environmentalists think the Northwest Forest Plan wasn't protective enough—the spotted owl population continues to decline by 3.7% a year—whereas the timber industry has complained that it's too restrictive (Science, 29 July 2005, p. 688). In 2002, the American Forest Resources Council (AFRC), an industry group, sued FWS for failing to complete a review of the owl's status, which it is required to do every 5 years. As part of a settlement, FWS agreed to reexamine its designation of critical habitat by 1 June 2008. A separate lawsuit also led the agency to finalize its recovery plan for the species.

In spring 2006, FWS formed a team to draft the recovery plan that included a broad range of expertise, including environmentalists and timber industry representatives, but lacked top scientists; some declined to participate in part because they feared the process would be politically charged, they told Science. The team's draft focused on the need to protect habitat and also dealt with the threat from barred owls, an invasive species that is competing with the spotted owl.

Politics did trump science, say observers and participants. After the first draft was sent to Washington, D.C., in September 2006, officials at the Department of the Interior (DOI) ordered the recovery team to add another management strategy, called Option 2, says recovery team member Dominick DellaSala, an ecologist who directs the National Center for Conservation Science & Policy in Ashland, Oregon. This option would reduce the amount of land set aside for owl conservation and give the Bureau of Land Management (BLM) and the Forest Service more flexibility to allow logging. The immediate goal was to make the recovery plan consistent with a BLM proposal to facilitate logging in Oregon, according to internal agency e-mails provided to Science by DellaSala. He says that officials also wanted the plan to list the barred owl threat as more dire than loss of habitat—over the objections of some of the recovery team members, as well as James Tate, DOI's own science adviser.

FWS released the draft plan, including Option 2, for public comment and requested scientific review in April 2007. Anonymous peer reviews, organized by the Society for Conservation Biology and two other science groups, raised many concerns. In August, for example, reviewers recruited by the Wildlife Society, a nonprofit association of wildlife conservation and management experts, called the draft plan “seriously flawed.” The reviewers deemed the draft's overall focus on barred owls rather than habitat conservation “incredibly risky.” And Option 2, they concluded, “drastically reduces protection for owl habitat and maximizes flexibility given to land managers by allowing them to operate under a series of nebulous rules.” The Wildlife Society urged the agency to start over.

In December 2007, the agency contracted with SEI to analyze all the reviews and suggest scientifically valid recovery options. In a 157-page report released on 21 April 2008, nine leading owl and forestry experts echoed many of the previous criticisms of the draft plan. They confirmed that barred owls are a threat, recommended experiments to determine how they can be controlled, and reiterated that the draft plan underestimated the importance of protecting habitat.

In a major departure from both previous reviews and the draft plan, the SEI panel called for much more aggressive thinning to reduce the risk of massive forest fires, especially in the dry, eastern part of the spotted owl's range. “We think the threat of wildfire is so great that we need to do thinning,” says lead author Steven Courtney of SEI. Ecological restoration is also necessary, Franklin adds. In contrast, DellaSala and reviewers for the Wildlife Society say that more needs to be learned about possible detrimental effects of thinning on spotted owls.

The recovery plan is now being finalized to meet the June deadline for revising critical habitat, says FWS spokesperson Joan Jewett. She expects that thinning will be addressed. The timber industry agrees that thinning, and barred owls, are a serious threat, says AFRC President Thomas Partin. But he dismisses the impact of logging large trees, because he says that has been relatively minimal in recent years. DellaSala counters that the science says every hectare of owl habitat matters. Given the pressure from BLM and the Forest Service, he's pessimistic about how much protection the final plan will afford old-growth forests. “It might get decided in the courts,” he notes.

2. GLOBAL WARMING

# Mother Nature Cools the Greenhouse, but Hotter Times Still Lie Ahead

1. Richard A. Kerr

As climate-change skeptics like to point out, worldwide temperatures haven't risen much in the past decade. If global warming is such hot stuff, they ask, why hasn't it soared beyond the El Niño-driven global warmth of 1998? Mainstream climate researchers reply that greenhouse warming isn't the only factor at work. And in a new paper, they put some numbers on that rebuttal. They show that regional and even global temperatures are being held down by a natural jostling of the climate system, driven in large part by vacillating ocean currents. The study “shows how natural climate variability can mask the global warming effect of greenhouse gases,” says climate researcher Adam Scaife of the Hadley Centre for Climate Prediction and Research in Exeter, U.K., “but only for a few years.”

The latest reminder of climate's confounding subtleties comes in climate forecasts that Noel Keenlyside of the Leibniz Institute of Marine Sciences in Kiel, Germany, and colleagues published this week in Nature. Rather than simply predicting temperatures at the end of the century, as most modelers do, they ran their simulations only 10 and 20 years into the future. At such a time range, short-term effects can override the contributions of rising greenhouse gases (Science, 10 August 2007, p. 746). For example, great, heat-carrying currents like the Gulf Stream can slow down and speed up, cooling and warming surrounding continents. As a result, temperatures rise and fall from decade to decade even in the absence of human interference.

To take account of such ocean-driven natural variability, Keenlyside and his colleagues began their model's forecasting runs by giving the model's oceans the actual sea surface temperatures measured in the starting year of a simulation. Providing the initial state of the ocean doesn't make much difference when forecasting out a century, so long-range forecasters don't usually bother. But an initial state gives the model a starting point from which to calculate what the oceans will be doing a decade hence and therefore what future natural variability might be like.

The added observations did in fact improve simulations of past climate variations. Looking into the future, the model forecasts a slowing of heat-carrying Atlantic currents and thus a cooling over the North Atlantic, North America, and western Europe in the next decade. It even predicts a slight cooling of the globe. But by 2030, forecast global temperatures bounce back up to the warming predicted with greenhouse gases alone.

The forecast is not the first to herald a slowing or even a temporary cessation of global warming. A study involving even more ocean observations inserted at the beginning of model runs reached similar conclusions last year. “The different approaches give slightly different results,” says climate modeler Douglas Smith of the Hadley Centre, who headed the earlier study, but “we do agree there's a temporary offset of global warming due to natural variability.” So if you're a climate-change activist pointing to year after year of mounting climate crises, you might want to rethink your approach.

3. LOUISE SLAUGHTER INTERVIEW

# How to Get a Genetic Protection Law Through Congress? Keep Trying

1. Jennifer Couzin

Persistence pays. After 13 years of rejection, a bill to protect individuals from employment and insurance discrimination based on their DNA, the Genetic Information Nondiscrimination Act (GINA), is poised to become law. Last week, the U.S. Senate voted unanimously to approve the bill; the House of Representatives was scheduled to vote on it as Science went to press. And President George W. Bush has vowed to sign it if it reaches his desk.

GINA has had many champions, but one of its most stalwart backers and the only co-author who trained in microbiology and public health is Representative Louise Slaughter (D-NY). Slaughter, 78, has introduced the bill in every Congress since 1995. She was first driven by concerns that individuals would be reluctant to participate in genetic research without the protections that GINA offers. But the changing genetic landscape has led her to expand her thinking, and she now hopes that GINA will put families more at ease about taking one of the growing number of genetic tests hitting the market.

Navigating the politics that held GINA up has been tricky. The bill didn't make it to a vote in the House until after the Democrats took control of Congress in 2006 and overcame the concerns of health insurers and other businesses that feared it would raise costs. GINA passed twice in the Senate, in 2003 and 2005, but was later blocked single-handedly by Senator Tom Coburn (R-OK). Coburn, who said the bill might expose businesses to legal liabilities, recently withdrew his objections. Slaughter was upbeat in a conversation with Science earlier this week.

Q: Why did you first introduce the bill back in 1995?

L.S.: I'm a scientist, and I believed in the Human Genome Project. We didn't know much about it at the time, but I really felt that since I was in Congress … that the social policy should keep up with the science. I wanted them to move in tandem.

Q: Why has it been such a long road?

L.S.: I think it's been a total misunderstanding. … In the Senate, a single man held up 85 bills including this one—he didn't understand it.

Q: You're talking about Senator Tom Coburn?

L.S.: Yes. [For many years,] this bill was held up in the House; there was never a hearing. We always had far more cosponsors than we needed to pass it but were not able to get it on the calendar.

Q: Why was that?

L.S.: The drug companies and the insurance companies didn't want it.

Q: Why not?

L.S.: They were mistaken. Everyone has bad genes. Dr. Francis Collins [head of the U.S. National Human Genome Research Institute in Bethesda, Maryland, and a longtime champion of GINA] had said to us early on, the president of the drug company and the president of the insurance company have bad genes. It's to their benefit and the benefit of their families that we know all we can.

L.S.: It started in research, but we knew the promise of this science was so great, being able to provide health care in an entirely new way. Not only to save billions of dollars in health care money, but we could save in human suffering. It never made any sense to me that we would have all these roadblocks.

Q: How will GINA help?

L.S.: People will not be afraid to allow themselves to be part of a research program because they will know it will not affect their employment or insurance. …

But the most important thing is the number of people [who worried about this]. …We had one constituent who had a lot of Alzheimer's in his family. He very much wanted to know if he had that gene in his family but was afraid to get tested. Many went to clinics [for genetic testing] under assumed names.

Q: Are more people going to buy genetic tests now? Is that problematic, given that there are concerns about the quality of these tests?

L.S.: We may get higher quality now. Once we got the bill passed, obviously protecting your information means getting a good test as well, in my book.

Q: What's next for you in this area?

L.S.: I'm in a wonderful spot for a microbiologist—given food that's unsanitary, outbreaks of E. coli. There's a number of things that we are very much aware of and worry about. We want the FDA [U.S. Food and Drug Administration] to be much stronger, much, much, much stronger.

4. ECOLOGY

# Yosemite: Protected but Not Preserved

1. Virginia Morell

The spectacular landscape of California's Yosemite Valley looks natural and has been protected for more than a century. Yet ecologists know that today's valley is not the one that enchanted naturalist John Muir when he promoted the formation of Yosemite National Park. One of the park's emblematic species, the California black oak (Quercus kelloggii), is in decline, a problem usually attributed to a lack of fire and an invasion of conifers. Now, two researchers have traced the trees' suffering through a complex chain of effects that starts with human influence and extends from cougars through mule deer to oaks and primroses. The result of that “trophic cascade” is that there are very few young oak trees to replace their elders, hurting other species from shrubs to birds and invertebrates and apparently reducing overall biodiversity, according to a report in the May issue of Biological Conservation.

“It's not just Yosemite,” says ecologist Thomas Rooney of Wright State University in Dayton, Ohio, who was not involved in the study. He notes that other oak species are in decline across the United States. “It shows that habitat protection alone is not enough. You need the predators.”

In Yosemite, the paper says, the direct cause of oak mortality is the high density of mule deer, which have been munching the oaks' basal sprouts and seedlings for nearly a century. Intriguingly, the mule deer's abundance stems not from a lack of predators—as with elk in Yellowstone National Park (Science, 27 July 2007, p. 438)—but from shy ones: elusive cougars (Puma concolor). The mountain lions keep the deer in check elsewhere in the park but avoid areas like the valley, where people congregate.

“There are higher deer densities now than in the 1850s,” when American settlers first entered Yosemite Valley, says ecologist William Ripple of Oregon State University, Corvallis, who co-authored the study with OSU colleague Robert Beschta, a forest hydrologist.

After Yosemite became a national park in 1890, visitors swarmed into the valley, eager to see its parklike landscape of black oaks and famed wildflower-filled meadows. Officials began eliminating cougars, bobcats, and coyotes and protecting mule deer from hunting. By 1925, deer were numerous, and park observers noted a marked decline in the most popular wildflower, the evening primrose (Oenothera hookeri). Although not apparent at the time, that's also when the black oak seedlings began to disappear, say Ripple and Beschta, who in 2006 measured the diameters of more than 3000 black oaks. They also took tree ring cores from 40 sites close to the valley's visitor center and sites 4 to 8 kilometers away. Black oaks can live as long as 500 years, but a “healthy stand includes a mix of young and old trees,” says Beschta. Oaks close to the visitor center have almost no young trees or basal sprouts. The deer are using people as “protective shields,” says Ripple. Deer had also nipped off the flower buds of nearly every evening primrose the scientists saw.

Today, Yosemite managers burn prescribed areas to keep out conifers and clear the way for oaks. But fires can't do what most needs to be done, says Ripple: “Get the baby oaks to grow.” That's not likely to happen until there are fewer deer. “It wouldn't be popular to have culling in a national park,” Rooney says, “but it may be necessary” if the valley's biodiversity is to be preserved.

5. TROPICAL DISEASES

# Dispute Clouds the Future of U.S. Naval Lab in Indonesia

1. Dennis Normile

A United States naval laboratory in Jakarta that studies tropical diseases may fall victim to Indonesia's determination to develop its own research capabilities and take control of its H5N1 viral samples. Negotiations over the lab's status have dragged on for more than 2 years, and now some Indonesian politicians are calling for it to be closed or taken over by Indonesia. Recently, the Indonesia Ministry of Health ordered institutions to stop sharing all medical samples with the facility, which has severely limited what it can do.

Opened in 1970, the U.S. Naval Medical Research Unit No. 2 (NAMRU-2) is one of five U.S. tropical disease research labs scattered around the world. NAMRU-2's main function is to study infectious diseases that might affect U.S. troops, although it has taken on a wider public health role in working with Indonesia's Ministry of Health, for instance, in supporting efforts to help curb malaria and dengue. In addition, NAMRU-2's staff of about 20 Americans and 150 Indonesians has trained hundreds of Indonesian health workers, researchers, and students in lab techniques. It also monitors emerging diseases throughout Southeast Asia in collaboration with local institutes.

Responding to e-mailed questions, Trevor Jones, NAMRU-2's commanding officer, wrote, “The U.S. benefits because American scientists have the opportunity to study tropical disease transmission where it is actually occurring.” He added that the restrictions have slowed the lab's work, but he hopes this is temporary.

One of the most worrisome diseases circulating in the world is the H5N1 strain of avian influenza, and Indonesia is by far the hardest-hit country, with 132 confirmed human cases and 107 deaths. NAMRU-2 performed the lab diagnosis of all human cases of H5N1 in Indonesia from June 2005 to January 2007, when Indonesia's Ministry of Health labs took over the duties. Since then, Indonesia has only sporadically shared samples of the H5N1 virus with the World Health Organization (WHO), demanding more equitable access to any vaccines or other benefits derived from those samples, which are used to monitor virus evolution, drug resistance, and pandemic risk (Science, 23 February 2007, p. 1065).

Now that dispute has extended to NAMRU-2 and is hampering its research on all tropical diseases. When the lab's Memorandum of Understanding expired in December 2005, U.S. officials assumed negotiating a new agreement would be routine. NAMRU-2 continued its work until the Ministry of Health halted all sample sharing on 31 March. Tensions escalated further after an early April visit by U.S. Health and Human Services Secretary Michael Leavitt, who wrote on his blog on 14 April that Indonesia's Minister of Health, Siti Fadilah Supari, “has used the sample-sharing debate and the negotiations over the status of NAMRU-2 in Indonesia to set herself up as an antagonist of the United States.” Last week, several Indonesian politicians joined the fray, calling for NAMRU-2 to be closed or taken over by Indonesia.

The U.S. Embassy and NAMRU-2 are now trying to reassure Indonesians of the lab's good intentions and negotiate a new agreement. John Heffern, deputy U.S. ambassador, held a press conference on 24 April to defend the lab's activities, and NAMRU-2 held a media open house on 25 April. One remaining point of contention, Heffern told the press, is providing diplomatic immunity for the Americans working at the lab, which the U.S. believes to be standard practice but many in Indonesia are now questioning. Local newspapers have reported that other issues pertain to technology transfers. (Calls seeking comment from Indonesia's Ministry of Foreign Affairs were not returned.)

The outcome of the negotiations will also affect NAMRU-2's status as a WHO collaborating center for emerging diseases. That designation, in effect since 1997, is on hold pending the resolution of its status. David Heymann, who heads WHO's pandemic influenza efforts, says the lab has been “very important” but adds that Indonesia's own lab capabilities are advancing rapidly. “We are encouraging developing country labs to become collaborating centers,” he says.

Jones declined to comment on the negotiations. But he emphasized that “we do not have plans to move to another Asian country. We are dedicated to a future here in Indonesia.”

6. PALEOECOLOGY

# Fossils Help Figure Out Food Webs Old and New

1. Elizabeth Pennisi

The watery world 540 million years ago abounded with exotic life forms rivaling those created by Dr. Seuss. Ecologists have long wondered how these worms, seaweeds, sponges, trilobites, snails, and meter-long beasts with rings of teeth interacted. Now, a daring analysis of fossils from China and Canada shows that these marine plants and animals from the Cambrian Period formed food webs on par with those existing today. These ancient networks follow the same rules seen among inhabitants of current reefs, deserts, and bays, report Jennifer Dunne, an ecologist at the Santa Fe Institute in New Mexico, and colleagues online 28 April in PLoS Biology.

“I was surprised that something like this [study] is really possible,” says Volkmar Wolters, an ecologist at Justus Liebig University in Giessen, Germany. “The result is a thought-provoking, highly informative, and breathtaking paper on the potential structure of Cambrian communities.”

Food webs are the complex networks of feeding interactions among the plants, animals, and microbes in a particular ecosystem. Over the past 30 years, researchers have learned that food webs share certain patterns, irrespective of the habitat or the particular species involved. The number of species making up the web and their degree of connectedness dictate certain elements—for example, the number of top predators, the percentage of omnivores, and so on. Dunne wondered how early these patterns emerged, but the fossil record had seemed too sketchy to provide detailed enough information about who eats whom.

Douglas Erwin, a paleontologist at the Smithsonian National Museum of Natural History in Washington, D.C., knew of two exceptions. After a lunchtime conversation with Dunne, he offered to help compile the necessary data from two Cambrian fossil beds, the Burgess Shale in Canada and the Chengjiang Shale in China. At both sites, soft as well as hard body parts were preserved. He and Rachel Wood of the University of Edinburgh, U.K., combed the literature for descriptions of gut contents, bite marks, teeth, and other structures indicative of particular eating habits. They established links among herbivores, predators, and prey for 142 plants and animals in the Burgess Shale and 85 in the Chengjiang formation and rated how confident they were of each link.

Dunne and her colleagues Richard Williams of Microsoft Research Limited in Cambridge, U.K., and Neo Martinez of the Pacific Ecoinformatics and Computational Ecology Lab in Berkeley, California, used these data to compare the two Cambrian ecosystems to eight modern ones, including a reef, an island, and a pond. They calculated the total number of links for each species in each ecosystem and the nature of those connections, for instance, how many were predator-prey, or herbivore-plant. They plugged the number of species and links into a mathematical model that describes modern feeding systems to see if it could accurately predict what the Cambrian food webs would look like.

It showed that the food-web structure—the numbers of organisms at each level, for example—was quite similar in all ecosystems studied. And it also showed that in some ways, the Chengjiang food web—15 million years older than Burgess—was more primitive. The researchers found more “loops” in the Chengjiang web, wherein the same species appears twice in a particular food chain. By contrast, the Burgess Shale and modern food webs tend to be more hierarchical, a trait considered important for stability, Dunne notes. Another analysis revealed that the Chengjiang food web was more loosely connected than the rest. Today, any species in a web is so closely connected to others that a change in one tends to affect most of the web members. In China, that may have not been the case.

But 15 million years made a difference. “The younger Burgess Shale web looks incredibly like the modern food webs,” Dunne points out.

The work is “an excellent contribution to both paleoecology and food-web theory, showing the relevance of the fossil record to understanding current ecosystem states,” says Peter Roopnarine, a paleontologist at the California Academy of Sciences in San Francisco.

However, he and Wolters wonder whether Dunne and colleagues were overly bold in assuming fossils at a particular locale really coexisted, as the beds cover millions of years, and whether the sampling was comprehensive enough for this sort of analysis. Just because the structure of the food webs seems similar doesn't mean they functioned the same way, cautions Roopnarine, who says that the paper “falls short on some of its claims.” Nonetheless, he thinks the work will have an impact: “The questions emerging from this paper should encourage paleontologists to think more seriously about the need to develop theoretical and modeling approaches to fossil ecologies.”

7. MEDICINE

# A Bruising Battle Over Lung Scans

1. Eliot Marshall

Doctors and researchers are sharply divided over the merits of screening smokers and others at high risk of lung cancer with costly computed tomography scans; a $200 million clinical trial has become a lightning rod. Doctors and researchers are sharply divided over the merits of screening smokers and others at high risk of lung cancer with costly CT scans; a$200 million clinical trial has become a lightning rod

Sheila Ross is known as a “two-time survivor” at the advocacy group where she works. Doctors found cancer in her lung, treated it, and declared the treatment a success—twice. “I am very lucky” to be alive, Ross acknowledges, because the odds of surviving 5 years after one such diagnosis are only about 15%. She lost a lung in the second treatment, and at that point, after 20 years as a congressional aide, she says, “I knew exactly what I had to do.” She joined a Washington, D.C., activist group, the Lung Cancer Alliance. Its style, like Ross's, is impatient.

The alliance kicked a hornets' nest last year, criticizing the cancer-research establishment for passivity and demanding that the U.S. government investigate two clinical leaders for alleged financial conflicts (see sidebar, p. 602). That was merely a side skirmish in a battle over early detection of lung cancer, a fight that has split the cancer-research and treatment communities. At issue is whether an advanced x-ray imaging technique known as low-dose spiral computed tomography (CT) scanning should be widely used to catch tumors when they're small, in hope of removing them before they spread. The alliance, along with an assertive group of scientists, argues that CT's effectiveness is already established and that it should be widely used to screen those who have a high risk of lung cancer, primarily smokers and former smokers.

An equally impassioned group of scientists argues that without better information, such a step would be folly. They say the benefits of CT screening for lung cancer aren't proven and that its widespread use could do more harm than good. It would be expensive, too, adding billions of dollars to the annual cost of Medicare and private insurance. Some argue that the government would get more reliable results by investing in tough measures to curb smoking, which is the biggest cause of lung cancer. “Essentially no one believes the current data support the hypothesis that screening is beneficial,” says outspoken skeptic Peter Bach, a lung-cancer specialist at the Memorial Sloan-Kettering Cancer Center in New York City.

Caught in the middle of this dispute is the National Cancer Institute (NCI). It is investing in a huge clinical trial, the largest cancer screening test it has ever run, in the hope of getting some guidance. The first data are due by 2010, although it may take longer to sift out key results. The trial itself has come under fire from the alliance, which wants to cut short what it considers an agonizing wait for action. The advocacy group has appealed to the U.S. Congress and government agencies to bypass the academic debate and move straight to screening. On the other side, the skeptics say that although the trial may not be perfect, it must be completed before a decision on screening is made.

A lot rides on the outcome. Lung cancer kills more than 160,000 people a year in the United States, and if more tumors can be caught early, many lives could be saved. But expanding CT screening would subject a large swath of the public each year to ionizing radiation, biopsies, and surgery; it would cause an unknown number of deaths itself. Every smoker in the United States is a prime candidate for lung imaging, and there are an estimated 45 million of them.

## Catching small tumors

Even people who disagree on screening agree that CT images are a vast improvement over the chest x-ray. Engineers and computational wizards came together in the mid-1990s to create machines that could probe large areas in the body noninvasively and with precision. In a CT scanner, the x-ray source and detectors spin in a ring around the patient, who slides past the moving beam to yield a spiral “slice” through the body. A CT scan can be quick—5 seconds— and uses less radiation than traditional methods. Speed also makes it possible to get an image in a single breath, reducing blur from motion. And unlike an old chest x-ray, the data can be tweaked by computer to tune certain densities in or out and reconstruct three-dimensional views of tissue and bone.

Starting in 1998, the race to upgrade imaging machines “really took off,” says Peter Kingma, U.S. vice president for CT imaging at Siemens, the German manufacturer. During a period known as the “slice wars,” companies increased the number of detectors rapidly; the capability to produce two simultaneous helical slices rose to 16 per scan when the big NCI screening trial began in 2002 and has continued climbing to 256 slices per scan and higher. Resolution also jumped from 10 millimeters in the early days to 0.625 mm; Kingma says low-dose images now resolve an area about 0.3 mm square.

Perhaps the most controversial pioneers in CT scanning for diagnosis of lung cancer are Claudia Henschke and David Yankelevitz, both of the Weill Cornell Medical College in New York City. In a landmark paper in the 26 October 2006 issue of The New England Journal of Medicine (NEJM), Henschke, Yankelevitz, and a group of international partners—known as the International Early Lung Cancer Action Project (I-ELCAP)—reported a stunning success: 88% of the early stage lung-cancer patients they detected lived at least a decade after diagnosis.

More than 40 participating centers in I-ELCAP adopted a method first developed on a small scale by Henschke and Yankelevitz in New York. They recruited subjects at somewhat elevated risk of lung cancer and gave them all CT scans. In I-ELCAP, those without tumors were asked to return for a scan in a year. The rest went on to a series of branching paths: If imaging found a specific solid nodule type with a diameter of at least 5 mm, or a nonsolid one of 8 mm diameter, they went under intensive study. Some got a biopsy and then possibly surgery or other treatment. Others, if their tumor looked less threatening based on texture or growth rate, were asked to come back in 3 months. The aim in the I-ELCAP studies is to continuously sift the population, identify fast-growing cancers, and remove them.

The group led by Henschke claims that with experience it has grown better and better at distinguishing slow-growing from fastgrowing tumors. This minimizes unnecessary biopsies, the clinicians say, but enables them to intervene early when a tumor changes. In the NEJM article, IELCAP reported that of the 31,567 people age 40 or older who initially enrolled in the study, 4186 (13%) were flagged by baseline CT scans as positive and required follow-up. But only 484 eventually received a diagnosis of lung cancer; 412 of these had the earliest form, stage I. It was in this latter group that the estimated 10-year survival rate was 88%. The authors say that only by providing good follow-up care and tracking people for more than 5 years can you see the full benefits of screening. Some studies that report no benefit didn't do aggressive follow-up or observe patients long enough, they argue.

The results had an impact in a field desperate for good news. “For the first time in my professional career, here is something that is offering to mitigate the largest cause of cancer [mortality],” says David Burns, a lung-cancer expert and professor of family medicine at the University of California, San Diego, who has a long career battling tobacco and defending federal antismoking measures. Now he is defending CT screening. “The logic is difficult to argue with,” he says: “You can clearly find [lung cancer] smaller, and you can clearly find it at an earlier stage. If you don't treat it, people die.” So why not try to make CT screening work?

## Doubts

Hope and enthusiasm are important, but “science has to be cold-eyed,” says Bach of Sloan- Kettering. He thinks the I-ELCAP screening study is riddled with flaws. He's also upset by the publicity it's received, which he worries may have increased the number of people seeking CT scans. The procedure is “incredibly lucrative,” he notes. The charge for an initial CT scan may be modest—about $200 to$300—but that's just the first installment. An anomaly “gins up all kinds of business,” says Bach, as clinicians follow up with positron emission tomography scans, biopsies, and other tests. Imaging is a gateway into high-cost medicine and has been flagged as a growing budget concern by the U.S. Medicare program.

The glaring weakness of I-ELCAP, according to Bach, who with epidemiologist Colin Begg of Sloan-Kettering and others published a study on lung screening last year in The Journal of the American Medical Association, is that it is not a randomized controlled trial. This makes it susceptible to bias. The best way to avoid bias in a screening trial, they argue, is to randomly assign patients to receive a CT scan or no CT scan and then keep track of who dies.

Without random selection, trial results can be dramatically skewed, for example, by “lead-time bias.” It produces the illusion that early diagnosis is responsible for extending the life of a patient when in fact the patient has just received a diagnosis earlier.

Other common problems, called “length bias” and “overdiagnosis,” arise from the imprecision of cancer biology. Too little is known about early stage tumors to predict which will become malignant; intensive screening can flag many that are benign or slow-growing as dangerous when they really are not. “Pseudodisease” is the term used by William Black, a radiologist and lung-cancer specialist at Dartmouth Medical School, to describe this byproduct of screening. He and his Dartmouth colleague, clinical epidemiologist H. Gilbert Welch, argue that this is a big medical risk that clinicians need to guard against. In addition to causing harm, overdiagnosis can boost the number of people who are diagnosed with cancer and appear to overcome it. “Everyone should know that when you go down this road [of cancer screening],” says Welch, “there is going to be harm; the question is, what will the benefit be.”

The study by Bach and others—a comparison of a validated model of clinical experience with data on 3246 patients from three CT screening trials—found “no evidence” that screening reduced the risk of death from lung cancer in a period of almost 5 years. But screening dramatically boosted medical workups. The authors found that biopsies increased three-fold above the expected level; lung surgeries, 10-fold.

Henschke and Yankelevitz claim that the extreme vigilance built into their approach keeps over-diagnosis and other biases to a minimum; clinicians intervene if “a malignant rate” of growth is evident. In addition, she and Yankelevitz write in the January 2008 issue of The Oncologist, a panel of pathology experts has examined all specimens removed by surgery and “confirmed that they are all genuine lung cancers and that 95% of them are already invasive.”

None of this satisfies the skeptics. Bach's doubts have grown so that he now says: “We worry that the basic principle [of CT screening] is wrong. … Most of the lung cancers that are claiming lives, we think, are coming like a meteor. They come out of nowhere and are everywhere.” Screening can't catch them. Yet others argue that Bach has gone overboard. Says James Mulshine, a leader of the Lung Cancer Alliance and associate provost for research at the Rush University Medical Center in Chicago, Illinois: “I haven't seen evidence in the literature that supports” Bach's view of meteorlike cancers.

Bruce Chabner, editor-in-chief of The Oncologist and clinical director of the Massachusetts General Hospital Cancer Center in Boston, says he's planning to air new concerns that go beyond study design in an editorial about the I-ELCAP results. For example, he claims that, unlike all clinical trials sponsored by drug companies and NCI, this privately funded project has not submitted its data to an outside audit. The Weill researchers did not respond to a request for comment.

NCI's proposed answer to the confusion is to look for help from a $200 million project it is now funding, the National Lung Screening Trial (NLST), a randomized controlled study. From 2002 to 2004, it enrolled and screened more than 50,000 individuals through a network of more than 30 study sites in the United States. The volunteers, all with an elevated risk for lung cancer, were randomly assigned to receive a chest x-ray or CT scan. Individual centers have been following up with standard monitoring and therapy. From 2008 on, researchers will be adding up deaths until they detect a statistically valid result showing that more people died in the x-ray group or the CT group—or neither. By 2010, the first results should be available from NLST. But CT screening advocates have already been taking shots at it. For example, some suggest that it was a mistake—perhaps unethical—to recruit people with the promise of high-quality diagnosis and then give chest x-rays, long viewed as a poor diagnostic tool. Henschke and Yankelevitz stopped using chest x-rays early in their study because, as they wrote in The Oncologist, it “missed” 76% of the screening-diagnosed cancers found by CT. The Lung Cancer Alliance also questioned whether patient follow-up was aggressive enough throughout the NLST network, because a slow response could make the diagnostic method look poor. NCI Director John Niederhuber responded in a letter last year that treatment “is not standardized in the NLST.” But he argued that this should not compromise the trial because “variations in treatment should occur equally in both arms.” According to Laurie Fenton Ambrose, president of the Lung Cancer Alliance, the emphasis on counting deaths rather than aggressively screening and treating patients is akin to “doing nothing” and is “just not acceptable.” Last year, Ambrose and other leaders of the pro-screening movement appealed to NCI for an interim view of CT screening, before NLST is done. They proposed combining data from I-ELCAP with data from NCI-funded trials, including NLST and another known as PLCO, in an attempt to get an early sense of the potential value of CT screening. Niederhuber met with the petitioners but decided it would not be “appropriate or fiscally responsible” for NCI to hold a review, he wrote to Ambrose. Otis Brawley, chief medical officer of the American Cancer Society (ACS), has agreed to serve as a broker. He is not an advocate of trying to get an early view of CT screening's benefits. (The idea was proposed by an epidemiologist at ACS, Robert Smith.) But Brawley says that he intends to host a meeting of experts on the topic; NCI and major international cancer institutions will be invited to participate. Brawley aims to bring investigators together in May or early June from four randomized trials, including three from Europe, and “perhaps” someone to represent the I-ELCAP study. It will be a kind of “grand jury,” he says, to review the trials and see whether it would be possible to use existing data to conduct a meta-analysis of CT screening. This grand jury may not lead to a new course of action, but it could help bring some calm to a hotly contested field of clinical research. 8. MEDICINE # A Bumper Crop of Conflicts 1. Eliot Marshall Some of the most contentious questions in the clash between clinical researchers over whether former smokers and others at high risk for lung cancer should be screened using computed tomography scans (see main text) have been about intellectual and financial conflicts. A clash between clinical researchers over whether former smokers and others at high risk for lung cancer should be screened using computed tomography (CT) scans (see main text) has turned bitterly personal. Some of the most contentious questions have been about intellectual and financial conflicts. The Lung Cancer Alliance (LCA), a patient advocacy group in Washington, D.C., cast the first allegation. LCA's president, Laurie Fenton Ambrose, is irate that the U.S. government has refused to endorse CT imaging for lung-cancer screening while it awaits results from a$200 million trial to evaluate the procedure, the National Lung Screening Trial (NLST). She has charged that some leaders of the trial revealed their bias against CT screening when they agreed in past years to testify for tobacco companies about how screening might do more harm than good. Ambrose and an ally at another advocacy group leveled these charges in a blitz of correspondence to federal agencies, targeting two distinguished NLST leaders—radiologists William Black of Dartmouth Medical School and Denise Aberle of the University of California, Los Angeles.

The letters prompted several inquiries, including one in the U.S. House of Representatives last fall led by Michigan Democrat John Dingell.* Dingell's probe, which made headlines and then faded from view, is “active and ongoing,” committee staff claim.

Black and Aberle have acknowledged that they agreed to testify for tobacco companies but said they did nothing improper. Aberle, who coordinates a large network of NLST clinical centers, provided testimony in 2003 in a class-action trial in Louisiana. As Aberle explained in a letter to the National Cancer Institute, she “violated no conflict of interest disclosure requirements,” and the checks she received—reportedly totaling about $30,000—went to her university. In the letter, she said she wanted to “articulate the uncertainties of CT screening and the potential risk” to people in Louisiana who might sign up for it. Black similarly agreed to provide testimony in 2006 for attorneys defending Philip Morris in a New York classaction suit by smokers who wanted the company to pay for their annual CT scans. As Dartmouth's general counsel explained in a letter to Ambrose, Black believes that widespread screening may “cause more harm than benefit” and prepared testimony about why it would be a mistake for the court to set a precedent for screening. But he changed his mind, withdrew, and returned a$700 payment because he realized his participation “might be misconstrued as support for the tobacco industry.” A review by the National Institutes of Health found that neither grantee had violated rules on disclosing conflicts of interest.

As the dust settled on this controversy, The Cancer Letter, a Washington, D.C., weekly, published an exposé of potential conflicts on the other side of the debate. It revealed that two well-known researchers who claim unprecedented success with CT screening for lung cancer—Claudia Henschke and David Yankelevitz, both of the Weill Cornell Medical College in New York City—have a financial stake in an invention that could be used in connection with CT screening. They have applied for 27 patents related to lung screening and have accepted royalty income from one license, but, The Cancer Letter charged, they did not properly disclose these interests in medical journal articles. In addition, The New York Times and The Cancer Letter reported in coordinated articles that most of the funds supporting the Weill project came from a tobacco company gift of $3.6 million. Henschke and Yankelevitz have since acknowledged that their widely cited 2006 article in The New England Journal of Medicine, for one, should have disclosed that they received royalties from their patented “methods to assess tumor growth and regression in imaging tests”—inventions that have been licensed to General Electric (GE), a maker of CT machines. In addition, they acknowledged that “virtually all” of the money from a foundation listed as a sponsor of their research actually came from an “unrestricted gift by the Vector Group, the parent company of Liggett Tobacco, which manufactures cigarettes.” In a separate statement, Weill says that Vector's original pledge was disclosed and reported in the national press 5 years ago and should be viewed in the same light as funding that “peer institutions and medical schools” received from antitobacco lawsuits. Even the group that first raised these questions may have a conflict of its own. Ambrose acknowledges that LCA, a tireless advocate for government action to expand CT imaging, has received funding from GE. Ambrose says the alliance always made known that it receives 40% of its funding from “corporate interests,” including the unrestricted GE grant and a larger one from a biotech company involved in lung-cancer research. 9. NEIL TUROK PROFILE # Wishing for an African Einstein 1. Daniel Clery* 1. With reporting by Robert Koenig. Hoping to nurture Africa's talent into becoming elite scientists, cosmologist Neil Turok has plans to clone the math institute he founded. Hoping to nurture Africa's talent into becoming elite scientists, cosmologist Neil Turok has plans to clone the math institute he founded CAMBRIDGE, U.K.—In 2001, mathematical physicist Neil Turok went back for the first time in 25 years to his childhood home, South Africa, to visit his parents. Dismayed by the lack of opportunities for math graduates in Africa and motivated by his father, a former antiapartheid activist, the University of Cambridge researcher took action. Over the next 2 years, Turok had a derelict building near Cape Town renovated into a new institute, enrolled 29 math graduates from 11 African nations, and persuaded mathematician colleagues to teach there for 3-week shifts. “It's a very inspirational venture, … a real flagship project,” says Britain's Astronomer Royal Martin Rees, who has visited the institute. The African Institute for Mathematical Sciences (AIMS) continues to grow (see sidebar, p. 605), but Turok isn't stopping there. He's leading the effort to replicate it at 15 centers across the continent, each focusing on a different area of applied math, such as economics. “When people hear about AIMS, they get very excited,” Turok says, “and people see the spark of something much, much bigger.” Turok's dream has gotten a massive boost from TED, a conference held every year in California, with about 50 invited speakers—originally just from the fields of technology, entertainment, and design, hence TED. Three of the annual speakers receive a$100,000 prize, and their talk must contain “one wish to change the world.” Last year, biologist E. O. Wilson wished for an online encyclopedia of life; it debuted a few months ago. This year, Turok was one of the chosen.

A prize winner's talk at the TED conference is essentially a sales pitch to philanthropists and companies; Wilson raised $50 million after his plea. Turok says he knew instantly what his wish would be: for the next Einstein to be African. Worrying that his Cambridge colleagues would ridicule such a notion, Turok says he tried it out on a few particularly hard-nosed individuals and didn't get one negative comment. Something is needed to jump-start African science, most agree. There have been many well-intentioned efforts to boost the continent's universities, but outside South Africa, world-class research still struggles to find a home and few pursue science professionally (Science, 9 February 2007, p. 748). “Africa has 922 million inhabitants, more than 40% of whom are under age 14. There may well be an Einstein here, but we must find better ways to tap into that vast pool of young talent,” says Sudanese mathematician Mohamed Hassan, president of the African Academy of Sciences and executive director of TWAS, the Academy of Sciences for the Developing World. Rees agrees: “It's important for Africa to develop its own intellectual elite and develop the conditions to keep them there.” Turok started off in Africa. He was born in 1958 in Johannesburg, where his parents, members of the African National Congress, fought apartheid. They were imprisoned and finally expelled from South Africa; Turok spent much of his childhood in Tanzania and Kenya before settling in London for high school. At Cambridge University, he says he “fell for theoretical physics.” After completing his doctorate, he worked at two U.S. universities and the Fermi National Accelerator Laboratory, returning to Cambridge in 1997. During this time, he carved a name for himself devising observational tests for current cosmological theories. In recent years, he's developed, in collaboration with Princeton University theorist Paul Steinhardt, a model for a cyclic universe in which the big bang is the result of two multidimensional membranes, known as brane worlds, colliding. In this model, the big bang is just the start of the latest phase of a universe that explodes, expands, contracts, implodes, and then starts all over again. Turok's own life split into two parallel universes during that 2001 visit to South Africa. He calculated that there must be thousands of math graduates emerging from African universities every year. “Most are not able to find work and are frustrated because they can't do the interesting stuff,” he says. Turok's father, who was by then a member of South Africa's parliament, provided “a huge stimulus” for AIMS. He refused to let his son sit down and watch an important rugby match until he'd put his idea for a mathematics institute down on paper; the father then promptly faxed the document to his political friends. “He forced me into it by pure embarrassment,” recalls Turok. The project won support from the universities of Cambridge, Oxford, and Paris-Sud XI, and the three universities of the Western Cape. The South African government and charitable foundations also pitched in; educating a graduate at AIMS costs roughly a fifth as much as sending them to Europe or the United States. Turok's experience with AIMS so far has confirmed his belief that the talent is there in Africa. It just needs a way to break out, he says. Most African math or science graduates would be overwhelmed, contends Turok, if they went straight to a European or North American research university. Having a postgrad center in Africa, he says, lets students “gain confidence. It changes their whole perspective.” By 2005, African politicians had taken notice of the AIMS model, and a plan was hatched to replicate it in universities across Africa. Under the auspices of the African Ministerial Council on Science and Technology, a group that included Turok began visiting math departments keen to become part of an African Mathematical Institutes Network (AMI-Net). Turok is convinced that rather than adapting existing structures, it would be better to create something new in each place: a separate institute, within the university infrastructure but autonomous, to insulate it from any political pressure. “There's a better chance of success if it's managed as a franchise,” he says. Not everyone agrees that starting from scratch makes financial sense. Although eager to replicate AIMS, Philippe Mawoko, a mathematician from the Democratic Republic of the Congo who works for the African Union's main development arm on science and technology initiatives, says it would be sensible “to use existing institutions as much as possible. Building completely new institutes would require more funding and legal work.” Since giving his TED talk in February, Turok says there has been “an amazing reaction.” He says he's had discussions with Google, the Bill and Melinda Gates Foundation, the U.K. government, accounting firm PricewaterhouseCoopers, and Barclays Bank, as well as endorsements from the likes of Richard Branson, Bob Geldof, and Forest Whitaker. With$2.3 million worth of donations committed, he's aiming to raise $150 million over the next 5 years. On 11 May, coinciding with the opening ceremony for a new research arm of AIMS, all parties involved in AMI-Net will gather in Cape Town to decide how to move forward. The first new institute, AIMS-Abuja, will open in July on the campus of the African University of Science and Technology, a World Bank project in Nigeria's capital. Turok says that once AMI-Net has enough money in place, it should quickly develop a model plan for how to set up new institutes and then embark on AIMS-3. “We have a huge responsibility to get it right. We have to make it work. It's the opportunity of a generation.” 10. NEIL TUROK PROFILE # An African Showcase for Math Studies 1. Robert Koenig The 53 students at the African Institute for Mathematical Sciences may speak an exotic mixture of languages, but they share a passion for numerical concepts and a determination to make a difference in their home countries. MUIZENBERG, SOUTH AFRICA—An exotic mixture of languages—Swahili, Amharic, Malagasy—echoes through the lobby of this former beach hotel that has been transformed into a hothouse for cultivating bright young African minds. But it is the universal language of mathematics that unites the diverse and ambitious group of students making a second home here. The 53 students at the African Institute for Mathematical Sciences (AIMS)—whether they come from the deserts of northern Africa, such as Esra Khaleel of the conflict-torn Darfur region, or from the lush southern island of Madagascar, like Mika Sidonie Ranaivomanana—share a passion for numerical concepts and a determination to make a difference in their home countries. “While I am at AIMS, my goal is to understand the difficult concepts,” says Proscovia Namayanja of Uganda. “I want to return to Uganda to teach those concepts.” AIMS's director, theoretical physicist Fritz Hahne, former dean of science at the University of Stellenbosch in South Africa, describes his charges as “creative and committed students” who “are here because they want to give something to Africa and to their home countries.” AIMS alumni surveys indicate that the vast majority of graduates go on to study for advanced degrees, mostly at South African universities but also in Europe and North America. It is unclear exactly how many return to their native lands, although every one of the dozen current students interviewed by Science said that they eventually would. Walter Mudzimbabwe of Zimbabwe, for example, plans to become an expert in a field that might benefit his hyperinflation-plagued homeland: financial mathematics. On the steps outside the AIMS building, Lydia Flore Mamoade—the first woman to attain a mathematics degree in her country—discusses her plans to earn a higher degree and then return home to the isolated Central African Republic. AIMS accepts only about one in five applicants. Students, who are given free room and board, take a series of intensive 3-week courses from visiting lecturers, who live in the building and make themselves available day and night. Although some courses focus on “pure” math or physics, most are in the problem-solving realm of what Hahne calls “relevant” mathematics—for example, related to bioinformatics, finance, or astronomy. When physicist Robert de Mello Koch of the University of the Witwatersrand in Johannesburg teaches electromagnetism, he avoids textbook tutorials, instead assigning challenging problems and projects such as building electroscopes out of soda cans. Understanding “the magic of AIMS” is only possible if you spend time with students, he says, “to see how hard they engage with the material, how far they manage to go, and how much it changes them.” Students appreciate their constant access to lecturers, a far cry from most of their university experiences. “The openness of lecturers, the nature of the material, the language—AIMS is completely different,” says Khaleel. This April, having completed their classes, the students were all working on final essays: Audry Ayivor of Ghana tackled topology in the library while Namayanja sipped tea as she explored a bioinformatics problem. Even though AIMS is at a beach resort near Cape Town, this year's class has no South African students—and previous classes had only a handful of them—in part because the nation's talented math graduates are quickly hired by industries to fulfill diversity goals. In an effort to attract more South Africans, AIMS plans to start a separate Honors program in biological mathematics next year. This month, Hahne says, AIMS is addressing another shortcoming by opening a new Research Center in two renovated buildings across the street. Visiting scientists, with joint appointments at AIMS and other universities, will try to create a synergy with the institute by enlisting students to help with research projects. Although the goal of developing an Albert Einstein in Africa may not happen in their lifetimes (see main text), AIMS students don't discount the possibility. “There is a mountain of talent on this continent, but young people need opportunities to excel,” says David Unuigbe of Nigeria, who wants to pursue solid-state physics at the University of Cape Town. And Ethiopian student Amasalework Ayele Ejigu believes AIMS will help students find such chances, ultimately benefiting the whole continent: “We are finding an African unity through mathematics.” 11. GENE THERAPY # Two Teams Report Progress in Reversing Loss of Sight 1. Jocelyn Kaiser The first safety trials of gene therapy for a degenerative eye disease produced good results in adults; researchers now intend to treat children. The first safety trials of gene therapy for a degenerative eye disease produced good results in adults; researchers now intend to treat children At a vision research meeting this week in Florida, scientists made a stunning announcement: In two independent studies, gene therapy has partially restored the sight of four young adults born with severe blindness. After receiving a single injection of a solution containing a curative gene months ago, the patients can see more light. Some of them can now read several lines of an eye chart. And two who had previously stumbled through an obstacle course can now navigate through it. The patients are still legally blind; they cannot read even a large-print book. Still, researchers are calling this first-ever test of gene therapy to treat a retinal disease a major advance. (The results were published online this week by the New England Journal of Medicine.) The patients' eyesight wasn't expected to improve much, if at all, because the disease was treated at an advanced stage. However, the two research teams—from the University of Pennsylvania (Penn) and Children's Hospital of Philadelphia (CHOP), and University College London—say the benefits may be much greater when they treat young children with this disorder, called Leber congenital amaurosis 2 (LCA2). Gene therapy researchers hailed the news as a boost for their field. “It's a very promising development,” says Arthur Nienhuis of St. Jude Children's Research Hospital in Memphis, Tennessee, president of the American Society for Gene Therapy. The studies were only designed to test safety. But if the hint of efficacy holds up in larger trials, the results could pave the way for using gene therapy for other inherited eye diseases as well as for common ones such as macular degeneration, eye researchers say. “It is a marvelous thing for the field and for the future,” says Paul Sieving, director of the U.S. National Eye Institute (NEI) in Bethesda, Maryland. Not everyone who helped develop this treatment is sharing in the celebration. A third group that was part of a team that pioneered the work in dogs 7 years ago has been sidelined. Although the group's members at Penn and the University of Florida have results from an adult clinical trial, they have been forbidden by NEI, which funded the work, to discuss their findings. According to Sieving, the agency is waiting for an examination of safety and efficacy data to be completed. ## Untreatable The eye has long been an attractive target for gene therapy because the test gene is expected to be confined within the eye and because improved sight can be measured with precision. Interest grew in the 1990s, as dozens of genes for inherited retinal diseases were identified. They include LCA, a catchall term for severe blindness disorders affecting about 3000 Americans that lead to atrophy of the light-sensing photoreceptor cells in the retina. Children with LCA begin losing their sight at birth or soon thereafter; by age 40, they are completely blind. There is no treatment. One form of LCA (LCA2) is caused by a defect in retinal pigment epithelium 65 (RPE65), a gene that is critical for the single layer of cells that line the back of the eye and nourish the photoreceptor cells. RPE65 codes for an enzyme that helps convert vitamin A into a form used to make rhodopsin, a pigment that absorbs light. Without rhodopsin, the photoreceptor cells eventually die. But this can take decades, so giving younger patients a working copy of RPE65 should restore photoreceptor function. The cells “just need to be resupplied [with pigment] and they're back in action again,” says Michael Redmond of the NEI intramural program, whose lab discovered RPE65 and studied it in mice. In 1998, gene therapists got a lucky break when researchers found that some Briard dogs carry the RPE65 mutation. A consortium formed to study gene therapy in these dogs. It included Cornell University veterinary researchers; scientists at the University of Florida, who developed the adeno-associated viral vector; and several Penn scientists. The Penn group included Samuel Jacobson, a clinician who runs a center for inherited retinal diseases, and two younger researchers: Albert Maguire, a retinal surgeon, and his wife, Jean Bennett, a molecular geneticist, who coordinated the studies. In 2001, the consortium reported in Nature Medicine that treating three young dogs with RPE65 mutations restored enough sight to allow them to avoid objects and walk through a maze. Their most famous subject, named Lancelot, became a poster dog, visiting members of the U.S. Congress to promote gene therapy. In June 2005, Jacobson proposed a clinical trial for LCA2 in young adults to the National Institutes of Health's Recombinant DNA Advisory Committee (RAC), which reviews gene therapy trials. Bennett was not part of the proposed trial. According to Jacobson, she preferred to begin the trials in children. “There was a real difference in medical-ethical feeling,” comments Jacobson, who received clearance to go ahead with the adult trial. In the meantime, Katherine High, director of a new gene therapy center at CHOP, made Bennett and Maguire an offer: CHOP's center would fund a separate trial on children with Bennett and Maguire as leaders. High had raised money to start the center after a company sponsoring her own work on hemophilia dropped out. She decided gene therapy for LCA2 in children was a good trial to start with. “The sooner you can intervene, the younger the patients, the better,” High says. In December 2005, Maguire proposed a pediatric trial to the RAC. He suggested that beginning with children made sense because it wasn't clear that adults would benefit. “Performing a study in older adults with devastated retinas may not provide adequate information to proceed with further studies in a pediatric population,” Maguire said in an archived Webcast of the meeting. Jacobson, who was surprised by this presentation, notes that Maguire's appeal included drama: comments from a couple whose blind 1-year-old son with LCA had been seen by Maguire's team. The father wept at the microphone while his wife stood by with their baby. RAC approved the proposal. The two groups—the Bennett-Maguire team, working through CHOP with private funding, and Jacobson's team with an NEI grant—moved ahead separately, although they briefly considered using a joint data and safety monitoring board (DSMB). NEI decided this wouldn't work, partly because the CHOP trial was moving faster, says NEI's Maryann Redford. The CHOP team and Italian collaborators began recruiting patients from the clinic of Francesca Simonelli at the Second University of Naples in Italy. Jacobson's NEI-funded team had a setback: It had to wait a year while the NEI approved a funding agreement for their trial. Then early last year, a late entrant joined the race. Robin Ali's team at University College London, which had been working on gene therapy for other retinal diseases, wanted to launch an LCA2 trial because “we knew this would be a very good system to test efficacy,” Ali says. They decided to treat adults first, Ali says, even though they knew results would be more dramatic with younger patients. It was the first time a human retina had been injected with a gene therapy vector, he notes. “If it had been a catastrophe,” he says, “we would have been open to the criticism that we rushed into this with an 8-year-old.” The U.K. team treated its first patient in February 2007. The two U.S. teams did not begin until October. In the end, CHOP started with adults, too—a 19-year-old and two 26-year-olds—because recruiting younger patients proved difficult and because adults can have more viable retinal tissue than they had thought, says Maguire. The injection caused no apparent immune response; one patient in the CHOP trial developed a pinhole-size macular hole, possibly from the injection surgery, but it didn't affect vision. A few weeks later, a test based on pupil constriction showed that the patients could detect three times more light. Performance on eye charts was compelling, too, although such tests can be influenced by the placebo effect. Two patients who had been limited to detecting hand motions were able to read several lines of an eye chart. And a 26-year-old patient who initially could not navigate an obstacle course was able to make his way through. * “I couldn't believe it,” Bennett says. The London team, which used a weaker promoter sequence to drive gene expression, saw improvement in only one patient, an 18-year-old who had the most intact retinal tissue, Ali says. He was no better at reading an eye chart, but his light perception improved 100-fold. In a video before treatment, he stumbles through a dimly lit, simulated night street scene, bumping into walls several times in 77 seconds. Six months after the injection, he breezes through in 14 seconds. “It's more than we could have hoped for,” Ali says. Jacobson says the results from his NEI-funded study's first patient, treated in October, were also “very encouraging.” His team drafted a case report in December, but NEI told them to wait until data for their next two patients—like the CHOP team's patients, treated in December and January 2008—had been reviewed in May by the study's DSMB. NEI is proceeding with caution, says Redford. The institute, however, issued a press notice last week describing the Penn trial and noting that NEI has spent$124 million on basic and clinical research on RPE65.

The other two teams say they sympathize. “It's very frustrating for everybody because all of us want to hear about their data,” says Bennett. The competition between three groups has been healthy despite the disappointment for some, says eye disease geneticist Stephen Daiger of the University of Texas Health Science Center in Houston, who knows both U.S. groups. “It pushed everyone to work harder and faster in getting trials going. In the long run, that is great for all concerned.”

The U.K. and CHOP teams are moving on to younger patients; Bennett and Maguire this month will treat an 8-year-old. They are also looking ahead to their next gene-therapy trial. Bennett says she and her Italian collaborators are thinking about recessive Stargardt disease, a disorder that affects about 27,000 Americans and involves a defect in a gene in photoreceptor cells.

Now that gene therapy for an eye disease has shown signs of helping the first set of patients, Bennett says, “it opens up so many opportunities.”

12. AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS MEETING

# Tuberculosis Jumped From Humans to Cows, Not Vice Versa

1. Ann Gibbons

At the American Association of Physical Anthropologists meeting, a DNA study showed that early humans were infected with strains of Mycobacterium tuberculosis, which cause tuberculosis, long before they began herding cattle, suggesting that it was humans who transmitted the disease to bovids and other animals.

AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS, 9–12 APRIL 2008, COLUMBUS, OHIO

Humans have been cozy with their cows for almost 10,000 years—milking them, herding them, and even sleeping with them for warmth. Many researchers have thought that cows also gave our ancestors a less welcome gift: their first exposure to the mycobacteria that cause tuberculosis (TB), a disease that kills 2 million people a year.

At the meeting, a DNA study of 10 species of mycobacteria showed that early humans were infected with strains of Mycobacterium tuberculosis, which cause TB, long before they began herding cattle. That suggests that it was humans who transmitted the disease to bovids and other animals. “TB spread from humans to animals,” perhaps when modern humans emerged from Africa to spread around the globe, reported graduate student Luz-Andrea Pfister of Arizona State University in Tempe in her talk.

Pfister compared the complete sequence of the same 60 genes in 10 species of mycobacteria from humans and animals. She sorted the bacteria into evolutionary trees and dated the splits among strains by assuming a constant mutation rate. She calculated that a proto M. tuberculosis split from the mycobacterium that causes leprosy about 36 million years ago, perhaps in a primate ancestor of humans. She also estimates that M. tuberculosis in humans gave rise to M. bovis, the strain that infects cattle, about 113,000 years ago, and to the strains in several other mammals about 90,000 to 100,000 years ago. Those dates are based on an Escherichia coli mutation rate, because no rate has been calculated for mycobacteria. So the precise dates may change, says Pfister, who is working to calculate a more precise mycobacterial clock.

Pfister's data suggest that humans gave the disease to cows; they also challenge a long-held view that M. tuberculosis descended from a single bug that flourished 20,000 to 35,000 years ago. That view was recently thrown into doubt by work by Cristina Gutierrez, an evolutionary mycobacteriologist at the Institut Pasteur in Paris, and others who found unexpectedly diverse DNA sequences in African strains of M. tuberculosis. That raised the possibility that M. tuberculosis was more ancient, in order to have accumulated such diversity (ScienceNOW, 19 August 2005: sciencenow.sciencemag.org/cgi/content/full/2005/819/2). Pfister's findings, says Gutierrez, confirm “the emergence of variants of the M. tuberculosis complex long before the domestication of animals.”

The genetic work also fits with recent fossil data showing putative TB lesions of great antiquity. For example, last year researchers reported such lesions in a 500,000-year-old archaic human fossil from Turkey, although without ancient DNA it's difficult to be sure TB was the culprit. Gutierrez thinks the main question now is “how could the first transmission from hominid to cow occur so long before their domestication?” TB transmission usually requires prolonged, close contact—but not always, says evolutionary osteopathologist Bruce Rothschild of the University of Kansas, Lawrence. He points out that “primates in zoos can get TB from humans who spit on them.” That raises the specter of a wheezing human ancestor who spat out an insult—in the form of TB—on some hapless bovid long ago.

13. AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS MEETING

# Australopithecus Not Much of a Nutcracker

1. Ann Gibbons

Several researchers reported at the American Association of Physical Anthropologists meeting that different analytical methods suggest that robust australopithecines didn't eat hard nuts and seeds routinely as had been thought, and that robust and gracile hominids actually ate similar fare.

AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS, 9–12 APRIL 2008, COLUMBUS, OHIO

In 1959, famous South African paleoanthropologist Phillip Tobias dubbed a newly discovered skull from Olduvai Gorge in Tanzania “Nutcracker Man,” because of its huge cheek teeth, thick tooth enamel, and massive jaw. Ever since, researchers have thought that the teeth and jaws of these robust australopithecines were specialized to chomp on hard nuts and seeds, whereas their more fine-boned cousins, the gracile australopithecines who eventually gave rise to humans, were generalists who ate fibrous fruit and plants.

Now, several researchers reported at the meeting, different analytical methods suggest that the diet of robust australopithecines wasn't so hard after all, and that robust and gracile hominids ate similar fare. Even Nutcracker Man's species, now called Paranthropus boisei, wasn't crunching nuts and small hard objects routinely, according to a new analysis of its tooth wear. “It looks more like it was eating Jell-O,” says paleoanthropologist Peter Ungar of the University of Arkansas, Fayetteville.

These results “show clearly that simplistic ideas about australopithecine diets need to be revised,” says paleoanthropologist David Strait of the University at Albany in New York.

Ungar and his colleagues exposed the soft side of Nutcracker Man's diet by scanning the molars of seven individuals of P. boisei from East Africa with a confocal microscope and analyzing microscopic wear and tear. The teeth lacked the telltale pits that come from eating small, hard nuts and seeds, as seen in graycheeked mangabeys and brown capuchin monkeys, the team reported in PLoS One this week. In fact, the robust australopithecine's pattern of wear resembles that of the gracile hominids, including A. afarensis from East Africa. At the meeting, Ungar reported in a talk that no australopithecines show signs of eating hard objects routinely, although A. robustus from South Africa may have done so seasonally.

A separate line of evidence from isotopic studies is also blurring the distinct ion between the diets of robust and gracile australopithecines. Various plants absorb atmospheric carbon dioxide differently, and so by measuring the ratio of carbon isotopes in teeth, researchers can detect whether ancient hominins ate tropical grasses and sedges rich in carbon-13, or woody fruits, shrubs, and herbs with less carbon-13.

The isotopic data from the gracile australopithecines matches that of the robusts and suggests that both groups ate more diverse diets than expected, said paleoanthropologist Matt Sponheimer of the University of Colorado, Boulder, in his talk. His team analyzed the carbon isotopes in teeth from 20 robust and 25 gracile australopithecines from South Africa. Both groups clustered in between animals that browse on fruit and leafy plants, such as fossil giraffes and fruit-eating antelopes, and animals that graze on fibrous grasses, such as fossil wildebeest and zebra. Sponheimer concluded that the robust australopithecines' cuisine included fruits from trees, bushes, and shrubs when they could get them but also less desirable grasses and sedges, similar to the diets of some baboons.

So why did robust australopithecines evolve such massive jaws and molars? The best guess at the meetings was that the extreme jawbreaker anatomy may have been an adaptation to eat less desirable “fallback” foods in times of drought or seasonal stress. Gorillas, for example, have evolved sharp teeth to eat fibrous leaves when needed for survival, although they prefer and consume far more fruit and soft leaves. “The anatomy shows not what an animal eats but what it has the potential to eat,” says Ungar.

14. AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS MEETING

# Snapshots From the Meeting

1. Ann Gibbons,
2. Elizabeth Culotta

Snapshots from the American Association of Physical Anthropologists meeting include the evolution of gliding, the divergence of Homo habilis and H. erectus, and Neandertal speech.

AMERICAN ASSOCIATION OF PHYSICAL ANTHROPOLOGISTS, 9–12 APRIL 2008, COLUMBUS, OHIO

Evolution of gliding. In the tropical forests of Southeast Asia, more animals glide from tree to tree than anywhere else in the world: leaping lizards and geckos, 45 species of frogs, and snakes that can glide 100 meters. Researchers have proposed that more gliding species evolved there because Asia's taller trees give animals enough time to gain lift after they fall, and there are few vines to use to move between trees. At the meetings, a researcher who mapped the canopies of 11 forests around the world with laser equipment offered another explanation: rugosity, or vertical gaps between treetops of different heights.

Biological anthropologist Nathaniel Dominy of the University of California (UC), Santa Cruz, reported that Asian forests showed the most extreme vertical drops from treetop to treetop, especially in high canopies of forests where trees grow more than 50 meters tall. Falling can be fatal, Dominy says, so natural selection would strongly favor animals with the anatomy to recover by gliding. As a result, says UC Berkeley physiologist Robert Dudley, Asia has many gliders whereas African and American forests have only a few. The steep plunges may even have shaped Asian apes: Dominy proposed in his talk that gibbons have developed their unique mode of ricochet brachiation (rapid, arm-over-arm swinging) to navigate the vertical drops. “There's lots of monkeying around in tall forests,” Dudley says.

Side by side. One of the hottest debates in paleoanthropology concerns the petite Homo habilis. Did it give rise to the long-legged human ancestor H. erectus about 1.8 million years ago in Africa? Or were the two species contemporaries who arose from an even earlier ancestor? At the meetings, paleoanthropologist Christopher Ruff of Johns Hopkins University in Baltimore, Maryland, offered the first analysis of the interior structure of arm and leg bones from H. habilis—and concluded that it and H. erectus moved in different circles.

By examining how bone density varies in cross sections of upper arm and thigh bones, researchers can see how mechanical loads are distributed on a hominin's limbs as it walks or climbs trees, for example. Ruff studied images of the cross sections from the arm and leg bones of H. habilis, H. erectus, chimpanzees, and 100 modern humans. His analysis is “convincing” that many features of H. habilis limbs are more primitive than those of H. erectus, says paleoanthropologist Henry McHenry of the University of California, Davis. Ruff suggests that H. habilis spent more time in the trees, supporting fossil evidence that the two species lived in different habitats. And that, in Ruff's view, makes it unlikely that H. habilis gave rise to H. erectus.

Neandertal voice-over. When Neandertals talked, what did they sound like? Robert McCarthy of Florida Atlantic University in Boca Raton played a brief synthesized version of a Neandertal voice, based on his reconstruction of their vocal tracts. The vowel sound of “e” in “see” was less precise than ours. It also sounds higher-pitched to our ears, although that's an artifact of this particular sound, McCarthy says.

McCarthy and Philip Lieberman of Brown University reconstructed the shape of the Neandertal vocal tract from fossils of the face and neck bones and inferences based on modern humans. The frequencies that emerged most clearly from their throats were different from ours. As a result, Neandertals would not have been able to form the precise, rapid-fire phonemes of human speech, the researchers found.

Perhaps the biggest surprise: A 100,000-year-old early modern human from Israel, known as Skhul 5, had the same vocal traits as Neandertals. “Full-blown modern speech came relatively late,” says McCarthy, perhaps between 100,000 and 40,000 years ago