News this Week

Science  20 Jun 1997:
Vol. 276, Issue 5320, pp. 1788

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution

  1. Causing Cancer by Remote Control?

    1. Trisha Gura
    1. Trisha Gura is a writer in Cleveland.

    Kaposi's sarcoma-associated herpesvirus may help drive the growth of the bone marrow cancer multiple myeloma, but from within neighboring cells, not the tumor cells themselves

    Add another trick to the sinister repertoires of viruses. For decades, researchers have been pursuing a trail of early clues suggesting that viruses might be culprits in human cancers. Only in a few malignancies have scientists been able to finger a viral suspect, however. Now, a group of researchers may have detected viral fingerprints on another cancer, a bone marrow tumor called multiple myeloma. But the group has found that in this case, the virus may contribute to tumor growth in a novel way, working behind the scenes like a cellular puppet master.

    On page 1851, a research team led by oncologists James Berenson and Matthew Rettig of the Veterans Affairs West Los Angeles Medical Center reports linking multiple myeloma to Kaposi's sarcoma-associated herpesvirus. KSHV is already under suspicion as the cause of Kaposi's sarcoma, a cancer that afflicts many AIDS patients. But unlike the malignant cells of Kaposi's and all the other cancers thought to be caused by viruses, multiple myeloma cells—derived from the bone marrow's antibody-producing plasma cells—don't seem to carry the virus. Instead, the Los Angeles team has found evidence that KSHV is lurking in adjacent dendritic cells, a subset of macrophages found in the bone marrow microenvironment.

    In those cells, the virus appears to crank out its own version of a human protein called interleukin-6 (Il-6) that is known to stimulate myeloma cell growth. This, the researchers propose, is what propels the runaway growth of myeloma tumors, or, as Berenson describes it, “The soil, the dendritic cell, is putting out a bunch of fertilizer, and that makes the seed, the tumor cell, germinate.”

    That kind of remote control, says Yuan Chang, who studies KSHV at Columbia University in New York City, is a “novel mechanism” for virally caused cancer: “This is really exciting, if [the authors] are right.” The finding could also steer researchers to new therapies for multiple myeloma, which strikes 13,000 people every year in the United States alone and usually kills its victims within 3 years. Drugs to block Il-6 might be one therapeutic avenue, and Berenson also suggests that it might be possible to devise therapies that specifically target the virus-infected dendritic cells themselves.

    Attacking the virus with drugs or a vaccine might also stave off full-blown multiple myeloma in the estimated 1 million people who have been diagnosed with an apparent precursor condition called monoclonal gammopathy of undetermined significance (MGUS). Others who might benefit are AIDS patients, especially homosexual men, who have a high risk of becoming infected with KSHV and getting Kaposi's sarcoma.

    Rettig and Berenson say that they were inspired to look for the virus in the nonmalignant bone marrow cells of multiple myeloma patients by the Chang team's discovery of a KSHV protein that closely resembles human Il-6, both structurally and functionally (Science, 6 December 1996, p. 1739). The Columbia workers also postulated that the viral Il-6 could feed tumor cell growth in Kaposi's sarcoma. At the time, however, there was no evidence that KSHV was involved in multiple myeloma. Further, when researchers, including Chang's group and that of Rettig and Berenson, looked for signs of KSHV in bone marrow samples from myeloma patients using the polymerase chain reaction (PCR)—a gene amplification technique that picks up very small amounts of specific DNAs—they detected no trace of KSHV DNA.

    Still, the finding of a viral Il-6 intrigued the Los Angeles researchers enough to entice them to look more specifically at the cells in the bone marrow stroma, a mix of cell types that provide support for developing plasma and other cells. “There was a large body of evidence saying that Il-6 was a very important and necessary growth factor for the persistence and propagation of myeloma,” Rettig says. “And we knew that Il-6 is, for the most part, produced in the stromal cells.”

    But because those cells are sticky and present in low numbers compared to neighboring malignant plasma cells, they are hard to get hold of. So Rettig and Berenson grew bone marrow samples under cell culture conditions that killed off the malignant cells but nurtured the nonmalignant stromal cells. The researchers then did PCR testing on both the uncultured bone marrow samples, made up of predominantly malignant cells, and on the stromal cell-enriched cultured cells.

    In the case of bone marrows obtained from 10 normal individuals and from 16 patients with blood cancers other than myeloma, both the uncultured and cultured cell samples came up negative. And as in the past, the researchers detected no KSHV DNA in uncultured bone marrow cells from all of the 15 myeloma patients tested. But all their stromal cells tested positive for the virus. “We have identified the virus consistently in 100% of myeloma patients,” says Berenson. “That percentage, in and of itself, is amazing.” Rettig and Berenson then went on to identify the virus's host cells as macrophage-derived dendritic cells, based on specific markers the cells carry on their surfaces.

    What's more, when the researchers looked at bone marrow samples from eight patients with MGUS, the myeloma precursor condition, they detected KSHV DNA in the dendritic cells from two. Because that is the same proportion—25%—as the fraction of people with MGUS who go on to develop myeloma itself, the observation raises the question of whether it's the virus that determines who will progress to the cancer—or something else that is yet unidentified.

    Another question Rettig, Berenson, and their colleagues have tackled is how KSHV might be stimulating tumor cell growth from adjacent cells. Their answer involves Il-6, whose messenger RNA was detected in infected dendritic cells from three out of three myeloma patients, but not in dendritic cells from two normal controls.

    But while researchers such as Chang are excited by the Los Angeles team's findings, they have reservations, partly because of past failures to detect KSHV in myeloma patients. Rettig doesn't see a conflict. “The bone marrow itself contains so few dendritic cells that the sensitivity of even a PCR assay was not adequate to detect any virus,” he says, adding that finding the virus required culturing the cells to enrich the dendritic population.

    Yet the culturing step itself raises concerns because what happens in culture may not reflect what is occurring in the cells of human beings, researchers say. Charles Rabkin, who studies the epidemiology of HIV-related cancers at the National Cancer Institute in Bethesda, Maryland, calls the new findings “extremely interesting,” but worries that the dendritic cells Berenson's group studied may have been contaminated with the virus after their removal from the patients or that the virus, although present, may not actually cause myeloma.

    Berenson and Rettig respond that contamination is very unlikely, as they did not detect KSHV in any of the samples from normal individuals or from patients with other cancers, even though all the samples were handled the same way. But even better evidence against contamination comes from as yet unpublished work in which their group looked for—and found—KSHV DNA in uncultured biopsies of bone marrow cells that were analyzed directly after removal from myeloma patients.

    Even with this evidence, however, an epidemiological conundrum posed by a link between KSHV and Kaposi's sarcoma needs to be sorted out. It is not clear whether groups prone to that cancer—Ashkenazi Jews and AIDS patients—have an increased risk of multiple myeloma, says Rabkin. That could mean that either the virus isn't involved in myeloma, or that an additional, as yet unidentified, factor is needed to tip a person over the edge so that the cancer can develop.

    In trying to resolve these issues, the researchers can look to the estimated 1 million people in the United States thought to have MGUS. They should make it possible to determine, for example, whether progression to myeloma correlates with signs of infection, such as having antibodies to the virus. Berenson says his group plans to test frozen blood samples from MGUS patients, some of whom have already developed myeloma.

    Other studies are likely to follow as well, Rabkin predicts. That is because proving a virus is the sole cause of cancer can be inherently difficult. The different bits of information often conflict. “But as far as KSHV is concerned,” Rabkin says, “this paper adds another piece to the puzzle … one that I think will be followed up.”


    How the Hectic Young Sun Cooked Up Stony Meteorites

    1. James Glanz

    WINSTON-SALEM, NORTH CAROLINA—Inspired by glimpses of the turmoil around young stars, a team of astrophysicists has presented a radically new theory of the solar system's most primitive—and perhaps most mysterious—solid objects. These meteorites, called chondrites, are thought to be shards of bodies like those that clumped together to form our planets. For more than a century, meteoriticists have puzzled over their bizarre composition—a stew of dust, roundish rocks that were “flash melted” and resolidified, and the remains of short-lived radioactive isotopes. The new theory holds that this stew was cooked up by explosive flares and powerful winds near the young sun.

    As Frank Shu of the University of California, Berkeley, explained here last week when he presented his team's results at a meeting of the American Astronomical Society, the two processes, in varying combinations, could explain all the ingredients of chondrites. Flares licking at the disk of gas and dust around the young sun could have irradiated the material with energetic particles, and the heat of the flares or the glare of the young sun could account for the melting. The winds might then have blown the molten material to the outer reaches of the solar system, where it mixed with dust to form chondrites. Simply put, the mechanism “takes material trying to get on the sun, heats it up, irradiates it, and plops it back onto the disk farther out,” says Donald Clayton of Clemson University in South Carolina, who organized the session at which Shu spoke.

    Some astrophysicists who have sought the origins of the isotopic anomalies outside the solar system—in radiation from a nearby supernova, for example—aren't convinced that Shu and colleagues have the full story. But others are embracing the scenario, among them Clayton, who says, “In my mind, his [explanation] is the leading contender right now.”

    The mystery dates from the 1870s, when researchers started cutting open meteorites. Inside some of them, says Glenn MacPherson of the department of mineral sciences at the Smithsonian Institution of Washington, was something that “looks a little bit like concrete.” These so-called carbonaceous chondrites consist of a dark, dusty matrix, which apparently never melted, sprinkled with once-molten rocks: centimeter-size pinkish-white ones called calcium-aluminum-rich inclusions (CAIs), and bluish-gray rocks measuring a millimeter or so, called chondrules.

    Strangely, the CAIs and chondrules in any particular chondrite are almost uniform in size, as if they had been sorted like peas in a factory. Even more surprising, their crystal structures show that the CAIs were molten for periods of days and the chondrules for just hours. The cool band of the protoplanetary disk now occupied by the asteroid belt, where chondrites are thought to have originated, seems an unlikely setting for such rapid melting and freezing.

    But the deepest mystery of all could be the residue of certain short-lived radioactive isotopes, found mainly in the CAIs. “There is evidence that these materials had live aluminum-26 in their crystalline structures when they formed,” says MacPherson. The aluminum-26 itself is long gone—it has a half-life of just a million years—but it left identifiable decay products, as did other slightly less mercurial isotopes such as manganese-53.

    Most theories have invoked separate processes to explain these physical and isotopic anomalies. A. G. W. Cameron of the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts, for example, has suggested that violent “x-ray flares,” like the ones now seen around young stars, might have flash-melted some material in the protoplanetary disk, while the strange isotopes might have come from a nearby supernova. The new theory, put together by Shu, Hsien Shang of Berkeley, Typhoon Lee of the Academia Sinica in Taiwan, and Alfred Glassgold of New York University, ties all the anomalies together in a single explanation. Parts of it appeared in Science a year ago (15 March 1996, p. 1545), but the group presented the full picture at the meeting.

    It had emerged as the Hubble Space Telescope, orbiting x-ray observatories, and other instruments revealed unexpected turmoil in the disks of material that surround young stars. The observations showed jets of gas and dust spewing into space, along with the flares that had inspired Cameron's scenario. The jets, Shu and others have suggested, indicate powerful winds, whipped up by the strong magnetic fields of young stars. Because a newborn star is likely to be spinning rapidly, Shu explains, its field lines would be whirling like the blades of an egg beater, flinging the ionized gases of the disk outward. The winds might blow some of the material entirely out of the disk, creating the jets.

    Other clumps of disk material, however, might be lofted out of the disk for just a few days, like dust balls caught by a slight breeze. Shu, Shang, Lee, and Glassgold calculated that at the radius of the disk's inner edge, that would be enough time for the heat of the sun to melt the dust before it settled back into the cooler disk and congealed. Flares, meanwhile, bombarded the molten material with particles, leaving it laden with radioactive isotopes. Later, at a more active point in the sun's magnetic cycle, much stronger winds might sweep up the CAIs and fling them far out into the disk, sorting them by size along the way—just as earthly winds carry dust farther than sand.

    As for the chondrules, Shu and his colleagues argue that they formed without ever leaving the disk when especially fierce flares melted these smaller clumps of material for a few hours. Later, they too rode the wind farther out into the disk.

    Shu points out that traces of ancient magnetism locked into the chondrules when they solidified suggest that they formed in the powerful magnetic field found close to a young star. The theory also received an unexpected boost during Clayton's session, when MacPherson reported that he, Kevin McKeegan of the University of California, Los Angeles, and others had found that CAIs and related inclusions all have a consistent ratio of oxygen isotopes. This, says MacPherson, suggests they all were forged in the same location in the early solar system, then “dispersed to the different regions of the asteroid belt where we find them.”

    Still, MacPherson and others, including Cameron and Gerald Wasserburg of the California Institute of Technology (Caltech), say it is far from clear that solar flares can explain the complex patterns of radioactive residues found in chondrites. “The sources of these isotopic anomalies are almost certainly from outside the solar system,” says Wasserburg. But others think the new model will overcome such objections. “It could be a synthesis that solves a decades-old problem,” says Eric Feigelson, an x-ray astronomer at Pennsylvania State University. “It smells right to me.”


    Is Warming Trend Harming Penguins?

    1. Jocelyn Kaiser

    One might think that penguins would be pretty resistant to shifts in their environment, living as they do in the coldest, iciest place on Earth. But scientists have found that even these tough birds have their sensitivities. A few good years for krill, penguins' main food, can push up penguin populations; the buzz of activity around research stations can nudge their numbers downward. Now ecologists are suggesting that the 4° to 5°C midwinter warming of the western Antarctic Peninsula climate observed over the last 5 decades is taking a toll on Adélie penguins.

    Ecologists had already proposed that a decrease in sea ice cover due to the warming might be responsible for a recent decline in penguin numbers. Now, ecologist William Fraser of Montana State University in Bozeman says he may have discovered a new way in which climate change could be affecting penguins. Fraser, who described his latest work last week at a seminar on Capitol Hill sponsored by the U.S. Global Change Research Program, argues that more snow on some islands near Palmer Station on the Antarctic peninsula may be making it harder for the birds to breed. Because a warmer atmosphere holds more moisture and could be causing heavier snowfall, Fraser thinks the Adélie decline could be a “canary in the coal mine”—a sign that warming is affecting Antarctic ecosystems.

    The warming may be just a natural fluctuation, not necessarily an early indicator of greenhouse warming. Still, other scientists find the idea intriguing. Gerald Kooyman of Scripps Institution of Oceanography in La Jolla, California, who studies Emperor penguins, cautions that Fraser's study area is “a pretty local area of Antarctica,” but says, “[Fraser] has a really interesting point about a warming trend and the effect of snowfall patterns.”

    Fraser, who has studied penguins on the five islands near Palmer Station for more than 2 decades, has been trying to understand why Adélie populations have plummeted from about 15,200 breeding pairs in 1975 to 9200 today. Five years ago, he and colleagues published a paper in Polar Biology suggesting that a gradual reduction in sea ice in the western Antarctic Peninsula was playing a role: In the mid-1900s, heavy sea ice formed in three or four of every 5 years, but it is now seen just once or twice every 5 years. And while less sea ice seems to have helped out chinstrap penguins, which prefer open water, Adélies, which feed near ice edges, appear to be getting squeezed.

    But “just as we were beginning to feel pretty smug” that sea ice changes accounted for the trends, Fraser says, he and his colleagues noticed an odd geographical pattern to Adélie rookeries. On Litchfield Island, where the number of breeding pairs dropped 43% between 1975 and 1992, the thriving nesting colonies were concentrated on the island's northeast side. The abandoned rookeries, by contrast, were on the southwest side of the island's rocky middle ridge, where more snow accumulates as storms sweep over the islands. The same pattern turned up on nearby islands, the researchers found.

    So Fraser and his co-workers have taken a closer look at the nesting Adélies, and they have observed that birds laying eggs in snowy depressions at the edges of colonies seem to lose more eggs and chicks to snow and slush. “If you're not breeding in the right place, you're in trouble,” Fraser says. He adds that once the colonies begin shrinking, the penguins are less able to fend off predatory birds called brown skuas, which steal chicks and eggs.

    He thinks that retreating sea ice probably is the main driving force behind the drop in Adélie populations, but “superimposed on that,” he says, may be the effects of more snowfall in early spring when the birds begin breeding. He cautions that 5 years' worth of observations of snowfall and rookeries isn't a whole lot of data. Because there are no long-term snowfall records for the region, he can't be certain yet that snowfall in the regions has truly increased over the past 2 decades. But more snowfall has been documented in other parts of Antarctica, notes Fraser.

    Others agree that it may be a while before scientists are sure why the Adélies are dwindling. Some factor that has increased populations of competitors for habitat, such as elephant seals, also could be taking a toll on the birds, says Steve Emslie of Western State College in Gunnison, Colorado. “If you don't look at all those confounding variables before you point a finger at any particular cause, you're going to get in a lot of trouble,” Emslie says. Still, says Kooyman, “These kind of things are an alert for us to start looking more closely at things that we might otherwise overlook.”


    Life on the Edge: Rainforest Margins May Spawn Species

    1. Martin Enserink
    1. Martin Enserink is a science writer in Amsterdam, the Netherlands.

    With millions of life-forms crawling, flying, and slithering about, the world's rainforests have become the icon of biodiversity, as well as a living laboratory for researchers studying how new species appear. But a study on page 1855 of this issue suggests that many of these species may arise not in the rainforests themselves, but rather in their frayed edges. Led by evolutionary biologist Thomas Smith of San Francisco State University, the study of small forest birds in Cameroon suggests that “ecotones,” the transition zones between rainforest and savanna, may be important to generating biodiversity. If so, the authors argue, these oft-forgotten habitats deserve some attention from conservationists.

    What's more, the study—“a clever blend of ecology, biogeography, and molecular genetics,” as University of British Columbia evolutionary ecologist Dolph Schluter puts it—offers empirical support for a controversial theory of speciation. The work suggests that populations may diverge into different species even when some interbreeding occurs between them. “This really is a major first observation to support that kind of hypothesis,” says evolutionary biologist John Endler of James Cook University in Townsville, Australia, who has been promoting this idea for 20 years. Indeed, says Guy Bush, an evolutionary biologist at Michigan State University in East Lansing, the study is “probably the best piece of work so far” on this question.

    Smith, with co-workers from San Francisco State, the University of California, Los Angeles, and the Zoological Society of London, set out to examine a contentious issue in evolutionary biology: How isolated must populations of a species be before they can start to develop physical differences—and so take the first steps on the road to becoming separate species? For decades, the dominant theory was that such a split could occur only when there was almost no interbreeding, or gene flow, between populations. Even when natural selection in different environments pushed populations apart, it was thought that a small amount of gene flow would mix the genetic pool enough to prevent speciation.

    However, a growing number of lab experiments with the fruit fly Drosophilahave suggested that different selection pressures can create populations with different traits even in the face of moderate gene flow, says Smith. When researchers exposed different populations of flies to different selection regimes and allowed some gene flow between them—for example, by connecting their cages via narrow tubes—the populations started to diverge. Over time, the flies began to mate preferentially with flies within their own population. The theory predicts that this would eventually lead to reproductive isolation and thus to the birth of a separate species. But most biologists thought that this style of speciation rarely if ever occurred in nature.

    To test this model in the wild, the researchers chose Cameroon's little greenbul (Andropadus virens), a small green bird that feeds on insects and fruits and inhabits both the tropical rainforest and patches of forest surrounded by grasslands in the ecotone. Although transition areas like the ecotone aren't as species-rich as the rainforest proper, they include different types of environments close together and are often zones of contact between subspecies or closely related species. So biologists have suspected since the 1930s that these transitional areas may somehow be important in speciation, although there have been several competing theories as to just how such species would arise.

    Using mist nets, the team trapped birds from six forest and six ecotone sites between 1990 and 1996. They measured a number of physical traits that are important to fitness and known to be under selection in other bird species—weight, the depth of the beak, and the length of the wing, leg bone, and upper jaw. They also drew a few drops of blood for genetic analysis.

    Most of the ecotone-dwelling little greenbuls turned out to be quite different in size and shape from their rainforest counterparts. They were heavier and had deeper bills and longer wings and legs. At least one of those traits, longer wing length, offers an advantage in open ecotone environment because it allows the birds to fly faster and so escape aerial predators, which are more common in the ecotone, says Smith. In some cases, the morphological gulf between little greenbuls from the forest and those from the ecotone was larger than that between two different bird species sharing the same habitat.

    But the results of the genetic part of the study, which compared birds from different sites using both mitochondrial and nuclear DNA, revealed many similarities between the populations. The team used the amount of similarity and a population-genetics model to estimate that there was considerable gene flow between the two habitats—about one to 10 migrants per generation in each population of birds.

    Taken together, Smith says, the findings mean that selection pressure in different environments can offset gene flow and create differences in populations, even when individuals travel back and forth frequently. Only in populations that experienced very high gene flow—eight to 10 migrants per generation—were the effects of natural selection washed away. In these cases, the morphological differences between an ecotone and a rainforest greenbul were much smaller, about the same as between two ecotone or two rainforest birds. Of course, morphological divergence isn't the same as speciation. “But it is evidence for strong differentiation in the presence of gene flow,” says Endler. “So it's a first step in speciation.”

    Other researchers say they are impressed by the study's hard data on what has been a mostly theoretical debate. “I think they're the first to go out and test some of these hypotheses,” says Bush. Endler, who in 1977 was one of the first to point out that selection might overcome the effects of gene flow, says it's “exciting” to have one of his theories confirmed.

    Still, evolutionary biologist Nick Barton at the University of Edinburgh in the United Kingdom warns that the study hasn't quite sewn up the case for this model of speciation. He points out that the amount of gene flow the team found is quite small and says he's not surprised that it wasn't enough to swamp selection: “Four to five migrants per generation … that's very weak relative to selection.” Smith acknowledges that because he doesn't know the little greenbul's population sizes, it's hard to prove just how significant the one to 10 migrants are. His team is now censusing the birds to find out.

    While its implications for evolutionary theory are under debate, the study, if it holds up, might also have implications for the fight to save the world's biodiversity. If gene flow isn't a barrier to new species formation, then the ecotone is a likely source of new diversity, argues Smith. Species spawned in the ecotone might use forest resources in new ways that would be advantageous in the rainforest proper. Moreover, because migration between ecotone and rainforest is so common, over millions of years many of these new species might have moved back from their ecotone cradle to the rainforest. Thus, the patchy ecotone may generate species that find long-term homes in the adjacent rainforest, says Smith: “Maybe the rainforest is a sink for new species, rather than an area where new species are generated.”

    Yet conservationists have tended to neglect ecotones, because they are typically much less diverse and also less pristine than the central rainforest. The Australian ecotone between dry forest and rainforest, for example, is sometimes considered “not pure” and therefore difficult to protect, says Endler. In Cameroon, the transition zone—which at some points is over 965 kilometers wide—is currently being devastated by cattle grazing, burning, and wood cutting, according to Smith. In the 6 years of the study, the team lost three sites to fire. In addition, Exxon and Shell oil companies are planning to build a huge pipeline from the oil fields in Chad to the Atlantic Ocean. In an attempt to protect the rainforest, the line is to run right through the Cameroon ecotone. While Smith says the pipeline itself probably won't harm the ecotone much, he is concerned about the increased access of people to the area. “If we are to protect biodiversity, we should also protect the processes that generate it,” he says. “The study suggests we need to look at what's happening at the periphery as well. … I'm hoping this paper might be a wake-up call.”


    Geophysicists Ponder Ancient Chills and Elusive Quakes

    1. Richard A. Kerr

    BALTIMORE—The spring meeting of the American Geophysical Union (AGU) here late last month attracted the usual assortment of presentations on climate and space science (Science, 13 June, p. 1648), but there was also much discussion of seismology, a subject usually more popular at the fall meeting in earthquake country in San Francisco. Here is a selection from each area: a series of reports that focuses new attention on “silent” earthquakes and a climate modeling study that takes the middle ground in a dispute over ice age cooling.

    Listening to Silent Earthquakes

    Seismologists have developed increasingly sophisticated means of watching and waiting, but they have yet to find a sure warning sign that a fault is about to rupture in an earthquake. At AGU, researchers had no magic solution to this problem, but they did offer persuasive new evidence of a potentially helpful phenomenon to which they had previously turned a deaf ear: so-called slow or “silent” earthquakes. These are fault movements so slow—taking days instead of seconds as in ordinary earthquakes—that they produce no seismic waves and hence can't be picked up by the listening ears of seismometers.

    Silent earthquakes had seemed little more than rare curiosities, but as a flurry of presentations at the meeting showed, reanalysis of old records and use of additional instruments are revealing many more of these slow groanings of the crust. The new data also suggest that these slow movements can provide a window into the secret life of faults. A silent quake may defuse a fault that has been building toward failure—or transfer stress to part of the fault already near failure and trigger a normal earthquake. Detecting a silent precursor might thus give a few days' warning before a quake strikes. “There's enough evidence of slow and silent earthquakes to make them credible,” says seismologist Paul Segall of Stanford University. “The grand prize in all this would be seeing something occurring before earthquakes” and so allow quake prediction.

    No one has captured that prize yet, but a new analysis of data collected southwest of Tokyo just before the magnitude-8.1 Tonankai earthquake of 1944 suggests that a silent quake presaged that killer. Geophysicists Alan Linde and Selwyn Sacks of the Carnegie Institution of Washington's Department of Terrestrial Magnetism reexamined data collected by Japanese seismologist A. Imamura of the University of Tokyo. Fearing a great quake, Imamura had requested a survey to gauge the slow flexing of the land southwest of Tokyo. On the night of 6 December 1944, the surveyors were surprised to note that one end of an 800-meter survey line was 3 millimeters higher than it had been 3 days before. They resurveyed the next morning, only to find an additional 4-millimeter difference. Either their surveying was badly awry, or the southeast end of the line was rising extraordinarily fast. They got their answer that afternoon when the great quake struck.

    At the meeting, Linde and Sacks reported that the survey data fit a silent quake that triggered the devastating quake from below. They found that the sharp rise of the land implied about 2 meters of silent slip along a deep, 60-by-50-kilometer patch of the fault during the few days before the earthquake. Such slip, equivalent to a magnitude-7.5 quake, would have loaded stress on the upper part of the fault, apparently bringing it to the point of failure.

    It's not clear how often silent quakes presage normal ones or whether Japan's next great quake will issue a silent warning. To find out, the Japanese monitoring network is today listening for such silent precursors using two types of instruments. Strainmeters buried in drill holes record the flexing of the crust, while Global Positioning System (GPS) receivers on the surface determine their location to millimeter accuracy by comparing radio signals from U.S. military satellites overhead (Science, 5 April 1996, p. 27).

    Whether or not silent earthquakes turn out to be useful warning signs, several other talks showed that these slow movements aren't as rare as they once seemed. Geophysicist Malcolm Johnston of the U.S. Geological Survey in Menlo Park, California, and colleagues reported that a slow earthquake struck the San Andreas fault 150 kilometers south of San Francisco in April of 1996, when a 6-kilometer-square patch of the fault slipped a few centimeters over the course of 2 days. If the same slip had taken a few seconds, as in a normal small quake, it would have delivered a magnitude-4.9 jolt to the residents of nearby San Juan Bautista; instead, they felt nothing and even nearby seismometers failed to note it. Johnston's team, however, had installed strainmeters in nearby drill holes, and two of these recorded slow crustal deformation as the silent quake relieved stress on that patch of fault.

    Across the Pacific, a GPS network detected a slow earthquake in May 1996, according to Takeshi Sagiya of the Geographical Survey Institute in Tsukuba, Japan. In the course of a week, about a dozen sites on the Boso Peninsula southeast of Tokyo moved as much as 15 millimeters to the south-southeast. With such limited data, Sagiya can't say which fault ruptured, but he suspects that it was a deeply buried fault, and estimates that the quake had a magnitude of 6.3.

    In many instances, these silent earthquakes are good news, apparently releasing strain that would otherwise erupt in a damaging quake. The San Juan Bautista event, which may have been triggered by a smaller, normal quake, harmlessly released strain on the San Andreas. A similar silent quake just to the south in 1992 did the same, as Linde and colleagues reported in the 5 September 1996 issue of Nature. Some of the strain loaded on the Sierra Madre fault just northeast of Los Angeles by the 1994 Northridge earthquake also was apparently dissipated in a silent earthquake, according to a presentation by geophysicist Michael Heflin and his colleagues at the Jet Propulsion Laboratory (JPL) in Pasadena, California. The GPS receiver at JPL—which sits right on the Sierra Madre—measured 35 millimeters of horizontal motion, apparently reflecting silent strain relief that reduced the risk of a second earthquake. If such harmless strain release is common, notes Heflin, it could significantly reduce the hazard projected for areas such as San Juan Bautista, which are constantly strained as the Pacific and North American plates grind past each other.

    For now, no one can be sure why some silent earthquakes harmlessly release strain and others trigger great quakes. But to catch more clues to how both noisy and silent earthquakes work, researchers will be listening ever more closely to Earth's most subtle whispers.

    Moderating a Tropical Debate

    Everyone knows that Earth was deeply chilled 20,000 years ago at the height of the last ice age, but how much did the tropics cool? For years, that question has divided researchers into two extreme camps, one seeing a much bigger drop than the other, with seemingly no middle ground between them. The gulf between these views may not be closing, but now at least it may be better understood.

    The estimates of modest cooling came from oceanographers tracking the fluctuating abundance of microfossils in marine sediments. They inferred only a 2°C average cooling across the tropics, implying that the climate system then was dramatically different from today's and somehow kept more heat trapped in lower latitudes. But paleoclimatologists consulting a range of mostly terrestrial records—pollen in lake muds, isotopes in corals, and noble gases in ground water, for example—estimated perhaps a 5°C tropical cooling (Science, 4 October 1996, p. 35).

    At the AGU meeting, climate modeler Anthony Broccoli of the Geophysical Fluid Dynamics Laboratory (GFDL) in Princeton, New Jersey, suggested that both groups' estimates may be reasonably accurate—but that neither is seeing the whole picture. His new computer modeling of ice age climate suggests that the cooling varied more from one region to another than indicated by previous models. This variation, Broccoli says, may explain much of the disparity in the data. Terrestrial temperature records are scattered so thinly that they could be missing areas of moderate cooling, he says, and even the more complete marine microfossil record probably has its blind spots. Completely reconciling the numbers is still going to be difficult, he says, but the model results could account for a few degrees of difference.

    Broccoli came to these conclusions after adding two key improvements. Most earlier simulations simply adopted oceanographers' estimates of glacial sea-surface temperatures, but in the new model the surface ocean temperature can vary in response to changes elsewhere in the climate system. Moreover, the resolution of the model is about twice as good.

    When this model was run under glacial conditions, such as increased ice coverage and reduced greenhouse gases, the tropics on average cooled only about 2°C, about as much as the oceanographers would have it. But the cooling was not uniform. For example, the model's Northern Hemisphere cooled more than the Southern Hemisphere due to the far larger area of ice in the north, an effect that extended into the northern tropics, where most of the best known terrestrial records are.

    In general, Broccoli found that the model's tropical land cooled about 1 degree more than the ocean. He says part of this is due to the 100-meter sea-level drop during the ice age, which in effect raised the land 100 meters. Because temperature drops with increasing altitude above sea level, the land cooled slightly. Also, land has less moisture available to moderate temperature swings and so would be expected to show more cooling during an ice age. All this suggests that regional variation was typical of the glacial climate system—and that researchers should “be careful extrapolating from a few sites to the entire tropics,” says Broccoli.

    That message got a varied reception. Thomas Guilderson of the Lawrence Livermore National Laboratory in Livermore, California, who has inferred a 5° cooling from Caribbean coral isotopes, sees increasing evidence for a large tropical cooling, but admits that it may have varied regionally, conceding that it may have been only 3°C in places.

    Paleoceanographer William Ruddiman of the University of Virginia, who was involved in the 1976 microfossil study that first proposed a moderate cooling, sees more hope for a reconciliation. Rather than requiring a glacial climate system radically different from today's, he can imagine “an accumulation of several really uninteresting 1° problems” of the sort Broccoli is finding that would account for much of the gap between the estimates. If so, there may be middle ground between land and sea after all.


    Does a Common Virus Give HIV a Helping Hand?

    1. Michael Balter

    PARIS—The major focus of basic AIDS research over the past 18 months can be summed up in one word: chemokines. The discovery that HIV hijacks the cell surface receptors for these immune system signaling molecules to force entry into its target cells has revolutionized the field and opened new avenues toward possible therapies. Now, research reported in this issue of Science suggests that HIV may have yet another port of entry into some cells. The findings implicate a common virus as a possible accomplice of HIV, helping the AIDS virus infect some types of cells and wreak havoc on the immune system.

    Deadly partner?

    Cytomegalovirus (above) may help HIV to infect target cells.


    On page 1874, a team led by Marc Alizon at the Institut Cochin in Paris, in collaboration with Michel Seman at the University of Paris, reports that HIV may use a protein called US28 to enter some types of cells. US28 is produced by cytomegalovirus (CMV)—a member of the herpesvirus family that has long been a leading suspect as an AIDS cofactor. The protein is expressed in cells experimentally infected with CMV, and it had previously been shown to act as a receptor for the same chemokines that bind to CCR5, the chemokine receptor used by HIV strains that dominate during the early phases of infection.

    Although the evidence supporting a cofactor role for CMV is contradictory and controversial, researchers say that if these new results hold up, they would imply a tighter symbiotic relationship between HIV and CMV than previously imagined. “The results are provocative and potentially important for HIV pathogenesis,” says Philip Murphy, a chemokine-receptor expert at the National Institute of Allergy and Infectious Diseases in Bethesda, Maryland.

    To determine whether US28 might act as HIV's accomplice, the French group inserted the US28 gene into a human laboratory cell line that HIV does not normally infect. Alizon's group then exposed these cells, which express the US28 protein on their surfaces, to various HIV strains. They also tested whether these modified cells fused with a second cell line engineered to carry proteins from HIV's outer viral coat. In these and related experiments, Alizon's team found that US28-bearing cells were easily infected by HIV strains that normally use the human chemokine receptor CCR5, and somewhat less easily by strains that use another human receptor, CXCR4.

    These findings are being greeted with surprise. Over the past year, several labs working on chemokine receptors—including Murphy's—had conducted experiments similar to those of the Alizon team to test whether US28 helps HIV enter cells. “We all got negative results,” Murphy says, “although none of us actually verified expression of US28 in the systems we were using, whereas Alizon et al. did.”

    Nevertheless, researchers who spoke to Science said there was no reason to doubt Alizon's results. “It is certainly believable,” says David Posnett, an immunologist at Cornell University Medical College in New York who has studied CMV. “I'm just wondering what this all means in real life.” Indeed, it is not yet clear whether the French group's results, which are restricted to laboratory cell lines, are relevant to HIV-infected people. “There is a total lack of information about when, where, and how much of US28 is expressed in people infected with CMV, and whether native US28 can support HIV entry into CMV-infected cells,” says Murphy.

    Then there is the basic question that has always dogged research on CMV and AIDS: Is CMV just an opportunistic organism taking advantage of the immune suppression caused by HIV, or a true partner in destroying the immune system? “There is still a lot of debate in this area,” says Thomas Folks, chief of the retrovirus diseases branch of the Centers for Disease Control and Prevention in Atlanta. Although numerous studies have shown that CMV can enhance transcription of HIV's genome in some cell types, epidemiological evidence that infection with both CMV and HIV leads to a worse outcome for patients has been hard to come by. The reason: Up to 80% of the general population, and almost all HIV-infected homosexual males, have been infected by CMV, which makes it very difficult to find CMV-free control groups to make comparisons. Infection rates are lower in children, however, and several recent studies have shown that children coinfected with HIV and CMV have a much higher death rate than that of those infected only with HIV.

    If CMV is indeed a partner in HIV's immune system destruction, researchers will have to explain yet another puzzle: Although CMV is known to infect the brain and retina, there is little evidence that coinfection of CD4 T cells—HIV's main target—by both CMV and HIV is a common event. For example, a study published in the Journal of Medical Virology last year by Sylvia Bertram and her colleagues at the University of Freiburg in Germany concluded that less than one in 100 HIV-infected CD4 cells also harbored CMV. But Alizon and his colleagues hypothesize that CMV might in some cases transfer US28 to cells without actually entering them, which would allow US28 to act as an HIV receptor even in cells not directly infected with CMV.

    Craig Gerard, a chemokine expert at Harvard Medical School in Boston, suggests another possible role for US28. Gerard proposes that CMV might allow HIV to expand the repertoire of cells it can attack by infecting cells that do not normally carry the chemokine receptors such as CCR5 or CXR4 which HIV normally hijacks. “If you add another chemokine receptor, you broaden the host range of infectible cells, in places like the central nervous system.”

    While the lack of hard evidence about US28's actual role in HIV-infected people makes most of these ideas purely speculative, the Alizon team's findings are sure to stimulate more research. “There is no doubt that many labs will act quickly” to try to confirm the results, says Murphy, adding: “If the authors are right, we will have to shift our thinking about CMV and AIDS from the present ‘welfare’ model—in which CMV is viewed as an opportunist taking advantage of the immunodeficiency caused by HIV—to a ‘workfare’ model, in which CMV earns its keep by providing HIV with another way to enter target cells. The idea that two different pathogens might act symbiotically this way to double-team the host is quite novel and interesting.”


    New Experiments Step Up Hunt for Neutrino Mass

    1. Dennis Normile

    TOKYO—Some of physicists' fondest hopes ride on the neutrino, a shadowy and seemingly massless particle. If the neutrino does have a trace of mass, swarms of them could account for some of the mass that cosmologists believe is missing from the universe as a whole. A massive neutrino might also point the way to a new theory of elementary particles and forces that would transcend the current Standard Model. Now experiments at underground laboratories around the world are striving to weigh the neutrino—without ever observing it.

    The new experiments monitor a radioactive material with sensitive detectors, watching for an excruciatingly rare—perhaps nonexistent—process known as neutrinoless double beta decay. These are far from the only efforts to measure neutrino mass. Other researchers, for example, are trying to catch neutrinos in the act of “oscillating” from one of the three neutrino types—called flavors—to another, a transformation that would be a sure sign of mass. But while the oscillation experiments might be sensitive to neutrino masses as low as thousandths of an electron volt (eV), they can only reveal the difference in mass between two flavors. Neutrinoless double beta decay, in which the energy of electrons flung from a decaying radioactive nucleus is measured, will give an absolute value for the mass of the neutrino. That makes the experiments “an essential part of the program of modern particle physics,” says astrophysicist John Bahcall of the Institute for Advanced Study in Princeton, New Jersey.

    The catch is that so far, neutrinoless double beta decay has never been observed. Ordinary double beta decay is a rare, although regularly observed, process in which two neutrons in a radioactive nucleus decay into two protons, emitting two electrons, or beta rays, and two antineutrinos (antimatter counterparts of the neutrino). In the elusive neutrinoless form, other protons within the nucleus would absorb the antineutrinos as neutrinos, and only the electrons would escape. By comparing the measured energy of the electrons with the total energy of the process predicted by theory, researchers could calculate the energy—and therefore the mass—of the neutrinos.

    Neutrinoless double beta decay can take place only if the neutrino has a nonzero mass. Its failure to appear in experiments so far means the process must be rare—and the neutrino mass minute. Current results already suggest a ceiling of roughly 1 eV—a tiny fraction of the mass of the electron. By monitoring larger quantities of radioactive material with more sensitive detectors while screening out sources of background radiation, the new generation of experiments aims to push the mass sensitivity as low as 0.1 eV. It's a level “people didn't think possible a few years ago,” says Rabindra Mohapatra, a theoretical physicist at the University of Maryland, College Park.

    Earlier this year, scientists from the Max Planck Institute for Nuclear Physics in Heidelberg, Germany, and the Russian Science Center Kurchatov Institute in Moscow reported lowering the upper limit of the possible mass of the electron neutrino to 0.48 eV. Mohapatra notes, however, that uncertainties in the “extremely complicated calculations” mean the results, from Gran Sasso National Laboratory in Italy's Apennines, could be off by a factor of 2.

    Now Osaka University's Research Center for Nuclear Physics in Japan is setting up a double-beta-decay experiment in its new Oto Cosmo Observatory in a never-used railroad tunnel 100 kilometers south of Osaka. Hiro Ejiri, a physicist at Osaka, says that steady winds blowing through the tunnel at the new observatory help reduce its natural radon concentrations, which can result in background events. “We hope to get the sensitivity down to 0.6 eV,” he says. That's no better than the Russian-German group has already claimed. But because the two groups are working with different isotopes (see table), Mohapatra and others say the Osaka experiment could provide independent confirmation of the recent Heidelberg-Moscow results.

    View this table:

    A second European group, including scientists from France's CNRS, several French universities, and the Joint Institute for Nuclear Research in Dubna, Russia, wants to push the limits even lower. The collaboration, called NEMO, hopes by the end of 1998 to bring a new detector online in the Frejus Underground Laboratory in the Alps along the French-Italian border. Francis Laplanche, a physicist at the University of South Paris and a member of the team, says that the group intends to reduce the upper limit on neutrino mass to 0.1 eV by using a large mass of an isotope of molybdenum, cutting background radiation, and improving detection schemes. The Heidelberg-Moscow group, however, believes it will reach that level first. Hans Klapdor-Kleingrothaus of the Max Planck Institute says the goal is to achieve an upper limit of 0.1 eV within 5 years “just by letting the experiment run.”

    If any of the experiments actually pin down the mass of the neutrino, physicists would have their first clear clue to a theory beyond the Standard Model. The various Grand Unified Theories make different predictions for neutrino mass, and the results to date, by lowering the ceiling on neutrino mass, are casting doubt on some theories, says Moscow's Alexei Smirnov. “These bounds [on neutrino mass] could forbid some schemes,” he says—not a bad payoff for an experiment that never sees its quarry.