News this Week

Science  14 Nov 1997:
Vol. 278, Issue 5341, pp. 1223

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


    The Architecture of Hearing

    1. Elizabeth Pennisi


    The M family of Costa Rica can trace its ancestry to Spanish explorers who settled in the Americas about 1600. They can trace another family legacy as far back as the mid-1700s: a tendency for many family members to go deaf, often beginning at age 10. Eight generations later, about half of 400 some members of this extended family—most of them still living in the same community—eagerly agreed to help Pedro León, a molecular biologist at the University of Costa Rica, track down the cause of their deafness.

    Now the 20-year quest, chronicled in Costa Rican newspapers and television, is over. Geneticists Eric Lynch and Mary-Claire King of the University of Washington, Seattle, working with León, have cloned the gene responsible for the family's deafness.

    The gene, described in this issue of Science, is only the latest harvest of an effort to track down genes for nonsyndromic deafness—deafness with no other symptoms—that accelerated in 1992, when King, Lynch, León, and their colleagues mapped the chromosomal location of the gene at fault in the Costa Rican family. King, Lynch, and León's achievement provided a big boost to the field, as it was the first to use family studies to narrow down the chromosomal location of these genes. “It was a kick in the field, the [result] we all cited in our grant applications,” says Cynthia Morton, who studies the genetics of deafness at Brigham and Women's Hospital in Boston.

    Precise alignment.

    As these hair cells mature, the projections will form a distinct V pattern.


    The King team's report, which appears on page 1315, brings to three the number of new, nonsyndromic genes pinpointed in the last year. And many more are on the way; the positions of some 30 other genes for nonsyndromic deafness have been mapped since the 1992 report. The new genes are not the first “deafness” genes. Researchers have already cloned more than two dozen genes that cause “syndromic” deafness, in which deafness comes with other symptoms, such as blindness or pigment abnormalities. But 70% of hereditary deafness is nonsyndromic, and up to 60% of the 28 million cases of hearing loss in the United States are thought to have a hereditary component. In addition, because the mutations in nonsyndromic deafness genes affect hearing alone, researchers expect these genes to offer clues to how the human auditory system works, and how it can go awry.

    View this table:

    Indeed, all three of the new genes seem to be involved in the operation of the sound-sensitive hair cells in the inner ear, or cochlea. Two of them, including the one that is mutated in the Costa Rican family, code for proteins that apparently help organize actin, a structural protein that stiffens the microscopic projections that crown these cells; the third helps build an electrical channel that may enable hair cells to reset themselves after they are exposed to sound. All three could offer glimpses into how hair cells normally develop and function, and point to treatments to keep hearing sharp.

    Genetic studies of nonsyndromic deafness have been slow in coming because the symptoms of individuals who are deaf, but don't have other problems, are so similar that geneticists have had a hard time sorting out people who are likely to have the same gene defects. Until the new crop of genes, geneticists had been able to link only three other genes to hearing loss without any symptoms. All three have distinctive inheritance patterns, which made them easier to track down: Two are found in the cellular organelles called mitochondria and the third on the X chromosome. But mutations in these genes are relatively rare.

    Mice to men

    Because of the difficulty of human family studies, some researchers took a different tack entirely. For example, mouse geneticist Karen Steel of the Medical Research Council (MRC) Institute of Hearing Research in Nottingham, United Kingdom, and her colleagues turned to mutant mice, hoping to ferret out deafness genes that would lead them to comparable genes in humans. She began with a mouse called Shaker1, which can't hear or keep its balance. With Steve Brown from the MRC Mouse Genome Centre in Harwell, she then interbred these mice, keeping track of how often various genetic markers were associated with hearing loss.

    This analysis showed that the animals' deafness is caused by a mutation in one of their genes for myosin, a protein that interacts with actin in many types of cells. The human counterpart of the gene, called myosin VIIA, resides at a site on the long arm of human chromosome 11 that had already been linked to a condition called Usher syndrome 1b, whose victims are both blind and deaf. In work completed about 2 years ago, Brown and Steel went on to show that the syndrome is indeed caused by a mutant myosin VIIA gene.

    But the researchers suspected that the gene might be responsible for other cases of deafness as well. Because Shaker1 mice are not blind, it seemed that a still less severe myosin VIIA mutation might produce hearing loss alone—a hunch Brown and Steel subsequently confirmed.

    The researchers first studied eight Chinese families in which deafness was a recessive trait. In two of these the condition could be traced to a defective myosin VIIA gene. At the same time, Christine Petit at the Pasteur Institute in Paris found the same recessive defect in the deaf members of a large Tunisian family. (Both sets of results appeared in the June issue of Nature Genetics.) And now, in this month's Nature Genetics, Brown and Steel show that this same gene, mutated yet another way, causes a progressive form of nonsyndromic deafness. The recessive deafness seen in both the Chinese and Tunisian families is present at birth. But the new mutation identified by Brown and Steel is dominant—only one copy of the gene needs to be mutated to cause the condition—and the deafness develops after birth and exposure to language. “myosin VIIA is involved in a spectrum of deafness,” Steel concludes.

    The key to its involvement, further work by Steel and her colleagues suggested, may be the effect that myosin VIIA mutations have on the structure of the hair cells. Normal hair cells are crowned with a V-shaped array of projections called stereocilia, which consist of an actin core and a myosin outer cover. The arrangement of stereocilia is important because they bend when sound vibrates the fluid surrounding the hair cells, causing those cells to fire a signal at the auditory nerve. But in mice with a mutated myosin VIIA gene, Steel says, “instead of a nice V shape, you get little clumps of stereocilia.”

    She and Brown think that the myosin may carry or anchor other molecules that are important for forming the precise arrangement of the stereocilia, and also for maintaining them over time. That double role may be why some people with the mutations are deaf from birth, while others develop the condition later in life, Steel says.

    Family history

    The new deafness gene described by Lynch and King in this issue may also lead to hair-cell disruptions. This group was able to surmount the difficulties of genetic linkage studies thanks to León's Costa Rican family. The family tree León worked out covered more than 190 people over eight generations and showed that about half the children born to deaf parents also became deaf. By analyzing DNA from 147 family members, 78 of whom had lost their hearing, the researchers were able to narrow the location of the faulty gene to an 800,000-base pair region on chromosome 5.

    Next, Lynch sequenced that portion of the chromosome and checked the resulting sequence against genes or partial genes on file in the public database called GenBank. The computer searches turned up more than 15 candidate genes. One of them—the human equivalent of a gene called diaphanous, previously identified in fruit flies and mice—is consistently mutated in the deaf members of the Costa Rican family, but not in unaffected members or in unrelated controls.

    Evidence that these mutations might also impair hair-cell function came when Lynch, with help from Morton's lab, showed that the diaphanous gene is active in the human cochlea. In addition, work in mice and fruit flies indicates that the gene's protein product serves as temporary scaffolding for actin as it rearranges to help a cell divide or form projections such as the stereocilia. “We've identified a gene that encodes a protein that when normal and healthy is probably critical to the maintenance of normal hearing,” King says.

    Serendipity aided the discovery of the third member of this year's trilogy of nonsyndromic deafness genes. The gene is connexin 26, which was already known to code for a protein that helps make gap junctions—electrical channels between adjacent cells—and genetic linkage studies by several research teams had already suggested that both dominant and recessive forms of nonsyndromic deafness might be due to mutations in the same region of chromosome 13 that contains the gene. But Irene Leigh, David Kelsell, and Howard Stevens of the University of London and their colleagues were pursuing a third form of deafness, which seemed to be inherited along with skin problems in a Caucasian family. They found that it, too, seemed to result from a mutation somewhere in the vicinity of connexin 26.

    The group then began looking for connexin 26 mutations in family members and found that this gene was indeed responsible for their deafness, but not for the skin problems, which must be caused by a mutation in a different, still undiscovered gene. When the Leigh group expanded their genetic analyses in other families with nonsyndromic deafness, they found that the connexin 26 gene was at fault in the two types of nonsyndromic deafness as well.

    As the team reported in the 1 May Nature, they confirmed that the connexin 26 protein is made in the human cochlea by treating tissue from the inner ear with mouse antibodies to the protein. She and her colleagues suspect that the gap junction it helps build normally creates a channel that pumps out the potassium ions that flood the hair cells when they are stimulated. Without the gap junctions, Kelsell speculates, the potassium levels in the cells may stay high, and the mechanism that enables the cells to respond quickly to sounds may not be reset.

    Researchers do not yet know what percentage of hereditary deafness is caused by myosin VIIA and diaphanous mutations, but connexin 26 mutations are turning out to be surprisingly common in families with a history of deafness. In this month's issue of Human Molecular Genetics, Petit's team at Pasteur reports that the gene was mutated in 39 of the 65 families tested. “It's clearly a major contributor to deafness,” says Thomas Friedman, a human molecular geneticist at the National Institute on Deafness and Other Communication Disorders (NIDCD) in Rockville, Maryland.

    What's more, the gene apparently has a mutational “hot spot.” Although the families studied by Petit come from all over the world—Tunisia, France, New Zealand, and the United Kingdom—half had the same mutation: the loss of one of a string of six guanine bases located near the beginning of the gene. Other teams made similar observations in deaf families in an isolated Israeli-Arab village and in the Mediterranean area. That finding, combined with the fact that connexin 26 is a simple gene, containing only one coding region, means it should be relatively simple to develop a genetic test for connexin 26 mutations. Such a test would allow rapid diagnosis of deafness in newborns so that they can be taught sign language from a very early age, keeping them from falling behind in their language acquisition skills.

    Other hearing researchers are optimistic that these gene finds will lead to treatments, especially for people who are born able to hear and then become deaf later. In addition, researchers speculate that subtle mutations in the genes might also cause the very common type of hearing loss associated with aging. If so, “you stand a much better chance of being able to do something about [progressive deafness],” says Steel.

    Indeed, researchers predict that the deafness genes found this year are just the beginning. In Boston, Morton has created the first cDNA library of genes expressed in the human cochlea, and Petit has created a mouse cochlea cDNA library. Each of these genes could be a candidate deafness gene. “The field is really just bursting at the seams,” says James Battey, scientific director of NIDCD. He predicts that a half dozen more will be cloned in the next year.*

    These new additions should help piece together the mystery of hearing as well as the puzzle of deafness. “We're on the brink of a whole new understanding of the molecular dynamics of hearing,” says Lynch.


    Gamma Rays Open a View Down a Cosmic Gun Barrel

    1. James Glanz

    Among the most spectacular displays in the cosmos are the colossal jets of material that blast out of certain galaxies, crackling with radiation across the entire electromagnetic spectrum. Most of the jets point away from Earth. But a few—so-called blazars—are aimed straight at us, bombarding our atmosphere with high-energy gamma rays. Lately, astronomers have been tracking the flashes of light produced when those gamma rays smash into the atmosphere. As Trevor Weekes of the Fred Lawrence Whipple Observatory in Amado, Arizona, puts it, they are “looking down the barrel of a gun” to see how it works. They got their best view yet early this year, when one of these celestial guns fired a volley straight into their detectors.

    Jet power.

    A jet of particles erupting from a black hole collides with photons from within the jet or sources close to it, boosting them to ultrahigh energies.


    At a meeting of the American Astronomical Society's High Energy Astrophysics Division in Estes Park, Colorado, last week, groups from Europe and the United States reported that they had nabbed a blazar called Markarian 501 in a violent outburst. Over several months starting in February, its brilliance in high-energy gamma rays sporadically flared to levels equivalent to 10 billion suns, making it the brightest object in the sky at those wavelengths. At the same time, observations from other instruments on the ground and in space showed that Markarian 501's intensity at longer wavelengths, such as x-rays and ultraviolet, danced roughly in step with the gamma rays.

    To theorists, the synchrony implies that the ultimate power source for all these forms of radiation is a beam of charged particles accelerated, perhaps by a spinning black hole, to nearly the speed of light within a small, magnetized region. The details of that dance, say these researchers, offer clues to the troupe of particles and interactions responsible for it. By watching it, says Alan Marscher of Boston University, “we can understand what's going on as close to the black hole … as anyone's ever been able to observe inside a jet.”

    To look down a cosmic jet, astronomers in this new field look up at the night sky for faint bursts of light that are the electromagnetic version of sonic booms. Gamma rays at energies of trillions of electron volts (TeV) don't reach the ground, and satellite-based detectors are ineffective in these energy ranges. But when a high-energy gamma ray crashes into the atmosphere, its interactions with air molecules trigger a narrow tube of Cerenkov light, which can be traced back to the gamma ray's ultimate source in the heavens. Among the first to study blazars by this technique, called atmospheric Cerenkov imaging, are researchers at the Whipple Observatory, which is run by the Harvard-Smithsonian Center for Astrophysics.

    Last year, the Whipple Observatory's 10-meter reflecting dish demonstrated the principle by seeing a blazar called Markarian 421 flickering on time scales as brief as 15 minutes. Since then, two new gamma ray detectors have joined the hunt: the French Cerenkov Array at Thémis (CAT) and the High Energy Gamma Ray Astronomy facility (HEGRA) at La Palma in the Canary Islands. HEGRA was built mainly with German backing and is now recovering from a recent fire (Science, 31 October, p. 807).

    As a result, there were plenty of eyes on Markarian 501 after the Whipple group noticed it brightening in February. Instruments at other wavelengths also joined the vigil, including the orbiting Compton Gamma Ray Observatory at less energetic gamma wavelengths, the Rossi X-ray Timing Explorer, and x-ray detectors aboard the Italian BeppoSAX satellite. “Markarian 501 was a big surprise,” says BeppoSAX investigator C. Megan Urry of the Space Telescope Science Institute (STScI) in Baltimore, who worked with Laura Maraschi of the Osservatorio Astronomico di Brera, Elena Pian of STScI, and others. “For many years, people dismissed it as boring; it didn't vary a whole lot.”

    By March and April, no one was yawning. Pian and her colleagues captured a tremendous x-ray flare in early April. In the TeV range, says Axel Lindner of HEGRA and the University of Hamburg in Germany, the blazar—although 300 million light-years distant—brightened for hours or minutes until it shone 10 times more brightly than the nearby Crab Nebula, a brilliant but steady source of gamma rays. “It had gone from the dimmest source we knew of in the gamma ray sky to the brightest,” says Michael Catanese of Iowa State University in Ames, who presented the Whipple team's results in Estes Park. “We'd never seen anything like that.” Changes across the spectrum occurred roughly together, and “the results of the three [TeV] experiments agree very nicely,” says Lindner.

    The roughly synchronous changes imply a single source for all the radiation, says Charles Dermer, a theorist at the Naval Research Laboratory in Washington, D.C. One candidate, he says, is the shock waves that would naturally form in a jet of material squirted out at nearly the speed of light along the axis of a spinning black hole. Intense electric fields in the shock waves would function as a staggeringly powerful particle accelerator.

    Dermer adds that the radiation's spectrum offers a clue about how the accelerated particles then ignite the brilliant displays seen from Earth. The spectrum, he says, has two humps, one at optical and x-ray wavelengths and the other in gamma rays. The low-energy hump probably comes from so-called synchrotron radiation, thrown off by fast electrons when their paths bend in magnetic fields. The high-energy hump probably results when the electrons crash into photons—either from the synchrotron process or from glowing gases outside the jet—boosting their energy.

    Raymond Protheroe of the University of Adelaide in Australia points out that there could be a simpler explanation for the highest energy radiation: Fast protons in the jet might smash into ambient material, shattering into a cascade of unstable, lighter particles. These would quickly decay into the gamma rays that reach Earth.

    Either process would have to be taking place at the very heart of the blazars. The rapid variability of the TeV radiation—indicating emission from a small region—and its remarkable intensity suggest that it emanates from within a tenth of a light-year of the putative black hole. Observers are seeing right into the chamber of the gun.


    Molecules Give New Insights Into Deadliest Brain Cancers

    1. Marcia Barinaga

    New OrleansGlioma is among the deadliest of tumors: Most of the 20,000 people diagnosed each year in the United States with this form of brain cancer die within 2 years. The tumors—which derive from support cells of the brain known as glia—are eerily invasive: Glioma cells spread freely into healthy surrounding brain tissue, making it impossible to remove all of the tumor surgically and virtually assuring that the cancer will soon return elsewhere in the brain. “We have so little to offer these patients,” says Bruce Ransom, chair of neurology at the University of Washington Medical Center in Seattle. “This is an area of neuroscience that is long overdue for a significant breakthrough.”


    Cells fan out into healthy tissue from the border of a human glioma (left). A breast cancer tumor that has metastasized to the brain shows no such invasiveness (right).


    Ransom and other researchers fighting glioma would settle for some real progress on two fronts: clues to what makes glioma insidiously invasive, and ways to direct drugs specifically to the malignant cells. Findings presented last month at the meeting of the Society for Neuroscience here could offer both. Two research teams have identified different proteins—one an ion channel and the other a secreted protein that sticks to the matrix of molecules surrounding brain cells—that are produced by all the human gliomas tested, and apparently by no other cells in normal adult human brains.

    Because they are so specific to glioma, the proteins could make good targets for directing antiglioma therapy right to the culprit cells. What's more, both teams, one led by Susan Hockfield of the Yale University School of Medicine and the other by Harald Sontheimer of The University of Alabama, have evidence that the proteins may help glioma cells invade surrounding brain tissue. The studies “both may be very important,” says neuro-oncologist Henry Brem, co-chair of a multicenter consortium for testing brain-tumor therapies headquartered at The Johns Hopkins University School of Medicine. “Anything we can do to better understand and control [glioma's] invasiveness is a major advance.”

    Hockfield's team was initially looking for proteins important in brain development, not glioma. Postdoc Diane Jaworski was searching for a brain-specific member of a class of proteins that bind hyaluronan—a sugar component of the extracellular matrix, the interlocking web of molecules surrounding cells. Previous work had shown that hyaluronan-binding proteins help cells to move around in tissues during development. Jaworski hit pay dirt with a protein she called BEHAB, for brain-enriched hyaluronan binding. Hyaluronan-binding proteins have also been associated with invasive cancers, so Jaworski checked and found that BEHAB is made and secreted by human gliomas, but by no other cells in the adult human brain. What's more, it appears to contribute to the invasiveness of glioma cells: Rat glioma cell lines that form invasive tumors make high levels of BEHAB, as do surgical samples of highly invasive human gliomas that she checked. Rat glioma cell lines that form tumors without invasive properties don't make the protein.

    BEHAB is normally cut in two as soon as it is secreted, and to see whether the protein's hyaluronan-binding portion could boost invasiveness, Hockfield's team put a gene encoding that fragment into a noninvasive rat glioma cell line. When the cell line was put back into rats' brains, expression of the BEHAB fragment “increased the invasive potential of these cells,” says Hockfield. She notes, however, that the cells making the BEHAB fragment aren't quite as invasive as the most invasive rat cell lines or human gliomas. “My sense is we don't have all the elements there that an invasive cell needs.”

    Indeed, tissue invasion probably depends on multiple proteins, and Sontheimer's team has discovered another one that seems to play a role: an ion channel that allows chloride and other negatively charged ions to pass through the glioma cell membrane. They found the ion channel by chance while they were doing electrophysiological studies on glioma cells. “Then the story got really exciting,” Sontheimer says. His group found no evidence of the channel in normal glial cells or normal brain tissue, but it was present in “every single [human glioma] we studied.” What's more, in the higher grade, more invasive, gliomas, nearly 100% of the cells had the channel, while in the lower grade gliomas, the channel turned up in just under 50% of the cells.

    Using a molecule from scorpion venom called chlorotoxin, which specifically blocks this new channel, Sontheimer's team probed the channel's role in the glioma cells' ability to slip through tiny spaces. They found that the chlorotoxin reduced the glioma cells' ability to penetrate a mesh of extracellular matrix proteins in a culture dish. Sontheimer speculates that for glioma cells to migrate straight through dense and healthy brain tissue, they must lose fluid so they can become “thin and spindly” and slip through small molecular openings. The chloride channels, he suggests, may help them lose salts, promoting water loss.

    Ransom says proteins like these “must contribute something to [the cells'] behavior, since [they are] so absolutely uniform across this whole class of cells that have in common unrestrained growth.” But glioma researcher David Louis of Harvard Medical School in Boston warns against jumping to conclusions because, he says, experimental tests of invasiveness don't truly replicate the kind of pernicious tissue invasion seen in human gliomas. Hockfield and other researchers agree that it will take a lot more study to learn whether blocking either protein will lead to therapies that can counter glioma invasiveness.

    On the other hand, Ransom notes, “it doesn't matter” whether you even know what a protein does if your aim is to use it as a target for anticancer drugs. Sontheimer is already pursuing that strategy. He and others have founded a Birmingham start-up company, Transmolecular Inc., which is planning to use chlorotoxin to direct cell-killing agents to glioma cells. In preclinical trials in animals, researchers at Transmolecular have shown that chlorotoxin can ferry radioactive iodine specifically to the tumors, and Sontheimer says they are currently juicing up the chlorotoxin molecule with enough radiation to actually kill the tumor cells when the toxin binds to them. If the animal tests go well, the team plans to apply for permission to conduct clinical trials in human glioma patients by next year.

    “We are very anxious in the [brain cancer] consortium to test this hypothesis,” says Brem. “If it shows promise clinically, it will be a major breakthrough”—and one that glioma researchers and patients would welcome as long overdue.


    HIV Survives Drug Onslaught By Hiding Out in T Cells

    1. Michael Balter

    As the war against HIV, the virus that causes AIDS, nears the end of its second decade, researchers and clinicians have wheeled powerful new cannons onto the battlefield. Potent cocktails of anti-HIV drugs have led the counterattack, pounding the virus down to undetectable levels in the blood of many HIV-positive patients. But a pair of papers published in this issue of Science (pp. 1291 and 1295) have both good and bad news for commanders on the front lines.


    T cells killed by HIV that was hiding in the latent reservoir of a patient on antiviral treatment.


    Two research teams—one led by immunologist Robert Siliciano at The Johns Hopkins University School of Medicine in Baltimore and the other by virologist Douglas Richman of the University of California, San Diego, School of Medicine in La Jolla—report that many patients taking the new drug cocktails, known as combination therapy, for as long as 30 months show no signs of developing drug-resistant strains of HIV. They do, however, still harbor latent virus in a small number of their T cells—immune cells that are HIV's primary target—despite having undetectable blood levels of HIV. And, in the test tube at least, these viruses can be induced to wake up and begin reproducing, simply by stimulating the T cells to become immunologically active—a condition known to be required for HIV to replicate.

    Although several research groups had previously demonstrated that HIV was still lurking in these cells, some scientists had speculated that it might exist in a damaged, nonviable form. But the new findings show that these viruses are fully capable of replicating and infecting other cells. Moreover, even in patients who adhered rigidly to their drug regimen for up to 30 months, the percentage of latently infected cells did not decrease significantly. This is not good news for hopes that combination therapies would be able to eradicate the virus quickly: Either the current drug regimens may take many years to eliminate HIV totally, or they are not powerful enough to do so. “The idea that the drugs can hit every infected site in the body is unrealistic,” says Simon Wain-Hobson of the Pasteur Institute in Paris.

    On the other hand, AIDS researchers are quick to stress the positive side of the results: The sequences of the viral genomes in the patients' T cells showed little evidence of mutational changes over the course of their treatment, as would be expected if drug-resistant strains were emerging. “The drugs are doing their job, but the reservoir is quite slow in its turnover,” says David Ho, director of the Aaron Diamond AIDS Research Center in New York City. Ho, who with Richman is a co-author on the Siliciano paper, adds that the next task will be to “think of strategies to flush out this cellular compartment.” And Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases in Bethesda, Maryland, says the results “should not have any impact on what we recommend to patients. We are reaping enormous benefits for patients by keeping the virus as low as possible for as long as possible.”

    Despite this optimism, new work from Fauci's laboratory—soon to be published in the Proceedings of the National Academy of Sciences—will sound yet another cautionary note. Fauci's group found the same hidden reservoirs as the Siliciano and Richman teams, but also uncovered evidence suggesting that some of the hidden virus may not be latent, but still replicating at a slow rate.

    The Siliciano and Richman teams adopted the same basic approach in looking for HIV in so-called memory T cells, which help lead the attack when the body encounters microbial invaders it has seen before. The researchers took blood cells from HIV-positive patients who were on a strict regimen of combination therapy and cultured them together with blood cells from HIV-negative donors, along with reagents that trigger memory T cells to become immunologically activated. The researchers observed virus from latently infected memory cells quickly replicating and infecting the HIV-negative cells, even though the original level of infection of the HIV-positive cells was very low—the Siliciano team measured no more than 16 infected cells per 1 million T cells.

    Researchers say these results clearly indicate that it is much too early to consider taking patients off combination therapy. “The results are not surprising,” says retrovirologist John Coffin of Tufts University in Medford, Massachusetts. “We knew that patients over this time frame would become virus positive if treatment is removed. The virus has to still be somewhere.” Adds Wain-Hobson: “We must assume that HIV infection is forever until we know to the contrary.”

    Researchers are far from sure how long infected memory T cells will live before finally dying out, perhaps taking the virus into oblivion with them. Several years ago, immunologists Angela McLean of Oxford University and Colin Michie at London's Ealing Hospital showed that while the average memory T cell lives for about 200 days, some individual cells could survive for many years. “If these cells are left on their own, they are going to last at least a decade, or maybe two or more,” says Richman. But leaving these cells on their own is the last thing AIDS researchers intend to do. “We must hasten the demise of these cells,” says Coffin.

    Fauci suggests that some patients, particularly those treated early in their infection, might be able to mount an effective immune response against the few infected memory cells once they are taken off combination therapy—especially if their immune systems could be boosted with an anti-HIV vaccine. Another approach, Fauci says, would be to develop new drugs against HIV enzymes called integrases—which allow the virus to take over the genetic machinery of its target cells—hence stopping the formation of hidden HIV reservoirs. Says Ho: “These new results tell us what we must do in the upcoming year. They shed light on the path to eliminating the virus from an infected person, the ultimate goal for those of us working on HIV therapeutics.”


    Receptor Offers Clues to How 'Good' Cholesterol Works

    1. Larry Husten
    1. Larry Husten is a science and health writer in New York City.

    Just about every health-conscious person knows by now that having lots of HDL cholesterol—the so-called good cholesterol—in the blood is just as important as having low levels of LDL cholesterol. LDL (low density lipoprotein), of course, predisposes to atherosclerosis and the problems it causes, including heart attacks and strokes, while HDL (high density lipoprotein) is protective. Exactly how HDL protects the arteries has been a mystery, but investigators now have their hands on a key to the answer: the receptor that enables cells to capture cholesterol from HDL particles in the bloodstream.

    The leading theory of how HDL safeguards arteries is that it somehow removes excess cholesterol from blood and tissue, including the cholesterol-loaded cells of atherosclerotic plaques, then carries the excess through the bloodstream to the liver and other tissues. They take in the cholesterol from HDL and use it to synthesize other substances such as steroid hormones and bile acids. But actually proving this scenario was difficult, says cell and molecular biologist Monty Krieger of the Massachusetts Institute of Technology (MIT): “What has been missing is an HDL receptor to give a molecular and cellular handle on the HDL system.”

    Now, in the 11 November issue of the Proceedings of the National Academy of Sciences (PNAS), the Krieger team provides the best evidence yet that a receptor identified during other studies is the key to HDL transport. The clinching evidence came when they knocked out the mouse gene encoding the molecule, designated SR-BI. They found—as expected—that the mice's blood cholesterol levels increased dramatically, while concentrations in organs that pick up cholesterol from HDL, such as the steroid-producing adrenal gland, dropped.

    With the HDL receptor positively identified, investigators should be able to study just how HDL protects against cholesterol deposition in atherosclerotic lesions and perhaps develop drugs that boost the protective effect. “This is a crucial step in obtaining a complete understanding of cholesterol transport,” says Michael Brown of the University of Texas (UT) Southwestern Medical Center in Dallas, who shared the 1985 Nobel Prize in medicine with his longtime Texas colleague Joseph Goldstein for discoveries that revealed how LDL helps cause atherosclerosis.

    Krieger and his colleagues didn't set out to find the HDL receptor. In their original work, which began in the mid-1980s, they were looking for so-called scavenger receptors, which macrophages use to vacuum up modified lipoproteins and other remnants of cells damaged by infection or disease. In 1994, they cloned the gene for what appeared to be the first of a new class of lipoprotein scavenger receptors, which they called SR-BI (for scavenger receptor BI).

    But it soon became clear that SR-BI wasn't just another scavenger receptor. First, the researchers were surprised to find that it binds LDL. And when they then tested SR-BI binding to other lipoproteins, they found that it could bind HDL tightly, which raised the possibility that SR-BI might be the long-sought receptor that enables HDL to deliver its cholesterol load to liver and steroid-synthesizing cells.

    Further studies supported that idea. When the team transferred the gene for SR-BI into cells that don't readily take up cholesterol from HDL, they found that the cells acquired that ability. What's more, the gene is expressed primarily in the right places for the proposed function, including the liver and glands that produce steroid hormones, such as the ovary and adrenals (Science, 26 January 1996, pp. 460 and 518). And in research reported in the 22 May issue of Nature, a group led by Krieger and one of his former graduate students, Karen Kozarsky, a gene therapy specialist at the University of Pennsylvania Medical Center in Philadelphia, used a modified adenovirus carrying the SR-BI gene to induce overexpression of the receptor in the liver cells of mice. As a result, nearly all circulating HDL disappeared from the animals' bloodstreams, while the concentration of cholesterol in their bile doubled.

    But the gold standard for proving the function of a gene is the knockout mouse. SR-BI has passed this test. In their PNAS paper, Krieger and MIT colleagues Attilio Rigotti, Bernardo Trigatti, Marsha Penman, and Helen Rayburn, along with UT Southwestern researcher Joachim Herz, report that they've inactivated the SR-BI gene in mice and found that plasma cholesterol levels in the knockout animals more than doubled. Indeed, knocking out the gene had approximately the same effect on total serum cholesterol levels as a known contributor to atherosclerosis: loss of the LDL receptor, which causes blood cholesterol concentrations to soar.

    At the same time, the knockout animals showed a dramatic drop in adrenal cholesterol, indicating that those tissues were being starved of HDL cholesterol. “Now it is doubly clear that this is the receptor” that takes up HDL cholesterol, says cholesterol researcher Daniel Steinberg of the University of California, San Diego.

    The next big question facing researchers concerns whether SR-BI is as important for cholesterol transport in humans as it is in mice. Researchers will also want to know whether defects in SR-BI contribute to the development of human atherosclerosis.

    But even if SR-BI is not directly involved in causing atherosclerosis, Krieger thinks that the receptor might still be a target for drugs to head off the condition. He notes that the work of Brown and Goldstein led directly to the enormously successful “statins,” which lower bad cholesterol by increasing LDL receptor levels in the liver. SR-BI could also be manipulated, he suggests, to make good cholesterol levels even better.


    Lung Fossils Suggest Dinos Breathed in Cold Blood

    1. Ann Gibbons

    When John Ruben first laid eyes on a high-quality photo of the so-called “feathered” dinosaur from China last year, he was stunned. It wasn't the featherlike structures that riveted his attention—he dismissed them as collagen fibers (see sidebar)—but the theropod dinosaur's innards, which were outlined in the slab of stone. “My eyes popped out,” recalls Ruben, a respiratory physiology expert at Oregon State University in Corvallis. “I realized that here was the first evidence in the soft tissue that theropods had the same kind of compartmentalization of lungs, liver, and intestines that you would find in a crocodile”—and not in a bird.

    Short of breath.

    Ruben (left) says dino lungs were inefficient.


    To prove that notion, Ruben and his graduate students sectioned crocodiles and other reptiles and found that their lung structures resembled the images of several flattened fossil dinosaurs from China. On page 1267, Ruben uses this lung evidence to argue not only that dinosaurs were incapable of the high rates of gas exchange needed for warm-bloodedness, but also that their bellowslike lungs could not have evolved into the high-performance lungs of modern birds. Thus, he challenges two of the reigning hypotheses concerning dinosaurs: that they were warm-blooded and that they gave rise to birds.

    Coming hot on the heels of another controversial paper that concludes that digits in bird wings could not have developed from dinosaur forelimbs (Science, 24 October, p. 666), Ruben's report is part of a “one-two punch to the dinosaur origins of birds hypothesis,” says paleontologist James Farlow of Indiana University-Purdue University in Fort Wayne. But while many dinosaur experts say they welcome Ruben's novel approach, few are willing to embrace his conclusions so far. “This is exactly the kind of research we need,” says Lawrence Witmer, an evolutionary biologist at Ohio University College of Osteopathic Medicine in Athens. And it's definitely weakening the case for warm-blooded dinosaurs. But many researchers, including Farlow and Witmer, think there's persuasive evidence that birds are descendants of dinosaurs. Says Farlow: “[This] is like a breath of fresh air, but it's going to ruffle a lot of feathers.”

    To test whether dinosaurs were really endotherms—warm-blooded animals able to generate their own heat—Ruben and graduate students Terry Jones and Nick Geist have sought to identify the signatures of endothermy, such as a scroll-like structure in the nose, in the bones of living animals. They have argued that dinosaurs lack such structures (Science, 30 August 1996, p. 1204). But what they really needed was improbable—a look at a dinosaur's lungs to see if they were efficient enough to power a warm-blooded animal.

    The improbable happened last year, however, when Ruben saw photos of several specimens of Sinosauropteryx, a small, meat-eating dinosaur from the 120-million-year-old Yixian formation in northeastern China. The fine silt from an ancient lake preserved the animals' soft structures, including a clear “silhouette of the lungs” of one dinosaur, says paleontologist Larry Martin of the University of Kansas, Lawrence, who has seen the fossils.

    When Ruben looked at the photos, it was “immediately apparent” to him that the dinosaur's lungs were arranged in a way that closely matched that of crocodiles. The theropods had two major cavities—the thoracic cavity containing the lungs, liver, and heart; and the abdominal cavity containing intestines and other organs. These were completely separated from each other by the diaphragm, as is the case in crocodiles. Birds have no such separation.

    In living crocodilians, the function of this separation is to provide an airtight seal between the cavities. Then, when the diaphragmatic muscles contract, they pull back the liver and create negative pressure in the thoracic cavity, allowing air to fill the bellows-type lungs. Birds don't need such a separation between the cavities, because air in their lungs moves one way through millions of tiny air passages, drawn by the expansion and contraction of air sacs throughout their bodies.

    Birds' flow-through lung system has plenty of surface area and is especially efficient at exchanging oxygen for carbon dioxide. (Mammals have yet another system that allows efficient gas exchange.) The bellowslike reptilian lung, however, provides much less area for gas exchange, and reptiles cannot absorb oxygen at the high rates needed to sustain intense activity. Ruben also showed that theropods and crocodiles share a distinct hip structure, linked to muscles that help bring air into the bellowslike lungs. All in all, says Ruben, it's “pretty solid evidence that theropods could not have had a modern, high-performance avian-style lung … and were stuck with an unmodified, bellowslike lung.” Says Martin: “Support for the hot-blooded dinosaur hypothesis now has the rigidity of a marshmallow.” The evolutionary implications are even more far-reaching. Ruben argues that a transition from a crocodilian to a bird lung would be impossible, because the transitional animal would have a life-threatening hernia or hole in its diaphragm. “There may well be a relationship between dinosaurs and birds, but it's not the linear relationship you see in museum displays,” he says.

    Ruben's analysis is “another nail in the coffin of the warm-blooded dinosaur theory,” says paleontologist Peter Dodson of the University of Pennsylvania, Philadelphia. But many other researchers say his case is not airtight. They point out that Ruben relied on photos showing a lung outline that is little more than “smudges on rock,” says Witmer. What's more, Ruben's inferences are based on a flattened, two-dimensional fossil. “You would expect some deformation when the organs squish out,” says Witmer, who suggests, only half-jokingly, that Ruben flatten his alligators with a steamroller for comparison. And the evolutionary transition from the actual theropod lung, rather than the modern crocodilian analog, might be easier.

    Indeed, even if Ruben's analysis of lung structure holds up, it would have to be weighed against “a mountain” of other evidence supporting the dinosaurian origin of birds, says Farlow. Still, he finds Ruben's findings of a crocodilian-type lung for theropods “compelling.” Fortunately, the Yixian formation is so rich in fossils that more specimens of Sinosauropteryx are likely to turn up. And if the same lung structure appears in enough fossils, Ruben's case will gather considerable weight.


    Plucking the Feathered Dinosaur

    1. Ann Gibbons

    Exactly 1 year ago, paleontologists were abuzz about photos of a so-called “feathered dinosaur,” which were passed around the halls at the annual meeting of the Society of Vertebrate Paleontology (Science, 1 November 1996, p. 720). The Sinosauropteryx specimen from the Yixian Formation in China made the front page of The New York Times, and was viewed by some as confirming the dinosaurian origins of birds. But at this year's vertebrate paleontology meeting in Chicago late last month, the verdict was a bit different: The structures are not modern feathers, say the roughly half-dozen Western paleontologists who have seen the specimens.

    Feathered friend?

    Collagen fibers in a sea snake's tail resemble feathers.


    The stiff, bristlelike fibers that outline the fossils lack the detailed organization seen in modern feathers, says Alan Brush, an ornithologist at the University of Connecticut, Storrs, who specializes in feather structure. Brush was part of a “dream team” sent to China this spring by The Academy of Sciences in Philadelphia to view the fossils.

    But just what the structures are—and whether they link birds and dinosaurs—is still under debate. Noting that the outline of the dinosaur skin is hard to discern in the fossilized stone, another dream team member, paleontologist Larry Martin of Kansas University, Lawrence, thinks the structures are frayed collagenous fibers beneath the skin—and so have nothing to do with birds. Zoologist John Ruben of Oregon State University in Corvallis dissected a sea snake's tail to show that such fibers can indeed look feathery (see photo). Others, including Brush and Philip Currie, a paleontologist at the Royal Tyrrell Museum of Palaeontology in Drumheller, Canada, describe the bristlelike fibers as “protofeathers”—fibers that may be hollow and made of the same kind of keratin as feathers.

    Meanwhile, Ji Qiang, director of the Chinese Geology Museum in Beijing, insists that the fibers are “obvious primitive feathers.” But a paper in press at Nature by another group of Chinese researchers doesn't make that claim, says Currie. Measuring the width of the fibers under a scanning electron microscope or testing whether they're made of collagen or keratin could resolve the debate. Some of these tests are under way, Currie adds.


    Galactic Disk Contains No Dark Matter

    1. Alexander Hellemans
    1. Alexander Hellemans is a writer in Naples, Italy.

    By studying the movement of stars in the disk of our Milky Way galaxy, two teams of French astronomers have concluded that what you see is what you get: The mass of the visible stars appears to account for all the material in the galactic disk. These findings, derived from data gathered by the European astrometric satellite Hipparcos, imply that the main body of our galaxy contains no “dark matter”—invisible material that astronomers believe accounts for up to 90% of the mass of the universe. “These studies confirm that the dark matter [presumed to be] associated with the galactic disc in fact doesn't exist,” says Honc-Anh Pham of the Paris Observatory at Meudon, whose doctoral thesis forms one of two studies that came to this conclusion. Instead, both groups argue, the dark matter must be lurking in the galactic halo, a large, spherical region encircling the galaxy containing dust, gas, and globular clusters of very old stars.

    No more than meets the eye.

    Visible stars seem to account for all the mass in the galactic disk.


    Pham studied the movements of 10,000 stars to get a fix on the gravitational forces pulling them around. She inferred from these movements that the local mass density in our galaxy is 0.11 solar masses per cubic parsec. (A parsec corresponds to 3.26 light-years.) A separate team, led by Michel Crézé of Strasbourg Observatory, reports in a forthcoming issue of Astronomy & Astrophysics that from a study of a smaller group of 100 stars they found an even lower value, 0.076 solar masses per cubic parsec. These values are close to estimates of the mass density of visible stars in the galactic disk and leave little room for dark matter. The new results confirm some earlier estimates, made before Hipparcos data became available, but they are much lower than values obtained by John Bahcall of the Institute for Advanced Study in Princeton, New Jersey, using velocity data from ground-based astrometric observations. He concluded that the local galactic density is between 0.15 and 0.20, enough to accommodate 30% to 50% dark matter.

    Astronomers have long surmised that dark matter provides some of the gravitational glue required to hold galaxies together: Most galaxies rotate so fast that they would fly apart if their visible stars provide the only sources of gravity. The stars in galaxies also orbit in a peculiar fashion: Unlike planets in the solar system, stars in the outer reaches of galaxies move as fast as those nearer the center. This suggests that the galaxies' mass must be spread out and not concentrated in the core, as it is in the solar system.

    Astronomers can estimate the total mass in a galaxy like our own from the forces needed to hold it together, but without an accurate knowledge of the mass density of the galactic disk, they could not tell how much of the dark matter resides in the disk and how much in the halo, Crézé says.

    Crézé and his Strasbourg colleagues Emmanuel Chereuil and Olivier Bienaymé, and Christophe Pichon of the University of Basel in Switzerland, looked at the Hipparcos data for 100 stars in a sphere of radius 125 parsecs around the sun. Hipparcos, launched in August 1989, cataloged over 4 years the precise position and motions of more than 100,000 stars. The team analyzed the distribution of the motions of their sample of “tracer” stars in the direction perpendicular to the galactic disk to assess the amount of gravitational pull dragging them back toward the galactic plane. Herwig Dejonghe of the University of Ghent in Belgium compares their method to “looking at a sample of high jumpers and deducing the mass of the Earth from the height they reach.”

    Pham's approach was somewhat different: Her larger sample within a sphere with radius 250 parsecs consisted of just one type of star, known as F-type, which are old and so have dissipated some of the motion associated with their births in swirling clouds of gas. From their distance from the galactic plane and their proper motions, she obtained her value of the local galactic density.

    The results were welcomed by Michael Merrifield of Britain's Southampton University, who with his colleague Robert Olling has argued from observations of the shape of the galactic disk, that all the dark matter in the galaxy should be found in a round halo. “We actually have run calculations … our [galactic] model with the round halo corresponds exactly to the kind of numbers they get,” Merrifield says.

Stay Connected to Science