News this Week

Science  22 May 1998:
Vol. 280, Issue 5367, pp. 1191

    Building a Better Aspirin

    1. Elizabeth Pennisi

    New aspirin-like compounds target a single enzyme to deliver pain relief without stomach and kidney damage. They may also slow the development of cancer and Alzheimer's disease

    A century-old wonder drug is about to get even better. Aspirin, first introduced in 1899, relieves pain and soothes aching joints and muscles. But this comfort comes at a cost: Aspirin and all the aspirin-like products, called nonsteroidal anti-inflammatory drugs (NSAIDs), can eat away the stomach lining, causing bleeding or ulcers, and damage the kidneys.

    Because these drugs constitute a $14 billion market, pharmaceutical companies have long sought anti-inflammatory compounds without these side effects. The problem with aspirin and its cousins, such as ibuprofen and acetaminophen, is that they aren't sufficiently specific. Their beneficial effects come from their ability to block an enzyme, called cyclooxygenase-2 (COX-2), that promotes inflammation, pain, and fevers. Unfortunately, the drugs are even more effective at inhibiting COX-1, a related enzyme essential for the health of the stomach and kidney. So researchers have been hard at work coming up with compounds that selectively inhibit COX-2, and two may be in drugstores within the next year. But a new wave of COX-2 inhibitors that could eventually be even more potent is also taking shape.

    Double action.

    NSAIDs prevent both COX enzymes from turning arachidonic acid from the cell membrane into prostaglandin hormones; COX-2 inhibitors are choosier.


    Unlike aspirin, which permanently inactivates the COX enzymes, the drugs in the first wave of inhibitors have only temporary effects. But on page 1268, biochemist Lawrence Marnett and his colleagues at Vanderbilt University School of Medicine in Nashville, Tennessee, report that they have developed compounds that chemically disarm COX-2—and only COX-2—instead of just blocking its activity, as the existing COX-2 inhibitors do. So far, these first irreversible COX-2 inhibitors are no more effective than the compounds now poised to enter the market. But they may be the vanguard of “the next generation” of these wonder drugs, says Stephen Prescott, a molecular biologist at the University of Utah, Salt Lake City. “Eventually all those [NSAID] molecules on the market will become dinosaurs,” predicts Philip Portoghese, a medicinal chemist at the University of Minnesota, Minneapolis.

    The market for these compounds could be huge, because they are making their debut just as the demand for anti-inflammatory drugs appears ready to burgeon. Over the past decade, epidemiological data have indicated that aspirin and NSAIDs can protect against certain cancers and Alzheimer's disease, and recent laboratory results suggest that COX-2 inhibition is a key factor in these effects (see sidebar). Preventing these diseases would require taking NSAIDs long-term—and selective COX-2 inhibitors seem tailor-made for such use because they appear safer than traditional NSAIDs. Their only apparent drawback is that, unlike aspirin, they can't lower the risk of cardiovascular disease, because that requires inhibiting the clot-promoting effects of the COX-1 enzyme.

    Enzyme busters.

    Altering its chemistry makes the new COX-2 inhibitor (right) more selective than aspirin (left).


    The search for selective COX-2 inhibitors began in 1991, when researchers first learned of the two COX enzymes. Both enzymes help produce hormones called prostaglandins, although COX-1 is present throughout the body, while COX-2 is made only under certain conditions. But researchers found that only the prostaglandins made by COX-2 lead to inflammation, pain, and fever. COX-1 primarily makes hormones that help keep the stomach lining intact and the kidneys functioning properly.

    Efforts to find COX-2 inhibitors heated up in 1996 when several teams determined the crystal structure of COX-2, information that has helped guide drug design (Science, 20 September 1996, p. 1660). So far, chemists have found or created about a dozen candidate drugs that block COX-2 alone; another compound that inhibits both enzymes but acts preferentially on COX-2 is already sold in Europe. All those drugs work differently from aspirin in that they bind temporarily to COX-2, thereby blocking its ability to generate prostaglandin. But once the drugs fall off, the enzyme becomes active again. Aspirin, in contrast, transfers an acetyl side group to the COX, permanently disabling it. “It's dead forever,” says Prescott. Thus prostaglandin production can resume only after the body produces more of these enzymes.

    At Vanderbilt, Marnett and his colleagues set out to make COX-2 blockers that would work in the same way as aspirin. To do this, Amit Kalgutkar in Marnett's group built a molecule that retained aspirin's original acetyl side group but also had a sulfur-containing group—a component of earlier COX-2 inhibitors that helps make those compounds specific for that enzyme. Getting the sulfur group right proved troublesome, however. After several dead ends, Kalgutkar added a sulfur to a ringed molecule and grafted the result to the acetyl group. When he tested this molecule against COX enzymes in a test tube, he found that it preferentially inactivated COX-2, although not as efficiently as he and Marnett would have liked.

    Nevertheless, “it proved to us that it could be done,” Marnett recalls. The researchers ultimately created a more potent COX-2 inhibitor by adding a tail to the ring, consisting of seven carbons, two of them joined with a triple bond. After building about 70 variations on this chemical theme, they found that the best compounds reduce COX-2 activity by as much as existing, reversible COX-2 inhibitors.

    “They've done quite a good job showing it works very well inhibiting COX-2,” comments Sir John Robert Vane, a pharmacologist and Nobel laureate at the William Harvey Research Institute in London. Marnett, whose university has patented this class of COX-2 inhibitors, sees the compounds as just a “proof of concept.” He hopes to team up with a pharmaceutical company to come up with an irreversible drug that binds more avidly to COX-2, making it even more effective at knocking out the enzyme.

    Meanwhile, at least two companies, Monsanto Corp. in St. Louis and Merck and Co., based in Whitehouse Station, New Jersey, are well on their way to getting their reversible COX-2 inhibitors to market. Recently, Monsanto finished a study comparing its drug, celecoxib, with existing NSAIDs in 12,000 people with arthritis. “It's fully as efficacious as the NSAIDs” without injuring the gut, says Philip Needleman, a pharmacologist at Monsanto. Merck's product, vioxx, also did well against arthritis and pain in early trials, says Merck's Barry Gertz.

    But although the new COX-2 inhibitors seem safe, “the potential long-term adverse consequences are not known,” warns John Breitner, an epidemiologist at the Johns Hopkins School of Public Health in Baltimore. He notes that because the drugs seem so safe, people are likely to use them at higher doses for much longer than they would aspirin, with its known risks. The irreversible inhibitors raise an additional concern about how the body might react over the long haul to chemically modified COX-2. Nonetheless, says neuroscientist Nicolas Bazan of the Louisiana State University Medical Center in New Orleans, the search for superaspirins “is a very exciting area of research, one that could benefit many areas of medicine.”


    Does Aspirin Ward Off Cancer and Alzheimer's?

    1. Elizabeth Pennisi

    The new “superaspirins” that are now emerging from the lab (see main text) are likely to have a warm welcome when they reach the market. They appear to offer most of the old benefits of aspirin and its relatives—combating pain, fever, and inflammation—without their stomach-ravaging side effects. And now it seems that both groups of aspirin-like drugs may offer some dramatic new benefits: slowing the progression of cancer and Alzheimer's disease.

    The first hints of these benefits came from epidemiological studies, which showed a 45% decrease in deaths from colon cancer and a much reduced incidence of Alzheimer's disease in people regularly taking aspirin or nonsteroidal anti-inflammatory drugs (NSAIDs) for arthritis or other conditions (Science, 5 July 1996, p. 50). A clue to why NSAIDs seem to inhibit cancer came from biochemist Raymond DuBois and cancer surgeon R. Daniel Beauchamp of Vanderbilt University in Nashville, Tennessee, in 1994. They found high levels of COX-2, an enzyme inhibited by aspirin and its relatives, in about 90% of the colon cancer tumors they examined, a finding that several other studies have confirmed. Monsanto Corp.'s Karen Seibert, a pharmacologist, reports that the COX-2 gene is also overactive in some squamous cell tumors, a skin cancer, and there are slight hints that it plays a role in breast cancer as well.

    In 1996, DuBois and his colleagues found a possible role for COX-2 in cancer. When the researchers used genetically engineered cells that made greater than normal amounts of COX-2, the cells became less susceptible to programmed cell death, or apoptosis, which normally removes mutated or damaged cells. “If you look in colon cancer tissue, you find it's blazing with [active] COX-2 gene and protein, and that it's lost programmed cell death,” pharmacologist Philip Needleman of Monsanto points out.

    Stephen Prescott of the University of Utah, Salt Lake City, thinks COX-2 might also cause trouble because it releases free radicals—such as reactive oxygen ions—as it performs its normal function, generating the hormones called prostaglandins. Because free radicals can cause mutations, they might increase the likelihood of the genetic changes that turn a cell cancerous. And DuBois's team has unpublished data pointing to a third possible mechanism. COX-2, they found, promotes the production of factors that encourage new blood vessels to grow into a tumor, giving it the nourishment it needs to grow.

    Why COX-2 inhibitors might protect people against Alzheimer's is less certain. Some researchers think that these drugs simply protect nerve cells against inflammation associated with amyloid plaques, the protein deposits found in the brains of people with Alzheimer's. The drugs might also lower the production of free radicals, suspected of having a role in Alzheimer's brain damage. Others surmise that COX-2 is part of a signaling pathway in nerve cells that—in contrast to the protein's effect on cancer cells—somehow causes apoptosis. In that case, inhibiting the enzyme might spare nerve cells.

    Neuroscientist Nicolas Bazan and his colleagues at the Louisiana State University Medical Center in New Orleans have evidence for that scenario. He found that nerve cell injury can turn on the COX-2 gene, increasing the level of COX-2 protein and accelerating apoptosis. When Bazan chemically shut down the COX-2 gene, reducing the amount of COX-2 in cells, “the amount of apoptosis [was] less,” he says.

    Uncertainties about the exact roles of COX-2 in Alzheimer's and cancer have not stopped researchers from exploring the promise of both traditional NSAIDs and the new COX-2 inhibitors in both conditions. At the American Health Foundation in Valhalla, New York, for example, Bandaru Reddy and his colleagues have shown that rats given celecoxib, a COX-2 inhibitor developed by Monsanto, developed 90% fewer tumors after receiving a chemical carcinogen than control rats did. The tumors that did develop in the treated rats were also less virulent, Reddy and his colleagues reported in the February issue of Cancer Research. Reddy's data “are incredibly striking,” says Ronald Lubet, a biochemist at the National Cancer Institute (NCI).

    Last month, gastroenterologist Russell Jacoby and clinical researcher Carolyn Cole at the University of Wisconsin, Madison, reported other promising results at the annual meeting of the American Association for Cancer Research in New Orleans. Jacoby's group mixed celecoxib with food given to mice bred to develop colon cancer spontaneously. After about 3 months, the mice had only one-third as many tumors as untreated mice, and those that did develop “were almost indistinguishable” from normal gut lining, he reports.

    Monsanto is now sponsoring human studies, including one, done in conjunction with the NCI, that is assessing whether celecoxib slows the development of precancerous growths called polyps in people genetically disposed to getting colon cancer. “By next year at this time, we'll know,” says Needleman.

    The company is also planning to study whether selective COX-2 inhibitors can keep people with Alzheimer's from getting worse, and other studies are planned to see whether the same is true for people who are suspected of having Alzheimer's but who are not yet sick enough to be diagnosed. Again, many researchers are optimistic. “I'd be willing to bet the mortgage that COX-2 [will be involved] in Alzheimer's disease,” says Prescott.


    The Fastest Counter of the Smallest Beans

    1. Robert F. Service

    Imagine a device that could nestle close to a current-carrying wire and eavesdrop on individual electrons as they speed through it. The makings of such a device exist already: transistors so small that they deal in individual electrons. Such single-electron devices could serve as the heart of futuristic computer chips (Science, 17 January 1997, p. 303). But they could also serve as the most sensitive and accurate detectors of electrical current around, because the influence of individual charges passing through a nearby wire is enough to switch them on and off. So far, however, investigators haven't been able to exploit that promise, because they have lacked an amplifier fast enough to capture the rapid-fire signals from the transistor.

    Now a team of U.S. and Swedish researchers has developed a new single-electron transistor (SET) architecture that includes an amplifier capable of recording the passage of electrons 1000 times faster than the previous record holder and about 1 million times faster than conventional SETs. The device, which the team describes on page 1238 of this issue, has drawn some rapid attention of its own, as it's likely to be fast enough to register individual charges in a current. “It's a real breakthrough in terms of speed,” says John Martinis, a physicist at the National Institute of Standards and Technology (NIST) in Boulder, Colorado. Martinis says he and his NIST colleagues are working to define a new common standard for electric current by counting the number of electrons flowing through a device each second. “This certainly will help us,” he says.

    To come up with their high-speed SET, the transatlantic team, led by physicists Rob Schoelkopf of Yale University and Peter Wahlgren of Chalmers University of Technology in Göteborg, Sweden, modified a standard SET design. Conventional SETs are themselves an offshoot of ordinary transistors, which switch the flow of electrons through a semiconducting channel on and off by applying a voltage to a “gate” electrode above it. SETs replace the semiconducting channel with insulating material, except for a tiny semiconducting or metallic island halfway along. Now electrons must hop first through the insulator to the island and then hop to the other side. But if an electron is already on the island, its negative charge will repel subsequent electrons, in essence keeping them where they are.

    A tiny voltage applied to a gate electrode over the island can increase the current through an SET, however. The voltage lowers the repulsive barrier felt by subsequent electrons, allowing them to jump onto the island and the resident electron to jump off. As a result, a stream of electrons passes through the channel, hopping one by one onto the island and off it again.

    The same setup can be used as an electrometer to detect tiny amounts of current flowing through a wire. To do so, researchers simply place a wire next to the SET's channel so that it runs past where the gate would normally be. As electrons whiz down the wire past the gate's usual position, each one creates a tiny voltage increase that acts like voltage on the gate. These minute voltage fluctuations from the wire switch the SET on and off, allowing tiny amounts of current to flow through its channel—each burst of current registering the passage of an electron along the wire. The problem is that the rate at which electrons flow down the wire can be extremely fast, and conventional electronic amplifiers are too slow to register the SET's tiny signals.

    To get around this problem, Schoelkopf and his colleagues connected an SET to a circuit known as a resonant amplifier. While the wire provides the voltage that turns the SET on and off, the resonant amplifier provides the impetus that pushes the electrons along the channel. An electromagnetic field that resonates at microwave frequencies within a circuit at the heart of the amplifier interacts with electrons to move them along. When the SET turns from off to on, this microwave energy pushes electrons through the channel, and the intensity of the microwave field resonating in the circuit drops—a change that is easy, and fast, to measure.

    The resonator is able to detect signals from the SET at a rate of 150 million cycles per second. The group has yet to measure individual electrons flowing down a wire. But Schoelkopf is hopeful: “We can put this on a chip next to something we want to study and watch electrons as they flow by in real time.”


    A Neutron Star That Got Revved Up

    1. James Glanz

    Astronomers have discovered a “missing link” that could explain the formation of the bizarre celestial beacons called millisecond radio pulsars. Neutron stars that spin hundreds of times a second and give off a radio blip with each rotation, millisecond pulsars are thought to acquire their high revs when material torn from a companion star spirals in and applies a twist. New observations made by NASA's Rossi X-ray Timing Explorer (XTE) satellite seem to show just that: a neutron star that has spun up like a top and is spewing out x-rays as infalling material spirals down onto its surface.

    Rudy Wijnands and Michiel van der Klis of the University of Amsterdam analyzed XTE observations of an x-ray beacon that lies perhaps 12,000 light-years away in the direction of the galactic center, captured shortly after the satellite saw the object brightening around 9 April. XTE's detectors, sensitive to rapid x-ray flickers, showed that x-rays from the object slightly dim and brighten every 2.5 milliseconds, or thousandths of a second. The variation probably arises as the “hot spot” from infalling material whirls around with the neutron star.


    X-rays from a neutron star surged in April.


    “The missing link has now been found,” says Frederick Lamb, a theorist at the University of Illinois, Urbana-Champaign. “It's kind of a dream come true.” Wijnands and Van der Klis have reported their result in International Astronomical Union (IAU) Circulars and in scientific talks. Van der Klis declined to comment on it, however, because the team has a paper on the topic under review at Nature.

    Radio pulsars are thought to consist of a spinning neutron star—a collapsed star about 10 kilometers across but more massive than our sun—sprouting a magnetic field about a billion times more intense than Earth's. Interactions between the whirling field and particles in the star's thin atmosphere generate radio waves, which stream like a lighthouse beam from the magnetic poles. Astronomers pick up a pulsar's radio blips each time a magnetic pole whirls past our line of sight, says Jonathan Arons of the University of California, Berkeley.

    Ordinary radio pulsars, which were discovered more than 30 years ago, emit a blip every second or so. But in the early 1980s, astronomers detected a new class of pulsars that spin nearly 1000 times faster. In a widely held theory, says Arons, the stars “get spun up” to such fantastic rates by accreting matter from a companion at an earlier stage of their lives. As the disk of material spirals inward, it applies a torque to the pulsar's magnetic field, revving up the star.

    Bursts of x-rays, emitting from the superheated material that crashes to the surface of the neutron star, should signal the spin-up. But until now, nearly all known x-ray pulsars have had periods longer than a second, probably because they have magnetic fields so powerful that they disrupt the disk of infalling material far from the surface of the star, where the disk is still spinning relatively slowly.

    The rotation rate of the new object, SAX J1808.4-3658, suggests that its field must be weak enough to allow the disk to come within about 20 kilometers of the surface. An earlier sighting supports that assumption. In September 1996, the Italian-Dutch BeppoSAX satellite observed x-rays emitted by thermonuclear explosions on what was apparently the same object. The explosions are thought to occur when pressure builds up within a widespread puddle of accreted material on the surface of a neutron star. Such extensive puddles of unburned material could never form on a star that has a powerful magnetic funnel to confine the infalling material, say astrophysicists. “This object has a spin period and an inferred magnetic field that would most likely allow it to become a millisecond radio pulsar when accretion shuts off,” says Lars Bildsten of Berkeley. “This is the first such example.”

    The rapid-fire x-ray pulses are only one sign of a millisecond pulsar being born. The midwife—a companion star—appears to be present as well. In a second result, reported in IAU circulars and accepted at Nature, Deepto Chakrabarty and Edward Morgan of the Massachusetts Institute of Technology describe slower variations in the x-ray emission, which seem to show that the pulsar is tightly orbiting a companion star. Chakrabarty, along with Paul Roche of the University of Sussex in the United Kingdom and others, may also have seen the companion star directly, through a ground-based telescope.

    Radio astronomers will now watch to see whether a radio pulsar emerges as accretion onto SAX J1808.4-3658, which is now fading, falls off even further. Until then, the infalling material will snuff out the radio bursts. But already, pulsar specialists regard the discovery as a triumph for their theory—and for XTE. “It's one of the things we hoped for from the Rossi mission,” says Lamb, “and it's delivered.”


    Can Great Quakes Extend Their Reach?

    1. Richard A. Kerr

    Earthquakes were once thought to keep to themselves, striking on a schedule determined only by the history of each particular fault. Then seismologists began to realize that every rupturing fault communicates with neighboring faults, instantly reaching out tens or hundreds of kilometers to hasten or delay distant earthquakes (Science, 16 February 1996, p. 910). Now a group of geophysicists suggests that these lines of communication extend even farther—and carry much, much slower messages.

    Big quakes can trigger other quakes thousands of kilometers away and decades later, according to calculations presented on page 1245 of this issue of Science. Geophysicists Fred F. Pollitz and Roland Bürgmann of the University of California, Davis, and seismologist Barbara Romanowicz of UC Berkeley simulated how stress travels through deep, viscous rock. They found that the great earthquakes that struck the far North Pacific in the 1950s and '60s could have set off wave that triggered a pulse of seismic activity in California in the 1980s.

    “It's an exciting possibility,” says seismologist Thomas Hanks of the U.S. Geological Survey (USGS) in Menlo Park, California. If the reach of big quakes extends that far, seismologists may be able to make more sense of the comings and goings of earthquakes worldwide, he adds. Researchers are intrigued, though not yet completely convinced. “It could be right,” says tectonophysicist Wayne Thatcher of the USGS, “but I think it has a ways to go before being a persuasive argument.”

    Researchers have long recognized a potential transmission route for long-distance messages among faults: the thin layer of soft rock at depths of 80 kilometers or more called the asthenosphere. There, temperature and pressure combine to soften rock so that, although still solid, it slowly flows over decades. The more rigid tectonic plates that make up Earth's surface, such as the great Pacific Plate, glide along on this softer layer.

    But plates don't slide smoothly at their edges. They stick to each other, build up stress, and then jerk forward in earthquakes. The quake redistributes stress nearby, adding stress in some places and relieving it in others. For example, between 1952 and 1965, four great quakes struck along the Aleutians and the Kamchatka Peninsula, where the Pacific Plate is diving beneath the North American Plate. After each quake, the Pacific Plate adjusted to the new plate positions immediately, stretching like a sheet of rubber and triggering flow in the asthenosphere below. Spreading outward through the asthenosphere like the ripple of a pebble dropped in a pond, the wave created by this flow could transmit the stress induced by the quakes.

    “That [stress] wave has to exist,” says Bürgmann. “The only question is how strong is it?” To find out, the group created a computer simulation of elastic plates, ductile asthenosphere, and large earthquakes in the northern North Pacific. In the model, the stress wave generated by the quakes moved southward across the Pacific and northward under the Arctic Ocean at a rate that depended on the viscosity of the asthenosphere. When researchers plugged in a viscosity that Romanowicz calls “reasonable but a bit on the low side” of current estimates, the crest of the stress wave entered the eastern Arctic Ocean in the 1970s; it passed off British Columbia around 1975, and California around 1985. Wherever the wave passed, it briefly accelerated plate motions, which could have spurred earthquake activity.

    The timing is a good fit to surges of seismic activity, say Pollitz and his colleagues. According to the model, the wave may have triggered the surge of magnitude 5 and greater quakes observed in the eastern Arctic Basin in the 1980s. To the south, the wave's progress—marked by accelerations of only a couple of millimeters per year—could be seen in pulses of increased seismicity in Northern California in the 1970s and Southern California in the 1980s.

    Even the types of earthquakes seemed to fit stress-wave triggering, says the group. The Southern California seismicity mostly took the form of quakes on faults other than the San Andreas. The sides of these faults move chiefly up and down rather than sideways, as the San Andreas does. That feature of the seismicity was noted in 1995 by seismologists Frank Press of the Washington Advisory Group in Washington, D.C., and Clarence Allen of the California Institute of Technology in Pasadena, who speculated that a stress wave oriented to favor vertical fault motions might be responsible. The wave set off by the great Alaskan quakes fits the bill, Pollitz's team says. “The whole thing seems to hang together,” says Press.

    But others point out that the correlation of the passing wave with a flurry of seismicity could be chance. “There have been many interesting patterns in seismology that have turned out to be wonderful coincidences,” says seismologist Lucile Jones of the USGS in Pasadena. And the stress wave, dampened by distance, seems too weak to trigger quakes, other researchers say. The extra strain added by the wave would be “really small,” says Thatcher, perhaps a factor of 10 smaller than what's assumed to trigger quakes at short range.

    “Yes, the strain changes are small,” concedes Romanowicz, “but I don't think anyone has a definite idea of how much strain you need to trigger an earthquake.” Bürgmann adds that a coincidence is unlikely, because seismicity along the entire west coast fits the stress-wave theory.

    Further tests of this bold idea are in the works. Although existing geodetic networks weren't sensitive enough to catch the subtle stress changes signaling the arrival of the wave, says Bürgmann, current systems should record its departure in the coming decade. Then it should become clear whether distant earthquakes are the interfering busybodies he and his colleagues suspect they are.


    Ecosystem 'Engineers' Shape Habitats for Other Species

    1. Joseph Alper
    1. Joseph Alper is a free-lance writer in Louisville, Colorado.

    In Israel's Negev desert, some microbes have developed a unique survival strategy: They pave the desert. Several species of bacteria and cyanobacteria secrete long-chain sugars that bind soil and sand into a black crust, which protects their damp colonies from the searing heat. But the microbes' labor benefits other species as well, according to Moshe Shachak and Bertrand Boeken of Ben-Gurion University in Sede Boker, Israel. After a downpour, the asphaltlike patches reduce water absorption by about 30%, increasing runoff, which pools in pits dug by desert porcupines and beetles. Windblown seeds germinate in the moist pits, giving rise to lush oases that can harbor dozens of species. “We see enormous effects … by a host of tiny organisms,” says Shachak.

    “Ecosystem engineers” like these microbes have sparked a new approach to assessing how species interact with one another. Shachak, together with ecologist Clive Jones of the Institute of Ecosystem Studies in Millbrook, New York, and John Lawton of Imperial College's Centre for Population Biology in Silwood Park, U.K., have proposed a new concept of how ecosystem engineers, by shaping habitats to their own needs, alter the availability of energy—food, water, or sunlight—and thus dictate the fates of other species.

    The concept has generated quite a stir among environmental scientists since it appeared in the journal Ecology last October. “Nobody had stepped back before and asked if this was a general phenomenon, then tried to put down some guiding principles,” says David Tilman, an ecologist at the University of Minnesota, St. Paul. “This is one of those rare papers that gets you thinking in a new way.” He and others think that after fine-tuning, the concept of ecosystem engineers may be ready to join an elite set of theories, such as natural selection and predator-prey theory, that help explain how species arise and interact.

    Missing from ecology's theoretical underpinnings has been a way to account for how species, by altering habitats, perturb other species—even though, as Jones explains, “we've known for a long time that there are things species do to their physical environment that have enormous knock-on effects throughout the ecosystem.” Then a few years ago, he and Lawton heard about the Negev story. The scientists soon grasped that ecosystem engineering was far more pervasive than humans erecting skyscrapers or beavers building dams. “Once we started looking in the literature and talking to people about this,” says Jones, he and his colleagues realized “how important ecosystem engineers are at affecting species diversity, distribution, and survival.”

    The concept's guiding principle is that engineers indirectly control the flow of energy within an ecosystem. These species, the ecologists say, can have just as great an influence on an ecosystem as keystone species, or top predators. The concept holds that ecosystem engineers alter habitats through two overarching mechanisms. Autogenic engineers transform ecosystems by their own growth and are integral to the altered environment. Corals, for example, build reefs for their own needs that also serve countless other species. Although some species feed on coral, most, including brittle stars, anemones, and sponges, use reefs only for shelter. Similarly, trees create habitat for myriad species that live in and among tree crotches, where large branches diverge from trunks. Without coral reefs or trees, says Jones, associated species would perish.

    The second class of organisms, allogenic engineers, alter the environment and then move on, leaving structures behind. Beavers, for instance, turn stream ecosystems into pond ecosystems by building dams that block stream flow. The pooling water drowns grasses and shrubs but provides marsh for herons and other species; crustaceans colonize debris from beaver dams. The Negev bacteria are also allogenic engineers.

    The ecologists list six factors—including population density of an engineering species and the types of resources it controls—to help assess an engineer's importance to an ecosystem. The researchers hope that this framework can be used to make predictions about how, for instance, engineers that invade an ecosystem might alter it.

    Researchers are already putting the concept to the test. Entomologist Bob Marquis and grad student John Lill of the University of Missouri, St. Louis, are studying how Pseudotelphusa caterpillars tie oak leaves together to form shelters. They have found that dozens of species—including spiders, weevils, and aphids—dwell in the shelters. By forcing researchers to look for those species that indirectly alter energy availability, the engineer concept “could help organize a great deal of what we're seeing in our experimental systems,” says Marquis. Indeed, he says, it has prompted him and Lill to revise their research plan. Instead of merely observing engineers in action, says Marquis, “we are going to manipulate the leaves ourselves to quantify the effects of the leaf ties on the resulting ecosystems.”

    Others hope to put the concept to predictive use. Lawton and mathematician William Gurney of the University of Strathclyde in Glasgow, U.K., are trying to devise robust computer models that forecast how an engineer's activities could affect other species. “Experiments are now getting started,” says Lawton, “but it will probably be a decade before we can really say what shape the models, and ultimately the theory, will take.” Such models could someday be useful for protecting or restoring habitats. “It's hard to think about conserving ecosystems without considering the effects that engineers have on a system,” says Shachak.

    Experts agree that the nascent concept needs sharpening to help researchers home in on the engineers that, like keystone species, are crucial to an ecosystem's overall health. “At some level you could say that every organism is engineering its ecosystem and that this activity affects other organisms,” says Alex Flecker, an ecologist at Cornell University in Ithaca, New York, who studies the ecosystem effects of fish that bulldoze sediments to find food in Andean streams. But “the important thing,” Flecker says, is that the new concept has “organized the different types of engineering behavior we see in the field into a useful, testable framework.”


    Antibodies Stage a Comeback in Cancer Treatment

    1. Steven Dickman
    1. Steven Dickman is a free-lance writer in Cambridge, Massachusetts.

    As investors in new cancer therapies ought to know, the history of cancer research is rife with reports of cancer “cures” that all too often turn out to be ephemeral. In 1982, for example, immunologist Ron Levy of Stanford Medical Center raised the hopes of many when he reported in The New England Journal of Medicine that he had vanquished cancer in a patient, Philip Karr, using antibodies custom-designed to attack Karr's own lymphoma cells. In the wake of the resulting optimism, Levy co-founded IDEC Pharmaceuticals to commercialize his discovery, and other white-hot biotech companies pounced on the idea too. Expectations—and stock values—soared. “It was assumed [that antibodies] would be the final answer, that we could just produce them and the rest of cancer research could close up shop,” recalls radiation oncologist Alan Lichter of the University of Michigan Medical School in Ann Arbor, who is president of the American Society of Clinical Oncology (ASCO).

    Then, in an object lesson in the dangers of hyping cancer therapies, hopes—and stock values—shriveled. Although Levy's antibody worked, the effects of other antibodies in humans didn't match those in mice, and unexpected toxicity even killed patients, bringing clinical trials to an abrupt halt. Antibodies vanished from page one, and many firms abandoned them.

    But now, after a decade and a half of hard work, the tide may be turning again. Last fall, IDEC of San Diego finally received approval from the U.S. Food and Drug Administration (FDA) for an antilymphoma antibody, a cousin to Levy's original preparation. Just last week, researchers announced some success with an antibody tailored to fight recalcitrant breast cancers that is nearing regulatory approval. A handful of antibodies are in earlier stage clinical trials, with a smattering of positive results. And dozens more are in preclinical testing around the world. “We're entering a period of cautious optimism,” says tumor immunologist Lloyd Old, director of the Ludwig Institute for Cancer Research in New York and co-organizer of an antibody meeting held in Manhattan last month.* Akhtar Samad, an analyst with the New York-based Mehta Partners, agrees: “We're in the early stages of renewed investor interest and confidence.”

    Researchers caution, though, that antibodies aren't the “magic bullets” hyped in the past, nor will they ever replace conventional cancer chemotherapy drugs. Indeed, so far results show that they may work best when combined with those drugs. “Typically in cancer treatment, you're looking at multiagent, multimodality therapy,” says clinical oncologist Antonio Grillo-Lopez of IDEC.

    The theory behind antibody therapy is straightforward. Antibodies are a first line of the body's defenses against infection. Each antibody grasps a specific target, or antigen, and holds on, meanwhile alerting the rest of the immune system to the intruder. Make antibodies that target antigens produced by tumors and inject them into the bloodstream, the theory went, and they would converge on a tumor and destroy it.

    Some antibodies lived up to that promise—and continue to do so. For example, in the longest running clinical trial of a therapeutic cancer antibody, immunologist Gert Riethmüller of the University of Munich in Germany and colleagues report success in preventing colon cancer from spreading by giving, after surgery, a mouse antibody called Panorex, which targets a protein found in both normal and cancerous gut cells; this protein helps cells stick together and, in the case of cancerous cells, may help metastases to form. After 7 years of study, Panorex continues to be significantly more effective than control treatments, Riethmüller's team reports in the current issue of the Journal of Clinical Oncology. Riethmüller observed a 32% reduction in mortality—that is, 63% of 76 patients in the control group died, while only 43% of the 90 patients treated with Panorex died. There are occasional short-term side effects—severe nausea and diarrhea, presumably from the antibody's attack on normal gut cells—but such toxicity was considered manageable, particularly when only a few doses are given.

    But other antibodies have been ineffective—or, worse, sparked lethal reactions. For example, in the early 1990s, several antibodies targeted against the blood infection sepsis went into late stage clinical trials—but patients who received them were found to be more likely to die than those who didn't. The trials were halted, and there is still no approved antibody treatment for sepsis. When it comes to antibody results, says oncologist Robert Cohen of Genentech, “it's hit or miss.”

    Some of the toxicity stems from antibodies attacking normal cells and some from the fact that most monoclonal antibodies are manufactured in mice, so human patients mount immune responses to the antibodies themselves. Researchers now have new strategies to get around that problem, by replacing varying amounts of the mouse antibody molecule with human antibody sequences. For example, IDEC's anti-B cell lymphoma antibody, Rituxan, is a half-mouse, half-human molecule. And it is aimed at a particular variant of a molecule called CD20, which is found only on B cells, especially cancerous ones. It causes only mild side effects such as fever, chills, and skin rash.

    Rituxan apparently works: It matches or bests standard treatments in non-Hodgkin's lymphoma patients with poor outlooks. At the meeting, Grillo-Lopez reported that when the antibody was given together with chemotherapy, every one of 40 patients responded. Thirty-one of 38 evaluable patients were still in remission at last count. Twenty-nine months after the study, the group has already slightly exceeded the median time before relapse seen in patients who receive chemotherapy alone. Even when given alone, Rituxan caused tumors to shrink or disappear in nearly half of 166 patients who had exhausted other methods of therapy. Six percent were in complete remission, and the median survival time was 13.1 months.

    Some immunologists consider lymphoma a special case, however, because cancer cells in the blood present an unusually accessible target for the injected antibodies. But recent work shows that well-designed antibodies can fight solid tumors, too.

    At Genentech, researchers designed an anti-breast cancer antibody called Herceptin to attack a specific target: HER-2/neu, a growth factor receptor that is present in larger than normal amounts on some breast cancer cells. Numerous studies have shown that the unlucky 25% to 30% of breast cancer patients whose tumors produce more HER-2/neu have worse prognoses and shorter life expectancies.

    Six years of clinical trials later, this design work has paid off. Last week at the ASCO meeting in Los Angeles, Genentech researchers announced that Herceptin can slow the progression of breast cancer in women whose cancer had already metastasized. When the antibody was combined with the chemotherapeutic drug taxol, 42% of 96 women with metastatic breast cancer responded, with tumors shrinking by half or more. That was much better than the 16% of 92 patients who improved with taxol alone. Addition of the antibody also seems to have extended the median time to relapse from 4 months to as much as 11 months. “Taxol is a pretty good drug,” observes Cohen, who led the Herceptin project at Genentech. “Here, for marginal additional toxicity [to the patient], you get a huge amount of benefit.”

    Boom, bust, and boom?

    This representation of antibodies' fortunes illustrates how they enjoyed years of hype, fell out of favor, and are now gaining respect again.

    Breast cancer patients who have exhausted other treatment options are already demanding the antibody, and the FDA has promised consideration within 6 months. “We estimate that 30,000 to 50,000 women would be eligible” to receive prescriptions for Herceptin if it is approved, says Cohen.

    Still, Herceptin is not necessarily the “breakthrough” that some company representatives claim, says the University of Michigan's Lichter. For one thing, although the drug slows disease progression, it is too early to say that it prolongs women's lives. Lichter calls the response “modest” but significant in that it targets solid tumors.

    And the Genentech team notes that Herceptin vindicates their strategy for reducing toxicity and humanizing antibodies. Herceptin began as a typical Y-shaped mouse monoclonal antibody. Genentech scientists replaced nearly all of it except the tiny pieces of the arms of the Y that actually contact the HER-2/neu antigen. “The antibody we're selling has about 5% original mouse sequences,” says Cohen. The researchers also “fiddled with the hinge” at the base of the arms so that they could wrap around the antigen correctly.

    Even so, researchers are aware that they must proceed with caution, for even “humanized” antibodies can be dangerous, as illustrated by a harrowing case study described at the meeting by oncologist Sydney Welt of the Memorial Sloan-Kettering Cancer Center in New York City. His team used a humanized version of a mouse monoclonal antibody, but the humanizing apparently hadn't gone far enough. Four out of 11 patients “developed significant anti-antibody activity,” Welt reported. Worse, one late-stage cancer patient experienced stronger and stronger immune responses to each dose of antibody. Her kidneys clogged with immune complexes made of her own antibodies clinging to the therapeutic antibody. She suffered kidney failure and died. Welt's team has abandoned that antibody and is preparing others against the same target.

    Still, the promising results with IDEC and Genentech antibodies, combined with new strategies for reducing toxicity, should draw new investment into the once-discredited field, predicts oncologist Lynn Schuchter of the University of Pennsylvania. “Clear-cut tumor responses” such as those in the Genentech study “will get industry interested in pursuing this again,” she says. Seventeen years after his miracle cure, Philip Karr is still alive and healthy. And so is the science that saved him.

    • * Antibodies 1998, the Cancer Research Institute, New York, April 22–24.

Log in to view full text

Via your Institution

Log in through your institution

Log in through your institution