News this Week

Science  24 Jan 1997:
Vol. 275, Issue 5299, pp. 476

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution

  1. Astrophysics

    Visions of Black Holes

    1. James Glanz

    A menagerie of black holes—forming, colliding, evaporating—is romping through the minds and computers of theorists, testing their command of Einstein's theory and hinting at new physics

    CHICAGO—When the talk is of black holes, everyone has a Chandra story. One of Kip Thorne's favorites comes from a cross-country drive he took in 1966 from Princeton University, where he had recently finished his graduate studies, to the California Institute of Technology, where he now specializes in general relativity. On the way, Thorne stopped at the University of Chicago to see the great astrophysicist Subrahmanyan Chandrasekhar—the man universally known as Chandra—who died two summers ago at the age of 84. Thorne says he drove out of his way simply because “I had a question for Chandra.” Thorne was studying an exotic concept, remote from anything ever observed: black holes. He made his pilgrimage to ask whether he might be wasting his time.

    Chandra had been in that situation himself. On a voyage from Bombay to England in 1930, he had realized that old, burnt-out white dwarf stars can't support themselves against gravity when they exceed 1.4 times the mass of the sun. Above what came to be called the Chandrasekhar limit, he declared, they collapse into denser and more exotic objects or blow apart completely. The idea met with disbelief at the time and even in 1966 had little observational support. “Chandra reassured me that [research on black holes was] at least as likely to make contact with observation as his white-dwarf work had been,” recalls Thorne.

    The advice was prescient. “Neutron stars were found a year later,” says Thorne—compact balls of matter, billions of times denser than ordinary stars, that are one endpoint of stellar collapse. Their discovery not only vindicated Chandra's ideas; it primed astrophysicists to accept the even more drastic vision of gravitational collapse that Thorne and others were exploring. “It was Chandra who started things off,” says Stephen Hawking.


    In a computer simulation, matter collapses into a donut-shaped black hole, contrary to predictions. The clock in each frame shows the fraction of time elapsed.

    S. L. Shapiro and S. A. Teukolsky

    Thirty years later, black-hole theorists are still following Chandra's example, as a conference* here showed: They are leaving present observations far behind as they create and study strange new beasts, confident that observers will eventually catch up. While the observational case for real black holes is firming up (see sidebar), theorists have already moved on to study the details of the holes' behavior. “There's an incredible level of activity,” says Ed Seidel of the Albert Einstein Institute in Potsdam, Germany. Lately, theorists have conjured up donut-shaped black holes, discerned clues to a fundamental graininess of space-time in the properties of black holes, and—thanks to Hawking's latest brainstorm—contemplated the possibility that invisible black holes might be gobbling information right under our noses.

    The exploding power of supercomputers is responsible for some of this frenzy, says Seidel. But the real spur is the intellectual challenge of trying to understand the purest creatures of general relativity, Einstein's theory representing gravity as curvature of space-time. In a black hole, that curvature is so extreme that it forms an “event horizon,” from which light itself can't escape. To theorists trying to master the many-armed complexity of Einstein's equations, says Matthew Choptuik of the Center for Relativity at the University of Texas, Austin, understanding violent interactions between black holes is the ultimate test.

    Hole in a hole

    One current goal is to simulate the interactions of two spinning black holes as they orbit around each other, gradually spiraling in and coalescing while spraying out gravitational waves. Tracing the process will eventually take a trillion bytes of memory and a week of computation time on a machine hundreds of times faster than most of today's supercomputers, says Choptuik, but “it's a watershed problem for general relativity.”

    Even before computational physicists get close to wrapping up that problem, the exercise has yielded unexpected theoretical insights. A year ago, a team including Seidel and Saul Teukolsky of Cornell University published the solution to a simplified version—the head-on collision of two nonspinning black holes—and got their first close look at how two event horizons could merge into one (Science, 10 November 1995, p. 941).

    The process can be visualized as a pair of pants stretched out in space-time, in which two “legs” merge at a “crotch.” The legs show the history of the converging event horizons, which are themselves traced out by the space-time trajectories of light rays that don't quite fall into the black holes but can never escape. The puzzle is how the light rays converge at the crotch, where two separate event horizons become one. The computer model showed that the rules of space-time don't allow light on the inseam of the pants to ride up the crotch onto the new event horizon. Instead, the simulation showed, some of the light rays trapped on the new boundary arrive from someplace else in the universe, flashing onto the event horizon during collision.

    Later, Teukolsky and Stuart Shapiro, of the University of Illinois, Urbana-Champaign, created what some mathematicians thought they had proved could not exist: a toroidal—or donut-shaped—black hole. Teukolsky and Shapiro did this by replacing the original “two-car collision” of black holes with an entire torus of collapsing matter, crashing inward like dozens of Renaults smashing together simultaneously in one of Paris's multistreet intersections. “We were unaware of these theorems, so we just tried to make [toroidal black holes],” says Teukolsky. “We succeeded”—although the donut hole always closed up soon after the collapse.

    Hot and cold black holes

    Another group of theorists is using the weirdness of black holes as a clue to physics beyond Einstein's equations entirely. No one has yet merged quantum mechanics—which describes the small-scale graininess of matter and energy—and relativity to make a successful theory of “quantum gravity” that would extend this graininess to space and time. One possible route could lie in the esoteric mathematics known as string theory (Science, 15 September 1995, p. 1511). But striking parallels between the mechanics of black holes and classical concepts of temperature and entropy—a system's degree of randomness—could also end up showing the path to quantum gravity, says Robert Wald of the University of Chicago and the principal organizer of the Chandra symposium.

    A century ago, he reminded his listeners, the second law of thermodynamics—the inevitable increase of entropy of a system such as gas in a piston's chamber—helped persuade physicists that all matter consists of atoms. By accounting for atomic motions, they found, they could actually prove the law. Lately, says Wald, theorists are finding resemblances between the textbook mechanisms and the workings of black holes that are “just too amazing, I think, to be some mathematical curiosity.”

    In this analogy, the area of the event horizon takes the place of entropy. Just as total entropy always increases when parcels of gas merge or get pushed around mechanically, the event horizon always expands when, say, two black holes interact or more matter is added to a black hole. Black holes have a temperature, too, as Hawking showed in 1974 when he found that black holes should radiate particles. This radiation is fed by the normally undetectable pairs of particles that, according to quantum mechanics, constantly pop in and out of existence throughout space: Near the event horizon, one member of a pair can get sucked into a black hole while the other flies away. And he and others have shown that this Hawking radiation should have a “thermal” spectrum of energies, shaped exactly like the spectrum of radiation from an object glowing at a particular temperature. The stronger a black hole's gravity at the event horizon, the higher its “temperature” would be.

    Physicists hope that such correspondences will let them understand the rest of the story. Working backward from the black hole's entropy, says Rafael Sorkin of Syracuse University in New York and ICN-UNAM in Mexico City, shows that “it's just as if the black-hole horizon was made out of many pieces of about the Planck size”—what would be the smallest imaginable dimension in quantum gravity. Sorkin is still sorting out exactly how that graininess might arise. But he thinks that ultimately, black-hole entropy may “[lead] us to the atoms of space-time itself.”

    Physicists acknowledge, however, that the analogy comes with deep mysteries, such as how the horizon area could ever take the place of the entire volume, the natural setting for classical thermodynamics. Even more distressing to some traditionalists is a black hole's apparent disregard for “real” information such as the kind of material that falls into it, which can figure in the standard kind of entropy. The Hawking radiation allows black holes to slowly “evaporate” and disappear without yielding any information about what they had swallowed up. That's uncomfortable for some physicists who, Hawking says, “seem to have a strong emotional attachment to information.”

    Hawking's talk only deepened this discomfort by proposing that the process could be ubiquitous: Microscopic pairs of black holes could be forming and evaporating throughout space, consuming small-scale order like Pac-Mans. Such black holes might, for example, eat one kind of particle and emit another as they evaporate, violating some of the most hallowed conservation laws of particle physics.

    But this mind-bending vision was also a fitting conclusion to the symposium, said Hugo Sonnenschein, president of the University of Chicago, where Chandra spent nearly 6 decades as a faculty member. “It's an amazing juxtaposition to call yourself a wanderer”—one of Chandra's favorite descriptions of himself—“and be in so many ways unsure, and yet feel that you can solve mysteries that are beyond imagination.”

    • * “Black Holes and Relativistic Stars: A Symposium in Honor of S. Chandrasekhar,” sponsored by the University of Chicago, 13-15 December 1996.

  2. Black-Hole Observations

    The Gathering Darkness

    1. James Glanz

    Black holes are flourishing in theorists' computers and on their blackboards, as a recent symposium held in honor of the late Subrahmanyan Chandrasekhar showed (see main text). But these objects—the children of Einstein's theory of gravity—are also alive and well in the universe, observers are finding: They are detecting increasingly strong signs of black holes at the centers of galaxies and, closer to home, in turbulent x-ray beacons. These new observations indicate that “the entities to which Chandra devoted so much of his life have a real, concrete existence and are not just theoretical constructs,” said Martin Rees of the University of Cambridge at the symposium.

    What Rees calls “the most convincing case for a black hole” comes from microwave measurements of a whirling, gaseous disk at the center of a galaxy called NGC4258, about 20 million light-years away. Two years ago, a team including Makoto Miyoshi of the National Astronomical Observatory in Japan and James Moran of the Harvard-Smithsonian Center for Astrophysics (CfA) in Cambridge, Massachusetts, observed NGC4258 with the Very Long Baseline Array, a far-flung set of radio telescopes. They found that as water molecules in the disk become excited by, say, starlight, they can amplify certain frequencies and act as a maser—the microwave equivalent of a laser. As the disk's material whirls toward or away from Earth, the maser's microwaves are shortened or lengthened. The magnitude of these Doppler shifts, along with other images of the galaxy, indicated that the gas is whirling around a compact, dark object with a mass of about 36 million suns.

    It's got that spin.

    An artist's impression shows the disk of gas at the heart of galaxy NGC4258. Frequency shifts in maser signals (bottom) from the disk suggest it is whirling in the grip of a giant black hole, which also drives the two jets.

    Prepared for Makoto Inoue by Joh Kagaya

    “It's dark, and the scale is so small”—less than half a light-year across—“that you could not conceivably hide a star cluster in that region,” says Rees. Moran, who presented new measurements at the 18th Texas Symposium on Relativistic Astrophysics, held just after the Chandra conference, says the team now has two more “solid candidates” for galaxies showing similar Doppler patterns.

    Similar observations of stars have pointed to black holes at the hearts of other galaxies, although stars—being scattered points—can't trace out a gravitational field with the smooth continuity of a disk of gas. Still, what Mitchell Begelman of the University of Colorado, Boulder, considers the next-best case for a black hole comes from stellar measurements close to home: at the center of the Milky Way. Doppler shifts in starlight from the Milky Way's core had suggested that a small, massive attractor resides there. But skeptics could argue that if some asymmetry led to slower velocities, on average, across the plane of the sky than along the line of sight to Earth, the mass of the central object might not be so great. In a paper in the 3 October issue of Nature, however, A. Eckart and Reinhard Genzel of the Max-Planck-Institut für Extraterrestrische Physik in Garching, Germany, “tightened the noose on the galactic nucleus,” says Begelman.

    Eckart and Genzel followed 39 stars near the galaxy's core across the sky for several years and showed that their velocities are roughly the same in all directions. The speeds indicate that they are orbiting a dark mass of several million suns somewhere within a fraction of a cubic light-year around the center—a density that strongly implies a black hole.

    What's true of our ordinary galaxy may be true of most others. During a 13 January press conference at an American Astronomical Society meeting in Toronto, for example, a team led by Douglas Richstone of the University of Michigan described a black-hole census of 15 nearby galaxies using the orbiting Hubble Space Telescope. Doppler measurements showed that at the centers of nearly all of the galaxies, stars are swirling at a frenzied pace, apparently rallied by the gravitational pull of giant black holes with anywhere from millions to billions of times the mass of the sun.

    Still, says Rees, all of this work begs the deeper question of “whether these black holes really have the properties that Einstein's theory says they have.” Here again, he says, “there's been progress.” One advance comes from an international team that used the Japanese x-ray satellite ASCA to collect data from “active” galaxies, which have mysterious, brilliant cores. X-ray spectra of the cores seem to show an imprint of the warped space-time that Einstein's theory predicts close to a black hole, in the form of peculiar frequency shifts that can be caused only by strong gravity.

    First reported 2 years ago, the signature has now been seen in more than a dozen new objects, says Richard Mushotzky of Goddard Space Flight Center in Greenbelt, Maryland, who did the work with a team including Andrew Fabian at Cambridge, Paul Nandra at Goddard, and Yasul Tanaka, recently retired from the Institute of Space and Astronautical Science in Japan. Says Mushotzky, “We think these [data] are the strongest evidence for the missing component in black holes—strong gravity.”

    But another set of ASCA observations described at the Toronto meeting may have trumped this claim. A team at the CfA announced uncovering possible evidence of the “event horizon” shrouding black holes. According to Einstein's equations, nothing—neither light nor matter—that passes within the event horizon ever emerges again. The CfA team looked at x-ray emissions from nine objects called x-ray novae, which are powered by gases peeled from a companion star and sucked toward a superdense object—either a neutron star or perhaps a black hole.

    X-ray novae don't drag in material at a steady pace, and theoretical models suggested that when the infalling material is especially tenuous, it should be slow to lose energy, because there are too few collisions for the atoms to radiate much. In fact, it should be unable to shed most of its energy until after it has made the plunge. The energy should still be able to escape from a neutron star, but it would be lost behind the event horizon of a black hole. And the team did find that the suspected black holes were unusually dim compared to the neutron stars during these periods of slow accretion. “We think we are seeing, for the first time, direct evidence that the event horizon really exists,” says Ramesh Narayan of CfA, who did the analysis with CfA's Jeffrey McClintock and Michael Garcia. Comments Yale University astronomer Charles Bailyn, “This moves some very exotic behavior into observational astronomy.”

  3. Planetary Science

    An Icy World Looks Livelier

    1. Richard A. Kerr

    WASHINGTON, D.C.—The latest images of Jupiter's moon Europa, released here last week at a NASA press conference, reveal a landscape in turmoil. To team members poring over images returned by the Galileo spacecraft, the wild jumble of ridges, grooves, pits, ice flows, and chaotic terrain shows ever more signs that heat forged Europa's icy surface. Even to a nonscientist, it's clear that this moon's surface is—or was—mobile. Pope John Paul II, who was shown images of the satellite last week after a scientific meeting in Italy, said simply, “Wow!”

    “These new images demonstrate that there was enough heat to drive [ice] flows on the surface,” says Galileo imaging team member Ronald Greeley of Arizona State University. And where there's water and enough heat, there might once have been life, he says. “Europa thus has a high potential” as a place where life could have gotten started.

    Space scientists had already raised the notion that Europa might at some time have been warm enough to support a water ocean, thanks to Voyager images taken in the 1980s and Galileo images from an earlier, more distant flyby (Science, 20 December 1996, p. 2015). Now, images showing details up to 20 times finer—as small as 36 meters—add weight to that possibility. Greeley points to several features in the accompanying image that he suspects depict various stages in the rising of plumes of warm ice or a wet, icy mush. The plume first bulges the surface and cracks it (middle of top block). Then, some unearthly geologic process takes over, and the broken terrain collapses into a chaotic jumble of ice blocks (right side of middle image blocks). Greeley speculates that sublimation of small amounts of ammonia or methane ice from the water ice weakens the terrain, and so it crumbles. In other images, rising water or icy mush appears to have burst through the surface like an ice volcano, flowing as 100-meter-thick lobes of ice for hundreds of kilometers.

    Even these features don't prove that an ocean lurks beneath the surface. Icy volcanism could be driven by scattered hot spots in a solid layer of ice rather than an actual ocean. Or an ocean may have existed in the moon's youth and frozen solid since, squeezing out these few dribbles of ice with its last gasp. But it's also possible that there may still be a warm, liquid sea—perfect for life—hidden below the ice. Galileo's future close encounters with Europa may tell.

  4. Ecology

    Much-Studied Butterfly Winks Out on Stanford Preserve

    1. Ellen McGarrahan
    1. Ellen McGarrahan is a free-lance writer in San Francisco.

    It might not be as big or as flashy as some of its cousins, but the bay checkerspot butterfly is among the most studied insects in science. Since 1960, when a young Stanford biology professor named Paul Ehrlich first took to the hills at the university's Jasper Ridge preserve, his sights set on Euphydryas editha bayensis, textbooks and scientific papers have been aflutter with checkerspot findings and theories. But last year, for the first time, graduate students engaged in the annual bay checkerspot count at Jasper Ridge came back empty-handed. After a long decline, it appears that the insect, which was officially listed as a threatened species, has disappeared from the same Stanford grasslands where the studies began. “The loss of the study population is a huge, huge thing because of the amount of information there was on that population,” says Daniel Simberloff, an ecologist at Florida State University (FSU) in Tallahassee. “It was really sort of a classic.”

    This local extinction isn't a certainty yet. Scientists at Stanford's Center for Conservation Biology say there's a remote chance that an odd E. editha bayensis may be spotted on the ridge this March. But the possibility of a Jasper Ridge local extinction has raised some entomological eyebrows. Ehrlich and others at Stanford chose to watch the population die off, saying more could be learned from watching it disappear than from intervening to try to save it. As Stanford biologist Dennis Murphy puts it, “We [decided to] observe the demise unimpassioned.”

    Local loss.

    Bay checkerspot butterfly in adult (left) and larval (right) stages.

    Photos by Ed Nelson/Jasper Ridge Biological Preserve

    That choice has reverberated in the ecology community. Some observers, including ecologist David Wright, who oversees endangered invertebrates at the U.S. Fish and Wildlife Service in Sacramento, California, wonder whether the wealth of information about the Jasper Ridge population might have justified efforts to revive it. And he and others say the controversy highlights a dilemma that many ecologists studying endangered species eventually face. “Increasingly, if we don't intervene, we will be watching our study materials disappear,” says Wright.

    Stanford scientists have observed the checkerspot since at least 1934 on the 485-hectare Jasper Ridge biological preserve, which is owned by the university and contiguous to the main campus. The preserve is now an island in a semisuburban sea, one of the few places in northern California where one can still find native grasses growing on stony, inhospitable, “serpentine” soils. Development is just one reason for the loss of serpentine grasslands. The invasion of exotic species and the cessation of livestock grazing also have reduced the habitat available to dependent species, including the bay checkerspot, which feeds on Plantago erecta, the California plantain, among other native plants. Even in this refuge, the checkerspot was confined to some 2 hectares of serpentine grasslands.

    Over the decades, checkerspot observations have yielded a treasure of entomological information. “It's pretty amazing where Ehrlich's initial study has taken us,” says Stuart B. Weiss, a Stanford postdoctoral fellow who has been studying the bay checkerspot since 1978. Ehrlich first counted the butterflies during their spring flight season in 1960. His discovery that they lived not as a single population but as three distinct populations (labeled G, C, and H) helped upset some ideas of the day about population dynamics. Observations of the butterflies and their host plants have helped scientists better understand coevolution, or the complementary evolution of closely interacting species. “There are very few organisms we know as much about as Euphydryas editha,” Ehrlich says. “And the more we learned, the more we became humble about what we didn't know.”

    Serpentine splendor.

    Native grasslands on Jasper Ridge during spring bloom.

    Jasper Ridge Biological Preserve and Alan Lerner (Center for Conservation Biology)

    During the droughts of the mid-1970s, the number of Jasper Ridge checkerspots nose-dived. In 1980, Murphy petitioned the U.S. Fish and Wildlife Service to place the butterfly under the protection of the Endangered Species Act (ESA). And in 1987, it was listed. During the interval, a previously unknown cache of hundreds of thousands of checkerspots was discovered to the south of Stanford near San Jose on land slated to become a landfill. The landfill developer, Waste Management Inc., agreed to set some of the property aside as a checkerspot preserve.

    Ironically, the protection of the Kirby Canyon butterflies south of San Jose helped seal the Jasper Ridge populations' fate. During the 1980s and the early 1990s, as the Jasper Ridge checkerspot populations staggered along, the Stanford biologists—mindful of the large pool of checkerspots at the new Kirby Canyon preserve—“agreed that we would just watch [the Jasper Ridge population] go belly up,” Murphy says. “The idea was, here's an opportunity to actually watch the extinction process,” says Weiss. The ESA required nothing different: The act prohibits harming protected species, but does not require that actions be taken to save them.

    Stanford biologists tick off a host of factors that could have contributed to the checkerspot's decline. First, there was the science itself. In the early 1980s, a massive mark and recapture study affected nearly every butterfly in the preserve, Weiss says. “It's concluded that we hurried the extinction of the population in area G and the rate at which C went extinct by sampling butterflies,” Ehrlich says. Then, there was the decision to remove cattle and horse grazing from the preserve in the early 1960s. Murphy, who has studied the checkerspot for 20 years, says, “There's no question that the removal of grazing up there contributed” to the butterfly's demise. Other events such as an inadvertent malathion spray in 1981 and local pesticide use may also have taken a toll.

    But ultimately it was bad weather and lack of topographic diversity that did in the butterfly, say Weiss and others. Confined to the ridge's small areas of serpentine grasslands, the bugs didn't have enough habitat choices to survive California's tempestuous climate. The plants on which the butterfly larvae feed wither early on warmer slopes, later on cooler slopes. Without adequate habitat variations to choose from, the butterflies died off. “It's sort of the Goldilocks effect,” says Weiss. “It has to be just right.”

    Indeed, that realization was part of the intellectual payoff of watching the bay checkerspot's slow slide toward extinction on Jasper Ridge, says Ehrlich. “My conclusion early on was that there were several large patches of suitable habitat which should be more than enough. That conclusion was dead wrong,” says Ehrlich. “What this adds is that it isn't just [total] area that matters. The topographic diversity of the habitat” is “an absolutely critical factor” as well, he says, as is “the timing of the plants, when they dry up.”

    But news of the loss of the Jasper Ridge checkerspot has highlighted for many the question of when, if ever, scientists studying a shrinking population should intervene to save it. “That's very controversial,” says FSU's Simberloff.

    The Fish and Wildlife Service's David Wright, for one, thinks more could have been done to help out the Jasper Ridge checkerspot, given its status as a threatened and much-studied species. “I don't intend this as a criticism of Stanford, but I regard this as a wake-up call … for anyone interested in conservation.” He says that “if a population has survived at a location for tens of thousands of years, it likely has the reproductive capability to recover from environmental vagaries”—if given a helping hand. Wright suggests that researchers might have experimented with increasing the amount of habitat available to the butterflies by restoring grazing or controlled burns or weeding part of the reserve.

    By most accounts, the case for intervening is strongest when the threatened population is genetically distinct. Murphy says that concern doesn't apply to Jasper Ridge, where “there were no unique alleles.” Ehrlich concurs, saying that the vast reserve of Kirby Canyon checkerspots is “very similar” to the Jasper Ridge populations. But Susan Harrison, an associate professor of environmental studies at the University of California, Davis, who did her graduate work with Ehrlich and Murphy on the Kirby Canyon checkerspots, asserts that “nobody really knows the answer to that because the studies weren't done.” Comprehensive genetic studies of Euphydryas populations along the western United States were done in the late 1960s and early 1970s, but used a method which has since been shown to be unreliable, says Alan Launer, research associate at Stanford's Center for Conservation Biology.

    Reed Noss, a population extinction expert and professor at Oregon State University in Corvallis, says that whether to intervene “really depends on the management goals for a particular area.” Small populations “have a high chance of going extinct,” he says, and “from a metapopulation standpoint, it probably doesn't matter” if the Jasper Ridge butterflies have disappeared because of the reservoir of bugs at Kirby Canyon. But where a whole species is winking out, and efforts to protect habitat haven't been effective, interventions may be warranted. “If a species is really on the brink, and we see that an intervention can be done, we have an obligation to do that, just ethically,” he says.

    Watching the Jasper Ridge checkerspots disappear was an important research opportunity, Murphy contends. “Watching this population hang on at about a dozen individuals was one of the more enlightening aspect of our study of the species,” he says. Indeed, says Ehrlich, “just trying to keep Euphydryas going on Jasper Ridge would give us less information” than observing the extinction.

    Either way, the question of observation versus intervention is “something the community needs to discuss more,” says Wright. He also would like to see Euphydryas editha back on Jasper Ridge. “I want to work with them to reintroduce the butterfly to Jasper Ridge. If there are special permits that are needed, I'm more than happy to put that on my list of priorities.”

    Ehrlich and Murphy share Wright's enthusiasm for a reintroduction. “We will be trying a reintroduction,” Ehrlich promises. But Murphy notes a possible snag: “Stanford is very sensitive to the legal implications of putting the butterfly back in the habitat. Stanford now has grasslands that are free of listed species. If they wanted to build on these habitats, they frankly could.”

    And while Stanford has no plans for construction on Jasper Ridge, says Stanford spokesperson Janet Basu, it isn't planning to reintroduce the checkerspot butterfly, either. “It looks like the reintroduction won't happen, at least [not] in the short term. … There's been no forward action on this,” Basu says.

    Come mid-March, graduate students will again be taking to Jasper Ridge to search for the checkerspot. If none are seen, then the extinction will be official. “It's the loss of a symbol, and it's another example of population extinction,” mourns FSU's Simberloff. “You now have more and more examples of a depressing trend.”

  5. Biochemistry

    Photons Add Up to Better Microscopy

    1. Elizabeth Pennisi

    The first view through the light microscope opened up the cellular world for 17th century biologists. Now, a new kind of microscope could do the same thing for the world of biochemistry, letting 20th century biologists follow molecules in real time within living cells.

    On page 530, applied physicist Watt Webb and his colleagues at Cornell University in Ithaca, New York, describe how they tapped photon physics to view a key brain chemical called serotonin inside living cells. The significance of the achievement goes beyond serotonin, however. Webb's method, which uses the additive energies of multiple photons to excite fluorescence from molecules that previously couldn't be observed without damaging or killing the cell, should open new vistas for any biologist interested in tracking specific molecules in tissue.

    The technique will enable researchers to probe deeper into cells and to monitor molecules in living samples much longer than previously possible, Webb says. Joseph Lakowicz, a biochemist at the University of Maryland School of Medicine in Baltimore who is also developing new fluorescence spectroscopy techniques, agrees: “Webb has really changed the paradigm of microscopy.”

    In current microscopic methods, biologists often visualize cellular components by tagging them with molecules that fluoresce when excited by light of the correct wavelength. But the fluorescence technique is limited because many dyes and cellular components, such as proteins, fluoresce only when excited by short-wavelength, high-energy photons that can overheat the cell or drive toxic chemical reactions. In addition, because the entire sample is illuminated, stray light or fluorescence can lower the contrast of the image. Confocal microscopy partially solves this problem with a pinhole, placed in front of the photodetectors, that blocks out all but the focused light.

    Webb got his first glimpse of a way around the wavelength limitation 9 years ago. A colleague had created synthetic molecular “cages” filled with the neurotransmitter acetylcholine that could be inserted in nerve cells to study whether this neurotransmitter's effects depend on the site of its release. He needed Webb to come up with a way to deliver enough energy to a particular part of the cell so the cages there would disintegrate and release the neurotransmitter—without killing the cells.

    Webb realized he could do this if he focused a laser beam and delivered the photons in short, intense pulses. The densely packed photons would have a good chance of hitting the cages in pairs, delivering a one-two photon punch that is equivalent, energetically, to a single hit by a photon with twice as much energy (or half the wavelength). The lower energy of the photons would help minimize the damage to the cells, and the photons would be sufficiently dense to deliver a double dose only along the plane of focus. “You don't burn up the cells above or below the plane of focus,” Webb says.

    But Webb—and later, others—quickly realized that those two photons could also stimulate fluorescence. “That was the conceptual breakthrough,” says John White of the University of Wisconsin, Madison. Indeed, in work reported in 1990, Webb demonstrated that two-photon excitation could be used in fluorescent imaging technology (Science, 6 April 1990, p. 73). He showed, for example, that he could follow the moving chromosomes in dividing cells by stimulating fluorescence of a dye attached to the DNA.

    In the current work, Webb and his colleagues have now gone a step further. By delivering light in shorter, brighter pulses from a titanium sapphire laser, they raise the odds that three photons will simultaneously strike individual molecules. The energy of all three add together, extending even further the range of fluorescence excitation and making possible the use of photons of even longer wavelengths.

    When tested on leukemia cells, which like nerve cells contain serotonin but are easier to work with, these triple hits were enough to make serotonin fluoresce, without the need of any external dye. Based on the amount of fluorescence, Webb was able to measure the amount of serotonin in the tiny granules that store the chemical until it's released. Others have tried to visualize serotonin molecules in the granules, but only in fixed tissue, not living cells, says Webb.

    While Webb was working out the logistics of three-photon excitation, others were hot on the same trail. In 1995, Maryland's Lakowicz began to test the potential of three photons to excite fluorescence from various dyes and biological molecules. At the same time, Victoria Centonze in White's lab made an unexpected observation. A cell that she expected to emit just red fluorescence under her microscope also emitted blue light. She and White didn't realize at first that the blue emission was the result of three-photon hits exciting a second dye that was also present, but their colleague David Wokosin did go on to demonstrate that that was indeed the case.

    White's team has now used a single laser to excite fluorescence by both two-photon and three-photon absorption in the same specimen. In work published in the September 1996 issue of Bioimaging, the researchers report that this allowed them to follow three different biological molecules, each tagged with a different dye. White estimates that the strategy will make it possible to follow up to five molecules simultaneously in living tissue.

    Also, the longer wavelengths of light that can be used in multiphoton excitation don't scatter on their way through tissue, as do shorter wavelengths, so “you can probe deeper into the cell,” White adds. Webb's group has looked 390 nanometers into skin and observed how sun-damaged elastin shatters into tiny pieces. And White says he can peer two to five times deeper into zebrafish embryos than he could with conventional confocal microscopy techniques.

    Currently, the only multiphoton instruments are those the researchers put together, but Cornell has granted Bio-Rad Laboratories in Hercules, California, a license to develop multiphoton excitation into a commercial instrument. This will likely cost several hundred thousand dollars, however, until laser technology improves. But White, who has consulted for the company but otherwise doesn't stand to gain from the new product, expects that researchers will quickly come to appreciate what this new microscope has to offer. “It has few disadvantages compared with the confocal microscope and quite a few advantages,” he says. “I suspect that it will largely supersede the confocal microscope.”

  6. Plasma Physics

    More Powerful Pulses Please and Puzzle

    1. Dennis Normile

    OSAKA, JAPAN—If results presented at an international meeting* here hold up, researchers will have taken a sizable step toward creating a new generation of compact particle accelerators powered by laser pulses. A team at the Japan Atomic Energy Research Institute (JAERI) led by physicist Kazuhisa Nakajima says it has succeeded in accelerating electrons to energies of from 100 million to more than 300 million electron volts. That's still well short of the energies needed for high-energy physics experiments, but it's more than three times higher than those reached in earlier experiments. The announcement has stirred both excitement and caution, however, because theorists can't explain the achievement.

    “The results, if you take them at face value, are extremely impressive,” says Chan Joshi, a University of California, Los Angeles (UCLA), electrical engineer and a pioneer in laser acceleration. “But there are aspects of the results that are hard to understand.” Among other things, he and his colleagues wonder how the laser pulse could have remained sharply focused for long enough to drive the intense accelerations reported by the group, which is part of JAERI's year-old effort to push the development and use of compact, short-pulse lasers (Science, 5 January 1996, p. 26).

    Wake-up call.

    Nakajima's results are controversial among short-pulse laser physicists.

    D. Normile

    In principle, accelerating electrons with such lasers sounds easy. A laser pulse shot into a gas ionizes it and creates a wake, much like a speedboat zipping across a pond. Electrons riding on the wake can be accelerated to high energies. The advantage over conventional accelerators is that the acceleration occurs over a much shorter distance, which could sharply cut the size, cost, and energy requirements of accelerators. But the effect occurs only if the laser pulse can somehow remain within a narrow channel for several centimeters instead of dispersing within a fraction of a millimeter, as it does normally.

    Theoretical studies predict that, at very high laser energies, the interaction of the laser and the plasma, or ionized gas, creates a sort of lens in the plasma that focuses the laser light and propagates it along the narrow path needed for high acceleration. The JAERI group, however, claims to be seeing these phenomena, called self-focusing and self-channeling, at much lower energy levels. As evidence, they cite the shape of the fluorescence excited by the laser in the plasma, the spectrum of the laser light scattered out of the plasma, and direct measurements of the diameter of the laser beam, which they say remained small for several centimeters. But UCLA's Joshi and others say the evidence is not conclusive. “What is lacking is a measure of laser intensity in the channel,” Joshi says.

    The experimental observations are bolstered by computer simulations by JAERI physicists Yasuaki Kishimoto and James Koga. When the researchers assumed that the laser pulse itself ionized the gas, which is the approach used in most experiments, they found that self-channeling did not occur even at the theoretically predicted energy levels. However, if the gas were ionized ahead of the pulse, then self-channeling occurred well below the critical power that theory predicts. Koga speculates that this occurs in some experimental setups when a spike of energy from the laser pulse precedes the rest of the pulse and ionizes the gas.

    His interpretation could explain why some groups, using laser pulses with different characteristics, have not detected self-channeling. But it leaves open the mechanism through which the self-focusing and self-channeling could be occurring. Says Koga, “We're wary about saying anything too strongly” about self-channeling. But Toshiki Tajima, a physicist at the University of Texas, Austin, who also does theoretical work at JAERI, says there is no other explanation for the results: “We may not understand the mechanism, but there has to be self-channeling.”

    The unexpected self-channeling laser isn't the only puzzle in the results, however. Wake-field acceleration is supposed to work only at a specific plasma density for any given laser pulse length, yet the JAERI group has seen electron acceleration over a broad range of plasma densities. They are convinced, however, that a wake field is responsible. In one set of experiments, the group varied the timing of the injection of the electrons relative to the laser pulse at different densities. Nakajima says the results—electrons injected too early or too late did not gain energy—support his contention.

    Kwan Je Kim, a physicist at Lawrence Berkeley National Laboratory on sabbatical at Kyoto University, says questions about the results indicate that the whole area of wake-field acceleration is “a bit immature.” But Tajima is more optimistic: “A year ago, we never thought the field would make such rapid progress.” Theoretical explanations, he adds, will come hand in hand with additional data.

    • * The 2nd Japan-U.S. Workshop on Interactions of High-Power Waves With Plasmas and Matters, Osaka, Japan, 16-18 December 1996.

  7. Cancer Research

    Designing Therapies That Target Tumor Blood Vessels

    1. Marcia Barinaga

    The word “cancer” conjures up images of a cohort of rampaging cells, burgeoning into life-threatening tumors that dispatch their metastatic offspring to ravage other parts of the body. Traditional cancer treatments have been based on attacking the rebel cells directly, by removing them surgically or attempting to destroy them with radiation or chemotherapy. But a new wave of potential cancer therapies aims to kill these hostile armies not by direct attack, but by shutting off their supply lines: the blood vessels through which tumors get the oxygen and nutrients they need to live and grow.

    Work reported today advances this anticancer strategy further, giving a boost to two different means of cutting tumors' lifelines. The most common one aims to prevent tumors from forming the new blood vessels necessary to nurture their growth. To block the process, called angiogenesis, researchers have identified agents that interfere with the endothelial cells that build the new vessels, by preventing them from responding to growth factors or suppressing their ability to chew their way through surrounding tissues. The second approach seeks to block blood vessels that have already formed.


    A human eye cancer attracts new blood vessels.


    These antiangiogenic and antivascular measures have already produced encouraging results on animal tumors, and today's reports add to the promise that has already launched a dozen or more candidate drugs toward the ultimate test in the clinic. In work described on page 547 of this issue, Philip Thorpe and his co-workers at the University of Texas Southwestern Medical Center in Dallas show that they can shrink or even eliminate tumors in mice by giving the animals agents that trigger blood clot formation in existing tumor-feeding vessels. And in today's issue of Cell, Judah Folkman and his colleagues at Harvard Medical School report their discovery of a factor called endostatin that is the most potent yet in a growing collection of molecules that block new blood vessel formation. Endostatin, Folkman's group reports, can shrink large tumors down to microscopic size in mice.

    To see tumors shrink so dramatically under treatment is “outstanding … better than my best hopes,” says Noel Bouck of Northwestern University, who is also working on antiangiogenesis drugs. “If it just works for human tumors, it will be fabulous.”

    That, of course, is a very big “if,” and one that applies to all the new strategies. As cancer researchers know only too well, many approaches that have looked promising in animals have died a quiet death after proving ineffective in humans. Still, antiangiogenic therapy may have unique strengths. For one, these drugs might avoid one of the main handicaps of conventional cancer therapies: the development of drug resistance, which ultimately leads to treatment failure.

    “Cancers have a formidable ability to acquire resistance to any therapeutic modality we throw at them … chemotherapy, radiation therapy, immunotherapy,” says tumor biologist Bob Kerbel of the Sunnybrook Health Science Center at the University of Toronto. But the cells of a tumor's blood vessels—which are the target of the new therapies—are normal and thus less prone to mutate than cancer cells. As added bonuses, an effective vessel-targeting therapy should be useful for many types of cancer because all tumor-feeding blood vessels are essentially the same, and delivering a drug to the vessels should be much easier than getting it into all the cells of a solid tumor.

    The idea of attacking a tumor's blood supply took some time to catch on. Back in the 1970s, when Folkman proposed that tumors have to induce new blood-vessel growth to obtain the nourishment they need, other researchers were skeptical. “The view of most scientists was that tumors didn't need blood-vessel growth at all, that they could grow with the supply that was there,” Folkman says. Over the next decade, it became increasingly clear that Folkman was right, as his group showed that tumors contain newly formed blood vessels and secrete diffusible factors that cause those vessels to grow. In the early 1980s, Folkman's lab and others isolated several of those factors and showed that they trigger blood-vessel growth. They also identified the first antiangiogenic agents, platelet factor 4, made by blood platelets, and fumagillin, a product of molds. Both showed ability to inhibit tumor growth in mice and have since entered clinical trials for cancer.

    But even today, no one knows for sure how fumagillin and platelet factor 4 block angiogenesis, or whether they might have other unknown actions that account for their effects on tumors. Indeed, “compelling proof” that blocking angiogenesis could halt tumor growth didn't come until 1993, Folkman says, when Napoleone Ferrara's team at Genentech Inc. in South San Francisco showed that antibodies that arrest the activity of an angiogenic protein called vascular endothelial growth factor (VEGF) slow the growth of several types of tumors in mice.


    New therapy induces a blood vessel-filling clot in a mouse's tumor.

    P. Thorpe et al.

    After the Genentech discovery, several companies began developing drugs that block the action of VEGF or basic fibroblast growth factor, both of which stimulate angiogenesis by enhancing endothelial cell growth. Some of the drugs, which include antibodies to the growth factors as well as molecules that clog the growth factors' receptors on endothelial cells, have been shown to slow or stop the growth of tumors in mice; several are nearing clinical trials.

    And recently, researchers in several labs uncovered another potential target for growth-factor inhibitors, a new family of receptors, called TIE receptors (for tyrosine kinase with immunoglobulin- and EGF-like domains) that exist almost exclusively on endothelial cells. George Yancopoulos and his colleagues at Regeneron Pharmaceuticals Inc. in Tarrytown, New York, have found a completely new family of naturally occurring molecules that they call angiopoietins. These either activate or inhibit the TIE receptors, which in turn influence endothelial cell growth. “We are working to move them forward” as potential therapies for diseases where angiogenesis needs to be either encouraged or inhibited, says Yancopoulos.


    View this table:

    Other avenues. Blocking growth-factor stimulation of endothelial cells is just one way to stop blood-vessel growth. Indeed, some familiar drugs have as-yet unexplained antiangiogenic activity. The notorious tranquilizer thalidomide is enjoying a renaissance as a potential cancer therapy, thanks to the Folkman group's report in 1994 that it inhibits angiogenesis. How it does this isn't known, but its antiangiogenic properties may explain why limbs failed to form in some of the children born to mothers who took thalidomide as a morning sickness drug in the 1960s, says Folkman.

    Other drugs that were already being tested for cancer therapy have also turned out to be angiogenesis inhibitors. These include the immune-signaling molecule interleukin-12 (IL-12), which was being developed as a cancer drug by Hoffmann-La Roche in Nutley, New Jersey, and the Genetics Institute in Cambridge, Massachusetts, because it activates killer T cells to attack tumors. Gary Truitt of Roche noticed that the tumors in IL-12-treated mice were pale and seemed to lack blood vessels, and he worked with Folkman's group to show that IL-12 blocks capillary growth. In at least one experimental system—mice with melanoma—it seems that the angiogenesis-inhibiting effects of the drug, rather than its immune-activating properties, may be what keep tumors at bay, says Maurice Gately, the IL-12 research leader at Roche.

    Another important group of crossover drugs is the metalloproteinase inhibitors. These compounds block enzymes secreted by cancer cells that break down proteins of the extracellular matrix, enabling the cells to slip through the surrounding tissue and spread. Angiogenesis requires the same breakdown of the surrounding tissue to make way for new blood vessels, and the inhibitors have turned out to be “very potent” antiangiogenic compounds as well, says Henrik Rasmussen, vice president of research at British Biotech in Annapolis, Maryland, one of at least four companies developing metalloproteinase inhibitors for cancer.

    In an early clinical trial, British Biotech's drug, Marimastat, seemed to slow disease progression and prolong life in patients with advanced colorectal, ovarian, and pancreatic cancer. But it is unclear, Rasmussen notes, how much of the drug's success depends on its ability to block blood-vessel growth rather than on its direct effects on the movement of cancer cells.

    While all these drugs seem to slow or stop tumor growth, at least in animals, few researchers expected that they would actually shrink tumors that already had a blood supply. But newer results have surprised researchers and revised those expectations.

    David Cheresh and his colleagues at the Scripps Research Institute in La Jolla, California, reported 2 years ago that they could halt tumor angiogenesis in chick embryos and mice by blocking the function of a molecule called integrin αvβ3, which is found on the surface of endothelial cells. Integrins interact with the proteins of the extracellular matrix and, in so doing, help the cells differentiate, migrate, and divide—just what they need to do to form new blood vessels.

    When Cheresh's group treated animals with molecules that bind to and block the activity of integrin αvβ3, tumors in those animals not only didn't grow; in some cases they disappeared partly or completely. “This was the first indication that an angiogenesis inhibitor could have such a profound effect on the existence or proliferation of a tumor,” Cheresh says. And it spurred the group to organize preliminary clinical trials of several integrin αvβ3-blocking drugs, which will soon begin in patients with advanced colon, lung, breast, and prostate cancer.

    In retrospect, says Folkman, it's not surprising that blocking the formation of new blood vessels can shrink tumors. The entire capillary bed that feeds a tumor is continually remodeling itself, he says: Blood vessels “are being recruited, disappearing, and being recruited again. If you have a tumor that is even several grams, for it to stay at that size, it is always recruiting new vessels. When you stop the angiogenesis, you begin to get capillary dropout.” And the tumor, starved of blood, begins to shrink.

    Potent inhibitors. Perhaps the most dramatic tumor-shrinking drugs have come from an unlikely source: tumors themselves. Cancer researchers have known for many years that removal of a primary tumor sometimes causes a burst of growth in distant metastases, suggesting that the primary tumor had been making a substance that kept the metastases in check. In the late 1980s, Northwestern's Bouck and her colleagues made a discovery that hinted at what might be going on. Some tumor cells, they found, although able to induce angiogenesis in the tissue around them, for some unknown reason also make angiogenesis inhibitors. Apparently, the balance of inducers to inhibitors is favorable for vessel growth in the vicinity of the tumor, but at sites distant from the tumor, the inhibitors seem to prevail.

    One inhibitor Bouck's team found to be made by some tumor cells is thrombospondin, a protein also found in platelets and in the extracellular protein matrix. It “makes endothelial cells unable to respond to any [angiogenesis] inducer we have tried,” says Bouck. How it does that remains a mystery, although Scripps's Cheresh notes that thrombospondin in the extracellular matrix binds to cell surface receptors including integrin αvβ3. That interaction, he suggests, may interfere with the endothelial cells' ability to form new blood vessels.

    Bouck's team has since found that thrombospondin blocks the metastasis of melanoma cells into the lungs of mice. And Pat Steeg and her colleagues at the National Cancer Institute have shown that breast cancer cells engineered to make high levels of thrombospondin lose some of their ability to promote angiogenesis and to metastasize.

    Two years after Bouck's 1989 finding, Folkman and then-postdoc Michael O'Reilly identified and then eventually purified another angiogenesis inhibitor, which was produced in mice by a metastasis-limiting tumor called Lewis lung carcinoma. Last June, Folkman's team reported that the inhibitor, angiostatin, not only stopped experimental tumor growth in mice, but shrank the tumors dramatically. Tumors that had been allowed to grow to 400 milligrams—about 2% of the weight of the mouse—shrank to microscopic size with angiostatin treatment.

    And in today's issue of Cell, the Folkman team reports purifying yet another antiangiogenic factor, from a different metastasis-limiting tumor. This protein, which they call endostatin, is even more potent. It shrank a wide variety of tumors in mice to microscopic sizes, keeping primary tumors and metastases in check as long as the mice received the drug.

    The means by which angiostatin and endostatin block angiogenesis is unknown, says Folkman. But Cheresh suspects that, like metalloproteinase inhibitors, they interfere with the remodeling of the extracellular matrix that has to happen for new blood vessels to grow. He notes that both proteins are pieces of larger proteins that play key roles in the extracellular matrix. Angiostatin comes from plasminogen, which is normally cleaved into the matrix-digesting protein plasmin, and endostatin comes from collagen. Cheresh suggests that angiostatin and endostatin, as free-floating fragments of these proteins, may clog the active sites of the enzymes that chop up plasminogen and collagen, and so prevent endothelial cells from forming new blood vessels.

    No one knows whether the compounds will produce such dramatic results in humans, and clinical trials are still a few years off, says Folkman. Toronto's Kerbel cautions that it may be unrealistic to expect tumor shrinkage in humans, because human tumors will have been growing for much longer than the tumors treated in mice. “When you have tumors that may have been there for 5 to 10 years, the nature of the vasculature might be quite different,” he says, and less vulnerable to a blockade of angiogenesis.

    Still, the mouse results suggest that the new angiogenesis inhibitors will not suffer from the bane of conventional cancer chemotherapies: the development of resistance. When Folkman's team took their mice off endostatin, the tiny tumors began to grow back. But they remained responsive, shrinking again when the drug was readministered. As of 18 November, when Folkman reported these unpublished results during a conference at the Institute of Human Virology in Baltimore, the mice had been through six cycles of endostatin administration and withdrawal; each time the tumors grew to hundreds of milligrams and shrank back to microscopic size. That result is “unprecedented,” says Kerbel. “That would just never happen with chemotherapy.” If it's an indication of how endostatin would work in humans, says Folkman, it would mean treatment could be restarted if metastases were to grow back.

    They probably would, because tumors smaller than 1 to 2 millimeters don't need an induced blood supply but can live on existing blood vessels. In practical terms, that might mean that a patient would have to take such inhibitors continually to hold metastases in check, unless they could be eliminated by combining the treatment with conventional chemotherapy. But some clinicians worry about possible complications of long-term antiangiogenesis treatment, because blood-vessel growth is critical to processes such as wound healing.

    Well-placed clots. Despite all the attention that antiangiogenesis therapy is getting, it isn't the only means of attacking a tumor's blood supply. Thorpe and his team at the University of Texas Southwestern Medical Center report in this issue that they have had success with what might be called an antivascular approach. Their goal was to cause blood clots specifically in the blood vessels feeding a tumor and so starve the tumor of blood and oxygen.

    Thorpe's team injected the mice with cancer cells engineered to make interferon g, which induces nearby endothelial cells to make class II antigen, a protein they normally don't make. The researchers then took antibodies to class II antigen and linked them to a shortened version of tissue factor (TF), a protein that normally sits outside blood vessels and triggers blood clotting at injury sites. Because the TF they used is missing a membrane-attachment region, it is powerless to form clots. The team reasoned that by linking it to the antibody, they could direct it to the class II antigen on the walls of the tumor blood vessels, thus bringing the TF back in contact with the vessel walls' membranes and causing clots to form specifically in those vessels.

    The experiment “worked like a charm,” says Thorpe. The antibody-TF complex “homed in on the tumor vessels as expected and caused rapid, selective, and complete thrombosis of the tumor's blood supply.” Indeed, clots formed in the tumor vessels within minutes of the injection, and the tumors were largely dead within 48 hours. In 38% of the mice, the tumors disappeared completely, while in others a rim of cells around the outside of the tumor, which don't depend on the tumor's blood vessels, survived. But those cells, Thorpe says, are rapidly dividing and are “sensitive to conventional cancer chemotherapeutic drugs.”

    Thorpe's approach “has promise,” says Robert Tressler, a tumor biologist developing anticancer drugs at Chiron Corp. in Emeryville, California. But he notes that before it can be used in humans, researchers will have to make sure that the drug doesn't cause clotting in nontumor vessels. Thorpe is optimistic that he can do that, because blood vessels around tumors are much more disposed toward blood clotting than normal vessels are. Even when his team deliberately mistargeted the truncated TF to a highly expressed marker on normal vessels, he says, they saw virtually no clotting.

    Thorpe points out that the tumors given to the mice had been engineered to induce the class II marker on the tumor-associated blood vessels, something that could not be done in human cancer patients. But, he says, there are several candidate proteins on newly formed blood vessels that could be used as targets in humans. “This isn't pie in the sky,” Thorpe says. “There are very real opportunities for use of these agents in humans.” He hopes to have the approach in clinical trials within 2 years.

    Whether antiangiogenic and antivascular therapies turn out to be pie in the sky or a realistic new approach for cancer rests on the results of the host of clinical trials planned for the next few years. But at this stage at least, a lot of researchers and companies are placing their hopes on this new strategy for cutting the lifeline to tumors.

Stay Connected to Science