Some long-term survivors of HIV infection produce rare and extremely potent antibodies that keep the disease from progressing to AIDS, and might point to a way to protect uninfected people from the virus, researchers reported yesterday in the closing hours of the 17th International AIDS Conference in Mexico City.

The antibodies, against a particular part of a much-studied HIV protein called gp120, might prove useful as a microbicide for blocking infection during sexual intercourse. If researchers could find a way to prompt the immune system to make its own supply of the antibodies before encountering the virus, they would have a vaccine.

"I think the road is long before we reach that point," cautioned Stephanie Planque, the researcher who presented two studies.

The search for both an AIDS microbicide and a vaccine has been particularly frustrating. None are in use, and some candidates tested in recent years have turned out to increase the risk of infection.

The antibodies described yesterday attack a small and crucial region of HIV's outer shell where the virus binds to its chief prey, immune-system cells called lymphocytes. Acting as an enzyme, the antibody clips the attachment point, and falls away undamaged and ready to do the job again.

These "catalytic antibodies" have been isolated from people with lupus, a disease in which the immune system malfunctions and produces a large number of unusual antibodies. Catalytic antibodies have also been seen in some long-term survivors of HIV infection, including three people with the blood disease hemophilia, whose cases were described yesterday.

if ( show_doubleclick_ad && ( adTemplate & INLINE_ARTICLE_AD ) == INLINE_ARTICLE_AD && inlineAdGraf ) { placeAd('ARTICLE',commercialNode,20,'inline=y;',true) ; }

Immunologists have paid great attention to the gp120 protein for more than a decade. The virus uses it to dock to two receptors on the surface of lymphocytes, starting the process of infection.

People infected with HIV make antibodies against gp120, but in most cases they are not enough to stop or slow the infection. A vaccine consisting of multiple injections of purified gp120 protein did not protect people in a large clinical trial run earlier this decade.

Recent research has focused on a 13-unit long stretch of the 500-unit protein, exactly where the virus attaches to the cells' receptors. (Each unit is an amino acid, which are strung in a chain like beads on a necklace to form the protein.)

Among the laboratories doing this work is one run by biochemist and immunologist Sudhir Paul at the University of Texas Medical School in Houston, where Planque is a graduate student. Others include labs at the National Institutes of Health and the Scripps Research Institute in California.

This small section of gp120 is identical in virtually all strains of HIV. But because it is partly hidden by other sections of the protein, the immune system "sees" it poorly and does not make good antibodies against it -- except in rare cases.

One of those exceptions is in lupus patients. AIDS researchers noticed years ago that HIV infection rarely occurred in them. "One hypothesis is that they mount an immune response that protected them," Planque said yesterday.

She screened a library of antibodies made by lupus patients, looking for ones with catalytic activity against the 13-unit stretch of gp120. She found some.

In one of yesterday's late sessions, she told scientists at the conference that lupus-derived antibodies killed HIV samples from five clades, or families, of the virus. In a related presentation, she described how antibodies purified from three people with hemophilia, who had each lived for more than 17 years with HIV infection, did the same.

The broad effect of the antibodies is important. The AIDS virus differs somewhat from one geographical region to another. To be useful, a microbicide or vaccine would have to protect against all the clades.

Antibodies are already in use as drugs for several chronic diseases, such as rheumatoid arthritis, but they are extremely expensive, in part because they must be given in large doses.

However, Paul, Planque's supervisor, speculated that a microbicide containing antibodies might be affordable even for people in the developing world. That is because only very small quantities of material would be needed to protect against sexual transmission, and the protection would have to last only hours, not days or weeks.

Finding a way to stimulate the immune system to make catalytic antibodies will be much harder, he said. Paul is researching strategies for "presenting" the crucial part of the protein to the immune system in a more effective way than occurs naturally. If successful, such a strategy could result in a useful vaccine.

Curiously, everyone already makes the catalytic antibodies against gp120 in small quantities. That is because they are among the roughly 2,500 antibodies whose production is directly driven by genes and, thus, part of the immune system's innate repertoire.

The immune system can make millions of more antibodies, capable of attacking virtually any foreign substance. But that process involves being exposed to the foreign substance, called an antigen, a complicated process of gene-shuffling, and a time-consuming refinement of antibody production by the immune system.

In contrast, the anti-gp120 catalytic antibody is part of an ancient, hard-wired protective response, whose evolutionary origins are largely a mystery. Whatever they are, the body's output of them is not quite good enough to protect most people from HIV.

Paul and others hope that, with some tinkering, it might be made so