WASHINGTON - For the next three years, Doug Pfaff and Rand Snell, both sexually active gay men, will roll up their sleeves for injections that scientists hope will protect them against AIDS.

Pfaff, a 39-year-old salesman for a computer training company from Fredericksburg, Va., and Snell, 43, a public relations consultant from Washington, have volunteered to be guinea pigs in the first large-scale test of an AIDS vaccine.

Both men are part of a national experiment that by next June will involve 5,000 high-risk participants, mainly sexually active homosexual men. Each volunteer will be inoculated seven times over 36 months with an experimental inoculation called AIDSVax, and all will have regular blood tests for any signs of HIV infection they may contract through unsafe sex.

The program, managed by vaccine maker VaxGen of South San Francisco, with help from the National Institutes of Health, is still seeking volunteers in the gay community. About 1,000 people have signed up so far.

San Francisco will be the next large city to join the effort. Two clinics - at the AIDS office of the Department of Public Health and at San Francisco General Hospital - will begin recruiting 300 volunteers sometime next month. The principal investigators are Drs. James Kahn of San Francisco General and Susan Buchbinder of the city Department of Public Health.

Of the 5,000 volunteers nationwide, about 3,300 will receive the vaccine, and the remaining 1,700 will receive a placebo consisting of a harmless mixture of sterile water and aluminum hydroxide.

Ethical dilemma

The researchers face this dilemma: There is no way to determine whether the vaccine is effective unless the volunteers are exposed to the HIV virus, which causes AIDS. On the other hand, medical ethics prevent researchers from promoting risky sexual behavior - in fact, volunteers are warned against unsafe sex when they sign up.

Underlying the research is the assumption that participants will stray from the safe-sex regimen and have unsafe sex sometime during the trial.

"They make it very clear from day one that we know what safe sex is," said Pfaff. "But they don't want anybody necessarily to change their behavior." Margaret McCluskey, the trial's project director for the Washington area, stressed that the volunteers were counseled regularly on safe sex practices. Yet, she said, "We all know that knowledge doesn't always change behavior. We are relying on human nature."

Both Pfaff and Snell say they practice safe sex and will continue to do so during the trial. But Snell acknowledged that, in the heat of the moment, some gay men opted for unprotected sex at times.

"We all take risks," he said. "Hopefully, they are educated risks. It's a reality that not everyone is perfect all the time."

Despite safe-sex counseling and the good intentions of participants, researchers estimate that during the three-year experiment, 1.5 percent of the placebo group will be infected with the deadly virus each year.

At the end of the trial, researchers will compare the infection rate of the placebo group with that of the vaccine group. If AIDSVax is effective, the vaccine group should have a lower infection rate.

"Anything under 1.5 percent, we'd be very happy about," said McCluskey.

"We'd have to ask, "Did the vaccine do that?'"

The expected 1.5 percent infection rate in the placebo group is based on previous studies, by the NIH and the Centers for Disease Control and Prevention, of gay men counseled on safe sex.

Risks of testing

Historically, vaccines have proved to be an extremely effective weapon against other diseases. Such once-deadly scourges as smallpox and polio have been all but wiped out in the developed world. Those successful vaccines were perfected in similar clinical trials, in which one group received the vaccine - and ended up healthy - and a control group got a placebo and ended up contracting the disease being studied.

Before being accepted onto the trial, volunteers have to meet several criteria:

*They must be sexually active with other men.

*They must be involved in a monogamous relationship with an HIV-positive man or have had sex with three or more men during the last year and have had anal sex during that period.

*They must be HIV negative at the outset and pass a physical exam.

Volunteers are required to read and initial each page of a 12-page consent form that warns them that they shouldn't assume the vaccine protects against HIV: That's what the experiment is attempting to determine. An informational pamphlet tells volunteers, "You will not know if you are receiving the placebo or the study vaccine."

Although the experiment focuses on gay men, who account for the majority of new HIV infections each year in the United States, it also includes a small percentage of sexually active women because they account for a growing segment of new cases. Female volunteers must have contracted a sexually transmitted disease sometime in their lives and have had at least two male sex partners in the last year.

The company and the Thai government are conducting a separate large-scale experiment in Thailand to determine the effectiveness of the vaccine against AIDS transmitted through intravenous drug use.

The U.S. trial is "double blind," meaning that neither the volunteers nor the experiment's administrators, nor the vaccine's manufacturer knows whether a volunteer is receiving the vaccine or the placebo.

Snell decided to volunteer because of the terrible toll AIDS has taken on the gay community. "I've had a lot of friends who have died," he said. "I felt that the potential benefits of the trial were significant. It's an opportunity to participate in something that expands the basic knowledge of science."

Not 100% effective

Bioethicists worry that if the vaccine shows only partial effectiveness, it could spark an increase in risky sexual behavior.

"When a vaccine is eventually licensed for widespread use, it is very unlikely to be 100 percent effective," Michael Langan and Chris Collins of the AIDS Research Institute at UC-San Francisco wrote in "Paving the Road to an HIV Vaccine," a paper published Dec. 1.

The experiment, a so-called Phase III human clinical trial, is the final rung on the research ladder before the federal Food and Drug Administration grants approval for widespread use. Phase III experiments are designed to test the overall safety and effectiveness of drugs that have already shown some benefit in prior studies.

Phase I and Phase II government-approved experiments showed earlier that AIDSVax protected chimpanzees from HIV infection and sparked an immune system response in humans.

"The big question is: How effective is it in real-life situations?" said Dr. Donald Francis, president of VaxGen and a former AIDS specialist at the CDC.

Researchers already are dampening expectations.

"We know that the vaccine didn't work 100 percent in earlier trials," said Joe Wright, a community coordinator for San Francisco's Department of Public Health AIDS office.

A key reason is that the vaccine isn't engineered to protect against all strains of HIV, a virus that mutates easily. "People should see this vaccine trial as a larger process toward a successful vaccine at some point," Wright said.

One deterrent to getting more volunteers, VaxGen's Francis said, has been a fear that the experimental vaccine may actually infect volunteers with HIV. Scientists stress that that's impossible. The vaccine is derived from a synthetic protein, not a live virus. In addition, the National Institute of Allergy and Infectious Diseases has already conducted six small trials on the vaccine to ensure its safety.

"There is zero chance of infection," said Francis, who himself has been inoculated with the vaccine.

For more information on the San Francisco vaccine trials, call (415) 554-9068.