A new vaccine has proven partially effective in protecting mice from malaria, apparently by inducing a type of immunity that may be a key to human vaccines for AIDS, leprosy and other still-unpreventable diseases.

Researchers from the Walter Reed Army Institute of Research described the oral mouse vaccine Thursday in an article in the journal Science. The vaccine was made by inserting a gene from the malaria protozoan into Salmonella bacteria that have been attenuated, or altered so they do not cause disease.The team of 11 scientists saw the genetically engineered vaccine dramatically increase the immunity of mice to malaria, apparently by inducing a response known as cell-mediated immunity that is believed to be the human body's main protection against many diseases that have eluded vaccine developers.

Dr. Jerald Sadoff, who led the research team, explained in a telephone interview, "There are lots of diseases where cellular immunity is important malaria, AIDS, parasitic diseases, tuberculosis, leprosy, and the rikettsial diseases like Rocky Mountain spotted fever, scrub typhus and other scourges that have wiped out entire armies.

"There is a whole series of classes of disease that we could protect people against if we could reliably produce cellular immunity," concluded Sadoff, who heads the bacterial diseases department at the Army research facility.

The scientists said they still need to improve the effectiveness of the mouse vaccine, which immunized 54 percent of the mice compared to 9 percent of those that received a Salmonella vaccine with no malarial genes. Still, they already are working on a human malaria vaccine, and Sa-doff said a candidate might be ready for preliminary tests by the end of the year.

The researchers are also trying to make a vaccine against the deadly acquired immune deficiency syndrome by inserting AIDS viral genes into the attenuated Salmonella. Prog-ress on that will be slower, however, because it is uncertain which part of the AIDS virus might trigger cellular immunity, Sadoff said.

"It's a long way away. That's the critical part of this puzzle," agreed Dr. Scott Koenig, a senior staff fellow at the National Institute of Allergy and Infectious Diseases.

Koenig, who had not yet read the Sadoff team's paper, said if the researchers indeed had induced cell-mediated immunity in the mice, "it is an interesting approach with application to other diseases."

"It's a big extrapolation from talking about Salmonella and malaria to AIDS," Koenig cautioned, "but just to find an appropriate way of treating malaria would be an accomplishment."

Worldwide there are 200 million cases of malaria, Sadoff said, and 2 million people, mostly children, die of the mosquito-borne sickness each year.

Sadoff said traditional vaccines against other diseases have worked primarily by inducing a humoral immune response, in which the body produces antibodies in the blood after being injected with killed cells, or pieces of the outside of the disease-causing organism.

Such immunity alone, however, has not proved effective against AIDS, malaria and other protozoan, fungal, bacterial or viral invaders of human cells.