GM3 is a ganglioside that has been biochemically identified as dominating the cell surface of several human tumours, but is also found on human normal cells at much lower density. Since GM3 is widely distributed in essentially all types of animal cells, there is a conflict with the concepts of tumour-associated antigen, immunogen, and toxicity. We have designed a GM3-based cancer vaccine for the treatment of human breast and melanoma tumours. Prior to the Phase I clinical trial, we carried out a 12-month dose repeated toxicity study in five male Macaca fascicularis monkeys. Four male monkeys were treated with placebo in a similar way. During the study, no differences were observed between control and treated monkeys related to daily clinical observations (other than local damage) including rectal temperature, blood pressure, respiratory and cardiac rates, weight gain, biochemical and hematological parameters (with the exception of transitory pathological changes), and anti-DNA and anti-nuclear antibodies, although treated monkeys consistently developed both IgM- and IgG-specific anti-GM3 antibodies. Sixty per cent of treated monkeys developed moderate local reactions at the injection site, which disappeared without sequels. We concluded that this GM3 cancer vaccine overcame in monkeys the natural tolerance to GM3 ganglioside evidenced by a strong immune response, while the local reactions elicited-were transitory without apparent important systemic toxicity effects.