In the development of targeted cancer immunotherapies, the choice of antigen is obviously critical to the design of any therapeutic strategy, but particularly so for tumor vaccines, which must distinguish malignant cells from normal cells. Investigations a decade ago focused on mutated tumor antigens, or viral tumor antigens, with the belief that these foreign or abnormal proteins would be best recognized by the host immune system. Within the last 10 years, however, several tumor antigens have been identified on the basis of recognition by infiltrating T cells in tumor samples. Studies on melanoma, in particular, have revealed that in addition to some mutated tumor antigens, several aberrantly expressed normal proteins, as well as tissue-specific differentiation factors, are recognized by the host immune system. Similar studies in other solid tumors have revealed that certain oncogenes overexpressed in malignant cells, such as p53 and HER-2/neu, are also recognized by host T cells. Our group has been investigating the HER-2/neu oncogenic protein as a vaccine target in patients with HER-2/neu-overexpressing cancers. However, several issues unique to the design of human clinical trials of cancer vaccines must be addressed when translating preclinical experiments to human clinical trials. First, HER-2/neu protein expression can vary depending on the tumor type. How would expression differences impact clinical trial design? Secondly, what are the issues in clinical trial design that are critical to the successful execution of a phase I study of a peptide-based vaccine? Thirdly, what types and amounts of clinical material are readily available for immunologic analysis and can be obtained with little distress and risk to the patients enrolled in the study? Finally, what steps must be implemented for a laboratory assay to evolve to meet the validation criteria needed for application as an immunologic monitoring tool?
