these studies. In general, these trials did not produce convincing evidence of 
therapeutic efficacy when they were appropriately controlled (45). 
The testing of human cancer vaccines involves uncertainties ranging from 
appropriate type, dose, route, and frequency of vaccine to suitable patient 
selection and evaluation of clinical response. Given the array of variables, a 
sequential analysis of effectiveness, if it were evaluated on clinical tumor 
response alone, would be extremely time consuming. What is required for the 
development of an immunogenic cancer vaccine are methods to assess 
effectiveness that are both rapid and objective and that guide the step-by-step 
process of vaccine construction and testing. With regard to vaccines against 
infectious diseases, serologic responses to bacterial and viral antigens have been 
an essential step in their development. The lack of comparable tests of humoral 
and cellular immunity to monitor the immunogenicity of cancer vaccines in 
humans has been a major impediment to investigating this approach to cancer 
therapy. 
Beginning with the 1970’s, attempts were made by many groups to demonstrate a 
specific immune response to vaccination in patients receiving cancer vaccines. 
Most of these studies were done in patients with melanoma, and involved tests 
for delayed cutaneous hypersensitivity, lymphocyte cytotoxicity, lymphocyte 
migration inhibition, and a variety of serologic assays. Because of the difficulties 
that were encountered in the analysis of the specificity of allogeneic reactions (in 
most cases the sera or lymphocytes tested and the melanoma cells used for 
vaccine construction or target cell preparation did not come from the same 
patient), those studies did not provide unequivocal evidence of a tumor specific 
immune response to vaccines (45). 
A clearer picture emerged when the analysis was restricted to autologous 
serologic reactions. Although this approach depended on establishing tumor cell 
lines from each patient to be studied, the testing of autologous combinations of 
tumor cells and sera eliminated the confusing reactions to normal allogeneic cell- 
surface antigens. Autologous typing defined three classes of surface antigens on 
melanoma cells. Class I antigens are restricted to the autologous tumor. Class II 
antigens are shared tumor antigens that are found on autologous as well as 
allogeneic tumor cells; some class II antigens are found on a restricted set of 
normal cell types and therefore, can be considered autoantigenic differentiation 
antigens. Class III antigens are widely distributed on malignant as well as normal 
cells (46). With the development of autologous typing systems for defining cell- 
surface antigens of melanomas, serologic tests of requisite sensitivity and 
specificity were available that could be used to gauge the immunogenicity of 
melanoma vaccines. A series of trials with vaccine containing class I or II 
melanoma antigens showed, however, that an antibody response to the relevant 
class I or II melanoma antigens were induced by vaccination only in exceptional 
Recombinant DNA Research, Volume 15 
[889] 
