HLA typing demands for peptide-based anti-cancer vaccine.

Immunological treatment of cancer has made some very promising advances during the last years. Anti-cancer vaccination using peptides or peptide-pulsed dendritic cells and adoptive transfer of in vitro generated, epitope-specific T cells depend on a well-fitting interaction of HLA molecule and epitope. Accurate HLA-typing is a key factor for successful anti-cancer vaccination. No comprehensive data and no suggestion exist on the HLA-typing in this setting. We performed a systematic review of PubMed analyzing HLA-typing data in cancer vaccination trials over the last 4 years (2004-2007). Then, using the SYFPEITHI database, we calculated the peptide binding prediction of the eight most often used HLA-A*0201 binding epitopes. Finally, high-resolution typing [by sequence-specific primers (SSP)] data of a HLA-A*02 or HLA-A*24 positive population in Berlin, Germany, were analyzed. Forty-five cancer vaccination trials with 764 patients were included. Eighteen studies were performed in the USA, 13 in Europe, 12 in Asia (mainly Japan), and two in Australia. Most common diseases targeted were melanoma, prostate cancer, colorectal cancer, renal cell cancer, and breast cancer. The trials tested protocols using peptide plus adjuvants without DC or protocols using peptide-pulsed DC. In 38 trials (84%) HLA-A2 positive patients were vaccinated, in 11 studies (24%) HLA-A24 positive patients were vaccinated. Nineteen studies with 291 patients (38%) presented the HLA type as four-digit code (high-resolution), 26 studies with 473 patients (62%) presented the HLA-type in a low-resolution code. The method of HLA determination was given in six out of 45 trials (13%). Using the SYFPEITHI database we calculated the peptide binding prediction of the eight most often used HLA-A*0201 binding tumor antigen-derived epitopes for binding to HLA-A*0203. While the epitopes had a binding score of 17-28 for HLA-A*0201, the score for binding to HLA-A*0203 was zero in seven out of eight tested peptides. Only for one peptide the score was eight. Finally, we analyzed high-resolution data of HLA-A*02 and HLA-A*24 positive patients in Berlin, Germany. We found the HLA-A*0201 allele and HLA-A*2402 allele in 95%, respectively. HLA-A*0201 and HLA-A*2402 are most commonly used for peptide based vaccine in cancer. Data on HLA-typing given in the included cancer vaccine manuscripts are fractional. Only 13% report the method of HLA typing and most HLA types are given as low-resolution code. Looking at the binding of specific peptides to both the alleles, it is important to perform high-resolution typing. Further suggestions for immunogenetic laboratories and clinical tumor immunologists regarding HLA-typing for cancer vaccine trials and adoptive T cell transfer approaches are discussed.