Abstract 753: In vitro veritas: Successful prediction of clinical immunogenicity of two cancer vaccines by a high-throughput in vitro priming platform

Abstract

Abstract Identification of tumor-associated peptides (TUMAPs) by HLA ligandome exploration or by means of reverse immunology allows the identification of high numbers of candidate substances for therapeutic cancer vaccines. However, the immunogenicity of such peptide antigens has to be validated by a separate step. Animal models are generally not appropriate for this task due to the high species-specificity of a) the HLA system, b) the T-cell repertoire and c) the expression of tumor antigens. Testing the immunogenicity of TUMAPs in a preclinical setting therefore requires the use of human cells. We assume that central tolerance (i.e. the deletion of T cells with distinct specificity to self-antigens during T-cell maturation) will be a major mechanism shaping the immunogenicity of TUMAPs and that “holes” in the T-cell repertoire may be identified by comparing in vitro priming efficiency of TUMAPs. Unfortunately, classical approaches for in vitro priming of human T cells are often difficult to standardize and scale. Here, we show that highly parallel and standardized in vitro priming using fixed numbers of isolated T cells and artificial antigen presenting cells (aAPCs) is scalable to determine the in vitro immunogenicity (% positive wells or derived estimated T-cell precursor frequencies) for hundreds of different TUMAPs from multiple donors. The method has been adopted to include the HLA alleles A*0201, A*2402, B*0702 and A*0301. We have validated the platform by two steps: First, the in vitro immunogenicity for literature-known natural ligand / altered ligands pairs (from gp100, tyrosinase and CEA) was determined and found to confirm previously reported results, with the altered ligand being the more immunogenic form in all cases. Second, the in vitro immunogenicity for individual class I binding TUMAPs as determined with the method described herein was used for the selection of two multi-peptide cancer vaccines, IMA901 for renal cell cancer and IMA910 for colorectal cancer. The drug products contained N=9 and N=10 class I binding TUMAPs, respectively. After clinical testing in two first-in-man studies (IMA901-101: N=27 and IMA910-101: N=65 evaluable patients), the in vivo immunogenicity measured in both studies (% of patients with vaccine-induced responses) was found to strongly correlate with the previously determined in vitro immunogenicity (Spearman correlation, r=0.82, p<0.05 and r=0.95, p<0.001). Recently, the in vivo immunogenicity of the peptides contained in IMA901 was confirmed by a further phase II study (IMA901-202: N=61 evaluable patients). To conclude, we believe that this method will be useful for the development of multi-epitope vaccines with high immunogenicity that can elicit broad immune responses in cancer patients. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 753. doi:10.1158/1538-7445.AM2011-753