4O - Enhanced anti-cancer vaccines with a new epitope improvement system

Date 04 November 2016
Event ESMO Symposium on Immuno-Oncology 2016
Session Proffered paper session
Presenter Cristian Capasso
Citation Annals of Oncology (2016) 27 (suppl_8): viii1-viii2. 10.1093/annonc/mdw525
Authors C. Capasso1, A. Magarkar1, V. Cervera Carrascon2, M. Müller1, M. Garofalo1, L. Kuryk1, M. Fusciello1, E. Ylösmäki1, A. Bunker1, V. Cerullo1
  • 1Division Of Pharmaceutical Biosciences, University of Helsinki, 00710 - Helsinki/FI
  • 2Cancer Gene Therapy Group, University of Helsinki, Helsinki/FI



Cancer vaccines induce an immune response against specific tumor antigens, which might also be self-antigens. Hence T-cells are either not trained to recognize them (i.e. low-avidity TCRs) or tolerant towards them. Previous studies have shown that mutated-epitopes, that resemble the original ones, are able to break the tolerance by exploiting the cross-reactivity of T-cells. Nevertheless, the selection of these analogue peptides is cumbersome and time consuming. We hypothesized that an optimized in silico framework might improve and speed up the whole selection process.


Our Epitope Discovery and Improvement System (EDIS) studies wild type epitopes and predict mutated forms suitable for cancer therapy. The novel aspect is the ability to interrogate different prediction servers and screen the whole mutational library. In addition, molecular dynamics simulations are performed on lead candidates to investigate the interactions between the peptides and the MHC-I binding pockets.


Starting with the model epitope SIINFEKL, we chose two mutated forms with an improved MHC-I binding affinity. First, we confirmed experimentally this finding; hence, we demonstrated in vivo that the two mutated forms are more effective in reducing the growth of established B16OVA tumors. We studied the epitope SVYDFFVWL from the tyrosinase-related protein 2 melanoma antigen. The in silico screening revealed that mutations at positions 8 or 7 would increase the MHC-I affinity, regardless of the chosen amino acid. On the contrary, mutations in positions 1, 2 or 3 would result in a lower affinity. Hence, we selected two analogues and we tested them by treating aggressive B16F10 tumors. Surprisingly, one of the two analogues reduced significantly the growth of the tumors compared to the native TRP2 epitope. The immunological analysis revealed that therapeutic vaccination with the 7A analogue increased the response against the wild type epitope, suggesting the presence of a beneficial cross-reactivity.


In conclusion, we developed an in silico framework that can unravel the properties of known MHC-I epitopes and improve their sequences to achieve an increased cross-reactivity among T-cells and circumvent the tolerance that often impairs the efficacy of cancer vaccines.

Clinical trial identification

Legal entity responsible for the study

University of Helsinki


University of Helsinki, Tekes


All authors have declared no conflicts of interest.