Defining key parameters of therapeutic activity using mouse cancer neoepitopes.

Cancer neoepitopes have emerged as strong candidates as cancer vaccines. Mouse models of cancer neoepitopes often test their activity in models of prophylaxis, while the human setting is always in therapy of preexisting disease. We have previously identified 7 MHC I-presented cancer neoepitopes of the BALB/c sarcoma Meth A by mass spectrometry; 3 of the 7 neoepitopes showed anti-tumor activity in assays of prophylaxis. Here, we test these neoepitopes for activity in treatment of preexisting 3-, 10- and 17-d old cancers, and observe that: (i) activity in prophylaxis is not predictive of efficacy in treatment of preexisting cancers. (ii) The higher activity of a neoepitope in therapy, albeit not in prophylaxis, correlates well with the high precursor frequency of CD8 T cells. (iii) Co-immunization with CD4 and CD8 neoepitopes is essential for optimal therapeutic efficacy. CD4 help can be obtained from an MHC II-restricted neoepitope of the tumor, or from an irrelevant MHC II-epitope. (iv) The CTL and helper epitopes, linked or un-linked to each other, are equally effective. These insights stress and stretch some of our current suppositions and shall inform designs for use of neoepitopes in cancer treatment.