Extensive variability in the composition of immune infiltrate in different mouse models of cancer.

Mouse models are invaluable tools for cancer immunology research. However, there are differences in the immune response to the tumour depending on the model used, and these differences are not often characterised on their own. Instead they are often only analysed in response to a therapeutic immune modulation. There are important issues with translatability into effective clinical research when considering the choice of mouse models. Here we analysed the tumour immune microenvironment and modified aspects of the tumour model to determine the effect on the composition of the immune infiltrate. Mice injected subcutaneously with the melanoma cell line, B16-OVA, had a higher frequency of T cells, especially CD8+ T cells, than mice injected subcutaneously with CT26 colorectal adenocarcinoma cells. We compared the same tumour cell line (CT26) delivered either subcutaneously and intracaecally. To minimise immunological impacts due to the invasive surgery procedure, we optimised an existing intracaecal injection protocol. Intracaecal tumours had a higher frequency of infiltrating CD3+ CD4+ T cells and a lower frequency of CD3-CD19- (putative NK cells) than subcutaneous tumours. In contrast, there was a higher frequency of F480+ macrophages in subcutaneous tumours than intracaecal tumours. These data demonstrate that variability between animals, between experiments and within tumour models, can lead to difficulty in interpreting the infiltrating immune response and translating this response to clinical research.