Immune escape at the onset of human colorectal cancer

The evolutionary dynamics of tumor initiation remain undetermined, and the interplay between neoplastic cells and the immune system is hypothesized to be critical in transformation. Colorectal cancer (CRC) presents a unique opportunity to study the transition to malignancy as pre-cancers (adenomas) and early stage cancers are frequently detected and surgically removed. Here, we examine the role of the immune response in tumor initiation by studying tumor-immune eco-evolutionary dynamics from pre-cancer to carcinoma. The integrated approach uses a computational model, ecological analysis of digital pathology, and multi-region neoantigen prediction. Model results indicate that there are several routes to malignancy, each of which uniquely shapes the tumor ecology and sculpts intra-tumor antigenic heterogeneity (aITH). These routes include combinations of evading detection via accumulating mutations with low antigenicity, the ability to block immune attack (e.g. PD-L1), and the ability to recruit immunosuppressive cells. Modeling predicts that, in general, the most common route from benign to malignant is the construction of an immunosuppressive niche. To determine which route is dominant in CRC initiation, we used a cohort of 21 colorectal adenomas, 15 carcinomas, and 26 adenomas with a focus of carcinoma (“ca-in-ad”) cases. The immune microenvironment was characterized using the spatial distribution of 17 markers across registered whole-slide images at 40x magnification, while patterns of intra-lesion aITH were described using multi-region neoantigen prediction. Observed changes in aITH, the tumor ecology, and spatial patterns of both cell associations and gene expression are consistent with simulations where immunogenic adenomas do not progress to CRC because they are under immune control. Conversely, adenomas that progress initially avoid detection through low immunogenicity, but gradually construct an immunosuppressive niche isolated from CD8+ cytotoxic T cells, thereby evading immune elimination and allowing for an increase in neoantigen burden. Both modeling and data indicate that immune blockade (e.g. PD-L1 expression) plays a secondary role to immune suppression in tumor initiation or progression. These results suggest that re-engineering the immunosuppressive niche may prove to be a most effective immunotherapy in CRC.