Investigating the innate and adaptive immune response in patients with metastatic colorectal cancer

Abstract

The immune response is strongly associated with outcome in CRC (stages I-III). Cytotoxic CD8+ T-cells are the most important subset of immune cells positively associated with outcome, in most solid malignancies and especially CRC. However, in the advanced stage of CRC, this is not always the case. Stage-IV CRC metastasises (mCRC) commonly to the liver, which this thesis addresses. The gold standard for treatment of colorectal liver metastasis (CRLM) is surgical liver resection. Indeed, improvements in surgical techniques have greatly improved the 5-year survival of these patients; however, up to 60% of patients still recur following surgical liver resection. Understanding the progression of mCRC in the context of the immune response is the main focus of this thesis.

To investigate the immune response at the primary site of CRC, a unique retrospective cohort of de novo or synchronous mCRC patients (n=109) was explored. Included in this cohort were patients that had microsatellite unstable tumours (MSI) (n=12), which in the metastatic setting have a reduced overall survival (OS) and have been found to respond to checkpoint blockade inhibition (CBI). I analysed the primary tumours of these patients in the context of tumour infiltrating lymphocytes (TILs), immune escape mechanisms and oncogenic potential of these primary tumours, where the patients had synchronous metastatic disease. Despite high frequencies of cytotoxic CD8+ T-cells in some tumours, there was no association with OS, indicating the tumour had surpassed immune control. Expression of PD-L1 >1% on tumour cells was independently correlated with OS in multivariate analysis suggesting that tumour cells have the ability to progress in part at least by evading the immune response.

To evaluate the immune response at the metastatic site in the liver, a prospective cohort of CRLM patients was recruited (n=11) to examine the immune context in these tumours. Patients undergoing liver resection were included in the study, and freshly isolated lymphocytes from the tumour; normal liver and peripheral blood were analysed by flow cytometry. These tumours were found to have a reduced infiltration of cytotoxic CD8+ T-cells and increased infiltration of CD4+ T-cells, including T-regulatory cells that are known to suppress immune responses. This immune milieu in these tumours alludes to a reduced cytotoxic and increased immunosuppressive environment. To investigate the functional capacity of these immune cells, a novel immune cytotoxic assay was developed. This assay involved co-culturing patient-derived tumouroids with expanded autologous TILs to assess dynamic interaction of function of these cells. TILs expanded from these tumours were able to kill matched tumouroids, further indicating that when removed from the immunosuppressive TME these cells have functional ability to kill tumouroids. To assess if these TILs respond to CBI, addition of anti-PD-1 antibody was included in the co-culture assays where no improvement in killing was observed. Further investigation of other immune cell subsets in CRLM tumours was undertaken to gain an insight into other immune cell populations that may contribute to the tumour microenvironment (TME).

One population of unconventional T-cells abundant in the liver are mucosal-associated invariant T-cells (MAIT cells). These cells play a role in bacterial infections and bridge a gap between innate and adaptive immune responses, and their role in tumour immunity is less defined. These cells are of interest in the context of the TME as they have cytotoxic capability and rapid produce cytokine. When assessing MAIT cell presence in CRLM (n=25) by flow cytometry, MAIT cell frequency was reduced in the tumour compared to surrounding normal liver. The phenotype of MAIT cells in the tissue was phenotypically distinct compared to the periphery, with high expression of PD-1 and CD69, both markers of activation. To assess the potential of MAIT cells, peripheral MAIT cells were isolated from healthy donors and co-cultured with patient-derived tumouroids in an unstimulated and stimulated state. MAIT cells in both states were able to kill patient-derived tumouroids. This is the first documentation of MAIT cell killing with patient-derived material. Despite PD-1 expression on these cells, addition of anti-PD-1 antibody did not enhance this killing. Even though at reduced frequency in the tumour, MAIT cells are activated and may contribute to the tumour microenvironment (TME) in CRLM.

Patients with advanced-stage CRC have a reduced survival compared to earlier stage CRC patients. In the context of the immune response, it is evident that these tumours have evaded the immune response to progress to metastasis. The work of this thesis highlights that at the primary tumour site of patients with de novo mCRC, despite high frequencies of cytotoxic CD8+ T-cells, there is an inability at controlling tumour progression. This has likely arisen through immune evasive mechanisms. Therefore there should be a focus on improving the immunogenicity of these tumours to again be recognised by the immune cells. Secondly, at the metastatic site, the TME is immunosuppressed and reinvigorating the function of cytotoxic immune cells present may restore improved immune responses. Importantly, understanding the immune biology of these tumours will provide greater guidance to improve potential immune therapies for these patients into the future.