Predicting clinical outcome in melanoma and breast cancer using blood and tumour-based immune biomarkers: a means to monitor anti-cancer immune responses

Abstract

In addition to conventional surgery, chemotherapy and radiotherapy, immunotherapy holds a great deal of promise as an effective form of cancer treatment. This utilises the patient’s own immune system to control tumour growth and thus relies on the presence of immune cells which can recognise and subsequently kill or otherwise control cancer. This notion is supported by many studies showing that the presence of T cells recognising tumour-associated antigens (TAAs) is associated with superior survival in a number of cancer types. In contrast, cancer patients with high levels of immune suppressor cells experience worse survival and respond more poorly to therapy. These studies demonstrate that the balance between different immune populations is closely associated with clinical outcome. Therefore, the aim of the work in this thesis was to identify new blood- and tissue-based prognostic markers that more accurately predict patient outcome. Monitoring patients for their possession of TAA-reactive T cells may be important for this, but standardising such biological assays is challenging. Therefore, the first part of this work attempted to identify surrogate markers that may more easily identify patients with beneficial T cell responses. To achieve this, T cells reactive to the shared TAAs MUC1, survivin and HER2 were measured in a cohort of breast cancer patients and were investigated for association with a broad set of immunological parameters. Considering HLA type, serum cytokines, tumour-infiltrating leukocytes and blood leukocyte populations, the latter were found to be the most informative for identifying patients with such antigen-reactive T cells. Moreover, by constructing composite immune profiles, we were able to achieve a sensitivity and specificity of up to 100 % for the identification of patients possessing these antigen-reactive T cells. In addition to predictive immune markers in the periphery, a number of studies has shown that the immune status within the tumour plays a major role in cancer progression. Therefore the second part of this work investigated immune features in the tumour microenvironment which may serve as prognostic markers for patient survival. Unlike in the majority of prior studies which only considered one type of immune feature, for example only leukocytes or only cytokines, this study assessed a combination of diverse immune parameters in the tumour microenvironment including lymphoid and myeloid cells, secreted cytokines and transcription factors. This investigation showed that patients whose tumours had high relative levels of putatively suppressive CD15+ cells had shorter overall survival. Furthermore, by combining the assessment of CD3+ and CD15+ cells, it was observed that patients with high levels of CD3+ T cells and low levels of CD15+ cells survived the longest. Other studies have yielded similar results regarding the impact of the state of the immune system on clinical outcome. These studies have shown that one of the mechanisms employed by melanoma to escape anti-tumour immunity is the induction of immunosuppressive myeloid-derived suppressor cells (MDSCs), the levels of which correlate with clinical outcome. The mechanisms by which tumour cells induce MDSCs remain unknown, but knowledge thereof may pave the way for new forms of cancer therapy. Therefore, an in vitro co-culture model was developed to uncover tumour-immune interactions, with the aim of identifying pathways that allow the induction of such suppressive cells to be prevented. The implementation of a three-way co-culture system employing melanoma cells, activated T cells and monocytes (precursors of MDSCs) allowed the re-creation of MDSC-induced immune suppression in vitro. Subsequent targeting of specific molecular pathways in melanoma cells revealed that proteins involved in cellular stress pathways (heat shock proteins) are involved in triggering the differentiation of normal monocytes into MDSCs. These studies show that the levels of immune populations in the periphery and in the tumour can be used to monitor the clinical course of cancer patients. Furthermore, this work identifies new mechanisms of immune suppression by revealing novel pathways used by tumour cells to suppress the immune system.