The influence of radio-immunotherapy on the tumor microenvironment of breast-to-brain metastasis and the investigation of novel adjuvant therapies

Abstract

With ongoing progress in cancer research and continuously improving treatment strategies for primary tumors, the incidence of brain metastasis is steadily increasing. The treatment options for brain metastasis patients, however, are limited and only prolong survival for a short duration. With the advent of immunotherapies, the cancer field was revolutionized. Checkpoint inhibitors, which reactivate T cell responses against cancer cells, show promising results even for aggressive cancers such as advanced metastatic melanoma. The brain was regarded as an immune privileged site for a long time. However, only recently a classical lymphatic system has been revealed in the brain. Moreover, the central nervous system harbors a greater variety of immune cells than previously assumed. Therefore, immunotherapies including checkpoint inhibitors, also gain interest for the treatment of brain metastases. To date, most research in the cancer field focusses on highly immunogenic cancer entities, whereas tumors of low immunogenicity such as breast cancer are less well investigated, as they are thought to be resistant to checkpoint inhibition. The development of strategies to convert 'cold' tumors of low immunogenicity into 'hot' tumors with a pro-inflammatory tumor microenvironment is of great interest, as they potentially sensitize highly immune suppressive tumor microenvironments to immune checkpoint inhibition. One strategy, that has shown promising results for different cancer entities is the combination of checkpoint inhibition with radiotherapy. The efficacy of this combination is investigated in clinical trials, including the treatment of brain metastasis. Current research focusses mostly on melanoma and lung cancer derived brain metastasis, and only little information is available on the efficacy of radio-immunotherapy in the treatment of breast cancer brain metastasis. Therefore, the aim of this thesis was to investigate if standard of care radiotherapy can sensitize breast cancer brain metastasis to immune checkpoint inhibition. In this context, the tumor microenvironment of the murine breast cancer brain metastasis model 99LN-BrM was investigated in detail. The purpose was to obtain an overview of proportions of cells, counteracting T cell responses, versus cells that are crucial for efficient checkpoint inhibition. It was confirmed that the tumor microenvironment of 99LN BrM is a typical 'cold' microenvironment dominated by myeloid cells. However, cell types, crucial for checkpoint blockade, such as T cells and dendritic cells were identified, too. The next step was the examination of the influence of ionizing radiation on brain homing breast cancer cells and on the immune cell composition in 99LN-BrM. It was revealed that brain homing breast cancer cells increasingly express inflammatory markers, such as TNFα and IL1β, after in vitro irradiation. In a preclinical trial, the treatment of 99LN BrM mice with fractionated whole brain radiotherapy, did not lead to increased recruitment of potentially immune suppressive cell types, such as blood borne myeloid cells or regulatory T cells. Moreover, radiosensitive cell types, crucial for efficient checkpoint inhibition, such as T cells and dendritic cells were not depleted. On the contrary, the infiltration of cytotoxic CD8+ T cells into 99LN-BrM lesion was increased by fractionated whole brain radiotherapy. To obtain a deeper understanding of the T cell compartment in 99LN-BrM, TCRβ-profiling was performed next. These data reveled, that T cells in 99LN-BrM lesions and in CNS draining lymph nodes, clonally expand, indicating prior tumor directed T cell activation. However, a negative correlation of T cell expansion with brain metastasis volume was observed. This indicates progressive inhibition of T cell responses, which was confirmed by in vivo T cell depletion experiments. The depletion of T cells in mice, injected with brain homing 99LN-BrM cells, did not shorten the time of brain metastasis onset. This demonstrates that T cell responses in the microenvironment of 99LN-BrM are sufficiently suppressed. A checkpoint axis, which often plays a crucial role in immune system evasion of cancer cells, is the PD 1/PD L1 axis. By expressing PD-L1, tumor cells can inhibit PD-1 positive T cells in the tumor microenvironment. Assessment of PD-L1 expression by brain homing breast cancer cell lines showed expression on the RNA- and protein level in vitro. In vivo analysis revealed that a high proportion of T cells in 99LN-BrM express PD-1, whereas PD-L1 is expressed by tumor cells, myeloid and T cells. Furthermore, analysis of the myeloid compartment demonstrated that a high proportion of infiltrating myeloid cells is PD L1 positive, which is not the case for brain resident microglia. In a preclinical trial, treatment of 99LN-BrM mice with anti-PD-1, whole brain radiotherapy or a combination of both exhibited superior efficacy of the radio-immunotherapy over the monotherapies. Tumor progression slowed down, which translated into significantly prolonged median survival. However, long term survival was not achieved. The flow cytometric and histological analysis of brain metastases from mice in the preclinical trial, revealed that only in the combination cohort, both the infiltration of CD4+ and CD8+ T cells was enhanced in the brain metastasis lesions. At the same time, increased infiltration of blood-borne PD-L1+ myeloid cells, especially monocyte derived macrophages, was observed. Moreover, this infiltration was most prominent in the combination treatment group and indicates a crucial role of monocyte derived macrophages in acquired resistance to radio-immunotherapy. Furthermore, the in vitro assessment of T cell inhibitory capacity of 99LN-BrM conditioned microglia versus monocyte derived macrophages, revealed that the latter inhibit T cells more efficiently. To develop strategies to induce long term efficacy, macrophages were targeted pharmacologically, in addition to radio-immunotherapy. One strategy included the inhibition of the chemokine receptor CXCR4, expressed by macrophages, with the CXCR4 inhibitor AMD3100. This approach aimed at inhibiting the recruitment of monocyte derived macrophages to brain metastases. The second approach was aimed at targeting all macrophages by inhibiting CSF1R, the receptor to the macrophage survival factor CSF1. In vitro, the components of both signaling pathways were expressed by brain homing breast cancer cell lines. However, in vivo, both macrophage targeting strategies did not induce long term efficacy of radio-immunotherapy with anti-PD-1. Analysis of 99LN-BrM lesions revealed that CXCR4 inhibition failed to inhibit the recruitment of monocyte derived macrophages and even increased the infiltration of 99LN BrM with PD L1+ immune cells. The pharmacological inhibition of CSF1R not only led to the depletion of most macrophages in brain metastasis, including microglia, but also significantly decreased T cell infiltration, which is crucial for the efficacy of checkpoint inhibition. To induce long term efficacy in the future, a deeper understanding of myeloid immune suppressive cells in breast cancer brain metastasis is crucial. Additionally, treatment strategies, targeting the recruitment of blood borne myeloid cells specifically, while sparing other immune cells, need to be developed.