J. Sia1,2, C. DSouza1, B. Castle1, Y. K. Huang1, H. Aw Yeang1, R. Idrizi1, M. Jana1, S. Siva2,3, C. Phillips1, and P. Neeson1,2; 1Peter MacCallum Cancer Centre, Melbourne, Australia, 2Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, VIC, Australia, 3Peter MacCallum Cancer Centre, Melbourne, VIC, Australia
Purpose/Objective(s): Stereotactic radiosurgery (SRS) is a treatment mainstay for brain metastases (BrM). Reports on whether SRS can evoke anti-tumour immune responses have been scarce and conflicting, largely due to challenges in accessing homogenous, unbiased tissue samples for study. Here, we addressed this issue by examining blood samples from a clinical trial of SRS for HER2+ breast cancer (HER2+BC) BrMs, matched with a unique longitudinal pair of HER2+BC BrMs resected from the same location in the same patient. Materials/
Methods: Blood samples from 10 patients with HER2+BC BrM taken pre- and 7-14 days post-SRS in the translational substudy of the Trans-Tasman Radiation Oncology Group (TROG) 16.02 trial were analysed by cytometry by time of flight (CyTOF). One patient received pre-operative SRS (20 Gy/1 fraction) for a HER2+BC BrM that recurred 7 months after resection, followed by planned re-resection 8 days post-SRS. Pre- and post-SRS tumors from this patient were analysed by bulk RNA sequencing, multiplex immunohistochemistry and T cell receptor (TCR) sequencing. Bulk transcriptome deconvolution was performed using a breast cancer BrM-specific signature matrix from published single-cell RNA sequencing data. Results: Monocytes, central memory CD8+ T and regulatory T cells were enriched in blood post-SRS, together with increased MHC-II expression on monocytes, conventional dendritic cells, and monocytic MDSCs. In tumour, SRS upregulated antigen presentation, T cell proliferation and T cell co-stimulation signatures. Transcriptome deconvolution revealed a loss of metastatic tumor cells post-SRS, alongside an influx of APOE+ tumor-associated macrophages (TAMs) and central memory CD4+ T cells. The proportion of CD8+ T cells in tumor was small and not altered by SRS. Spatially, an immune-excluded phenotype was seen pre- but not post-SRS. Specifically, TAMs and CD4+ T cells, but not CD8+ T cells, demonstrated spatial co-localisation post-SRS. These TAMs were lowly PD-L1 expressing, but CD4+ T cells showed increased PD-1 expression. TCR clonality in tumor was increased following SRS. A sizeable proportion of T cell clonotypes were retained post-SRS, and four clones demonstrated significant, non-stochastic expansion. These clones did not match known sequences against common viral and bacterial epitopes. Conclusion: In this homogenous HER2+BC patient cohort, systemic and local changes, particularly in transcriptional programs, spatial patterns, and T cell clonal repertoire, suggest that SRS may facilitate MHC-II-restricted T cell priming involving the monocyte-macrophage lineage and CD4+ T cells. This may provide mechanistic support for clinical observations of improved intracranial disease control with concurrent SRS and immunotherapy.
