A targetable, myeloid orchestrated immunosuppressive microenvironment characterizes bone metastatic human prostate cancer Ninib Baryawno 1,2,3,4,*, Youmna Kfoury 1,2,3,*, Nicolas Severe 1,2,3,*, Shenglin Mei 5,*, Karin Gustafsson 1,2,3,*, Taghreed Hirz 1,2,3,*, Thomas Brouse 1, Elizabeth W. Scadden 1, Anna A. Igolkina 6, Bryan D. Choi 7, Nikolas Barkas 5, John H. Shin 7,†, Philip J. Saylor 8,†, David T. Scadden 1,2,3,†, David B. Sykes 1,2,3,†, Peter V. Kharchenko 2,5,†, as part of the Boston Bone Metastasis Consortium 1 Center for Regenerative Medicine, Massachusetts General Hospital, Boston, MA, USA 2 Harvard Stem Cell Institute, Cambridge, MA, USA 3 Department of Stem Cell and Regenerative Biology, Harvard University, Cambridge, MA, USA 4 Childhood Cancer Research unit, Dep. Of Children’s and Women’s Health, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden 5 Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA 6 St. Petersburg Polytechnical University, St. Petersburg, Russia 7 Department of Neurosurgery, Harvard Medical School, Boston, MA, USA 8 Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, USA *,† Contributed equally Bone is the third most common site for solid tumor metastases with the highest incidence for prostate and breast cancer. Current treatments are rarely curative (Migliorini et al., 2020) and patients are at a disadvantage when it comes to response to immune check point therapies (Beer et al., 2017). It is not clear what elements of the complex bone microenvironment contribute to it being a favored and treatment refractory metastatic site. We conducted a high-resolution analysis, using droplet-based single cell RNA sequencing of patient bone metastatic prostate cancer samples (Tumor=9) paired with bone marrow (BM) from the same patient that was adjacent (Involved=8) or distant (Distal=8) to the site of metastases. These were contrasted with BM samples from patients undergoing hip replacement surgery representing a non-malignant, inflammatory microenvironment (Benign=7) in addition to a published data set of healthy BM. This multi-layered approach allowed us to control for inter-patient variation and identify cellular and molecular changes that are specific to a malignant metastatic microenvironment rather than a general reflection of an inflammatory state. Analyses included hematopoietic lymphoid and myeloid cells, in addition to non-hematopoietic clusters representing tumor cells, mesenchymal and endothelial lineages. The presence of metastases reshaped the immune cell composition of the BM when compared to benign samples as reflected by a striking depletion of B cells and B cell progenitors in all three patient derived fractions, a substantial population shift within the myeloid compartment of the tumor fraction and an expansion of dysfunctional T cell clusters characterized by decreased cytotoxicity and increased exhaustion signatures. In particular, tumor associated M2 polarized macrophages (TAM) and tumor inflammatory monocytes (TIM) were selectively enriched in the tumor fraction. Their gene expression characteristics suggested an immune suppressive function. Indeed, the increased proportion of TAMs and TIMs correlated with cytotoxic T cell exhaustion and a reduced cytotoxicity signature respectively. Ligand/ receptor interaction analysis identified 241 potential channels of interaction between the myeloid and lymphoid compartments. To validate our findings, we developed a C57BL/6 syngeneic mouse model of highly penetrant bone metastatic prostate cancer that produces bone osteolytic and osteoblastic lesions. Using this model, we demonstrated that disrupting one of the channels mediating myeloid/lymphoid interaction by either blocking the ligand or deleting the receptor, the survival of bone metastases bearing mice improved. Further, T regulatory cells decreased as did T cell exhaustion as demonstrated by single cell RNA sequencing, phenotypic and functional assays. These data provide a high resolution landscape of the bone metastatic prostate cancer microenvironment that is distinct from benign inflammation and demonstrate the existence of a targetable immunosuppressive milieu that may guide novel therapeutic strategies to this devastating medical condition.