OMIP‐054: Broad Immune Phenotyping of Innate and Adaptive Leukocytes in the Brain, Spleen, and Bone Marrow of an Orthotopic Murine Glioblastoma Model by Mass Cytometry

Abstract

Here we present a 42 parameter panel to characterize myeloid immune cell subsets and T lymphocyte activation status in cryopreserved and barcoded single cell suspensions obtained from brain, spleen, and bone marrow of an orthotopic immunocompetent glioblastoma mouse model in a C57BL/6 background (Table 1). This panel is designed for mass cytometry by time of flight (CyTOF) and combines 34 antibodies against diverse cell surface and intracellular targets together with cisplatin for live/dead discrimination, iridium for cell identification, and six cellular barcodes 1 to enable simultaneous multiplexed acquisition of up to 20 samples (Table 2). The panel is designed for a comprehensive evaluation of the immune system in different organs during murine in vivo studies in the field of glioblastoma immunology, but could also be applied to other disease models in the central nervous system (CNS) with possible systemic involvement, such as brain metastasis arising from other types of cancer, experimental autoimmune encephalomyelitis (EAE), or neurodegenerative disease models. Marker selection was partly based on a combination of previously reported panels studying CNS immune infiltrates 2-7. The selected set of markers captures T lymphocytes (CD8+, CD4+, and regulatory T lymphocytes), dendritic cells (DC), monocytes, macrophages, microglia, tumor cells (when GFP positive), and granulocytes, and contains a set of antibodies to detect cell activation, migratory capacity and immune checkpoints (Table 2). The panel has been optimized with respect to marker selection, antibody clone usage, antibody-metal pairing, and antibody concentration, and has room for additional markers by filling in a number of free channels as listed in Table 2. In the last decades enormous efforts have been made to develop therapeutic interventions that boost antitumor immunity, including adoptive cell transfer, anti-cancer vaccination, application of oncolytic viruses, and immune checkpoint inhibition 8, 9. A wide range of immunotherapeutic strategies are being tested in glioblastoma, advanced by successes in other tumor types. In these approaches particularly T cells are targeted as final effector cells by, for example, dendritic cell vaccination or immune checkpoint inhibition, but durable responses remain limited to case reports 10. Boosted T cell responses as a result of, for example, vaccination strategies encounter functional impairment due to immune inhibitory mechanisms both in the tumor microenvironment and systemically. To overcome T cell inhibition hope was pinned on immune checkpoint inhibition with monoclonal antibody therapy against PD1 and CTLA-4. Nevertheless, to date clinical trials testing immune checkpoint inhibition in glioblastoma have not shown impressive results 11. Clearly glioblastoma-specific immune suppression is a complex and unsolved problem that remains an obstacle for successful immunotherapies. Therefore, a better understanding of glioblastoma immune escape and contributing factors is warranted, alongside the development of biomarkers for patient selection and prediction of clinical response. Immune suppression in glioblastoma is largely mediated by infiltration of monocytes into the glioblastoma microenvironment 12. Myeloid immune cells dominate immune infiltrates and are involved in glioblastoma disease progression 13-15. Although some subsets have been studied in isolation, the different types of infiltrating myeloid and lymphoid cells, their recruitment from the bone marrow and spleen, and their phenotypic distribution need further investigation. Mass cytometry provides for the simultaneous measurement of more than 40 parameters at single cell resolution, improving the ability of cytometry to characterize the complexity of the immune system 16. Similar to the development of antibody panels for multichromatic fluorescence cytometry, mass cytometry panel development requires optimization of antibody-metal pairing, conjugation of antibodies with metal polymers, determination of optimal antibody concentrations, and optimization of buffers and staining conditions. Here we developed a mass cytometry immunophenotyping panel, which was designed to quantify population frequencies and to infer functional states of T cells, including activation, differentiation, exhaustion, or anergy in the murine glioblastoma microenvironment (Fig. 1) and spleen. Furthermore, our panel helps to differentiate and quantify a multitude of glioblastoma infiltrating and bone marrow derived myeloid cell subsets and immune cells resident to the bone marrow. Although this panel is optimized for single cell suspensions obtained from mouse brain, spleen, and bone marrow it could also be applied to the study of innate and adaptive components of the immune system in other mouse organs. Single cell suspensions were generated using mechanical dissociation and enzymatic digestion using the Neural Tissue Dissociation kit P (Miltenyi, Germany) 17. After live/dead staining and barcoding 1 cells were cryopreserved until antibody staining and mass cytometry analysis (Fig. 1A). Following data pre-processing including bead normalization 18 and debarcoding 1, the first gates aimed at exclusion of normalization beads, cell debris, doublets, and dead cells (140Ce−, 193Ir+, and 194Pt−). Next, we plotted CD45 against CD11c to identify CD45hiCD11chi DCs (population 1). Plotting CD45 against CD11b for non-DCs resulted in four main populations: CD45− non-immune cells (population 2), CD45+CD11b− lymphocytes, CD45+CD11b+ infiltrating myeloid cells, and CD45dimCD11b+ microglia. CD45− cells include tumor cells and non-immune non-tumor brain resident cells, such as glial cells or neurons. In the CD45+CD11b−CD3+TCRb+ cells we identified CD8+ T cells, CD4+ T cells, and CD4+FoxP3+ regulatory T cells (populations 3, 4, and 5, respectively). Next, we identified microglia as being the CD44−CD49d− cells in the CD11b+CD45dim population (population 6). Based on MHC-II, CCR2, and PD-L1 expression we identified three clusters of microglia with different activation states (6A, 6B, 6C). CD45+CD11b+ infiltrating myeloid cells contain granulocytes (population 7), Ly6ChiCCR2hi recently extravasated monocytes (population 8), and some clusters of macrophages with different phenotypes and/or states of activation (9A, 9B, 9C, 9D, 9E, and 9F) based on their expression of CCR2, MHC-II, PD-L1, CD38, and CD88. Gated populations were identified based on two-dimensional embeddings of the high dimensional space visualized using the viSNE algorithm in Cytobank (Fig. 1B,C). Enzymatic digestions often result in the cleavage of molecules on the membrane of immune cells, which could have detrimental effect in their recognition by antibodies. In order to allow direct comparison amongst different organs (brain, spleen, and bone marrow) all samples were subjected to the same enzymatic treatment as it was required for the processing of the brain samples. The panel presented in this OMIP has been optimized to the single cell suspension preparation procedure as suggested by the manufacturer with some minor adjustments (see Supporting Information “Online Protocol”). Care should be taken to specifically address any effects of alterations in the tissue dissociation protocol to the performance of this panel. This OMIP is partly overlapping with OMIP-032, which describes two murine multicolor immunofluorescence panels including the following markers to detect B and T lymphocytes, natural killer cells, dendritic cells, macrophages, monocytes, and neutrophils: CD45, TCRβ, CD4, CD8, Ly6C, CD49b, Ly6G, CD11c, MHCII, CD206, NKp46, CD62L, and CD44 19. These markers overlap with our panel except for CD49b, NKp46, and CD62L. Since our panel is more focused on activation and exhaustion in T lymphocytes, we included more antibodies against activation markers and co-inhibitory molecules (immune checkpoints) rather than the presence of naïve or memory T cells. Also OMIP-031 has overlap with a panel aimed at immune checkpoint expression 20, which reports the following markers: CD3, CD4, CD8, CD69, CD44, CD45RA, CD27, CD62L, KLRG127, CD127, PD-1, CTLA4, TIM-3, LAG-3, and CD45. Our panel includes all of these markers except for CD45RA, CD27, CD62L, KLG127, and LAG-3. Previously, three other mass cytometry panels have been published. They focus on the characterization of human peripheral leukocytes (OMIP-34), human head and neck cancer (OMIP-45), and the quantification of calcium sensors and channels in lymphocyte subsets 21-23. This is the first mass cytometry OMIP for a comprehensive characterization of the mouse immune system. All data used generated during the optimization and verification of this panel can be found in FlowRepository under ID: FR-FCM-ZYUF. We thank Prof. van Kooyk (Amsterdam UMC, Location VUmc) for fruitful discussions and laboratory resources. We also thank S. Schetters, Dr. Crommentuijn (both from Amsterdam UMC, Location VUmc), E. Abels (Massachusetts General Hospital), and Dr. Broekman (Leiden University Medical Center) for discussions and suggestions during panel design. We thank E. Abels for providing samples for optimization experiments. SAD and JV performed experiments. SAD and JV developed and optimized the panel. SAD, JV, and JJGV discussed panel design and optimization. SAD and JJGV wrote the article. The authors declare that they have no conflict of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.