Mapping the Cellular Landscape of Brain Tumors Using Imaging Mass Cytometry

Transitional meningioma Gliosarcoma Malignant ependymoma αSMA CD44 Vimentin Iba1 Collagen 1 CD20 CD4 Iba1 CD31 CD34 MAP2 CD44 GFAP Iba1 Fibronectin CD8 F4/80 CD45 Iba1 p-tyrosine pERK1/2 Ki-67 Vimentin pS6 β-catenin CD34 CD31 β-actin Neuroblastoma A B Standard BioTools Inc. 2 Tower Place, Suite 2000, South San Francisco, CA 94080 USA +1 650 266 6000 • Toll-free in the US and Canada: 866 359 4354 standardbio.com For Research Use Only. Not for use in diagnostic procedures. Limited Use Label License and other terms may apply: www.standardbio.com/legal/salesterms. Patent and License Information: www.standardbio.com/legal/notices. Trademarks: www.standardbio.com/legal/trademarks. Any other trademarks are the sole property of their respective owners. ©2023 Standard BioTools Inc. (f.k.a. Fluidigm Corporation). All rights reserved. Mapping Neuro-Oncological Cellular Landscape of Human and Mouse Brain Tumors Using Imaging Mass Cytometry Nick Zabinyakov, Qanber Raza, Christina Loh | Standard BioTools Canada Inc., Markham, ON, Canada Introduction Brain cancer research presents challenges that require comprehensive assessment of the structural and cellular organization of the tumor microenvironment (TME). Imaging Mass Cytometry™ (IMC™) offers unprecedented insight into the TME by uncovering the spatial distribution of 40-plus distinct molecular markers without data artifacts caused by autofluorescence or tissue discohesion. We developed high-plex proteomic analysis tools to thoroughly characterize the TME of both human and mouse brain tissue using IMC. Here we present a deep phenotypic spatial analysis of various mouse and human brain tumors and identify cellular composition and activation of immuno-oncological processes within the TME. Material and Methods A 39-parameter human antibody panel and a 36-parameter mouse antibody panel, both designed to highlight central features of normal and diseased brain tissue, are presented in this poster. These panels contain antibodies that identify major cell populations in normal brain and the specific states of tumor and immune cell populations in the diseased brain. Tissue slides were prepared and stained using optimized antibody dilutions and were ablated using the Hyperion Imaging System at 200 Hz with 1 µm pixel size. Qualitative data analysis, multiplexed image rendering, and single-channel image extractions were performed using MCD™ Viewer. Resulting images were rendered in MCD Viewer and exported for singlecell analysis. Segmentation was performed using CellProfiler™ v4.2.1. Nuclei and cell membrane were detected using Cell-ID™ Intercalator-lr (PN 201192B) and the Maxpar® IMC Cell Segmentation Kit (ICSK; PN 201500), respectively. Cell masks were subsequently exported for Antibody panels for human and mouse FFPE tissue 39-parameter human antibody panel and 36-parameter mouse antibody panel including 2 DNA and 3 ICSK channels designed to highlight central features of the TME. These panels are subdivided into 6 modules, each revealing critical insights about normal and tumor tissue composition, state, and biology. The tissue architecture module identifies the underlying cellular and structural markers of the tumors. The brain cell process module identifies activation of signaling pathways, metabolism, and growth in brain cells. The lymphoid and myeloid modules delineate lymphoid and myeloid cell subtypes of immune cell infiltrates in brain tissue. The immune activation module assesses the functional state of immune cells in brain tissue. E-cad – E-cadherin, GRNZB – granzyme B, Pan-CK – pan-cytokeratin Ordering information for products listed in this poster Download this poster Conclusion Imaging Mass Cytometry workflow. (1) Experiment is initiated by obtaining antibodies conjugated with specific heavy metal isotopes. (2) An antibody cocktail is assembled and used to stain tissue in a single step, all at once. (3) The stained tissue is imaged using the Hyperion Imaging System. (4) The multidimensional data is collected and images are rendered demonstrating localization of marker of choice. (5) Data analysis is conducted to identify specific cell populations and tissue structures present in the TME. Imaging Mass Cytometry workflow with neuro-oncology panels enables high-quality neurological and immunological readouts of over 40 protein markers simultaneously without autofluorescence, tissue discohesion, or signal spillover in normal brain tissue and in brain tumors of glial, non-glial, and mixed origin. Results Our data demonstrates performance of markers with high specificity and signal intensity and absence of autofluorescence data artifacts in human and mouse normal and diseased brain tissue. Presence or absence of primary brain cell constituents such as microglia, astrocytes, neurons, endothelial cells, and oligodendrocytes is apparent in both human and mouse tissue. Both infiltrating and resident myeloid and lymphoid immune cells are distinguished, and their spatial locations within the TME are demarcated. Following sections discuss the performance of neuro-oncology panels and the tumor-specific immuno-oncological insights they offer. Figure 1. Application of Imaging Mass Cytometry using neurophenotyping panel demonstrates cellular heterogeneity in normal human and mouse brain tissue. A 7-parameter neurophenotyping panel consists of human and mouse cross-reactive clones, which enables flexible panel design for brain-specific research goals. IMC generated images of highly autofluorescent normal FFPE human cerebral cortex tissue (A) and normal FFPE mouse cerebral cortex, cerebellum, and hippocampus tissue (sagittal cut) (B) that were stained with neurophenotyping panel are shown. The following brain tissue constituents can be identified: neuronal cell bodies and axonal extremities (NeuN, MAP2), oligodendrocytes (Olig2), resting and activated astrocytes (GFAP, S100β), resting and activated microglia (Iba1), and brain vasculature (CD34). In human brain sections, 3 tissue cores obtained from different locations demonstrate the contrasting cellular composition of the tissue microenvironment. In mouse brain sections, images of 3 distinct brain tissue demonstrate unique structural and cellular architecture of tissue compartments. Neurophenotyping panel can be applied to identify composition of both central and peripheral human nervous system tissue. Figure 2. Application of high-parameter neuro-oncology IMC panels on human and mouse glioblastoma tissue. IMC generated images of human glioblastoma (grade 4) with overt hemorrhage (bleeding) (A) and mouse glioblastoma (B) are shown. In human tissue (A), hemorrhagic tumor areas are demarcated in red. Smooth muscle cells (αSMA), immune cells (CD45, CD11b), tumor cells (GFAP, Olig2), and activated tumor cells (pERK, vimentin, Ki-67, pS6) can be identified. Inset A’ marks infiltrating T cells (CD45RO+, CD45RA+) and monocytes (CD11b) in the hemorrhagic tumor tissue. Inset A’’ depicts cells of myeloid origin, both resident and infiltrating (CD14+, CD16+), in the TME. Inset A’’’ demonstrates one of few pseudopalisade areas with extensive tumor scaffolding expressing CD44 and CD34. Inset A’’’’ demonstrates presence of immune cells of myeloid lineage (CD68, CD163, Iba1) within the alternative pseudopalisade area. In mouse tissue (B), tissue area located at the margin of tumor is shown. Tumor cells (Olig2, S100β), endothelial cells (CD34), metabolically active cells (pS6), T cells (CD3), and stem cell-like tumor cells (CD44) can be identified. Inset B’ demonstrates extracellular matrix composition (collagen 1, fibronectin) and smooth muscle covered vasculature (αSMA, CD31). Inset B’’ demonstrates activation of receptor tyrosine kinase activity (p-tyrosine), Ras signaling cascade, and cytoskeletal structure (β-actin) in tumor and immune cells. Inset B’’’ shows presence of T cell subtypes, cytotoxic T cells (CD3, CD8), regulatory T cells (CD3, FoxP3), and areas with cytotoxic immune cell activation (GzmB). Inset B’’’’ depicts presence of macrophages (F4/80), monocytes (CD11b), and microglia (Iba1). Figure 3. Application of neuro-oncology IMC panel on human and mouse tumors of mixed or non-glial origin. IMC generated images of human transitional meningioma, gliosarcoma, malignant ependymoma (A), and mouse neuroblastoma (B) are shown. Diverse abnormal composition of connective tissue (αSMA, collagen 1) can be observed in human transitional meningioma. A rare cluster of infiltrating B cells (CD20) and T helper cells (CD4) around a blood vessel (CD34, CD31) and microglia (Iba1) scattered around the tissue can be observed in human gliosarcoma. Presence of activated amoeboid microglia (Iba1) can be noted in human malignant ependymoma. In mouse neuroblastoma (B) tissue, scattered deposits of fibronectin and presence of myeloid cells (F4/80, Iba1) are noted. Only a few cytotoxic T cells (CD8; green arrowheads) can be observed, indicating a suppressed tissue immune microenvironment. Activation of receptor tyrosine kinase (p-tyrosine), Ras signaling (pERK1/2), mTOR signaling (pS6), replication (Ki-67), activation, and cytoskeletal structure (vimentin, β-actin) across tumor cells can be detected. Relative spatial positioning of vascular network (CD34, CD31, β-catenin; white arrowheads) within the TME (β-actin) can be distinguished. 3. Neuro-oncology panel allows deciphering of cell populations in various neoplasms in human (A) and mouse (B) diseased brain 1. Neurophenotyping panel identifies major cell types in normal human (A) and mouse (B) brain tissue Cerebrum, occipital lobe Cerebrum, temporal lobe Cerebral cortex Cerebellum Hippocampus A B MAP2 NeuN S100β Iba1 GFAP CD34 Olig2 2. High-parameter neuro-oncology panels delineate tumor cell states and immune cell infiltration within TME of human (A) and mouse (B) glioblastoma Human glioblastoma (grade 4) with bleeding Pseudopalisade Pseudopalisade CD14 CD16 DNA CD45RA CD45RO CD11b CD34 CD44 CD31 CD68 CD163 Iba1 αSMA CD45 CD11b GFAP pS6 pERK1/2 Ki-67 Olig2 Vimentin Overt hemorrhage pERK p-tyrosine β-actin S100β Olig2 CD34 pS6 B’’’ Mouse glioblastoma CD3 CD44 β-catenin GzmB CD8 CD3 FoxP3 Iba1 F4/80 CD11b CD31 Fibronectin αSMA Collagen 1 A B A’’ A’’’’ A’’’ A’ A’’ A’’’ A’’’’ B’ B’’ B’’’’ B’’’ B’ B’’ B’’’’ A’ Figure 4. Single-cell analysis identifies specific cell populations in human brain neoplasms. ICSK was used to demarcate cellular boundaries in cells to enhance cell segmentation capabilities (A, A’). t-SNE and PhenoGraph analysis identified 18 separate cellular clusters from 2 regions of interest including 58,119 cells for glioblastoma (A) and 15 separate cellular clusters from 2 regions of interests including 22,290 cells for transitional meningioma (A’) defined by the expression of a single or multiple markers from the neuro-oncology panel. Spatial positioning of cells belonging to individual cellular clusters is shown overlaid on tumor image (B, B’). t-SNE maps further demonstrate the separation of individual cellular clusters by color (C, C’). Heat maps of examples of individual marker expression are displayed as cell masks overlaid on tumor tissue and t-SNE graphs (D, D’). GFAP and S100β expression indicates positioning of tumor cells and activated astrocytes. Iba1 expression highlights microglial cells, which elicit an anti-tumor immune response (D). NeuN expression indicates positioning of perikarya. S100β expression highlights astrocytes in activated state. Iba1 expression highlights microglial cells, which can elicit an anti-tumor immune response (D’). Figure 5. Single-cell analysis identifies specific cell populations in mouse glioblastoma and neuroblastoma. ICSK was used to demarcate cellular boundaries in cells to enhance cell segmentation capabilities (A, A’). t-SNE and PhenoGraph analysis identified 27 separate cellular clusters from 7 regions of interest including 204,070 cells for glioblastoma (A) and 15 separate cellular clusters from 2 regions of interest including 24,619 cells for neuroblastoma (A’) defined by the expression of a single or multiple markers from the neurooncology panel. Spatial positioning of cells belonging to individual cellular clusters is shown overlaid on tumor image (B, B’). t-SNE maps further demonstrate the separation of individual cellular clusters by color (C, C’). Heat maps of examples of individual marker expression are displayed as cell masks overlaid on tumor tissue and t-SNE graphs (D, D’). Olig2 expression indicates positioning of tumor cells. S100β expression highlights astrocytes in activated state. Iba1 expression highlights microglial cells, which can elicit an anti-tumor immune response (D). MAP2 expression indicates positioning of axons. CD34 shows the presence of vasculature, and Iba1 expression highlights microglial cells that elicit an anti-tumor immune response (D’). Cell segmentation markers Mask with cell populations A B C Human glioblastoma t-SNE graphs ICSK1 ICSK2 ICSK3 D GFAP S100β Iba1 t-SNE heat maps of selected markers Human transitional meningioma ICSK1 ICSK2 ICSK3 B’ A’ NeuN Iba1 C’ D’ 4. Single-cell analysis of human brain tumor demonstrates cellular composition of TME 5. Single-cell analysis of mouse brain tumor demonstrates cellular composition of TME ICSK1 ICSK2 ICSK3 Mouse neuroblastoma A’ B’ MAP2 C’ CD34 D’ Iba1 ICSK1 ICSK2 ICSK3 Mouse glioblastoma A B C D Olig2 S100β Iba1 Mouse neuro-oncology panel Human neuro-oncology panel Neurophenotyping Imaging Mass Cytometry and single-cell analysis workflow S100β single-cell phenotyping and cellular clustering into histoCAT™ v1.76. t-SNE maps and corresponding PhenoGraph cell clusters were identified to assess relevant and specific tumor and immune cell subpopulations. Cell masks highlighting specific clusters were generated and overlaid on the corresponding IMC image. Ordering information and a digital version of this poster can be accessed via QR codes at the bottom of the poster.