Learning cell identity from single-cell data presently relies on human experts. Here, we present Marker Enrichment Modeling (MEM), an algorithm that objectively describes cells by quantifying contextual feature enrichment and reporting a human and machine-readable text label. MEM outperformed traditional metrics in describing immune and cancer cell subsets from fluorescence and mass cytometry. MEM provides a quantitative language to communicate characteristics of new and established cytotypes observed in complex tissues.

Background-Mass cytometry measures 36 or more markers per cell and is an appealing platform for comprehensive phenotyping of cells in human tissue and tumor biopsies. While tissue disaggregation and fluorescence cytometry protocols were pioneered decades ago, it is not known whether established protocols will be effective for mass cytometry and maintain cancer and stromal cell diversity.

Advances in single-cell biology have enabled measurements of >40 protein features on millions of immune cells within clinical samples. However, the data analysis steps following cell population identification are susceptible to bias, time-consuming, and challenging to compare across studies. Here, an ensemble of unsupervised tools was developed to evaluate four essential types of immune cell information, incorporate changes over time, and address diverse immune monitoring challenges. The four complementary properties characterized were: 1) systemic plasticity, 2) change in population abundance, 3) change in signature population features, and 4) novelty of cellular phenotype. Three systems immune monitoring studies were selected to challenge this ensemble approach. In serial biopsies of melanoma tumors undergoing targeted therapy, the ensemble approach revealed enrichment of double-negative (DN) T cells. Melanoma tumor resident DN T cells were abnormal and phenotypically distinct from those found in non-malignant lymphoid tissues, but similar to those found in glioblastoma and renal cell carcinoma. Overall, ensemble systems immune monitoring provided a robust, quantitative view of changes in both the system and cell subsets, allowed for transparent review by human experts, and revealed abnormal immune cells present across multiple human tumor types.

Background Mass cytometry measures 36 or more markers per cell and is an appealing platform for comprehensive phenotyping of cells in human tissue and tumor biopsies. While tissue disaggregation and fluorescence cytometry protocols were pioneered decades ago, it is not known whether established protocols will be effective for mass cytometry and maintain cancer and stromal cell diversity. Methods Tissue preparation techniques were systematically compared for gliomas and melanomas, patient derived xenografts of small cell lung cancer, and tonsil tissue as a control. Enzymes assessed included DNase, HyQTase, TrypLE, collagenase (Col) II, Col IV, Col V, and Col XI. Fluorescence and mass cytometry were used to track cell subset abundance following different enzyme combinations and treatment times. Results Mechanical disaggregation paired with enzymatic dissociation by Col II, Col IV, Col V, or Col XI plus DNase for 1 hour produced the highest yield of viable cells per gram of tissue. Longer dissociation times led to increasing cell death and disproportionate loss of cell subsets. Key markers for establishing cell identity included CD45, CD3, CD4, CD8, CD19, CD64, HLA-DR, CD11c, CD56, CD44, GFAP, S100B, SOX2, nestin, vimentin, cytokeratin, and CD31. Mass and fluorescence cytometry identified comparable frequencies of cancer cell subsets, leukocytes, and endothelial cells in glioma (R = 0.97), and tonsil (R = 0.98). Conclusions This investigation establishes standard procedures for preparing viable single cell suspensions that preserve the cellular diversity of human tissue microenvironments.

Sudden unexplained death in childhood (SUDC) affects children >1-year-old whose cause of death remains unexplained following comprehensive case investigation and is often associated with hippocampal abnormalities. We prospectively performed systematic neuropathologic investigation in 20 SUDC cases, including (i) autopsy data and comprehensive ancillary testing, including molecular studies, (ii) ex vivo 3T MRI and extensive histologic brain samples, and (iii) blinded neuropathology review by 2 board-certified neuropathologists. There were 12 girls and 8 boys; median age at death was 33.3 months. Twelve had a history of febrile seizures, 85% died during apparent sleep and 80% in prone position. Molecular testing possibly explained 3 deaths and identified genetic mutations in TNNI3, RYR2, and multiple chromosomal aberrations. Hippocampal abnormalities most often affected the dentate gyrus (altered thickness, irregular configuration, and focal lack of granule cells), and had highest concordance between reviewers. Findings were identified with similar frequencies in cases with and without molecular findings. Number of seizures did not correlate with hippocampal findings. Hippocampal alterations were the most common finding on histological review but were also found in possibly explained deaths. The significance and specificity of hippocampal findings is unclear as they may result from seizures, contribute to seizure pathogenesis, or be an unrelated phenomenon.

Single cell mass cytometry is revolutionizing our ability to quantitatively characterize cellular biomarkers and signaling networks. Mass cytometry experiments routinely measure 25–35 features of each cell in primary human tissue samples. The relative ease with which a novice user can generate a large amount of high quality data and the novelty of the approach have created a need for example protocols, analysis strategies, and datasets. In this chapter, we present detailed protocols for two mass cytometry experiments designed as training tools. The first protocol describes detection of 26 features on the surface of human peripheral blood mononuclear cells. In the second protocol, a mass cytometry signaling network profile measures 25 node states comprised of five key signaling effectors (AKT, ERK1/2, STAT1, STAT5, and p38) quantified under five conditions (Basal, FLT3L, SCF, IL-3, and IFNγ). This chapter compares manual and unsupervised data analysis approaches, including bivariate plots, heatmaps, histogram overlays, SPADE, and viSNE. Data files in this chapter have been shared online using Cytobank (http://www.cytobank.org/irishlab/).

Mass cytometry is a single cell biology technique that samples >500 cells per second, measures >35 features per cell, and is sensitive across a dynamic range of >104 relative intensity units per feature. This combination of technical assets has powered a series of recent cytomic studies where investigators used mass cytometry to measure protein and phospho-protein expression in millions of cells, characterize rare cell types in healthy and diseased tissues, and reveal novel, unexpected cells. However, these advances largely occurred in studies of blood, lymphoid tissues, and bone marrow, since the cells in these tissues are readily obtained in single cell suspensions. This protocol establishes a primer for single cell analysis of solid tumors and tissues and has been tested with mass cytometry. The cells obtained from this protocol can be fixed for study, cryopreserved for long term storage, or perturbed ex vivo to dissect responses to stimuli and inhibitors.

A goal of cancer research is to reveal cell subsets linked to continuous clinical outcomes to generate new therapeutic and biomarker hypotheses. We introduce a machine learning algorithm, Risk Assessment Population IDentification (RAPID), that is unsupervised and automated, identifies phenotypically distinct cell populations, and determines whether these populations stratify patient survival. With a pilot mass cytometry dataset of 2 million cells from 28 glioblastomas, RAPID identified tumor cells whose abundance independently and continuously stratified patient survival. Statistical validation within the workflow included repeated runs of stochastic steps and cell subsampling. Biological validation used an orthogonal platform, immunohistochemistry, and a larger cohort of 73 glioblastoma patients to confirm the findings from the pilot cohort. RAPID was also validated to find known risk stratifying cells and features using published data from blood cancer. Thus, RAPID provides an automated, unsupervised approach for finding statistically and biologically significant cells using cytometry data from patient samples.