Abstract 2244: Evaluation of large antibody panels in single-cell genomic immunophenotyping of fresh and preserved human leukocytes

Abstract Background: Multi-omic single cell studies are revolutionizing knowledge of the tumor immune microenvironment and positioned to detect rare cell subsets that correlate with response to immunotherapy. DNA-barcoded antibody panels enable the unambiguous classification of cell subsets via simultaneous measurement of surface proteins and transcriptome. However, single-cell sequencing performs best on fresh samples while multi-center clinical trials routinely obtain frozen specimens. Peripheral blood mononuclear cells (PBMCs) are commonly used tissues for charactering immune cell types, dynamics, and signaling in human studies. Our group has pioneered analyses of the tumor micro-environment in several cancer types, and also developed genomic capabilities with the resolution of flow cytometry over thousands of markers. We hypothesized that the development of advanced bioinformatics tools and large antibody panels (137 antibodies) can augment the analysis of frozen samples. Methods: Single cell genomics on matched fresh and frozen human PBMCs was performed using the TotalSeq-C human universal cocktail (BioLegend) and the 10x Genomics Chromium single-cell platform. We created bioinformatics pipelines to compare the quality control metrics, isotype matched control performance, and cell subtype analyses for fresh and frozen samples. Advanced visualization and analysis tools were developed to allow interrogation of the data in a flow cytometry paradigm, including complex gating strategies to identify rarer cell subtypes. Results: In terms of library quality, fresh and frozen cells had similar quality control parameters, although mean reads per cell were moderately lower for frozen vs. fresh tissue in both the RNA and antibody arms of the study. RNA and antibody data were integrated, clustered, and annotated for cell type and subtype. Cell type proportions were comparable across fresh and frozen samples. A panel of 7 antibody isotype controls demonstrated limited non-specific binding. Comparison of individual surface proteins across fresh and frozen samples revealed similar distributions in most cases, although signal was attenuated for a small number of epitopes in frozen samples. Our bioinformatics approaches allowed rare cell subtype identification with sensitivity and selectivity rivaling flow cytometry. Conclusions: Multi-modal single cell genomics, including a 137 antibody panel, and novel bioinformatics analytics reveal that fresh and frozen PBMC samples are largely concordant at both the transcriptome and cell-surface protein levels. Our pilot study is being expanded to larger scale clinical settings and may help characterize important cell populations that are associated with disease status, pharmacodynamics, and therapeutic response. Citation Format: Nathan O. Siemers, Jasmine Chen, Parminder Mankoo, Shazia Ilyas. Evaluation of large antibody panels in single-cell genomic immunophenotyping of fresh and preserved human leukocytes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2244.