A Single-cell Approach to Identify Biomarkers of Pulmonary Toxicity for Immune Checkpoint Blockade

The investigators will apply single cell RNA- and TCR-sequencing on up to 5,000 single cells per sample. Additionally, cell surface protein expression can be integrated with the transcriptional information. Various bioinformatics pipelines, including Seurat, will be used to identify different cell clusters, which through marker gene expression will be assigned to known cell types, cellular subtypes or phenotypes. For instance, this will make it possible to monitor the abundance of PD-1/PD-L1 expressing T-cells, cytotoxic T-cells, immune-suppressive myeloid cells etc. The following parameters at single-cell level will be relevant, amongst others:

• The composition and relative abundancies of established immune cell types (e.g. T cells (CD4+, CD8+ and regulatory subsets), NK cells, B cells, MDSCs, macrophages, neutrophils, dendritic cells). Transcriptomic data for each of these immune cell subtypes will be analyzed, allowing characterization of specific gene expression programs that define specific phenotypic states.
• Composition of all stromal cellular subtypes identified by single-cell transcriptomics.
• A gene regulatory network for each cell type and cellular subtype (or cell state) will be established and master transcriptional regulators will be identified. Individual T-cells and T-cell subclusters will be classified based on interferon activation, high rates of proliferation and transcription and increased granzyme expression, which are all indicative of T-cell activation.

Blood samples will be subjected to similar single-cell experimental procedures. First, peripheral blood mononuclear cells (PBMC) are isolated using Ficoll density gradient centrifugation. Single-cell transcriptome analysis in combination with CITE-seq will be performed on 5000 PBMC. Cellular composition will be determined using the same bioinformatic pipelines as used for processing the BAL fluid cells.