Feuchtinger, T. | bioRxiv Science

bioRxiv2022

Modular Pooled Discovery of Synthetic Knockin Sequences to Program Durable Cell Therapies

Chronic stimulation can cause T cell dysfunction and limit efficacy of cellular immunotherapies. CRISPR screens have nominated gene targets for engineered T cells, but improved methods are required to compare large numbers of synthetic knockin sequences to reprogram cell functions. Here, we developed Modular Pooled Knockin Screening (ModPoKI), an adaptable platform for modular construction of DNA knockin libraries using barcoded multicistronic adaptors. We built two ModPoKI libraries of 100 transcription factors (TFs) and 129 natural and synthetic surface receptors. Over 20 ModPoKI screens across human TCR and CAR T cells in diverse conditions identified a transcription factor AP4 (TFAP4) construct to enhance long-term T cell fitness and anti-cancer function in vitro and in vivo. ModPoKIs modularity allowed us to generate a [~]10,000-member library of TF combinations. Non-viral knockin of a combined BATF-TFAP4 polycistronic construct further enhanced function in vivo. ModPoKI facilitates discovery of complex gene constructs to program cellular functions. HighlightsO_LIModular pooled knockins of hundreds of TF and surface receptor constructs combined with different antigen receptors C_LIO_LIChronic stimulation screens discover programs to improve T cell persistence C_LIO_LICombinatorial knockin screens with [~]10,000 transcription factor combinations C_LIO_LIBATF-TFAP4 dual knockin construct improves CAR T cell function in vitro and in vivo C_LI

immunology↗

bioRxiv2017

Systematic identification of cancer-specific MHC-binding peptides with RAVEN

Immunotherapy can revolutionize anti-cancer therapy if specific targets are available. Recurrent somatic mutations in the exome can create highly specific neo-antigens. However, especially pediatric cancers are oligo-mutated and hardly exhibit recurrent neo-antigens. Yet, immunogenic peptides encoded by cancer-specific genes (CSGs), which are virtually not expressed in normal tissues, may enable a targeted immunotherapy of such cancers. Here, we describe an algorithm and provide a user-friendly software named RAVEN (Rich Analysis of Variable gene Expressions in Numerous tissues), which automatizes the systematic and fast identification of CSG-encoded peptides highly affine to Major Histocompatibility Complexes (MHC) starting from publicly available gene expression data. We applied RAVEN to a dataset assembled from more than 2,700 simultaneously normalized gene expression microarrays comprising 50 tumor entities, with a focus on sarcomas and pediatric cancers, and 71 normal tissue types. RAVEN performed a transcriptome-wide scan in each cancer entity for gender-specific CSGs. As a proof-of-concept we identified several established CSGs, but also many novel candidates potentially suitable for targeting multiple cancer types. The specific expression of the most promising CSGs was validated by qRT-PCR in cancer cell lines and by immunohistochemistry in a comprehensive tissue-microarray comprising 412 samples. Subsequently, RAVEN identified likely immunogenic peptides encoded by these CSGs by predicting the affinity to MHCs. Putative highly affine peptides were automatically crosschecked with the UniProt protein-database to exclude sequence identity with abundantly expressed proteins. The predicted affinity of selected peptides was validated in T2-cell peptide-binding assays in which many showed similar kinetics to a very immunogenic influenza control peptide.\n\nCollectively, we provide a comprehensive, exquisitely curated and validated catalogue of cancer-specific and highly MHC-affine peptides across 50 cancer entities. In addition, we developed an intuitive and freely available software to easily apply our algorithm to any gene expression dataset (https://github.com/JSGerke/RAVENsoftware). We anticipate that our peptide libraries and software constitute a rich resource to accelerate the development of novel immunotherapies.

immunology↗