Accelerating target deconvolution for therapeutic antibody candidates using highly parallelized genome editing

Abstract

Therapeutic antibodies are transforming the treatment of cancer and autoimmune diseases. Today, a key challenge is finding antibodies against new targets. Phenotypic discovery promises to achieve this by enabling discovery of antibodies with therapeutic potential without specifying the molecular target a priori. Yet, deconvoluting the targets of phenotypically discovered antibodies remains a bottleneck; efficient deconvolution methods are needed for phenotypic discovery to reach its full potential. Here, we report a comprehensive investigation of a target deconvolution approach based on pooled CRISPR/Cas9. Applying this approach within three real-world phenotypic discovery programs, we rapidly deconvolute the targets of 38 of 39 test antibodies (97%), a success rate far higher than with existing approaches. Moreover, the approach scales well, requires much less work, and robustly identifies antibodies against the major histocompatibility complex. Our data establish CRISPR/Cas9 as a highly efficient target deconvolution approach, with immediate implications for the development of antibody-based drugs.

Fig. 1. Antibody target deconvolution using CRISPR/Cas9 screening.

a Schematic outline of the target deconvolution process. Cells staining positive with the antibody of interest are transduced with a lentiviral sgRNA/Cas9 knockout library resulting in a heterogenous cell pool with a small population of antigen-negative cells. These cells with gene knockouts leading to lost or diminished antibody binding are isolated by FACS, the genomic DNA is extracted, and the sgRNA-encoding DNA is sequenced on the Illumina NextSeq 500 platform. Genes with sgRNAs enriched in the antigen-negative cells are identified, resulting in a proposed antibody target. b–d Representative example of flow cytometry data for mAb binding to transduced cells detected with anti-human IgG-APC. b Input cells before sort. c Pre-enriched cells after one sort with a distinct fraction of antigen-negative cells. d Sequenced cell pool after two sorts with a highly enriched fraction of antigen-negative cells. e Representative example of MAGeCK gene enrichment scores for two sort replicates, showing that the same genes are identified for both replicates. Sorts were performed in 3 to 5 replicates. Source data are provided as a Source Data file.

Fig. 2. Identification of antibody targets.

a MAGeCK gene enrichment scores for all genes across the 39 test mAbs. Target genes selected using our three criteria indicated in red, MHC class I dependency genes in pink. b Representative example showing enrichment of high MAGeCK scores among MHC class I dependency genes for a mAb against the MHC class I complex (mAb19). MAGeCK false discovery rate (x-axis) vs gene enrichment score for identified MHC class I specific antibody mAb19 (y-axis). Target genes selected using our three criteria indicated in red, MHC class I dependency genes in pink. c MHC class I enrichment score, calculated as ‒log₁₀(P-value) for one-sided Wilcoxon rank-sum test for MAGeCK scores for our 11 selected MHC class I dependency genes vs other genes in the genome. As shown, we observed significantly higher enrichment scores for mAbs confirmed to target MHC class I (n = 9) compared to antibodies with non-MHC targets (n = 25; **** Indicates P < 0.0001 by one-sided Wilcoxon rank-sum test; exact P-value 5.5 × 10⁻⁶). Based on this observation, we infer that four antibodies (right) that showed high MHC class I enrichment scores but whose specificity could not be validated experimentally (n = 4) also target MHC class I. d Histogram of the rankings of the best-scoring gene representing the target in the MAGeCK analysis across the 38 resolved mAbs. Source data are provided as a Source Data file.

Fig. 3. DepMap scores showing gene essentiality.

Since deconvolution by CRISPR/Cas9 requires that the target protein encoding gene is not essential for survival of the test cells, we investigated if cell membrane genes are more, or less, essential than genes in general. To this end, we used dependency scores from CRISPR/Cas9 proliferation screens for 436 cell lines from DepMap. In essence, a negative DepMap score indicates that a gene is essential for the survival/proliferation of a large fraction of cell lines. Gray lines indicate all genes, blue lines investigated gene set. a Distributions of DepMap scores for membrane protein genes and cluster designation (CD) antigen genes. As shown, these gene sets are significantly depleted of negative DepMap scores relative to other genes in the genome, indicating that they are less essential. This observation indicates that failure to deconvolute a target because of essentiality is unlikely. b By contrast, gene sets that are essential for cell survival are enriched for negative DepMap scores, exemplified here by ribosomal protein genes. In all panels, P-values for enrichment/depletion were calculated using RenderCat using default settings (two-sided Zhang C goodness-of-fit test).