Late Breaking Abstract - High-throughput in vivo functional genomics screen identifies therapeutic candidates that regulate lung physiology and ECM synthesis in pulmonary fibrosis

<bold>Intro:</bold> Pulmonary fibrosis (PF) is a culmination of complex interplay between injured epithelium, immune cells, and activated fibroblasts. Here we present in vivo pooled screens to assess therapeutic efficacy of 100s of targets directly in the complex disease environment. <bold>Obj:</bold> Demonstrate the predictive power of high-throughput in vivo functional genomics screen for therapeutic discovery. <bold>Methods:</bold> A multi-dose bleomycin mouse model of PF received a barcoded AAV library intratracheally, testing 250 therapeutically relevant genetic perturbations. After 3 weeks, transduced fibroblasts and epithelial cells were sorted and profiled by single-cell RNA-seq. Predictive algorithms for physiological outcomes nominated candidates were validated using human precision-cut lung slices (PCLS) and mouse model of chronic PF. <bold>Results:</bold> Disease cell state were characterized by 50 transcriptomic predictors for the range of PF pathophysiology. 10 perturbations including JAK1, TGFBR2, and SMAD4, produced robust transcriptomic shifts (p<0.05), correlating with improved pulmonary physiology and reduced ECM gene expression. In PCLS and in vivo, knockdown of novel targets led to reduced Sircol levels similar to TGFBR2 and SMAD4, downregulation of COL1A1, ACTA2, COL3A1, FN1, and upregulation of genes linked to improved FVC. <bold>Conclusion:</bold> Our in vivo mosaic screening platform enables efficacy testing of >200 therapeutic targets in a single experiment, broadening the search space for novel targets. The transcriptomic resolution and validation in human PCLS demonstrate its predictive power and utility for accelerating target discovery for complex diseases like PF.