PROJECT SUMMARY: Chronic lung allograft dysfunction (CLAD) is the primary driver of morbidity and mortality in lung transplant recipients. Currently there is a need to identify clinical and molecular biomarkers of CLAD, where the latter has the potential to inform targetable pathways for intervention. There is growing evidence to hypothesize that a key component of CLAD pathobiology is the recruitment of profibrotic monocyte-derived alveolar macrophages (MoAM) upon injury to the lung epithelium. Profibrotic MoAMs stimulate the activation, differentiation, and proliferation of myofibroblasts. With sustained injury, MoAMs are continually maintained through colony stimulating factor 1 (CSF1) signaling through its cognate receptor (CSF1R), leading to progressive fibrosis. T-regulatory cells dampen ongoing injury and have been shown to mitigate CLAD in preclinical models and are also associated with more favorable lung transplant outcomes in humans. Our group demonstrated that recipient-derived MoAMs express profibrotic genes in mice and humans and that administering a CSF1R antagonist improved fibrosis in a murine model of CLAD. Translating these findings in humans requires analyzing longitudinal data collected before CLAD diagnosis. Where most studies focus on measurements made at one or two points in time, often when CLAD has significantly progressed, the proposed work will leverage a machine learning approach developed by our group to determine clinical states and their sequences that develop after lung transplantation. This work will examine associations between these clinical states with flow cytometry, single cell and spatial transcriptomic analysis of BAL fluid across time to identify early, predictive indicators of CLAD. Specifically, the proposed work will address the hypothesis that the emergence of pathogenic MoAM and loss of tissue-resident donor-derived T-regs in serially sampled BAL predict CLAD and ACR respectively. Aim 1 will determine whether the CSF1-driven maintenance of profibrotic MoAMs precedes the clinical diagnosis of CLAD. Aim 2 will determine whether the paucity of tissue-resident donor-derived T-regs is associated with CLAD after ACR. Both aims consist of 1. Combining flow cytometry and single-cell transcriptomics to quantify cell abundances and ligand receptor analyses relevant to either aim unique to CLAD BAL and 2. Integrating these molecular features with clinical data in machine learning models for CLAD (aim 1) and ACR (aim 2) prediction. In leveraging these data-driven and machine learning approaches, the long term goal of the proposed work is to reveal therapeutic targets and elucidate early signs of ACR and CLAD for timelier intervention and to ultimately reduce lung transplant failure. The candidate and her mentors have designed a detailed training plan that utilizes the support of diverse mentors and resources in immunology, single-cell genomics, machine learning, and translational research. Ultimately, this train...