Long-lasting circuits between T cells and macrophages drive acute cellular rejection in lung transplant patients - PROJECT SUMMARY/ABSTRACT Lung transplantation is the only therapeutic option available for patients with advanced lung diseases, yet it has the lowest long-term survival rate among solid organ transplants, with only 59% of recipients surviving beyond five years. Acute cellular rejection (ACR), a frequent complication post-transplant, significantly contributes to graft failure by driving immune-mediated lung injury. Despite protocol-driven immunosuppression, many patients experience persistent ACR episodes whose underlying mechanisms remain unclear. This underscores an urgent need for more sensitive, accurate, and non-invasive diagnostic tools to detect ACR early, thereby improving patient outcomes. Previous studies have identified CD8+ T cells as central mediators of lung transplant rejection. However, the cellular and molecular pathways that sustain their prolonged activation and infiltration despite immunosuppression are poorly understood. Preliminary data from mouse models of lung transplant rejection suggest a critical role for monocyte-derived interstitial macrophages (MoIM) in promoting the persistence of autoreactive CD8+ T cells through chemokine signaling, particularly via CCL13 and CXCL9. Moreover, preliminary human data revealed a similar MoIM population in lung transplant recipients experiencing graft failure, suggesting a conserved mechanism in human ACR pathology. This proposal hypothesizes that CCL13+CXCL9+ MoIM form specialized immune niches that facilitate the sustained recruitment and activation of CD8+ T cells, driving persistent rejection despite protocol-driven immunosuppressive treatment. To test this, I will leverage clinical biopsy samples, digital pathology, single-cell spatial transcriptomics, high-resolution computed tomography (HRCT) imaging, and electronic health records (EHR) data from lung transplant patients. Aim 1 will investigate whether increased signaling by CCL13+CXCL9+ MoIM is associated with persistent CD8+ T cell activation during ACR. Using single-cell spatial transcriptomics and digital pathology, spatial immune niches will be characterized in biopsies from patients experiencing ACR versus matched controls. Aim 2 will evaluate whether machine learning (ML)-based analysis of HRCT imaging can non-invasively detect structural lung changes indicative of ACR. Transformer and Mamba deep learning models will analyze longitudinal HRCT data, focusing on early detection and interpretability of structural lung alterations associated with rejection. Aim 3 will integrate digital pathology, imaging, and EHR data into a unified multimodal ML model designed to enhance diagnostic accuracy for ACR detection. Fusion strategies, including early, intermediate, and late integration approaches, will be systematically compared, alongside interpretability analyses to ensure clinical relevance and actionable insights. The long-term goal of the proposed work is to elucidate immune mechanisms driving persistent ACR, identify novel imaging biomarkers for non-invasive diagnosis, and develop multimodal ML tools for improved clinical management of lung transplant recipients, ultimately aiming to enhance patient outcomes.
