PROJECT SUMMARY Lupus nephritis (LN) is the most common severe manifestation of systemic lupus erythematosus (SLE) Disease progression is associated with tubulointerstitial hypoxia and metabolic dysfunction, capillary rarefaction, accumulation of immune infiltrates and fibrosis. Complete remission rates in patients with LN are <50% (and often <30%) even in the setting of rigorous clinical trials and responses cannot currently be predicted by clinical and histologic features at initial biopsy. We currently have insufficient understanding of why clinical outcomes do not always correlate with histologic changes, why only some patients with interstitial kidney inflammation progress to ESRD, and how to predict responses to therapeutic intervention. We do not know which human in situ disease mechanisms are manifest in which murine LN models. Indeed, murine models have often failed to predict clinical utility in human LN. While no murine model provides a holistic picture of human LN, this would not be expected as the human disease is very heterogeneous. Rather, we propose an overall hypothesis that specific pathogenic human LN immune states are quantitatively replicated in select murine models of lupus nephritis. We will test this hypothesis using innovative high- dimensional confocal microscopy and AI-driven analysis of both human and murine LN tissue. These studies will be complemented by directed mechanistic studies in relevant LN murine models. Specific Aims: Aim 1. To define prognostically important in situ autoimmune states in human LN. We hypothesize that specific in situ immunological architectures, associated with specific CD4- T cell and myeloid cell populations, will define therapy-resistant, progressive renal disease. Aim 2. Quantify in situ T cell states in murine LN models and their relationship to human LN. We hypothesize that in situ T cell architectures implicated in progressive human LN will be approximated in select murine LN models and that these areas reflect sites of pathogenic CD8+ T cell clonal expansion. Aim 3: To quantify in situ myeloid immune states in murine LN models and their relationship to human LN. We hypothesize that in situ myeloid cell architectures implicated in progressive human LN will be identified in mouse models, and that only some myeloid subsets will be associated with tissue injury and fibrosis. This information will inform complementary functional studies in mice.