PROJECT SUMMARY The focus of the University of Chicago ACE (UCACE) is in situ human autoimmunity. During the last funding cycle, we successfully developed techniques to fully characterize the transcriptional state of single B cells and plasmablasts sorted from tissue samples and to pair this with analyses of functional immunoglobulin repertoire. Application of this approach to lupus tubulointerstitial inflammation (TII) and renal acute mixed allograft rejection (AMR) suggest that the activation state, antigenic repertoire and mechanisms of antigenic-driven B cell selection in human inflammation is fundamentally different than that typically observed in secondary lymphoid organs (Collaborative Project). In Celiac disease, B cells expressing transglutaminase 2 (TG2) specific antibodies are massively expanded in the duodenal mucosa. A highly-restricted repertoire of VH genes encode these antibodies, most notably VH5-51. Remarkably, this VH gene is also over-represented in the recirculating IgA+ B cell pool which is rich in anti-microbial activity. These findings suggest a model in which microbial antigens select for pathogenic TG2 reactive antibodies in susceptible hosts (Principle Project). Our second technical innovation is unique to the UCACE (Collaborative Project). Previous work has demonstrated that by quantifying the distance between T and B cells in multicolor confocal images (Cell Distance Mapping, CDM) we could identify competent TFH cells and functional relationships with B cells. We have now implemented a deep convolutional neural network (DCNN) that accurately identifies both cell position and shape in multicolor confocal images. In mice, analysis of the DCNN output (CDM3) indicates that T cell shape as a function of distance from dendritic cells (DCs) discriminates between cognate and non- cognate T cell:DC interactions with a sensitivity and specificity approaching that of two-photon emission microscopy (TPEM). In lupus TII, CDM3 both confirmed that myeloid DCs present antigen to CD4+ T cells in situ and identified plasmacytoid DCs as an important antigen presenting cell (APC) in severe TII. Finally, in the Pilot Project, we are applying microfluidics to develop in vitro culture systems capable of studying cognate interactions between single cells. These projects demonstrate a novel pipeline of methodologies to identify in situ cell populations, characterize their function and quantify the adaptive cell networks through which they cooperate to drive local adaptive immunity and inflammation.