Quantifying In Situ Adaptive Immunity in Human Tubulointerstitial Lupus Nephritis
Abstract
The most common, severe complication of lupus is inflammation of the kidneys referred to as lupus nephritis. In many lupus patients, nephritis is difficult to treat and ultimately progresses to kidney failure requiring dialysis. This occurs even in patients receiving our strongest drugs such as cyclophosphamide. It is not understood why lupus nephritis can be so difficult to treat. Several years ago, we postulated that there might be factors intrinsic to the kidney in lupus that make it resistant to current medications. Before this time, it had been assumed that lupus nephritis was a manifestation of the systemic autoimmunity that is so well studied in both mice and humans. In systemic autoimmunity, the disease-causing cells and proteins circulate through the blood and deposit in affected organs. However, we demonstrated that there was an immune response in the kidney itself, apparently separate from that in the blood, associated with irreparable kidney damage and progression to kidney failure. Furthermore, this immune response, while common in human lupus, is not accurately recapitulated in mouse models of lupus. Therefore, to understand the unique features of human lupus nephritis, we had to develop new techniques to study immune responses in kidney biopsies from patients with lupus nephritis. To understand autoimmunity in the kidneys, it is critical to understand how the immune system, and inflammation, is organized. However, when we started our studies, there were no tools available to understand how immune cells were organized in human inflamed tissue and to identify which cells were coordinating their efforts to drive local tissue damage. It was possible to image limited numbers of cell types in human tissue using a technique called multicolor confocal microscopy. However, there were not computational tools available to quantitatively analyze these images in a way that allowed the spatial and functional relationships between cells to be assessed. Recently, we have successfully used deep machine learning (Convolutional Neural Networks, CNN) to develop a novel set of computational tools that captures both the position and shape of cells in human tissue. In a series of experiments in mice and humans, we have validated that this approach, which we refer to as Cell Distance Mapping version 3 (CDM3). CDM3 can discriminate between cells that are productively interacting versus those cells that are just close to each other. Therefore, for the first time, the architecture of inflammation in human tissue can be quantified and compared across different patients and even different diseases. Such knowledge will allow us to identify which cells in specific diseases, and in individual patients, are cooperating to cause disease. The promise of CDM3 is that it will allow tailoring of therapies to target critical pathogenic mechanisms operative in individual patients. Through CDM3, we can bring personalized medicine to organ specific autoimmunity. The specific goals of this grant are to use high-dimensional (16 primary antibodies or more) multicolor confocal microscopy to describe the frequency and spatial distributions of the different types of immune cells infiltrating the lupus kidney (Aim 1). We will then use machine learning approaches, including CDM3, to quantify the true complexity of inflammation and identify functional relationships between different T cell populations and the cells that are presenting antigen, and therefore activating, pathogenic T cells (Aim 2). These studies will be performed on left-over material from diagnostic renal biopsies. Therefore, there is no risk to the patient associated with this study. This grant application addresses two of the three focus areas. Most directly, the proposal is fully focused on examining the pathobiology of lupus in a target organ, the kidney. In addition, we will address disease heterogeneity by identifying mechanistically different subsets of patients with lupus nephrit
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 19, 2019
- Source ID
- W81XWH1910642
Entities
People
- Marcus R Clark
Organizations
- United States Army
- University of Chicago