Multiplexed High-Resolution Imaging Approach to Decipher the Cellular Heterogeneity of the Kidney and its Alteration in Kidney Disease and Nephrolithiasis

Abstract

Kidney disease and nephrolithiasis both present a major burden on the health care system in the US and worldwide. The cellular and molecular events governing the pathogenesis of these diseases are not fully understood. We propose that defining the cellular heterogeneity and niches in human and mouse kidney tissue specimens from controls and various models of renal disease could provide unique insights into the molecular pathogenesis. For that purpose, a multiplexed fluorescence imaging approach using co-detection by Indexing (CODEX) was used, using a panel of 33 and 38 markers for mouse and human kidney tissues, respectively. A customized computational analytical pipeline was developed and applied to the imaging data using unsupervised and/or semi-supervised machine learning and statistical approaches. The goal was to identify various cell populations present within the tissues, as well as identify unique cellular niches that may be altered with disease and/or injury. In mice, we examined disease models of acute kidney injury (AKI) and in human tissues we analyzed specimens from patients with AKI, IgA nephropathy, chronic kidney disease, systemic lupus erythematosus, and nephrolithiasis. In both mice and humans, the disease and reference samples show similar broad cell populations for the main segments of the nephron, endothelium, as well as similar groups of immune cells, such as resident macrophages and neutrophils. When comparing between health and disease, however, a change in the distribution of few sub-populations occurred. For example, in human kidney tissues, the abundance and distribution of a subpopulation of proximal tubules positive for THY1 (a marker of differentiation and repair), was markedly reduced with disease. Changes observed in mouse tissues included shifts in the immune cell population types and niches with disease. We propose that our analytical workflow and the observed changes in situ will play an important role in deciphering the pathogenesis of kidney disease.