DRGquant: A new modular AI-based pipeline for 3D analysis of the DRG

BACKGROUND: The dorsal root ganglion (DRG) is structurally complex and pivotal to systems processing nociception. Whole mount analysis allows examination of intricate microarchitectural and cellular relationships of the DRG in three-dimensional (3D) space. NEW METHOD: We present DRGquant a set of tools and techniques optimized as a pipeline for automated image analysis and reconstruction of cells/structures within the DRG. We have developed an open source software pipeline that utilizes machine learning to identify substructures within the DRG and reliably classify and quantify them. RESULTS: Our methods were sufficiently sensitive to isolate, analyze, and classify individual DRG substructures including macrophages. The activation of macrophages was visualized and quantified in the DRG following intrathecal injection of lipopolysaccharide, and in a model of chemotherapy induced peripheral neuropathy. The percent volume of infiltrating macrophages was similar to a commercial source in quantification. Circulating fluorescent dextran was visualized within DRG macrophages using whole mount preparations, which enabled 3D reconstruction of the DRG and DRGquant demonstrated subcellular spatial resolution within individual macrophages. COMPARISON WITH EXISTING METHOD(S): Here we describe a reliable and efficient methodologic pipeline to prepare cleared and whole mount DRG tissue. DRGquant allows automated image analysis without tedious manual gating to reduce bias. The quantitation of DRG macrophages was superior to commercial solutions. CONCLUSIONS: Using machine learning to separate signal from noise and identify individual cells, DRGquant enabled us to isolate individual structures or areas of interest within the DRG for a more granular and fine-tuned analysis. Using these 3D techniques, we are better able to appreciate the biology of the DRG under experimental inflammatory conditions.