ScoMorphoFISH: A deep learning enabled toolbox for single‐cell single‐mRNA quantification and correlative (ultra‐)morphometry

Abstract

Increasing the information depth of single kidney biopsies can improve diagnostic precision, personalized medicine and accelerate basic kidney research. Until now, information on mRNA abundance and morphologic analysis has been obtained from different samples, missing out on the spatial context and single‐cell correlation of findings. Herein, we present scoMorphoFISH, a modular toolbox to obtain spatial single‐cell single‐mRNA expression data from routinely generated kidney biopsies. Deep learning was used to virtually dissect tissue sections in tissue compartments and cell types to which single‐cell expression data were assigned. Furthermore, we show correlative and spatial single‐cell expression quantification with super‐resolved podocyte foot process morphometry. In contrast to bulk analysis methods, this approach will help to identify local transcription changes even in less frequent kidney cell types on a spatial single‐cell level with single‐mRNA resolution. Using this method, we demonstrate that ACE2 can be locally upregulated in podocytes upon injury. In a patient suffering from COVID‐19‐associated collapsing FSGS, ACE2 expression levels were correlated with intracellular SARS‐CoV‐2 abundance. As this method performs well with standard formalin‐fixed paraffin‐embedded samples and we provide pretrained deep learning networks embedded in a comprehensive image analysis workflow, this method can be applied immediately in a variety of settings.

Super-Resolution Microscopy: A Technique to Revolutionize Research and Diagnosis of Glomerulopathies

“…Using 3D-SIM, podocytes with elongated podocin-stained filtration slits showed ACE2 expression (arrowheads in Fig. 6d, e), illustrating the combination of SRM and single-cell expression analysis [26].…”

“…3c, unpublished 3D-SIM data of a murine kidney section by our group), the width of the GBM can be easily evaluated in parallel to PEMP analysis. To date, PEMP has emerged as an important tool and has been used in several different publications to determine the FSD in animal models [18,[20][21][22][23][24][25], human samples [26], or even organoids [27]. Additional interesting development was presented by Matsumoto and colleagues, who imaged stained kidney sections with regular histological dyes periodic acid Schiff (PAS) and Elastica-Masson trichrome on FFPE material.…”

“…In a first proof-of-principle work, we trained DL networks to segment glomeruli and single cells from kidney biopsies using single glomerular markers and nuclear counterstains, respectively [26]. Subsequently, single-mRNA localizations derived from multiplexed single-mRNA in situ hybridization were annotated for every single cell.…”

High-throughput image analysis with deep learning captures heterogeneity and spatial relationships after kidney injury

“…U-Net requires GPU acceleration to operate efficiently, but its ImageJ plugin was designed to use remote servers to perform these calculations, which are inexpensive and readily available. U-Net recently has been deployed in the kidney for specific applications such as defining the spatial context for transcripts detected by in situ hybridization 17 and performing podocyte morphometrics in human biopsy samples 18 . These examples reinforce the utility of U-Net for the more generalized analysis of immunofluorescence-stained kidney samples outlined here.…”

Virus-response proteinuria underlies rapid changes in kidney filtration barrier function in COVID-19

“…So far, direct kidney infection by SARS-CoV-2 remains controversial, as proof of direct infection came largely from autopsy kidney tissues but rarely seen in biopsy tissues [26][27][28][29] . We explored if kidney cells require direct infection by the virus or if circulating viral proteins can be sufficient in causing proteinuria.…”