Deep learning-based vessel extraction in 3D confocal microscope images of cleared human glioma tissues.

Comprehensive visualization and accurate extraction of tumor vasculature are essential to study the nature of glioma. Nowadays, tissue clearing technology enables 3D visualization of human glioma vasculature at micron resolution, but current vessel extraction schemes cannot well cope with the extraction of complex tumor vessels with high disruption and irregularity under realistic conditions. Here, we developed a framework, FineVess, based on deep learning to automatically extract glioma vessels in confocal microscope images of cleared human tumor tissues. In the framework, a customized deep learning network, named 3D ResCBAM nnU-Net, was designed to segment the vessels, and a novel pipeline based on preprocessing and post-processing was developed to refine the segmentation results automatically. On the basis of its application to a practical dataset, we showed that the FineVess enabled extraction of variable and incomplete vessels with high accuracy in challenging 3D images, better than other traditional and state-of-the-art schemes. For the extracted vessels, we calculated vascular morphological features including fractal dimension and vascular wall integrity of different tumor grades, and verified the vascular heterogeneity through quantitative analysis.

Sudan Black B treatment for reducing autofluorescence in human glioma tissue and improving fluorescent signals of bacterial LPS staining.

The 3D visualization based on tissue clearing technology allows us to have a deeper understanding of the 3D spatial information of deep molecules in the tissue. Tissue clearing and bacterial labeling methods have been used for in situ 3D microbiota imaging, and we have developed a pipeline for 3D visualization of in situ microbiota in human gliomas. Anti-LPS antibodies are appropriate to label and characterize bacteria in situ within tumors. However, autofluorescence (AF) is common in biological tissues, especially in brain tissues filled with lipofuscin-like (LF) substances. This natural fluorescent signal is usually considered to be a problem because it affects the 3D visualization of fluorescent signals in bacterial LPS staining. Here, we used Sudan Black B (SBB) to mask the AF of human glioma tissue and explored in detail the optimal quencher concentration, which allows 3D visualization of intratumoral bacteria to reduce AF and maintain the intensity of intratumoral bacteria-specific LPS fluorescent signals.

Comprehensive histological imaging of native microbiota in human glioma.

Mounting evidence suggests that distinct microbial communities reside in tumors and play important roles in tumor physiology. Recently, a previous study profiled the composition and localization of intratumoral bacteria using 16S ribosomal DNA (rDNA) sequencing and histological visualization methods across seven tumor types, including human glioblastoma. However, their results based on traditional histological examinations should be further validated considering potential sources of contamination originating from sample collection and processing. Here, we aim to propose a three-dimensional (3D) in situ intratumoral microbiota visualization and quantification protocol avoiding surface contamination and provide a comprehensive histological investigation on local bacteria within human glioma samples. We develop a 3D quantitative in situ intratumoral microbiota imaging strategy, combining tissue clearing, immunofluorescent labeling, optical sectioning microscopy, and image processing, to visualize bacterial lipopolysaccharide (LPS) within gliomas in a direct, contaminant-free, and unambiguous manner. Through an automated statistical algorithm, reliable signals can be distinguished for further analysis of their sizes, distribution, and fluorescence intensities. In tandem, we also combined 2D images obtained from thin-section histological methods, including immunohistochemistry and fluorescence in situ hybridization, to provide comprehensive histological imaging for local bacterial components within human glioma samples. We have, for the first time, achieved 3D quantitative imaging of bacterial LPS colonized in gliomas in a contamination-free manner within human glioma samples. We also built the multiple histological evidence chain demonstrating the irregular shapes and sparse distribution of bacterial components within human glioma samples, mostly localized near nuclear membranes or in the intercellular space. This study provides favorable evidence for the presence of microbiota in human gliomas and provides information on the feature and distribution of bacterial components. The results, along with the integrated 3D quantitative intratumoral microbiota imaging method, are promising to provide insightful information into the direct interactions between the microbial community and the host in the tumor microenvironment.