Background

Vasculature plays a key role during the growth, maintenance, and repair of tissues throughout the body. Accumulating evidence suggests that the slow and protracted reorganization of vascular networks can promote recovery of function (Felmeden et al., 2003; Evans et al., 2021) but may also contribute to neuropathological hallmarks of injury and disease (Ostergaard et al., 2016).

A critical point to consider is that standard methods to study vasculature changes in laboratory settings have proven inaccurate and labor-intensive, lacking the inherent three-dimensional structure of the vasculature. Conventionally, visualization of blood vessels in tissues of interest often employs immunohistochemical (IHC) techniques. This approach uses thin tissue sections to detect endothelial cells through vessel wall stains. The resulting two-dimensional images do not capture the inherent three-dimensional, tortuous structure of vasculature networks (Rust et al., 2020). Furthermore, IHC is a multistep process, increasing the likelihood of introducing error, non-experimental variables (e.g., lot-to-lot variability that contributes to inconsistent results in assays using antibodies), or experimenter bias to analysis, as only a small fraction of the tissue is typically processed. Thus, quantifying the vascular response to injury or disease is often limited to the mean fluorescent intensity of IHC stains, which is highly dependent on technical consistency and subject to bias (Marien et al., 2016; Cheung et al., 2020). Methods of three-dimensional imaging of vasculature networks are currently available. However, they are niche in their applications throughout the body and are largely constrained by expensive or otherwise inaccessible imaging equipment and methodologies including tomography, magnetic resonance imaging, and photoacoustic imaging (Xiong et al., 2017; Pac et al., 2022; Menozzi et al., 2023).

To overcome these barriers, we have developed a practical and accessible pipeline that permits robust and highly reproducible qualitative and quantitative three-dimensional analysis of vascular networks that can resolve changes to the capillary level. Initially, we detail our approach for vasculature tracing and preparing unsectioned tissue samples throughout the body. Then, we present a methodology for optically clearing neuronal and non-neuronal tissues (Susaki et al., 2015) and provide the necessary steps to acquire three-dimensional images using conventional laser scanning confocal microscopy. Lastly, we detail our pipeline for vessel segmentation using AI-based convolutional reconstruction (Imaris software, paid subscription) or semi-automated three-dimensional reconstruction (ImageJ software, freely available). This framework can consistently capture subtle changes in capillary structure such as total length, branching points, branch length, and diameter of vessels, providing a robust and user-friendly approach to study vascular networks in health and disease. Additionally, this scalable method can be adapted for volumetric imaging and analysis of vasculature structures in a variety of preclinical models, thus accelerating translation of promising findings to large animals and non-human primates.