Correlative microscopy on tissue: Development applied to the skin and the nervous system

Abstract

Correlative Light and Electron Microscopy (CLEM) combines two imaging techniques at different scales to figure out the precise localization of rare elements in a well‐defined biological context. Light microscopy (LM) allows positional mapping of the sample labeled with a fluorochrome. Electron microscopy (EM) provides the nanometer resolution of the mapped area. The general principle of CLEM is to collect different informations from a single region in a sample. The data are then combined toward a global understanding of the sample ultrastructure. The precise localization of molecules of interest in their biological context serves to define their functional role. In the present study, we aimed to apply CLEM to tissue samples. We have chosen define regions of interest (ROI) by fiducial laser marks surrounding the ROI. For this purpose, we applied the “Near infrared branding” (NIRB) method (Bishop et al., 2011). This NIRB technique consists to create easily detectable fiducial marks in three dimensions in a fixed tissue sample. A specific region of the sample is imaged with LM, the fiducial marks are performed, and the same region is analysed with EM. The fiducial marks are detectable both in optical and electron microscopy. They are made with a femtosecond pulsed titanium‐sapphire laser. The laser is used to create a user defined pattern with line or points scans. The size of the marks depends on the fraction of laser power used or the number of laser swipes. The NIRB marks can be placed with a three‐dimensional micrometer precision in close proximity to the ROI. Laser‐made marks are visualized through tissue photo‐oxidation that induces autofluorescence in LM and is a characteristic marker for EM. When using Green Fluorescent Protein (GFP) labelling, no GFP photo‐oxidation is noticed when drawing NIRB. The present study focused on adjustments that are necessary to adapt the NIRB method to nervous tissue and skin. These two types of tissue present a very different composition and are therefore well suited to compare the conditions of NIRB implementation. The NIRB technique has been set using immunostaining of easily detectable and strongly represented antigens with a specific location in well‐defined regions. In nervous tissue, the constitutive extra‐membranous mitochondrial protein TOM20 was used as a target antigen in spinal cord sections. For the skin, the development relied on Langerin immunostaining (a type II transmembrane, C‐type lectin receptor on Langerhans cells). The goal is to locate the Langerhans cells in the epidermis. All immunostaining strategies are performed with a pre‐embedding method. It consists in immunostaining on chemically fixed samples before embedding in resin and ultra‐sectioning for TEM. First, we set the method to achieve the ROI localization. Then, we defined our own parameter settings using our multiphoton system. Finally, it is essential to adapt the marks to the size of the region of interest, and the type of sample. The delimited ROI can size up to 70μm3. Future perspectives are to develop three‐dimensional CLEM approach using electron tomography.