Block Face Scanning Electron Microscopy of Fluorescently Labeled Axons Without Using Near Infra-Red Branding

Abstract

In this article, we describe the method that allows fluorescently tagged structures such as axons to be targeted for electron microscopy (EM) analysis without the need to convert their labels into electron dense stains, introduce any fiducial marks, or image large volumes at high resolution. We optimally preserve and stain the brain tissue for ultrastructural analysis and use natural landmarks, such as cell bodies and blood vessels, to locate neurites that had been imaged previously using confocal microscopy. The method relies on low and high magnification views taken with the light microscope, after fixation, to capture information of the tissue structure that can later be used to pinpoint the position of structures of interest in serial EM images. The examples shown here are td Tomato expressing cortico-thalamic axons in the posteromedial nucleus of the mouse thalamus, imaged in fixed tissue with confocal microscopy, and subsequently visualized with serial block-face EM (SBEM) and reconstructed into 3D models for analysis.

Keywords: axons; correlative light and electron microscopy (CLEM); neuron ultrastructure; scanning electron microscopy; serial block-face electron microscopy (SBEM).

Figure 1

Locating fluorescent structures in an 80-micrometer thick coronal brain section from a P18 Rbp4-Cre;Ai14 mouse. (A) Low magnification image of the thalamic region in the vibratome section. The yellow box indicates the region shown in (B). (B,C) Higher magnification views of the region containing the fluorescent axons of interest showing a prominent blood vessel indicated with a white dotted line. Yellow box in (B) indicates region shown in (C). (D) Confocal image of corticothalamic axonal boutons in the region indicated in (C). The dark shadow of the same blood vessel is seen at a depth of 30 micrometers. (E) At a shallower depth of 8 microns (z = −8 μm) a specific bouton is singled out from imaging with electron microscopy (EM). (F) Higher magnification view of the region indicated with the yellow box in (E). White arrowhead indicates a single bouton that is targeted for ultrastructural analysis. Scale bar in (A) is 1 mm; (B), 200 μm; (C), 50 μm; (D) 25 μm; (F), 10 μm.

Figure 2

The region of interest can be located in the tissue section using blood vessels and the edge of the tissue section as guides. The resin-embedded tissue is trimmed according to these landmarks. (A) The wet section is imaged with transmitted light to show the position of the blood vessel in relation to the entire section. The arrowhead indicates the same highlighted blood vessel shown in Figures 1A–D. The white arrowhead in (A–C) indicates the exact same position in the tissue. (B) After the section has been heavy-metal stained, and resin embedded, the same region can be localized by overlaying the image of the wet section (shown in A) with that of the resin-embedded section, whether or not the blood vessels are still visible. (C) This region is trimmed from the rest of the section by sequentially removing one side of a square that contains the region of interest. The yellow boxes in the first three trims indicate the final region stuck to the stub. After each trim, an image is taken and this is overlaid with the previous so that the yellow box can be positioned precisely, indicating the region of interest. The small yellow box shown in the 4th trim shows the final region that remains after trimming in the ultramicrotome. This region corresponds with the block face shown in Figure 3. (D) The schematic diagrams show the initial and final trimming of the block from the side that is mounted on the pin. The block is initially trimmed to leave only the region of interest of approximately 250 × 250 μm. This is then surrounded in the conductive glue again, and this is trimmed away again until some of the resin remains around the edge to help with conduction. Scale bar in (A,B) is 1 mm.

Figure 3

The final trimmed block is scanned in the microscope revealing the position of cell bodies and blood vessels seen with fluorescence microscopy. (A) The block is carefully trimmed so that the region of interest lies just within its borders. An image of this final block can be overlaid precisely with an image of the entire region taken prior to trimming. Then, a box (yellow) is able to indicate the precise region where images can be taken. (B) The block is then placed in the microscope and once a large part of the face has been revealed with the diamond knife, an scanning EM (SEM) image is taken. (C) The region of interest is then confirmed by matching the position of cells and vessels (arrowheads) with those seen previously in the confocal image stacks. Their height in the image stacks is also noted. Scale bar in (B) is 50 μm and in (C; right-hand images) is 20 μm.

Figure 4

The matching of structures seen in both fluorescence and EM images allows features of interest to be identified and serially imaged. (A) Fluorescent image of labeled axons and their boutons (arrowed) shows their location in relation to the shadows of a cell body and blood vessel (labeled). Bouton indicated with a white arrow is the same bouton from Figure 1. (B) EM micrograph of the same region shown in (A) with the corresponding features labeled. These features are used to pinpoint the position of the axons and axon terminals (pseudo-colored in purple). (C) Higher magnification view of the highlighted region in (B). (D) The axonal bouton shown in three images from the series, and from which the reconstruction is made. The depth of each image in the stack is shown at the bottom left of each image. (E) Reconstruction of the bouton (purple) shown in (D) and the dendrite (yellow) to which it synapses. Lower image, side view shows the position of each of the images shown in (D). Scale bar in (A,B; left hand image) is 10 μm; in (C) is 2.5 μm; in (D) is 1 μm.