Skull optical clearing window for in vivo imaging of the mouse cortex at synaptic resolution

Imaging cells and microvasculature in the living brain is crucial to understanding an array of neurobiological phenomena. Here, we introduce a skull optical clearing window for imaging cortical structures at synaptic resolution. Combined with two-photon microscopy, this technique allowed us to repeatedly image neurons, microglia and microvasculature of mice. We applied it to study the plasticity of dendritic spines in critical periods and to visualize dendrites and microglia after laser ablation. Given its easy handling and safety, this method holds great promise for application in neuroscience research.


Imaging the dendritic spines in infantile and adult mice through the SOCW
We keep the skull of infantile mice intact without thinning (Supplementary Fig. 1a). The composition and thickness of the skull changes with age, which lead to strong scattering and further reduce the two-photon imaging performance. To visualize the dendritic spines of mice older than P30, we have to thin the skull to about 100 μm before clearing (Supplementary Fig. 1b). Supplementary Fig. 1c-d shows that the maximum projections of Thy1-YFP neurons across 10-15 µm images below the surface through the skull of mice aged P19 (Supplementary Fig. 1c) and P60 (Supplementary Fig. 1d) before and after skull optical clearing. We can see that the image quality is obviously improved and it is sufficient for imaging dendritic spines after treatment with OCAs.

Using one-photon to illuminate
We used one-photon microscopy to image the dendrites of Thy1-YFP neurons (P30) through the intact skull.
After clearing, the image quality is obviously improved and the imaging depth increases from 20 µm to about 60 µm (Supplementary Fig. 2). Therefore, this method is also effective for confocal microscopy. The images were acquired by Zeiss 780 LSM confocal microscope with a water-immersion objective (20×, numerical aperture = 1.0, working distance = 1.8 mm, Zeiss). And the 488 nm laser was used to illuminate.

Imaging the interneurons expressing RFP with the SOCW
We used transgenic mice Sst-IRES-Cre::Ai14 to image the cortical interneurons expressing red fluorescent proteins (RFP). Before clearing, we could hardly get any information about interneurons, but after clearing, the interneurons across 100-200 µm below the surface could be obtained. Supplementary Fig. 3 shows the images of mice aged P30. Thus, this method is also compatible for RFP as well.

Dual-color imaging
Dual-color imaging could also be realized under SOCW (Supplementary Fig. 4). The blood vessels of transgenic mice were labeled with fluorescence dye (Tetramethylrhodamine-conjugated dextran) via tail vein injection. Supplementary Fig. 4a-b shows that the cerebral vasculature can be imaged at the same time as 1 dendrites ( Supplementary Fig. 4a) and microglia (Supplementary Fig. 4b).

Repeated imaging of microglia and brain micro-vessels
And we also repeatedly imaged microglia in mice Cx3cr1 EGFP/+ (P28) and subsurface brain micro-vessels in C57BL/6 mice (P28) in which the plasma had been stained with fluorescein-conjugated dextran. Supplementary Fig. 5a-b shows representative fluorescence images of GFP-expressing microglia cells, FITC-dextran labeled cerebrovascular in the cortex 0 d, 2 d and 21 d after clearing. Thus the clearing technique enables us to repeatedly image microglia and micro-vessels.

Safety assessment of the SOCW in adult mice
The microglia and astrocytes through the SOCW in adult mice remain in non-active state. We found that the microglia in both the SOCW and the control hemisphere remained in non-active state with a highly branched morphology (Supplementary Fig. 6a-b). Supplementary Fig. 6c shows that the GFAP immunostaining patterns exhibit similar levels of GFAP expression for the both sides, which means the astrocytes are not activated. It means that the optical clearing does not induce an inflammation response in adult mice.

Safety assessment of the craniotomy technique
First, the distribution of microglia 0 and 2 d after craniotomy was completely different (Supplementary Fig.   7a). Second, the morphology of microglia (Supplementary Fig. 7b) and the expression of the glial fibrillary acidic protein in astrocytes (Supplementary Fig. 7c) obviously changed. The results show that the craniotomy may induce inflammatory responses. After craniotomy, microglia become active, and their distribution after 2 d is obviously different from that of 0 d. Scale bar, 25 µm. (b) Histological images of microglia under craniotomy. Two days after craniotomy, GFP-labeled microglia appeared abnormal, with many ramified branches projected from somata. Scale bar, 1 mm (above), 25 µm (below). (c) Ten days after craniotomy, immunostaining showed little GFAP expression in astrocytes on the contralateral control side, but extensive GFAP expression in the entire hemisphere of the cortex subjected to open-skull surgery. Scale bar, 1 mm (above), 50 µm (below).