Specific in vivo staining of astrocytes in the whole brain after intravenous injection of sulforhodamine dyes

Abstract

Fluorescent staining of astrocytes without damaging or interfering with normal brain functions is essential for intravital microscopy studies. Current methods involved either transgenic mice or local intracerebral injection of sulforhodamine 101. Transgenic rat models rarely exist, and in mice, a backcross with GFAP transgenic mice may be difficult. Local injections of fluorescent dyes are invasive. Here, we propose a non-invasive, specific and ubiquitous method to stain astrocytes in vivo. This method is based on iv injection of sulforhodamine dyes and is applicable on rats and mice from postnatal age to adulthood. The astrocytes staining obtained after iv injection was maintained for nearly half a day and showed no adverse reaction on astrocytic calcium signals or electroencephalographic recordings in vivo. The high contrast of the staining facilitates the image processing and allows to quantify 3D morphological parameters of the astrocytes and to characterize their network. Our method may become a reference for in vivo staining of the whole astrocytes population in animal models of neurological disorders.

Figure 1. Astrocytes are progressively stained in vivo after an intravenous injection of sulforhodamine B.

A) 10 min after an iv injection of SRB (P18 rat) only blood vessels were observed, 40 min after the injection both vessels and astrocytes were stained and 90 min after the injection only astrocytes could be imaged. Scale bar = 50 µm. B) Curve of the fluorescence intensity changes in both astrocyte and blood compartments as a function of time. C) In vivo two-photon imaging of cortical astrocytes at a 250 µm depth (P21 rat) 30 minutes after SRB (20 mg/kg) iv injection. Arrows show non-stained regions corresponding to neuron soma and the arrowhead corresponds to an astrocytic end-foot linked to a blood vessel. Scale bar = 50 µm.

Figure 2. Sulforhodamine stained cells are only astrocytes.

A) Two-photon excitation spectra of SRB and SR101 shows the possibility to excite one dye without exciting the other. Z-projection (standard-deviation) of a 100 µm stack acquired on an acute coronal brain slice of rat somatosensory cortex (P23) after iv injection of SRB/SR101 mix (1∶1) and FITC-dextran to stain vasculature. The left merged image was acquired with an 800 nm laser excitation wavelength, corresponding to SRB (orange) and FITC-dextran (green) emissions. The central merged image was acquired with a 900 nm laser excitation wavelength, corresponding to SR101 (red) and FITC-dextran (green) emissions. The right panel shows a merge of SRB, SR101 and FITC-dextran (pink corresponds to SRB and SR101 colocalization). Scale bar = 50 µm. B) Z-projection (standard-deviation) of two-photon microscopy images of cortical brain slices from GFAP-GFP transgenic mouse (P37) 2 h after iv injection of SRB. Left image shows SRB staining (red), central image shows GFAP-GFP expression (green) and right panel is a merge of left and central images with double-stained cells appearing in yellow. C) Z-projection (standard-deviation) of TPLSM images of cortical brain slices from a P18 rat. Left image: SRB astrocytes staining after iv injection (red), central image: slice immunostained with S100B antibody (green). Right panel is a merge of left and central images with double-stained cells appearing in yellow. D) Merge images of SRB-stained slices (red) secondarily immunostained with NG2 antibody (green, left panel), or NeuN antibody (green, central panel), or CD11b antibody (green, right panel). Arrowheads show NG2 cells or microglial cells which are not SRB-stained. Scale bar = 20 µm.

Figure 3. Astrocytes and blood vessels staining in acute cortical brain slices (P23 rat).

A) Astrocytes are SRB labeled (red) whereas the vessels are stained by FITC-dextran (green). Scale bar = 50 µm. B) Higher magnification of image A showing that most of the surface of the vessels was covered by either astrocytic endfeet (1) or astrocytic cell bodies (2) or pericytes (3). Scale bar = 20 µm.

Figure 4. Comparison of two methods for astrocytes staining in acute brain slices (thickness of 300 µm, P18 rat).

TPLSM images: A–D) brain slicing performed 2 h after intravenous injection of SR101 (20 mg/kg) and E–F) brain slices incubated 15 minutes with SR101 (final concentration 1 µM) in aCSF. Z-projections (standard-deviations) of acquired stacks: A) in the cortical L1 and L2/3 (stack thickness 300 µm); B) in the dentate gyrus (stack thickness 150 µm); C) in the substantia nigra pars reticulata (stack thickness 100 µm) and D) in the cerebellum (stack thickness 50 µm). Images collected after incubation with SR101: E) at the surface of the slice, and F) 80 µm deeper from the brain slice surface. Scale bar = 50 µm.

Figure 5. Comparison of two in vivo methods for astrocytes staining (P21 rat).

TPLSM images acquired: A_1–3) 30 minutes after SRB (20 mg/kg) intravenous injection and B_1–3) 5 minutes after SR101 application (100 µM) on the cortical surface. (A₁) and (B₁) Images were taken at 100 µm below the pia.mater (A₂) and (B₂) Images were taken at 250 µm below the surface of the cortex. (A₃) and (B₃) show a 3D reconstruction (V3D) using the entire stack of images. Scale bar = 50 µm.

Figure 6. Calcium signaling in neocortical astrocytes stained with SR101.

A) Left panel: merged confocal images of astrocytes labeled with SR101 (iv injection; red) and incubated with 5 µM Fluo-4 AM (green) in acute brain slice (P17 rat). Right panel: example of typical fluorescence variations measured in three SR101-stained astrocytes within the somatosensory cortex. B) Left panel: merged confocal imaging of astrocytes stained with both SR101 (red) and Fluo-4 AM (green). Right panel: example of fluorescence increase induced by ATP (5 µM) measured in three typical SR101 positive cells. C) Left panel: confocal imaging of astrocytes loaded with Fluo-4 AM (green). Right panel: example of fluorescence increase induced by ATP (5 µM) measured in three loaded astrocytes. D) Left panel: TPLSM imaging of cortical astrocytes in vivo 1 hour after an iv injection of SR101 (P18 rat, depth = 200 µm). Right panel: example of fluorescence variations measured in two SR101-stained astrocytes within the somatosensory cortex layer 2/3 labeled with Fluo-4 AM. Scale bar = 50 µm.

Figure 7. 3D morphological analysis of the astrocytic network after iv injection of SRB.

A) A multistacks mosaic acquired in the somatosensory area of a coronal acute brain slice (P17 rat). The region of interest (ROI; white rectangle) shows colored astrocytes detected using ImageJ plugins. Each color corresponds to single astrocytes cell body. Scale bar = 150 µm. B) Frequency histogram showing of astrocyte densities as a function of cortical depth calculated from the ROI described above. Depth was divided in 15 bins with 100 µm increments. C) Frequency histogram of normalized astrocytes radial densities in the cortical layer 1 and E) in layers 2/3. D) TPLSM images showing astrocytes in the cortical layer 1 and F) in layers 2/3. Scale bar = 25 µm.

Figure 8. Rearrangement of the astroglial network in a mouse model of mesiotemporal lobe epilepsy.

A–B) Bright field microscopy imaging (Nikon Multizoom AZ100, France) of acute hippocampal slices from adult mice 2 weeks after a unilateral intrahippocampal kainate injection. A) Contralateral non-injected hippocampus. B) Ipsilateral hippocampus. Note the absence of CA1/CA3 areas and the enlargement of the dentate gyrus. C–D) TPLSM imaging of regions indicated by white squares on (A) and (B) images, respectively, after iv injection of SR101. Scale bar = 50 µm. E) Column scatters representation showing the distribution of astrocytes cell body volumes at both sides of the hippocampus. The bar corresponds to the mean value.