2.2.1.

Sulforhodamine B binding to albumin

Dye binding to plasma proteins (principally albumin) prevents BBB crossing. To measure the unbound dye fraction in the presence of albumin, we prepared an aqueous solution of albumin ( $30\mathrm{g}∕\mathrm{l}$ , Molekula, Germany) in $0.1\text{-}\mathrm{M}$ phosphate buffer (pH 7.4). SRB (Radiant Dye, Germany) was added to the albumin solution (final concentration: $0.2\mathrm{g}∕\mathrm{l}$ ). The solution was dialyzed for $24\mathrm{h}$ at $37°\mathrm{C}$ using cellulose dialysis tubing (cutoff: $1000\mathrm{Da},$ Fisher Scientific AG, Germany). Finally, the SRB concentration was measured in the dialysate using a spectrofluorophotometer (lambda 9, Perkin-Elmer, Waltham, Massachusetts) and the unbound plasma dye concentration was deduced. Control measurements were performed on a SRB solution without albumin.

2.2.2.

Sulforhodamine B clearance from the mouse vascular compartment

To study the SRB clearance from the mouse vascular compartment, mice (Swiss/IOPS CD1, $38\text{to}43\mathrm{g}$ , Charles River Laboratories, France) were anesthetized using a gas mixture of isoflurane/air (5% for induction, 2% for maintenance). The catheter was inserted into the jugular vein and $100\mu \mathrm{l}$ of a SRB solution ( $5\mathrm{mg}.{\mathrm{ml}}^{-1}$ in saline buffer) was injected. At different times for $3\mathrm{h}$ after injection (see Fig 1 ), a blood sample of $40\mu \mathrm{l}$ was collected. The blood sample was diluted 40 times and homogenized in heparinized physiological saline. One milliliter of every sample was centrifuged and absorbance of the supernatant was measured at $560\mathrm{nm}$

Fig. 1

SRB clearance from the mouse vascular compartment. Evolution of SRB concentration in the mouse vascular compartment after injection of $100\mu \mathrm{L}$ of a SRB solution $(5\mathrm{g}∕\mathrm{L})$ at time 0. Square and round dots are associated with two different mice whose SRB clearances are very similar.

2.3.1.

Primary astrocytic cultures

Astrocytic cultures were performed as described by Raponi ²⁶ Hemispheres from whole forebrains of P2 newborns mice (Charles River Laboratories, France) were separated and dissociated by trypsinization and mechanical trituration in Hanks balanced saline solution. Microglial cells were removed by preplating the cell suspension on noncoated petri dishes for $15\min$ . The resulting cell suspension was then replated on noncoated Falcon petri dishes in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). After $4\text{days}$ , oligodendrocyte progenitor cells and neuroblasts growing on top of the developing astroglial layer were selectively removed by flushing.²⁷ The percentage of nonastrocytic cells never exceeded 2.5% of the cells.

2.3.2.

Primary neuron cultures

Primary cerebellar granular neuron cultures were prepared from $6\text{-}\mathrm{day}$ -old S/IOPS NMRI mice (Charles River Laboratories, France) as described by Trioulier ²⁸ Cerebella were removed and cut before a $10\text{-}\min$ incubation at $37°\mathrm{C}$ in 0.25% trypsin EDTA in DMEM. The cerebella were dissociated by trituration, and dissociated cells were centrifuged for $5\min$ at $500\mathrm{g}$ . Cells were plated on poly-D-lysine precoated dishes and incubated at $37°\mathrm{C}$ in $\mathrm{K}25+\mathrm{S}$ medium. One day later, cytosine- $\beta$ -D-arabinoside ( $10\mu \mathrm{M}$ , Sigma, France) was added in the medium to prevent the growth of non-neuronal cells.

2.3.3.

In vitro studies

$5\mu \mathrm{l}$ of a SRB solution $(10\mathrm{mg}∕\mathrm{ml})$ was added to the $5\text{-}\mathrm{ml}$ culture medium as previously described. Observations were made by two-photon microscopy (see next). A $60\times$ water-immersion objective (0.90 NA, LUM Plan Fl/ IR Olympus, Japan) was directly immersed in the medium, and acquisition was performed from $3\min \text{to}2\mathrm{h}$ after addition of the SRB on each petri dish. The temperature was maintained at $37°\mathrm{C}$ . Cell nuclei were stained by adding $40\mu \mathrm{l}$ of a Hoechst 33342 solution ( $1\mathrm{mg}∕\mathrm{ml}$ –Sigma, France). Observations on six different petri dishes with astrocytes and three with neurons are presented. The SRB dye in the extracellular compartment diffuses inside the whole intracellular compartment of the astrocytes, thus a distinction of the cell shapes from the background is impossible. Therefore, on a different petri dish of the same cell culture, the cell membranes were stained with the lipophilic tracer $1,1^{\prime }$ -dioctadecyl- $3,3,3^{\prime },3^{\prime }$ -tetramethylindodicarbocyanine, $4^{\prime }$ -chlorobenzenesulfonate (DiD, Invitrogen France).

2.4.1.

Microscopy setup

Two-photon laser scanning microscopy was performed with a confocal microscope consisting of a Biorad MRC 1024 scanhead and an Olympus BX50WI microscope. Fluorescence was directly epicollected. An $800\text{-}\mathrm{nm}$ excitation beam from a Tsunami femtosecond Ti:sapphire laser ( $5\text{-}\mathrm{W}$ pump, Millenia V, Spectra-Physics, Mountain View, California) was focused in the sample using a $20\times$ water-immersion objective (0.95 NA, Xlum Plan Fl Olympus, Japan).

The beam was scanned in the $x\text{-}y$ plane to acquire $512\times 512\text{pixel}$ images in $0.9\mathrm{s}$ . The z-scan for variation of the observation depth was realized by vertical motion of the motorized objective. The incident laser intensity was adjusted by using a half-wave plate and a polarizer placed before the microscope, so that the total average power delivered at the surface ranged from $1\text{to}150\mathrm{mW}$ . Two channels were simultaneously observed using two added external photomultiplicator tubes (PMT) and an appropriate filter set made up of a HQ620/60 Red filter for SRB and a HQ535/30 Green filter for FITC-dextran (staining vasculature) associated with a 585 DCRX Dichroic filter, all from Chroma Technology (Rockingham, Vermont).

Images were acquired using the Biorad exploitation system and displayed using ImageJ (ImageJ., v.1.33, Public Domain Software, available on http://rsb.info.nih.gov/ij/, 2005).

2.4.2.

Animal preparation

Mice were anaesthetized with a gazeous inhalation of a mixture of isoflurane/air (2%). For dye injections, craniotomy realization, and two-photon imaging, mice were placed on a stereotaxic frame modified to allow a longitudinal rotation of the animal to optimize laser penetration. Core temperature was maintained at approximately $36°\mathrm{C}$ using warm water circulating through a pad.

2.4.3.

In vivo two-photon tridimensional imaging

The images of the SRB distribution were acquired at successive depths in the cortex (down to $500\mu \mathrm{m}$ below the dura) from $2\mathrm{h}30\text{to}5\mathrm{h}$ ( $\mathrm{n}=3$ Swiss nude mice, Charles River Laboratories, France) after injection of the SRB solution ( $50\mu \mathrm{l}$ , $10\mathrm{mg}.{\mathrm{ml}}^{-1}$ in saline buffer) through the tail vein. Fifteen minutes before two-photon imaging, a $3\text{-}\mathrm{mm}$ -diam craniotomy was carried out above the left parietal cortex; the bone was removed and the exposed cerebral cortex was filled with a 0.9% saline water solution. Approximately $10\min$ after surgery, a $70\text{-}\mathrm{kDa}$ FITC-Dextran solution ( $100\mu \mathrm{l}$ , $100\mathrm{mg}.{\mathrm{ml}}^{-1}$ in saline buffer, Sigma, France) was injected via the tail vein to visualize simultaneously microvessels and SRB staining.

3.1.1.

Sulforhodamine B clearance from the mouse vascular compartment

The results are presented in Fig. 1. The SRB-concentration decrease has been fitted with the sum of two exponential functions. The two characteristic times were $t_1=3.6\pm 0.4\min$ and $t_2=29.9\pm 7.1\min$ . The initial SRB concentration in the blood circulation measured approximately $30\mathrm{s}$ after injection was $0.3\mathrm{g}∕\mathrm{l}$ .

3.1.2.

Sulforhodamine B binding to plasma proteins

After $24\mathrm{h}$ of dialysis of the solution with SRB and albumin, a dialyzable fraction of SRB of 0.6 was measured. This indicates that the majority of the dye in the blood was free. Surprisingly, the control solution (SRB in $\text{water}+\text{PBS}$ ) showed a smaller dialyzable fraction. This could be explained by the tendency of SRB molecules to aggregate and form dimers in an aqueous solution.²⁹ So it appears that the albumin may efficiently prevent the dimer formation.

3.3.1.

Sulforhodamine B crosses the blood brain barrier

From $2\mathrm{h}30$ after injection of the SRB solution, a specific staining of cortical cells was observed (see Fig. 3 ). In these in vivo studies, any uptake of SRB due to a traumatic increase of BBB permeability after craniotomy cannot be excluded. The SRB, however, was injected more than two hours before the surgery, when the SRB blood concentration is inferior to $0.02\mathrm{g}∕\mathrm{l}$ (see Fig. 1) and would contribute little or nothing to the astrocytes staining. This demonstrates that the SRB crosses most probably the healthy BBB. SRB labeled cells are detected until $5\mathrm{h}$ after injection.

Fig. 3

Images acquired in the left parietal cortex of a nude mouse at $2\mathrm{h}30\min$ after injection of $50\mu \mathrm{l}$ of a SRB solution $(10\mathrm{mg}.{\mathrm{ml}}^{-1})$ . Fifteen minutes before two-photon imaging, a $2\text{-}\mathrm{mm}$ -diam craniotomy was performed and a $70\text{-}\mathrm{kDa}$ FITC-Dextran solution ( $100\mu \mathrm{l}$ , $100\mathrm{mg}.{\mathrm{ml}}^{-1}$ , Sigma) was injected intravenously. (a) Z projection of the maximum intensity of a stack of images obtained between 80 and $120\mu \mathrm{m}$ below the dura. (b) A slice obtained at $85\mu \mathrm{m}$ below the dura. Cells stained by the SRB (in red) present characteristic features of astrocytes: star-shaped with many extensions and end-feet encircling capillaries. Note thinner red spots coming from astrocyte processes outside the focal plane. The capillaries appeared in yellow (merge of red and green colors).

3.3.2.

Sulforhodamine B stains specific mouse cortical cells

All cells stained by SRB had common and specific morphological features, such as starshapes with many extensions and end-feet encircling capillaries (see Fig. 3). These features are very similar to those described for neocortex astrocytes in Refs 19, 30. The volume density of SRB stained cells calculated on a volume of $299\times 299\times 40\mu {\mathrm{m}}^3$ between 80 and $120\mu \mathrm{m}$ under the dura is ${11.10}^3\text{cells}∕{\mathrm{mm}}^3$ .