SorCS2 Is Important for Astrocytic Function in Neurovascular Signaling

Abstract

Introduction: The receptor SorCS2 is involved in the trafficking of membrane receptors and transporters. It has been implicated in brain disorders and has previously been reported to be indispensable for ionotropic glutamatergic neurotransmission in the hippocampus.

Aim: We aimed to study the role of SorCS2 in the control of astrocyte-neuron communication, critical for neurovascular coupling.

Methods: Brain slices from P8 and 2-month-old wild-type and SorCS2 knockout (Sorcs2^-/-) mice were immunostained for SorCS2, GFAP, AQP4, IB4, and CD31. Neurovascular coupling was assessed in vivo using laser speckle contrast imaging and ex vivo in live brain slices loaded with calcium-sensitive dye. Bulk and cell surface fraction proteomics was analyzed on freshly isolated and cultured astrocytes, respectively, and validated with Western blot and qPCR.

Results: SorCS2 was strongly expressed in astrocytes, primarily in their endfeet, of P8 mice; however, it was sparsely represented in 2-month-old mice. Sorcs2^-/- mice demonstrated reduced neurovascular coupling associated with a reduced astrocytic calcium response to neuronal excitation. No differences in vascularization or endothelium-dependent relaxation ex vivo between the 2-month-old groups were observed. Proteomics suggested changes in glutamatergic signaling and suppressed calcium signaling in Sorcs2^-/- brains from both P8 and 2-month-old mice. The increased abundance of glutamate metabotropic receptor 3 in Sorcs2^-/- astrocytes was validated by PCR and Western blot. In cultured Sorcs2^-/- astrocytes, AQP4 abundance was increased in the bulk lysate but reduced in the cell surface fraction, suggesting impaired trafficking.

Conclusion: The results suggest that SorCS2 expression is important for the development of neurovascular coupling, at least in part by modulating glutamatergic and calcium signaling in astrocytes.

Keywords: SorCS2; astrocytes; calcium signaling; glutamate signaling; neurovascular coupling.

FIGURE 1

SorCS2 is localized to astrocyte endfeet of newborn mouse brain. In 8‐day‐old wild‐type mice (P8), SorCS2 staining overlapped with the astrocytic marker, glial fibrillary acidic protein (GFAP), and surrounded vascular endothelial cells labeled with α‐d‐galactosyl‐specific isolectin B4 (IB4), as indicated (A). SorCS2‐specific staining was not detected in the brain of P8 Sorcs2 ^−/− mice, while there was no difference in GFAP and IB4 labeling in Sorcs2 ^−/− and wild‐type brains (bars correspond to 20 μm). Representative images for 7 slices from Sorcs2 ^−/− mice and for 9 slices from wild‐type mice; 2–4 cortical fields of view per slice. The intensive aquaporin‐4 (AQP4) and SorCS2 co‐labeling suggest that Sorcs2 is localized in the endfeet processes of astrocytes surrounding IB4 positive vascular endothelial cells in the brain of P8 wild‐type mice (B; bars correspond to 50 μm). Merged image suggests an overlap of AQP4 and SorCS2 intensities in astrocytic endfeet in brain slices from P8 wild‐type mice, as indicated by arrows in the magnified image (the bar indicates 10 μm). Two to four cortical fields of view are taken per slice (n = 9). The fluorescence intensity profiles over the cross‐section, as indicated with a dashed rectangle (B), illustrate an overlay of AQP4 and SorCS2 labeling that is outside of the blood vessel lumen labeled with IB4 (C; the traces representative for P8 brains of 9 wild‐type mice).

FIGURE 2

SorCS2 is sparsely detected in astrocytic soma in 2‐month‐old mouse brain. Brain slices were stained for SorCS2 together with glial fibrillary acidic protein (GFAP) to label astrocytes, and vascular endothelial cell maker α‐d‐galactosyl‐specific isolectin B4 (IB4), as indicated. In 2‐month‐old wild‐type mice, sparse GFAP‐SorCS2 co‐labeling was detected as indicated by white arrows. No SorCS2 signal was detected in the brain slices from 2‐month‐old Sorcs2 ^−/− mice. Representative images for 2–4 fields of view per slice of 4 slices from Sorcs2 ^−/− and 4 slices from wild‐type mice. Bars correspond to 50 μm. See also Figure S1.

FIGURE 3

2‐month‐old Sorcs2 ^−/− mice exhibited reduced neurovascular responses in vivo. An averaged blood flow map (grayscale) with an overlaid relative blood flow index (BFI) changes (color scale) in response to mechanical whisker stimulation, which induced a local change in blood perfusion of the primary somatosensory cortex in age‐matched wild‐type (A) and Sorcs2 ^−/− mice under ketamine and xylazine anesthesia (B). Bars correspond to 1000 μm. A relative increase in parenchymal BFI of the somatosensory cortex in response to contralateral whisker stimulation was reduced in Sorcs2 ^−/− mice (n = 9) compared to wild‐type mice (n = 8; C). Responses in (C) were compared using two‐way ANOVA. *p < 0.05. See also Figure S4.

FIGURE 4

2‐month‐old Sorcs2^−/− mice showed a reduced increase in vascular diameter and blood flow in response to sensory stimulation. Laser speckle contrast imaging was used to assess changes in arterial diameter and arterial blood flow increase in response to whisker stimulation in age‐matched 2‐month‐old wild‐type (WT) (A) and Sorcs2^−/− mice (B). The color scale shows the blood flow index (BFI). Bars correspond to 1000 μm. The red rectangle identifies the 3rd branch of the middle cerebral artery used for single vessel analysis. Representative single artery segmentation profiles show changes in diameter and blood flow in response to whisker stimulation, as indicated on the X‐axis (C). Color code corresponds to BFI. The Y‐axis indicates a cross‐section coordinate with respect to the center of the vessel, i.e., the inner diameter can be identified. There was no difference in baseline diameter of the 3rd branch of the middle cerebral artery between 2‐month‐old Sorcs2^−/− and WT mice (D). Averaged peak responses to whisker stimulation revealed that Sorcs2^−/− mice showed a smaller increase in diameter (E) and blood flow (F) in the 3rd branch of the middle cerebral artery compared to WT. *p < 0.05 and **p < 0.01 (unpaired t‐test). n = 7–9.

FIGURE 5

Astrocytes showed a reduced increase in intracellular Ca²⁺ in response to neuronal excitation of the brain slices from 2‐month‐old Sorcs2 ^−/− mice in comparison with age‐matched wild types. Brain slices were loaded with Ca²⁺‐sensitive dye, Calcium Green‐1/AM. The dye was preferentially allocated to astrocytes and, to a lesser extent, to neuronal tissue (red arrows). White arrows indicate astrocytic endfeet surrounding a parenchymal arteriole (A). Three seconds of electric field stimulation (EFS) increased intracellular Ca²⁺ in astrocytes and neuronal tissue, as shown in representative images from baseline recording and 6 s after initiation of EFS (A). The increase in astrocytic endfeet Ca²⁺ was larger in the brains of 2‐month‐old wild‐type mice (n = 5) than in the brain slices from age‐matched Sorcs2 ^−/− mice (n = 4), while no difference in the neuronal Ca²⁺ response was seen (B). **p < 0.01 (two‐way ANOVA with Sidak's multiple comparisons test).

FIGURE 6

Endothelium‐dependent vasorelaxation of the middle cerebral artery was unchanged in 2‐month‐old Sorcs2 ^−/− mice. The diameter of isolated middle cerebral arteries from age‐matched wild‐type (WT) and Sorcs2 ^−/− mice was not different (A). There was no difference in the vasorelaxation in response to 10⁻⁵ M carbachol between wild‐type and Sorcs2 ^−/− arteries (B). n = 8–11.

FIGURE 7

Volcano plot of differentially abundant proteins in astrocytes from wild‐type and Sorcs2 ^−/− mice. These plots depict fold changes in protein abundance in the lysate of astrocytes from Sorcs2 ^−/− versus wild‐type mice correlated to the probability that the protein is differentially abundant. Comparisons were made for bulk astrocyte lysates from P8 mice (A) and 2‐month‐old mice (B). p < 0.05 was set as the threshold for differential abundance and depicted on the graphs with a horizontal dotted line. Dots in red denote significantly upregulated proteins, and dots in blue denote significantly downregulated proteins. Gray dots indicate proteins that were not significantly different. Comparisons were made without correction for the False Discovery Rate, although mGluR3 level remains significantly increased after correction. See Tables S1 and S2 for more detailed information. Venn diagram (C) shows the number of proteins, whose abundance was changed in only one of the age groups, i.e., in P8 or 2‐month‐old mice, and those that were changed in both ages. n = 4. Volcano plot of differentially abundant proteins (D) in cell surface fractions of primary Sorcs2 ^−/− astrocytes. Comparisons were made without correction for False Discovery Rate. See Table S4 for detailed information; n = 4 for (A, B) and n = 6 for (D).

FIGURE 8

The relative mRNA expression profile of astrocytes from P8 and 2‐month‐old wild‐type and Sorcs2 ^−/− mice. (A) P8 Sorcs2 ^−/− mice (n = 5) have, in comparison with age‐matched wild types (n = 5), diminished expression of Sorcs2. Sorcs2 expression in astrocytes from 2‐month‐old wild types (n = 6) was suppressed in comparison with P8 wild types. (B) Astrocytes from Sorcs2 ^−/− mice showed an increased expression of glutamate metabotropic receptor 3 (mGluR3) mRNA in comparison with age‐matched controls at both P8 and 2‐month ages. Moreover, mGluR3 expression was lower in astrocytes from 2‐month‐old than in P8 Sorcs2 ^−/− mice. (C) mRNA for glial fibrillary acidic protein (Gfap) was increased in astrocytes from 2‐month‐old Sorcs2 ^−/− mice in comparison with age‐matched wild types and P8 Sorcs2 ^−/− astrocytes. (D) mRNA for aquaporin‐4 (Aqp4) was increased in 2‐month‐old wild‐type mice in comparison with P8 mice. (E) Glycerol‐3‐phosphate dehydrogenase 2 (Gpd2) mRNA expression had only a tendency for increased expression in P8 Sorcs2 ^−/− mice in comparison with age‐matched wild type. *p < 0.05, **p < 0.01 and ****p < 0.0001.

FIGURE 9

Loss of SorCS2 altered astrocytic cell surface expression of glutamate metabotropic receptors 2 and 3 (mGluR2/3) and aquaporin 4 (AQP4). Primary astrocytes (wild type, WT, and Sorcs2 ^−/− , KO) were subjected to a cell surface protein biotinylation assay, and the biotinylated proteins were captured using streptavidin beads prior to Western blot analysis. See also Figure S7 (A) Representative results of Western blot bands of mGluR2/3 receptor (the primary antibody detects both mGluR2 and mGluR3, multimers > 200 kDa)¹⁰ and AQP4 levels in total astrocytic lysates and in cell surface fraction. E‐cadherin serves as a loading control and is enriched in the cell surface fraction, while tubulin is depleted from this fraction. SorCS2 is detected in the WT samples. (B) Quantification of the results in (A). Signal intensities for mGlur2/3 and AQP4 were normalized to E‐cadherin signals. Graphs show mean ± SEM. n = 5–6 biological replicates (independent primary cultures). *p < 0.05 in two‐tailed t‐test for independent samples.