Astrocytes sense glymphatic-level shear stress through the interaction of sphingosine-1-phosphate with Piezo1

Abstract

Astrocyte endfeet enwrap brain vasculature, forming a boundary for perivascular glymphatic flow of fluid and solutes along and across the astrocyte endfeet into the brain parenchyma. We evaluated astrocyte sensitivity to shear stress generated by such flow, finding a set point for downstream calcium signaling that is below about 0.1 dyn/cm². This set point is modulated by albumin levels encountered in cerebrospinal fluid (CSF) under normal conditions and following a blood-brain barrier breach or immune response. The astrocyte mechanosome responsible for the detection of shear stress includes sphingosine-1-phosphate (S1P)-mediated sensitization of the mechanosensor Piezo1. Fluid flow through perivascular channels delimited by vessel wall and astrocyte endfeet thus generates sufficient shear stress to activate astrocytes, thereby potentially controlling vasomotion and parenchymal perfusion. Moreover, S1P receptor signaling establishes a set point for Piezo1 activation that is finely tuned to coincide with CSF albumin levels and to the low shear forces resulting from glymphatic flow.

Keywords: Biological sciences; Biophysics; Cell biology; Neuroscience.

Graphical abstract

Figure 1

Astrocytes are highly sensitive to shear stress (A) Diagram of dual-wavelength Fura-2 imaging of astrocytes in parallel flow chambers. (B) Representative ratiometric fluorescence micrographs of Ca²⁺ response time course in astrocytes at moderate (0.19 dyn/cm²) and high (5.78 dyn/cm²) shear forces. Scale bars, 100 μm. (C) Representative change in [Ca²⁺]_i for a series of shear stresses using 10 s stimuli and (D) 30 s stimuli. Orange and blue arrows indicate when the first shear force was applied. (E) Normalized change in Ca²⁺ concentration amplitude across shear rates with 10 and 30 s stimuli. (F) Percentage of cells responding to a range of shear rates for 10 and 30 s stimuli. (G) Time from stimulus onset to peak Ca²⁺ concentration across a range of shear stresses for 10 and 30 s stimuli. All data shown as mean ± SEM. n ≥ 3 experiments for all treatments.

Figure 2

Astrocyte response to shear stress is dependent on albumin concentration (A) Representative ratiometric fluorescence micrographs of Ca²⁺ response time course in astrocytes at 0 and 90 μM BSA using 2 mL/min flow rate (2.57 dyn/cm²). Scale bars, 100 μm. (B) Normalized change in Ca²⁺ concentration amplitude across shear rates with 10 and 30 s stimuli. (C) Area under normalized Ca²⁺ response curve (AUC) for 30 s stimuli in 0 and 45 μM BSA. Unpaired Mann–Whitney test: ∗∗p < 0.01. (D) Representative traces of normalized change in Ca²⁺ concentration in 0 and 45 μM BSA. Green trace indicates experiment in which cells were alternately exposed to 10 and 30 s stimuli in the presence of 45 μM BSA, then valve was switched to another syringe containing the same BSA concentration, stimuli repeated, valve switched back to original syringe and stimuli repeated. Note that responses to each stimulus (denoted by arrows) are similar when BSA concentration is the same. Gray trace shows experiment in which 0 and 30 s stimuli were repeatedly applied, first in 45 μM BSA, then after switching to 0 μM BSA and then switching back. Note that responses to stimuli (denoted by arrows) in media lacking BSA are severely attenuated and that this is reversed by return to 45 μM. (E) Percentage of cells responding to a range of shear rates for 10 and 30 s stimuli. All data shown as mean ± SEM. n ≥ 3 for all experiments.

Figure 3

S1P recovers astrocyte mechanosensitivity in albumin-free condition (A) Representative ratiometric fluorescence micrographs of Ca²⁺ response time course in astrocytes in 0 μM BSA plus vehicle (methanol) and 0 μM BSA plus S1P across a range of shear stresses. Scale bars, 100 μm. (B) Normalized change in Ca²⁺ concentration amplitude across shear rates with 10 s stimuli. (C) Area under normalized Ca²⁺ response curve for vehicle control and after 1 h incubation prior to and for the duration of flow experiments with S1P conditions at 5.8 dyn/cm². Unpaired Mann–Whitney test: ∗∗∗p < 0.001. (D) Representative traces of normalized change in Ca²⁺ concentration in 0 μM BSA with vehicle or S1P. (E) Percentage of cells responding to a range of shear rates for 10 s stimuli. All data shown as mean ± SEM. n ≥ 3 for all experiments.

Figure 4

S1P receptor antagonist, FTY720, blunts astrocyte response to shear stress (A) Representative ratiometric fluorescence micrographs of Ca²⁺ response time course in astrocytes in 45 μM BSA plus vehicle (DMSO) and 45 μM BSA plus FTY720 in response to 10 s 2.9 dyn/cm² shear. Scale bars, 100 μm. (B) Normalized change in Ca²⁺ concentration amplitude across shear rates with 10 s stimuli. (C) Area under normalized Ca²⁺ response curve for vehicle control and FTY720 conditions at 2.9 dyn/cm². Unpaired Mann–Whitney test: ∗p < 0.05. (D) Representative trace of normalized change in Ca²⁺ concentration in 45 μM BSA with vehicle or FTY720. (E) Percentage of cells responding to a range of shear rates for 10 s stimuli. All data shown as mean ± SEM. n ≥ 3 for all experiments.

Figure 5

PIEZO1 determines the mechanosensitivity of astrocytes (A) Representative experiment showing that GsMTx4 (1 μM), a PIEZO1 antagonist, reversibly inhibits the astrocyte calcium responses to flow. 5.8 dyn/cm² shear flow for 10 s (thin arrows) was applied at 300-s intervals before, after drug additions and after washout. Wash-in and washout of drug (thick arrows) were done at 0.2 mL/min for 2 min. (B) Summary of 3 independent experiments showing the reduction of shear-induced response by 1 μM GsMTX4 (red bar) and partial recovery (gray bar) upon washout. One-way ANOVA followed by Tukey’s post hoc test: ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. (C and D) Representative experiments showing the amplification of astrocyte calcium shear response by Yoda1, a PIEZO1 activator. A descending series of 10-s shear forces was applied in flow medium containing 0 or 45 μM BSA (C and D, respectively), then 10 μM Yoda1 was added and an ascending series of 10-s shear forces was applied. Shear forces (dyn/cm²) are indicated above each response. Note: substantially enhanced responses in the presence of Yoda1. (E) Summary data comparing Yoda1 effect in various BSA concentrations (0, 7.5, and 45 μM) at moderate shear forces. Note: significant enhancement of sensitivity in the presence of Yoda1 at all BSA concentrations. Two-way mixed ANOVA: 45 μM BSA with vs. without Yoda: p value = 0.0024; F(1,13) = 14.15; 7.5 μM BSA with vs. without Yoda: p value = 0.0021; F(1,21) = 12.28; 0 μM BSA with vs. without Yoda: p value = 0.0202; F(4,28) = 3.468. (F) Representative experiment of calcium response of astrocytes incubated in FTY720 for 1.5 h. A descending series of 10-s shear forces was applied in flow medium containing 45 μM BSA, then 10 μM Yoda1 was added (bold black arrow) and an ascending series of shear forces was applied; open arrowhead indicates refocusing. (G) Summary of 3 independent experiments showing the block of flow-induced stress after FTY720 incubation and the recovery of response upon the addition of Yoda1. Unpaired Student’s t test: ∗p < 0.05; ∗∗p < 0.01. n ≥ 3 for all experiments. All data shown as mean ± SEM.

Figure 6

PLC inhibition eliminates astrocyte Ca²⁺ response (A) Representative ratiometric fluorescence micrographs of Ca²⁺ response time course in astrocytes in 45 μM BSA plus PLC inhibitor, U-73122, across a range of shear rates. Scale bars, 100 μm. (B) Representative traces of normalized change in Ca²⁺ concentration showing the blockade of shear response after incubation in U-73122 for 1.5 h. A descending series of 10-s shear forces was applied in flow medium containing 45 μM BSA and U-73122, then 10 μM Yoda1 was added (bold black arrow) and an ascending series of shear forces was applied. (C) Summary of 6 independent experiments showing the block of flow-induced stress after U-73122 incubation and the recovery of response upon the addition of Yoda1. Unpaired Student’s t test: ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. All data shown as mean ± SEM.

Figure 7

Attenuation of responses by heparan sulfate proteoglycan degradation (A) Representative single-slice confocal images of HSPG (green), F-actin cytoskeleton visualized with phalloidin (red) and their merge in WT astrocytes. Cells were cultured in DMEM without (CTRL) and with heparinase III (Heparinase) for 2 h. The nuclei were stained blue with DAPI. Scale bar, 50 μm. (B) Histogram showing HSPG fluorescence intensity quantification expressed as a percentage relative to that of the control (CTRL, black), considered as 100% after 2 h treatment with heparanase III (Hep., red). Data are expressed as mean ± SEM. Four images per experiment, from a total of four experiments, were taken for each condition. Unpaired Mann–Whitney test: ∗p < 0.05. (C) Left, confocal maximum intensity projections of control WT astrocytes immunostained for HSPG (green), F-actin (red), and nuclei (blue). Scale bar, 20 μm. Right, xz cross-sectional views of the stacked confocal images from the right showing the degree of cell surface distribution between HSPG (green), actin cytoskeleton (red), and nuclei (blue). (D) Left, confocal maximum intensity projections of WT astrocytes after 2 h treatment with heparanase immunostained for HSPG (green), F-actin (red), and nuclei (blue). Scale bar, 20 μm. Right, xz cross-sectional views of the stacked confocal images from the right showing the degree of cell surface distribution between HSPG (green), actin cytoskeleton (red), and nuclei (blue). (E) Reconstructed 3D surface plot images of (C) and (D). Color range from black to green indicates the relative level of fluorescence intensity in pixel. (F) Summary of normalized astrocyte calcium responses to a range of shear forces for control and after enzymes degrading heparan sulfate proteoglycan or chondroitin sulfate proteoglycan. n = 3, 4, 8 for control, HSPG degradation, CSPG degradation, respectively. Unpaired Student’s t test: ∗p < 0.05; ∗∗p < 0.01. All data shown as mean ± SEM.