The Ca2+-gated channel TMEM16A amplifies capillary pericyte contraction and reduces cerebral blood flow after ischemia

Abstract

Pericyte-mediated capillary constriction decreases cerebral blood flow in stroke after an occluded artery is unblocked. The determinants of pericyte tone are poorly understood. We show that a small rise in cytoplasmic Ca2+ concentration ([Ca2+]i) in pericytes activated chloride efflux through the Ca2+-gated anion channel TMEM16A, thus depolarizing the cell and opening voltage-gated calcium channels. This mechanism strongly amplified the pericyte [Ca2+]i rise and capillary constriction evoked by contractile agonists and ischemia. In a rodent stroke model, TMEM16A inhibition slowed the ischemia-evoked pericyte [Ca2+]i rise, capillary constriction, and pericyte death; reduced neutrophil stalling; and improved cerebrovascular reperfusion. Genetic analysis implicated altered TMEM16A expression in poor patient recovery from ischemic stroke. Thus, pericyte TMEM16A is a crucial regulator of cerebral capillary function and a potential therapeutic target for stroke and possibly other disorders of impaired microvascular flow, such as Alzheimer's disease and vascular dementia.

Keywords: Calcium; Cell Biology; Ion channels; Neurological disorders; Vascular Biology.

Figure 1. Cortical pericytes express functional TMEM16A channels.

(A) TMEM16A expression in the soma and circumferential processes (yellow arrows) of NG2-labeled pericytes (white arrows) in a representative fixed cortical slice of a P21 rat. Scale bar: 10 μm. (B) TMEM16A expression in a pericyte labeled using isolectin B4 (IB4) in a fixed cortical slice of a 40-year-old human (representative of data from 5 participants). Scale bar: 5 μm. (C) Representative family of whole-cell TMEM16A currents recorded from individual rat cortical pericytes, using pipette solutions designed to isolate Cl^– currents and various free [Ca²⁺]_i (nominally 0, 0.25, or 1.3 μM), in the absence or presence of Ani9 (2 μM). The voltage protocol is illustrated at the top, shown after correction for liquid junction potential. (D) Mean whole-cell TMEM16A current density versus voltage relationships in cortical pericytes (n = 9–14), with various [Ca²⁺]_i, in the absence or presence of Ani9.

Figure 2. Vasoconstricting G_qPCR agonists raise pericyte [Ca²⁺]_i and constrict capillaries at pericyte somata in acute cortical slices.

(A) Representative bright-field images of a live rat cortical capillary pericyte before and after 15 minutes exposure to endothelin 1 (ET-1; 10 nM). Red arrowheads indicate the pericyte soma, and yellow lines indicate where the internal capillary diameter was measured. Scale bar: 10 μm. (B) Mean internal capillary diameter at pericyte somata during exposure to ET-1 (10 nM), normalized to the diameter measured in the absence of ET-1 (n = 10). The ET-1–evoked capillary constriction was not dependent on the sex of rats (Supplemental Figure 3A). (C) Representative bright-field images of a live capillary pericyte, as in A. The thromboxane A2 analog U46619 (200 nM) was applied. Scale bar: 10 μm. (D) Mean internal capillary diameter at pericyte somata during exposure to U46619 (200 nM), normalized to the diameter measured in the absence of U46619 (n = 8). (E) Two-photon microscopy images (maximum intensity projections) of SMCs on a PA and pericytes on first- to third-order capillary branches in acute cortical slices obtained from NG2-Cre^ERT2-GCaMP5G mice. ET-1 raised the somatic [Ca²⁺]_i of SMCs and pericytes (encircled with white dashed lines). Scale bar: 20 μm. The pericyte [Ca²⁺]_i rise coincides with capillary constriction, as indicated by the white line across the vessel lumen in the higher magnification image (scale bar: 10 μm). (F) ET-1 significantly raises [Ca²⁺]_i in first- to third-order pericyte somata and evoked the greatest [Ca²⁺]_i rise in third-order pericytes. The mean GCaMP5G fluorescence in pericyte somata (points indicate individual pericytes from 5 mice; baseline, n = 22; first, n = 13; second, n = 4; third, n = 5) was normalized to the mean GCaMP5G fluorescence of the 17 minutes baseline (F_baseline) with aCSF (1-way ANOVA with Tukey’s post hoc test).

Figure 3. Pericyte contraction evoked by G_qPCR activation requires Ca²⁺ entry via Cav channels.

(A) Two-photon microscopy images (maximum intensity projections) of pericytes on first- to third-order capillary branches from the PA in an acute cortical slice of a NG2-Cre^ERT2-GCaMP5G mouse. Dashed white circles denote pericyte somata. Scale bar: 10 μm. (B) Removing extracellular Ca²⁺ abolished the ET-1–evoked [Ca²⁺]_i rise. Time course of GCaMP5G fluorescence (F) in pericyte somata (red trace; n = 32) and processes (other traces; n = 79) normalized to mean baseline fluorescence with 2 mM [Ca²⁺]_o (the last 10 minutes of 15 minutes in 0 [Ca²⁺]_o are shown). [Ca²⁺]_i changes in processes were quantified less than 5 μm from pericyte somata centers (“at soma”) and at 10 μm and 20 μm along the vessel from the soma center. (C) 15 minutes of 0 [Ca²⁺]_o did not affect pericyte [Ca²⁺]_i (compare bottom 2 plots). Reintroducing Ca²⁺ in the continuous presence of ET-1 raised pericyte [Ca²⁺]_i in processes and somata (see also, B). (D) Capillary constriction at pericyte somata (n = 40) coincides with the [Ca²⁺]_i rise upon 2 mM [Ca²⁺]_o reperfusion in B. There was no significant change in capillary diameter away from pericyte somata at 10 μm (n = 18) or 20 μm (n = 11). (E) Capillary diameter is larger at baseline and constricts in response to 2 mM [Ca²⁺]_o reperfusion at pericyte somata. In D and E, diameter is from tdTomato channel. (F) Time course of ET-1–evoked [Ca²⁺]_i change in pericyte somata (n = 23), normalized to aCSF baseline. Nimodipine (3 μM) (n = 27) or vehicle (n = 17) were applied 15 minutes before ET-1 application. (G) Nimodipine greatly attenuated the ET-1–evoked pericyte [Ca²⁺]_i rise (note that 0 [Ca²⁺]_i is at 1 on the y axis) (Kruskal-Wallis test with Dunn’s post hoc test). The inset shows that in the presence of nimodipine, the ET-1–evoked [Ca²⁺]_i rise was similar in pericytes on first-order (n = 9) versus second- and third-order branches (n = 10) (unpaired 2-tailed Student’s t test).

Figure 4. Effects of Cl^– on TMEM16A-mediated control of pericyte [Ca²⁺]_i and tone.

(A) Capillary diameter at pericyte somata during Ani9 (2 μM) or MONNA (5 μM) superfusion in acute rat cortical slices, normalized to baseline diameter (left). Mean normalized capillary diameter after exposure to Ani9 (n = 13) or MONNA (n = 6) (right). (B and C) Ani9 or MONNA reduced capillary constriction evoked by (B) ET-1 (10 nM) or (C) U46619 (200 nM) (n = 6–8). (D) Ani9 (n = 19) reduced the ET-1–evoked [Ca²⁺]_i rise in pericyte somata (n = 23) in cortical slices of NG2-Cre^ERT2-GCaMP5G mice. GCaMP5G fluorescence (F) was normalized to baseline fluorescence in aCSF. The inset shows that in the presence of Ani9, the ET-1–evoked [Ca²⁺]_i rise was similar in pericytes on first-order (n = 9) versus second- and third-order (n = 10) branches from the PA. (E) Whole-cell currents of rat cortical pericytes in acute rat cortical slices (left). The voltage protocol is illustrated at the top, shown after correction for the liquid junction potential. Free [Ca²⁺]_i was 0 or 0.25 μM. Mean whole-cell current density (I_MAX) at the end of the ramp (i.e., +87 mV) in 0 (n = 8) or 0.25 μM (–Ani9: n = 10; +Ani9: n = 5) [Ca²⁺]_i (right). (F) Normalized capillary diameter at pericyte somata during bumetanide (40 μM) superfusion (left). Mean normalized capillary diameter at pericyte somata after exposure to bumetanide (40 μM) (n = 9) (right). (G) Effects of low [Cl^–]_o, (n = 7) and bumetanide (n = 8) on rat cortical capillary diameter. (H) Changes in capillary diameter in response to ET-1 after 15 minutes of preincubation with bumetanide (n = 7) with (n = 7) or without (n = 8) a lower [Cl^–]_o. (G and H) Representative capillary responses (left). Normalized capillary diameter (right). Numbers of animals are detailed in Supplemental Table 2. (A) Mann-Whitney test; (D and G) unpaired 2-tailed Student’s t test; (B, C, E, and H) 1-way ANOVA with Bonferroni’s post hoc multiple comparisons test.

Figure 5. TMEM16A KO in pericytes reduces endothelin-1–evoked pericyte contraction, and depolarizing the membrane potential induces pericyte contraction independent of TMEM16A block.

(A) Mean internal capillary diameter at pericyte somata during exposure to endothelin-1 (ET-1) (10 nM), normalized to the diameter measured in the absence of ET-1, in acute cortical slices of wild-type or ANO1-KO mice. (B) TMEM16A KO reduced the average ET-1–evoked pericyte contraction during the last 5 minutes of the experiment (Ano KO, n = 27; Wild-type, n = 19). (C) Mean internal capillary diameter at pericyte somata during exposure to 92.5 mM extracellular potassium ([K⁺]_o), normalized to the diameter measured in the presence of 2.5 mM [K⁺]_o in acute rat cortical slices. (D) Raising [K⁺]_o evokes pericyte contraction, and this is not reduced by Ani9 (2 μM). Points indicate individual pericytes from 5 rats per condition (–Ani9: n = 5; +Ani9: n = 6). (B and D) Unpaired 2-tailed Student’s t test.

Figure 6. Blocking TMEM16A slows the ischemia-evoked [Ca²⁺]_i rise in pericytes, delays capillary constriction, and reduces pericyte death in acute cortical slices.

(A) Mean normalized capillary diameter at pericyte somata during perfusion with aCSF (control, n = 6) or oxygen and glucose deprivation (OGD) solution in the presence (n = 6) or absence (n = 10) of Ani9 (2 μM) in acute rat cortical slices. (B) Ani9 reduces the ischemia-evoked pericyte-mediated capillary constriction at 30 minutes of OGD. Points indicate individual pericytes from 5 to 8 rats per condition (1-way ANOVA with Bonferroni’s post hoc multiple comparisons test). The OGD-evoked capillary constriction was not dependent on the sex of rats (Supplemental Figure 3B). (C) Time course of the mean change in normalized GCaMP5G fluorescence (F) in pericyte somata during OGD with or without Ani9 (2 μM) in NG2-Cre^ERT2-GCaMP5G mice. (D) Ani9 reduces the [Ca²⁺]_i rise (measured from the y axis value of 1) at 16 minutes of perfusion with OGD. Points indicate individual pericytes (OGD, n = 14; OGD+Ani9, n = 17) from 3 mice per condition (unpaired 2-tailed Student’s t test with Welch’s correction). (E) Confocal images of rat cortical capillaries labeled with isolectin B4 (IB4) to visualize pericytes labeled by the necrosis marker propidium iodide (PI) after a 1-hour exposure to aCSF or OGD in the presence or absence of Ani9 (2 μM). The inset illustrates examples of necrotic (PI +) and healthy (PI –) pericytes. Scale bar: 30 μm. (F) Ani9 reduces the OGD-evoked pericyte death. The percentage of dead pericytes was quantified by dividing the number of PI-labeled pericytes by the total number of pericytes in images as shown in (E) (aCSF: n = 32; OGD: n = 33; OGD+Ani9: n = 35) (Kruskal-Wallis test with Dunn’s post hoc test). Number of animals are detailed in Supplemental Table 2.

Figure 7. Blocking TMEM16A reduces pericyte contraction and improves CBF after CCAO.

(A) In vivo barrel cortex pericytes in P36–P40 NG2-Cre^ERT2-GCaMP5G mice before and after approximately 7.5 minutes of CCAO, following a 1-hour exposure to Ani9 (10 μM) or vehicle. Yellow triangles indicate sites of diameter changes. Scale bar: 10 μm. (B) Mean GCaMP5G fluorescence (F) in pericyte somata normalized to baseline prior to sham operation or CCAO. Data are measured at 70–90 minutes of reperfusion (sham, n = 11; CCAO, n = 9; CCAO+Ani9, n = 10) (1-way ANOVA, Tukey’s post hoc test). (C) Time course of normalized CBF from laser Doppler during sham operation or CCAO (black bar) without (aCSF, n = 19) or with Ani9 (10 μM, n = 17). (D) Normalized CBF at 70–90 minutes of reperfusion. Points are from individual animals (baseline CCAO, n = 19; sham, n = 8; CCAO, n = 19; Ani9, n = 17) (paired 2-tailed Wilcoxon’s test with continuity correction for “baseline” and “during CCAO aCSF” conditions; 1-way ANOVA with Dunnett’s post hoc test for other conditions). (E) In vivo barrel cortex of NG2-dsRed mouse with FITC-dextran in blood. Capillary branching orders are given from the PA or AV. Red and blue tracing indicate vessels on arteriole or venule side, respectively; white boxes denote occlusion sites; yellow arrows represent pericytes. Scale bar: 20 μm. (F) Mean baseline capillary diameter versus distance from pericyte soma for all branch orders, during CCAO and at 70–90 of minutes of reperfusion. Capillary branch orders are first to third order from arteriole (aCSF, n = 14; Ani9, n = 13); fourth to sixth order from arteriole or venule (aCSF, n = 4; Ani9, n = 7); or first to third order from venule (aCSF, n = 13; Ani9, n = 11). P values compare slope of the linear regression line with 0. (G) Mean first-order capillary diameter from PAs at 70–90 minutes of reperfusion (aCSF, n = 18; Ani9, 10 μM, n = 14). Points indicate individual capillary segments at 0–5 μm from pericyte somata (Mann-Whitney test). (H) As for G, for pericytes on first to third capillary branch order from AVs (aCSF, n = 32; Ani9, n = 33) (unpaired 2-tailed Student’s t test).

Figure 8. Blocking TMEM16A partially restores capillary perfusion after CCAO.

(A) Isolectin B4–labeled cortical capillaries in fixed slices (purple); FITC-albumin in gelatin (recolored red) shows perfused vessels 1.5 hours after sham operation or CCAO without or with Ani9 (10 μM). Yellow arrows indicate pericyte somata less than 5 μm from capillary blocks. Right images show 3D tracing of FITC-albumin in perfused (white) or unperfused (magenta) capillaries. Scale bar: 20 μm. (B) Cumulative probability distribution of distance of 110 pericyte somata to capillary occlusions (black). The red line denotes the predicted distribution assuming pericytes are uniformly spaced along the capillary (see Supplemental Figure 5E) and blocks occur randomly (2-sample Kolmogorov-Smirnov test). (C) Extent of perfusion of 3D-traced capillaries. Points indicate individual confocal stacks from P30–P72 CCAO (aCSF) (n = 25), P31–P83 CCAO Ani9 (n = 16), and P39–P62 sham aCSF (n = 9) mice (Kruskal-Wallis test with Dunn’s post hoc test). (D) Ly6G-labeled neutrophil in a third order capillary in a fixed cortical slice at 1.5 hours after CCAO. The neutrophil obstructs blood flow (revealed by FITC-gelatin staining). Zoomed-in image of the right end of the block and left end of neutrophil (“Block start”) shows a possible red blood cell (RBC) to the left of the neutrophil. Scale bar: 10 μm. (E) Distribution of 109 neutrophils versus distance from the nearest pericyte soma. (F) Neutrophils in cerebral capillaries per confocal stack at 1.5 hours after CCAO in the absence (n = 18) or presence (n = 33) of Ani9 (10 μM) (Mann-Whitney test). (G) CD41-labeled platelets (or aggregates thereof) in third and fourth order capillary branches in a fixed cortical slice at 1.5 hours after CCAO. (H) Platelets (or platelet aggregates) in cerebral capillaries per confocal stack at 1.5 hours after CCAO in the absence (n = 71) or presence (n = 60) of Ani9 (10 μM) (Mann-Whitney test). The numbers of animals are specified in Supplemental Table 2.

Figure 9. Blocking TMEM16A improves CBF and reduces neuronal hypoxia and infarct size in aged mice.

(A) Normalized CBF from laser Doppler during CCAO (black bar) in the absence (aCSF, n = 7) or presence of 10 μM Ani9 (n = 7). (B) Normalized CBF during CCAO or reperfusion (average of last 5 minutes of traces in A) in 15 month-old mice (n = 7 for each, each point is 1 Doppler recording) (paired 2-tailed Wilcoxon’s test with continuity correction and Mann-Whitney test). (C) TTC-stained brain sections at 6 hours of reperfusion after CCAO in aged mice. (D) Infarction quantified from mean intensity of TTC-stained sections (Mann-Whitney test) (vehicle, n = 8; Ani9, n = 6). Confocal images of fixed cortical (E) and striatal (F) slices from aged mice that were injected with pimonidazole (Hypoxyprobe) in vivo at 70 minutes after CCAO. Mice underwent 6 hours of reperfusion. Bar graphs indicate Hypoxyprobe intensities in the cortex (E) and striatum (F) from mice treated with aCSF (cortex, n = 40; striatum, n = 13) or Ani9 (cortex, n = 30; striatum, n = 9) (Mann-Whitney test and unpaired 2-tailed Student’s t test). (G) Confocal image of layers II and III in a fixed cortical slice from a mouse undergoing 6 hours of reperfusion after CCAO. (H) Proportion of cells with a glial or neuronal morphology labeled with Hypoxyprobe in the cortex quantified from images, as in H. (I) Cortical NeuN fluorescence intensity at 6 hours reperfusion (aCSF, n = 16; Ani9, n = 12) (unpaired 2-tailed Student’s t test). (J) Schematic of mechanisms revealed. G_qPCR activation triggers the IP₃ pathway; the resulting [Ca²⁺]_i rise stimulates TMEM16A, cell depolarization, and Cav-mediated Ca²⁺ entry. In ischemia, low ATP slows Ca²⁺ pumping, leading to TMEM16A activation, pericyte contraction, and death. Neutrophils and platelets become trapped as pericytes contract and capillaries narrow, further lowering CBF. TMEM16A inhibition enhances capillary reflow and reduces tissue damage. Animal numbers are provided in Supplemental Table 2. Scale bar: 50 μm.