Astrocyte regulation of cerebral blood flow during hypoglycemia

Abstract

Hypoglycemia triggers increases in cerebral blood flow (CBF), augmenting glucose supply to the brain. We have tested whether astrocytes, which can regulate vessel tone, contribute to this CBF increase. We hypothesized that hypoglycemia-induced adenosine signaling acts to increase astrocyte Ca²⁺ activity, which then causes the release of prostaglandins (PGs) and epoxyeicosatrienoic acids (EETs), leading to the dilation of brain arterioles and blood flow increases. We used an awake mouse model to investigate the effects of insulin-induced hypoglycemia on arterioles and astrocytes in the somatosensory cortex. During insulin-induced hypoglycemia, penetrating arterioles dilated and astrocyte Ca²⁺ signaling increased when blood glucose dropped below a threshold of ∼50 mg/dL. Application of the A_2A adenosine receptor antagonist ZM-241385 eliminated hypoglycemia-evoked astrocyte Ca²⁺ increases and reduced arteriole dilations by 44% (p < 0.05). SC-560 and miconazole, which block the production of the astrocyte vasodilators PGs and EETs respectively, reduced arteriole dilations in response to hypoglycemia by 89% (p < 0.001) and 76% (p < 0.001). Hypoglycemia-induced arteriole dilations were decreased by 65% (p < 0.001) in IP3R2 knockout mice, which have reduced astrocyte Ca²⁺ signaling compared to wild-type. These results support the hypothesis that astrocytes contribute to hypoglycemia-induced increases in CBF by releasing vasodilators in a Ca²⁺-dependent manner.

Keywords: Hypoglycemia; IP3R2 KO; adenosine; astrocytes; calcium signaling; cerebral blood flow.

Figure 1.

Measurement of arteriole diameter and astrocyte Ca²⁺ activity during hypoglycemia. (a) Two-photon microscopy imaging of the awake mouse cortex. The head-fixed mouse rests on a treadmill. Drawing created with BioRender.com. (b) Experimental timeline; insulin or insulin vehicle is injected at time = 0. Blood glucose collection points are shown above and interleaved two-photon imaging sessions below the timeline. Blood glucose was sampled at 15 min intervals at the beginning of an experiment and more frequently as the experiment progressed. (c) Example two-photon image of a penetrating arteriole labeled with Texas Red dextran (red) and astrocyte Ca²⁺ activity indicated by membrane-tethered GCaMP6f (green). Scale bar, 20 µm. (d) Results from one experiment showing arteriole diameter (red) and Ca²⁺ activity (green) increasing over the course of the experiment as blood glucose levels (black) decrease. Diameter and Ca²⁺ activity results are normalized to the average euglycemic value.

Figure 2.

Hypoglycemia increases arteriole diameter and astrocyte Ca²⁺ activity. (a) Example images of the dilation of a penetrating arteriole during euglycemia (left) and during moderate hypoglycemia (right). White lines denote vessel diameter during hypoglycemia. Scale bar, 20 µm. (b) Change in arteriole diameter as blood glucose decreases. Arteriole diameter is normalized to the diameter prior to insulin administration. Light and dark shaded regions indicate mild and moderate hypoglycemia ranges, respectively. (c) Summary showing changes in arteriole diameter during euglycemia (green; >70 mg/dL), mild hypoglycemia (blue; 50–70 mg/dL) and moderate hypoglycemia (red; 30–50 mg/dL); n = 30, 18, and 19 trials, respectively. (d) Example images showing Ca²⁺ events in a single image frame during euglycemia (left) and moderate hypoglycemia (right) as determined by AQuA. Scale bar, 20 µm. (e and f) Calcium activityContinued.as blood glucose decreases, measured in astrocyte endfeet surrounding an arteriole (e) and in astrocyte processes (excluding endfeet; f). Calcium activity was normalized to pre-insulin activity levels and was measured as the summed “area under the curve” (see Methods section). (g) Summary showing Ca²⁺ activity in astrocyte endfeet (triangles) and processes (circles) during euglycemia (green), mild hypoglycemia (blue) and moderate hypoglycemia (red); n = 26, 27, and 15 trials (endfeet) and 15, 16, and 16 trials (processes), respectively. (h and i) Correlation between Ca²⁺ activity and arteriole diameter in endfeet (h) and processes (i). r = 0.22 and 0.23, respectively. Summary data shown are mean ± SD; *p < 0.05, **p < 0.01; two-way ANOVA.

Figure 3.

Insulin administration does not induce increases in arteriole diameter and astrocyte Ca²⁺ activity when co-administered with dextrose to maintain euglycemia in mice that were initially euglycemic. (a) Change in vessel diameter 0 to 30 minutes and 30 to 60 minutes after administration of insulin and dextrose. (b) Change in astrocyte Ca²⁺ activity in endfeet (triangles) and processes (circles) 0 to 30 minutes and 30 to 60 minutes after administration of insulin and dextrose. For all categories, n = 8. Data normalized to pre-insulin + dextrose measurements. T-test used to compare difference from the null condition of 1. Data are mean ± SD; *p < 0.05.

Figure 4.

Astrocyte Ca²⁺ signaling is reduced in IP3R2 KO mice in both endfeet (triangles) and processes (circles). Astrocyte Ca²⁺ activity, as measured by the area under the curve of Ca²⁺ events (a), the total number of Ca²⁺ events (b), the average area of Ca²⁺ events (c), and the duration of Ca²⁺ events (d). The area under the curve of Ca²⁺ events was reduced in processes of IP3R2 KO mice and the size and number of events was reduced in both endfeet and processes of IP3R2 KO mice. For (a–d), n = 38, 44, 32, and 32 for WT endfeet, WT processes, IP3R2 KO endfeet and IP3R2 KO processes, respectively. One-way ANOVA used to compare WT to IP3R2 KO mice. Seven outlier values were removed from (a), 6 from (b), 13 from (c), and zero from (d). Data are mean ± SD; ***p < 0.001.

Figure 5.

Arteriole diameter and astrocyte Ca²⁺ activity are altered during hypoglycemia in IP3R2 KO mice and with pharmacological inhibitors. (a) Arteriole diameter change from euglycemia in control mice (n = 15, 18 for mild and moderate hypoglycemia), IP3R2 KO mice (n = 25, 28), and mice given SC-560 (n = 7, 11), miconazole (n = 9, 8), combined SC-560 and miconazole (n = 6, 8), and ZM-241385 (n = 9, 10). (b) Astrocyte Ca²⁺ activity (endfoot and process, combined) for control mice (n = 14, 16) and mice given ZM-241385 (n = 9, 9). Data are normalized to euglycemia data from the same session. For statistical significance, data are compared to control mice of the same glycemic status using one-way ANOVAs. Data are mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 6.

Inhibition of PGs and EETs synthesis does not constrict arterioles. Arteriole diameter before and after administration of SC-560 to inhibit PG synthesis (n = 5), miconazole to inhibit EETs synthesis (n = 4), and administration of combined SC-560 and miconazole (n = 4). Individual trials (gray circles) and means (black asterisks) are shown. Student’s t-test comparing pre-drug to post-drug arteriole diameter. No results were significantly different.

Figure 7.

Proposed mechanism of an astrocyte contribution to hypoglycemia-induced CBF increases. Hypoglycemia-induced increases in brain adenosine (ADO) activate adenosine receptors on astrocytes, including A_2A receptors (A_2AR), leading to the release of Ca²⁺ from the endoplasmic reticulum (ER) through IP3 type 2 receptors (IP3R2). The resulting increase in astrocyte Ca²⁺ leads to the synthesis and release of prostaglandins (PGs) via cyclooxygenase-1 (COX-1) and epoxyeicosatrienoic acids (EETs) via P450 epoxygenase (P450), leading to the dilation of brain arterioles and to increases in cerebral blood flow (CBF). Astrocyte Ca²⁺ increases may also result in the release of adenosine, either by direct release of adenosine or via release of ATP, which is rapidly converted to adenosine by ectonucleotidases. In addition to stimulating astrocytes, hypoglycemia-evoked adenosine can directly activate receptors on arterioles to cause vasodilation.