Astrocytic endfoot Ca 2+ correlates with parenchymal vessel responses during 4-AP induced epilepsy: An in vivo two-photon lifetime microscopy study

Abstract

Neurovascular coupling (NVC) underlying the local increase in blood flow during neural activity forms the basis of functional brain imaging and is altered in epilepsy. Because astrocytic calcium (Ca²⁺) signaling is involved in NVC, this study investigates the role of this pathway in epilepsy. Here, we exploit 4-AP induced epileptic events to show that absolute Ca²⁺ concentration in cortical astrocyte endfeet in vivo correlates with the diameter of precapillary arterioles during neural activity. We simultaneously monitored free Ca²⁺ concentration in astrocytic endfeet with the Ca²⁺-sensitive indicator OGB-1 and diameter of adjacent arterioles in the somatosensory cortex of adult mice by two-photon fluorescence lifetime measurements following 4-AP injection. Our results reveal that, regardless of the mechanism by which astrocytic endfoot Ca²⁺ was elevated during epileptic events, increases in Ca²⁺ associated with vasodilation for each individual ictal event in the focus. In the remote area, increases in Ca²⁺ correlated with vasoconstriction at the onset of seizure and vasodilation during the later part of the seizure. Furthermore, a slow increase in absolute Ca²⁺ with time following multiple seizures was observed, which in turn, correlated with a trend of arteriolar constriction both at the epileptic focus and remote areas.

Introduction

The brain has high-energy demand and requires a constant and continuous supply of oxygen and glucose for normal function. To ensure that blood supply matches metabolic needs, the brain possesses a major control mechanism, namely neurovascular coupling. Functional hyperemia^1–4 is defined as a local increase in cerebral blood flow in response to neuronal activity. Part of this vascular regulation, via the synaptic activation of astrocytes, remains to be clearly defined. Astrocytes are a subtype of glial cells, and a significant part of them are in close proximity of cerebral blood vessels with endfeet processes almost completely enveloping cerebral blood vessels. The interplay between the astrocytic endfoot and the cerebral vasculature is an area of intense investigation. Recent evidence suggests that neuronal activity is also encoded by astrocytes in the form of dynamic intracellular calcium (Ca²⁺) signals, which travel to astrocytic endfoot encasing the arterioles in the brain. Astrocytic Ca²⁺ signaling has been implicated in the dilatory response of adjacent arterioles, linking neuronal activity to enhanced local blood flow.^5–10

Epilepsy is a common neurological disease characterized by recurrent unprovoked seizures, which result from abnormal and excessive neuronal activity in the brain. At a cellular level, these events reflect intense and highly synchronous discharges that involve large numbers of cortical neurons. ¹¹ In both animal models and patients, the epileptic discharges can evoke drastic increases in cerebral blood flow (CBF) to meet the high metabolic demand caused by this intense neuronal activity.^12–15 However, observations of the hemodynamic response during seizures have also displayed nonlinear phenomena, ¹⁶ with the first nonlinear term having an inhibitory contribution to the expected linear response. Mechanisms underlying these observations are to be identified. In brain slices, using uncaged Ca²⁺ in astrocytic endfeet, it was shown that a high increase in the absolute concentration of Ca²⁺ leads to vasoconstriction. ⁶ Whether this phenomena contribute to the nonlinear inhibitory effects observed in epilepsy is an open question. The relationship between absolute Ca²⁺ concentrations in astrocytic endfeet and the vascular response during ictal events is unknown, and it is unclear whether or not changes in endfeet Ca²⁺ can account for the full spectrum of vascular responses to neuronal activity in epilepsy.

In contrast to the common intensity-based measurements, fluorescence lifetime imaging (FLIM) techniques that use specific indicators to monitor nanometer-scale molecular interactions in live cells have been emerging.^17,18 It was also demonstrated that the fluorescence lifetime of some commonly used Ca²⁺ sensitive dyes, such as Oregon Green 488 BAPTA-1 (OGB-1), is sensitive to free Ca²⁺ in the physiological nanomolar range.^19,20 Compared to ratiometric methods to evaluate absolute concentration, the lifetime technique is immune to absorption and fluorescence bleaching effects making it more suitable for in vivo imaging. This property of OGB-1 has led to the successful evaluation of Ca²⁺ changes in astroglia of normal and Alzheimer’s disease mice models.^18,21 In this work, we designed a two-photon FLIM system that enables imaging of deep brain tissue in live animals with single cell spatial resolution. We adapted the FLIM technique to investigate the role of astrocytes in the response of cerebral blood vessels to epileptiform discharges. By using a 4-aminopyridine (4-AP) model of focal seizures in vivo, we found that ictal, seizure-like discharges were rapidly followed by large Ca²⁺ increases in astrocyte endfeet. This concentration increase was accompanied by nearby arteriolar dilation, but there was a negative relationship between the baseline [Ca²⁺]_i prior to seizures and the amplitude of vasodilation.

Materials and methods

Animal preparation

Animals were used according to the ARRIVE guidelines and the recommendations of the Canadian Council on Animal Care. The Animal Research Ethics Committee of the Montreal Heart Institute approved all procedures. Fourteen male C57/BL6 mice (Charles-River, postnatal 8 weeks old, 20–25 g weight), of which three died before data acquisitions due to movement of the catheter in the femoral artery, were deeply anesthetized with 1–1.6 g/kg urethane and body temperature was maintained at 37℃ with a controlled physiological monitoring system that also monitored heart and respiration rates continuously (Labeotech, Inc., CA). Mice breathed via a tracheal tube to reduce the risk of respiratory depression often seen with the use of this anesthetic. A moderate flow of ambient air lightly supplemented with oxygen was supplied next to the tracheotomy (10% oxygen, 90% air, 1 L/min). Animals were placed in a stereotaxic frame. A ∼ 2 mm × 2 mm cranial window was opened over one hemisphere to expose the somatosensory cortex and surrounding brain (AP: −1.5 mm, DV: + 1.5 mm). A small hole was drilled next to the cranial window for the injection of 4-AP (Figure 1(c)). After injection of the Ca²⁺ indicators, the cranial window was sealed with 1% agarose in artificial cerebrospinal fluid (aCSF, 125 mM NaCl, 10 mM HEPES, 10 mM glucose, 5 mM KCl, 1,5 mM CaCl₂, 1 mM MgSO₄) using a 150 µm-thick microscope coverslip. During the experiment, a catheter in the femoral artery was used to monitor the blood gases (pCO₂, 36–39 mmHg, and pO₂, 110–160 mmHg). The average blood pressure (80–110 mmHg) was obtained noninvasively by a tail-cuff blood pressure system (Kent Scientific). At the end of the surgery, 500 µL of saline was injected subcutaneously to avoid animal dehydration during imaging. Surgery was started in the morning and imaging sessions debuted around noon. Prior to procedures, mice were kept in a 12:12 h light-dark cycle in ventilated cages.

Figure 1.

(a) Schematic of the two-photon lifetime microscopy system. Excitation light is provided by a laser oscillator (MaiTai-BB) through an acousto-optic modulator (AOM) followed by a polarizer (P) to adjust the gain. A telescope (L₁ and L₂) expands the galvo-mirrors image onto the microscope objective pupil for illumination. Emitted fluorescence is separated using a first dichroic mirror (DM₁). The return beams are then split by a second dichroic mirror (DM₂) sending the signal to detectors centered at wavelengths of 520 nm (F₁) and 593 nm (F₂). The Ca²⁺ concentration was monitored in the 520 nm channel using a photon-counter for FLIM imaging. (b) Top: In vivo fluorescence staining of neurons in green, astrocytes in yellow and vasculature in red. Bottom: OGB-1 stained neurons and astrocytes (left panel) while SR101 stained astrocytes alone and Rhodamine B localized the vasculature (right panel). (c) Pictogram of measurement areas on the mouse brain. The craniotomy was done on the left side. The 4-AP injection location and electrode recordings were done at the same site. The remote area was defined to be that further than 1.5 mm from the focus. (d) Calibration of the fluorescence decay of OGB-1 at 10 different buffer Ca²⁺ concentrations. In this range, fit lifetime varied from ∼4.86 ns to ∼0.65 ns for high/low concentrations, respectively. (e) Typical images at 593 nm for longitudinal vessel scan used to measure the diameter. The right figure shows measurements of absolute [Ca²⁺]_i in one astrocytic endfoot during resting state.

Ca²⁺ indicator loading

Multi-cell bolus loading was performed to load neurons and glial cells with the Ca²⁺ sensitive fluorescence indicator OGB-1 and the astrocyte specific fluorescence marker, Sulforhodamine 101 (SR101). A patch pipette with a tip diameter of 30–60 µm was inserted into the cortex to a depth of ∼ 300 µm from the surface. OGB-1 (50 µg, O-6807, Molecular probes-Invitrogen, CA, USA) was dissolved in dimethyl sulfoxide (DMSO) containing 20% pluronic acid (F-127, Sigma–Aldrich) and mixed in 412 µM SR101 (Sigma–Aldrich) to a final concentration of 1 mM.^22,23 One µL OGB-1 and SR101 were injected with a micropipette using a microsyringe pump controller (UMP3, World Precision Instruments, Sarasota, FL) at a rate of 100 µL/min. After injection, we allowed 1 h for loading. To label vasculature, Rhodamine B isothiocyanate-Dextran (200 µL of 100 mg/mL solution, molecular weight ∼ 70,000 Da, Sigma-Aldrich), which highlights blood plasma, was injected into the tail vein. Serial images from the pial surface to cortex layers 2/3 (∼200 µm deep) showed the OGB-1 signals localized in neurons and astrocytes while SR101 stained astrocytes (Figure 1(b)).

Epileptogenesis and electrophysiology

Ictal discharges were induced by injecting the potassium channel blocker 4-aminopyridine (4-AP; Sigma; 15 mM, 0.5 µL) through a glass microelectrode using a syringe pump controller into a small hole next to the cranial window ¹² (Figure 1(c)), similar to the injection of mixed OGB-1 and SR101 described above. Extracellular local field potentials (LFP) were recorded with a tungsten electrode (impedance, 0.5–2 MΩ), and lowered to a depth of ∼300 µm into the neocortex. The signal was filtered by a band-pass filter between 1 and 5000 Hz, amplified 1000 times with a microelectrode AC amplifier (model 1800, A-M system, Sequim, WA), and digital filtered between 0.2 and 130 Hz.^24,25

Two-photon fluorescence lifetime setup

Measurements were collected using a custom-built 2-photon laser scanning fluorescence microscope (Figure 1(a)) with 80 MHz, 150 fs pulses from a MaiTai-BB laser oscillator (Newport Corporation) with a maximum output of ∼2W through an acousto-optic modulator (ConOptics) to adjust the gain for depth-dependent two-photon excitation intensity. The laser pulses were scanned in a raster pattern by galvanometric mirrors. Reflected light was collected by a water-immersion objective (20 × , 1.0 NA; Olympus), and then separated into two beams by dichroic mirrors. The beams were split and filtered around center wavelengths of 520 and 593 nm, and measured on two distinct photomultiplier tubes (H10682-210 for photon counting at 520 nm, R3896 for CW measures at 593 nm, Hamamatsu photonics, Japan). To monitor Ca²⁺ changes, the excitation was set to 800 nm and the laser power reduced to limit the count rate to a maximum of 5% of the total number of laser pulses to remain in single-photon counting regime (<70 mW). Emission counts at 520 nm were recorded using a photon counter (PicoHarp 300, PicoQuant). Scanning and data recordings were controlled by custom-designed software written in Matlab (MathWorks, USA).

Experiments were performed in two steps. Focus was first achieved on the cortex, and then the objective was moved to a depth of ∼200 µm. A point on the endfoot of interest and a perpendicular line on the encased vasculature by the endfoot were selected for measuring Ca²⁺ concentration ([Ca²⁺]_i) and diameter of the encased vessel. The field of view was 50 µm × 50 µm and the pixel density was 0.125 µm during the measurement. A custom scanning sequence, rapidly alternating between each type of measure was designed (gating the photon counter to only count while the beam was sitting on the endfoot) to have a simultaneous assessment of Ca²⁺ concentration and diameter (Figure 1(e)). Multiplexed Ca²⁺ measurements and line scan (200 lines) at a given location had a temporal resolution of 1 s.

Fluorescence lifetime calibration

OGB-1 was used to monitor dynamic changes in intracellular Ca²⁺, which can be measured through changes in a lifetime. For calibration, we used the standard calibration method provided by the Invitrogen Ca²⁺ calibration buffer kit manual (Invitrogen, Thermo-Fisher, Missisauga, ON, Canada). Absolute [Ca²⁺]_i was determined independently of variations in dye concentration by measuring bound and unbound Ca²⁺ decay curves with fluorescence lifetime microscopy, which exhibit different lifetimes. The lifetime decay curves for OGB-1 in each of 10 Ca²⁺ buffers were measured using samples of free dye in glass capillary tubes at varying Ca²⁺ concentrations. ²⁶ In the absence of Ca²⁺, OGB-1 had a single, fast decay leading to an effective lifetime of ∼0.65 ns accounting for the PMT response function. At saturating levels of Ca²⁺ (1.35 µM), OGB-1 had a single, slow decay leading to a lifetime of ∼4.86 ns. While decays are bi-exponential within this range, we adopted an effective strategy of fitting a single decay curve on the first 80% of the decay, which was found to be robust at lower counts that in turn allowed faster in vivo recordings. Using this technique, effective calibration curves were generated for each buffer lifetime (Figure 1(d)) that were found to be reproducible from calibration to calibration (see error bars) and also when decreasing the total number of counts.

Data analysis

All analyses were performed with Matlab using in-house code. The relationship between the fluorescence lifetime and [Ca²⁺]_i was obtained from the calibration (Figure 1(d), right). The resulting time decay curves obtained by the photon counter were fit with a single-exponential function from the maximum counts to 80% counts. The [Ca²⁺]_i was then calculated by fluorescence lifetime at each second using calibration results.

Due to the injected fluorescent dye (Rhodamine B), the plasma appears bright in the images while red blood cells appear as dark shadows (Figure 1(e) left). Imaging plasma through successive line scans over the same region is the principle for measuring diameter. ²⁷ The vessel diameter was defined by fitting to a Gaussian function whose full-width at half-maximum from the perpendicular scans was used as the diameter (which may underestimate the real diameter). ²⁸ Small pre-capillary arterioles can capillaries were targeted by size and by choosing low branching order while identifying a surrounding endfoot process. A total of N = 34 vessels were studied with a mean diameter of D = 6.68 ± 0.44 µm. The diameter data were converted to percent change by subtracting and then dividing the median value of the scan.

Results

Using two photon fluorescence lifetime measurements, [Ca²⁺]_i was measured in the somatosensory cortex astrocytes of anesthetized mice during normoxia. Resting Ca²⁺ concentration in glial cells was spatially heterogeneous; resting Ca²⁺ concentration in somatic regions was significantly higher than in endfoot regions from 14 astrocytes in seven mice (4-AP was injected after basal measurements) (Figure 2, paired t-test p < 0.001) with mean values of 90.15 ± 1.62 nM in the somata and 79.44 ± 1.63 nM in endfeet which is in agreement with the sensitivity range of the dye and values previously obtained by in vivo two-photon fluorescence imaging^18,21 but slightly lower than ex vivo preparations.^6,29 We also quantified fluctuations of concentration over time, and we observed no significant differences in fluctuations between somata and endfeet. The fluctuations estimations provided guidelines for statistical evaluation of the changes measured below.

Figure 2.

Paired t-test of [Ca²⁺]_i in astrocytic somata and endfeet during resting state. (a) Boxplot of the average of [Ca²⁺]_i in somata and endfeet in basal state. The [Ca²⁺]_i in somata was significantly larger than endfeet (p < 0.001). (b) Boxplot of the standard deviation (SD) of [Ca²⁺]_i in somata and endfeet during baseline over time. There was no difference in SD of [Ca²⁺]_i between the somata and endfeet.

Diameter changes in the epileptic focus and remote areas during seizure-like activity

Seizure-like activity was elicited with an injection of 4-AP and recorded by local field potentials with a tungsten electrode. They were characterized by first rhythmic spiking of increasing amplitude and decreasing frequency, evolving into rhythmic spikes and slow wave activity prior to gradual offset (see e.g., Figure 3(a) top, and Figure 3(b) top).

Figure 3.

Simultaneous measurements of absolute Ca²⁺ in astrocytic endfoot and diameter of parenchymal vessels during epileptic events in local (a) and remote (b) areas (raw data with 0.1 Hz filtered signals). The time course of LFP indicates seizure initiation (top of (a) and (b) panels). (a) Nearby the injection site, simultaneous measures of Ca²⁺ and diameter show that Ca²⁺ and diameter display a monophasic increase with ictal discharge. (b) In remote areas, parenchymal vessels constricted at the onset of ictal event, then dilated while Ca²⁺ remains elevated throughout seizures. (c) Raw line scan images of a vessel whose diameter changes during seizure (for seizure shown in (a)).

We simultaneously measured [Ca²⁺]_i in astrocytic endfeet and diameter of parenchymal vessels in the somatosensory cortex of mice during epileptic events (Figure 3) to study how absolute [Ca²⁺]_i in astrocytic endfeet correlates with the diameter of parenchymal vessels in local (<1.5 mm) and remote areas from 4-AP injection site (>1.5 mm) ²⁴ (Figure 1(c)). A typical time-course of changes in astrocytic endfoot Ca²⁺ and diameter at the epileptic focus (Figure 3(a)) and remote areas (Figure 3(b)) is shown in Figure 3. During epileptic events, Ca²⁺ significantly increased, returning to the baseline after seizure both at the epileptic focus and remote areas. At distances over 1.5 mm from the site of injection, we observed early constriction of the parenchymal vessel followed by delayed dilation (Figure 3(b)) during seizure. These results were in agreement with previous works where similar vascular responses was observed with a two-photon microscope.^13,14

Astrocytic endfoot baseline Ca²⁺ determines the level of arterial response during seizures

To investigate the relationship between astrocytic endfoot Ca²⁺ and parenchymal vessel diameter changes at different distances from the epileptic focus, seven mice were recorded with measurements close to the injection site and four mice were measured far away from the injection site (>1.5 mm) (Figure 1(c)). Seizure-like activities evoked an increase in neuronal activity and a widespread increase in astrocytic Ca²⁺, and were also associated with vasodilation. Figure 4(a) shows representative data from one mouse comparing the relative changes in observed seizures and indeed shows a linear relationship between relative increase in diameter and relative increase in Ca²⁺ (rCa²⁺) for each seizure. We then analyzed the data by calculating the relative diameter and rCa²⁺ changes as a function of absolute baseline concentration [Ca²⁺]_i (calculated between seizures) over seven mice (90 seizures, Figure 4(b) and (c)). With increasing baseline [Ca²⁺]_i, the relative diameter and rCa²⁺ changes decreased in amplitude either suggesting a constrictive modulation associated with increasing baseline [Ca²⁺]_i or the fact that with increasing baseline [Ca²⁺]_i, the vessels partially dilate decreasing remaining reserve to dilate further.

Figure 4.

The relationship between [Ca²⁺]_i in astrocytic endfoot and the parenchymal vessel changes in diameter during epileptic events ((a), (b) and (c)) as well as the statistical analysis of all mice (seven mice) in the focus. (a) The relationship between the relative [Ca²⁺]_i (rCa²⁺) and relative changes in diameter (rDiameter) during epileptic events at the focus over eight seizures from one animal. Note that during ictal events, [Ca²⁺]_i increases in the encasing astrocytic endfoot are accompanied by parenchymal vessel dilations. (b) Correlation between the absolute baseline [Ca²⁺]_i and relative changes in diameter at the focus (from seven mice, 90 seizures). (c) Relationship between the relative [Ca²⁺]_i and absolute [Ca²⁺]_i at the focus (from seven mice, 90 seizures). (d) Paired t- test of the average and SD of [Ca²⁺]_i values during baseline and epileptic seizures for all mice. The average [Ca²⁺]_i during ictal discharges was higher than baseline (p < 0.001). The SD of [Ca²⁺]_i did not significantly vary between the seizure and basal level. (e) Paired t-test of the average and SD of diameter in all mice. The mean diameter had a significant increase with ictal events (p < 0.001), and there were no significant differences in the SD of diameter between seizures and baseline. (f) Bar plots of the slopes of linear fits (as shown in Figure 5(a) and (b)) in 21 astrocytes of seven mice over 90 seizures. Slopes were significantly negative (both for basal (p = 0.05) and for seizures (p = 0.006), one sample t-test).

To validate whether this observation reflected a difference in the resting state tone due to elevated Ca²⁺, which would lead to a reduced reserve for dilation, we analyzed the absolute diameter during seizures across the recording session for the same animal. By plotting diameter during seizures against the absolute [Ca²⁺]_i (Figure 5(a), two sessions, diameter normalized to its median), we observed a negative trend between single parenchymal vessel diameter and [Ca²⁺]_i: at higher Ca²⁺ concentrations during seizures, the corresponding diameter decreased contradicting the reserve hypothesis. We also performed a correlation between the absolute Ca²⁺ concentration in all astrocytic endfeet and the median diameter over the time course of the experiments at baseline and observed no correlation (Figure 5(c) correlation coefficient: R = 0.09). Furthermore, we observed that in remote areas, the relative constriction with respect to baseline diameter also increased with increasing [Ca²⁺]_i (despite a decrease in absolute diameter).

Figure 5.

Scatter plot illustrating the relationship between absolute [Ca²⁺]_i in one astrocytic endfoot and the parenchymal vessel diameter. (a) Correlation between the absolute [Ca²⁺]_i and diameter during epileptic events in local area (over two recording sessions, eight seizures from one animal). Linear fit is: y = −0.004x + 1.70, R²= 0.13. (b) Correlation between the absolute [Ca²⁺]_i and relative diameter change during epileptic seizures in remote areas (over two sessions, eight seizures, from one animal). The linear fit for undershoot is: y = −0.004x + 1.241, R²= 0.22 while for the overshoot: y = −0.005x + 1.676, R²= 0.35. (c) The relationship between the absolute [Ca²⁺]_i and median of diameter over all measurements at baseline. The Spearman correlation coefficient is R = 0.09.

Consolidating data from all mice with measures at the focus (seven mice, 21 astrocytes, 90 seizures), the average of [Ca²⁺]_i in astrocytic endfeet during each epileptic seizure was significantly larger than baseline level (Figure 4(d), p < 0.005). As expected, the increase in parenchymal vessel diameter was also significant during ictal discharges (Figure 4(e), p < 0.001). However, when analyzing relative diameter versus basal absolute [Ca²⁺]_i over the time course of the session, we observed a significant negative association between absolute [Ca²⁺]_i and the relative diameter as quantified by their slope (Figure 4(f), one sample t-test of slope against zero), in all cases preserving the negative relationship between the absolute [Ca²⁺]_i and relative diameter or [Ca²⁺]_i seen in Figure 4(b) and (c).

Astrocytic endfoot baseline Ca²⁺ determines the level of arteriole constriction first then of dilation with seizures in remote areas

To study the relationship between astrocytic endfoot [Ca²⁺]_i and parenchymal vessel diameter changes during epileptic events in remote areas, measurements at a distance over 1.5 mm from the injection site were done in four mice. Seizure-like activities evoked an increase in astrocytic Ca²⁺ that was associated with a constriction at the beginning of the seizure followed by dilation (representative data from one mouse shown in Figure 3(b)). We analyzed the data in a similar way than at the epileptic focus: first by calculating the changes of relative diameter and the relative [Ca²⁺]_i (Figure 6(a)). Figure 6(a) shows that relative changes during seizures indeed show a biphasic vascular response (constriction followed by a vasodilation) while [Ca²⁺]_i remain elevated at each seizure. However, when the full data (Figure 6(b)) are analyzed using absolute Ca²⁺ basal concentration, we observed a positive trend between relative parenchymal vessel constriction and Ca²⁺ increase, and a negative trend between relative parenchymal vessel dilation and Ca²⁺ increase. These results indicate that the higher the absolute [Ca²⁺]_i, the weaker is the dilation or the stronger is the constriction for a given seizure. We also calculated the relationship between the relative [Ca²⁺]_i and absolute [Ca²⁺]_i over four mice (58 seizures). Figure 6(c) shows that the relative [Ca²⁺]_i decreases as the absolute [Ca²⁺]_i increases similarly to the focus. To characterize the relationship between the size of the undershoot and overshoot in diameter during epileptic events, Figure 7(a) plots that the negative relationship of relative diameter changes between undershoot and overshoot.

Figure 6.

The relationship between [Ca²⁺]_i in the astrocytic endfeet and parenchymal vessel diameter changes during epileptic seizures ((a), (b), and (c) as well as statistical analysis over all mice (four mice) in remote areas. (a) The relationship between the relative [Ca²⁺]_i and relative changes in diameter during epileptic seizures in a remote area over two recording sessions (eight seizures) from one mouse. For each seizure, [Ca²⁺]_i increased in the endfoot with parenchymal vessel constricting at the beginning of seizures followed by dilation. (b) Correlation between the absolute [Ca²⁺]_i and relative diameter changes in remote areas (from four mice, 58 seizures). (c) The relationship between the relative [Ca²⁺]_i during seizures and absolute [Ca²⁺]_i in remote areas (from four mice, 58 seizures). (d) Paired t-test of the average and SD of absolute [Ca²⁺]_i values during undershoot or overshoot and baseline for all mice. The average of [Ca²⁺]_i during the beginning of seizure and the duration of seizure was higher than baseline (p = 0.025 and p = 0.028). The SD of [Ca²⁺]_i did not significantly vary between the undershoot or overshoot and basal level. (e) Paired t-test of the average and SD of the percent change of diameter in all mice. The mean diameter had a significant decrease during the beginning of seizure (p = 0.001) and the average of diameter during the duration of seizure was higher than the baseline (p = 0.038). There were no significant differences in the SD of diameter between undershoot or overshoot and baseline. (f) Bar plots of slopes of linear fits between relative diameter changes and absolute [Ca²⁺]_i (as shown in Figure 5(b)) in 13 astrocytes of four mice over 58 seizures. Slopes were negative (For basal, p = 0.014, for seizure, p = 0.009, one sample t-test).

Figure 7.

Relationship between changes of relative diameter during undershoot and overshoot in remote area. (a) Scatter plot illustrating the correlation of relative diameter changes between undershoot and overshoot in one pair of endfoot and parenchymal vessel located in the remote area (one recording session, 9 seizures). Linear fit is: y = −1.17x + 13.542, R²= 0.77. (b) Bar plot of slopes of linear fits between relative diameter changes during undershoot and overshoot (as shown in (a)) in 13 pairs of astrocytic endfeet and parenchymal vessels (from four mice, 58 seizures) (p = 0.003, one sample t-test).

Using data from all mice with measures in remote areas (four mice, 13 astrocytes, 58 seizures), we observed that the average of [Ca²⁺]_i in astrocytic endfeet during the undershoot and the overshoot, was significantly larger than at baseline level (Figure 6(d), p = 0.025 and p = 0.028). The undershoot and overshoot periods were associated with a decrease and an increase in vascular diameter, respectively. The average decrease in parenchymal vessel diameter was significant during the beginning of seizures (Figure 6(e), p = 0.001) as well as their ensuing increase (Figure 6(e), p = 0.038). When measuring the diameter versus absolute [Ca²⁺]_i, we observed a significant constrictive trend quantified by their slope (Figure 6(f)), in all cases preserving the negative relationship seen in Figure 5(b). When characterizing the relationship of relative diameter changes during the undershoot and overshoot, a negative trend shown in Figure 7(a) was obtained in all cases (Figure 7(b), p = 0.003). This data indicate that seizure-like activities increase absolute [Ca²⁺]_i which in turn, associates with an increased constriction during the undershoot, and a decreased dilation during the overshoot.

Discussion

In this work, we showed that absolute [Ca²⁺]_i of astrocyte endfeet and diameter of parenchymal vessels can be simultaneously measured in the mouse cortex in vivo using two-photon fluorescence lifetime measurements with an appropriate excitation regime. In similar brain studies, the key advantages of two-photon fluorescence lifetime measures of Ca²⁺ are their relative immunity to changes in fluorophore concentration, light attenuation, and bleaching thus opening the door to investigating quantitatively changes in Ca²⁺ at different depths in the cortex with good spatial resolution. Here, we used this technique to investigate absolute [Ca²⁺]_i changes in astrocytic endfeet and diameter of parenchymal vessels during epileptic event induced by 4-AP. Our findings provide new correlative evidence regarding potential Ca²⁺ mediated diameter changes during epileptic activity, by characterizing the relationship between [Ca²⁺]_i and diameter.

The key findings of this study were the following. (1) The free [Ca²⁺]_i in astrocytes was measured, and a spatially heterogeneous distribution of Ca²⁺ was observed in astrocytic soma and endfoot. (2) Seizure activity induced a Ca²⁺ elevation in endfeet with simultaneous parenchymal vessel dilation at the epileptic focus. In remote areas, the Ca²⁺ increase was associated with parenchymal vessel constriction at the onset of seizure and dilation later during the seizure. (3) A negative relationship between basal Ca²⁺ in astrocytic endfeet and diameter changes in parenchymal vessels was found. These novel observations have implications for understanding the relationship between astrocytic endfeet [Ca²⁺]_i and parenchymal vessel diameter.

Center-surround phenomena during acute seizure activity

Using epileptic seizures induced by 4-AP in vivo, we investigated in this study the [Ca²⁺]_i in astrocytic endfoot and diameter of parenchymal vessels simultaneously. Our data showed that, in the epileptic focus, the elevation of Ca²⁺ was accompanied by vessel dilation (Figure 3(a)). In the remote area, a clear transient decrease in diameter despite a sustained increase of Ca²⁺ in astrocytic endfoot was shown (Figure 3(b)), showing that the elevation of Ca²⁺ could not predict a vasoconstriction or vasodilation in this area. Similar diameter changes during ictal discharges were investigated by several investigators with various methods.^13,14 These data suggest that other elements, for example, vascular tone, ³⁰ lactate, ⁹ and nitric oxide (NO) ³¹ could also influence the behavior of the vessel. The simplest explanation for this inhomogeneous response is a passive model whereby vasodilation response propagates upstream in a stepwise manner. Early in the seizure, vasodilation in the focus shunts blood from the surround to the focus. As vasodilation propagates further upstream, vessels dilate in the surround as well. This observation, however, seems to be in contradiction with our observed decrease of dilation at the focus and increase of constriction in the surround with increasing absolute [Ca²⁺]_i. In the passive model, one would expect constriction to be reduced in the surround when dilation is reduced in the focus, which is not the case here. Alternatively, our observations could be due to neurotransmitters released by interneurons, for example, neuropeptide Y (NPY), as 4-AP is known to induce the release of NPY, which is a strong constrictor in remote areas. These findings could also be a result of factors that switch the astrocyte response, for example, glutamate, which raises [Ca²⁺]_i. in astrocytes by generating messengers to regulate vessels dilation or constriction.

The relationship between calcium and diameter

The elevation of [Ca²⁺]_i in astrocytic endfoot associated with dilation of the parenchymal vessel with epileptic events at the epileptic foci. Our data are similar to results from sensory stimulation. ³² Meanwhile, the slow increase of basal [Ca²⁺]_i was also accompanied by a decrease of relative dilation during seizures over the whole time course of the experiment. This may be linked by the basal tonic arteriole blood flow control by astrocytes. ³³

The link between astrocytic [Ca²⁺]_i levels and arteriole dilation has been the subject of much debate ³⁴ an whether astrocytes modulate vasodilation remains controversial: initial studies indicated that increase in levels of [Ca²⁺]_i in astrocytes led to a release of arachidonic acid-derived messengers, prostaglandins, epoxyeicosatrienoic acids (EETs), and 20-hydroxyeicosatetraenoic acid (20-HETE), each known to have a modulating role on vascular smooth muscles.^9,35,36 However, this initial interpretation was mitigated by evidence showing that [Ca²⁺]_i signals in astrocytes were slow when compared to blood flow elevation caused by neuronal activity.^37–39 In more recent work, this was again questioned with the observation that astrocyte processes had distinct [Ca²⁺]_i transients that are more frequent than that found in the somata. ⁴⁰ For these reasons, we focused our measures on these processes. As our sampling rate (1 second) for the dual ([Ca²⁺]_i and diameter) measures was slow, our work does not contribute new data related to the neurovascular response associated with astrocytes, rather we focus on slow fluctuations of basal [Ca²⁺]_i and observed a negative correlation between response and absolute calcium values. One hypothesis is that this negative correlation corresponds to the relationship between vascular tone and extravascular potassium concentrations released by large-conductance calcium sensitive potassium channel, as was demonstrated in slices at higher endfeet [Ca²⁺]_i.. ⁶ This would suggest that the transition between vasodilation and vasoconstriction described in Girouard et al. ⁶ is gradual and proportional to [Ca²⁺]_i.

Nonlinear hemodynamic responses to seizures

Among original motivations for this work was the observation of inhibitory nonlinear hemodynamic phenomena ¹⁶ in epilepsy. Our data correlating the increase of basal [Ca²⁺]_i to an inhibitory dilation of parenchymal vessels suggests that [Ca²⁺]_i may indeed contribute to these observations. In the context of epilepsy, this decrease response may contribute to local hypoxia during seizures. More work is required, however, using different epileptogenic agents and models as well as in vivo uncaging of calcium and chelators to further understand the relative importance of [Ca²⁺]_i in this inhibitory phenomenon.

Limitations

There is a possibility that our 4-AP injections diffuse into tissue and that the compound itself play a role in measured calcium values. Given that 4-AP is not known to interact with calcium and that observations were acquired in regions that span beyond the expected diffusion (∼300 µm) of the compound (e.g., > 1.5 mm in remote areas), we believe this is unlikely. For recordings done at the focus, we did not observe any effect of grouping astrocytes closer to the injection site when compared to that located farther. Results could also be specific to the epileptic model used (4-AP), and more experiments would be required to validate similar findings in other epileptic models.

Conclusion

We developed a technique that simultaneously provided absolute values of [Ca²⁺]_i in astrocytic endfoot and diameter in adjacent arteriole in the brain cortex by means of two-photon fluorescence lifetime microscopy. The technique was applied to study how the astrocytic endfoot controls parenchymal vessels during epileptic activity in mice. To our knowledge, this is the first report of simultaneously measurements of absolute [Ca²⁺]_i values and diameter changes during epilepsy. In our work, following 4-AP injection in the somatosensory cortex of mice, we observed significant changes of [Ca²⁺]_i in endfoot and diameter of parenchymal vessel at the focus and in the remote areas. In addition, we found a negative correlation between basal [Ca²⁺]_i in astrocytic endfoot and the amplitude of parenchymal vessel dilation for all measurements, while, in the remote areas, there was a positive correlation between [Ca²⁺]_i in endfoot during the onset of seizures and the level of vasoconstriction.