Advanced in vitro approach to study neurovascular coupling mechanisms in the brain microcirculation

Abstract

An understanding of the signalling events underlying neurovascular coupling mechanisms in the brain is a crucial step in the development of novel therapeutic approaches for the treatment of cerebrovascular-associated disorders. In this study we present an enhanced in vitro brain slice preparation from male Wistar rat cortical slices that incorporates haemodynamic variables (flow and pressure) into parenchymal arterioles resulting in the development of myogenic tone (28% from maximum dilatation). Moreover, we characterized flow-induced vascular responses, resulting in various degrees of vasoconstrictions and the response to 10 mm K⁺ or astrocytic activation with the mGluR agonist, t-ACPD (100 μm), resulting in vasodilatations of 33.6 ± 4.7% and 38.6 ± 4.6%, respectively. Using fluorescence recovery, we determined perfusate velocity to calculate diameter changes under different experimental pH conditions. Using this approach, we demonstrate no significant differences between diameter changes measured using video microscopy or predicted from the velocity values obtained using fluorescence recovery after photobleaching. The model is further validated by demonstrating our ability to cannulate arterioles in two brain regions (cortex and supraoptic nucleus of the hypothalamus). Altogether, we believe this is the first study demonstrating successful cannulation and perfusion of parenchymal arterioles while monitoring/estimating luminal diameter and pressure under conditions where flow rates are controlled.

Key points

• We show that parenchymal arterioles of less than 30 μm in diameter can be successfully cannulated and perfused in cortical and hypothalamic brain slices.

• Increased flow/pressure induced parenchymal arteriolar constriction.

• Alternatively to video microscopy, diameter measurements of cannulated arterioles could be obtained from perfusion velocity values obtained using fluorescence recovery after photobleaching.

• Perfused and pressurized parenchymal arterioles responded with dilatation to two well-established signals in neurovascular coupling, K⁺ (10 mM) and activation of metabotropic glutamate receptors (mGluRs).

Introduction

A number of elegant studies have addressed the signalling events underlying neurovascular coupling in the brain. To this end, the in vitro brain slice preparation has been used to measure vascular responses to neuronal or astrocytic stimulation (Anderson & Nedergaard, 2003; Filosa et al. 2004, 2006; Mulligan & MacVicar, 2004; Blanco et al. 2008). Such studies have contributed significantly to our understanding of the vasoactive signals (i.e. prostaglandins, K⁺, epoxyeicosatrienoic acids) leading to vasodilatation during neuro-glial activation. In addition, a few groups have addressed myogenic reactivity and the mechanisms that control vascular tone in small (<100 μm) parenchymal arterioles (Bryan et al. 2001a; Cipolla & Bullinger, 2008; Hannah et al. 2011). The use of isolated pressurized (and sometimes perfused) vessels has advanced our understanding of the cellular mechanisms by which endothelial cells (ECs) and vascular smooth muscle cells (VSMCs) control myogenic reactivity (Duling et al. 1981; Dacey & Duling, 1982; Ngai & Winn, 1995, 1996; Bryan et al. 2001b) and participate in important physiological mechanisms such as brain autoregulation and neurovascular coupling or functional hyperaemia. However, the absence of a physiological neuro-glial microenvironment around parenchymal arterioles, as vessels in this preparation are pulled away from their native environment, limits our comprehensive understanding of the intricate cell-to-cell communication modalities participating in neurovascular coupling.

Clearly, the various techniques used to study neurovascular coupling come with their own set of advantages and limitations. For example, two-photon microscopy has provided fundamental information with regards to the participation of astrocytes, pericytes and neurons in blood flow responses in vivo (Schummers et al. 2008; Nimmerjahn et al. 2009; Fernandez-Klett et al. 2010). However, limitations such as the temporal and spatial resolution of the system, its high cost, its limitation to the surface of the brain, and the need to use high doses of pharmacological agents, limits advancement for further studies underlying cellular mechanisms governing neurovascular coupling. An alternative and sometimes complementary approach has been the use of brain slices. An advantage of this preparation is the study of signalling mechanism from both neurons and astrocytes in a more controlled manner in which all the elements of the neuro-glial microenvironment are preserved. A major limitation in these studies, however, has been the lack of luminal pressure and/or flow, having important consequences such as keeping vascular cells from being at their resting/active state, a condition that may significantly alter the characteristics of the vascular response generated following glia or neuronal stimulation and vice versa. An attempt to circumvent the limitation of pressure/flow within parenchymal arterioles has been the addition of a vasoconstrictor agent (i.e. U46619 (Brown et al. 2002; Filosa et al. 2004; Metea & Newman, 2006; Blanco et al. 2008; Straub et al. 2009; Girouard et al. 2010), l-NAME (Zonta et al. 2003; Mulligan & MacVicar, 2004), prostaglandin (Staunton et al. 1998, 2000) or vasopressin (Fergus & Lee, 1997a,b)) to the perfusate. While this later step increases vascular tone, bath application of these agents may inadvertently alter the state of key players in neurovascular coupling, creating a new problem in data interpretation of signalling mechanisms.

Given that luminal flow and pressure are significant determinants of vascular tone (Bevan & Laher, 1991; Ngai & Winn, 1995; Ward et al. 2000; Nystoriak et al. 2011), in this study we sought to: (1) optimize the brain slice preparation model by incorporating these haemodynamic variables into cortical parenchymal arterioles with the endpoint of achieving endogenous myogenic tone and bringing ECs and VSMCs to their resting, but yet active state, (2) characterize the response of parenchymal arterioles to haemodynamic variables in the presence of an intact neurovascular unit, and (3) determine if pre-capillary arterioles (vessels less than 30 μm) respond similarly to haemodynamic variables as those of higher calibre, as previously reported (Bryan et al. 2001b; Toth et al. 2011).

To our knowledge, only one published study has addressed neurovascular coupling mechanisms in the presence of luminal pressure/flow in the slice preparation (Lovick et al. 2005). However, the parameters of pressure/flow were not tightly controlled. The effects of flow on vascular reactivity has also been evaluated under in vivo conditions (Takano et al. 2006; Nishimura et al. 2007), in larger vessels (Shapiro et al. 1971) as well as in excised parenchymal arterioles of greater diameters (>60 μm) (Shimoda et al. 1996, 1998; Bryan et al. 2001a; Horiuchi et al. 2001, 2002; Cipolla et al. 2004, 2009; Cipolla & Bullinger, 2008; Toth et al. 2011). As the effects of flow on vessel diameter are dependent on the calibre of the vessel, understanding the effects of flow on pre-capillary parenchymal arterioles in situ is essential, and forms the basis for the development of this novel experimental approach.

Methods

Experimental set-up

Experiments were conducted using the Andor Revolution system (Andor Technology Belfast, UK) attached to a Nikon microscope (Eclipse FN 1, Nikon, Tokyo, Japan). The microscope was connected to a FRAP-PA (fluorescence recovery after photobleaching-photoactivation) module used for photobleaching experiments. The FRAP module was in turn attached to a laser confocal spinning unit (CSU-X1, Yokogawa, Tokyo, Japan) connected to a Sutter filter wheel and an ultrasensitive EMCCD camera (iXon, Andor Technology, Belfast, UK). Figure 1 includes a sketch of the experimental set-up. The microscope chamber was continuously perfused with artificial cerebrospinal fluid (aCSF) using a peristaltic pump (Miniplus 3, Gilson, Middleton, WI, USA) at a rate of ∼2–3 ml min⁻¹. Chamber temperature was maintained at 33 ± 2°C using a single line solution heater (SH-28B, Warner Instruments, Hamden, CT, USA) connected to a DC power supply (1735A, BK Precision, Yorba Linda, CA, USA). Temperatures were maintained at this range as it helps preserve the viability of the slice preparation (Hrabitova & Nicholson, 2007). Finally, micromanipulators were employed to hold and manoeuver the glass pipette (cannula) used to cannulate arterioles.

Figure 1. Experimental set-up and identification of cortical microvessels.

A, schematic representation of the experimental set-up. B, representative DIC images from a cortical brain slice showing an arteriole (left), a venule (middle) and a capillary (right). FRAPPA, fluorescence recovery after photo-bleach and photo activation. Calibration bar, 10 μm.

Brain slice preparation

Cortical brain slices were prepared from juvenile (P23–31) Wistar rats following protocols approved by the animal care and use committee of the Georgia Health Sciences University and also comply with the policies and regulations as required by The Journal of Physiology (Drummond, 2009). Following anaesthesia with sodium pentabarbitol, the brain was removed and cut into 250–300 μm thick coronal slices using a vibratome (Leica VT 1200S, Leica Microsystems, Wetzlar, Germany) in cold aCSF (in mm): KCl 3, NaCl 120, MgCl₂ 1, NaHCO₃ 26, NaH₂PO₄ 1.25, glucose 10 and CaCl₂ 2, plus 400 μm l-ascorbic acid, with osmolarity at 300–305 mosmol l^-1, equilibrated with 95% O₂–5% CO₂. Slices were kept at room temperature (RT) in aCSF equilibrated with 95% O₂–5% CO₂ (pH 7.4) until transferred to the microscopy chamber.

Vessel cannulation

Parenchymal arterioles were visualized using a 60× Nikon objective (NIR Apo, 60×/1.0w, DIC N2, ∞/0 WD 2.8) equipped with infrared differential interference contrast (IR-DIC) optics. The first step was the identification of a parenchymal arteriole in a cortical brain slice. Figure 1B shows representative images of a non-cannulated arteriole, a venule and a capillary. As noted in the figure, arterioles were characterized by the presence of a continuous thick smooth muscle cell layer, whereas venules and capillaries had thinner walls with no smooth muscle. The identification of the vessel type is a critical component of the experimental procedure since, once flow is introduced, the arteriolar wall becomes thinner in appearance as the muscle layer stretches, and may be inadvertently confused for a venule. Cannulas (ID 1.17 mm and OD 1.50 mm, G150TF-3, Warner Instruments, Hamden, CT, USA) were pulled with a micropipette puller (P-97 puller Sutter Instruments, Novato, CA, USA) and mounted onto a micromanipulator. Luminal flow was controlled with the use of a syringe pump (PHD 2000, Harvard Apparatus, Holliston, MA, USA). A pressure transducer was placed just prior to the cannula for constant pressure monitoring. The internal cannula solution consisted of (in mm): KCl 3, NaCl 135, MgCl₂ 1, glucose 10, Hepes 10 and CaCl₂ 2, plus 1% albumin (Duling et al. 1981; Dacey & Duling, 1982) with osmolarity at 300–305 mosmol l^-1 and pH 7.4 adjusted with NaOH. The resistance of each cannula was obtained prior to the cannulation procedure and determined over a given range of flow rates as shown in Fig. 2A. Next, the tip of the cannula (having minimal flow) was manoeuvered towards the entrance of the arteriole and slowly introduced into the vessel lumen (Fig. 2B inset). Following cannula insertion, flow rate was increased to 0.1 μl min⁻¹ and pressure continuously monitored with a Servo Pump (PS/200, Living System Instrumentation, Burlington, VT, USA). To attain relative physiological haemodynamic values we first calculated shear stress (SS) at the lowest flow rate using equation t = 4 η/πr³, where t equals shear stress (in dynes cm^-2), η equals viscosity, equals flow rate and r equals radius, as previously defined (Koller et al. 1993; Bryan et al. 2001a). The SS value at the lowest flow rate possible provides the estimated range of flow rates that can be used to control luminal pressure. Flow ranges were then altered so as to increase luminal pressure while keeping SS preferably below 100 dynes cm⁻². With the desire flow rate in place and the pressure reading from the transducer, we then calculated the resistance_total= (resistance_cannula+ resistance_vessel) using Ohm's law (resistance _electrical= drop in voltage / flow of current) R=ΔP/ where R equals the total resistance and P equals pressure. The pressure in each arteriole was calculated from the pressure measured in the cannula, the resistance in the cannula alone and the total resistance in the cannula plus the arteriole. In a few cases, to increase resistance without changing luminal flow rate, a second cannula pulled and fire polished was use to occlude the distal end of the arteriole (Supplementary movie 1). This later step was not needed in cortical arterioles which had longer lengths; however, it is shown here as it becomes a useful tool to manipulate the resistance of shorter vessels.

Figure 2. Cannulation of cortical microvessels and development of myogenic tone.

A, cannula and tubing resistance as defined by the change in pressure over a range of flow rates. B, representative trace showing the time course for the development of myogenic tone following cannulation and perfusion of a parenchymal arteriole (a; resting state before cannulation, b; immediately after cannulation and perfusion at 0.1 μl min⁻¹ and c; steady state at a perfusion rate of 0.1 μl min⁻¹). Inset shows a DIC image of a parenchymal arteriole with the cannula. Calibration bar, 10 μm. C, averaged number of constriction from single VSMC throughout the development of myogenic tone. D, summary data of the development of myogenic tone over time at 0.1 μl min⁻¹. Data are means ± SEM.

Video microscopy

Diameter changes in cortical arterioles (8–20 μm internal diameters) were recorded using an EMCCD camera (Andor Tech). Images were acquired at 1 frame s⁻¹, visualized and stored using Andor IQ software (Andor Tech).

Fluorescence recovery after photobleaching (FRAP)

Using a similar approach to that described by Flamion et al. 1991, FRAP recordings were conducted to measure axial velocity, and to estimate in this way changes in arteriolar diameter as defined by the equation ν=/πr², where ν equals axial velocity, equals flow rate and r equals vessel radius. Velocity values where determined from equation 6 in Flamion et al. 1991 were ν=L/(slope⁻¹+½t_b+t_s). For our experimental conditions, L (defined as the length of the photobleached area) is 9 μm, the slope (rate of fluorescence recovery and thus a component of velocity) was the measured tau, t_b (time of bleach pulse) was 1 ms and t_s (discarded time from end of bleach pulse to beginning of acquisition) was 2 s. To this end, 20 μm Alexa 488 was added to the intraluminal solution. Following a successful cannulation and flow/pressure equilibration, a region of interest (ROI, 37 × 37 pixels) was drawn and placed within the arteriolar lumen image. The ROI was bleached using a 1 ms pulse with a 488 nm laser set at 100% intensity. Under these conditions, light-induced photoactivation was not observed. Fluorescence intensity was recorded before, during and after the bleached pulse at a rate of ∼5–6 frames s⁻¹. Images were stored for later analysis using Andor IQ software. Using flow rates and pressures defined in the study, no dye extravasation from the arteriole into the extracellular space was observed suggesting an intact blood brain barrier.

Immunohistochemical studies

Following anaesthesia (sodium pentobarbital, 60 mg kg⁻¹i.p.) rats were subjected to transcardiac perfusion with 0.01 m PBS (150 ml) and then 4% paraformaldehyde (0.01 and 0.3 m PBS), pH 7.2, (∼350 ml). The brain was then removed and postfixed in 4% paraformaldehyde for 3 hrs at 4°C and cryoprotected (0.01 m PBS containing 30% sucrose for 72 h at 4°C). Hypothalamic coronal sections (Leica cryostat CM3050, 25 μm) were collected in 0.01 m PBS at 4°C. Sections were incubated with 0.01% Triton X-100, 0.04% NaN₃ and 10% normal horse serum for 1 h. For immunofluorescence reactions, sections were incubated for 24 h in a mixture of primary antibodies that included a polyclonal guinea pig anti-oxytocin (1:200,000; Bachem) and polyclonal guinea pig anti-(Arg⁸)-vasopressin (1:200,000, Bachem). Reactions with primary antibodies were followed by 4 h incubation with secondary antibodies (donkey anti-guinea pig-Cy3 labelled (1:250 dilution) diluted in PBS containing 0.01% Triton X-100 and 0.04% NaN₃, Jackson ImmunoResearch Laboratories). In addition, the vasculature was stained using Isolectin-GS-IB₄ (Invitrogen). Following the immunohistochemical steps, slices were incubated overnight at 4°C in 0.01 m PBS, 1 mm CaCl₂, 10 mm Hepes, 0.1% Triton X-100, pH 7.4 with 1 m NaOH and 5 μl/500μl Isolectin-GS-IB₄ Alexa-568 from 1 mg ml⁻¹ stock solution. Following incubation, slices were washed with 0.01 m PBS for 15 min, three times at RT.

Drugs and chemicals

All drugs with the exception of t-ACPD (Tocris, Ellisville, MO, USA) and bovine serum albumin (Fisher Scientific, Pittsburgh, PA, USA), were purchased from Sigma (St Louis, MO, USA). t-ACPD was initially diluted in 1 m NaOH and then dissolved further to give a final dose of 100 μm in aCSF.

Data analysis

Arteriolar diameter data using infrared differential interference contrast (IR-DIC) and FRAP experiments were analysed using Andor IQ. Changes in the internal (luminal) arteriolar diameter were determined from averaged measurements taken at multiple points across the arteriolar lumen. Arteriolar diameter values are expressed as per cent (%) from maximum dilatation in the presence of zero Ca²⁺ and papaverine (100 μm). Tau (τ), which defines the fluorescence recovery rate, was determined from a single exponential fit using Clampfit 10.2 software (Molecular devices, Sunnyvale, CA, USA). Summary data are expressed as mean ± SEM. Differences between two means from the same arteriole were determined using two-tailed Student's paired t-test. One-way ANOVA (Dunnett's multiple comparison) was used for multiple group comparisons against control data. P < 0.05 was considered significant.

Results

Experimental configuration and cannulation procedure

Details of the experimental set-up are described in the Methods section and schematically depicted in Fig. 1A. Brain slices were placed on the recording chamber and held in place with a nylon grid as previously reported (Blanco et al. 2008). The resistance of the cannula and tubing was calculated prior to cannulation. Pressure was linearly increased with increasing flow rates from 0.5 to 2.5 μl min⁻¹ (R = 0.9994, n = 9–12, P < 0.0001) (Fig. 2A).

Next, a cannula having minimal outflow and filled with luminal solution was introduced into the arteriole (Fig. 2B inset). The diameter of the parenchymal cortical arterioles studied ranged from 8 to 20 μm. To prevent leakage at the proximal end of the arteriole, the cannula was pushed inwards until the outside wall of the glass pipette made a tight contact with the inside lumen of the arteriole. Following a successful cannulation, total resistance (resistance_total i.e. resistance_cannula+ resistance_vessel) was calculated from at least two flow rates applied to the arteriole (0.1 and 0.2 μl min⁻¹). The resulting value was used to determine the estimated per cent drop in luminal pressure. On average, luminal pressure between these flow rates dropped 21.0 ± 2.9% (n = 19). Using these parameters we could then estimate the flow rate needed to achieve luminal pressures in the order of 40 mmHg, the preferred physiological range (Golding et al. 1998). However, because SS is directly proportional to flow rate, in some cases, high flow/SS levels may be needed to achieve luminal pressures in the order of 40 mmHg. To overcome this problem, vascular resistance was increased through the placement of a second pipette on the distal end of the arteriole (Supplementary movie 1). In the presence of constant flow, the fire-polished pipette was gently pressed against the arteriolar wall so as to decrease the opening at the distal end. The increased resistance led to a slow increase in luminal pressure without interfering with flow/SS. Once a successful cannulation was achieved, luminal perfusion rate was set to 0.1 μl min⁻¹ and the arterioles were allowed to stabilize (∼20 min), reaching steady-state myogenic tone (Supplementary movie 2, Fig. 2B and D). At a flow rate of 0.1 μl min⁻¹ luminal diameter decreased to 9.1 ± 2.1% (Fig. 2B and D, n = 3, P = 0.0481) from maximum diameter in zero Ca²⁺ and papaverine (100 μm). Once tone begun to develop, the frequency of individual VSMC constriction increased during the first 5 min to a plateau value of 6 constrictions min⁻¹, indicative of steady-state tone (Fig. 2C, n = 11, P < 0.0001, ANOVA).

Flow rate responses

In order to determine if the modified in vitro model also enabled the efficient assessment of diameter changes when performing fluorescence imaging acquisitions rather than through video microscopy (i.e. such as during simultaneous measurements of Ca²⁺ dynamics in astrocytes and VSMC), we used a similar approach to that described by Flamion et al. 1991. To this end, the cannulated arteriole was perfused with a solution containing 20 μm Alexa 488. Following the equilibration period, flow rate was set to 0.5 μl min⁻¹ and arterioles were subjected to a photobleaching pulse within a predefined ROI. The rate of fluorescence recovery within the ROI, and corresponding to the axial velocity, was used to estimate diameter changes under different conditions (Flamion et al. 1991).

Figure 3 shows representative DIC and fluorescence images of a perfused arteriole subjected to bath application of an acidified (pH 6.8) or alkalinized (pH 7.65) aCSF. As previously shown, parenchymal arterioles dilate in response to acidification and constrict in response to alkalinization (Ngai & Winn, 1995; Nakahata et al. 2003). Fluorescence was monitored before, during and after the bleach pulse at each different pH condition (Fig. 3A left and Supplementary movie 3). The rate of fluorescence recovery was measured by fitting a single exponential function and determining τ during the recovery phase (Fig. 3A right and B). Under control (pH 7.4) conditions τ was 1.95 ± 0.19 s, increased to 3.12 ± 0.18 s at pH 6.8 and decreased to 1.32 ± 0.16 s at pH 7.65. The calculated axial velocity values were 2.28 ± 0.11 μm s⁻¹ under control conditions and 1.75 ± 0.06 and 2.70 ± 0.11 μm s⁻¹ at pH 6.8 and 7.65, respectively (Fig. 3C). Using the equation ν=/πr² we then calculated the predicted diameter changes under each different pH condition. Based on the calculated axial velocity, acidification induced a 12.50 ± 2.14% increase and alkalinization induced an 8.15 ± 1.80% (n = 6) decrease in arteriolar diameter (Fig. 3E). The per cent diameter changes from the predicted values were closely matched to the measured diameter values using video microscopy where acidification induced a diameter increased of 12.36 ± 4.23% (n = 6, P = 0.0309) and alkalinization induced a diameter decreased of 10.78 ± 2.9% (n = 6, P = 0.0136) (Fig. 3D and E). The data show that under conditions where video microscopy cannot be used to measure diameter changes, diameter values derived from the measurements of axial velocity using FRAP can readily overcome this limitation.

Figure 3. Estimated diameter changes using FRAP.

A, representative DIC (upper) and fluorescence images (below) of a parenchymal arteriole filled with Alexa 488 in the presence of acidic (pH 6.8), control (pH 7.4) and alkaline (pH 7.65) solutions, respectively (left panel). Calibration bar, 10 μm. Trace of fractional fluorescence showing a single exponential fit (τ, grey dashed line) during the recovery and plateau phase (right panel). B, bar graph of averaged τ, % velocity (C) and % diameter changes (D) in response to acidification (open) and alkalinization (black) and under control conditions (gray). E, summary data of % changes in diameter from the actual measurements using video microscopy and the estimated values from the measured velocity values using the FRAP technique. Data are means ± SEM.

Flow-dependent arteriolar responses

As previously reported, cerebral arterioles have myogenic reactivity, i.e. the ability to adjust diameter to pressure changes in order to maintain constant blood flow (Heistad & Kontos, 1979; Hurn & Traystman, 1997; Cipolla et al. 2004). In addition, it has been reported that the response of cerebral arterioles to increased flow varies according to the size of the arterioles, with larger vessels having the least responsiveness to pressure changes (Faraci & Heistad, 1990). One of the goals of this study was to determine if parenchymal arterioles of less than 30 μm in diameter within an intact neuro-glial microenvironment, responded in the same manner as previously reported in excised arterioles (Dacey & Duling, 1982; Koller et al. 1993; Cipolla et al. 2004). Arterioles were subjected to step increases in luminal flow rates (0.1–1 μl min⁻¹) while diameter was monitored using video microscopy. Following the equilibration period, flow rate was increased up to 1 μl min⁻¹. Increasing flow rate to 0.4 and 0.5 μl min⁻¹ led to a reduction in diameter of 15.4 ± 2.8% (n = 13, P = 0.0245) and 17.7 ± 2.6% (n = 11, P = 0.0169), respectively, from baseline, indicative of flow-mediated vasoconstriction. However, at flow rates greater than 0.5 μl min⁻¹ arteriolar diameter became unstable resulting in an increase in diameter (Supplementary movie 4, Fig. 4A), probably due to forced distension (Ngai & Winn, 1995). These data suggest that under the conditions used, parenchymal arterioles develop myogenic tone and adjust their diameter in concert with changes in luminal flow. Increasing flow rate also resulted in a linear increase in SS up to a certain level where SS increased exponentially probably due to turbulent flow (Fig. 4B, R = 0.9577, n = 4–13, P < 0.0001). As expected, luminal pressure was linearly increased in response to flow rate increases (Fig. 4C, R = 0.9590, n = 4–13, P < 0.0001).

Figure 4. Effects of flow rate on arteriolar diameter changes.

A, representative example of diameter changes over time from a parenchymal arteriole exposed to increasing perfusion rates (left panel); summary data of per cent changes in diameter (from maximal dilatation) in response to flow rates from 0.1 to 1 μl min⁻¹ (right panel). B, shear stress changes in response to increases in flow rates from 0.1 to 1 μl min⁻¹ (left panel); summary data of per cent shear stress changes as a function of flow rate (right panel). C, estimated luminal pressure changes in response to increasing levels of flow rate from 0.1 to 1 μl min⁻¹ (left panel). Summary data of the estimated luminal pressure changes as a function of flow rate (right panel). R, correlation coefficient. Data are means ± SEM.

K⁺ and astrocytic activation-induced vascular responses in perfused parenchymal arterioles

A unique advantage of studying vascular physiology in the slice preparation is that it allows the assessment of the precise contribution of neuronal and/or glial activation to the neurovascular coupling process. However, an important observation is that the level of tone in non-perfused parenchymal arterioles determines the polarity of the vascular response to glial-derived signals (Blanco et al. 2008), stressing the importance of vascular tone. We, thus, evaluated whether in the presence of myogenic tone, parenchymal arterioles responded with dilatations or constrictions following 10 mm K⁺ or astrocyte stimulation through activation of metabotropic glutamate receptors via t-ACPD (100 μm) (Fig. 5). Following the equilibration period, flow rate was increased to 0.5 μl min⁻¹ resulting in an averaged SS of 35 dynes cm⁻² from a range of 8 to 92 dynes cm⁻², an averaged luminal pressure of 34 mmHg and an averaged tone of 28% (Fig. 4), comparable to previously reported values (Kroll et al. 1996). Bath application of 10 mm K⁺ resulted in dilatations of 33.6 ± 4.7% (n = 8, P < 0.0022) while activation of metabotropic glutamate receptors resulted in dilatations of 38.6 ± 4.6% (n = 8, P < 0.0038) (Fig. 5A and C and Supplementary movie 5). Figure 5B shows the relationship between arteriolar responses and percent baseline tone. As previously reported (Blanco et al. 2008), the degree of arteriolar responses to K⁺ or t-ACPD were dependent on the initial level of tone of the arterioles (R = 0.6523, n = 8, P = 0.0153 for K⁺ and R = 0.7517, n = 8, P = 0.0053 for t-ACPD). Of note, no K⁺/t-ACPD-induced vasoconstrictions were observed in perfused arterioles with myogenic tone.

Figure 5. Effects of K⁺ and astrocytic activation on arteriolar diameter changes.

A, diameter changes in response to 10 mm K⁺ and 100 μmt-ACPD at 0.5 μl min⁻¹ (upper trace); corresponding shear stress changes in response to 10 mm K⁺ and 100 μmt-ACPD at 0.5 μl min⁻¹ (lower trace). B, per cent vascular responses (K⁺ and t-ACPD) as a function of the % baseline tone. R, correlation coefficient (P = 0.0153 (○, dashed line) and P = 0.0053 (•, continuous line)). C, summary data of per cent vascular responses to 10 mm K⁺ and 100 μmt-ACPD at 0.5 μl min⁻¹. Data are means ± SEM.

Validation of the use of perfused cerebral arterioles in other brain regions

An important improvement of the described methodology is the ability to study the neurovascular coupling mechanism in areas that lie deep within brain regions inaccessible using two-photon imaging technology, providing in this manner a tool to assess and compare the neurovascular coupling mechanism throughout the brain. A unique area, for example, is the supraoptic nucleus (SON) of the hypothalamus which contains oxytocin (OT) and vasopressin (VP) neurons and is critically involved in reproductive and fluid–electrolyte homeostasis. The SON is characterized by having a dense capillary network (Ambach & Palkovits, 1979; Bourque, 1990; Badaut et al. 2000) as well as astrocytes with dynamic processes which retract/expand in an activity-dependent manner (Theodosis et al. 2008), providing a unique model to study the consequences of neuro-glial structural remodelling in neurovascular coupling. As proof of principle, we cannulated and perfused cerebral arterioles in the SON of the hypothalamus. Figure 6A shows immunoreactivity against OT and VP neurons (green) and vascular staining using Isolectin-GS-IB₄ (red) in the SON. As shown, the SON receives its blood supply from arteries originating from the circle of Willis (i.e. anterior cerebral artery, lateral supraoptic and lateral thalamostriatal arteries) (Ambach & Palkovits, 1979; Bourque, 1990) that can be easily seen to enter the brain at the base of the nucleus. As in the cortex, cannulation and perfusion of SON arterioles lead to the development of myogenic tone (Fig. 6B and C and Supplementary movie 6). The unique vascular organization of the SON, relative to neuronal cell bodies/processes and astroglial elements (Fig. 6), calls for studies in distinctive brain regions in which neurovascular coupling (NVC) mechanisms may significantly differ from those in the cortex and hippocampus.

Figure 6. Validation of the technique in another brain region, the supraoptic nucleus of the hypothalamus.

A, the SON is clearly delineated by OT and VP immunoreactive neurons as well as by the well-defined vasculature network (Isolectin-GS-IB₄). B, representative DIC image of a large arteriole (left) and venule (right) penetrating the SON region; note the cannula within the arteriole. C, diameter changes from an SON arteriole in response to increasing perfusion levels. Calibration bar, 10 μm.

Discussion

In this study we demonstrate: (1) our ability to successfully cannulate and perfuse parenchymal arterioles in two distinct brain regions (cortex and hypothalamus) while monitoring diameter, SS and luminal pressure; (2) flow/pressure-dependent responses of pre-capillary parenchymal arterioles; and (3) the response of perfused and pressurized parenchymal arterioles to two well-established signals in neurovascular coupling, K⁺ (10 mm) and activation of mGluR.

Among the limitations encountered throughout the development of this technique was our inability to precisely measure luminal pressure; vessel branching and leakage was also a challenge. However, both of these limitations were successfully addressed. Luminal pressure was estimated based on the percentage drop in resistance from the cannula/tubing and that of the total system (Bryan et al. 2001a). In a few cases, potential pressure drops or leakages were controlled by the introduction of a second cannula just prior to the first branching point of the arteriole, a few micrometres below the site of diameter recordings. Our data show that following a successful cannulation, arterioles develop myogenic tone ∼30% from maximum diameter, consistent with values of tone reported from studies in excised arterioles cannulated at both ends and where the pressure is tightly controlled (Dacey & Duling, 1982; Ngai & Winn, 1995; Bryan et al. 2001a; Cipolla et al. 2004). Moreover, once arterioles developed tone, stable diameters were maintained for long periods of time (>1 h), suggesting pressure leakage was minimal and did not have an effect on the ability of the arterioles to sustain tone. While this is not a trivial limitation, the successful achievement of spontaneous myogenic tone within these very small arterioles, the ability to perfuse them, in combination with imaging and electrophysiological techniques already well established using brain slices (Filosa et al. 2006; Hamel, 2006; Blanco et al. 2008), promises significant advancements towards studies that will expand our current understandings of cell-to-cell communication at the neurovascular unit. While the parameters used in this study allowed for the development of myogenic tone of parenchymal arterioles, a caveat to our experiments is that they were conducted at subphysiological temperatures (33 ± 2°C) with the aim of preserving the viability of the slice preparation. However, given temperature-dependent changes in vascular tone (Winquist & Bevan, 1980; Ogura et al. 1991; Sagher et al. 1997) and cerebral blood flow (Busija & Leffler, 1987) it is recommended that future experiments are conducted at normothermic conditions for the animal species used.

Importantly, we were able to characterize pressure/ flow-dependent responses in cortical parenchymal arterioles, indicative of proper myogenic reactivity. Due to the fact that our model does not dissociate between pressure and flow, it is unclear whether the observed vasoconstriction responses are the result of flow or pressure changes as these variables were simultaneously altered. Similarly, it would be difficult to dissociate between pressure and flow in an in vivo setting. While data from our study are in agreement with previous reports showing flow-induced constrictions (Madden & Christman, 1999; Bryan et al. 2001a) and flow plus pressure-induced constrictions in excised arterioles (Toth et al. 2011), it is important to point out that biphasic responses (dilatations and constrictions) have been also reported in vitro (Garcia-Roldan & Bevan, 1990, 1991; Ngai & Winn, 1995; Shimoda et al. 1996, 1998; Thorin-Trescases & Bevan, 1998). Only flow-induced dilatations have been reported in vivo (Fujii et al. 1991), suggesting that more studies are needed to better understand the physiological relevance of these mechanisms as well as the intricate interplay between pressure, flow and vasoactive signals. In light of these observations and as previously suggested (Bevan & Joyce, 1990; Bevan & Laher, 1991), it is possible that upon changes in pressure/flow multiple mechanisms are activated and the net vascular response is the result of the competitive action of the vasoactive signals released. On the other hand, increases in pressure/flow in the presence of neuronal activation will result in the release of vasodilatory signals that overcome flow/pressure-induced constrictions. For example, using forepaw stimulation Devor et al. (2007 and 2008) reported a centre–surround pattern of activation (Devor et al. 2007) where net neuronal depolarizations and arteriolar vasodilatations correlated with increased oxygenation and net neuronal hyperpolarizations and arteriolar vasoconstrictions away from the stimulus-evoked region correlated with a decrease in oxygenation (Devor et al. 2007). Interestingly, arterioles exhibited biphasic (dilatory and constrictive) responses even at the site of neuronal depolarization suggesting that opposing vascular responses are indeed activated during functional hyperaemia in vivo (Devor et al. 2007, 2008). We hypothesize that pressure/flow-induced constriction can participate in the functional hyperaemic response in the following ways: (1) in the absence of a dilator signal (when the signal is washed away, concentration diminished or release stopped) pressure/flow-induced constriction could act as an active mechanism to rapidly return vascular tone to its original state and/or (2) flow-induced constriction away from the site of neuronal stimulation (where its influence is more dominant due to minimal vasodilatory influences) helps to spatially delineate the area of increased flow contributing to the well-defined regional/focalized increase in cerebral blood flow following neuronal activation. The latter will also protect the brain from large changes in volume and intracranial pressure as discussed in Bryan et al. (2001a). While Devor's studies suggest that the vasoconstriction phenomenon is associated with the activation of inhibitory neurons and the release of vasoconstrive substances (Kleinfeld et al.; Devor et al. 2007, 2008), other possibilities such as a direct effect of flow (and or pressure) on the vasculature (i.e. stretch-activated channels, integrins; Madden & Christman, 1999; Bryan et al. 2001a) cannot be ruled out.

Our data show that following the development of myogenic tone, arterioles predictably responded to stimuli such as K⁺ and/or astrocyte stimulation through activation of mGluRs. Intriguingly, astrocytic activation has been shown to induce both dilatation and constriction depending on the level of tone of the arteriole (Blanco et al. 2008; Gordon et al. 2008) or the oxygen concentration in the preparation (Devor et al. 2007; Gordon et al. 2008), suggesting the possibility that changes in the resting state of the vascular cells may indeed alter the functional outcomes of the neurovascular unit. In our hands, and using as close to physiological haemodynamic variables as this preparation will allow, astrocyte-induced vascular responses on arterioles that had tone were only associated with vasodilatations and not constrictions, consistent with in vivo observations (Takano et al. 2006).

A growing body of evidence suggests an important contribution from glial and neuronal elements to vascular reactivity, in particular arterioles that lie deep within the neuropile. The approach presented here advances previously established in vitro techniques, which lacked controlled luminal pressure and/or flow (Lovick et al. 2005). While in the present study we present data from parenchymal arterioles from the cortex and SON, these technical approaches can be readily applied to other brain regions, which we expect to facilitate a significant advancement in the field, given limited knowledge on neurovascular coupling mechanisms in brain areas inaccessible with in vivo approaches. Moreover, the versatility of studying arterioles in specific locations provides a unique opportunity to assess the impact that distinct neuro-glial microenvironments have on neurovascular coupling function. This includes, for example, areas that lack a blood–brain barrier (i.e. circumventricular organs), areas where the release of vasoactive substances is common (i.e. hypothalamus), or areas where rapid and reversible rearrangement of the neuro-glial microenvironment takes place. From a mechanical perspective, the presence of pressure/flow within the arteriole would further our understanding of the functional implications stretch-activated channels, particularly those expressed at the neurovascular junction (Corey, 2003; Vriens et al. 2005; Benfenati et al. 2007; Morita et al. 2007; Fernandez-Fernandez et al. 2008; Earley et al. 2009), have on neurovascular coupling as the presence of haemodynamic variables may alter the types of ion channels that are constitutively active at rest, the resting membrane potential of vascular cells and the polarity of the vascular response to signals released during neuronal activation.

We believe this improved methodology will help elucidate signalling mechanisms at the neurovascular unit in brain areas currently inaccessible with in vivo approaches. Moreover, this approach will enable addressing important questions as to whether haemodynamic changes per se can alter vascular responses during neuronal and glia stimulation. In addition, the ability to perfuse parenchymal arterioles with physiologically active substances or conditions, will help differentiate signals that can cross and/or alter the blood–brain barrier, and whether these effects are region specific, and/or whether these factors influence neuronal and glia function. It is our hope that this improved preparation will be also applied to other slice preparations such as the pancreas (Rupnik, 2009; Huang et al. 2011), lung (Moreno et al. 2006; Liberati et al. 2010) and other organs (Vickers & Fisher, 2005), to better understand intercellular communication between the microcirculation and other cell types. Findings from such studies will provide new answers and clarify the complexity of cell-to-cell signalling. Moreover, the characterization of responses in the slice preparation may provide a bridge towards the study of human organ slices and a way to extrapolate dysfunction from a clinical perspective (Rupnik, 2009).

Glossary

Abbreviations

EC

endothelial cell

mGluR

metabotropic glutamate receptor

VSMC

vascular smooth muscle cell

Author contributions

Both authors (K.J.K. and J.A.F.) contributed to design and execution of the experiments, data analysis and writing of the manuscript. Both authors approved the final version for publication. The work was conducted at Georgia Health Sciences University.

Supplementary material

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Supplementary Movie 4

Supplementary Movie 5

Supplementary Movie 6

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors