INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CONDUCTION of a dilation or constriction response along the length of an arteriole is a well-recognized behavior (6, 9, 22). This form of cell-to-cell communication is thought to occur as a result of an electrical signal originating at the point of stimulus and conducting along cells of the vessel wall (27, 28). Such conduction may play a critical role in the response of arterioles and arteries to the changing milieu and oxygen demands of tissue.

There is evidence that pharmacological stimulation of capillaries can cause responses in upstream arteriolar segments (7, 26). If capillaries were to respond to pharmacological stimulation with changes in endothelial cell membrane potential in vivo, this might provide a mechanism linking capillary function to tissue demand.

We have tested the capacity of the capillary endothelial cell membrane potential to respond to a variety of stimuli using the voltage-sensitive dye 1-(3-sulfonatopropyl)-8-[-[2-(di-n-butylamino)-6-naphthyl]vinyl]pyridium betaine (di-8-ANEPPS). We found that KCl, acetylcholine (ACh), and phenylephrine (PE) can all cause changes in capillary membrane potential.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animal preparation. Male Golden hamsters (90-150 g body wt) were prepared for in vivo examination of the cheek pouch (8). Each animal was anesthetized with an intraperitoneal injection of pentobarbital sodium (80-100 mg/kg). Body temperature was maintained at 36.5-37.5°C by radiative and conductive heating. A tracheostomy tube was placed, and the animal was allowed to breathe room air spontaneously during the course of the experiment. The left femoral vein was cannulated for administration of fluid and anesthesia. Intravenous anesthesia (pentobarbital in normal saline, 4 mg/ml) was infused continuously with a syringe pump (Sage Instruments) at a rate of ~0.45 ml/h during the experiment. Supplemental intraperitoneal injections of 0.1 ml of 70% pentobarbital were administered as needed during dissections. The left cheek pouch was exteriorized, opened, cleared of connective tissue, mounted on the stage of a microscope (Zeiss), and observed with a ×20 (NA = 0.33) air (Zeiss) or ×20 (NA = 0.55) water immersion (Olympus) objective.

4

Labeling of capillaries. An 8 mM stock solution of di-8-ANEPPS (Molecular Probes) was prepared in dimethyl sulfoxide. Labeling solution was prepared by diluting stock solution in 3-(N-morpholino)propanesulfonic acid (MOPS)-buffered saline containing (in mM) 145 NaCl, 4.7 KCl, 2.0 CaCl₂, 1.2 MgSO₄, 1.2 NaH₂PO₄, 5.0 glucose, 2.0 pyruvate, 0.02 EDTA, 2.8 NaOH, and 2.0 MOPS to a final concentration of 12 µM. Pluronic F-127 detergent was added (0.1%) to increase the solubility of the dye. The final concentration (%vol) of dimethyl sulfoxide in the labeling solution was 0.12%. The dye solution was vortexed and stored in the dark until use. Micropipettes for dye perfusion were prepared from stock glass capillary tubes (1.2 mm in diameter; World Precision Instruments) pulled with a pipette puller (David Kopf Instruments) and tripled-beveled to ~5-µm tips. Cheek pouch arterioles and capillaries were labeled by perfusing dye into an arteriole with a micropipette using a picospritzer (PV-820, World Precision Instruments). During perfusion, the proximal segment of the arteriole was occluded with a micropipette with a rounded tip (~30 µm in diameter) and perfusion pressure was kept at the lowest level, usually 5-25 lb/in.², that would displace blood and label associated capillary networks. Time of infusion was ~15-20 min, and the amount of dye solution infused varied between 0.04 and 0.1 ml. Adequacy of dye staining was determined by periodic visual inspection of the preparation under epi-illumination. Time of dye infusion and the amount of vessel area dyed had greater impacts on the adequacy of capillary labeling than the amount of dye infused. When labeling was found to be adequate, the perfusion pipette and occluder were removed and blood flow resumed through arterioles and capillaries. After a 10-min equilibration period, studies were begun.

Drug application. Drugs were applied topically to short capillary segments by pressure ejection using a picospritzer (KCl, 140 mM in Ringer bicarbonate superfusate solution) or iontophoresis (1 M ACh or PE in distilled water). For KCl applications, a pipette containing drug and a second pipette containing vehicle solution were positioned next to a capillary segment by micromanipulators (Leitz), thus allowing easy selection of either vehicle or drug stimulus. Pressure ejection was accomplished without disturbing the capillary or the blood flow within it. For iontophoresis of ACh and PE, pipette tips were left unbeveled (size of tip < 1 µm). Ejection was driven by an iontophoresis programmer (3 s, 600 nA; World Precision Instruments). Antagonist studies were accomplished by adding atropine 10⁶ M for ACh or prazosin 10⁷ M for PE to the superfusate and allowing 5-10 min for stabilization before iontophoresis of drug.

Recording of membrane potential. Epi-illumination was restricted to capillary segments 75-150 µm long (Fig. 1) by adjusting the iris of a diaphragm in the light path. There was no fluorescence noted from nonvascular cells in these recording windows. Dye was excited using a 75-W xenon arc lamp filtered for excitation at a 475-nm center with 15-nm band pass. Emission wavelengths of 560 and 620 nm were selected using a 500-nm long-pass filter (Zeiss) and recorded using a dual-wavelength photomultiplier system (Nikon CCTV dual-image adapter and Hamamatsu PMT model 1104). Capillary segments demonstrating a steady fluorescence ratio (F₆₂₀/F₅₆₀) were selected for study. The fluorescence ratio was recorded for 10 s before drug was applied and for 30 s after application. In the case of KCl, results of some rapid sequential applications were recorded. In the cases of ACh and PE, 2-3 min were allowed to pass between each measurement. Changes in capillary endothelial membrane potential were assessed based on a prior calibration in which a change in F₆₂₀/F₅₆₀ (%R) of 1% represented an ~10-mV change (2). A negative value represented depolarization, and a positive value represented hyperpolarization. Raw data were recorded at 25 Hz with a 12-bit data acquisition board (Workbench, Strawberry Tree) and analyzed off-line.

View larger version (85K): [in this window] [in a new window]   Fig. 1.   A: capillary network labeled with di-8-ANEPPS. Apparent variations in dye loading of capillaries in same field are explained by different depths of capillaries within tissue or by variations in amount of dye perfusion of capillaries branching off different arteriolar sources. B: a capillary segment (75-150 µm) could be isolated for ratiometric study by adjusting a diaphragm iris over segment. Bars = 20 µm.

Data analysis. Statistical analysis was performed for paired sequential applications of KCl (same duration) on the same capillary. Mean (± SE) differences in %R, time of initiation of KCl application to time of initial response, and slope of depolarization over the first 3 s were determined for the first and second applications and compared using a paired two-tailed t-test, with significance denoted by P < 0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

KCl application. Pulsed applications of KCl (range: 0.5-11 s in duration) onto labeled capillary segments resulted in depolarization in 48 of 49 instances (17 capillaries, 5 animals) (Fig. 2). Among different capillaries, the maximum %R was quite variable and did not correlate well with the duration of the KCl application. Representative tracings of responses are shown in Fig. 3. In those instances in which application of the vehicle was paired with application of KCl, the vehicle had no effect on membrane potential (18 applications, 12 capillaries, 4 animals). In individual capillaries, the fluorescence change correlated well with pulse duration, and a steady state could be achieved after 8-11 s (Fig. 4).

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Distribution of change in fluorescence ratio F620/F560 (%R) obtained from KCl applications of varying time duration to capillaries.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   %R resulting from KCl application (bar denotes 5-s duration) to 6 different capillaries.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   %R from sequential applications of KCl (bars denote times of 2, 5, and 15 s) to same capillary. A steady-state change in F620/F560 is achieved at ~8 s during last application.

Acetylcholine and phenylephrine application. Application of ACh (57 applications, 22 capillaries, 6 animals) caused a hyperpolarization (Fig. 5A) that began immediately after application. The change in fluorescence corresponds to a maximal change in membrane potential of ~3.5 mV (0.35 %R) ~6 s after drug application. Membrane potential returned to baseline ~20 s after drug application. The fluorescence signal showed a slight, though not statistically significant, overshoot.

View larger version (18K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Fig. 5.   Mean (± SE) changes in membrane potential (Vm) in capillary endothelial cells after application of acetylcholine (ACh, n = 57 applications; A) and application of ACh in presence of atropine in superfusate (n = 18 applications; B). Time 0 indicates initiation of drug application, and bars indicate duration of application (3 s). Data points were generated every 0.8 s.

6

View larger version (18K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Fig. 6.   Mean (± SE) changes in membrane potential in capillary endothelial cells after application of phenylephrine (PE, n = 36 applications; A) and application of PE in presence of prazosin in superfusate (n = 22 applications; B). Time 0 indicates initiation of drug application, and bars indicate duration of application (3 s). Data points were generated every 0.8 s.

7

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

KCl, ACh, and PE were chosen for this study because each is known to induce conducted arteriolar responses (6, 9, 22, 23, 27, 28). These responses are thought to be due to changes in membrane potential originating in and/or conducted through endothelial or smooth muscle cells. KCl, ACh, and norepinephrine have also been reported to induce changes in arteriolar blood flow when applied to capillaries (7, 26), presumably from signals generated in the capillaries and then conducted upstream. Also, each of the drugs chosen has a potentially different mechanism of action for inducing changes in membrane potential in capillary endothelial cells.

KCl causes depolarization of endothelial cells by altering K⁺ distribution across the cell membrane and thus would be expected to elicit a change in membrane potential in capillaries. In arteriolar endothelial cells, depolarization and repolarization in response to KCl are characteristically rapid (2, 27, 28). In our experiments, responses in capillaries were slightly slower than those in arteriolar cells. The degree of depolarization was on the same order as that found in experiments by Miao and Joyner (18). They recorded changes in membrane potential in capillaries and venules in hamster mesentery using the fluorescent probe bis-oxonol. In their experiments, suffusion of capillaries and venules with a 100 mM K⁺ solution resulted in a 50% change in fluorescence (~8-9 mV depolarization), whereas suffusion and perfusion together of those vessels resulted in a 225% fluorescence change (~36 mV depolarization).

There were substantial differences in the rate and magnitude of KCl responses between capillaries or within the same capillary during different viewing periods in our experiments. This might be explained by a number of potential factors. First, attempts were made to avoid any significant disturbance of connective tissue or superfusate during application of drug to avoid altering the capillary image. Thus the injection pressure (and therefore the amount of drug applied) was necessarily adjusted from capillary to capillary to achieve this goal. Second, capillaries vary in their depths within tissue. Thus the amount of, and the rapidity with which, drug actually reached each capillary likely varied. Minor inconsistencies in washout of drug from superfusate flow may also have been present. Third, with the use of a voltage-dependent dye to measure changes in membrane potential, only a relative change in membrane potential from baseline can be determined, not the actual membrane potential. Thus variability in resting membrane potential between capillary segments, or variations in a particular capillary segment between viewing periods, could result in variations in the signal generated by drug application.

ACh binds to muscarinic receptors and activates a potassium current, which causes endothelial cell hyperpolarization and possibly increased intracellular Ca²⁺ (3, 20). This hyperpolarization of membrane potential in turn is conducted to surrounding smooth muscle cells through gap junctions (3, 21). Previous investigators have determined certain characteristics of endothelial response to ACh stimulation in in vitro studies. The magnitude of change in membrane potential in response to ACh stimulation may depend on the baseline endothelial cell resting potential (3, 16). In addition, depolarization of membrane potential after an initial hyperpolarization has been noted (16, 17). Finally, changes in membrane potential induced by ACh in endothelial cells recover slowly, and rapidly repeated stimulations result in responses that are significantly attenuated, if present at all, compared with the initial response (3, 4).

Our results are consistent with these previous findings in that the initial signals obtained were of a hyperpolarizing nature, with a subsequent slight depolarization. Recovery of the membrane potential to baseline was gradual. In addition, when attempted in preliminary studies, we could not reliably obtain repetitive responses to rapid, sequential applications of ACh as we could with KCl.

Phenylephrine acts primarily through adrenergic receptors, and there is evidence to suggest that some endothelial cells may possess ₁-receptors (13, 24). Some investigators have suggested that PE may induce production of endothelium-derived relaxing factor (1, 13, 21), although this could not be shown in isolated arterioles (28). Given the limited evidence for ₁-activity in endothelial cells, the possibility should also be considered that the capillary responses to PE may be secondary to an action on surrounding pericytes or perhaps nerves. Pericytes grown in culture have demonstrated depolarization in response to stimulation of ₁-receptors (12), and there is evidence that pericytes connect to capillary endothelial cells via gap junctions (11, 14, 25). In addition, in arterioles, changes in membrane potential generated in smooth muscle cells by PE are conducted to endothelial cells (28). From these collective findings, one could hypothesize that, in our experiments, PE may cause depolarization in pericytes which is conducted to capillary endothelial cells.

We found that typically membrane potential did not fully recover in the time that a capillary was observed after application of PE. Thus we did not attempt rapid, repeated stimulations of capillary segments with PE. A relatively long period of depolarization compared with duration of drug application may be characteristic of adrenergic binding sites on capillary endothelial cells or surrounding pericytes.

Because we were attempting to detect direct capillary responses, care was taken to minimize the possibilities that the changes in membrane potential observed resulted from stimulation of upstream arterioles with conduction back into capillaries. Capillary segments studied were a minimum of 125-200 µm from the nearest arteriole. We determined, by periodically applying drugs at various distances from arterioles under transillumination, that an arteriolar response due to diffusion over this distance was not present or was delayed on the order of 5-10 s. Multiple random observations under transillumination also revealed that either no change or an extremely delayed and minor change in the diameter of upstream arterioles occurred during application of drug. Multiple observations were also performed during the course of our experiments to determine that local application of drug had either no effect on blood flow within the capillary being visualized or that any changes in flow were delayed well beyond the time course of membrane potential signals being obtained. Time courses showed that changes in capillary membrane potential occurred immediately after application of drug (Figs. 5 and 6). These changes were faster than what might be expected from diffusion effects and suggest that changes in membrane potential were being generated directly from the capillary or pericapillary cells.

The recording windows for viewing capillary segments encompassed connective tissue and other cells in the extravascular space. Membrane potential signals could therefore have been recorded from nonvascular cells if significant amounts of dye had entered the cell membranes after leakage through the vascular wall during loading or from dye that had dissociated from capillaries over time. Because the fluorescence yield from the dye is greatly enhanced in the cell membrane (10), any labeled extravascular cells present in the recording window would be visible. Because we did not observe fluorescence from nonvascular cells either adjacent to capillaries or in the connective tissue within recording windows, we conclude that membrane potential responses originated solely from labeled capillary endothelial cells. The dye can be present in aqueous and nonmembrane compartments; however, this will contribute only a background intensity that is not voltage dependent. The effect of background fluorescence on arteriolar recordings has been addressed by Beach et. al. (2). In addition, nonspecific effects of KCl on the dye signal have been addressed by Loew and co-workers (10, 15, 19). The voltage response obtained was shown to follow a Nernst relation in lipid vesicles and cell cultures in response to changes in K⁺ concentration, and hence it is unlikely that KCl produced a nonspecific effect on the dye signals that we obtained in our experiments.

In summary, these experiments demonstrate the ability to detect changes in capillary membrane potential in vivo. These changes may be in either a depolarizing or hyperpolarizing direction, and they may be induced by stimuli that act either directly on capillaries or possibly on pericytes or nerves. Changes in membrane potential may be the means by which capillaries communicate with upstream vessels to optimize delivery of blood flow and oxygen to downstream tissue.