INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

SEVERAL STUDIES have focused on the ability of the microvasculature to conduct vasomotor responses over distances of millimeters when vasoactive stimuli are locally applied (10, 17-20). More recently, conduction has been reported between capillary networks and arterioles. Dietrich and Tyml (6-8, 21) showed that changes in blood flow velocity and vasomotor activity in arterioles could be obtained during agonist stimulation within capillary networks and that these responses were not mediated by neural activity or by transport of the stimulant to the arterioles. Thus there is evidence that signals regulating flow of blood into capillary networks could arise from receptor-activated mechanisms in the capillary, possibly by interaction of metabolites with cell receptors (21).

Segal and Duling (18) hypothesized that conduction of vasomotor activity was mediated by communication between cells of the arteriolar wall and that the conducted responses coordinated the distribution of resistance in vascular segments. Subsequent studies showed that the mechanism of conducted vasomotor responses in arterioles is largely explained by passive electrical coupling between cells in the vessel wall (17, 19, 20). Because arteriolar endothelial and smooth muscle cells are electrically coupled (4, 16), and capillary endothelial cells are an extension of the arteriolar intima, it is possible that endothelial cell, coupling unites capillaries and arterioles.

The aim of this study was to show that electrical continuity between arteriolar and capillary levels of the microcirculation can be measured in situ. We used the fast voltage-sensitive dye 1-(3-sulfonatopropyl)-8-[-[2-(di-n-butylamino)-6-naphthyl]vinyl]pyridium betaine (di-8-ANEPPS) with dual-wavelength recording (2, 14, 15) to measure membrane potential responses in capillaries elicited by pulsed stimulation of proximal feed arterioles with potassium, phenylephrine (PE), and ACh.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animal preparation. Male golden hamsters (120 g) were anesthetized with pentobarbital sodium (70 mg/kg ip). The animal was tracheotomized, and polyethylene tubing was inserted into the bronchial tube to maintain an open airway. The left femoral vein was cannulated, and the anesthetic level was maintained by infusion of pentobarbital sodium in isotonic saline (10 mg/ml). The animal was mounted on a platform, and the left cheek pouch was exteriorized and pinned to a Plexiglas disk for microscopic observation. The preparation was continuously superfused (5 ml/min) with bicarbonate-buffered Ringer solution (37°C) composed of (in mM) 131.9 NaCl, 4.7 KCl, 1.2 Mg₂SO₄, 2.0 CaCl₂, and 20.0 NaHCO₃. We maintained the pH of the solution at 7.4 by bubbling it with a gas mixture of 5% CO₂-95% N₂.

Application of voltage-sensitive dye. The dye solution containing 12 µM di-8-ANEPPS (Molecular Probes) in MOPS-buffered saline was prepared by dilution of dye stock (8 mM dye in DMSO) in Ringer buffer with 5 mM MOPS with the same ionic composition as the superfusate. Pluronic F-127 was added to achieve a final concentration of 0.1% to increase solubility of the dye (2). The final concentration of DMSO was 0.12% of total volume. Regions confined to single second-order arterioles, the higher order branching arterioles and connected capillary segments, were labeled with each application of dye. After each vascular network was loaded, measurements were carried out for up to 1 h. Between one and three vascular networks were labeled during an experiment. Vessels were labeled by pressure ejection of the dye (12-30 psi) from a micropipette into an arteriole while blood flow was stopped by occlusion with a blunt pipette. Vessels were monitored under transillumination to insure stable perfusion of staining solution, and the micropipette pressure was adjusted to the minimum necessary to observe mixing of dye and blood in venules draining capillary beds. Dye perfusion was maintained for at least 15 min or until bright staining of capillary beds was observed by epi-illumination. After perfusion, blood flow was restored and the vessel tone was allowed to return to baseline.

Stimulation of arterioles and capillaries. Arterioles feeding labeled capillaries were stimulated with micropipettes positioned within 10 µm of the vessel wall. Figure 1 diagrams the relationship between stimulatory and recording sites and the microvascular tree. Potassium (140 mM) was applied by pressure ejection (8-22 psi, WPI PicoPump) from 5-µm tipped pipettes at sites ~150 µm proximal or distal to side branches connected to capillary networks. A similar micropipette filled with Ringer solution provided a vehicle control stimulation. The duration of stimulus pulses was 3 s, long enough to produce strong local constriction of the arteriole. For each recording, the spot epi-illumination was on for 90 s, during which one or two stimulus pulses were applied. Epi-illumination was turned off for 2 min between recordings, and up to three records were obtained for each vessel network. ACh and PE were applied by iontophoresis (300- to 600-nA ejection, 50- to 60-nA retention) using unpolished microelectrodes filled with 1 M solutions in distilled water. The stimulation duration was 0.3 or 3 s with 35 s allowed between pulses and 3 min between recordings. Possible effects of electrical current flow rather than agonist were checked by increasing the retain current and observing any resulting vasomotor activity. The strength of KCl and agonist stimuli were chosen to elicit local and conducted vasomotor responses along arterioles.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Relationship between stimulus and measurement sites used for study of capillary membrane potential responses in hamster cheek pouch. Stimulation micropipettes were positioned either proximally or distally to vessel feeding capillary network under study. Ringer pipette served as a vehicle control for conducted responses. Vessel networks were chosen such that direction between site of stimulation and measured capillaries ran counter to direction from application of superfusion solution to site of drainage. Circles mark typical recording sites in capillary networks.

Optical recordings. Dye fluorescence was excited with a 100-W xenon arc lamp (exciter: 475-nm center, 15-nm band pass; emission: 500-nm long-pass filter). This excitation intensity did not produce changes in arteriolar tone over 30-s exposures. Selected vessel segments were viewed using a water immersion lens (Olympus, ×20, NA 0.55), and the image was recorded with a Silicon-intensified target camera under transillumination. Measurements were made within 1 h after the preparation stabilized following dye loading.

Data analysis and calibration. Distance along the vascular tree between stimulus sites on arterioles and recording sites on capillaries, as well as the number of vessel branches between the two sites, was determined from videotaped images of labeled vessels viewed under epi-illumination. Averages over time of membrane potential responses for arterioles and capillaries were obtained from individual responses by using the stimulus marker as a reference point and averaging responses at each time point. The resulting averaged signal represents the waveform of the membrane potential change from multiple vessels and animals. F₆₂₀/F₅₆₀ changes were calibrated to membrane potential changes using a factor of 10% change in ratio per 100 mV. The voltage sensitivity of this technique in the arteriolar endothelium was determined previously (2). Before each day's measurements, responses of the dual-wavelength recording system to preset changes of 5 and 10% in the ratio of red versus green light were checked using a two-color LED calibration source positioned in focus under the microscope objective (1). The vessel electrical length constant was determined by fitting an exponential decay curve to plots of estimated membrane potential depolarization versus separation along vessels between stimulus and recording sites.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Labeling of vessel networks. Dye uptake in capillary endothelial cell membranes during intraluminal loading produced bright fluorescence with a low background intensity. Images obtained at two magnifications within 20 min (Fig. 2) of loading showed continuous staining along interconnected capillary segments. Figure 2, right, shows bright fluorescence from the edge of the capillary and weaker, out-of-focus fluorescence at the center coming from the vessel wall overlying and underlying the plane of focus. At this higher magnification, variations in brightness are caused by differences in the depth of the capillary network within the tissue relative to the focal plane. Figure 2, left, shows longer segments of labeled capillaries and larger collecting vessels. After dye loading, the intensity of capillary fluorescence decreased by up to 50% during the first hour of measurement, partly as a result of photobleaching and dissociation of dye from cell membranes. During this same period the background intensity increased slightly because of dye accumulation. Typically, the ratio of capillary to background light intensity remained above 100:1 over this period, enabling recordings of membrane potential changes to be carried out without significant drift in our calibration. Similar intense staining of arteriolar endothelium could be seen, and immediately after labeling there was no evidence of stain in the smooth muscle cells. Staining of smooth muscle cells did appear slowly over ~1 h after termination of dye loading, and tissue background light remained very low near arterioles upstream from capillaries over periods of several hours (2).

View larger version (105K): [in this window] [in a new window]   Fig. 2.   Di-8-ANEPPS-labeled microvessels in cheek pouch. Right: capillary fluorescence (×20 objective). Dye signals were recorded from capillary segments within a window of typical size shown by circle. Left: low-magnification image (×10 objective) showing stain pattern of longer vessel segments including capillaries (arrows) and collecting venules. Labeled arterioles are outside field of view. Calibration bars, 10 µm.

Conduction of potassium-induced depolarization. Potassium stimulation of FCS-perfused arterioles 0.5-1 mm proximal to capillary networks transiently depolarized the arteriolar endothelium. The first detectable membrane potential changes started within 1 s from the beginning of the stimulus (Fig. 3A, averaged from 6 vessels in 3 hamsters). The average maximal amplitude of the depolarization response was 21 mV. These electrical responses were followed by vasomotor responses that conducted in both directions from the site of stimulation. In capillary networks stimulation of arterioles produced a similar although attenuated depolarization of endothelial cells followed by a hyperpolarizing afterpotential (Fig. 3B, averaged from 17 recordings obtained from 12 capillaries in 5 hamsters). We were unable to determine any systematic change over the duration of the experiment in the amplitude of membrane potential changes from capillaries. Responses obtained from blood-perfused capillaries after direct potassium stimulation near the capillary segment (Fig. 3C, representative individual recording) caused almost the same degree of depolarization as was observed in the arteriole, although the peak response was delayed for several seconds (15).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Comparison of averaged vessel membrane potential changes in response to 140 mM potassium stimulation. A: response from FCS-perfused arteriole. B: conducted capillary response after stimulus of arteriole during normal blood flow. C: capillary response to direct application of potassium during normal blood flow. Beginning and end of stimulus are marked with vertical lines. Vertical calibration bar represents 10-mV membrane potential change and is equivalent to a change () in fluorescence ratio (R) of 1%.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Averaged membrane potential response in capillary obtained with 140 mM potassium pulse stimulation of arterioles. Top: early part of response. Dotted line, ensemble-averaged response; solid line, after smoothing with moving average filter. Middle: complete averaged response with depolarizing and hyperpolarizing components. Left y-axes in top and middle panels give membrane potential changes (Vm) in mV after applying a calibration of 10% R per 100-mV Vm; right y-axes give recorded changes in fluorescence ratios (F620/F560). Bottom: original recording without averaging, showing individual conducted membrane potential responses in capillary. Ratio changes correspond to depolarizations of 6.8 mV in both responses and hyperpolarizations of 9.7 and 4.5 mV for first and second response, respectively. y-Axis is fluorescence ratio. Stimulus times are indicated by horizontal lines.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Conducted capillary responses to different flow conditions. A: stimulus pipette proximal to branching arteriole (n = 12 vessels). B: pipette distal to branch (n = 10 vessels). C: an unaveraged response when feed arteriole was fully occluded (occlusion marked with arrows). Note hyperpolarization with occlusion. Stimulus times are indicated by horizontal lines. Vertical calibration bar represents 5-mV Vm and is equivalent to R of 0.5%.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Response of capillary red blood cell lineal density (means ± SE) after potassium stimulation of arterioles (n = 4 vessels). Stimulus time indicated by horizontal line.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Attenuation of conducted capillary depolarization versus distance from stimulus site. , Arteriolar response at stimulus site. Capillary responses () show an inverse relationship between amplitude and distance. An exponential decay curve fit (excluding data point for local arteriolar response) gives a 1/e attenuation distance () of 590 µm. Number of responses at each time is indicated below each point; number of vessel branches between the stimulus and recording site is shown above conducted capillary points. Values are means ± SE.

Conducted electrical responses to agonists. Agonists applied iontophoretically to arterioles (3-s pulse) also caused membrane potential changes in capillary segments. The change could be detected within 1 s after the start of the stimulus pulse. ACh, an endothelial M₁-receptor agonist, produced hyperpolarization, whereas PE, a smooth muscle ₁-agonist, produced depolarization. Agonist-induced responses recovered to the baseline membrane potential in ~25 s. Signal-averaged membrane potential responses are shown for both agonists in Fig. 8 (inset emphasizes early part of responses). Mean amplitudes and durations of agonist responses are included in Table 1. We did not observe vasomotor activity when iontophoresis stimulation with the opposite polarity and equal current strength was applied to arterioles.

View larger version (36K): [in this window] [in a new window]   Fig. 8.   Averaged membrane potential changes in capillaries from agonist stimulation of arterioles. Top trace: hyperpolarization with ACh. Bottom trace: Depolarization with PE. Drugs were applied by iontophoresis (300-600 nA, 3 s). Vertical calibration bar represents a 5-mV Vm and is equivalent to a R of 0.5%. Start and duration of stimulus pulse are marked by vertical and horizontal lines, respectively. Responses to both agonists recovered to prestimulus levels after 25-35 s. Early portions of responses are shown in inset. Response parameters are shown in Table 1.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study we have used an optical method for measuring membrane potential changes in response to vasoactive stimuli in the intact microcirculation. The sensitivity of this method enabled us to observe depolarizations in capillaries as small as 3 mV. The optical method permitted recording from capillaries in situ, a measurement difficult or impossible using microelectrode techniques.

The value we obtained for potassium-induced depolarization in FCS-perfused arterioles is in approximate agreement with previous measurements using microelectrodes (16, 24). There is no microelectrode measurement in situ of capillary depolarization with which to compare our results; however, our estimates using the dye method are consistent with attenuation of passive conducted membrane potential changes as described below.

Several observations and controls rule out the possibility that capillary responses after arteriolar stimulation were the result of direct action of drugs carried by superfusion solutions. Although the direction of superfusion flow could never be determined with certainty during capillary recordings, we could establish that convection of drugs did not produce vascular responses that could have led to the capillary membrane potential changes obtained several hundred micrometers from the site of arteriolar stimulation. First, we assumed that vasoconstriction by potassium is a reliable indicator for vessel depolarization that could be conducted toward capillaries. Our results showed that with our stimulus pulses, under all prevailing superfusion conditions, vascular responses were not obtained within 5 s when the distance between the vessel and the pipette exceeded ~120 µm. In our study the only arteriole located within this distance was immediately adjacent to the pipette. Second, our results showed that pipette application of potassium to the tissue surface far from an arteriole does not produce capillary depolarization over distances greater than ~60 µm from the measured capillary, indicating that the effect of convection on the membrane potential of capillaries occurs over a smaller area than that for vasomotor activity in arterioles. When surface stimulation was adjacent to the measured capillary, the beginning of depolarization was delayed no more than 1 s (Fig. 3C); however, the time to maximal depolarization was appreciably longer than when stimulus was applied next to the feed arteriole located several hundred micrometers from the capillary (compare Fig. 3, B and C). We believe that this additional delay is caused by the time needed for potassium concentration to increase deeper in the tissue (15). Any convection from the arteriolar stimulus site would have also required a similar time to penetrate the tissue to produce a direct depolarization. These two observations make it unlikely that arteriolar stimulation produced capillary depolarization via convection by the superfusion flow. Third, we note that potassium and agonist stimulation at arterioles elicited capillary responses in much shorter times than could be accounted for by diffusion of drugs. With the use of the Einstein-Smoluchowski relation for diffusion in two dimensions, D = 1/4d²/, where , D, and d are time delay, potassium ion diffusion constant, and diffusion distance, respectively, and assuming D = 1 × 10⁵ cm²/s (22), we calculate that the time required for potassium to diffuse over 300 µm is 22.5 s. This time was significantly longer than the subsecond delay observed for capillary responses.

The possibility that capillary responses arose from convection of potassium in the blood was also ruled out by two different experiments. Responses similar to those recorded during normal blood flow were obtained when blood flow was interrupted by vessel occlusion proximal to the site of drug application. During this procedure, occlusion per se also caused a rapid hyperpolarization. This hyperpolarizing response could represent the effect of reduced PO₂ on the capillary membrane potential because ischemia can cause tissue PO₂ to decrease significantly within seconds. Alternatively, this response may represent a conducted change in membrane potential from an arteriole after myogenic dilation induced by the occlusion. Future measurements may determine whether either of these mechanisms causes capillary hyperpolarization. To determine in the absence of any other membrane potential activity whether downstream capillary responses are influenced by blood flow, we resorted to a second test in which KCl was locally applied to arterioles supporting flow either toward or away from the capillary network (Fig. 5). Because capillary responses were similar in the two cases, we can exclude convection of potassium in the blood as a contributing factor.

It is possible that red blood cells directly affect the capillary fluorescence ratio if hemoglobin light absorption at 560 nm significantly reabsorbs fluorescence from endothelial cells located behind red blood cells. In our recordings, most of the light was collected from labeled cells on the side of the capillary, which are in the plane of focus (Fig. 2), and from cells overlying the capillary. For there to be any stimulus-induced change in the fluorescence ratio, a change in the density of circulating capillary red blood cells must occur. If the density were to decrease, as expected during vasoconstriction upstream, fewer red blood cells would be present in the recording window and reabsorption of dye fluorescence at 560 nm would decrease, causing a concomitant decrease in F₆₂₀/F₅₆₀. We observed no change in capillary hematocrit during the depolarizing component of the response and noted a transient decrease in capillary hematocrit that overlapped the early portion of the hyperpolarizing component of the response. Therefore, it is unlikely that the early phase of the ratio response was caused by fluorescence reabsorption by red blood cells. For the second phase of the conducted response, the ratio change is opposite to that which would occur from reduction in density of capillary red blood cells, and hence it is also unlikely that this portion of the response represents effects of red blood cells, although the magnitude of the response could have been reduced by this effect.

Two characteristics of capillary responses to arteriolar stimulation are consistent with conduction along vessels between the stimulus and recording positions. First, attenuation of the membrane potential depolarization with distance showed a characteristic length (distance to 1/e value) of ~600 µm. Although this distance is shorter than the length constant measured by Hirst and Nield (10) and later by Xia et al. (24) for arteriole-conducted depolarization in straight vessel segments, it is approximately the same as that noted by Segal and Nield (20) in arterioles near branch points from pooled electrophysiological data and is thus consistent with their model of passive electrical conduction in branched vessel networks. The lack of any relationship between the number of intervening vascular bifurcations and response size in our data may indicate that passive currents do not simply divide between each daughter vessel. Second, the biphasic membrane potential response obtained at the capillary in response to stimulation of the arteriole with potassium was similar to conducted responses recorded previously by Xia et al. (24) with microelectrodes from arterioles showing two distinct components for the membrane potential change.

The origin of the second hyperpolarizing component of the capillary response is not presently understood. Hyperpolarization of vascular cells has been demonstrated in the range of 1-3 mM extracellular potassium (5), which could account for our second component as potassium concentration falls back to that of interstitial fluid. However, the microelectrode recordings obtained near the site of pulse stimulation with a potassium micropipette do not show hyperpolarization after the initial depolarization (24), and this lack of hyperpolarization could result if potassium concentration decreased nonuniformly over the stimulated volume after the pulse. Microelectrode recordings have shown hyperpolarizing afterpotentials after initial potassium-induced depolarization in responses obtained over 500 µm from the site of stimulation (24). The lack of hyperpolarization at the stimulation site with appearance of hyperpolarization distal to the stimulation site would be consistent with a mechanism for hyperpolarization that competes with the high concentration of extracellular potassium in the stimulated volume, producing a net depolarization. As the response spreads from the stimulation site, the hyperpolarizing mechanism returns the membrane potential back to baseline more quickly and produces a transient hyperpolarization after the initial depolarization. In the arteriole the hyperpolarizing component is smaller than the initial depolarization, whereas in responses from capillaries these two components are of approximately the same amplitude. This observation is consistent with recent measurements in isolated arterioles showing that the length constant for conducted hyperpolarization is significantly longer than that for depolarization (9). The membrane mechanism of hyperpolarization may involve activation of either calcium-dependent potassium channels (11) or voltage-gated potassium inward rectifier channels (23); some may result from an unknown mechanism.

Stimulation of arterioles with ACh and PE produced capillary membrane potential responses of similar duration but smaller amplitude than responses measured previously in the arteriole (3, 13, 16, 24). In contrast, stimulation of arterioles with PE produced responses in capillaries that were significantly shorter than responses produced by direct capillary stimulation by PE (15). These data are consistent with different ₁-receptor-activated mechanisms on the arteriole and capillary and electrical continuity between the capillary endothelium and cells containing the receptor at both sites.

Our measurements demonstrate that membrane electrical changes in arterioles are coupled to responses in capillary segments. The pathway for coupling of the electrical signal through the arteriolar network could include either endothelial or smooth muscle cell layers (3, 13, 16). In capillaries a cellular pathway is likely to be the endothelium, although pericytes cannot be excluded. The simplest interpretation of our results is that electrical continuity within the endothelial cell layer extends from arterioles into capillaries, possibly through gap junctional connections present in these vessels (12).

Because this pathway would support spread of current in either direction along the vascular tree, our data suggest the possibility that membrane potential changes integrated over capillary networks in response to the local environment may influence blood flow by spreading to arterioles. A direct demonstration of arteriolar vasomotor response accompanied by conducted membrane potential changes in response to capillary stimulation would further test this hypothesis.