INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DESCENDING VASA RECTA (DVR) supply blood flow to the medulla of the kidney (25, 27). These microvessels arise from juxtamedullary efferent arterioles in the outer medulla and collect into vascular bundles as they descend toward the papilla. DVR on the periphery of each vascular bundle peel off to supply the outer medulla, whereas central DVR supply the inner medulla.

DVR are vasoactive and might regulate the distribution of blood flow within the renal medulla. ANG II constricts DVR isolated from rat kidneys and perfused in vitro, apparently by contracting smooth muscle-like pericytes surrounding the endothelium of these microvessels (24, 25, 30). This vasoactivity suggests that DVR could redistribute blood flow between the outer and inner medulla, given the anatomy described above (20, 27). The distribution of renal medullary blood flow may influence natriuresis and the long-term control of arterial pressure (7).

The electrophysiology of constriction of the efferent arterioles from DVR appears to be different from that in most types of vascular smooth muscle. Most descriptions of the actions of ANG II and other vasoconstrictors have shown that contraction of vascular smooth muscle involves depolarization of plasma membranes, often through activation of chloride channels (13, 17-19, 26, 27, 36). This depolarization promotes smooth muscle contraction by activating voltage-sensitive calcium channels in the plasma membrane and so increasing Ca²⁺ influx, even though it reduces the electrical force driving cation entry (1, 9). This seems true of smooth muscle in systemic vessels (17-19, 36) and renal afferent arterioles (5) and also of mesangium in glomeruli (22, 33). In contrast, ANG II constricts renal efferent arterioles without consistently depolarizing their smooth muscle (21) and in the presence of chloride channel blockers (6). We have examined how ANG II constricts isolated outer medullary DVR (OMDVR), using a voltage-sensitive dye, patch-clamp recording, and measurement of vasoactivity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation of OMDVR. Kidneys were removed from Sprague-Dawley rats (70-150 g; Harlan), which had been anesthetized with intraperitoneal thiopental sodium (50 mg/kg body wt). Renal slices were stored at 4°C in a solution of (in mM): 150 NaCl, 10 Na acetate, 5 KCl, 1.2 MgSO, 1.2 Na₂HPO, 1 CaCl₂, 5 HEPES, 5 L-alanine, and 5 D-glucose, plus bovine serum albumin (0.5 g/dl). OMDVR for imaging of fluorescence or vasoactivity were dissected in this solution, with and without voltage-sensitive fluorophore, respectively (24, 26, 29). OMDVR for patch-clamp recording were dissected in calcium-free solution and treated with collagenase to assist sealing. All OMDVR were transferred to a custom-built chamber on an inverted microscope (Nikon Diaphot), within several hours of kidney removal.

Fluorescence microscopy. Changes of membrane potential (m) in OMDVR were monitored by fluorescent microscopy of isolated, nonperfused microvessels exposed to bis[1,3-dibutylbarbituric acid-(5)] trimethineoxonol [DiBAC₄(3)] in the bath solution. DiBAC₄(3) is a lipophilic anion that becomes fluorescent on binding proteins or cell membranes (4, 10, 14, 35). Hyperpolarization reduces the fluorescence of cells exposed to DiBAC₄(3), and depolarization increases fluorescence by changing the intracellular concentration of the anionic dye. We included DiBAC₄(3) in the solution used to store renal slices before fluorescence investigations, because when OMDVR were exposed to DiBAC₄(3) for the first time during experiments, continuous loading rapidly increased fluorescence and obscured our observation of responses of m.

Fluorescence calibration. The relationship between m and DiBAC₄(3) fluorescence was calibrated in OMDVR exposed to gramicidin D, an ionophore for monovalent cations. m was calculated using the Goldman equation, by assuming that gramicidin D equalizes membrane permeability to K⁺ and Na⁺ (10, 14, 35),

(1)

where R is the gas constant, T is absolute temperature, F is Faraday's constant, and subscripted o and i denote concentrations outside and inside cells, respectively. Intracellular K⁺ concentration [K⁺]_i is assumed to be 100 or 150 mM and [Na⁺]_i to be 10 mM. Calibration solutions contained [Na⁺]_o (in mM) of 2.5, 5, 25, 50, 100, or 150, due to isosmolar substitution of N-methyl-D-glucamine chloride (Sigma, St. Louis, MO) for NaCl. Each calibration solution also contained (in mM): 4 KCl, 1 Na₂HPO, 1 MgCl₂, 1 CaCl₂, 1 HEPES, and 5 glucose, plus gramicidin D (2 mg/ml).

Limitations of fluorescence. Changes of m in OMDVR bathed in DiBAC₄(3) can be obscured by variations in fluorescence due to fluctuations in fluid level during solution exchanges. DiBAC₄(3) in the bath typically emitted background fluorescence of ~25% of the total signal, even without albumin. Exchange artifacts were minimized in two ways. First, ANG II was delivered by merging two streams. One stream provided bath flow to the chamber. A second pump was turned on at the required point in the experiment to deliver a solution of appropriate ANG II concentration (in bathing buffer) to be diluted into the first stream. The ratio of the flow of the two streams was 1:23. Second, when it was necessary to completely exchange the bath, bathing buffer was delivered from two 30-ml syringes whose plungers were driven simultaneously on a single syringe pump. The streams from the syringes were directed to either a collection flask or the perfusion chamber. By switching a stopcock, we could rapidly alternate the destinations of the two streams without affecting bath flow rate.

View larger version (80K): [in this window] [in a new window]   Fig. 1.   Fluorescence from isolated outer medullary descending vasa recta (OMDVR) in bis[1,3-dibutylbarbituric acid-(5)] trimethineoxonol [DiBAC4(3)]. A: images show vessels loaded with DiBAC4(3) from the bath. Corresponding white light and fluorescent images are provided for comparison. Fluorescence arises from both endothelium and pericytes (P). Upper two panels and lower two panels show different OMDVR. These unperfused vessels are ~10 µm in diameter. B: images show selective loading of DiBAC4(3) into pericytes on isolated OMDVR. A pericyte cell body was drawn into a holding pipette and held in place by gentle suction. The surface of the pericyte was superfused with DiBAC4(3) using a second internal concentric pipette. Top and bottom: corresponding white light and fluorescent images from a loaded pericyte cell body. The unperfused vessels are ~10 µm in diameter.

Patch-clamp recording. Whole cell m and single-channel activity were monitored by patch-clamp recording from pericyte cell bodies on isolated, nonperfused OMDVR at room temperature. These OMDVR had been treated to digest basement membrane and assist gigaohm sealing of patch pipettes onto pericytes (2, 31). Small wedges of renal outer medulla were separated from kidney slices in calcium-free PBS containing EGTA (in mM): 145 NaCl, 5.4 KCl, 1 MgCl₂, 0 CaCl₂, 10 HEPES, 10 glucose, and 0.1 EGTA. The tissue was incubated in the same solution for 1 h at 4°C and then transferred to PBS without EGTA for an additional hour. After this, the wedges were placed in calcium-free PBS containing papain (0.6 mg/ml) and dithiothreitol (0.5 mg/ml) for 20 min at 4°C. The tissue was then digested at 37°C for 20 min in calcium-free PBS containing collagenase 1A (0.5 mg/ml; Sigma), protease XIV (0.4 mg/ml; Sigma), and albumin (0.8 mg/ml). OMDVR from the digested tissue were stored in PBS containing 0.1 mM CaCl₂ and transferred to the cell bath on an inverted microscope for patch-clamp recordings.

Vasoactivity. Changes in diameter of OMDVR were monitored by white light microscopy of isolated, perfused microvessels and recorded on videotape (24). The inverted microscope supporting the chamber contained a beam splitter and a side port for attachment of a video camera (Panasonic WV-BL90) linked to a Panasonic model AG 1960 VCR with a microphone for voice recording. OMDVR were imaged using a ×40 objective, and the final magnification on the video screen was ×1,300. Internal diameters were measured using calipers at the position where the vessel constricted most. Diameter changes are expressed as percent constriction, given by [1  (D/Do)] × 100, where D is experimental diameter and Do is initial baseline diameter. The perfusion and bath solutions used for monitoring vasoactivity had the same composition as dissection solution (with CaCl₂). The pH was adjusted with NaOH to 7.55 at room temperature to yield a pH of ~7.4 at 37°C.

Reagents. ANG II, bovine serum albumin (A2153, Cohn fraction V), gramicidin D, collagenase 1A, and protease XIV came from Sigma. ANG II (10⁵ M) in water was stored in 100-µl aliquots at 20°C and diluted on the day of experiment. DiBAC₄(3) (5 mM; Molecular Probes, Eugene, OR) in DMSO was stored at 20°C in lightproof containers and diluted to a concentration of 4 µM in experimental solutions. The chloride channel blockers, indanyloxyacetic acid 94 (IAA-94) and niflumic acid, were also stored in aliquots in DMSO and diluted on the day of an experiment. Reagents were frozen and thawed once only. Excess reagents were discarded at the end of each experimental day.

Statistics. Except where otherwise specified, data in the text or figures are given as means ± SE. The significance of differences between means was calculated using Student's t-test (paired or unpaired, as appropriate) and repeated-measures analysis of variance. In the figures, straight lines through groups of data were drawn by regression analysis. In figures that show DiBAC₄(3) fluorescence, data were sampled every 2 s, but the error bars associated with most of the points have been suppressed for clarity.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DiBAC₄(3) calibration. The calibration of DiBAC₄(3) with gramicidin D is shown Fig. 2. Fluorescence increases with depolarization, as calculated from equation 1. The slope is 0.37%/mV, which resembles published values for this class of dyes (14, 35). Calculation of m from equation 1 requires [K⁺]_i. As illustrated in Fig. 2, predictions are insensitive to [K⁺]_i when it is taken to be between the reasonable limits of 100 and 150 mM.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Calibration of DiBAC4(3) fluorescence against membrane potential (m). OMDVR bathed in DiBAC4(3) and gramicidin D were exposed to solutions of varying NaCl concentration, rendered isosmotic by substitution with N-methyl-D-glucamine chloride (n = 6). Extracellular Na+ concentration ([Na+]o) was increased and decreased or visa versa in each OMDVR. The background-subtracted fluorescence (F) at each [Na+]o was averaged and normalized to the maximum observed fluorescence (Fo). m was calculated using the Goldman equation (equation 1), assuming an intracellular ([Na+]i) of 10 mM and a [K+]i of either 100 () or 150 mM (). The change in F/Fo with calculated m was 0.37%/mV with either assumption.

Effect of K+ and Cl alterations on resting potential. Increasing bath KCl concentration from 5 to 30 mM by isosmolar substitution for NaCl increased fluorescence, indicating depolarization of DiBAC₄(3) loaded cells (Fig. 3). When normalized fluorescence for sham-exchanged controls was subtracted, an ~18% increase in fluorescence resulted from KCl corresponding to a 48-mV depolarization. This must be taken as a weighted average for the endothelia and pericytes. If one assumes that K⁺ conductance is the sole determinant of m, then changing from 5 to 30 mM KCl in the bath yields a theoretical depolarization of 47.7 mV when [K⁺]_i is assumed to be either 100 or 150 mM. Chloride channel activity influences m in many cell types, including smooth muscle (19, 35) and endothelia (1, 23). To test for a possible role of Cl conductance in OMDVR, bath [Cl] was reduced from 150 to 30 mM by isosmolar substitution with Na acetate (Fig. 3). No significant change in fluorescent signal was observed in the absence of ANG II stimulation.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Effects of [K+]o and [Cl]o on fluorescence from OMDVR bathed in DiBAC4(3). The effect of changing [K+]o from 5 to 30 mM is shown along with sham exchange of controls (n = 7, each group). Increasing [K+]o increased fluorescence from OMDVR (P < 0.05 vs. 150 mM NaCl for all times >53 s), indicating depolarization, but lowering [Cl]o from 150 to 30 mM had no significant effect (n = 7). Fluorescence was measured by averaging for 2-s intervals.

Effects of ANG II on m. Compared with controls, abluminal application of ANG II (10⁸ M) from the bath increased the fluorescence of fully DiBAC₄(3)-loaded vessels by 8.0% (Fig. 4), corresponding to depolarization of 21.6 mV. When DiBAC₄(3) was selectively superfused onto the pericytes (Fig. 1B), ANG II (10⁸ M) also depolarized the cells by 9.2%, corresponding to depolarization of 24.9 mV. To corroborate the finding of pericyte depolarization, we measured the effect of ANG II on m by whole cell patch-clamp recording. In these experiments, performed at room temperature, ANG II (10⁸ M) depolarized the pericyte plasma membrane by an average of 29.4 mV, from a resting potential of 55.6 ± 2.6 to 26.2 ± 5.4 mV, P < 0.05 (Fig. 5).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effects of ANG II on DiBAC4(3) fluorescence. A: OMDVR bathed in DiBAC4(3) (Fig. 1) and exposed to abluminal ANG II (108 M) or vehicle from time = 0. ANG II increased fluorescence, indicating depolarization of pericytes and/or endothelium (P < 0.05 vs. controls for all time >56 s). N = 7 and 9 for controls and ANG II, respectively. B: pericyte cell bodies selectively loaded with DiBAC4(3) (Fig. 2) and exposed to abluminal ANG II (108 M) or vehicle from time = 0. ANG II increased fluorescence, indicating pericyte depolarization (P < 0.05 vs. control for all time >62 s). The fluorescence of control microvessels, also loaded by individual pericyte superfusion, did not change significantly over the duration of observation (n = 7).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Whole cell-permeabilized patch-clamp recording of pericyte m. Left: depolarization of a pericyte by bath ANG II (108 M). Right: m is shown as the resting value at time = 0 before and as the peak value after ANG II (108 M) exposure of 5 pericytes from different vessels (*P < 0.05 vs. ANG II).

Effect of chloride channel inhibitors on m. To determine whether ANG II-induced depolarization is mediated by Cl channels, the effects of IAA-94 on DiBAC₄(3) fluorescence were examined in ANG II-pretreated vessels. In support of the lack of effect of Cl substitution observed in Fig. 3, IAA-94 did not alter DiBAC₄(3) fluorescence in OMDVR that had not been exposed to ANG II. In contrast, IAA-94 repolarized ANG II-pretreated vessels, implying that Cl conductance contributes significantly to m when ANG II receptors are occupied (Fig. 6). We tested the effects of IAA-94 and the calcium-activated chloride channel blocker niflumic acid on vasoactivity, but we could not examine the effect of niflumic acid on DiBAC₄(3) because it seemed to destroy DiBAC₄(3) fluorescence.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Effect of indanyloxyacetic acid 94 (IAA-94) on DiBAC4(3) fluorescence. OMDVR bathed in DiBAC4(3) were treated with abluminal ANG II (108 M) or vehicle for 300 s before and during the period of observation. Groups were then exposed to the chloride channel inhibitor IAA-94 (3 × 105 M) or vehicle at time = 0 on the abscissa as labeled on the graph. IAA-94 repolarized OMDVR previously exposed to ANG II but had no effect in the absence of ANG II. P < 0.05 vs. vehicle for all time >143 s.

Effect of chloride channel inhibitors on ANG II-induced vasoconstriction. To test whether chloride channel activity is necessary for ANG II-induced vasoconstriction, the effects of two chloride channel blockers, IAA-94 and niflumic acid, were examined. In a first series, ANG II (10⁸ M) was introduced into the bath for 5 min, following which either IAA-94 (3 × 10⁵ M) or vehicle was added for an additional 5 min and then removed (Fig. 7A). As is typical for ANG II, vasoconstriction maximized and then waned somewhat in vessels treated with ANG II alone. In comparison, reversal of vasoconstriction occurred in vessels treated with IAA-94. Partial return to the vasoconstricted state was achieved after removal of IAA-94 from the bath, but washout was slow. Slow reversal of the effects of this lipophilic agent has been described by others (3). We also tested the effects of the chloride channel inhibitor niflumic acid. At both 1 mM (data not shown) and 30 µM, niflumic acid inhibited ANG II-induced OMDVR constriction (Fig. 7B). The reversal of ANG II-induced vasoconstriction by niflumic acid was rapidly eliminated with washout (Fig. 7B). Neither IAA-94 nor niflumic acid had significant effects on OMDVR that had not been preconstricted by ANG II (Fig. 7, A and B).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Effect of chloride channel blockers on constriction of OMDVR by ANG II. A: effect of IAA-94 on ANG II-induced constriction. OMDVR were exposed to abluminal ANG II (108 M) or vehicle from 2 to 17 min. OMDVR exposed to ANG II were then also treated with IAA-94 (3 × 105 M) or vehicle from 7 to 12 min (n = 5, each group). IAA-94 dilated OMDVR preconstricted by ANG II, it but did not affect control microvessels (n = 9). *P < 0.05, IAA-94 vs. vehicle. B: effect of niflumic acid (Nif Acid) on constriction of OMDVR by ANG II. OMDVR were exposed to Nif Acid (, 3 × 105 M, n = 8) or vehicle (, n = 9) from 2 to 5 min to assess the effects of Nif Acid on baseline diameters. Subsequently, the vessels were constricted with ANG II (108 M) and exposed again to Nif Acid () or vehicle () from 10 to 15 min. Nif Acid dilated vessels that had been preconstricted by ANG II, but it did not affect OMDVR before ANG II exposure (*P < 0.05, Nif Acid vs. vehicle).

Effect of Cl substitution on ANG II-induced vasoconstriction. To test further whether chloride conductance is important for ANG II-induced vasoconstriction, we enhanced the electrochemical driving force favoring Cl efflux from the pericytes by reducing the [Cl]_o of the bath from 150 to 30 mM. In these experiments, the collection end of the vessel was crimped to minimize perfusion and therefore any bath to lumen gradients of Cl. As shown in Fig. 8, the waning of vasoconstriction observed with ANG II in controls was completely reversed when extracellular Cl was reduced. Subsequent return of [Cl]_o to 150 mM at 15 min dilated the vessels.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Effect of [Cl]o on constriction of OMDVR by ANG II. OMDVR were exposed to bath ANG II (108 M) from 2 to 20 min. From 10 to 15 min, the bath was either exchanged to reduce [Cl]o from 150 to 30 mM by isosmolar substitution of Na acetate for NaCl or sham exchanged. To pressurize the lumen without inducing transmural gradients of [Cl], the collection end of each microvessel was crimped. Low [Cl]o reversibly increased vasoconstriction by ANG II (, n = 7), compared with sham exchange (, n = 9, *P < 0.05).

Single channels activated by ANG II in pericytes on OMDVR. Single-channel currents from cell-attached and -excised patches are illustrated in Fig. 9, A and B, respectively. ANG II activated these channels in pericytes on isolated OMDVR, during recording from cell-attached patches (Fig. 9, A and C). Single-channel currents in cell-attached and -excised patches are summarized in Fig. 10. In excised patches exposed to nearly symmetrical chloride concentrations in the pipette and bath and to Cs⁺ and Na⁺ as cations, reversal potential was near zero (Fig. 10, A and B). This is consistent with currents through chloride channels or poorly selective cation channels. Interestingly, this channel appeared to activate on excision without rundown over 10 to 15 min (Figs. 9B and 10C). In cell-attached patches, with CsCl in the pipette (but no Na⁺ or K⁺), currents also reversed near a pipette potential of zero (Fig. 10D). This indicates that Cl is the main conducting ion, because for [Cl]_i, ~20- to 30-mM reversal is expected to occur at approximately 50 mV, which is the resting m of these cells (Fig. 5). These channels were too rarely found in patches to permit a more detailed study of their properties. When observed, it was common to find two or three channels per patch.

View larger version (15K): [in this window] [in a new window]   Fig. 9.   Modulation of chloride channel from pericytes on isolated OMDVR. A: ANG II (108 M) in bath activates the channel in a cell-attached patch. Vertical and horizontal bars are 0.5 pA and 100 ms, respectively. B: patch excision into extracellular buffer also activates the channel. Vertical and horizontal bars are 2 pA and 500 ms, respectively. C: summary of the effects of ANG II on channel activity. With ANG II exposure, the probability of a channel being open (Popen) increased from 0.0104 ± 0.0044 to 0.0433 ± 0.0082 (n = 5, *P < 0.05). Popen was measured from 2-min records.

View larger version (21K): [in this window] [in a new window]   Fig. 10.   Single channels in OMDVR pericytes. A: channel activity is shown in excised patches for various pipette-holding potentials. Vertical and horizontal bars are 1 pA and 250 ms, respectively. B: single-channel currents in excised patches are shown (n = 3, means ± SD) as a function of transmembrane potential (Em). The pipette and bath contained CsCl and extracellular solutions, respectively (see METHODS). In these solutions, slope conductance is 11.0 pS. C: Popen determined from 2-min records is shown shortly after excision and at various times thereafter. D: single-channel currents in cell-attached patches in the absence of ANG II (n = 5, means ± SD) are shown vs. pipette-holding potential. In cell-attached patches, slope conductance is 16.8 pS.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Substantial evidence has been provided to show that ANG II exerts a tonic influence on perfusion of the renal medulla (8, 11, 32). Motivated by this, we have investigated the mechanisms involved in constriction of OMDVR by ANG II. Our data indicate that ANG II constriction involves depolarization of the contractile pericytes, at least, in part, through activation of chloride channels. Thus the electrophysiology of the pericytes resembles that of smooth muscle from most blood vessels. The evidence that ANG II depolarizes OMDVR is that it reduces fluorescence from microvessels loaded with DiBAC₄(3) (Figs. 1 and 4), a voltage-sensitive dye. The evidence that ANG II depolarizes pericytes is that it reduces fluorescence from pericyte cell bodies selectively loaded with DiBAC₄(3) (Figs. 1B and 4), and this is supported by patch-clamp measurements (Fig. 5).

ANG II depolarizes OMDVR apparently by activating chloride channels, because a chloride channel blocker inhibits this depolarization (Fig. 6). Chloride channels in OMDVR seem inactive without ANG II, because resting m is insensitive to [Cl]_o (Fig. 3) or IAA-94 (Fig. 6). At least some of these channels may be in pericytes, where patch-clamp recording reveals probable chloride channels barely active until stimulated by ANG II (Figs. 9 and 10). ANG II constriction of OMDVR apparently involves a depolarizing chloride current, because chloride channel blockers inhibit constriction (Fig. 7) and reduction of [Cl]_o to favor a depolarizing current enhancing constriction (Fig. 8). Given the incomplete reversal of vasoconstriction by chloride channel blockade (Fig. 7), other actions besides depolarization may be necessary for ANG II to induce OMDVR vasoconstriction.

This evidence has limitations. First, monitoring m by patch clamping requires enzymatic digestion to enable gigaohm sealing, and this might modify receptors or ion channels in pericytes. Second, DiBAC₄(3) fluorescence depends to an unknown extent on a contaminating signal from intracellular membranes. Third, to minimize background fluorescence, monitoring m using DiBAC₄(3) requires removal of albumin from the bath, and this may modify endothelial responses (12). Fourth, we do not know the contributions of pericytes and endothelium to fluorescence from OMDVR bathed in DiBAC₄(3) (Figs. 1 and 4). Even after selective loading of pericytes (Fig. 1B), some DiBAC₄(3) might diffuse to the endothelium.

Constriction of OMDVR by ANG II might involve depolarization of endothelium as well as pericytes. In endothelia, which generally do not express voltage-dependent calcium channels (1, 9), depolarization would reduce Ca²⁺ influx by reducing the electrical force driving cation entry. This would be expected to reduce endothelial [Ca²⁺]_i and to inhibit release of calcium-dependent vasodilators, such as nitric oxide and prostacyclin. We have recently provided evidence that ANG II reduces endothelial [Ca²⁺]_i in OMDVR and inhibits bradykinin and thapsigargin-induced [Ca²⁺]_i elevations (28).

If pericytes surrounding OMDVR contract when depolarized via activation of chloride channels, then they resemble many types of vascular smooth muscle. Chloride channels influence m in smooth muscle (17-19, 36, 37) and mesangium (22, 33) as well as in endothelium (23). Vasoconstrictors act on renal vascular smooth muscle by activating chloride currents and inducing depolarization (13, 34). Gordienko and colleagues (13) provided an extensive dissection of whole cell currents in smooth muscle cells isolated from larger resistance arterioles from the renal cortex. In CsCl-loaded cells, calcium influx occurred through voltage-regulated pathways and induced a secondary current carried by Cl. The latter was blocked by chelation of intracellular calcium or the chloride channel blocker DIDS.

Cl channel blockade with niflumic acid or IAA-94 reversed ANG II constriction of OMDVR (Fig. 7). Cl channel blockers seem to reverse constriction in renal arterioles as well. Jensen and colleagues (15, 16) found that DIDS had little effect on unstimulated vessels but reversed ANG II-induced afferent arteriolar constriction. With the use of the juxtamedullary nephron preparation and the isolated, perfused hydronephrotic kidney, respectively, Carmines (5) and Takenaka et al. (35) showed that IAA-94 blocked ANG II-induced afferent but not efferent arteriolar constriction. In our hands, reduction of [Cl]_o enhanced ANG II-induced constriction (Fig. 8). The effects of reducing extracellular chloride on ANG II-induced constriction of the afferent arteriole have also been investigated. Takenaka et al. (35) reduced external Cl from 114 to 51 mM and enhanced ANG II-induced constriction by 20%. Restoration of intracellular chloride to arterioles, previously depleted by incubation in gluconate, revived contractile responses to high [KCl]_o (15).

With the use of single-channel recording techniques, we identified a probable Cl channel in the plasma membrane of OMDVR pericytes (Figs. 9 and 10). Several of this channel's characteristics suggest that it mediates depolarization of pericytes by ANG II, although we recognize that this is not proven. First, the low open probability of this channel in the absence of ANG II stimulation (Fig. 10) mirrors the lack of effect of chloride substitution and channel blockers on resting potential (Figs. 3 and 6) and vessel diameters (Fig. 7). Second, ANG II activates this channel (Fig. 10) and depolarizes pericytes (Figs. 4 and 5). The difficulty in finding this channel in membrane patches has thus far made more extensive study of its control impractical.

Regulation of blood flow and microvascular pressures in the kidney is vital to enable fine control of glomerular filtration, salt and water handling, extracellular fluid volume, and blood pressure. Given this fact, it is perhaps not surprising that vasoconstrictors exert their actions on renal resistance vessels through various signal-transduction pathways involving varied ion channel architecture. Membrane depolarization through chloride channel activation has been described in the renal afferent arteriole (5, 13, 15, 16). In contrast, the efferent arteriole constricts normally in the presence of chloride channel blockers (5) and fails to depolarize when treated with ANG II (21). On this basis, the actions of ANG II in OMDVR more closely resemble those in the afferent arteriole, a result that contrasts with the origination of OMDVR from juxtamedullary efferent arterioles. Thus these data confirm that vasoconstrictors exert segment-specific actions not only in resistance vessels from the cortex, but also in DVR from the renal medulla.

In summary, the principal findings of this study are that constriction of OMDVR by ANG II involves activation of chloride channels in the plasma membrane of smooth muscle or pericytes. This leads to membrane depolarization that can be reversed by chloride channel blockade. OMDVR constriction is at least partly dependent on this process, because chloride channel blockers not only prevent depolarization but also reverse vasoconstriction. Finally, a chloride channel activated by ANG II has been identified in the pericyte plasma membrane and is likely to contribute to outward chloride currents and membrane depolarization.