INTRODUCTION

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

THE MICROCIRCULATION OF THE kidney is regionally specialized. In the cortex, afferent and efferent arterioles govern the driving forces that promote glomerular filtration. A dense peritubular capillary plexus arising from efferent arterioles surrounds the proximal and distal convoluted tubules to accommodate enormous reabsorption of glomerular filtrate. In contrast, vasa recta serve needs specific to the medulla. Through the counterflow arrangement of descending (DVR) and ascending vasa recta (AVR), countercurrent exchange traps NaCl and urea deposited to the interstitium by collecting ducts and the loops of Henle. This is vital to maintain corticomedullary osmotic gradients but conflicts with the need to supply nutrient blood flow to medullary tissue. Metabolic substrates that enter the medulla in DVR blood diffuse to the AVR to be shunted back to the cortex. To deal with the threat of medullary hypoxia resulting from this process, the kidney has evolved a capacity to exert subtle control over regional perfusion of the outer and inner medulla. The details are far from clear, but much experimental evidence points to the complex interactions of many autocoids and paracrine agents to modulate vasomotor tone at various sites along the microvascular circuit. The goal of this review is to summarize recent insights into the control of medullary perfusion and the cell biology and mechanisms that govern vasa recta transport and vasoactivity.

	RENAL MEDULLARY MICROVASCULAR ANATOMY

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

Studies of regional perfusion have consistently shown that the fraction of total renal blood flow that is distributed to the inner cortex and medulla is subject to regulation (2, 17, 18, 28-30, 73-80, 88, 90, 110, 113, 180-185). To understand the sites at which such regulation might occur, microvascular anatomy will be briefly reviewed (Fig. 1). For greater detail, the interested reader is directed to a number of well-illustrated sources (6, 7, 48, 49, 55, 56, 60, 81, 85, 86, 99, 105, 110, 113, 163).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Anatomy of the medullary microcirculation. In the cortex, interlobular arteries arise from the arcuate artery and ascend toward the cortical surface. Juxtamedullary glomeruli arise at a recurrent angle from the interlobular artery. The majority of blood flow reaches the medulla through juxtamedullary efferent arterioles; however, some may also be from periglomerular shunt pathways. In the outer medulla, juxtamedullary efferent arterioles in the outer stripe give rise to descending vasa recta (DVR) that coalesce to form vascular bundles in the inner stripe. DVR on the periphery of vascular bundles give rise to the interbundle capillary plexus that perfuses nephrons (thick ascending limb, collecting duct, long looped thin descending limbs; not shown). DVR in the center continue across the inner-outer medullary junction to perfuse the inner medulla. Thin descending limbs of short looped nephrons may also associate with the vascular bundles in a manner that is species dependent (not shown). Inner medulla: vascular bundles disappear in the inner medulla, and vasa recta become dispersed with nephron segments. Ascending vasa recta (AVR) that arise from the sparse capillary plexus of inner medulla return to the cortex by passing through outer medullary vascular bundles. DVR have a continuous endothelium (inset) and are surrounded by contractile pericytes. The number of pericytes decreases with depth in the medulla. AVR are highly fenestrated vessels (inset). As blood flows toward the papillary tip, NaCl and urea diffuse into DVR and out of AVR. Transmural gradients of NaCl and urea abstract water across the DVR wall across aquaporin-1 water channels.

The medulla of the kidney is perfused by the efferent arteriolar blood flow that leaves juxtamedullary glomeruli. In the outer stripe of the outer medulla, juxtamedullary efferent arterioles give rise to many DVR (Fig. 1). There is also strong anatomic evidence that periglomerular shunt pathways give rise to some DVR (16). The inner stripe of the outer medulla is characterized by its separation into vascular bundles and the interbundle region. Vascular bundles contain all DVR destined to perfuse the interbundle region and those that eventually penetrate beyond the inner-outer medullary junction to the inner medulla. DVR on the bundle periphery give rise to a capillary plexus that perfuses the interbundle region, where metabolically demanding, salt-transporting epithelia of the thick ascending limb and collecting duct are located. DVR in the bundle center traverse the inner stripe of the outer medulla to reach the inner medulla. The latter may be larger and more muscular than peripheral DVR. Vascular bundles contain all AVR returning from the inner medulla and, to a degree that varies with species, short looped thin descending limbs of Henle (6, 7, 56, 60, 99, 110). Based on anatomic considerations alone, it seems evident that the vascular bundles place DVR and AVR into close apposition to favor efficient equilibration. They are also likely to be an important site for regulation of the regional perfusion of the outer and inner medulla because preferential vasodilation of DVR on the bundle periphery or constriction of DVR in the bundle center should enhance perfusion of the interbundle region.

DVR occupy a functional niche that is partially arteriolar and partially capillary in nature. The DVR wall is characterized by smooth muscle remnants that surround a continuous endothelium and impart contractile function. The pericytes persist into the inner medulla but eventually disappear (111, 120). Near their termination, DVR become fenestrated and give rise to a sparse capillary plexus. The plexus coalesces to form AVR that are characterized by a high degree of fenestration (117, 138). The fraction of the AVR wall occupied by fenestrations is larger in the inner medulla than in the outer medulla.

	ROLE OF DVR IN REGULATION OF MEDULLARY BLOOD FLOW

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

Regulation of blood flow to the renal medulla has been a subject of interest for decades. The methods used to measure blood flow and the conclusions derived from the associated studies have been frequently reviewed (17, 73, 76, 90, 105, 110). Laser-Doppler is now the dominant method for measuring regional perfusion of the kidney. That method relies on the Doppler shift imparted to monochromatic light by backscatter from moving red blood cells (RBCs) in a localized area of tissue. Many regional perfusion studies have been performed in animals that have had optical fibers acutely or chronically implanted into the renal parenchyma (2, 17, 28-30, 70, 73, 76). Studies of the extent of autoregulation of medullary perfusion with variation in renal perfusion pressure have produced variable results (17, 70, 76, 90), but much evidence points to a possible role for variation of medullary autoregulation with extracellular fluid volume status to regulate "pressure natriuresis" (17, 76, 78, 131). Laser-Doppler studies with implanted flow probes yield a signal from a fixed volume of tissue and thus the probes are sensitive to probe orientation. Direct measurement of DVR blood flow using dual-slit videomicroscopy has also been performed. Cupples and Marsh (19) found that flow in a single descending vasa rectum was regulated between 85 and 160 mmHg. The latter does not rule out the possibility that recruitment of flow might occur through previously unperfused DVR, as described by Roman et al. (see below) (131). In the latter case, single-vessel flow might be autoregulated while overall flow is not. Apart from the issue of regional autoregulation, a number of consistent themes have emerged. Blood flow to the medulla is dependent on the tonic vasodilatory influence of prostaglandins and nitric oxide (NO) (74, 75, 79, 83, 84, 88, 119). Blockade of renal medullary vasodilator synthesis leads to a reduction of medullary blood flow, salt retention, and hypertension. Thus, apart from the anticipated role of medullary blood flow to contribute to urinary concentration and water balance, a role in the regulation of sodium balance may also exist. The effector mechanisms that connect medullary perfusion to regulation of epithelial sodium transport are not well established, but various possibilities exist. It has been demonstrated that increases in perfusion pressure cause a secondary increase in renal interstitial pressure and that the latter leads to inhibition of proximal reabsorption (128). Some forms of hypertension are associated with regulation of sodium transport pathways (71), but this would not account for the immediate ability of increased perfusion pressure to cause saliuresis. If medullary perfusion modulates local NO release, a secondary effect on salt and water excretion could result (94, 95).

Given the importance of medullary perfusion in influencing salt and water balance, we were motivated to determine which location(s) along the microvascular circuit provides control of medullary blood flow. Afferent and efferent arterioles of juxtamedullary glomeruli could constrict and dilate to serve this purpose, but that hypothesis conceivably conflicts with the need for them to simultaneously control juxtamedullary glomerular filtration pressures. The possibility that periglomerular pathways for perfusing the medulla are important in this scheme has been proposed (16, 17). Based on anatomic considerations (Fig. 1), it seems probable that DVR are an important site of regulation. The latter has been impossible to test directly because the outer medulla and most of the inner medulla are inaccessible to observation in vivo. When DVR are isolated from rats and examined in vitro, contractile pericytes are observed and the vessels exhibit vasoreactivity, responding to numerous constrictors and dilators (Fig. 2). Vasoactive agonists include many paracrine agents that are synthesized within the medulla (104, 111-113, 129, 144-146). Because DVR are branches of efferent arterioles, it seems logical to conclude that their vasomotion could affect glomerular filtration pressures; however, compensatory changes in afferent arteriolar tone could hypothetically offset that effect with the response to signals arising from tubuloglomerular feedback and myogenic response. The details are uncertain, but it is likely that DVR are an important regulator of medullary perfusion.

View larger version (46K): [in this window] [in a new window]   Fig. 2.   Vasoconstriction of isolated, microperfused outer medullary DVR. A: DVR isolated and microperfused in vitro is exposed to ANG II (10 nM) by abluminal application from the bath. a and b, Vessel before and after constriction, respectively. Two cell types can be seen. Pericyte cell bodies project from the abluminal surface, and endothelia line the lumen. B: quantification of DVR constriction through measurement of luminal diameter. Results are expressed as %constriction = 100 × (Do  D)/Do, where Do is basal diameter and D is diameter after constriction. The mean luminal diameter of perfused DVR is ~14 µm. Constriction has been induced by abluminal exposure to endothelin 1 (0.1 nM, n = 6), ANG II (10 nM, n = 15) or by raising of extracellular K+ concentration from 5 to 100 mM by isosmotic substitution for NaCl (n = 6). Data are reproduced from Refs. 104, 146, and 176.

The parallel arrangement of DVR within outer medullary vascular bundles raises two questions. First, does global constriction of DVR regulate the rate at which blood crosses from the cortex to the medulla, i.e., influence total medullary perfusion? Second, does DVR vasomotion within vascular bundles regulate distribution of blood flow between the interbundle region of the outer medullary inner stripe and the inner medulla? Those actions are not mutually exclusive. Some quantitative and qualitative observations are germane. Roman and colleagues (131) reported that variation in inner medullary tissue perfusion is related both to changes in blood flow through single vessels and to the recruitment of blood flow into previously nonperfused DVR. Given that the luminal diameter of DVR is similar to that of RBCs, it seems plausible that flow through them could be stopped by intense foci of constriction in the outer medulla. If that occurs, excessive back-pressure might not result because blood from the efferent arteriole could still traverse adjacent vessels that lie in parallel within the vascular bundles (Fig. 1). This possibility is qualitatively supported by the observation that bolus spurting of RBCs through the lumen of DVR can sometimes be seen on the surface of the papilla (inner one-third of the inner medulla) when the papilla is exposed for micropuncture (Pallone, unpublished observations). Goligorsky and colleagues (34, 61) have postulated that membrane fluidity changes related to local production of NO contribute to such processes. Finally, differential regulation of outer vs. inner medullary perfusion by vasopressin has been observed when optical laser-Doppler probes were inserted into the medulla at various depths (28).

	VASOACTIVITY OF DVR

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

When DVR are isolated from outer medullary vascular bundles, perfused in vitro, and exposed to contractile agonists, they constrict at various foci along their length (Fig. 2). Of the various agents thus far examined, endothelin-1 and -2 yield the most intense and durable constriction (146). They have threshold effects at 10¹⁴ M and most often obliterate the vessel lumen at higher concentrations (100% constriction, Fig. 2). These agents induce durable constriction that does not wane and is only slowly reversible after washout. ANG II is also a consistent constrictor that, on average, reduces luminal diameter of microperfused vessels by 40- 60% (104, 114, 129). As in cortical vessels, thromboxanes are partially responsible for mediation of ANG II DVR constriction (140, 143, 164, 165). In contrast to endothelins, vasoconstriction by ANG II tends to maximize 5-10 min after application and then slowly wane toward a stable baseline (Fig. 2). Microperfused DVR seem to exhibit a minimum of intrinsic tone and to show no myogenic activity when pressurized (104). It must be recognized, however, that these observations are obtained from isolated vessels placed into artificial buffers without supporting interstitum, often hours after the death of the rat.

Pericytes are the smooth muscle remnants that surround DVR and presumably impart contractile function. Recently, the mechanisms by which ANG II induces vasoconstriction have been evaluated using fluorescent probes of intracellular Ca²⁺ concentration ([Ca²⁺]_i) and membrane potential and by electrophysiological recording. As expected for signaling via the ANG II AT₁ receptor, a classic peak-and-plateau intracellular Ca²⁺ response is elicited in fura 2-loaded pericytes (Fig. 3) (130, 176). Both whole cell electrophysiological recording and measurements with a potentially sensitive fluorescent probe showed that ANG II depolarizes the pericyte cell membrane (Fig. 4) (107, 130, 175, 176). This seems to occur through activation of a Ca²⁺-sensitive Cl conductance that shifts membrane potential away from the equilibrium potential of the K⁺ ion toward that of Cl (107, 175). An 11-pS Cl channel has been identified in the pericyte cell membrane. This channel has low basal open probability but is activated by ANG II or excision into high-Ca²⁺ buffers (Fig. 5). Membrane potential of ANG II-treated pericytes often oscillates, and voltage-clamped cells held at 70 mV exhibit classic spontaneous transient inward currents typical of various smooth muscle preparations (Fig. 4C) (35, 43, 46, 91).

View larger version (61K): [in this window] [in a new window]   Fig. 3.   Measurement of intracellular Ca2+ transients in DVR pericytes. A: appearance of an isolated DVR after exposure to collagenase. B: isolated collagenase-treated vessel has been drawn into a glass micropipette with a heat-polished opening of ~6 µm diameter, stripping pericytes from the abluminal surface. This process can be continued to isolate a group of pericytes for loading with the Ca2+-sensitive fluorescent indicator fura 2 (130). Bar = ~10 µm. C: intracellular Ca2+ concentration ([Ca2+]i) response of fura 2-loaded DVR pericytes to ANG II (10 nM) is shown in the presence and absence of diltiazem, n = 6 and 7, respectively. In the control group, diltiazem was added to the bath for 11-16 min. Diltiazem inhibits the plateau phase of the pericyte Ca2+ response. Data are reproduced from Refs. 130 and 176.

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Nystatin-perforated-patch whole cell electrophysiological recordings from DVR pericytes. A: membrane potential recorded from pericytes as they are exposed to ANG II (10 nM) for 2-10, 2-3, or 20-25 min. Data were sampled at 10 Hz, averaged to 1 Hz, and then averaged for n = 6 cells/group. ANG II depolarizes the cells but cannot be reversed after prolonged exposure. B: membrane potential oscillations in a pericyte exposed to 10 nM ANG II. C: spontaneous transient inward currents in a DVR pericyte exposed to ANG II. The cell is held at 70 mV and then exposed to ANG II (10 nM, arrow). Data are reproduced from Ref. 107.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   11-pS Cl channel recorded from DVR pericyte. A: recording showing channel activity in excised patch held at potentials shown on the right. O and C , open and closed, respectively. B: excision of patch into 1 mM Ca2+ buffer activates the channel. ANG II activates this channel in cell-attached patches (data not shown). Data are reproduced from Ref. 175.

The role of membrane depolarization and changes in Cl conductance are well established in the afferent arteriole (13, 39, 47, 63) as a means of gating Ca²⁺ entry (14, 15, 64). Until recently, however, the existence of voltage-gated Ca²⁺ entry pathways in the efferent circulation has been uncertain. Rigorous examination of this issue with RT-PCR, immunochemistry, and the assessment of vasoreactivity in isolated arterioles has been reported by Hansen and colleagues (38). T-type subunits, Ca(V)3.1 and Ca(V)3.2, and an L-type subunit, Ca(V)1.2, were found in efferent arterioles of juxtamedullary glomeruli but not superficial glomeruli. These subunits were also identified in DVR (38). Based on this finding, membrane potential of DVR pericytes is expected to control voltage-gated Ca²⁺ entry and modulate [Ca²⁺]_i, a prediction that received recent experimental support. The L-type channel blocker diltiazem vasodilates ANG II-constricted DVR and reduces [Ca²⁺]_i of ANG II-treated pericytes. Both high external K⁺ concentration and the L-channel agonist BAY K 8644 are weak DVR vasoconstrictors. Finally, agents that repolarize pericytes, bradykinin (BK) and the ATP-sensitive K⁺ channel opener pinacidil, are effective vasodilators (Fig. 6) (176). The many downstream effects of pericyte ANG II receptor activation remain unknown; however, there are hints that important effects result from actions independent of [Ca²⁺]_i elevation. Principally, depolarization in the absence of agonist induces far less intense constriction than does ANG II or endothelins (Fig. 2). Phosphorylation events that sensitize the intracellular contractile machinery to the effects of Ca²⁺ are likely to be implicated (133).

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Repolarization of ANG II-depolarized pericytes by vasodilators. A: recording of membrane potential from a DVR pericyte successively exposed to ANG II (10 nM) and bradykinin (100 nM). Resting membrane potential of 48 mV depolarizes after ANG II. A biphasic repolarization occurs after exposure to bradykinin. B: similar recording from a DVR pericyte exposed to ANG II (10 nM) and then the ATP-sensitive K+ channel opener pinacidil (10 µm). Both bradykinin and pinacidil repolarize pericytes and vasodilate preconstricted DVR (109, 176). Data are reproduced from Ref. 176.

	ROLE OF DVR ENDOTHELIA IN THE REGULATION OF VASOACTIVITY

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

As a fortuitous consequence of the fact that the Ca²⁺-sensitive fluorophore fura 2 loads almost exclusively into DVR endothelia (to the near exclusion of pericytes), it has been relatively easy to examine global Ca²⁺ transients generated by endothelium-dependent vasodilators (Fig. 7). As expected, BK generates a peak-and-plateau Ca²⁺ response, enhances NO generation (Fig. 8), and induces vasodilation (112, 129, 130). An unexpected finding is that the vasoconstrictor ANG II suppresses basal Ca²⁺ and inhibits BK-, acetylcholine-, thapsigargin-, and cyclopiazonic acid-induced Ca²⁺ responses in DVR endothelia (Fig. 7, B and C) (114, 130). This is surprising for several reasons. First, AT₁ receptors, which mediate the vast majority of the effects of ANG II, signal through inositol 3,4,5-trisphosphate generation and Ca²⁺ mobilization. Second, infusion of ANG II has been observed to lead to secondary enhancement of NO levels within the medulla (182, 185) and in isolated cortical microvessels (153, 154). Given that endothelial nitric oxide synthase (eNOS)/NOS3 is a Ca²⁺-dependent isoform of NOS, suppression of Ca²⁺ would be expected to block rather than enhance endothelial NO generation. As a possible answer to this paradox, we pointed out that adjacent nephrons also express NOS isoforms, so that ANG II elevation of Ca²⁺ in those structures would favor NO generation on the vascular bundle periphery. Hypothetically, this could provide a feedback loop (in addition to adenosine) through which the medullary thick ascending limb (mTAL) can regulate its own perfusion. As previously discussed, ANG II might suppress DVR endothelial Ca²⁺ signaling as a means of turning regulation of DVR vasomotion away from the endothelium to the mTAL (114, 130).

View larger version (43K): [in this window] [in a new window]   Fig. 7.   Measurement of [Ca2+]i transients in DVR endothelia. A: white light and fluorescent images from an isolated DVR loaded with fura 2. Abluminal pericytes (arrows, left) are not visible in the fluorescent image (right). In contrast, endothelial cells load fura 2 and emit a strong fluorescent signal. B: example of DVR endothelial [Ca2+]i response after exposure to bradykinin (BK; 100 nM). C: means ± SE of n = 7 DVR exposed to ANG II (10 nM). Slight suppression of [Ca2+]i is seen. D: suppression of DVR endothelial [Ca2+]i after exposure to ANG II (10 nM) is dramatic when [Ca2+]i has been previously increased by exposure to the sarcoplasmic endoplasmic reticulum Ca2+ ATPase inhibitor cyclopiazonic acid (CPA; 10 µm). Data are reproduced from Refs. 112, 114, and 130.

View larger version (39K): [in this window] [in a new window]   Fig. 8.   Measurement of nitric oxide (NO) generation in isolated DVR using 4,5-diaminofluorescein (DAF)-2. A: white light and fluorescent image shows that DAF-2 loads into both endothelia and pericytes. B: fluorescent emission at 535 nm during excitation at various wavelengths (abscissa). Successive recordings are obtained at 1-min intervals from DVR exposed to 2.5 mM sodium nitroprusside (SNP) as a NO donor. Fluorescence increases without a shift in spectra. C: DAF-2 emission reflecting endogenous NO production in isolated DVR. In the control group, fluorescence declines due to leakage of DAF-2 from the cytoplasm. Fluorescence is greater on exposure to BK (100 nM) and increases further when the bath contains the superoxide dismutase mimetic tempol (1 mM). Data are reproduced from Ref. 129. **P < 0.01, *P < 0.05 vs. control.

The concentrations of ANG II required to influence endothelial Ca²⁺ signaling or to maximally constrict DVR in vitro (nanomolar) exceed circulating levels (picomolar). Compartmental ANG II concentrations within the kidney can be as high as 10⁹ to 10¹⁰ M (89, 142). The renal outer medulla is inaccessible to micropuncture, so that direct sampling of outer medullary DVR plasma directly downstream of juxtamedullary glomeruli cannot be performed. Seikaly and colleagues (142) found that ANG II concentrations in star vessel plasma sampled downstream of superficial glomeruli exceed circulating levels by as much as 1,000-fold. On this basis, it seems possible that DVR could be exposed to nanomolar ANG II in vivo.

The vasodilatory influence of NO cannot be completely understood without considering its interaction with O₂ free radicals. These are generated by one-electron reductions of O₂ to generate superoxide (O·), hydrogen peroxide (H₂O₂), hypochlorous acid, and hydroxyl radical (·OH), the "reactive oxygen species" (ROS). ROS favor vasoconstriction and have been implicated in various forms of hypertension (53). Mechanistically, this is at least in part because O· reacts with NO to form peroxynitrite (ONOO), a product that, compared with NO, is a weak vasodilator. ROS are generated by the "leak" of electrons from the mitochondrial electron transport chain as well as a variety of enzymatic processes. Intrinsic mechanisms limit cellular levels of ROS. Several isoforms of superoxide dismutase (SOD) convert O· to O₂ and H₂O₂. In turn, H₂O₂ is decomposed to O₂ and H₂O by catalase and other peroxidases. By limiting reaction of NO with O·, the extracellular isoform of SOD found in plasma and endothelia has been identified as a principal regulator of NO bioavailability (62, 96). Endogenous antioxidants are also responsible for scavenging ROS. For example, hemeoxygenase (HO) is a microsomal enzyme that degrades heme. In the process, it forms CO, a vasodilator, and biliverdin, an antioxidant (27, 180). Both HO-1 and HO-2 isoforms are expressed in renal smooth muscle and nephrons (4, 40, 45).

Renal generation of ROS favors vasoconstriction and may contribute to some forms of hypertension. Tempol is a cell-permeant SOD mimetic that reduces hypertension in the spontaneously hypertensive rat (139, 141). Of the several sources of O·, NADPH oxidase appears to be important. Some NADPH oxidase subunit isoforms are upregulated and activated by vasoconstrictors (36, 57, 126, 159, 165, 177). It has been shown that both ANG II receptor blockers (ARBs) and the combination of hydrochlorothiazide, hydralazine, and reserpine (triple therapy) can normalize blood pressure in the SHR rat; however, only ARBs reduce excretion of ROS reaction product 8-iso-PGF₂. Interestingly, PO₂ values are lower in the SHR, and this too is normalized by ARB therapy or treatment with the SOD mimetic tempol (1, 164). Tempol also has been shown to enhance medullary perfusion (183). NO production by isolated DVR and the mTAL is enhanced by tempol, and this agent blunts ANG II-induced DVR vasoconstriction (94, 129). Given the importance of medullary blood flow in the regulation of blood pressure, it is inviting to speculate that some forms of hypertension might be related to an increase in "oxidative stress" in the renal medulla.

	MEDULLARY PO2 AND PERFUSION OF THE MEDULLA

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

PO₂ in the medulla of the kidney is low, in the range of 10-25 mmHg (11, 12, 22). This is predicted to be a consequence of the countercurrent arrangement of vasa recta (173) because O₂ in DVR blood diffuses to AVR to be shunted back to the cortex. Evidence supports the notion that O₂ consumption by the salt-transporting mTAL makes it vulnerable to ischemia (12). Several hormonal systems play a role in the protection of the medulla from ischemic insult. Each shares the ability to enhance medullary blood flow and inhibit salt reabsorption along the nephron. Hypothetically, this should have a dual effect of enhancing the supply of O₂ and simultaneously reducing the demand for its consumption. A first example is the generation of vasodilatory prostaglandins through activation of cyclooxygenase (COX) (58). It has been observed that perfusion of the renal medulla is sensitive to COX inhibition (110, 113, 119). Furthermore, renomedullary interstitial cells have receptors for ANG II and BK and release PGE₂ in response to these agents (21, 72, 178, 179, 186). Recently, Qi et al. (127) have shown that the COX-2 isoform is expressed in those cells and is responsible for this action. In their study, ANG II reduced medullary blood flow in COX-2-, but not COX-1-, deficient mice (127). The ability of PGE₂ to promote natriuresis may be explained in part by the ability of PGE₂ to inhibit Cl reabsorption by the mTAL (41, 146). PGE₂ is a vasodilator of in vitro perfused DVR (104, 146).

In most vascular beds, ischemia favors generation of adenosine, a paracrine agent that enhances blood flow through local vasodilation. The actions of adenosine in the renal cortex are unusual because it induces vasoconstriction, accompanied by a reduction of glomerular filtration rate (2, 90). Within the medulla, however, adenosine acts as a vasodilator and inhibits salt reabsorption by the mTAL (2, 184). This presumably serves to reduce O₂ consumption while enhancing O₂ delivery (2, 22). It is a reasonable hypothesis that adenosine produced by the mTAL diffuses to and dilates outer medullary DVR on the periphery of vascular bundles. Outer medullary DVR on the bundle periphery supply the mTAL with blood flow, so such a mechanism represents a feedback system that would protect the mTAL from hypoxia. The mTAL has the capacity to produce adenosine (9) and that A₁ and A₂ receptor mRNA is expressed in DVR (54). Adenosine A₁ and A₂ receptor stimulation favors DVR vasoconstriction and vasodilation, respectively (144, 145).

Studies have shown that renal medullary NOS activity and NO production exceed that in the cortex (74, 75, 79, 80, 87, 166, 167, 182). Evidence is accumulating that NO acts in an autocrine and paracrine fashion to modulate both vasomotor tone and epithelial NaCl reabsorption. Inhibition of NOS in the renal medulla has isoform-specific effects. NOS1 inhibition reduces NO levels in the medulla and induces salt-sensitive hypertension but fails to alter medullary perfusion (50, 74, 77). Global inhibition of NOS1, NOS2, and NOS3 isoforms with nonselective blockers decreases medullary NO levels and induces salt retention and hypertension. In addition, global NOS inhibition reduces medullary blood flow and tissue oxygenation (11, 42, 78, 124). NO generation may be important to abrogate tissue hypoxia that would otherwise arise from release of vasoconstrictors. ANG II, norepinephrine, and vasopressin stimulate release of NO in the medulla (121, 150, 151, 181, 185). Subpressor infusion of N^G-nitro-L-arginine methyl ester into the renal interstitium does not affect medullary blood flow or PO₂ but enables otherwise ineffective doses of ANG II (185) norepinephrine (151, 181), or vasopressin (150) to induce a fall in these parameters. Taken together, the data support the conclusion that medullary NO production has a tonic effect in maintaining perfusion and protecting the medulla from ischemic injury. In addition to vascular effects, NO inhibits solute and water reabsorption in the collecting duct and thick ascending limb (31-33, 95, 122, 123, 148, 149). Studies in knockout mice implicated NOS3 as the isoform responsible for autocrine stimulation of NO in the thick ascending limb (122). Thus, like prostaglandins and adenosine, NO is a vasodilator that also inhibits salt reabsorption and therefore O₂ consumption in the thick ascending limb.

	TRANSPORT OF SOLUTES AND WATER ACROSS VASA RECTA

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

DVR occupy a functional niche that is partially that of a vasoactive arteriole and partially that of a transporting microvessel. As such, it is not surprising that evidence has emerged to show that DVR endothelia have specialized characteristics that reflect the requirements endowed by their anatomic location and dual role. DVR have a continuous endothelial lining and zona occludens. It is expected that the paracellular pathway conducts diffusive transport of NaCl and other small hydrophilic solutes (82, 109, 158). In addition to this, transcellular pathways have been identified that conduct transport of urea and water (51, 52, 92, 93, 97-99, 106, 108, 117, 118). The principal transport functions of vasa recta will be briefly discussed. For a detailed explanation and historical perspective, the reader is referred to other reviews (25, 110).

The first evidence that DVR endothelia express a carrier for urea was that individual vessels, perfused simultaneously with [²²Na]- and [¹⁴C]urea, often exhibited low or moderate Na⁺ permeability but always had very high permeability to urea (117). This was surprising because the diffusivities of Na⁺ and urea in water are nearly identical, so that transport via diffusion in an aqueous pore predicts identical permeabilities to these tracers. The ability of phloretin and urea analogs to reduce DVR urea permeability supported the existence of a carrier in DVR endothelia, implying a major contribution of transcellular, facilitated diffusion to overall urea transport (97, 117). That urea carrier was eventually identified as the one expressed by the RBC (UTB), a form that is distinct from the vasopressin-sensitive and -insensitive splice variants of the epithelial carrier (UTA1-UTA4) (5, 8, 44, 51, 52, 125, 134, 135, 155-157, 168, 170). Taken together, the expression of epithelial, endothelial, and RBC-facilitated carriers seems to ensure that urea diffusing from AVR plasma will be efficiently recycled by diffusing into thin descending limbs and DVR. Yang and Verkman (171, 172) have shown that UTB exhibits water channel activity and that UTB conducts water across the aquaporin-1 (AQP1)-deficient RBC membrane. A role for UTB in the transport of water across DVR has not been established; however, it is notable that urea, glucose, and raffinose can drive substantial water flux across AQP1-deficient DVR (106).

It has long been known that DVR plasma protein concentration increases with distance into the renal medulla due to water efflux from the DVR lumen to the papillary interstitium. That transport proceeds in a direction opposite to inwardly directed Starling forces (hydraulic and oncotic pressure), thus implicating transmural small-solute osmotic gradients as the responsible driving force (98, 118, 137, 138). For NaCl and urea gradients generated by the lag in equilibration between DVR plasma and medullary interstitium to induce water efflux, the water must traverse a pathway of sufficiently small pore size for these solutes to be osmotically active. The missing piece of that puzzle was provided by the cloning of the AQP1 water channel, followed by the demonstration that it is expressed in DVR and other endothelia (3, 92, 93, 132). When AQP1 was blocked by p-chloromercuribenzenesulfonate in the rat (108) or deleted in the mouse (66, 106), the water flux driven by transmural NaCl (but not albumin) gradients was nearly eliminated, such that osmotic water permeability fell from ~1,100 µm/s to nearly 0. Another surprising finding was that AQP1 deletion was accompanied by a marked increase in DVR diameter, the first demonstration of the capacity of vasa recta to remodel (Fig. 9). The question remained, How does the efflux of water from DVR to the medullary interstitium benefit urinary concentration? Insight derived from mathematical simulations showed that shunting of water from DVR to AVR in the superficial medulla reduced blood flow to the deep medulla, secondarily improving plasma-interstitial equilibration. This was predicted to improve the efficiency of inner medullary countercurrent exchange and enhance interstitial osmolality (23, 24, 106, 152).

View larger version (35K): [in this window] [in a new window]   Fig. 9.   Osmotic water permeability (Pf) of DVR from aquaporin-1 (AQP1)-deficient (/) and wild-type (+/+) mice. A: photomicrographs show the increase in diameter of AQP1-deficient (/) murine DVR. B: Pf measurements obtained from in vitro perfused DVR of wild-type (+/+), heterozygous (+/) and AQP1 null (/) mice. Pf was measured by imposing transmural gradients of NaCl to drive water flux across the DVR wall. Data are reproduced from Ref. 106 with permission.

Nephrons, collecting ducts, and DVR deposit water to the medullary interstitium. Thus, for mass balance, removal is delegated to AVR. The latter are highly fenestrated (110, 138) and have very high hydraulic conductivity and solute permeability that exceeds that of DVR (68, 69, 102, 103, 116). AVR are less well characterized than DVR because they cannot be isolated for in vitro perfusion. AVR have been postulated to serve an unusual role in the clearance of macromolecules from the medullary interstitium. The renal medulla is devoid of lymphatics, so albumin must be removed from the interstitium into AVR (10, 20). The reflection coefficient of AVR to albumin is relatively low (0.58-0.78) (66-68, 101). Based on this and the ability of AVR to withstand a hydraulic pressure gradient without collapsing, it has been postulated that albumin is cleared from the interstitium by convective solvent drag across the AVR wall (67-69, 100, 160-162). A detailed simulation by Zhang and Edwards (174) verified the feasibility of that mechanism and predicted the presence of an axial gradient of albumin concentration in the medullary interstitium.

	SUMMARY AND CONCLUSIONS

TOP ABSTRACT INTRODUCTION RENAL MEDULLARY MICROVASCULAR... ROLE OF DVR IN... VASOACTIVITY OF DVR ROLE OF DVR ENDOTHELIA... MEDULLARY PO2 AND PERFUSION... TRANSPORT OF SOLUTES AND... SUMMARY AND CONCLUSIONS REFERENCES

The microcirculation of the renal medulla serves several roles. The classic depiction of vasa recta as passive filters that provide countercurrent exchange and solute trapping belies the complexity of their structure and function. DVR are arterioles that respond to an array of vasoactive agents (111). They are also transporting microvessels that express specific transporters for water (AQP1) and urea (UTB) (92, 93, 168). The ability to isolate and study these vessels has made it possible to examine transport properties, pericyte contractile mechanisms, and endothelial interactions. The physiology of AVR is less well characterized because they have only been studied on the surface of the exposed papilla in vivo.

We have recently learned that DVR pericytes behave in ways that are typical of smooth muscle from the afferent arteriole and other microvascular beds. When exposed to ANG II, pericytes depolarize by activating a Cl conductance, a process that gates voltage-activated Ca²⁺ entry pathways (38, 107, 130, 176). It is surprising is that this process apparently occurs in the juxtamedullary efferent circulation that supplies the renal medulla with blood flow but not in smooth muscle of superficial efferent arterioles (13-15, 63, 64). The physiological purpose of this axial heterogeneity is uncertain, but prior observations that L-channel antagonists enhance medullary blood flow seem better explained (26, 37, 65, 169). The ability to access DVR pericytes for electrophysiological examination and measurement of intracellular Ca²⁺ transients is a recent development (107, 130, 175, 176). Only cells from outer medullary vessels have been studied, and it remains possible that those from the inner medulla will be found to have unique properties. In the inner medulla, extracellular Na⁺, K⁺, Cl, urea, and osmolyte concentrations are high, and this might have forced the evolution of unusual mechanisms to control and gate membrane potential and Ca²⁺ entry.

It is clear that DVR endothelia are unusual. These cells express the AQP1 water channel that is responsible for small-solute-driven efflux of water to the medullary interstitium (118, 137). Thus, in opposition to the classic view of purely diffusive countercurrent exchange, water abstraction is an important mode of DVR equilibration. Furthermore, it has been predicted that the latter serves to lower blood flow and optimize interstitial solute concentrations in the deep medulla, where axial gradients are largest (106). In addition to AQP1, DVR endothelia express the same urea carrier as red blood cells, UTB (125, 168, 170-172). Presumably, this provides for rapid equilibration of urea in DVR plasma, RBC interior, and medullary interstitium. As expected, DVR endothelia generate NO, cellular levels of which may be regulated in part through generation of superoxide anion (129). What has been more surprising is that an elevation of endothelial [Ca²⁺]_i , expected to stimulate the NOS3 isoform, does not occur on ANG II stimulation. In fact, ANG II is found to reduce endothelial [Ca²⁺]_i and suppress [Ca²⁺]_i responses to endothelium-dependent vasodilators. It has been hypothesized that this turns modulation of pericyte constriction away from the endothelium to NO diffusion from adjacent mTALs of the outer medullary interbundle region (114, 130). Because NOS3 is subject to regulation by a variety of influences, confirmation of this hypothesis awaits sensitive measurement of DVR NO generation.

Although isolation of DVR for in vitro study has provided the key technique for delineation of the cell biology of these vessels, barriers to our understanding remain. It is uncertain how well observations translate to the in situ condition, where vessels are supported by interstitium and lie in close proximity to paracrine influences arising from adjacent interstitial cells and epithelia (59). The notion that DVR play a role in the modulation of total and regional perfusion of the medulla remains a matter of anatomic inference. Until methods specifically enable observation of blood flow redistribution within vascular bundles in response to specific agonists, uncertainty concerning their precise role will continue. Given the importance of medullary perfusion to salt and water balance, tissue oxygenation, and the pathophysiology of analgesic nephropathy and acute renal failure, the motivation to resolve the details of pericyte-endothelial interactions will be substantial.