Physiologic and Pathophysiologic Roles of Cyclooxygenase-2 in the Kidney

Abstract

In the mammalian kidney, prostaglandins are important mediators of physiologic processes, including modulation of vascular tone and salt and water. Prostaglandins arise from enzymatic metabolism of free arachidonic acid (AA), which is cleaved from membrane phospholipids by phospholipase A2 activity. The cyclooxygenase (COX) enzyme system is a major pathway for metabolism of arachidonic acid in the kidney. Cyclooxygenases are the enzymes responsible for the initial conversion of AA to PGG2 and subsequently to PGH2, which serves as the precursor for subsequent metabolism by specific prostaglandin and thromboxane synthases. In addition to high levels of expression of the “constitutive” rate-limiting enzyme responsible for prostanoid production, COX-1, the “inducible” isoform of cyclooxygenase, COX-2, is also constitutively expressed in the kidney and is highly regulated in response to alterations in intravascular volume. Prostaglandins and thromboxane A2 exert their biological functions predominantly through activation of specific 7-transmembrane G-protein–coupled receptors. We and others have shown that COX-2–derived prostaglandins exert important physiologic functions in maintenance of renal blood flow, mediation of renin release, and regulation of sodium excretion. In addition to physiologic regulation of prostanoid production in the kidney, increases in prostanoid production are also observed in a variety of inflammatory renal injuries, and we have found a role for COX metabolites to serve as mediators of inflammatory injury in renal disease.

INTRODUCTION

Prostaglandins regulate vascular tone and salt and water homeostasis in the mammalian kidney and are involved in the mediation and/or modulation of hormonal action. Prostaglandins arise from enzymatic metabolism of free arachidonic acid (AA), which is cleaved from membrane phospholipids by phospholipase A2 activity. Cyclooxygenase (prostaglandin synthase G₂/H₂, COX) is the rate-limiting enzyme in metabolizing arachidonic acid to prostaglandin G₂ and subsequently to prostaglandin H₂, which serves as the precursor for subsequent metabolism by prostaglandin and thromboxane synthases. Two isoforms of COX exist in mammals, “constitutive” COX-1 and inflammatory-mediated and glucocorticoid-sensitive COX-2. COX-1 is expressed in mammalian kidney in vasculature, glomerular mesangial cells, and the collecting duct. We first reported COX-2 to be present at low but detectable levels in a normal adult rat kidney, with localized expression in the cortex in the macula densa (MD) and adjacent cortical thick ascending limb (cTAL) and in the medulla restricted to lipid-laden medullary interstitial cells (1) (Fig. 1). Subsequent studies have documented COX-2 expression in MD/cTAL and medullary interstitial cells in kidney of mouse, rat, rabbit, dog, and human, as well as lower levels of expression in podocytes and renal arterioles (2).

Fig. 1.

Proposed roles of COX-2 in renal physiology.

REGULATION OF MACULA DENSA COX-2 EXPRESSION

In the mammalian kidney, the macula densa is involved in regulating renin release and afferent arteriolar tone (via tubuloglomerular feedback). Administration of non-specific cyclooxygenase inhibitors or selective COX-2 inhibitors prevents the increases in renin release mediated by macula densa sensing of decreases in luminal NaCl. Induction of a high renin state by imposition of a salt deficient diet, ACE inhibition, diuretic administration, or experimental renovascular hypertension all significantly increase macula densa/cTAL COX-2 expression (3–7).

Decreased intraluminal chloride concentration is the signal for macula densa stimulation of renin secretion. Macula densa sensing of luminal chloride concentration is dependent upon net apical transport, mediated by the luminal Na⁺/K⁺/2Cl^- cotransporter, NKCC2 (8). NKCC2 possesses a high affinity for Na⁺ and K⁺, such that minimal alterations in transport occur with physiologic changes of Na⁺ or K⁺ concentrations; however, the affinity for chloride is lower and falls within the range of loop chloride values, thereby resulting in an uptake mechanism that is very sensitive to any change in luminal chloride. Loop diuretics, which inhibit NKCC2, increase renin activity, even in the absence of volume depletion. Studies in primary cultured cTAL and in an immortalized mouse macula densa cell line indicate that COX-2 expression increases in response to decreased extracellular chloride (9–11). Studies by Bell et al. using isolated perfused macula densa have directly shown increased PGE₂ from macula densae perfused with decreased NaCl concentrations (12).

Recent in vivo studies compared the effects of diuretics that act by inhibiting sodium reabsorption in different nephron segments. Acetazolamide acts as a diuretic primarily due to its inhibition of carbonic anhydrase activity in the proximal tubule, resulting in inhibition of proximal tubules reabsorption and increased NaCl delivery to the macula densa/cTAL. Hydrochlorothiazide (HCTZ) and amiloride and spironolactone inhibit sodium reabsorption distal to the macula densa/cTAL in distal convoluted tubules and connecting tubules and collecting tubules, respectively. In rats, urine volume and sodium excretion increased comparably in response to all diuretics tested, indicating comparable volume depletion. Acetazolamide suppressed cortical COX-2 expression, whereas HCTZ, spironolactone, and amiloride all stimulated cortical COX-2 expression. Although loop diuretics may increase cortical COX-2 expression in part through direct inhibition of macula densa/cTAL NKCC2 (5), the increased cortical COX-2 expression seen with HCTZ, amiloride, or spironolactone presumably results from systemic volume depletion. Therefore, in the acetazolamide-treated animals, the stimulating effect of volume depletion on cortical COX-2 expression appears to have been counteracted by increased luminal NaCl delivery to macula densa/cTAL due to inhibition of proximal tubule reabsorption (13). The overall suppressive effect of acetazolamide on cortical COX-2 expression indicates the importance of proximal salt and fluid reabsorption in the regulation of cortical COX-2 expression.

INTERACTIONS OF MACULA DENSA COX-2 WITH THE RENIN-ANGIOTENSIN SYSTEM

Renin is normally expressed in the juxtaglomerular (jg) cells, granular smooth muscle-like cells in the afferent arteriole. Lesser amounts of renin expression have also been reported in efferent arteriole and glomerulus, and recent studies have shown that renin is expressed in connecting segment and collecting duct (14, 15). However, by far the greatest expression of renin is in the jg cells. Renin secretion is mediated by a number of stimuli, including β-adrenergic stimulation, alterations in arteriolar perfusion pressure and macula densa–mediated regulation. Prostaglandins have been implicated as modulators for both macula densa–dependent (16) and –independent (17) renin secretion (Fig. 2). Increased renal renin expression in response to low salt, ACE inhibitors, or renovascular hypertension is inhibited by highly selective COX-2 inhibitors, in addition to non-selective COX inhibitors. In global COX-2 knockout mice, neither dietary salt deficiency nor ACE inhibition increased renal renin expression (18, 19), whereas in global COX-1 deficiency, the same stimuli led to normal increases in renin expression (20, 21). Recent studies have indicated that macula densa COX-2 expression plays an essential role in the tonic expression or jg cell renin expression rather than acting as an acute regulator of stimulated renin production/release in response to macula densa–derived signals as well as other signals for renin release (β adrenergic stimulation or renal perfusion pressure) (22, 23).

Fig. 2.

Regulation of the renin-angiotension system by COX-2. Prostaglandins derived from macula densa COX-2 can regulate expression of renin, which then mediates production of angiotensin II and aldosterone.

There is increasing evidence that complex interactions exist between macula densa COX-2 expression and components of the renin-angiotensin system, with both positive and negative feedback mechanisms. In vivo studies indicate that angiotensin II, through AT₁ receptors, inhibit macula densa COX-2 expression. ACE inhibitors or ARBs increase cortical COX-2 mRNA and immunoreactive protein (4). In adult wild type mice on a control diet, minimal renal cortical COX-2 immunoreactive protein was detected, whereas in AT₁R knockout mice (Agtr1a-/-; Agtr1b-/-), abundant COX-2 immunoreactivity was observed in the macula densa (4). Furthermore, chronic administration of either hypertensive or nonhypertensive concentrations of angiotensin II inhibits macula densa COX-2 expression, suggesting a direct inhibition of enzyme expression through AT₁ receptors (24), which have been shown to be present in macula densa cells. Unexpectedly, in addition to COX-2 inhibition by AT₁ receptor activation, angiotensin II can also stimulate macula densa COX-2 expression via AT₂ receptor signaling, although the AT₂-mediated effects are only detected in the presence of AT₁ inhibition, suggesting that under physiologic conditions, the AT₁-mediated inhibitory effects predominate (24). Of interest, in metanephric ovine kidney, AT₂ receptors are highly expressed in macula densa (25) and are present in cultured mouse macula densa cells. COX-2 expression is stimulated by an AT₂ receptor agonist in these cells (24).

REGULATION AND FUNCTION OF MEDULLARY COX-2

The renal medulla is a rich source of COX-2, with the greatest concentrations in the medullary interstitial cells. COX-2-derived prostanoids appear to be involved in modulating sodium excretion. Salt loading downregulates COX-2 expression in renal cortex, but upregulates its expression in renal medulla (3, 26). In experimental animals, systemic or selective medullary COX-2 inhibition leads to sodium retention (27, 28). COX-2 inhibitors have also been reported to induce sodium retention in a minority of human subjects without renal impairment (29), and when measured in balance studies, decreased urinary sodium excretion is observed for the first 72 hours of administration of COX-2 inhibitors (30, 31).

There are also indications that alterations in COX-2 expression and activity may be involved in modulation of vasopressin-mediated water reabsorption. Prostaglandins are known to antagonize vasopressin-mediated water reabsorption. In lithium-induced nephrogenic diabetes insipidus, lithium-induced decreases in GSK-3β activity lead to enhanced renal COX2 expression and COX2-derived urine PGE₂ excretion. Suppression of COX2-derived PGE₂ blunts lithium-associated polyuria (32). PGE₂ synthesis is also altered in central diabetes insipidus. Medullary COX-1, COX-2, and microsomal prostaglandin E synthase-1 (mPGES) are all markedly reduced, and treatment with ddAVP markedly increases medullary COX-2 expression (33).

INTERACTIONS OF RENAL COX-2 AND DOPAMINERGIC SYSTEMS

In addition to its role as a neurotransmitter, dopamine also serves important physiologic functions outside of the central nervous system. Dopamine has an important physiologic role in the kidney to regulate net salt and water excretion, and intrarenal activation of dopaminergic pathways serves to prevent or mitigate the development and consequences of hypertension. Furthermore, abnormalities in renal dopaminergic signaling may be an underlying factor in the development of essential hypertension in some individuals.

Our studies have highlighted potential interactions between the intrarenal dopaminergic and COX-2 systems. These studies suggest that increased macula densa COX-2 expression observed in low salt–treated animals is due in part to decreased proximal tubule dopaminergic activity, resulting in increased net salt and fluid reabsorption in the proximal tubule and decreased luminal NaCl delivery to macula densa. Dopamine modulates cortical COX-2 expression at least in part by regulating proximal tubule reabsorption through a DA₁ receptor mediated pathway in response to alterations in dietary salt (13). It is possible that interactions between the intrarenal dopaminergic and cortical COX-2 systems may comprise an important physiologic feedback regulation in responses to variations of salt intake.

INTRARENAL DOPAMINERGIC SYSTEM

The kidney has an intrarenal dopaminergic system that is distinct from any neural dopaminergic input. Circulating concentrations of dopamine are normally in the picomolar range, whereas dopamine levels in the kidney can reach high nanomolar concentrations (34). The dopamine precursor L-DOPA (L-dihydroxyphenylalanine) is taken up by the proximal tubule via multiple amino acid transporters, including rBat, LAT2, and ASCT2 (35, 36) from the circulation or after filtration at the glomerulus and is then converted to dopamine by aromatic amino acid decarboxylase (AADC), which is also localized to the proximal tubule (37) (Fig. 3). Intrarenal dopamine production increases when dietary salt intake increases (38, 39).

Fig. 3.

Intrarenal production of dopamine. The rate-limiting step is conversion of L-DOPA to active dopamine by AADC, which is localized in the proximal tubule.

Dopamine serves as a counterregulatory factor to angiotensin II in the kidney (40, 41). Dopamine inhibits renal renin expression (42) and inhibits angiotensin II–mediated proximal tubule reabsorption and AT₁ expression (43–46).

Dopamine receptor activation leads to decreases in salt and water reabsorption in the mammalian kidney, mediated at least in part by inhibition of specific tubule transporter activity along the nephron, including NHE3, NaPi-II, NBC, and Na/K-ATPase in the proximal tubule; NKCC2 in TAL; and ENaC and AQP2 in the collecting duct (47–51). A general characteristic of essential hypertension is a relative defect in renal sodium and water handling. Because it is estimated that the intrarenal dopaminergic system is responsible for regulating more than 50% of net renal salt and water excretion when salt intake is increased (52), dysfunction of this system could have profound consequences for regulation of intravascular volume and systemic blood pressure.

Dopamine stimulates prostaglandin production in the kidney medulla and increases urinary prostaglandin excretion (53–55). Specifically, increased intrarenal dopamine leads to increased medullary COX-2 expression and activity (56). Medullary COX-2–mediated prostaglandins promote renal salt and water excretion (57), suggesting that dopamine-mediated effects on natriuresis and diuresis may be at least partly the result of increased renal medullary prostaglandin production.

COX-2 INHIBITION AND HYPERTENSION

In a high salt/elevated aldosterone model of hypertension, COX-2 inhibitors exacerbate hypertension (58). COX-2 inhibition increases systolic BP in both SHR and WKY rats on a high-salt diet, but not in rats on low-salt intake, indicating that the hypertension induced by COX-2 inhibition can occur independently of a genetic predisposition to hypertension and can be prevented by salt deprivation (59). In animals on a high-salt diet, selective medullary infusion of COX-2 inhibitors will also induce hypertension. (26, 28). It has also been reported that deletion of mPGES, one of the major PGE₂ synthases associated with COX-2, leads to salt-sensitive hypertension.

DOPAMINE AND HYPERTENSION

Mice deficient in COMT, one of the major dopamine metabolizing enzymes in the kidney, have increased intrarenal dopamine levels and blunted elevations in blood pressure in response to DOCA/high salt (56), as well as increases in nocturnal blood pressure in response to a high-salt diet (60). Conversely, mice with selective intrarenal deletion of intrarenal dopamine due to AADC deficiency develop salt-sensitive hypertension (61) (Fig. 4). Abnormalities in dopamine production and receptor function have been associated with human essential hypertension and several forms of rodent genetic hypertension (62–65). In animal models of genetic hypertension, there is impairment of the proximal tubule dopaminergic pathway and the ability to increase urinary sodium excretion is impaired (66).

Fig. 4.

Renal dopamine deficiency results in salt-sensitive hypertension. (A) Proximal tubule AADC was deleted in ptAADC-/- mice. (B) Salt-sensitive hypertension was observed in ptAADC-/- mice [Whelton (31)].

In summary, COX-2 expression in the kidney is regulated by both physiologic and pathophysiologic perturbations, with effective volume depletion upregulating macula densa expression and effective volume expansion upregulating medullary expression. Macula densa COX-2 is a modulator of juxtaglomerular renin expression, and there is increasing evidence that COX-2 expression is modulated by multiple components of the renin-angiotensin system, including not only angiotensin II but also renin itself. Medullary COX-2 metabolites modulate salt and water excretion, and COX-2 inhibitors lead to sodium and volume retention and may predispose to or exacerbate hypertension. The kidney has a robust intrarenal dopaminergic system, and there is increasing evidence that alterations in intrarenal dopamine signaling may underlie essential hypertension. Our studies indicate important interactions between the intrarenal COX-2 and dopaminergic systems in regulation of renal salt and water homeostasis, the renin-angiotensin system, and systemic blood pressure. Given cardiovascular and renal side effects, COX-2 inhibitors are not a feasible intervention for long-term therapy against progressive renal damage, but further delineation of the downstream receptors and synthases involved may provide therapeutic targets.

DISCUSSION

Mitch, Houston: Thank you, Ray, as always you‘ve taken a very complex area and made it quite clear. I was curious, so how do these three systems interact to solve the problem that when you have a patient with hypertension and you salt-restrict them and they actually salt-restrict themselves, that actually the blood pressure falls more with renin inhibitors, et cetera.

Harris, Nashville: That's a very interesting question. I think that the drop in blood pressure probably has more to do with the fact that you're restricting the renin-angiotensin system, which still may predominate over any potential activation by cyclooxygenase-2, but I think that each patient may be a little different.

Schreiner, Los Altos: Ray, I don't want to add a fourth complexity to this wonderful portrayal of three complicated systems, but there are also alternative sources of oxidized lipids in the kidney, specifically in these areas, proximal tubule, as you know, the cytochrome p450 capacity. Could you just briefly put your work in the context of the emerging studies on the physiologically active oxidized lipids that are coming out of the cytochrome enzymes and the proximal tubule?

Harris, Nashville: Interestingly enough, dopamine has been shown to mediate production of 20-HETE in the proximal tubule. Although we have not yet looked at the potential role of 20-HETE in the AADC knockout mice, we would predict that there may also be effects from p450 arachidonic acid metabolites that are important, especially since 20-HETE has been shown to be natriuretic. Therefore, it's possible that dopamine's effects, at least in the proximal tubule and possibly in the thick limb as well, may be mediated in part by cytochrome p450 metabolites.