Nonsteroidal anti-inflammatory drugs (NSAIDs) are some of the most widely used therapeutic agents in clinical practice today. Although the gastrointestinal toxicity of these medications is well known, it has become increasingly apparent that the kidney is also an important target for untoward clinical events. The renal toxicity associated with the use of NSAIDs can be divided into one of several distinct clinical syndromes. These include a form of vasomotor acute renal failure, nephrotic syndrome associated with interstitial nephritis, chronic renal injury, and abnormalities in sodium, water, and potassium homeostasis. The common link in these syndromes is a disruption in prostaglandin metabolism, the class of compounds whose synthesis is inhibited by these agents.

Prostaglandins are members of a class of compounds termed eicosanoids. Eicosanoids are biologically active fatty acids that are all derived from the oxygenation of arachidonic acid. The particular enzyme involved in the oxygenation process dictates which class of eicosanoid will be synthesized. Oxygenation of arachidonic acid by the enzyme cyclooxygenase is responsible for prostaglandin and thromboxane synthesis (Fig. 32.1). The enzyme lipoxygenase converts arachidonic acid to leukotrienes, lipoxins, and eventually, to hydro fatty acid derivatives such as hydroxy eicosatetraenoic acid (HETE). Finally, oxygenation by the cytochrome P-450 system generates epoxyeicosatrienoic acids (EETs).

The availability of free arachidonic acid is the ratelimiting step in eicosanoid biosynthesis. Normally, arachidonic acid is found esterified to membrane phospholipids, where it undergoes deacylation primarily under the influence of phospholipase A₂. Phospholipase A₂-mediated arachidonic acid release is a calcium-calmodulin-dependent step that is stimulated by vasopressin, bradykinin, angiotensin, and norepinephrine.¹ Once released, free arachidonic acid is either re-esterified back into membrane lipids or is converted into one of the biologically active eicosanoids.

The first step in the synthesis of prostaglandins and thromboxanes is a cyclooxygenase reaction in which arachidonic acid is converted into the cyclic endoperoxide prostaglandin G₂ (PGG₂). PGG₂ then undergoes a peroxidase reaction to form a second endoperoxide called PGH₂, which is accompanied by the formation of a superoxide radical. Both of these reactions are catalyzed by the enzyme cyclooxygenase (COX), also known as prostaglandin endoperoxide H synthase.²,³ The cyclooxygenase and peroxidase reactions occur on distinct but neighboring sites on the COX enzyme. Once formed, PGH₂ has a short half-life and is rapidly acted on by a series of enzymes that produce biologically active prostaglandins or thromboxane. Prostacyclin synthase acts to form prostacyclin (PGI₂), thromboxane synthase forms thromboxane A₂, and isomerases are responsible for the formation of PGE₂, PGD₂, and PGF_2α.

Prostaglandins are synthesized on demand and exert physiologic effects in discrete microenvironments along the nephron in close proximity to their points of synthesis (Table 32.1). Due to the virtual absence of distant effects, these compounds are best regarded as autacoids rather than hormones. Variations in the synthetic and degradative machinery along the length of the nephron account for the differing types and amounts of prostaglandins found in any given segment.⁴ PGI₂ is the most abundant prostaglandin produced in the cortex and is primarily synthesized in cortical arterioles and glomeruli.⁵ This location corresponds to the known effects of PGI₂ in regulating renal vascular tone, the glomerular filtration rate (GFR), and renin release. PGE₂ and thromboxane A₂ are also produced in the glomerulus and therefore may exert effects at this site.

The most abundant prostaglandin found in the tubules is PGE₂.⁵ The cortical and especially the medullary portion of the collecting duct are the dominant sites of PGE₂ synthesis. Lesser amounts are found in the thin descending
and thick ascending limb with the least amount of synthesis found in the proximal tubule. Medullary interstitial cells are also a rich source of PGE₂ production. This distribution provides the anatomic basis for PGE₂ to modulate sodium and chloride transport in the Henle loop, regulate arginine vasopressin-mediated water transport, and control vasa recta blood flow. PGF_2α is synthesized primarily by medullary interstitial cells and less by the papillary collecting tubule and glomeruli. Prostaglandin-degradative enzymes are found in both the cortex and medulla but are most abundant in the cortex. Except for PGI₂, which undergoes spontaneous hydrolysis to 6-keto-PGF_2α, prostaglandins are rapidly metabolized into inactive products by a 15-prostaglandin dehydrogenase. An increased concentration of this enzyme in the proximal nephron may facilitate the degradation of prostaglandins delivered to the proximal tubule by glomerular filtration.⁶

Under baseline euvolemic conditions, prostaglandin synthesis is negligible, and as a result, these compounds play little to no role in the minute-to-minute maintenance of renal function. However, a major role occurs in the setting of a systemic or intrarenal circulatory disturbance. This interaction is best illustrated when examining the renal function under conditions of volume depletion (Fig. 32.2). In this setting, renal blood flow is decreased while sodium reabsorption, renin release, and urinary concentrating ability are increased. To a large extent, these findings are mediated by the effects of increased circulating levels of angiotensin II (AII), arginine vasopressin (AVP), and catecholamines. At the same time, these hormones stimulate the synthesis of renal prostaglandins, which in turn act to dilate the renal vasculature, inhibit salt and water reabsorption, and further stimulate renin release. Prostaglandin release under these conditions serves to dampen and counterbalance the physiologic effects of the hormones that elicit their production. As a result, renal function is maintained near normal levels despite the systemic circulation being clamped down. Predictably, the inhibition of prostaglandin synthesis will lead to unopposed activity
of these hormonal systems, resulting in exaggerated renal vasoconstriction and magnified antinatriuretic and antidiuretic effects. In fact, many of the renal syndromes that are associated with the use of NSAIDs can be explained by the predictions of this model.

Aspirin and other NSAIDs exert their prostaglandin-inhibitory effects by inhibiting the COX enzyme. The COX enzyme exists as two isoforms termed COX-1 and COX-2. These enzymes are encoded by two different genes and differ significantly in their regulation. The COX-1 enzyme is constitutively expressed in most tissues and is responsible for producing prostaglandins involved in maintaining normal tissue homeostasis. The COX-2 enzyme is principally an inducible enzyme rapidly upregulated in response to a variety of stimuli such as growth factors and cytokines typically found in the setting of inflammation.³ With the discovery of COX-2, a great deal of effort was put forth to develop compounds to selectively block the activity of this isoform without affecting the activity of COX-1. The availability of a COX-2-specific inhibitor would provide a therapeutic tool to inhibit the synthesis of arachidonic acid metabolites at sites of inflammation and yet leave unperturbed COX-1-derived prostanoids involved in normal homeostasis. In this manner the analgesic, anti-inflammatory, and antipyretic effects of an NSAID could be obtained with minimal to no side effects. Although the initial experience with specific COX-2 inhibitors has been associated with a reduction in gastrointestinal complications, this paradigm is not applicable to the kidney.

COX-1 and COX-2 are both constitutively expressed in the kidney. COX-1 is localized to mesangial cells, arteriolar endothelial cells, parietal epithelial cells of the Bowman capsule, and throughout the cortical and medullary collecting duct.⁷ COX-2 is primarily expressed in the macula densa and the adjacent cells in the cortical thick ascending limb with lesser amounts in the podocytes and the arteriolar smooth muscle cells.⁸,⁹ COX-2 is also abundantly expressed in interstitial cells in the inner medulla and the papilla.

The expression of COX-2 in different regions of the kidney varies in response to alterations in intravascular volume. This variation is particularly evident in the macula densa, where studies show COX-2 plays an important stimulatory role in the release of renin via the tubuloglomerular feedback mechanism. Under conditions of low renal perfusion when the chloride concentration at the level of the macula densa is low, renin release is inhibited by a COX-2 selective inhibitor but unaffected by a COX-1 inhibitor.¹⁰ In genetically engineered mice lacking COX-2, there is a failure of renin release in response to a low salt diet, whereas renin release is intact in animals lacking COX-1.¹¹,¹²

Stimulation of renin with the subsequent formation of angiotensin II is part of a feedback loop because angiotensin II exerts an inhibitory effect on COX-2 synthesis in the macula densa via the angiotensin type 1 (AT₁) receptor.¹³ In contrast to effects at the macula densa angiotensin II upregulates COX-2 and prostaglandin synthesis in vascular smooth muscle cells and mesangial cells.¹⁴ This latter effect provides a mechanism for COX-2 to both facilitate the tubuloglomerular feedback response to low salt delivery to the macula densa by increasing angiotensin II levels and preserve the glomerular filtration rate through the generation of vasodilatory prostaglandins to antagonize the vasoconstrictive effect of angiotensin II.¹⁵

The expression of COX-2 is also responsive to changes in volume. COX-2 expression decreases with salt depletion and increases with a high salt diet and dehydration.⁹ COX-2-derived prostaglandins may play an important role in facilitating a natriuretic response to salt loading and may help protect against volume overload. The increase in COX-2 in response to dehydration is thought to provide a cytoprotective effect in the setting of hypertonic stress.¹⁶,¹⁷ Treatment of water-deprived animals with a selective COX-2 inhibitor is associated with apoptotic patches of renal medullary interstitial cells. By contrast, no such changes are seen in animals treated with the inhibitor alone or in animals undergoing water deprivation without pharmacologic treatment.

In summary, COX-2 is constitutively expressed in the kidney and is highly regulated in response to physiologic perturbations in intravascular volume. The majority of experimental and clinical studies to date suggest that the specific COX-2 inhibitors may not offer any distinct advantage
over traditional NSAIDs with regard to renal toxicity. In fact, most of the renal syndromes that have been linked to nonselective COX inhibitors have now been described with the selective COX-2 inhibitors. The only exception is the development of chronic kidney disease and papillary necrosis. The failure to link COX-2 inhibitors use to these complications is not surprising because these agents have only been available for clinical use for a relatively short time. As with traditional NSAIDs, the COX-2 inhibitors need to be used cautiously and require close monitoring of renal function in patients at high risk for adverse renal outcomes.

Prostaglandins primarily exert a vasodilatory effect on the renal vasculature. This vasodilatory effect alters the renal circulation in two major ways. First, these compounds influence the distribution of renal blood flow to different regions of the kidney. Prostaglandin stimulation results in a preferential increase in blood flow to the more juxtamedullary nephrons.¹⁸,¹⁹ By contrast, the inhibition of prostaglandin synthesis results in a selective reduction of flow to the inner cortical nephrons while flow remains well preserved in the outer cortex.²⁰ Second, prostaglandins exert a vasoregulatory effect on the renal microcirculation to include the interlobular, afferent, and efferent arterioles as well as the glomerular mesangium. In isolated renal arterioles, both PGE₂ and PGI₂ attenuate AII-induced and norepinephrine-induced afferent arteriolar vasoconstriction. On the efferent side of the circulation, PGI₂ similarly antagonizes AII-induced and norepinephrine-induced vasoconstriction, but PGE₂ is without effect.²¹ In addition to local production, vascular reactivity of the efferent arteriole appears to be influenced by prostaglandins produced in the upstream glomerulus. In this regard, Arima and associates²² find that the orthograde infusion of AII (afferent arteriole-glomerulus-efferent arteriole) results in less vasoconstriction of the efferent arteriole as compared to when infused in a retrograde fashion (efferent arterioleglomerulus-afferent arteriole). Pretreatment with indomethacin markedly increases the vasoconstrictive effect during an orthograde infusion but is without effect during the retrograde infusion.

Prostaglandins have also been shown to attenuate mesangial cell contraction induced by AII, endothelin, AVP, and platelet-activating factor.²³,²⁴ Contraction of these cells will normally cause a decrease in the total glomerular capillary surface area and result in a fall in the GFR. Mesangial cell synthesis and the release of PGI₂ in humans and PGE₂ in rats dampen the constrictor effects of these hormones such that the glomerular capillary surface area is maintained, thereby minimizing any fall in GFR. Thus, in the setting of enhanced hormonal constrictor activity, prostaglandins play a major role in maintaining glomerular hemodynamics by exerting a vasodilatory effect at the level of the afferent and efferent arteriole as well as within the glomerular mesangium.