- Top of page
- Conflict of interest
COXs are enzyme complexes that catalyse the formation of prostanoids from arachidonic acid. To date, two distinct isoenzymes, COX-1 and COX-2, which differ in their expression pattern and regulation, have been identified (Habenicht et al., 1985; Herschman, 1996). Enzymatic activities of these COX isoforms produce prostanoids that play a pro-inflammatory role by mediating fever, hyperalgesia and vasodilatation. COX-1 is expressed in most tissues and is regarded as a constitutive enzyme involved in physiological functions such as mucus secretion in the stomach (Ikari et al., 1999), whereas COX-2 is inducible (Sirois and Richards, 1993; Crofford et al., 1997; Herschman et al., 1997). However, COX-2 has been reported to be constitutively expressed in certain tissues such as the kidney (Harris et al., 1994; Ferreri et al., 1999), brain and spinal cord (Breder et al., 1995). Non-steroidal anti-inflammatory drugs (NSAIDs) can alter renal function by reducing glomerular filtration rate, renal blood flow, and sodium and potassium excretion (Catella-Lawson et al., 1999). Both COX enzymes are known to be present in the kidney and their related prostanoids are involved in the regulation of fluid balance and blood pressure (Warner and Mitchell, 2002). An increase in COX-2 expression has been demonstrated in ureteral obstruction in the rat (Norregaard et al., 2006), and prostanoids have also been found to increase the contractility of electrically-stimulated human ureters (Cole et al., 1988).
Recently, a number of inhibitors relatively selective for COX-1 and COX-2 have been discovered, each of them demonstrating differing degrees of potency in their ability to inhibit COX activity (Riendeau et al., 2001). In addition, inhibition of COX-1 has been found to cause intestinal hypermotility, bacterial invasion, inducible NOS expression, and upregulation of COX-2 (Tanaka et al., 2002).
Acute ureteral colic resulting from ureteral obstruction is a common and painful event in urolithiasis. Pharmacological relaxation of the ureter smooth muscle would facilitate the treatment of ureter stone colic as well as prepare the ureter for easier endoscopic access. Smooth muscle relaxant drugs may reduce smooth muscle contraction and spasm on the level of the stone, causing obstruction and increase of the intraluminal pressure, which is finally responsible for the severe colic pain.
NSAIDs, such as diclofenac, are used to treat patients suffering from renal colic caused by ureteric calculi, symptomatically. These compounds have been demonstrated to have anti-inflammatory and analgesic effects, whereas little is known about the possible effects of NSAIDs on ureteral smooth muscle. Indomethacin has been shown to reduce the frequency of severe colic attacks (Grenabo and Holmlund, 1984), whereas diclofenac, despite its analgesic effect, does not improve stone passage rate in renal colic (Kapoor et al., 1989; Laerum et al., 1995).
However, diclofenac has been shown to inhibit the spontaneous activity of sheep ureters (Thulesius et al., 1987), to cause relaxation of KCl-contracted human ureters (Sivrikaya et al., 2003), and of porcine 5-hydroxytryptamine-stimulated ureter segments (Mastrangelo et al., 2000). Furthermore, diclofenac has been shown to almost abolish the contractile response of electrically-stimulated human ureter muscle in vitro (Cole et al., 1988), and to induce a faster release of ureteric calculi and reduce the pain of patients suffering from renal colic (Hetherington and Philp, 1986).
Although the effects of COX-2 inhibitors on the contractility of ureteral smooth muscle of several species have been evaluated (Mastrangelo et al., 2000; Jerde et al., 2005), little information on their effects on human ureter are available in the literature until recently. Hence, the aim of this study was to investigate the expression of both COX-1 and COX-2 receptors in the human ureter and kidney. An in vivo model for partial ureteral obstruction was developed in piglets so that the effects of selective COX-2 inhibitors and non-selective inhibitors on ureteral contractility could be evaluated in samples of human ureters, in vitro, and porcine ureters, in vivo. As prostanoids have been shown to increase the contractility of human ureters (Cole et al., 1988), it was hypothesized that inhibitors of COX-1 and COX-2 would decrease the contractility of these ureters.
- Top of page
- Conflict of interest
In the present study, we demonstrated that COX-1 and COX-2 receptor proteins are expressed in urothelium, smooth muscle cells and in the tunica muscularis and the tunica media of blood vessels (small arterioles) of human ureter. Furthermore, COX-2 was present in renal tubule and macula densa. In kidneys, COX-1 and COX-2 were observed in renal tubule epithelium. Our results on the functional effects of the selective COX-2 inhibitor, valdecoxib, and the non-selective COX inhibitor, diclofenac, demonstrate that diclofenac induces a concentration-dependent decrease in the amplitude of contractions in electrically-stimulated human ureters in vitro. Valdecoxib had no effect on ureter contractility when compared with solvent.
A significant increase in the amplitude of contractions was seen over time in porcine non-obstructed ureters compared with partially obstructed ureters in vivo. In comparison to the in vitro data, a dose-dependent effect of the non-selective COX inhibitor, diclofenac, on ureteral contractility was not obtained in pig ureters in vivo. The COX-2 inhibitor, parecoxib, reduced the amplitude of ureteral contractions in non-obstructed ureters, but had no effect in the partially obstructed porcine ureters.
Our immunohistochemistry results are in agreement with those obtained by Fornai et al. (2005). This group demonstrated that COX-1 and COX-2 are constitutively expressed in colonic tunica muscularis (Fornai et al., 2005). Also, our findings on the intracellular localization of both COX proteins in human ureters are consistent with the results of Murakami et al. (2003). These authors demonstrated that both COX-1 and COX-2 are located in the perinuclear envelope, and the ER membrane, with COX-1 being dispersed into the cytoplasm along the ER membrane and COX-2 being more prevalent in the perinuclear area in HEK293 cells.
In rodent kidney, Campean et al. (2003) demonstrated that COX-1 is located in mesangial cells of the glomerulus, in terminal distal tubule, in connecting tubule and cortical, and medullary collecting ducts. In our study, COX-2 was found to be present in renal tubule and macula densa. Furthermore, staining in vascular smooth muscle cells was absent. These differences could be due to species differences. In contrast to our findings, in humans, COX-1 has been shown to be expressed in collecting duct cells, interstitial, endothelial and smooth muscle cells of pre- and postglomerular vessels, whereas COX-2 is localized in endothelial and smooth muscle cells of arteries, veins and intraglomerular regions of podocytes (Komhoff et al., 1997). In adult humans, COX-2 was shown to be associated with the endothelium of vasa recta and medullary capillaries, whereas COX-1 is only localized in collecting ducts and macula densa.
Nakada et al. (2002) demonstrated that COX-2 mRNA and protein levels are upregulated in chronically obstructed human ureters. We used non-obstructed ureter samples, and both COX-1 and COX-2 proteins were detected in samples from human ureter.
NSAIDs are currently considered a first-line treatment of renal colic. Their action has been ascribed to the inhibition of renal prostaglandin synthesis, which decreases renal blood flow and diuresis, and consequently lowers the pressure in the renal pelvis and ureter.
Diclofenac and the selective COX-2 inhibitor NS-398 have been shown to be almost equipotent in reducing neurokinin-A-induced contractions in the porcine isolated ureter in vitro (Mastrangelo et al., 2000). We also demonstrated that diclofenac decreases the contractility of human ureters in vitro. However, in our study, valdecoxib did not decrease the amplitude of contractions of human ureters when compared with solvent controls. This inconsistency might be due to different methods of stimulation of the ureteral specimens (EFS versus neurokinin-A- or 5-HT-induced contractions) or the fact that Mastrangelo et al. (2000) used a different selective COX-2 compound (valdecoxib versus NS-398). In addition, our experiments were performed in human ureter samples, whereas Mastrangelo et al. (2000) used porcine samples.
Indomethacin has been demonstrated to cause a concentration-dependent inhibition and/or suppression in EFS-evoked contractions in guinea-pig ureter, but these effects were obtained at high concentrations (100–500 μM) (Santicioli et al., 1995). Indomethacin and NS-398 completely abolished the frequency of contractions in human and porcine ureters in vitro (Nakada et al., 2000). Furthermore, celecoxib, a selective COX-2 inhibitor, indomethacin and NS-398 have been demonstrated to inhibit TNFα-induced ureteral contractions in porcine ureter in vitro (Jerde et al., 2005). However, the effect of these COX inhibitors was investigated on spontaneous ureteral contractions occurring at a frequency of 5–10 contractions per 5 min. Again, different methodological parameters might explain the differences in the results. Specimens were suspended in the organ bath in a longitudinal manner, whereas we used ureter rings. However, Jerde et al. (1999) did ascertain that no differences in contractions were evident when ureter rings were compared with spiral cut and longitudinal segments (Jerde et al., 1999). Furthermore, in their study, the effect of only one concentration of each compound was investigated, whereas we constructed concentration–response curves. Inhibition of COX-1 has also been shown to increase the amplitude and the frequency of contractions in rat ureter but to have no effect in the guinea-pig ureter (Davidson and Lang, 2000). Davidson and Lang (2000) also showed that NS-398, a selective COX-2 inhibitor, reduced the motility index (amplitude × frequency) in the guinea-pig upper urinary tract but had no effect on this variable in rat upper urinary tract.
COX inhibitors have been found to cause smooth muscle relaxation in a variety of smooth muscle tissues. In contrast to our data on ureteral smooth muscle, diclofenac has been demonstrated to inhibit distension-induced rhythmic contractions in rat urinary bladder in vivo, and this effect was found to be proportional to its effectiveness as an anti-inflammatory agent (Maggi et al., 1984). From these results, it was suggested that arachidonic acid metabolites could regulate micturition by enhancing the amplitude of myogenic contractions in the bladder, but no analysis of prostaglandin synthesis was performed in this study.
Indomethacin and NS-398 have been shown to decrease contractility in guinea-pig trachea in vitro (Charette et al., 1995). COX-1 sparing drugs have been reported to exert tocolysis in human myometrium in vitro, and this effect was less marked when non-selective COX inhibitors were used (Slattery et al., 2001). Indomethacin has also been shown to inhibit myometrial contractility via mechanisms independent of COX (Sawdy et al., 1998). In guinea-pig isolated small intestine, SC-560, a selective COX-1 compound, and NS-398 did not affect contractility in vitro, whereas indomethacin disturbed the regular pattern of propulsive motility in this species by an effect unrelated to COX inhibition (Shahbazian et al., 2001). Therefore, differences in function of COX inhibition occur within species but also smooth muscle tissues. Different expression levels within tissues and/or species may explain the inconsistent effects of selective COX-2 inhibitors on smooth muscle contractility.
In our in vivo model, an acute partial obstruction was induced about 30 min before the compounds were administered to healthy piglets. A significant difference in Amax was obtained between partially obstructed and non-obstructed ureters under physiological conditions. Therefore, the model used to investigate the effects of compounds is important with respect to clinical consequences or conclusions. COX-2 expression has been shown to be upregulated in chronically obstructed human (Nakada et al., 2002) and porcine ureters (Jerde et al., 2006) in vivo.
In our study, diclofenac decreased the contractility of pig ureters in vitro in a concentration-dependent manner, whereas no dose-dependent effect of diclofenac on contractility was seen in vivo. Recent findings from Davenport et al. (2007) support this finding; they demonstrated that diclofenac had no effect on the contraction frequency in the human ureter in vivo. In our study, the selective COX-2 inhibitor, parecoxib reduced the amplitude of contractions in porcine non-obstructed ureters when compared with solvent, whereas no effect on contractility was seen in partially obstructed ureters. Parecoxib has been shown to attenuate the increase in pelvic pressure that occurs during obstruction in rats in vivo, but the dose used in that study was considerably higher than those used in our study (Norregaard et al., 2006). Also, Norregaard et al. (2006) found an increased expression of COX-2 protein in dilated ureter compared with non-dilated ureter suggesting that COX-2 activity contributes to increased pressure after obstruction. Furthermore, urinary prostaglandin-E2 excretion was increased after release of the obstruction. Interestingly, inhibition of prostanoid synthesis has been found to reduce ureteral contractility rates in vivo and in vitro, in support of the decrease in contractility seen with parecoxib in non-obstructed ureters. However, it has recently been demonstrated that prostaglandin-E2 increases contractility in chronically obstructed ureters but inhibits contractility in non-obstructed ureters (Lowry et al., 2005).
In the in vitro studies, differences in harvesting, localization, storage, preparation and experimental design could explain the different results reported. In the in vivo studies, small alterations in the renal pelvis can affect peristaltic activity in the ureter and the smooth muscle is particularly susceptible to physical interferences (Ancill et al., 1972). Drugs used for general anaesthesia might also cause relaxant effects on the ureter (Young et al., 1994), and, therefore, the in vivo results may not represent the true effects of the COX inhibitors.
In addition to the analgesia and anti-inflammatory potential of COX-2 inhibitors (Mehlisch et al., 2003), a relaxant effect on the ureter might also be beneficial in patients with acute stone colic. Ureteric peristalsis has always been assumed to be essential for the spontaneous passage of stones, and our in vivo findings showed that diclofenac did not affect ureter contractility in pigs. In acute ureteral colic resulting from ureteral obstruction, pharmacological relaxation of the ureter smooth muscle would facilitate the treatment of ureter stone colic as well as prepare the ureter for easier endoscopic access. Whereas our in vitro findings suggest that diclofenac has potential spasmolytic properties and might be beneficial with respect to treatment of ureter stone colic, the in vivo findings in pigs did not demonstrate an effect on ureteral contractility. Hence, as we demonstrated that COX-2 inhibitors have no effect on the contractility of human isolated ureters or of porcine partially obstructed ureters in vivo, apart from their known analgesic effect, it is unlikely that they would be useful in the treatment of ureteric stone colic.