Effects of oral prostaglandins on indomethacin-induced renal failure in patients with cirrhosis and ascites.

Nonsteroidal antiinflammatory drugs (NSAID), such as indomethacin, reduce production of renal prostaglandins and markedly impair renal function in patients with cirrhosis and ascites. To determine if simultaneous administration of oral prostaglandin analogs minimizes the renal impairment, 10 patients received indomethacin and either misoprostol or placebo in a double-blind, crossover study. Indomethacin reduced urinary sodium from 23 +/- 9 to 8 +/- 4 microEq/min in 4 hours. Misoprostol with indomethacin tended to prevent the fall in urinary sodium (from 35 +/- 15 to 46 +/- 21 microEq/min at 4 hours), but sodium excretion fell to the same level in both groups by 8 hours (6 +/- 3 microEq/min). Indomethacin reduced creatinine clearance in 4 hours by 49%; misoprostol plus indomethacin reduced creatinine clearance by only 34%. Misoprostol tended to minimize or delay the nephrotoxic effects of indomethacin, suggesting that more potent prostaglandin analogs may prevent the renal impairment induced by NSAID.

Interleukin 1 suppresses inflammation in rabbit colitis. Mediation by endogenous prostaglandins.

Pretreatment with low-dose IL-1 has protective effects in animal models of inflammation or tissue injury, but the mechanisms of these protective effects are not established. To determine if prostaglandins are involved, we administered human recombinant IL-1 beta and measured rectal PGE2 production in rabbits with formalin-immune complex colitis. IL-1 beta (0.3 micrograms/kg) administered 24 h before induction of colitis increased PGE2 (231 +/- 36 to 1,299 +/- 572 pg/ml, P less than 0.01) and reduced subsequent inflammatory cell infiltration index (from 2.8 +/- 0.3 to 1.4 +/- 0.3, P less than 0.02) and edema (from 2.5 +/- 0.3 to 1.3 +/- 0.3, P less than 0.01) compared with vehicle-matched animals. Administration of ibuprofen (10 mg/kg i.v.) together with IL-1 beta prevented the stimulation of PGE2 and the reduction in inflammation. Colonic PGE2 production correlated inversely with subsequent severity of inflammation (P less than 0.02, r = -0.39) and edema (P less than 0.04, r = -0.35). IL-1-administration 30 min before induction of colitis did not affect the severity of inflammation. Similarly, pretreatment with a noninflammatory synthetic peptide (fragment 163-171) of human IL-1 beta, either 30 min or 24 h before colitis induction, did not reduce inflammation or increase prostaglandin synthesis. These data demonstrate that pretreatment with IL-1 beta 24 h before the induction of colitis reduces inflammation by a mechanism that requires prostaglandin synthesis.

Regulation of eicosanoid production in rabbit colon by interleukin-1.

Prostaglandins and thromboxanes are increased in human and experimental colitis, but the factors that regulate this enhanced production are unclear. The present studies evaluate the effects of the monokines, interleukin-1 alpha and beta on eicosanoid production in rabbit colon. In tissue incubations the peak dose response of eicosanoid release to human recombinant interleukin-1 is 50 ng/ml. Interleukin-1 alpha increases prostaglandin E2 (PGE2) by 4.5 +/- 1.9 ng/g tissue, 6-keto PGF1 alpha by 6.2 +/- 2.7 ng/g, and thromboxane B2 by 2.1 +/- 0.3 ng/g compared to placebo. In isolated rabbit colons perfused with Krebs' solution, 10-h infusion of interleukin-1 alpha (50 ng/ml) progressively increases production of PGE2, 6-keto PGF1 alpha, and thromboxane B2. Bolus injections of bradykinin increase production of PGE2, but not 6-keto PGF1 alpha and thromboxane B2, and these responses are markedly augmented by interleukin-1 alpha: at 10 h bradykinin-stimulated PGE2 production is 518 +/- 104 vs. 95 +/- 18 ng/5 min (p less than 0.005), 6-keto PGF1 alpha is 172 +/- 88 vs. 8 +/- 2 ng/5 min (p less than 0.02), and thromboxane B2 is 60 +/- 14 vs. 13 +/- 4 ng/5 min (p less than 0.02) for interleukin-treated colons vs. placebo-treated colons, respectively. The response is greater with interleukin-1 alpha than interleukin-1 beta. This study demonstrates that interleukin-1 stimulates prostaglandin and thromboxane production in normal colon tissue. These data are consistent with the concept that interleukin-1 production by inflammatory cells may augment prostaglandin and thromboxane production in colitis.

Tissue origin of peptide-responsive eicosanoid production in rabbit intestine.

Different layers of rabbit large and small intestine display different peptide sensitivity and different profiles of eicosanoid release. Isolated perfused mesenteric pedicle alone, with muscularis/submucosa or with muscularis and mucosa from normal small bowel, normal colon, or inflamed colon were stimulated with bradykinin (BK) or n-formyl-methionyl-leucyl-phenylalanine (fMLP). Released prostaglandin (PG)E2, thromboxane (Tx)B2, and leukotriene (LT)B4 were assayed using extensively validated radioimmunoassays. In rabbit colon, PGE2 arises primarily from the mesentery, while in small intestine the muscularis/mucosa releases 70-80% of the total PGE2. BK releases no significant thromboxane from healthy colon, although both muscularis/submucosa and mucosa respond in inflamed colon. In contrast, fMLP stimulates thromboxane from muscularis/submucosa and mucosa of even healthy colon, while release is greatly potentiated in inflammation. Lipoxygenase in the colon is regulated differently than cyclooxygenase; it is not stimulated by BK in either healthy or inflamed colon. fMLP releases equal amounts of LTB4 from healthy and inflamed colon, but release was primarily from healthy colonic mucosa, whereas it was distributed throughout mesenteric pedicle, muscularis, and mucosa in inflamed colon. The ability of normal colonic mucosa to release proinflammatory LTB4 in response to a chemotactic factor (fMLP) produced by enteric bacteria suggests a possible role for these compounds as a stimulus for inflammation in some patients with inflammatory bowel disease.

Comparative acute effects of diflunisal and indomethacin on renal function in patients with cirrhosis and ascites.

Nonsteroidal anti-inflammatory drugs such as indomethacin induce a rapid reduction in renal perfusion and blunt the effects of diuretics in patients with cirrhosis and ascites. Nonacetylated salicylates reportedly cause less reduction in renal prostaglandins than do aspirin and other nonsteroidal anti-inflammatory drugs. To determine whether nonacetylated salicylates affect renal function, we compared diflunisal with indomethacin in nine patients with cirrhosis and ascites. One 50-mg dose of indomethacin reduced inulin clearance (91 +/- 11 to 76 +/- 11 ml/min) and blunted furosemide-stimulated natriuresis (58 +/- 12 to 36 +/- 9 mEq/h) and diuresis (1103 +/- 148 to 809 +/- 170 ml/h, all p less than 0.05). Three doses of diflunisal had no effect on inulin clearance (94 +/- 16 ml/min), natriuresis (60 +/- 12 ml/h), or diuresis (1041 +/- 112). Indomethacin caused greater reduction in urinary prostaglandin E2 (50% vs. 10%) and in serum thromboxane (94% vs. 80%) than diflunisal (p less than 0.05). Thus, nonacetylated salicylates avoid renal impairment and may be the preferred nonsteroidal anti-inflammatory drug in patients with cirrhosis and ascites.

Comparative acute effects of aspirin, diflunisal, ibuprofen and indomethacin on renal function in healthy man.

Nonsteroidal anti-inflammatory drugs (NSAIDs), such as indomethacin, suppress renal prostaglandins and markedly reduce renal perfusion and diuretic response in some renal disorders. Mild renal impairment may occur in healthy subjects. Pharmacodynamic characteristics of certain NSAIDs, such as the nonacetylated salicylates, suggest that they may have less deleterious renal effects. We compared the renal effects of standard therapeutic doses of indomethacin, ibuprofen, aspirin, and the nonacetylated salicylate, diflunisal, in 6 healthy supine volunteers. Only indomethacin significantly reduced creatinine clearance (by 13%) and renal plasma flow (by 23%; p less than 0.05). Indomethacin also tended to reduce furosemide-induced diuresis and natriuresis, and this effect was significantly greater than with diflunisal (p less than 0.05). Serum thromboxane, a reflection of platelet cyclo-oxygenase activity, was reduced by 99% with aspirin, ibuprofen and indomethacin, but by only 78% with diflunisal. Nonacetylated salicylates may be the preferred drugs, at least in short-term usage, when it is necessary to minimize the effects of NSAIDs on platelet or kidney function.

Effects of 16,16-dimethyl prostaglandin E2 and indomethacin on leukotriene B4 and inflammation in rabbit colitis.

The role of increased prostaglandin production and the effects of exogenous prostaglandins on inflammation of colitis are not established. We administered intramuscular 16,16-dimethyl prostaglandin E2 (DiM-PGE2) and indomethacin to rabbits with formalin immune-complex colitis and measured leukotriene B4 (LTB4), prostaglandin E2 (PGE2) and severity of inflammation. DiM-PGE2 (100 micrograms/kg/BID) reduced LTB4 production (from 401 +/- 108 to 216 +/- 58 pg/ml) and infiltration of neutrophils, mucosal necrosis, inflammatory exudate and edema (all P less than 0.05). Other studies determined that parenteral DiM-PGE2 did not reduce the initial chemical damage induced by formalin, suggesting that cytoprotection of chemical insult was not the mechanism of suppressed inflammation in the immune colitis model. Indomethacin (10 mg/kg/d) reduced endogenous PGE2 by 80%, but did not reduce leukotriene production or inflammation. Exogenous prostaglandins cause a dose-dependent suppression of inflammation in experimental colitis, by a mechanism other than cytoprotection of chemical-induced mucosal injury.

Visceral surface oxygen tension in experimental colitis in the rabbit.

The impairment of bowel healing that is characteristic of inflammatory bowel disease (IBD) is poorly understood. Because bowel healing is related to the adequacy of perfusion in other circumstances, we studied bowel surface oxygen tension (PSO2), which is related to bowel perfusion, in rabbits with IBD. Both cell-mediated (n = 17) and immune complex-mediated (n = 10) colitis caused marked attenuation of colon PSO2. Control (n = 13) left colon PSO2 was 36 +/- 5 (SEM) torr. In mild colitis, left colon PSO2 fell to 11 +/- 5 torr, and in severe colitis it fell to 4 +/- 1 torr (p less than 0.01 for each compared with control). These changes occurred irrespective of the mechanism of induction of colitis. Gastric and small intestinal PSO2 were unaffected. Hepatic and renal PSO2 were decreased in severe colitis only. The presence of decreased PSO2 was a better marker for the presence of IBD than was histologic evaluation. It is suggested that attenuation of PSO2 may be a marker for the physiologic activity of IBD. If this is so, PSO2 may prove a useful adjunct in the operative management of IBD.

Glucagon and prostaglandins are mediators of amino acid-induced rise in renal hemodynamics.

An oral protein load or infusion of amino acids induces a rise in renal hemodynamics in normal subjects, but the mechanisms mediating this phenomenon are unknown. We investigated whether glucagon may mediate the increase in RPF and GFR induced by an arginine infusion and whether prostaglandins are required for this effect. In four different studies, normal subjects underwent 13 inulin and PAH clearances of 30 minutes each. During the fourth and tenth clearance periods arginine HCl, 250 mg/kg, was infused over 30 minutes. At the beginning of the fifth clearance period several subjects ingested indomethacin, 150 mg, (N = 8) or ibuprofen, 800 mg (N = 6). Control subjects (N = 4) did not receive cyclooxygenase inhibitors. Six subjects underwent a similar protocol except that they were infused with glucagon, 6 ng/kg/min, instead of arginine, for 30 minutes during the fourth and tenth periods. They also ingested indomethacin, 150 mg, in the fifth period. In all four studies, a transient and significant rise in RPF and GRF and fall in RVR occurred during the first arginine or glucagon infusion. These changes in renal hemodynamics were blocked when the arginine or glucagon infusion was repeated after administration of indomethacin or ibuprofen. Urinary excretion of 6-keto-PGF1 alpha did not rise with either arginine infusion in the control subjects or in the individuals who received indomethacin. As predicted, urinary 6-keto-PGF1 alpha fell significantly after ingestion of indomethacin before the second infusion of arginine. Plasma norepinephrine and epinephrine concentrations were unaffected by the arginine infusions or by indomethacin.(ABSTRACT TRUNCATED AT 250 WORDS)

Exaggerated prostaglandin production by colonic smooth muscle in rabbit colitis.

Enhanced production of arachidonic acid metabolites by colonic mucosa has been reported in ulcerative colitis as well as in experimental models of colitis. However, production of these compounds by colonic smooth muscle from colitis subjects has not been described. To evaluate arachidonic acid metabolism in colonic tissue, we studied the production of prostaglandin E2 (PGE2) by mucosa and muscularis propria in two experimental models of acute colitis in which inflammation was virtually confined to the mucosa. Colitis was induced in New Zealand white rabbits by either of two methods, dinitrochlorobenzene (DNCB) sensitization or formalin followed by intravenous soluble immune complexes (F-IC). Arachidonic acid metabolites were identified from in vitro incubations of tissue with [14C] arachidonic acid by thin layer chromatography followed by autoradiography. The major eicosanoid metabolites of colitis mucosa and muscularis were 14C-labeled prostaglandin E2, prostaglandin F2a and 6-keto prostaglandin F1 alpha. PGE2 was quantitated from incubations without labeled arachidonic acid by radio-immunoassay. PGE2, expressed as picograms per milligram protein per 20 min (mean +/- SEM), was increased in F-IC mucosa (1093 +/- 141 vs 645 +/- 189, P less than 0.05) and DNCB mucosa (1354 +/- 487 vs 527 +/- 222, P less than 0.05) compared to normals. PGE2 production by uninflamed colitis muscularis propria was also increased five- to eightfold compared to normals for F-IC muscularis (1594 +/- 329 vs 189 +/- 35, P less than 0.005) and DNCB muscularis (1287 +/- 171 vs 225 +/- 72, P less than 0.005). Thus, the adjacent inflammation in colonic mucosa may induce increased eicosanoid production by the uninflamed smooth muscle.(ABSTRACT TRUNCATED AT 250 WORDS)

Role of eicosanoids in human and experimental colitis.

Prostaglandins, thromboxanes, and leukotrienes (collectively called eicosanoids) are increased at sites of inflammation and contribute to the manifestations of inflammation, such as hyperemia, hyperalgesia, edema, and inflammatory cell infiltration. Inhibition of eicosanoid production is the basic mechanism of action of corticosteroids and of nonsteroidal antiinflammatory drugs. Eicosanoid synthesis is also increased in human and experimental inflammatory bowel disease. Leukotriene B4 is the most potent proinflammatory eicosanoid, and in vivo production of this compound is the predominant eicosanoid in colitis. Recent experimental data demonstrate that selective inhibition of leukotrienes may be a therapeutic strategy to reduce inflammation in inflammatory bowel disease.

Immunohistochemical distribution of renal prostaglandin endoperoxide synthase and prostacyclin synthase: diminished endoperoxide synthase in the hepatorenal syndrome.

To evaluate possible causes of the diminished prostaglandin production in advanced hepatorenal syndrome, prostaglandin endoperoxide synthase and prostacyclin synthase were localized and semiquantitated by immunofluorescence in postmortem, biopsy and nephrectomy renal tissues. In normal kidneys, antiprostacyclin synthase serum caused intense staining in peritubular capillaries, in the adjacent renal interstitial cells and in glomerular mesangial regions. Antiprostaglandin endoperoxide synthase serum caused staining of collecting duct epithelial cells, cells of the thin ascending limb and possibly glomerular mesangial cells. Prostacyclin synthase-positive staining was graded 5+ (scale of 0+ to 5+) in all kidney samples. Medullary collecting tubule prostaglandin endoperoxide synthase-positive staining was graded 4+ or 5+ in kidney samples from patients with acute tubular necrosis or acute tubulointerstitial nephritis and from patients with liver failure without the hepatorenal syndrome. However, prostaglandin endoperoxide synthase-positive staining was markedly diminished or absent (average 1+) in patients with the hepatorenal syndrome. These data suggest that loss of the medullary prostaglandin endoperoxide synthase is the cause of diminished urinary prostaglandin E2 excretion in the hepatorenal syndrome.

Enhanced urinary immunoreactive thromboxane in neonatal necrotizing enterocolitis. A diagnostic indicator of thrombotic activity.

Urinary thromboxane B2 levels increased threefold to 20-fold in infants with neonatal necrotizing enterocolitis compared with healthy infants and infants with benign causes of heme-positive stools. Increased urinary thromboxane B2 levels were detected coincidently with the initial signs of necrotizing enterocolitis, and values paralleled the course of the illness. Infants with serious illnesses other than necrotizing enterocolitis had increased urinary thromboxane B2 levels but with lower values than those of infants with necrotizing enterocolitis. Indomethacin therapy appeared to reduce urinary thromboxane B2 levels and reduce their usefulness as a marker of illness. Another product of platelet activation, beta-thromboglobulin, was increased in the urine of infants with necrotizing enterocolitis. Decreased platelet counts in infants with necrotizing enterocolitis correlated inversely with urinary thromboxane. Results of beta-thromboglobulin and platelet studies are consistent with the concept that platelet consumption due to ischemic thrombosis was the source of enhanced thromboxane excretion.

Effects of captopril on renal function in patients with cirrhosis and ascites.

Blockade of angiotensin-converting enzyme has been variously reported to increase or to decrease sodium excretion in patients with cirrhosis and ascites. We administered captopril (50-150 mg) to 11 patients with cirrhosis and ascites to determine the effects on blood pressure, renal blood flow and sodium excretion. Plasma renin activity increased and mean blood pressure fell (by 14 mm Hg). Para-aminohippurate clearances increased from 321 +/- 53 to 559 +/- 83 ml/min (P less than 0.005), but inulin clearances were minimally altered (73 +/- 8 to 76 +/- 7 ml/min), suggesting preferential dilation of glomerular efferent arterioles. Despite unchanged glomerular delivery of sodium, urinary sodium excretion fell in all subjects (from 2.70 +/- 1.00 to 0.48 +/- 0.21 mEq/h), urinary volume was reduced (377 +/- 55 to 182 +/- 42 ml/h, P less than 0.005), and the natriuretic effect of furosemide was blunted. The antinatriuretic effect of captopril may be mediated by reduced angiotensin II-mediated sodium excretion, by decreased prostaglandin production, and/or by indirect effects of reduced blood pressure. Captopril impairs rather than promotes sodium excretion.

Immediate effects of furosemide on renal hemodynamics in chronic liver disease with ascites.

Furosemide occasionally causes azotemia in patients with ascites, independently of induced volume depletion. To define this effect, we measured renal clearances in patients with chronic liver disease and ascites and in nonascitic controls. Furosemide (80 mg i.v.) transiently increased p-aminohippurate clearance in controls (from 693 +/- 67 to 928 +/- 93 ml/min) and in 11 patients with ascites (from 418 +/- 81 to 526 +/- 80 ml/min). In contrast, in 13 patients with ascites, p-aminohippurate clearance fell by 34% (from 545 +/- 51 to 360 +/- 24 ml/min) within 20 min and by 41% within 60 min, and inulin clearance fell by 19% at 20 min and by 30% at 60 min. The renal effects lasted approximately 4 h. The renal response could not be predicted by renin activity, urinary prostaglandin excretion, urinary sodium, or clinical characteristics. In all 14 patients who received oral furosemide, p-aminohippurate clearance fell within 90 min (by 24%) and remained suppressed for at least 4 h. These immediate effects of furosemide on renal perfusion may contribute to azotemia in some patients with ascites.

Chemotactic peptide stimulation of leukotrienes from healthy and inflamed rabbit colons.

Prostaglandins, thromboxanes and leukotrienes are increased in human and experimental colitis. To evaluate the biosynthesis of these eicosanoids, colon inflammation was induced in rabbits by formalin enema followed by i.v. immune complexes, and the distal colon was perfused ex vivo. Bradykinin increased synthesis of prostaglandin E2 and thromboxane B2 more from colitis than from control colons (both P less than .001) but had no effect on leukotriene synthesis. The inflammatory cell agonist N-formylmethionyl-leucyl-phenylalanine (30 ng) also induced greater synthesis of prostaglandin E2 (70 +/- 13 vs. 14 +/- 6) and thromboxane B2 (84 +/- 22 vs. 20 +/- 11) from colitis than from control colons (P less than .01), but leukotriene B4 (416 +/- 68 vs. 438 +/- 128 ng/5 min) and leukotriene C4 (171 +/- 50 vs. 203 +/- 25 ng/5 min) synthesis were greatly augmented in both colitis and control colons. In vitro incubations demonstrated similar dose-dependent stimulation of leukotriene B4 by N-formylmethionyl-leucyl-phenylalanine in both colitis and control colons. These studies demonstrate that healthy colon tissue as well as colitis tissue can produce proinflammatory leukotrienes in response to bacterial peptides. Leukotriene production may contribute to the induction or mediation of colon inflammation.

Altered furosemide pharmacokinetics in chronic alcoholic liver disease with ascites contributes to diuretic resistance.

Some patients with chronic alcoholic liver disease and ascites have an impaired natriuretic response to furosemide. To elucidate the mechanism of this diuretic resistance, we measured para-aminohippurate and inulin clearances and urinary excretion of electrolytes, prostaglandin E2, and furosemide after intravenous administration of 80 mg of furosemide in 26 patients. The natriuretic response was variable (3.3-172 mEq/h) and was unrelated to basal sodium excretion, renal clearances, or urinary prostaglandin E2. Natriuresis correlated negatively with plasma aldosterone (r = -0.54, p less than 0.01), and strongly with urinary furosemide (range 5.5-76 mg/h, r = 0.71, p less than 0.001). As urinary furosemide excretion reflects the amount of furosemide reaching the active site on the luminal side of the tubule, the data demonstrate markedly reduced amounts of furosemide at its primary site of action in patients with diuretic resistance. Plasma furosemide was higher in patients with reduced furosemide excretion and impaired natriuresis, suggesting that the defect was an impairment of furosemide transport into the tubule. Thus, a major factor in diuretic resistance is altered furosemide pharmacokinetics.