Role of Microsomal E Synthase 1 in the Kidney

Helene Francois,* Carie Facemire,* Anil Kumar,* Laurent Audoly,† Beverly Koller,‡ and Thomas Coffman* *Divisions of Nephrology, Duke University and Durham Veterans Affairs Medical Centers, Durham, North Carolina, †MedImmune, Gaithersburg, Maryland, and ‡Department of Genetics, University of North Carolina, Chapel Hill, North Carolina

Prostaglandin E2 (PGE2) is one of the most ubiquitous in the kidney, where it may influence a wide range of physiologic functions. PGE2 is generated through enzymatic of endoperoxides by specific PGE synthases (PGES). Several putative PGES have been identified and cloned, including the membrane-associated, inducible microsomal PGES1 (mPGES1), which is expressed in the kidney. To evaluate the physiologic role of mPGES1 in the kidney, mice with targeted disruption of mPges1 gene were studied, with a focus on responses where PGE2 has been implicated, including urinary concentration, regulation of blood pressure, and response to a loop diuretic. The absence of mPGES1 was associated with a 50% decrease in basal excretion of PGE2 in urine (P < 0.001). In female but not male mPGES1-deficient mice, there was a reciprocal increase in basal excretion of other prostanoids. Nonetheless, urinary osmolalities were similar in mPges1؉/؉ and mPges1؊/؊ mice at baseline and after 12 h of water deprivation. Likewise, there were no differences in blood pressure between mPGES1-deficient and wild-type mice on control or high- or low-salt diets. The furosemide-induced increase in urinary PGE2 excretion that was seen in wild-type mice was attenuated in mPGES1-deficient mice. However, furosemide-associated diuresis was reduced only in male, not female, mPGES1-deficient mice. Stimulation of renin by furosemide was not affected by mPGES1 deficiency. These data suggest that mPGES1 contributes to basal synthesis of PGE2, but there are other pathways that lead to renal PGE2 synthesis. Moreover, there are significant gender differences in physiologic contributions of mPGES1 to control kidney function. J Am Soc Nephrol 18: 1466–1475, 2007. doi: 10.1681/ASN.2006040343

rostaglandin E2 (PGE2) is one of the major prostanoid the actions of phospholipases, conversion of to products generated by the kidney (1), and it has diverse unstable endoperoxide intermediates by actions in the kidney, affecting both vascular tone and (COX-1 and -2), and, finally, isomerization of PGH to PGE by P 2 2 epithelial functions. For example, infusion of PGE2 into the terminal PGE synthases (PGES) (8). PGES activity has been kidney commonly causes renal vasodilation (2,3). PGE2 also characterized biochemically in a number of tissues (9). In most modulates renal sodium and water excretion (1,4). Studies in cell populations, the highest levels of specific PGES activity are isolated perfused nephron segments suggest that PGE2 acting associated with microsomal membrane fractions. Jakobsson through E-prostanoid (EP) receptors such as EP1 may directly and associates (10,11) provided the first molecular character- stimulate solute flux in renal epithelia (5). PGE2 may also ization of a membrane-bound form of human PGES (mPGES1), influence water permeability in the distal nephron by opposing showing that it is a member of the MAPEG (membrane-asso- the actions of the stimulated G-protein–linked vasopressin re- ciated proteins involved in and metab- ceptor (4). Moreover, enhanced generation of PGE2 seems to olism) gene family. Expression of mPGES1 is highly inducible contribute to the actions of loop diuretics (6). It has been sug- in states of injury and inflammation, often in parallel with gested that the propensity of nonsteroidal anti-inflammatory expression of the inducible COX-2 (12,13). More recently, a drugs (NSAID) to cause fluid retention and hypertension is due second putative microsomal PGES (mPGES2) that can also to inhibition of PGE synthesis and consequent attenuation of 2 generate PGE from endoperoxides in vitro with high affinity its actions in the kidney (7). 2 and specificity was identified (14,15), but its capacity to gener- PGE is produced by a series of three enzymatic reactions: 2 ate PGE in vivo is not clear. A third form of PGES has been release of arachidonic acid from membrane phospholipids by 2 isolated from cytoplasmic extracts (16). This cytosolic PGES (cPGES) is expressed ubiquitously in a constitutive manner Received April 11, 2006. Accepted March 30, 2007. (16). However, because of its low substrate affinity compared with the microsomal , the relevance of cPGES to the Published online ahead of print. Publication date available at www.jasn.org. generation of PGE2 in vivo has also been questioned. Address correspondence to: Dr. Thomas M. Coffman, Building 6/Nephrology (111I), VA Medical Center, 508 Fulton Street, Durham, NC 27705. Phone: 919-286- Very little is known about the functions of the individual 6947; Fax: 919-286-6879; E-mail: [email protected] PGES in the kidney. Nonetheless, several studies have evalu-

Copyright © 2007 by the American Society of Nephrology ISSN: 1046-6673/1805-1466 J Am Soc Nephrol 18: 1466–1475, 2007 mPGES1 and Kidney Function 1467 ated regional expression of mPGES1 in the kidney by in situ in phosphate and carbonate buffer prepared with deionized water and hybridization and immunocytochemistry (5,17–19). There is a incubated overnight at 37°C to convert intact PGE2 and its intermediate general agreement among these studies that mPGES1 is ex- metabolites to the stable PGE2 metabolite (PGEM). Concentrations of pressed in the distal nephron from the connecting tubule PGEM in urine were determined by using a specific ELISA assay (Cayman Chemicals, Ann Arbor, MI). Previous studies have shown through the cortical and medullary collecting duct (5,17–19). In that urinary PGEM is a reliable indicator of urinary PGE excretion in this region, the distribution of mPGES1 parallels COX-1 expres- 2 vivo (24). Thromboxane B (TxB ), the stable metabolite of TxA , and sion (18,19). This distribution is consistent with previous stud- 2 2 2 6-keto- PGF1␣, the stable metabolite of prostacyclin (PGI2), were mea- ies indicating that the collecting duct is one of the nephron sured in fresh urine using specific ELISA (Cayman Chemicals). segments with the highest capacity for PGE2 generation (1); abundant EP receptors, including EP1, EP3, and EP4, are also Assessment of Urinary Concentrating Capacity expressed in the collecting duct (4,20). Moreover, this segment Mice were placed in metabolic cages for collection of urine and of the nephron plays a critical role in the final adjustments of monitoring of water intake. After an initial adaptation period, water sodium excretion that have a major impact on fluid volume and intake and urine volumes were measured while the mice had free ϩ ϩ Ϫ Ϫ BP homeostasis. However, the capacity of individual mPGES to access to water. Urine of male and female mPges1 / and mPges1 / influence renal functions is not known. mice was collected by bladder massage before and after a 12-h period In studies that used mPGES1-deficient mice, we show here of water deprivation. Urine osmolalities were then immediately mea- that mPGES1 contributes to basal and stimulated PGE2 gener- sured using a vapor pressure osmometer (Wescor Instruments, Logan, ation by the kidney. In male mice, augmented generation of UT) as described previously (25).

PGE2 by mPGES1 contributes to furosemide-induced diuresis. However, the relative contribution of mPGES1 to synthesis and Responses to the Loop Diuretic Furosemide functional consequences of PGE2 in the kidney differs signifi- A 5-mg/ml furosemide (Sigma Chemicals, St. Louis, MO) solution cantly between genders. was prepared in 10% DMSO in deionized water at pH 7. The 10% DMSO solution alone was used for control injections. Mice were ad- Materials and Methods ministered an injection of 0.1 ml/10 g body wt of the furosemide Animals solution or vehicle control and were immediately placed into individ- Production of mice with targeted disruption of the mPges1 gene has ual metabolic cages, where they were provided free access to tap water. been described previously (21). The mPges1 mutation was originally Urine was collected for exactly 4 h. Urine volumes were measured generated by homologous recombination in embryonic stem (ES) cells using tared containers, and urine sodium concentrations were mea- that were derived from DBA/1 inbred mice. Resulting chimeras were sured using a flame photometer (Instrumentation Laboratory, Lexing- crossed with DBA/1 mice (Jackson Laboratory, Bar Harbor, ME) to ton, MA). For definition of the contribution of prostanoids to the generate inbred DBA/1 mice that bear the mPges1 mutation. Inbred furosemide response, separate groups of inbred male and female mice Ϫ Ϫ ϩ ϩ DBA/1 mPges1 / and mPges1 / mice that were generated by inter- were given (10 mg/kg; Sigma Chemicals) in drinking water ϩ Ϫ crossing DBA/1 mPges1 / mice were used for the studies described for 24 h before the furosemide injection. here. Mice were bred and maintained in the animal facility of the Durham Veterans Affairs Medical Center, and genotypes were deter- Reverse Transcription and Real-Time Quantitative PCR mined as described previously (21). The experimental procedures de- Total RNA was extracted from whole kidneys using TriReagent scribed next were approved by the respective institutional animal care (Sigma) according to the manufacturer’s protocol. RNA was DNAse- and use committees of the Durham Veterans Affairs and Duke Univer- treated using Turbo DNA-free (Ambion, Austin, TX) to remove genomic sity Medical Centers. DNA contamination. RNA yield was quantified by ultraviolet spectro- photometry, and integrity was verified by 1% agarose gel electrophore- Measurement of Systolic BP in Mice sis and staining with ethidium bromide. Only RNA with A260/280Ͼ1.7 Systolic BP was measured in conscious mice using a computerized and displaying no significant degradation was used for reverse tran- tail-cuff system (Hatteras Instruments, Cary, NC) after 2 wk of daily scription. training as described previously (22). Data were recorded at baseline cDNA were synthesized from 5 ␮g of total RNA using random for2wkandthen5daweek throughout the study period (one set of hexamers and SuperScript II reverse transcriptase (RT; Invitrogen, 10 measurements per day). Measurements with a SD Ͼ20 mmHg for Carlsbad, CA). “No RT” samples that lacked RT were prepared during the systolic BP were discarded. This method was validated previously each RT reaction for use as negative controls during PCR. Real-time and correlates well with direct measurements of intra-arterial pressure quantitative PCR was performed using the fluorogenic 5Ј-exonuclease (23). Baseline BP was measured on a control diet (0.4% NaCl). For assay (26). Primers and dual-labeled probes (5Ј-FAM, 3Ј-TAMRA) tar- determination of whether mPGES1 plays a role in BP homeostasis when geting mPGES1 (assay ID Mm00452105_m1; exons 2 to 3 boundary, dietary sodium intake is altered, BP was also monitored while mice amplicon length 87 bp), mPGES2 (assay ID Mm00460181_m1; exons 2 to were fed low-salt (LS; Ͻ0.02% NaCl) or high-salt (HS; 6% NaCl) diets 3 boundary, amplicon length 73 bp), cPGES (assay ID Mm00727367_s1; for 3 wk (Harlan Teklad, Madison, WI). For these studies, the experi- exon 8, amplicon length 70 bp), COX-1 (assay ID Mm00477214_m1; ϩ ϩ mental groups consisted of both male mPges1 / (n ϭ 14) and exons 2 to 3 boundary, amplicon length 66 bp), and COX-2 (assay ID Ϫ Ϫ ϩ ϩ Ϫ Ϫ mPges1 / (n ϭ 11) and female mPges1 / (n ϭ 6) and mPges1 / (n ϭ Mm00478374_m1; exons 5 to 6 boundary, amplicon length 80 bp) were 7) mice. purchased from Applied Biosystems (Foster City, CA). PCR reactions were performed in duplicate on an iCycler real-time detection system Measurements of Urinary Prostanoid Excretion (BioRad, Hercules, CA). cDNA and negative control (no RT, water) Urinary prostanoid excretion was measured using individual ELISA templates (1 ␮l) were added to 20 ␮l PCR reaction mixtures that ϫ ϫ assays. For measurement of PGE2, fresh urine samples were incubated consisted of 1 ABI TaqMan Universal PCR master mix and either 1 1468 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 1466–1475, 2007

Ϫ Ϫ human eukaryotic 18S rRNA primer-probe mix or 1ϫ primer-probe tabolites in urine that was collected from mPges1 / and ϩ ϩ mix for the gene of interest. Gene expression was quantified using the mPges1 / mice. In general, as shown in Figure 1, urinary excre- two standard curve method for relative quantification (27). Briefly, a tion of all prostanoid moieties was significantly higher in female five-point dilution series of positive control cDNA was prepared, and compared with male mice irrespective of mPges1 genotype (P Ͻ standard curves of threshold cycle versus relative template concentra- 0.001 for urinary PGEM excretion and TxB excretion between tion were generated for the gene of interest and housekeeping gene 2 male and female mice; P Ͻ 0.05 for urinary 6 keto-PGF ␣). Excre- (18S rRNA). Template concentrations of unknown samples were then 1 determined, based on their threshold cycle values, from these standard tion of PGE2 metabolites was reduced by approximately 50% in Ϫ/Ϫ curves. Expression of mPGES1, mPGES2, cPGES, COX-1, and COX-2 male mPges1 mice compared with wild-type (WT) controls Ϯ Ϯ Ͻ was normalized to 18S rRNA expression in the same tissue. (from 368 50 to 196 34 pg/24 h; P 0.01; Figure 1A); 24-h In a separate groups of mice, furosemide or vehicle was administered urine volume was similar in both groups (1.1 Ϯ 0.6 versus 1.1 Ϯ 0.5 Ϯ as described. Four hours later, kidneys was harvested, RNA was ex- ml). In contrast, urinary excretion of TxB2 (1579 244 versus tracted, and expression of mPGES1 and COX-2 was determined. 1509 Ϯ 253 pg/24 h) or 6 keto-PGF1␣ (1941 Ϯ 417 versus 1422 Ϯ 132 pg/24 h) were unaffected by the mPges1 mutation in male mice.

Furosemide-Stimulated Renin Expression As shown in Figure 1B, PGE2 excretion was also reduced by Ϫ Ϫ To examine the role of mPGES1 in regulation of renin, we compared approximately 50% in female mPges1 / mice compared with baseline and furosemide-stimulated renin expression in male Ϯ Ϯ Ͻ ϩ ϩ Ϫ Ϫ controls (2527 248 versus 1306 217 pg/24 h; P 0.01). mPges1 / and mPges1 / mice. Groups of mice of were given furo- However, basal excretion of PGE2 metabolites was almost four- semide (2.5 mg/kg per d) in the drinking water (n ϭ 12) or tap water Ϫ Ϫ fold higher in female mPges1 / mice than in WT male mice alone (n ϭ 7). After 5 d, kidneys were harvested, RNA was isolated, and (1306 Ϯ 217 versus 368 Ϯ 50 pg/ml; P Ͻ 0.01). Moreover, in renin mRNA levels were determined by real-time PCR as described. Ϫ Ϫ contrast to the male mPges1 / mice, urinary excretion of other prostanoid species was increased in the mPGES1-deficient female Statistical Analyses Ϯ The values for each parameter within a group are expressed as the mice. For example, urinary TxB2 excretion was 5473 635 pg/24 Ϫ/Ϫ Ϯ means Ϯ the SEM. For comparisons between groups, statistical signif- hinthemPges1 female mice compared with 3960 377 pg/24 Ͻ icance was assessed using ANOVA followed by unpaired t test ad- h in controls (P 0.05). Similarly, excretion of 6 keto-PGF1␣ was Ϫ/Ϫ justed for multiple comparisons. A paired t test was used for compar- 4088 Ϯ 660 pg/24 h in the mPges1 female mice compared with isons within groups. only 2177 Ϯ 250 pg/24 h in female controls (P Ͻ 0.05).

Results Expression of PGES and Cyclooxygenases in Kidneys of mPGES1 and Urinary Prostanoid Excretion mPGES1-Null Mice To determine the contribution of mPGES1 to urinary prostanoid To determine the effect of mPGES1 deletion on other com- excretion, we measured excretion of PGE2, TXA2, and PGI2 me- ponents of the prostanoid synthetic pathway and to test for

Figure 1. Microsomal synthase 1 (mPGES1) and urinary prostanoid excretion. Prostanoid excretion was deter- Ϫ Ϫ mined in 24-h urine collections from male wild-type (WT; Ⅺ; n ϭ 15) and mPges1 / (f; n ϭ 11) mice (A) and female WT (Ⅺ; n ϭ Ϫ/Ϫ f ϭ 6) and mPges1 ( ; n 7) mice (B) and is expressed in pg/24 h. Prostaglandin E2 (PGE2) metabolite excretion is decreased by Ϫ/Ϫ Ϫ/Ϫ 50% in both male and female mPges1 mice. Thromboxane B2 (TxB2) and 6 keto-PGF1␣ are increased in female mPges1 versus female WT mice (P Ͻ 0.05). Prostanoid excretion is increased in female compared with male mice regardless of gender and Ͻ Ͻ Ͻ † Ͻ genotype (P 0.001 versus male mice for PGEM and TxB2; P 0.01 for 6 keto-PGF1␣). *P 0.01 versus WT; P 0.05 versus WT; ‡P Ͻ 0.001 versus male mice; §P Ͻ 0.01 versus male mice. J Am Soc Nephrol 18: 1466–1475, 2007 mPGES1 and Kidney Function 1469

ϩ ϩ potential gender effects on their expression in the kidney, we compared with male mPges1 / mice (P Ͻ 0.01). There were no used real-time quantitative RT-PCR to compare mRNA abun- differences in expression of mPGES2 and cPGES between male dance of COX-1, COX-2, mPGES1, mPGES2, and cPGES in and female WT mice. ϩ ϩ Ϫ Ϫ groups of male and female mPges1 / and mPges1 / mice. As shown in Figure 2A, the absence of mPGES1 had no Among WT mice, there seemed to be a trend toward greater effect on COX-1 expression in male or female mice. By con- COX-1 mRNA abundance in WT female compared with male trast, COX-2 mRNA expression was increased significantly Ϫ Ϫ mice, but the difference did not reach statistical significance in mPGES1 / male mice compared with WT male controls (Figure 2A). Conversely, levels of COX-2 mRNA in kidneys of (P ϭ 0.01), and these levels were similar to those of ϩ ϩ WT female mice were approximately two-fold higher than mPGES1 / female mice. However, the absence of mPGES1 those of WT male mice (P Ͻ 0.01), corresponding with the did not cause further augmentation of COX-2 expression in significantly higher urinary prostanoid excretion that we ob- female mice as COX-2 mRNA expression was similar in ϩ ϩ Ϫ Ϫ served in female mice. Similarly, as shown in Figure 2B, levels female mPges1 / and mPges1 / mice. As expected, of mPGES1 mRNA were also significantly higher in female mPGES1 expression was not detected in kidneys from male

Figure 2. Renal PGES and mRNA expression. Relative mRNA expression levels of mPGES1, mPGES2, and cPGES (A), as well as COX-1 and COX-2 (B) were assessed by real-time quantitative reverse transcriptase–PCR in kidneys of male and Ϫ Ϫ female mPges1 / and WT control mice. Target mRNA expression levels were normalized to expression of 18S ribosomal RNA in the same tissue. Statistical significance was assessed by one-way ANOVA followed by Holm-Sidak post hoc test. *P Ͻ 0.01 versus male; #P Ͻ 0.01 versus WT. 1470 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 1466–1475, 2007

Ϫ Ϫ or female mPges1 / mice (Figure 2B). Moreover, the ab- sence of mPGES1 had no effect on expression of the other putative PGES, mPGES2 or cPGES. mPGES1 and Urinary Concentration

Previous studies have suggested that PGE2 may directly oppose the actions of vasopressin to promote water flux in the distal nephron, thereby modulating urinary concentrating mechanisms (4). To determine whether mPGES1 might contrib- ute to this response, we compared urinary osmolality in Ϫ Ϫ ϩ ϩ mPges1 / and mPges1 / mice with free access to water and ϩ ϩ Ϫ Ϫ after 12-h of water deprivation. Baseline urine osmolalities Figure 3. Furosemide responses in mPges1 / and mPges1 / Ϫ Ϫ ϩ ϩ were similar in mPges1 / (n ϭ 8) and mPges1 / (n ϭ 12) mice. Urinary sodium excretion (4h-UNa; A) and urine volume mice (1672 Ϯ 180 versus 1876 Ϯ 134 mOsm; NS). After 12 h of (4h-UV; B) were measured during the first 4 h after vehicle or ϭ Ϫ/Ϫ ϭ water deprivation, urine osmolalities increased significantly in furosemide injection to male WT (n 15) and mPges1 (n 10) Ϫ/Ϫ Ͻ mice or in female WT (n ϭ 12) and mPges1 (n ϭ 7) mice. *P Ͻ both groups (P 0.001), and the maximal urine osmolalities † ‡ Ϫ Ϫ 0.001 versus vehicle; P Ͻ 0.01 versus WT; P Ͻ 0.05 versus WT. after thirsting were virtually identical in mPges1 / and ϩ ϩ mPges1 / mice (3431 Ϯ 109 versus 3461 Ϯ 73 mOsm; NS). Thus, the absence of mPGES1 does not significantly alter the capacity of the renal concentrating mechanisms. PGE2 was higher in the WT female compared with male mice, Ϯ PGE2 excretion increased further after furosemide, from 889 Role of mPGES1 in BP and Adaptation to Altered Dietary 196 to 1664 Ϯ 122 pg/4 h (P Ͻ 0.01). Similar to male mice,

Sodium Intake excretion of TxB2 and 6 keto-PGF1␣ was not altered by furo- On the control (0.4% NaCl) diet, there were no differences in semide in female WT mice (Figure 4). BP between male (117 Ϯ 2 versus 118 Ϯ 2 mmHg) and female To determine whether furosemide effected expression of syn- Ϯ Ϯ Ϫ/Ϫ ϩ/ϩ (121 1 versus 119 2 mmHg) mPges1 and mPges1 thetic enzymes in the PGE2 pathway, we compared mRNA mice, respectively. BP was not significantly affected by low-salt levels for mPGES1 in WT DBA/1 mice 4 h after administration feeding in male (120 Ϯ 2 versus 118 Ϯ 2 mmHg) or female of vehicle or furosemide. Levels of mPGES1 mRNA were sim- Ϫ Ϫ ϩ ϩ (120 Ϯ 1 versus 121 Ϯ 2 mmHg) mPges1 / and mPges1 / ilar in vehicle- (1.48 Ϯ 0.22) and furosemide-treated mice mice, respectively. Feeding the high-salt diet caused BP to (1.32 Ϯ 0.10), suggesting that furosemide had no effect on increase significantly in all of the groups (P Ͻ 0.0001 by paired mPges1 expression during the interval tested. ϩ ϩ t test), but the magnitude of BP increases was similar in the WT As shown in Figure 3, compared with male mPges1 / mice, and mPGES1-deficient mice so that BP were similar at the end response to furosemide was significantly attenuated in the male Ϫ Ϫ of the high-salt period in male (133 Ϯ 3 versus 129 Ϯ 2 mmHg) mPges1 / mice. Urine flow (0.95 Ϯ 0.05 versus 1.2 Ϯ 0.07 ml/4 Ϫ Ϫ and female (132 Ϯ 3 versus 134 Ϯ 3 mmHg) mPges1 / and h; P Ͻ 0.05) and sodium excretion (117 Ϯ 6 versus 147 Ϯ 8 ϩ ϩ mPges1 / mice, respectively. ␮mol/4 h; P Ͻ 0.01) after furosemide injection were approxi- mately 20% lower in male mPGES1-deficient mice than in WT Role of mPGES1 in the Diuretic Response to Furosemide controls. However, both urine flow and sodium excretion were To determine whether mPGES1 might be the source of the increased significantly compared with baseline conditions (P Ͻ enhanced PGE2 generation with furosemide and to define its 0.001). Furthermore, unlike in WT male mice, furosemide had Ϫ/Ϫ contribution to facilitating the actions of furosemide in the no effect on PGE2 metabolite excretion in the male mPges1 kidney, we compared furosemide responses in groups of male mice (73 Ϯ 18 versus 74 Ϯ 10 pg/4 h). Accordingly, as shown in ϩ/ϩ Ϫ/Ϫ and female mPges1 and mPges1 mice. As shown in Figure 4, excretion of PGE2 metabolites after administration of Figure 3, administration of furosemide caused a brisk diuresis furosemide was significantly lower in the male mPGES1-defi- in male WT mice, with urine volume increasing from 0.31 Ϯ cient mice (74 Ϯ 10 pg/4 h) than in male, WT controls (152 Ϯ 28 0.07 ml/4 h before treatment to 1.2 Ϯ 0.07 ml/4 h after furo- pg/4 h; P Ͻ 0.05). Ϫ Ϫ semide (P Ͻ 0.001). Urine sodium also increased significantly Unlike the male mPges1 / mice, the extent of furosemide- Ϫ Ϫ from 23 Ϯ 7to147Ϯ 8 ␮mol/4 h (P Ͻ 0.001). Along with this induced diuresis was not altered in female mPges1 / mice. As Ϯ Ϯ diuresis and natriuresis, excretion of PGE2 metabolites in- shown in Figure 3, urine flow (0.28 0.03 versus 1.22 0.07 creased in male WT mice from 73 Ϯ 18 to 152 Ϯ 28 pg/4h (P Ͻ ml/4 h; P Ͻ 0.001) and sodium excretion (18 Ϯ 7 versus 145 Ϯ ␮ Ͻ 0.05), whereas excretion of TxB2 and 6 keto-PGF1␣ was unaf- 11 mol/4 h; P 0.001) increased significantly. Moreover, after fected (Figure 4). furosemide, urine flow (1.22 Ϯ 0.07 versus 1.24 Ϯ 0.1 ml/4 h) Furosemide had very similar effects in WT female mice (Fig- and sodium excretion (145 Ϯ 11 versus 132 Ϯ 12 ␮mol/4 h) were ϩ ϩ Ϫ Ϫ ure 3). Urine volume increased from 0.26 Ϯ 0.07 ml/4 h before virtually identical in female mPges1 / and mPges1 / mice. treatment to 1.24 Ϯ 0.1 ml/4h with furosemide (P Ͻ 0.001). As shown in Figure 4, despite this preservation of furosemide Ϯ Ϯ Ϫ/Ϫ Sodium excretion also increased from 13.6 4to132 12 diuresis in the mPges1 female mice, PGE2 metabolite excre- ␮mol/4 h (P Ͻ 0.001; Figure 3). Although basal excretion of tion was modestly but not significantly augmented after furo- J Am Soc Nephrol 18: 1466–1475, 2007 mPGES1 and Kidney Function 1471

Figure 4. Prostanoid excretion after furosemide administration. Prostanoid excretion was determined in 4-h urine collections after Ϫ Ϫ Ϫ Ϫ administration of furosemide or vehicle in male WT and mPges1 / mice (A) and female WT and mPges1 / mice (B) and is expressed in pg/4 h. *P Ͻ 0.05 versus WT; §P Ͻ 0.01 versus vehicle; °P Ͻ 0.001 versus vehicle; •P Ͻ 0.05 versus vehicle; **P Ͻ 0.01 versus WT; †P Ͻ 0.05 versus WT; ‡P Ͻ 0.001 versus WT.

semide (387 Ϯ 85 versus 585 Ϯ 65 pg/4 h; NS). Thus, compared densa, reduced solute flux through the Na-K-2Cl co-transporter with female WT mice, PGE2 metabolite excretion after furo- (NKCC2) regulates the PGE2-dependent signal for renin (28). semide was significantly lower in the mPGES1-deficient female To examine the role of mPGES1 in this pathway, we compared mice (599 Ϯ 55 versus 1664 Ϯ 122 pg/4 h; P Ͻ 0.001). Nonethe- the effects of chronic NKCC2 blockade with furosemide on ϩ ϩ Ϫ Ϫ less, as a consequence of the marked gender difference in basal renin stimulation in mPges1 / and mPges1 / mice. There prostanoid excretion, the levels of PGE2 excretion after furo- was no significant difference in renin expression between un- ϩ ϩ Ϫ Ϫ semide were significantly higher in mPGES1-deficient female treated mPges1 / (1.03 Ϯ 0.17) and mPges1 / mice (0.71 Ϯ mice than in WT male controls (585 Ϯ 65 versus 301 Ϯ 52 pg/4 0.26). Administration of furosemide to WT mice for 5 d caused Ͻ Ϯ ϭ h; P 0.01). Although PGE2 excretion was not significantly a marked stimulation of renin expression (5.68 0.56; P affected by furosemide, excretion of both TxB2 (840 Ϯ 65 versus 0.002). Furosemide caused a similar stimulation of renin in the Ϯ Ͻ Ϯ Ϯ ϭ 2143 4 pg/4 h; P 0.05) and 6 keto-PGF1␣ (881 166 versus mPGES1-deficient mice (6.44 1.02; P 0.006). However, the 4397 Ϯ 637 pg/4 h; P Ͻ 0.01) increased after furosemide in the extent of furosemide-stimulated renin expression was not sig- ϩ ϩ Ϫ Ϫ female mPGES1-deficient mice (Figure 4). nificantly different between the mPges1 / and mPges1 / To explore further the relationship between furosemide-in- mice, suggesting that mPGES1 does not have a unique, nonre- duced diuresis and prostanoid generation in female mice, we dundant role to stimulate renin mRNA expression via this examined the consequences of a more complete inhibition of mechanism. prostanoids by pretreating mice with the COX inhibitor diclofe- nac (10 mg/kg). Pretreatment of female WT mice with diclofe- nac caused a significant attenuation of furosemide-induced Discussion natriuresis and diuresis in WT female mice (Figure 5, A and B). PGE2 is a key product of the COX pathway of arachidonic Urine flow was decreased by approximately 35%, from 1.38 Ϯ acid metabolism. This lipid mediator has diverse actions that 0.13 ml/4 h without pretreatment to 0.89 Ϯ 0.09 ml/4 h with include cytoprotective functions in the gastrointestinal tract, diclofenac (P Ͻ 0.05). Sodium excretion was similarly reduced regulating smooth muscle tone in airways and vasculature, from 145 Ϯ 13 to 108 Ϯ 12 ␮mol/4 h (P Ͻ 0.05) with the COX modulating the immune system, and triggering febrile re- inhibitor. As shown in Figure 5C, along with its effects on the sponses in the brain (29–31). PGE2 is a major arachidonic acid diuretic response, diclofenac caused a broad inhibition of furo- metabolite in the kidney, where it may affect control of renal semide-stimulated prostanoid excretion (575 Ϯ 48 versus 2044 Ϯ blood flow, sodium excretion, and BP (1). It is produced by a Ͻ Ϯ Ϯ 252 pg/4 h for PGE2 [P 0.001]; 237 42 versus 561 123 pg/4 series of three enzymatic reactions: release of arachidonic acid Ͻ Ϯ Ϯ h for TxB2 [P 0.05]; 248 36 versus 511 126 pg/4 h for 6 from membrane phospholipids by the actions of phospho- keto-PGF1␣ [P Ͻ 0.05]). lipases, conversion of arachidonic acid to unstable endoperox- ide intermediates by COX-1 and -2, and, finally, isomerization

Furosemide-Stimulated Renin Expression of PGH2 to PGE2 by terminal PGES. The temporal and spatial

PGE2 is a potent secretagogue for renin, where it operates as properties of PGE2 synthesis are determined by the regulated part of a signaling pathway that includes COX-2. In the macula expression of each of the enzymes involved in this cascade. Of 1472 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 1466–1475, 2007

Figure 5. Role of prostanoids in furosemide diuresis. Sodium excretion (A) and urine volume (B) were measured during 4 h Ϫ Ϫ immediately after furosemide injection in female WT (Ⅺ; n ϭ 8) and mPges1 / (f; n ϭ 7) mice and in a group of WT female mice (u; n ϭ 8) that had been pretreated with the COX inhibitor diclofenac (Diclo). (C) Urinary prostanoid excretion after furosemide Ϫ Ϫ was also measured and is expressed as pg/4 h. *P Ͻ 0.05 versus furosemide and mPges1 / ; §P Ͻ 0.001 versus WT; **P Ͻ 0.05 versus WT; †P Ͻ 0.01 versus WT and WTϩDiclo; ‡P Ͻ 0.001 versus WT ϩ Diclo.

the several putative PGES that have been described, mPGES1 is ing PGI2 (prostacyclin). This may result from shunting of en- the best characterized, and a role for mPGES1 in inflammation doperoxides toward other prostanoid synthetic pathways. This and febrile responses has been clearly demonstrated (21,32,33). phenomenon has been observed in other circumstances when Although mPGES1 is expressed in the kidney (17–19,34), its prostanoid synthetic enzymes are genetically deleted (39) or contribution to PGE2 synthesis in the kidney and to the conse- pharmacologically inhibited (40). Because PGI2 and PGE2 have quent regulation renal functions has not been characterized some redundant functions (8), this increase in PGI2 generation previously. might act as a compensatory mechanism to ameliorate the

Relative to other COX products, PGE2 is generated in sub- physiologic consequences of mPGES1 deficiency. Similar re- stantial quantities by the kidney (1). Major sites of PGE2 gen- ductions in basal PGE2 generation with shunting have been Ϫ Ϫ eration by the kidney are renal medullary interstitial cells and reported in other tissue preparations from mPges1 / mice collecting duct (1,35–37). Studies using in situ hybridization and (41,42). We also found prominent differences in renal prosta- immunocytochemistry have documented mPGES1 expression noid generation between male and female mice in our studies in the distal nephron from the connecting tubule through the with basal levels of prostanoid excretion that are almost four- cortical and medullary collecting duct (17–19,34). In this region, fold higher in female than male mice. These were accompanied the distribution of mPGES1 parallels expression of COX-1 by enhanced expression of mRNA for synthetic enzymes, in- (18,19). cluding COX-2 and mPGES1 in kidneys of female compared Our studies suggest that mPGES1 makes a nonredundant with male mice. contribution to renal PGE2 production, because urinary PGE Although our data assign a unique role for mPGES1 in renal metabolite excretion is reduced by approximately 50% in PGE2 generation, we find significant residual levels of PGE2 mPGES1-deficient animals in the basal state. Our findings are production in the absence of mPGES1. Moreover, we and oth- consistent with another recent article that showed a marked ers have previously shown that ductus arteriosus closure is reduction of urinary PGE2 metabolite excretion, measured by dependent on activation of the EP4 for PGE2 (43,44), gas chromatography–mass spectrometry, in mPGES1-deficient yet mPGES1-deficient mice undergo normal closure of the duc- mice compared with controls (38). Although the magnitude of tus arteriosus (21,32). Taken together, these data indicate that the impact of mPGES1 deficiency was similar across genders, in there are alternative, mPGES1-independent pathways for PGE2 Ϫ Ϫ Ϫ Ϫ mPges1 / female but not mPges1 / male mice, there was a synthesis in the intact animal. The nature of these pathways is proportional increase in excretion of other prostanoids, includ- not clear, but our unpublished data suggest that another puta- J Am Soc Nephrol 18: 1466–1475, 2007 mPGES1 and Kidney Function 1473

tive PGE synthase, mPGES2, does not make a significant con- ment in PGE2 (53). Our studies indicate that mPGES1 is the tribution to PGE2 generation in vivo (Jania et al., submitted). primary source of furosemide-stimulated PGE2 release, because

Among its wide range of physiologic actions, PGE2 has a the enhancement of PGE2 generation after furosemide is virtu- number of effects that could affect BP homeostasis. It is a potent ally abolished in mPGES1-deficient mice. Blockade of NKCC2 vasodilator in both the peripheral (45) and renal vasculature by furosemide stimulates prostanoid generation in part via

(46). Moreover, PGE2 has potent natriuretic effects in the kidney upregulation of COX-2 (54). Our data suggest that mPGES1 is through direct epithelial effects on salt and water transport and also linked to this pathway and, at least in males, directly as a consequence of its hemodynamic actions to increase renal contributes to diuresis and natriuresis. However, we find no blood flow (3,8). Accordingly, it has been suggested that inhi- effect of furosemide on mPges1 expression in the kidney, sug- bition of PGE2 could explain the actions of NSAID and COX-2 gesting that stimulation of prostanoid excretion by furosemide inhibitors to cause edema and hypertension. However, PGE2 is occurs at a level upstream of mPGES1. a potent secretogogue for renin, and the consequences of acti- mPGES1 is also essential for furosemide-stimulated PGE2 vation of the renin-angiotensin system by PGE2 are in direct generation in female mice. However, this augmentation of opposition to its effects on renal vasculature and epithelia that PGE2 generation is not required for the full diuretic response in lead to vasodilation and natriuresis. However, our data do not female mice. Because COX inhibition reduces the extent of support a role for mPGES1 in furosemide-stimulated renin furosemide diuresis in WT female mice, there are at least two release. These findings are consistent with previous studies that possible explanations for the preserved furosemide response in suggested that PGI2, acting through the I-prostanoid (IP) recep- female mPGES1-deficient mice. First, despite their failure to tor, may be the critical pathway mediating stimulation of renin augment PGE2 excretion with furosemide, the levels of PGE2 with NKCC2 blockade (47). excretion in female mPGES1-deficient mice exceed those seen in

The role of individual pathways for PGE2 synthesis in regu- male WT mice after furosemide administration. Thus, there lating BP is not known. Our data suggest that mPGES1 does not may be a threshold level of PGE2 excretion required to facilitate play a major role in BP homeostasis, because BP was virtually diuresis with furosemide, and this level may be exceeded even ϩ ϩ Ϫ Ϫ identical in mPges1 / and mPges1 / mice across a range of in the basal state in mPGES1-deficient female mice. Alterna- dietary sodium intakes. This finding is consistent with the tively, the increase in PGI2 generation that is seen in the female recent report by Cheng et al. (38), who also found no difference mPGES1-deficient mice may act as a compensatory mechanism Ϫ/Ϫ in BP between mPges1 and WT control mice on either nor- to blunt the impact of their failure to augment PGE2 generation mal- or high-salt diet. Furthermore, our previous studies im- after furosemide. plicated alterations in PGI2 in NSAID-associated hypertension Until the recent development of mPGES1-deficient mouse (48). Nonetheless, it is still possible that mPGES1 might have lines (21,32), very little was known about the in vivo functions Ϫ Ϫ played a role in modulating BP and vascular tone on a different of the different forms of PGES. Studies using mPges1 / mice genetic background or in models of hypertension. indicate that mPGES1 is responsible for the late phase of PGE2 Among its actions in the kidney, PGE2 may influence water synthesis during inflammation via a pathway that depends on permeability in the distal nephron by opposing the actions of COX-2 (21,32). Febrile responses to LPS are also mediated by the Gs-linked vasopressin receptor (49). Because of its linkage mPGES1 (33), and mPGES1 is the source of PGE2 that facilitates to inhibitory G-protein (Gi) and its high levels of expression in acute inflammatory pain (21). Expression of mPGES1 is mark- renal medulla, the EP3 receptor was suggested as a candidate to edly upregulated in synovial tissues in arthritis (12,55), contrib- mediate the actions of PGE2 in opposing vasopressin-stimu- uting to the pathogenesis of joint inflammation and swelling lated water flow (50), yet in a previous study, we found only (21). Taken together, these observations suggest that inhibition subtle changes in urinary concentrating and diluting mecha- of mPGES1 might be useful for the treatment of inflammation nisms in EP3-deficient mice (51). In these studies, we find that and pain, providing a more precise therapeutic target than the absence of mPGES1 has no effect on urinary concentrating NSAID or COX-2 inhibitors. Nonetheless, the utility of this capacity. Accordingly, insofar as PGE2 may modulate vasopres- approach would depend on whether some of the untoward sin actions in the intact animal, mPGES1 is not absolutely effects that are associated with global COX inhibition, such as required. It is possible that the residual level of PGE2 generated gastrointestinal toxicity, hypertension, and adverse renal ef- in kidneys of mPGES1 mice is sufficient to fulfill this function. fects, would be avoided by specific mPGES1 inhibition. Our Alternatively, the capacity of PGE2 to influence vasopressin findings predict that there would be a preservation of a residual actions in vivo may be less robust than suggested by studies in capacity for PGE2 generation in the kidney with mPGES1 inhi- cell culture or isolated perfused nephron segments (50,52). bition that might protect against renal complications of COX However, alterations in urinary concentrating capacity did not inhibition, such as hypertension. However, some resistance to become apparent in cytosolic (cPLA2)-defi- loop diuretics might be expected. cient mice until Ͼ320 d of age (24), and our studies were carried out in mice that were Ͻ6 mo of age. Evidence from clinical studies indicates a contribution of Acknowledgments PGE2 to the sodium excretion that is induced by furosemide This work was supported by National Institutes of Health grant (6,53). For example, in healthy humans, a correlation was found DK069896 and by funding from the Medical Research Service of the between sodium excretion and the furosemide-induced incre- Veterans Administration. 1474 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 1466–1475, 2007

We acknowledge outstanding administrative support from Norma gawa K, Kojima M, Ito S, Watanabe K: Identification and Turner. characterization of a novel type of membrane-associated prostaglandin E synthase. Biochem Biophys Res Commun Disclosures 291: 884–889, 2002 None. 16. Tanioka T, Nakatani Y, Semmyo N, Murakami M, Kudo I: Molecular identification of cytosolic prostaglandin E2 syn- thase that is functionally coupled with cyclooxygenase-1 in References immediate prostaglandin E2 biosynthesis. J Biol Chem 275: 1. Bonvalet JP, Pradelles P, Farman N: Segmental synthesis 32775–32782, 2000 and actions of along the nephron. Am J 17. Fuson AL, Komlosi P, Unlap TM, Bell PD, Peti-Peterdi J: Physiol 253: F377–F387, 1987 Immunolocalization of a microsomal prostaglandin E syn- 2. Pallone TL: Vasoconstriction of outer medullary vasa recta thase in rabbit kidney. Am J Physiol Renal Physiol 285: by angiotensin II is modulated by prostaglandin E2. Am J F558–F564, 2003 Physiol 266: F850–F857, 1994 18. Campean V, Theilig F, Paliege A, Breyer M, Bachmann S: 3. Roman RJ, Lianos E: Influence of prostaglandins on papil- Key enzymes for renal prostaglandin synthesis: Site-spe- lary blood flow and pressure-natriuretic response. Hyper- cific expression in rodent kidney (rat, mouse). Am J Physiol tension 15: 29–35, 1990 Renal Physiol 285: F19–F32, 2003 4. Breyer MD, Breyer RM: Prostaglandin E receptors and the 19. Vitzthum H, Abt I, Einhellig S, Kurtz A: Gene expression of kidney. Am J Physiol Renal Physiol 279: F12–F23, 2000 prostanoid forming enzymes along the rat nephron. Kidney 5. Guan Y, Zhang Y, Breyer RM, Fowler B, Davis L, Hebert Int 62: 1570–1581, 2002 RL, Breyer MD: Prostaglandin E2 inhibits renal collecting 20. Breyer MD, Jacobson HR, Breyer RM: Functional and mo- duct Naϩ absorption by activating the EP1 receptor. J Clin lecular aspects of renal prostaglandin receptors. JAmSoc Invest 102: 194–201, 1998 Nephrol 7: 8–17, 1996 6. Kirchner KA, Martin CJ, Bower JD: Prostaglandin E2 but 21. Trebino CE, Stock JL, Gibbons CP, Naiman BM, Wacht- not I2 restores furosemide response in indomethacin- mann TS, Umland JP, Pandher K, Lapointe JM, Saha S, treated rats. Am J Physiol 250: F980–F985, 1986 Roach ML, Carter D, Thomas NA, Durtschi BA, McNeish 7. Frishman WH: Effects of nonsteroidal anti-inflammatory JD, Hambor JE, Jakobsson PJ, Carty TJ, Perez JR, Audoly drug therapy on blood pressure and peripheral edema. LP: Impaired inflammatory and pain responses in mice Am J Cardiol 89: 18D–25D, 2002 lacking an inducible prostaglandin E synthase. Proc Natl 8. Breyer MD, Breyer RM: G protein-coupled prostanoid re- Acad Sci U S A 100: 9044–9049, 2003 ceptors and the kidney. Annu Rev Physiol 63: 579–605, 2001 22. Francois H, Athirakul K, Mao L, Rockman H, Coffman TM: 9. Murakami M, Nakatani Y, Tanioka T, Kudo I: Prostaglan- Role for thromboxane receptors in angiotensin-II-induced din E synthase. Prostaglandins Other Lipid Mediat 68–69: hypertension. Hypertension 43: 364–369, 2004 383–399, 2002 23. Ito M, Oliverio MI, Mannon PJ, Best CF, Maeda N, Smithies 10. Jakobsson PJ, Thoren S, Morgenstern R, Samuelsson B: O, Coffman TM: Regulation of blood pressure by the type Identification of human prostaglandin E synthase: A mi- 1A angiotensin II receptor gene. Proc Natl Acad Sci U S A crosomal, glutathione-dependent, inducible , con- 92: 3521–3525, 1995 stituting a potential novel drug target. Proc Natl Acad Sci 24. Downey P, Sapirstein A, O’Leary E, Sun TX, Brown D, USA96: 7220–7225, 1999 Bonventre JV: Renal concentrating defect in mice lacking 11. Thoren S, Weinander R, Saha S, Jegerschold C, Pettersson group IV cytosolic phospholipase A(2). Am J Physiol Renal PL, Samuelsson B, Hebert H, Hamberg M, Morgenstern R, Physiol 280: F607–F618, 2001 Jakobsson PJ: Human microsomal prostaglandin E syn- 25. Oliverio MI, Delnomdedieu M, Best CF, Li P, Morris M, thase-1: Purification, functional characterization, and pro- Callahan MF, Johnson GA, Smithies O, Coffman TM: Ab- jection structure determination. J Biol Chem 278: 22199– normal water metabolism in mice lacking the type 1A 22209, 2003 receptor for ANG II. Am J Physiol Renal Physiol 278: F75– 12. Mancini JA, Blood K, Guay J, Gordon R, Claveau D, Chan F82, 2000 CC, Riendeau D: Cloning, expression, and up-regulation of 26. Heid C, Stevens J, Livak K, Williams PM: Real time quan- inducible rat prostaglandin E synthase during lipopolysac- titative PCR. Genome Res 6: 986–994, 1996 charide-induced pyresis and adjuvant-induced arthritis. 27. Bustin S: Absolute quantification of mRNA using real-time J Biol Chem 276: 4469–4475, 2001 reverse transcription PCR assays. J Mol Endocrinol 25: 169– 13. Murakami M, Naraba H, Tanioka T, Semmyo N, Nakatani 193, 2000 Y, Kojima F, Ikeda T, Fueki M, Ueno A, Oh S, Kudo I: 28. Bell PD, Lapointe J, Peti-Peterdi J: Macula densa cell sig- Regulation of prostaglandin E2 biosynthesis by inducible naling. Annu Rev Physiol 65: 481–500, 2003 membrane-associated prostaglandin E2 synthase that acts 29. Tilley SL, Coffman TM, Koller BH: Mixed messages: Mod- in concert with cyclooxygenase-2. J Biol Chem 275: 32783– ulation of inflammation and immune responses by prosta- 32792, 2000 glandins and thromboxanes. J Clin Invest 108: 15–23, 2001 14. Murakami M, Nakashima K, Kamei D, Masuda S, Ishikawa 30. Smith WL, Garavito RM, DeWitt DL: Prostaglandin endo- Y, Ishii T, Ohmiya Y, Watanabe K, Kudo I: Cellular pros- peroxide H synthases (cyclooxygenases)-1 and -2. J Biol taglandin E2 production by membrane-bound prostaglan- Chem 271: 33157–33160, 1996 din E synthase-2 via both cyclooxygenases-1 and -2. J Biol 31. Narumiya S, Sugimoto Y, Ushikubi F: Prostanoid recep- Chem 278: 37937–37947, 2003 tors: Structures, properties, and functions. Physiol Rev 79: 15. Tanikawa N, Ohmiya Y, Ohkubo H, Hashimoto K, Kan- 1193–1226, 1999 J Am Soc Nephrol 18: 1466–1475, 2007 mPGES1 and Kidney Function 1475

32. Uematsu S, Matsumoto M, Takeda K, Akira S: Lipopo- 44. Nguyen M, Camenisch T, Snouwaert JN, Hicks E, Coffman lysaccharide-dependent prostaglandin E(2) production is TM, Anderson PA, Malouf NN, Koller BH: The prostaglan- regulated by the glutathione-dependent prostaglandin E(2) din receptor EP4 triggers remodelling of the cardiovascular synthase gene induced by the Toll-like receptor 4/MyD88/ system at birth. Nature 390: 78–81, 1997 NF-IL6 pathway. J Immunol 168: 5811–5816, 2002 45. Audoly LP, Tilley SL, Goulet J, Key M, Nguyen M, Stock 33. Engblom D, Saha S, Engstrom L, Westman M, Audoly LP, JL, McNeish JD, Koller BH, Coffman TM: Identification of Jakobsson PJ, Blomqvist A: Microsomal prostaglandin E specific EP receptors responsible for the hemodynamic synthase-1 is the central switch during immune-induced effects of PGE2. Am J Physiol 277: H924–H930, 1999 pyresis. Nat Neurosci 6: 1137–1138, 2003 46. Audoly LP, Ruan X, Wagner VA, Goulet JL, Tilley SL, 34. Guan Y, Zhang Y, Schneider A, Riendeau D, Mancini JA, Koller BH, Coffman TM, Arendshorst WJ: Role of EP(2) Davis L, Komhoff M, Breyer RM, Breyer MD: Urogenital and EP(3) PGE(2) receptors in control of murine renal distribution of a mouse membrane-associated prostaglan- hemodynamics. Am J Physiol Heart Circ Physiol 280: H327– din E(2) synthase. Am J Physiol Renal Physiol 281: F1173– H333, 2001 F1177, 2001 47. Fujino T, Nakagawa N, Yuhki K, Hara A, Yamada T, 35. Beck TR, Hassid A, Dunn MJ: The effect of arginine vaso- Takayama K, Kuriyama S, Hosoki Y, Takahata O, Tanigu- pressin and its analogs on the synthesis of prostaglandin chi T, Fukuzawa J, Hasebe N, Kikuchi K, Narumiya S, E2 by rat renal medullary interstitial cells in culture. J Phar- Ushikubi F: Decreased susceptibility to renovascular hy- macol Exp Ther 215: 15–19, 1980 pertension in mice lacking the prostaglandin I2 receptor IP. 36. Jackson BA: Prostaglandin E2 synthesis in the inner med- J Clin Invest 114: 805–812, 2004 ullary collecting duct of the rat: Implications for vasopres- 48. Francois H, Athirakul K, Howell D, Dash R, Mao L, Kim sin-dependent cyclic AMP formation. J Cell Physiol 129: HS, Rockman HA, Fitzgerald GA, Koller BH, Coffman TM: 60–64, 1986 Prostacyclin protects against elevated blood pressure and 37. Schlondorff D: Renal prostaglandin synthesis. Sites of pro- cardiac fibrosis. Cell Metab 2: 201–207, 2005 duction and specific actions of prostaglandins. Am J Med 49. Nadler SP, Zimpelmann JA, Hebert RL: PGE2 inhibits wa- 81: 1–11, 1986 ter permeability at a post-cAMP site in rat terminal inner 38. Cheng Y, Wang M, Yu Y, Lawson J, Funk CD, FitzGerald medullary collecting duct. Am J Physiol 262: F229–F235, GA: Cyclo-oxygenases, microsomal prostaglandin E syn- 1992 thase 1, and cardiovascular function. J Clin Invest 116: 50. Sonnenburg WK, Zhu JH, Smith WL: A prostaglandin E 1391–1399, 2006 receptor coupled to a pertussis toxin-sensitive guanine 39. Yokoyama C, Yabuki T, Shimonishi M, Wada M, Hatae T, nucleotide regulatory protein in rabbit cortical collecting Ohkawara S, Takeda J, Kinoshita T, Okabe M, Tanabe T: tubule cells. J Biol Chem 265: 8479–8483, 1990 Prostacyclin-deficient mice develop ischemic renal disor- 51. Fleming EF, Athirakul K, Oliverio MI, Key M, Goulet J, ders, including nephrosclerosis and renal infarction. Circu- Koller BH, Coffman TM: Urinary concentrating function in lation 106: 2397–2403, 2002 mice lacking EP3 receptors for prostaglandin E2. Am J 40. Zoja C, Perico N, Corna D, Benigni A, Gabanelli M, Morigi Physiol 275: F955–F961, 1998 M, Bertani T, Remuzzi G: Thromboxane synthesis inhibi- 52. Nakao A, Allen ML, Sonnenburg WK, Smith WL: Regula- tion increases renal prostacyclin and prevents renal disease tion of cAMP metabolism by PGE2 in cortical and medul- progression in rats with remnant kidney. J Am Soc Nephrol lary thick ascending limb of Henle’s loop. Am J Physiol 256: 1: 799–807, 1990 C652–C657, 1989 41. Boulet L, Ouellet M, Bateman KP, Ethier D, Percival MD, 53. Brater DC, Beck JM, Adams BV, Campbell WB: Effects of Riendeau D, Mancini JA, Methot N: Deletion of microso- indomethacin on furosemide-stimulated urinary PGE2 ex- mal prostaglandin E2 (PGE2) synthase-1 reduces inducible cretion in man. Eur J Pharmacol 65: 213–219, 1980 and basal PGE2 production and alters the gastric prosta- 54. Kammerl MC, Nusing RM, Richthammer W, Kramer BK, noid profile. J Biol Chem 279: 23229–23237, 2004 Kurtz A: Inhibition of COX-2 counteracts the effects of 42. Trebino CE, Eskra JD, Wachtmann TS, Perez JR, Carty TJ, diuretics in rats. Kidney Int 60: 1684–1691, 2001 Audoly LP: Redirection of eicosanoid metabolism in 55. Claveau D, Sirinyan M, Guay J, Gordon R, Chan CC, mPGES-1-deficient macrophages. J Biol Chem 280: 16579– Bureau Y, Riendeau D, Mancini JA: Microsomal prosta- 16585, 2005 glandin E synthase-1 is a major terminal synthase that is 43. Kobayashi T, Narumiya S: Function of prostanoid recep- selectively up-regulated during cyclooxygenase-2-depen- tors: Studies on knockout mice. Prostaglandins Other Lipid dent prostaglandin E2 production in the rat adjuvant-in- Mediat 68–69: 557–573, 2002 duced arthritis model. J Immunol 170: 4738–4744, 2003