J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Receptors in Podocytes

MARTIN BEK,* ROLF NUSING,¨ † PASCAL KOWARK,* ANNA HENGER,* PETER MUNDEL,‡ and HERMANN PAVENSTADT*¨ *Department of Medicine, Division of Nephrology, University of Freiburg, Germany; †Department of Medicine, University of Marburg, Germany; and ‡Division of Nephrology, Albert Einstein College of Medicine, Bronx, New York.

Abstract. participate in the regulation of impor- extracellular solution with a reduced ClϪ concentration (from tant glomerular functions and are involved in the pathogenesis 145 to 32 mM). PGE2 and the , but not of glomerular diseases. This study investigates the influence of the IP and the EP2 receptor agonist prostaglandins on membrane voltage, ion conductances, cAMP butaprost, induced a time- and concentration-dependent cAMP 2ϩ accumulation, and cytosolic calcium activity ([Ca ]i) in dif- accumulation in podocytes. In fura-2 fluorescence experi- ferentiated podocytes. (PGE2) caused a con- ments, PGE2, , PGF2␣, fluprostenol (a potent FP centration-dependent depolarization and an increase of the agonist), and U-46619 (a selective A2 agonist) Ϸ 2ϩ whole cell conductance in podocytes (EC50 50 nM). Com- induced a biphasic increase of [Ca ]i in 60 to 80% of podo- pared with PGE2, the EP2/EP3/EP4 receptor agonist 11-deoxy- cytes. In reverse transcription-PCR studies, podocyte mRNA PGE1 caused an equipotent depolarization, whereas the DP for the EP1,EP4, FP, and TP receptor could be amplified. receptor agonist BW 245 C, the EP1/EP3 receptor agonist These data indicate that in podocytes, PGE2 regulates distinct sulprostone, and the IP receptor agonist iloprost were at least cellular functions via the EP1 and EP4 receptor, thereby in- 2ϩ 100 to 1000 times less potent than PGE2. The EP2 receptor creasing [Ca ]i and cAMP, respectively. Furthermore, PGF2␣ 2ϩ agonist butaprost did not change membrane voltage of podo- and U-46619 increase [Ca ]i via their specific receptors. cytes. The depolarizing effect of PGE2 was increased in an

Within the kidney, prostaglandins have been implicated to focal segmental glomerulosclerosis (7). Podocytes are also regulate important cellular functions in a variety of different believed to initiate and maintain the progression of glomeru- cell types through specific prostanoid receptors. They influ- losclerosis after subtotal nephrectomy, a model of chronic ence glomerular hemodynamics by increasing GFR and renal renal failure (8). blood flow, especially under conditions with an increased An imbalance of prostaglandin action seems to play a role in vascular tone (1,2). Their effects on GFR may partly be me- glomerular diseases, which are at least in part induced by diated by a change of the ultrafiltration coefficient (Kf). In this podocyte injury. Changes of prostaglandin action have been regard, it has been assumed that prostaglandins antagonize reported in several glomerular diseases such as membranous vasoactive hormone-induced mesangial cell contraction, which nephropathy (9), toxin-induced glomerular injury (10), and might lead to a decrease of Kf (3,4). development of glomerulosclerosis after subtotal renal ablation Knowledge about the influence of prostaglandins on cellular (11,12). functions of the cells that form the glomerular filtration barrier, Because prostaglandin-mediated changes of podocyte func- i.e., the glomerular endothelial cell and the podocyte, is lim- tion might be involved in the pathogenesis of these glomerular ited. Podocytes play a crucial role as a filter restricting the diseases, we attempted to investigate whether prostaglandins passage of macromolecules, and they maintain a large filtration influence cellular functions of differentiated cultured podo- surface through the slit membranes (5). It has been speculated cytes, and we characterized the prostanoid receptors involved. that they might modify Kf by contraction or dilation of their foot processes (6). Under pathologic conditions, podocytes are the target cell of injury in several forms of glomerular injury, Materials and Methods such as minimal change and membranous nephropathy, and Cell Culture In the present study, we used conditionally immortalized mouse podocyte clone cells harboring a temperature-sensitive mutant of immortalizing SV-40 large T- coupled with a promoter whose Received July 1, 1998. Accepted April 30, 1999. ␥ Correspondence to Dr. Hermann Pavensta¨dt, Medizinische Universita¨tsklinik, activity is enhanced by -interferon (13). Podocytes were maintained Abteilung IV, Labor C4, Hugstetter Strasse 55, 79106 Freiburg, Germany. in RPMI 1640 (Life Technologies, Eggenstein, Germany) supple- Phone: ϩ49 76 1270 3270; Fax: ϩ49 76 1270 3245; E-mail: paven@ mented with 10% fetal calf serum (Boehringer Mannheim, Mann- mm41.ukl.uni-freiburg.de heim, Germany), 100 U/ml penicillin, and 100 ␮g/ml streptomycin

1046-6673/1010-2084 (Life Technologies) in a humid atmosphere with 5% CO2. To prop- Journal of the American Society of Nephrology agate podocytes, the culture medium was supplemented with 10 U/ml Copyright © 1999 by the American Society of Nephrology mouse recombinant ␥-interferon (Sigma, Deisenhofen, Germany) to J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Prostanoid Receptors 2085

enhance the expression of T-antigen, and cells were cultivated at 33°C ny). The autofluorescence signal of cells, which had not been loaded (permissive conditions). To induce differentiation, podocytes were with fura-2, was measured, and the results and the noise were sub- cultured on type I collagen (Biochrom, Berlin, Germany) at 37°C tracted from the results obtained in fura-2-loaded cells. This had no without ␥-interferon (nonpermissive conditions) and with 1% fetal effect on the bandwidth of the measurements. A calibration of the calf serum for 2 to 3 wk. Arborized cells were immunologically fura-2 fluorescence signal was attempted at the end of each experi- characterized and stained positive for the Wilms’ tumor antigen and ment using the Ca2ϩ ionophore ionomycin (1 ␮M) and low and high 2ϩ 2ϩ synaptopodin, a characteristic marker for podocyte foot processes in Ca buffers. [Ca ]i was calculated from the fluorescence ratio vivo and in vitro (13,14). Podocytes between passages 15 and 25 were 340/380 nm according to the equation described by Grynkiewicz et al. used in all experiments. (18).

Patch-Clamp Experiments Expression of Prostanoid Receptor mRNA in Mouse The patch-clamp method used in these experiments has been de- Glomeruli and Podocytes scribed in detail (15,16). In brief, podocytes were placed in a bath The RNA preparation, the reverse transcription, and the PCR chamber on the stage of an inverted microscope, kept at 37°C, and amplification (PCR) were performed according to the method recently were superperfused with a phosphate-buffered Ringer’s-like solution. described (19). In brief, total RNA from cultured mouse podocytes 2ϩ To vary free Ca activity, the solutions were prepared according to was isolated with guanidinium/acid phenol/chloroform extraction, and ␤ established techniques with ethyleneglycol-bis( -aminoethyl ether)- the amount of RNA was measured by spectrophotometry. For first Ј Ј 2ϩ N,N,N N -tetra-acetic acid and nitrilotriacetic acid as Ca buffers. strand synthesis, 2 ␮g of total RNA from podocytes was mixed in 5ϫ 2ϩ The Ca activity was calculated from a standard equation and was reverse transcription buffer and incubated with5UofDNase I for 15 determined with a Ca2ϩ-selective electrode (Radiometer, Copenha- ϩ min. After termination, the reaction was completed with 0.5 mM gen, Denmark). In ion replacement studies, the extracellular Na or ␮ Ϫ ϩ dNTP, 0.5 M sequence-specific primers, 10 mM dithiothreitol, and Cl was replaced by N-methyl-D-glucamine (NMDG ) or gluconate, 200 U Superscript II transcriptase. (Reverse transcriptase was omitted respectively. The patch pipettes were filled with a solution containing in some experiments to control for amplification of contaminating (in mM): 95 K-gluconate, 30 KCl, 4.8 Na2HPO4, 1.2 NaH2PO4, 0.73 DNA.) The reverse transcription was performed at 42°C for 1 h, ␤ Ј CaCl2, 1.03 MgCl2, 1 ethyleneglycol-bis( -aminoethyl ether)-N,N - followed by a denaturation at 95°C for 5 min. cDNA was purified 2ϩ Ϫ7 tetra-acetic acid, and 5 D-glucose, pH 7.2; Ca activity 10 M. The from amplification reaction and dissolved in 30 ␮l of 10 mM Tris ⍀ patch pipettes had an input resistance of 2 to 3 M . A flowing (10 buffer, pH 8. PCR was performed in duplicate in a total volume of 20 ␮ l/h) KCl (2 M) electrode was used as a reference. The data were ␮l, each containing 2 ␮l of reverse transcription reaction and 18 ␮lof recorded using a patch-clamp amplifier (Fro¨be and Busche, Physiolo- PCR master mixture containing 10 pmol each of sense and antisense gisches Institut, Freiburg, Germany) and continuously displayed on a primer and 2.5 U of Taq DNA polymerase. The cycle profile included pen recorder. The access conductance (Ga) was monitored in most of denaturation for 30 s at 94°C, annealing for 30 s at 52°C, and the experiments by the method recently described (16). The mem- extension for 30 s at 72°C. The amplification products of 10 ␮lof brane voltage (Vm) of the cells was recorded continuously using the each PCR reaction were separated on a 1.5% agarose gel, stained with current clamp mode of the patch-clamp amplifier. To obtain the whole ethidium bromide, and visualized by ultraviolet irradiation. cell conductance (Gm), the voltage of the respective cell was clamped The primers were: in the voltage clamp mode (V )toV . Starting from this value, the Ј Ј c m EP1 receptor [D16338]: f-5 -GCTGTACGCCTCGCATCGTGG-3 whole cell current was measured by depolarizing or hyperpolarizing r-5Ј-CCTTGAGCCCAGCCCAGAATG-3Ј V in steps of 10 mV to Ϯ40 mV. G was calculated from the Ј Ј c m EP2 receptor [D50589]: f-5 -CCTGCCGCTGCTCAACTACG-3 measured whole cell current (I), Ga and Vc using Kirchhoff’s and r-5Ј-GTCTCCTCTGCCATCGAAGTCCTC-3Ј Ohm’s laws. Ј Ј EP3 receptor [D10204]: f-5 -GCTGGCTCTGGTGGTGACCTT-3 r-5Ј-AAGCCAGGCGAACTGCAATTAGAA-3Ј Ј Ј Measurements of Intracellular cAMP EP4 receptor [D13458]: f-5 -TGCTCATCTGCTCCATTCGCG-3 Podocytes were cultured in 6-well plates. They were kept at 37°C r-5Ј-CTGCTGATCTCCTTTAACTCCC-3Ј and rinsed with a physiologic Ringer’s-type solution. After preincu- FP receptor [U47287] : f-5Ј-CTTCCTTCCTGCTTTTGGCT-3Ј bation with 0.5 M 3-isobutyl-I-methylxanthine (Research Biochemi- r-5Ј-GTTCTGGACGCTTGGATTTG-3Ј cals Inc., Cologne, Germany) for 5 min, cells were exposed to the TP receptor [D10849]: f-5Ј-CCTTGTTCTCACCGACTTCC-3Ј added agents. To terminate the assay, the supernatants were rapidly r-5Ј-GCTGAACCATCATCTCCACC-3Ј removed and cells were rinsed with ice-cold ethanol (70 g/L). After an IP receptor [D26157]: f-5Ј-TTGCCTTGCCTTCCATCTAC-3Ј ethanol extraction, cAMP concentrations were measured with an r-5Ј-CAACCAGAACTTGAGGCGTT-3Ј -linked immunosorbent assay (Amersham Buchler, Braun- The National Center for Biotechnology Information accession no. schweig, Germany). of the respective nucleotide sequence is given in brackets. PCR amplification of reverse transcription (RT) reactions without reverse Measurements of Intracellular Ca2ϩ Activity transcriptase revealed no PCR product, thereby excluding amplifica- tion of genomic DNA. Identity of amplification products was deter- Measurements of intracellular Ca2ϩ activity ([Ca2ϩ] ) with the i mined by dideoxy sequencing. Ca2ϩ sensitive dye fura-2 (Sigma) were performed in podocytes on an inverted fluorescence microscope (17). The system allows fluores- cence measurements at the single cell level at three excitation wave- Chemicals

lengths. The field of measurement could be chosen by an adjustable The following agents were used: (PGE1), 11- ␮ pinhole between 2 and 300 m diameter. A time resolution of up to deoxy-PGE1, BW 245 C, sulprostone, AH-6809, SC 51089, U-46619, 200 Hz was achieved using a high-speed filter wheel and a single fluprostenol, (all from Cayman Chemical, Ann Arbor, MI), PGE2, photon counting tube (Hamamatsu H63460-04, Herrsching, Germa- PGF2␣, forskolin (Sigma), iloprost (kindly donated by Schering, Ber- 2086 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2084–2093, 1999 lin, Germany), butaprost (kindly donated by Dr. Stu¨nkel, Bayer AG, Leverkusen, Germany).

Statistical Analyses The data are given as mean values Ϯ SEM, where n refers to the number of experiments. The average of the effect of the agonist before and after the experiment was taken as control. Paired t test was used to compare mean values within one experimental series. A P value Յ0.05 was considered statistically significant.

Results Prostaglandins Depolarize the Membrane Voltage of Podocytes In 106 experiments in the slow whole cell configuration, a Ϫ Ϯ stable resting Vm of 59 1 mV was obtained for the control period. Addition of PGE2 caused a long-lasting and reversible ϭ depolarization of podocytes (n 106). PGE2 (100 nM) depo- larized podocytes by 14 Ϯ 1mV(n ϭ 54). During the depo- larization, the conductances for the inward and outward current increased significantly from 2.0 Ϯ 0.3 to 2.6 Ϯ 0.3 nS and 1.7 Ϯ 0.2 to 2.3 Ϯ 0.3 nS, respectively (n ϭ 50). Figure 1 shows a typical recording of the effect of 100 nM PGE2 on Vm and Gm of a podocyte.

Effect of Various Prostanoid Agonists on Membrane Voltage of Podocytes To characterize the prostanoid receptor involved in the re- sponse to PGE2, the Vm response to different prostaglandins and synthetic prostaglandin agonists was tested (20,21). Figure

2 shows the concentration response curves of the Vm response to PGE1, PGE2, 11-deoxy-PGE1 (EP2/EP3/EP4-receptor ago- nist), BW 245 C (DP receptor agonist), and sulprostone (EP1/ EP3 receptor agonist). Compared with PGE2, PGE1 induced a slightly stronger depolarization. The depolarizing effect of

PGE2 and 11-deoxy PGE1 were equipotent, whereas the DP receptor-selective agonist BW 245 C and the EP1/EP3 receptor- selective agonist sulprostone were less potent than PGE2. One Figure 1. Typical recording of the effect of prostaglandin E2 (PGE2) micromolar iloprost ( agonist) depolarized podo- (100 nM) on membrane voltage (Vm, top panel) and the whole cell Ϯ ϭ ␮ conductance (G , bottom panel) in podocytes. To obtain G , the cytes by 5 1mV(n 6), whereas 1 M of the EP2 receptor m m agonist butaprost had no effect on membrane voltage of podo- voltage of the respective cell was clamped to the resting Vm. Starting ϭ at this value, the whole cell current was measured by depolarizing or cytes (n 5) (data not shown). Ϯ hyperpolarizing Vm in steps of 10 mV to 40 mV. Influence of a Reduced Extracellular ClϪ or Naϩ Concentration on the Depolarization Induced by PGE2 ϩ Under resting conditions, PGE2 depolarized podocytes by lular Na , the effect of PGE2 was not inhibited, i.e., PGE2 (100 13 Ϯ 1 mV. Reduction of the extracellular ClϪ concentration nM) depolarized podocytes by 20 Ϯ 2 mV (data not shown). (from 147 to 32 mM) caused a slight hyperpolarization by 8 Ϯ 1 mV. Paired experiments showed that the depolarizing effect Prostaglandins Stimulate cAMP Accumulation in Ϫ of PGE2 in the presence of 32 mM Cl (100 nM) was signif- Podocytes Ϯ Ϯ icantly augmented from 13 1to27 3 mV (six paired Because multiple PGE2 receptors exist that are coupled to experiments) (Figure 3). different signaling pathways, we first investigated the effect of

Next, we examined the effect of a removal of the extracel- PGE2 on intracellular cAMP. Incubation of podocytes with 100 ϩ lular Na on the Vm response to PGE2 in seven paired exper- nM PGE2 resulted in a time-dependent accumulation of cAMP iments. Under resting conditions, PGE2 depolarized podocytes (Figure 4A). The effect of PGE2 on cAMP accumulation was by 13 Ϯ 1 mV. When Naϩ was removed from the extracellular concentration-dependent (Figure 4B). ϩ bath solution (0 mM versus 145 mM Na ), Vm hyperpolarized Figure 5 shows the effect of different prostanoid agonists on Ϯ ϭ transiently by 4 1mV(n 7). In the absence of extracel- cAMP increase in podocytes. Like PGE2, PGE1 (50 nM), J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Prostanoid Receptors 2087

Figure 2. Concentration–response curves of the effect of prostanoid Figure 3. Summary of the effect of a reduction of the extracellular ClϪ concentration (Cl , from 145 to 32 mM). n ϭ 6 for each data agonists on the change of membrane voltage (Vm). (No. of experi- 32 ments for 0.1, 1, 10, 100, 1000, and 10,000 nM, respectively, for point.

PGE1:5,5,7,9,6,0;PGE2: 6, 4, 5, 54, 5, 3; 11-deoxy-PGE1:5,5, 6, 6, 5, 3; BW 245 C: 0, 0, 3, 6, 4, 6; sulprostone: 0, 0, 0, 3, 4, 3). in podocytes (change of the fluorescence ratio from 1.0 Ϯ 0.1 to 1.5 Ϯ 0.2, n ϭ 4). In the continuous presence of 100 ␮M 2ϩ ␮ AH-6809, the [Ca ]i response to 0.1 M PGE2 was com- ␮ 11-deoxy-PGE1 (500 nM), BW 245 C (10 M), and sulpros- pletely inhibited After removal of AH-6809 and rinsing cells ␮ 2ϩ tone (10 M) stimulated an increase of cAMP accumulation, with control solution, PGE2 failed to increase [Ca ]i, indicat- whereas iloprost (1 ␮M) and butaprost (1 ␮M) did not increase ing that the effect of AH-6809 was irreversible (n ϭ 4, data not cAMP in podocytes (n ϭ 6 for each agonist). shown). Lower concentrations of AH-6809 (10 ␮M) did not 2ϩ ␮ change resting [Ca ]i. In the presence of 10 M AH-6809, the 2ϩ 2ϩ ␮ Effect of on [Ca ]i in Podocytes [Ca ]i response to 0.1 M PGE2 was significantly inhibited Ϯ ϭ In 113 of 185 experiments, addition of PGE2 (10 nM to 10 by 39 14% (n 10). After rinsing cells with control ␮ 2ϩ M) resulted in a reversible and concentration-dependent in- solution, a following [Ca ]i response to PGE2 was still re- 2ϩ crease of [Ca ]i in podocytes, whereas stimulation of cAMP duced. Figure 7A shows an original experiment of the effect of ␮ 2ϩ ϭ 2ϩ production by 1 M forskolin had no effect on [Ca ]i (n 5, PGE2 on [Ca ]i in the absence and presence of AH-6809, and 2ϩ data not shown). The [Ca ]i increase induced by PGE2 and all Figure 7B summarizes the data. other tested prostanoids (see below) was biphasic with a tran- PGF2␣ and the potent FP agonist fluprostenol (both 10 nM to 2ϩ 2ϩ ␮ sient [Ca ]i peak, which was followed by a [Ca ]i plateau. In 10 M [20]) increased reversibly and concentration-dependently 2ϩ these and the following experiments, cells that did not respond [Ca ]i in podocytes in 81 of 109 experiments (Figure 8, A and to PGE2 or different prostanoids responded to other vasoactive B). ϭ ϭ agonists such as (n 58) or extracellular ATP (n Figure 9A shows that U-46619, a selective 2ϩ 49). receptor agonist (20), reversibly increased [Ca ]i in a podo- 2ϩ In 12 of 20 experiments, a [Ca ]i increase was observed cyte. This could be reproduced in 34 of 54 cells. Figure 9A also ␮ after the addition of 1 M of the EP1/EP3 receptor agonist shows that in contrast to PGE2 and fluprostenol, after removal sulprostone (20). Figure 6A shows a fluorescence recording of of U-46619 and rinsing cells with a Ringer’s-like solution for 2ϩ 2ϩ the effect of PGE2 and sulprostone on [Ca ]i in podocytes, up to 10 min, U-46619 did not evoke a second [Ca ]i re- and Figure 6B summarizes the data. sponse (n ϭ 5). In contrast, the addition of fluprostenol after a 2ϩ To identify the EP receptor involved in the PGE2-mediated U-46619-mediated [Ca ]i increase still led to a second 2ϩ 2ϩ ϭ [Ca ]i response, experiments with the putative EP1 receptor [Ca ]i response (n 5). Figure 9B shows the concentration antagonists SC 51089 and AH-6809 (20) were performed. SC dependence of the effect of U-46619. 2ϩ 51089 itself increased [Ca ]i in podocytes (change of the Ϯ Ϯ ␮ ϭ fluorescence ratio from 1.1 0.1 to 1.4 0.2 [1 M, n 8] Podocytes Express mRNA for the EP1,EP4, FP, and and from 1.3 Ϯ 0.2 to 2.0 Ϯ 0.4 [10 ␮M, n ϭ 6]). SC 51089 TP Receptor ␮ 2ϩ (1 or 10 M) did not inhibit PGE2-mediated [Ca ]i response In another approach, we studied the expression of the EP ␮ 2ϩ (data not shown). AH-6809 (100 M) itself increased [Ca ]i receptor subtypes in mouse podocytes on the mRNA level. 2088 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2084–2093, 1999

Figure 5. Summary of the effect of iloprost (1 ␮M), butaprost (1 ␮M), ␮ ␮ sulprostone (10 M), BW 245 C (10 M), PGE1 (50 nM), and ϭ 11-deoxy-PGE1 (500 nM) on cAMP accumulation in podocytes (n 6 for each data point). After preincubation with 0.5 M IBMX for 5 min, podocytes were exposed to the indicated agent and the assay was terminated after an incubation time of 5 min. n ϭ 6 for each data point.

to any of the EP receptors, indicating an amplification artifact. Even after 40 cycles, an IP receptor could not be detected in podocytes by RT-PCR. However, we detected the IP receptor in mouse aorta and mouse spleen, indicating that the primers chosen were able to amplify the IP receptor (data not shown).

Discussion Prostaglandins are produced within the glomerulus and are known to modulate glomerular functions by acting locally on cells in which they are synthesized (22). Recently, we have shown that the prostaglandin-generating cyclooxy- genase-1 (COX-1) and -2 (COX-2) are primar-

Figure 4. PGE2 stimulates cAMP accumulation in podocytes. (A) ily expressed in the podocyte of the human kidney (23). Very Time dependence of cAMP increase induced by PGE2 (100 nM). recently it has been demonstrated that COX-2 expression is Mean values Ϯ SEM. Podocytes were preincubated with 0.5 M exclusively enhanced in podocytes during the course of Thy-1 3-isobutyl-I-methylxanthine (IBMX) for 5 min and stimulated with nephritis, suggesting that mesangial cell injury stimulates ϭ PGE2 for the indicated time. n 3 for each data point. (B) Concen- COX-2-induced prostaglandin synthesis in podocytes (24). The tration–response curve of the cAMP increase induced by PGE (n ϭ 2 mechanism of proteinuria in Thy-1 nephritis is not clearly 3 to 15). understood, but it may be speculated that the increase of prostanoid synthesis in podocytes might, in an autocrine man- ner, change their cellular functions leading to an impairment of RNA was analyzed by RT-PCR technique using sequence- podocyte barrier function. The present study demonstrates that specific primers to the different EP receptor subtypes. Figure prostaglandins regulate cellular functions of the podocyte via 10 shows ethidium bromide-stained agarose gel electrophoresis specific prostanoid receptors. Prostanoid receptor agonists de-

of PCR products for the EP1 to EP4, FP, and TP receptor in polarized podocytes and increased their whole cell conduc- mouse podocytes. Expression of ␤-actin was used as an inter- tance. The depolarization was probably mediated by an open- Ϫ nal standard. Sequence analysis of the resulting amplification ing of a Cl conductance because the effect of PGE2 on products revealed the sequence identity for the fragments. An membrane voltage of podocytes was accompanied by an in-

EP2 receptor could not be detected in cultured podocytes (lane crease of the whole cell conductance, and PGE2-induced de- Ϫ 2). The faint band in the EP3 lane had no polarization was augmented in a solution with a reduced Cl J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Prostanoid Receptors 2089

␮ Figure 7. The EP1 AH-6809 (AH, 10 M) inhibits 2ϩ ␮ Figure 6. PGE2 and the EP1/EP3 receptor agonist sulprostone increase the [Ca ]i response to PGE2 (0.1 M). (A) Original recording of the 2ϩ 2ϩ the cytosolic calcium activity ([Ca ]i) in podocytes. (A) Fluores- effect of PGE2 on [Ca ]i in the absence and presence of AH-6809. ␮ cence recording of the effect of PGE2 (1 M) and sulprostone (Sulp, (B) Summary of the experiments. ␮ 2ϩ 1 M) on [Ca ]i in podocytes. (B) Concentration–response curve of ␮ 2ϩ the effect of PGE2 and of the effect of 1 M sulprostone on [Ca ]i in podocytes. Number of observations is shown in parentheses. A star EP ) of the EP receptor have been characterized (20,22). To indicates significant differences between control and peak values. 4 identify the prostanoid receptor involved in the depolarization

induced by PGE2, we examined the influence of synthetic prostaglandin analogs on the membrane voltage response in

concentration. In contrast, the membrane voltage response to podocytes. Compared with PGE2, 11-deoxy PGE1,anEP2/EP3 ϩ PGE2 was unchanged in the absence of extracellular Na , and EP4 receptor agonist, was equipotent, whereas the DP ϩ indicating that it was not due to an influx of Na . A cAMP- receptor agonists BW 245 C, the EP1/EP3 receptor agonist and Ca2ϩ-dependent ClϪ conductance has been detected re- sulprostone, and the IP receptor agonist iloprost were much

cently in podocytes (25). Within the kidney, a PGE2-mediated less potent than PGE2. The selective EP2 receptor agonist activation of ClϪ channels has been demonstrated in principal butaprost did not depolarize podocytes. This indicates that the

cells (26), whereas up to now nothing is known about PGE2- depolarizing effect of PGE2 was most likely mediated by an Ϫ mediated regulation of Cl channels in glomerular cells. The EP4 receptor. Ϫ precise cellular role for the activation of the Cl conductance The suggested involvement of an EP4 receptor in PGE2 in podocytes is unknown, but it may be speculated that the action was further corroborated by measurements of cAMP. Ϫ activation of the Cl conductance by PGE2 contributes to the 11-deoxy PGE1 induced a marked stimulation of cAMP accu- cell volume regulation in these cells (27). mulation in podocytes, whereas butaprost had no effect. The

The prostanoid receptor family includes the so-called DP, existence of an EP4 receptor in podocytes was further con- EP, FP, IP, and TP receptors, which preferentially bind pros- firmed by the detection of EP4 receptor mRNA in podocytes, taglandin D (PGD), PGE, PGF, PGI, and thromboxane A whereas EP2 receptor mRNA could not be detected. These data (TXA), respectively. Moreover, four subtypes (EP1 through are in agreement with in situ hybridization studies by others, 2090 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2084–2093, 1999

Figure 8. PGF2␣ and fluprostenol, a potent FP agonist, increase Figure 9. U-46619, a selective thromboxane A2 receptor agonist, 2ϩ 2ϩ 2ϩ [Ca ]i in podocytes. (A) Original recording of the effect of PGF2␣ increases [Ca ]i in podocytes. (A) Original recording of the [Ca ]i ␮ 2ϩ ␮ and fluprostenol (both 1 M) on [Ca ]i in podocytes. (B) Concen- response to 0.1 M U-46619. After a control period, U-46619 did not 2ϩ tration–response curves of the agonists. induce a second [Ca ]i increase, whereas cells still responded to PGF2␣. (B) Concentration–response curve of the effect of U-46619 on 2ϩ [Ca ]i. which demonstrate that the EP4 receptor is highly expressed in the mouse and human glomerulus (28,29). The existence of the

EP4 receptor in podocytes suggests that it might, at least in In RT-PCR studies, a fragment for the EP4 and for the EP1 part, mediate the effects of PGE2 on glomerular functions. receptor could be amplified, whereas a fragment for the EP3 Micromolar concentrations of the DP receptor agonist BW receptor could not be detected. The possible existence of an

245 C (20) only had a weak effect on membrane voltage and EP1 receptor in podocytes was surprising because mouse and cAMP generation in podocytes, indicating that the effects of human glomerular mRNA for the EP1 receptor has not been PGE2 are not mainly mediated via a DP receptor. Future detected by in situ hybridization (29,30). The lack of expres- studies need to clarify whether this receptor is expressed in sion of the glomerular EP1 receptor in the latter studies might podocytes at all. be explained by the absence of EP1 mRNA under physiologic In the present study, the IP receptor agonist iloprost has no conditions, i.e., an expression of the EP1 receptor might occur effect on cAMP accumulation in podocytes, and an IP receptor only under pathologic or under in vitro conditions. Alterna- could not be detected in RT-PCR studies, indicating that cul- tively, it is possible that a low expression of EP1 receptor tured mouse podocytes do not possess an IP receptor. The mRNA in podocytes, which represent only a very small portion slight depolarizing effect of iloprost may be explained by a of the glomerular cell population, is hardly detectable by in situ small affinity of iloprost to other EP receptors (21). Very hybridization. Recently, it has been reported that the mouse recently in kidney slices an IP receptor has been demonstrated for the EP1 receptor and the PKN protein kinase overlap in human podocytes (30). The lack of expression of the IP (31). However, in this study the possibility that the EP1 am- receptor in mouse podocytes may be explained by species plification products in podocytes represented PKN protein differences. Alternatively, podocytes might lose the IP receptor kinase mRNA expression has been excluded because, com- under in vitro conditions. pared with the study by Båtshake et al., we used a different J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Prostanoid Receptors 2091

PGE2. The results are in agreement with reports in mesangial cells, in which PGE2 also induces a weak and inconsistent 2ϩ increase in [Ca ]i (32). To further characterize the EP receptor involved in the 2ϩ PGE2-mediated [Ca ]i response, experiments with the EP1/ EP3 receptor agonist sulprostone and two different EP1 recep- tor antagonists, SC 51089 and AH-6809, were performed (20). 2ϩ Like PGE2, sulprostone increased [Ca ]i in podocytes, indi- cating the existence of an EP1 or EP3 receptor. SC 51089 (1 and 10 ␮M) and higher concentrations of AH-6809 (100 ␮M) 2ϩ itself increased [Ca ]i, suggesting that the EP1 receptor an- tagonists had an intrinsic receptor activity or induced some nonspecific cellular effects. AH-6809 (10 ␮M) did not influ- 2ϩ 2ϩ ence [Ca ]i and slightly inhibited the [Ca ]i response to PGE2. These results are in agreement with binding studies showing that SC 51089 and AH-6809 have a very low affinity

to the EP1 receptor of the mouse, suggesting that there are species differences (21). Thus, although the EP1 antagonists 2ϩ had only small effects in this study, the [Ca ]i response to PGE2 may be—also because of the lack of an EP3 receptor expression in RT-PCR experiments—mediated by an EP1 re- ceptor.

It has been assumed that PGE2 acts predominantly as a glomerular vasodilator, but the ability of PGE2 to increase 2ϩ [Ca ]i in podocytes and mesangial cells (33) suggests an additional role for PGE2 as a glomerular vasoconstrictor. We recently reported that the ClϪ conductance of podocytes is 2ϩ regulated by cAMP and by [Ca ]i (25). Therefore, we cannot exclude that in some patch-clamp experiments the membrane Figure 10. Expression of mRNA for the EP through EP , FP, and TP 1 4 voltage change induced by PGE2 is partly due to an activation receptor in mouse podocytes. Fragments for the different prostanoid ofaCa2ϩ-dependent ClϪ conductance. receptor subtypes were amplified using specific primers and subjected Not only PGE , but also PGF ␣, fluprostenol (a highly to agarose gel electrophoresis. Lanes on the left and right side indicate 2 2 selective FP receptor agonist), and U-46619 (a TP receptor length. Note that bands with the expected sizes are only 2ϩ agonist) (20) increased [Ca ]i in podocytes. In contrast to present in the EP1 through EP4, FP, and TP lanes, but not in the EP2 and EP lanes. The faint band in the EP lane was identified as an PGE2 and the FP agonists, repeated addition of U-46619 could 3 3 2ϩ amplification artifact. not induce a second [Ca ]i response, suggesting that the TP receptor involved might be rapidly desensitized by the agonist. 2ϩ Compared with the FP agonists, the [Ca ]i response to approach of RT-PCR. Båtshake et al. used oligo-dT for reverse U-46619 was about 100 times more potent and, after addition transcription and thereby transcribed both the mRNA species of U-46619 cells, still responded to fluprostenol, suggesting from sense and from antisense strand. The following PCR that the effect of fluprostenol and U-46619 was mediated by resulted in the fragments as described by Båtshake et al. (31). different receptors. In addition, the results of the RT-PCR Instead of using this approach, we replaced oligo-dT by se- studies indicate that both an FP and a TP receptor are expressed quence-specific primers for reverse transcription, which tran- in podocytes. Little is known about the glomerular function of scribe only the mRNA corresponding to the sense strand but the FP receptor. It has been reported in mesangial cells and not the mRNA corresponding to the antisense strand. In addi- therefore it might be involved in the regulation of mesangial tion, the RNA was DNase-treated to exclude the possibility of cell contraction (32,33). accidental amplification of contaminating genomic DNA. Thromboxane A2 has been suggested to regulate the ultra-

The EP1 receptor is known to be coupled to phosphatidyl- filtration coefficient, and it is involved in a variety of glomer- 2ϩ inositol hydrolysis, resulting in an increase of [Ca ]i (20,22). ular diseases such as different forms of glomerulonephritis and To clarify the possible functional existence of an EP1 receptor diabetic nephropathy (34,35). In Heymann nephritis, an exper- 2ϩ in podocytes, measurements of [Ca ]i were performed in imental model of membranous nephropathy in which the podo- three different cloned podocyte cell lines. In some of the cyte is the primary target cell of injury, inhibition of throm- 2ϩ experiments, PGE2 failed to increase [Ca ]i in podocytes, but boxane A2 synthesis prevented alterations of glomerular in about 60% of the experiments, a weak, but concentration- morphology and proteinuria (36). The TP receptor in podo- 2ϩ dependent, biphasic [Ca ]i increase could be elicited by cytes might therefore be involved in the regulation of physio- 2092 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 2084–2093, 1999 logic functions like the control of the ultrafiltration coefficient Kriz W, Zeller R: Rearrangements of the cytoskeleton and cell and might also play a role during podocyte injury. contacts induce process formation during differentiation of con- 2ϩ ditionally immortalized mouse podocyte cell lines. Exp Cell Res The lack of a [Ca ]i response toward the agonists in some of the cells suggests that the expression of the receptor proteins 236: 248–258, 1997 is low. Alternatively, the signaling system activated by PGE , 14. Mundel P, Reiser J, Kriz W: Induction of differentiation in cultured 2 rat and human podocytes. J Am Soc Nephrol 8: 697–705, 1997 PGF2␣, and thromboxane A via their receptors may be inhib- 2 15. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ: Im- ited in some cells by a yet unknown mechanism. Future studies proved patch-clamp techniques for high-resolution current re- are needed to clarify the reasons for the unstable functional cording from cells and cell-free membrane patches. Pflu¨gers 2ϩ expression of the [Ca ]i mobilizing receptors. Arch 391: 85–100, 1981 In conclusion, the data indicate that prostanoids regulate 16. Greger R, Kunzelmann K: Simultaneous recording of the cell important cellular functions of podocytes via different prosta- membrane potential and properties of the cell attached membrane noid receptors: They depolarize podocytes, increase their of HT 29 colon carcinoma and CF-PAC cells. Pflu¨gers Arch 419: 2ϩ 209–211, 1991 cAMP, and stimulate [Ca ]i in a part of podocytes. The coexpression of the prostaglandin receptors might balance op- 17. Nitschke R, Fro¨be U, Greger R: Antidiuretic hormone acts via V1 receptors on intracellular calcium in the isolated perfused posite cellular effects of PGE2 in podocytes. The activation of the prostanoid receptors in podocytes might influence glomer- rabbit cortical thick ascending l imb. Pflu¨gers Arch 417: 622– 632, 1991 ular hemodynamics, and it might be an important event in the ϩ 18. Grynkiewicz G, Poenie M, Tsien RY: A new generation of Ca2 pathogenesis of proteinuria. indicators with greatly improved fluorescence properties. J Biol Chem 260: 3340–3350, 1985 Acknowledgments 19. Greiber S, Mu¨nzel T, Ka¨stner S, Mu¨ller B, Schollmeyer P, This work was supported by the Deutsche Forschungsgemeinschaft Pavensta¨dt H: NAD(P)H oxidase activity in cultured human (DFG Pa 483/1-5). We thank Temel Kilic for his excellent technical podocytes: Effects of adenosine triphosphate. Kidney Int 53: assistance. 654–663, 1998 20. Coleman RA, Smith WL, Narumiya S: VIII. International Union References of Pharmacology Classification of prostanoid receptors: Proper- 1. Garrick RE: The renal . In: Contemporary Issues in ties, distribution and structure of the receptors and their subtypes. Nephrology, Vol. 23, edited by Stein JH, New York, Churchill Pharmacol Rev 46: 205–229, 1994 Livingstone, 1991, pp 231–261 21. Kiriyama M, Ushikubi F, Kobayashi T, Hirata M, Sugimoto Y, 2. Schor N, Ichikawa I, Brenner BM: Mechanism of action of Narumiya S: binding specificities of the eight types and various hormones and vasoactive substances on glomerular ul- subtypes of the mouse prostanoid receptors expressed in Chinese trafiltration in rats. Kidney Int 20: 442–451, 1981 hamster ovary cells. Br J Pharmacol 122: 217–224, 1997 3. Andrews PM: Investigations of cytoplasmatic contractile ele- 22. Breyer MD, Jacobson HR, Breyer RM: Functional and molecular ments and cytoskeletal elements in the kidney glomerulus. Kid- aspects of renal prostaglandin receptors. J Am Soc Nephrol 7: ney Int 20: 549–562, 1981 8–17, 1996 4. Mene´P, Simonson MS, Dunn MJ: Physiology of the mesangial 23. Komhoff M, Gro¨ne HJ, Klein T, Seyberth HW, Nu¨sing RM: cell. Physiol Rev 69: 1347–1423, 1989 Localization of cyclooxygenase-1 and -2 in adult and fetal hu- 5. Drumond MC, Kristal B, Myers BD, Deen WM: Structural basis man kidneys: Implication for renal function. Am J Physiol 272: for reduced glomerular filtration capacity in nephrotic humans. F460–F468, 1997 J Clin Invest 94: 1187–1195, 1994 24. Hirose S, Yamamoto T, Feng L, Yaoita E, Kawasaki K, Goto S, 6. Drenckhahn D, Franke RP: Ultrastructural organization of con- Fujinaka H, Wilson CB, Arakawa M, Kihara I: Expression and tractile and cytoskeletal proteins in glomerular podocytes of localization of cyclooxygenase isoforms and cytosolic phospho- chicken, rat, and man. Lab Invest 59: 673–682, 1988 lipase A2 in Anti-Thy-1 glomerulonephritis. J Am Soc Nephrol 9: 7. Kerjaschki D: Dysfunction of cell biological mechanisms of 408–416, 1998 visceral epithelial cells (podocytes) in glomerular diseases. Kid- 25. Huber T, Gloy J, Henger A, Schollmeyer P, Greger R, Mundel P, ney Int 45: 300–313, 1994 Pavensta¨dt H: Catecholamines modulate podocyte function. JAm 8. Kriz W: Role of the podocyte in the pathogenesis of glomerulo- Soc Nephrol 9: 335–345, 1998

sclerosis. Kidney Blood Press Res 20: 180–183, 1997 26. Ling BN, Kokko KE, Eaton DC: Prostaglandin E2 activates 9. Stahl R, Adler S, Baker PJ, Chen YP, Pritzl PM, Couser WG: clusters of apical ClϪ channels in principal cells via a cyclic Enhanced prostaglandin formation in experimental membranous adenosine monophosphate-dependent pathway. J Clin Invest 93: nephropathy. Kidney Int 31: 1126–1131, 1987 829–837, 1994 10. Sraer JD, Baud L, Sraer J, Delarue F, Ardaillou R: Stimulation of 27. Lang F, Busch GL, Vo¨lk H: The diversity of volume regulatory

PGE2 synthesis by mercuric chloride in rat glomeruli and glo- mechanisms. Cell Physiol Biochem 8: 1–45, 1998 merular cells in vitro. Kidney Int 2: 63–68, 1982 28. Sugimoto Y, Namba T, Shigemoto R, Negishi M, Ichikawa A, 11. Heifets M, Morrissey JJ, Purkerson ML, Morrison AR, Klahr S: Narumiya S: Distinct cellular localization of mRNAs for three Effect of dietary lipids on renal function in rats with subtotal subtypes of receptor in kidney. Am J Physiol nephrectomy. Kidney Int 32: 335–341, 1987 266: F823–F828, 1994 12. Barcelli UO, Weiss M, Pollack VE: Effects of a dietary prosta- 29. Breyer MD, Davis L, Jacobson HR, Breyer RM: Differential glandin precursor on the progression of experimentally induced localization of prostaglandin E receptor subtypes in human kid- chronic renal failure. J Lab Clin Med 100: 786–797, 1982 ney. Am J Physiol 270: F912–F918, 1996 13. Mundel P, Reiser J, Borja AZM, Pavensta¨dt H, Davidson GR, 30. Komhoff M, Lesener B, Nakao K, Seyberth HW, Nu¨sing RM: J Am Soc Nephrol 10: 2084–2093, 1999 Characterization of Prostanoid Receptors 2093

Localization of the in human kidney. Kid- 34. Takahashi K, Schreiner GF, Yamashita K, Christman BW, Blair

ney Int 54: 1899–1908, 1998 I, Badr KF: Predominant functional roles for thromboxane A2 31. Båtshake B, Sundelin J: The mouse genes for the EP1 prostanoid and prostaglandin E2 during late nephrotoxic serum glomerulo- receptor and the PKN protein kinase overlap. Biochem Biophys nephritis in the rat. J Clin Invest 85: 1974–1982, 1990 Res Commun 227: 70–76, 1996 35. Craven PA, Melhem MF, DeRubertis FR: Thromboxane in the 32. Mene´P, Dubyak GR, Abboud HE, Scarpa A, Dunn M: Phos- pathogenesis of glomerular injury in diabetes. Kidney Int 42:

pholipase C activation by prostaglandins and thromboxane A2 in 937–946, 1992 cultured mesangial cells. Am J Physiol 255: F1059–F1069, 1988 36. Nagao T, Nagamatsu T, Suzuki Y: Effect of DP-1904, a throm-

33. Mene´P, Dunn M: Eicosanoids and the control of mesangial cell boxane A2 synthase inhibitor, on passive Heymann nephritis in contraction. Circ Res 62: 916–925, 1988 rats. Eur J Pharmacol 316: 73–80, 1996