Concerted EP2 and EP4 Receptor Signaling Stimulates Autocrine Prostaglandin E2 Activation in Human Podocytes

Glomerular hyperfiltration is an important mechanism in the development of albuminuria. During hyperfiltration, podocytes are exposed to increased fluid flow shear stress (FFSS) in Bowman’s space. Elevated Prostaglandin E2 (PGE2) synthesis and upregulated cyclooxygenase 2 (Cox2) are associated with podocyte injury by FFSS. We aimed to elucidate a PGE2 autocrine/paracrine pathway in human podocytes (hPC). We developed a modified liquid chromatography tandem mass spectrometry (LC/ESI-MS/MS) protocol to quantify cellular PGE2, 15-keto-PGE2, and 13,14-dihydro-15-keto-PGE2 levels. hPC were treated with PGE2 with or without separate or combined blockade of prostaglandin E receptors (EP), EP2, and EP4. Furthermore, the effect of FFSS on COX2, PTGER2, and PTGER4 expression in hPC was quantified. In hPC, stimulation with PGE2 led to an EP2- and EP4-dependent increase in cyclic adenosine monophosphate (cAMP) and COX2, and induced cellular PGE2. PTGER4 was downregulated after PGE2 stimulation in hPC. In the corresponding LC/ESI-MS/MS in vivo analysis at the tissue level, increased PGE2 and 15-keto-PGE2 levels were observed in isolated glomeruli obtained from a well-established rat model with glomerular hyperfiltration, the Munich Wistar Frömter rat. COX2 and PTGER2 were upregulated by FFSS. Our data thus support an autocrine/paracrine COX2/PGE2 pathway in hPC linked to concerted EP2 and EP4 signaling.


Introduction
Podocytes are terminally differentiated epithelial cells that form the third layer of the glomerular filter with their interdigitating foot processes [1]. Their high degree of differentiation permits podocytes to accomplish their highly specialized functions. However, it limits their regenerative capacity, making them particularly vulnerable to pathological conditions such as glomerular hyperfiltration. When nephron number is reduced, compensatory changes of the remaining functional nephrons lead to adaptation of glomerular hemodynamics, resulting in increased glomerular filtration rate (GFR) in After stimulation or inhibition experiments, supernatants were removed and stored at −80 • C. Cells were washed twice with cold phosphate buffered saline (PBS), PBS was completely aspirated from wells, and the 12-well plates were immediately stored at −80 • C until further processing. Before analysis, cells were scraped from plate and suspended in 500 µL water. A 50 µL aliquot was taken for total protein measurement following the Lowry protocol.
The total prostaglandins were released using phospholipase A2 from honey bee Apis Mellifera (Sigma-Aldrich, Taufkirchen, Germany) as described previously [46]. After pH adjustment to 6, acetic acid samples were extracted by solid phase extraction (SPE) using Bond Elute Certify II columns (Agilent Technologies, Santa Clara, CA, USA), which were preconditioned with 3 mL methanol, followed by 3 mL of 0.1 mol/L phosphate buffer containing 5% methanol (pH 6). SPE-columns were then washed with 3 mL methanol/H 2 O (40/50, v/v). For elution, 2 mL of n-hexane:ethyl acetate 25:75 Cells 2020, 9, 1256 5 of 19 with 1% acetic acid was used. The extraction was performed with an SPE Vacuum Manifold. The eluate was evaporated on a heating block at 40 • C under a stream of nitrogen to obtain a solid residue which was dissolved in 100 µL methanol/water 60:40 and transferred in an HPLC autosampler vial (HPLC, high performance liquid chromatography).
Experiments for analysis of prostaglandins in hPC consisted of n = 3-6 replicates per experimental group and. Experiments were performed once or in triplicate (on different cell passages and on different days) as indicated.
Rat glomeruli obtained by differential sieving of one kidney as described below were divided into 3 parts and stored at −80 • C until further analysis. One aliquot with approximately 1/3 total kidney was prepared as described for cells, but without application of phospholipase A2.
For rat plasma, 200 µL plasma were spiked with internal standard and BHT. In addition, 20 µL glycerol and 500 µL acetonitrile was added and shaken vigorously. pH was adjusted at 6 with 2 mL phosphate buffer (0.1 mol/L). The samples were centrifuged and the clear supernatant was extracted using SPE as described above.
Experiments for analysis of prostaglandins in rat glomeruli or plasma consisted of n = 8-10 glomerular isolated or plasma samples of n = 8-10 different animals per rat strain.

LC/ESI-MS/MS
The residues were analyzed using an Agilent 1290 HPLC system with binary pump, multisampler and column thermostat with a Zorbax Eclipse plus C-18, 2.1 × 150 mm, 1.8 µm column using a solvent system of aqueous acetic acid (0.05%) and acetonitrile. The elution gradient was started with 5% organic phase, which was increased within 0.5 min to 32%, 16 min to 36.5%, 20 min to 38%, 28 min to 98% and held there for 5 min. The flow rate was set at 0.3 mL/min, the injection volume was 20 µL. The HPLC was coupled with an Agilent 6495 Triplequad mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) with electrospray ionisation source. The source parameters were Drying gas: 115 • C/16 L/min, Sheath gas: 390 • C/12 L/min, Capillary voltage: 4300 V, Nozzle voltage: 1950 V, and Nebulizer pressure: 35 psi.
Analysis was performed with Multiple Reaction Monitoring in negative mode. For details, see Table S1. Unless stated otherwise, all solvents and chemicals were purchased from VWR International GmbH, Darmstadt, Germany.

FFSS
For FFSS, 1 × 10 5 6 × 10 5 cells where seeded on collagen IV coated Culture Slips ® (cat. no. CS-C/IV, Dunn Labortechnik GmbH, Asbach, Germany), which are glass slides coated with collagen type IV and rimmed with a 1.0 mm wide polytetrafluoroethylene border to limit cell culture growth to the portion of the slip exposed to fluid flow. FFSS experiments were performed as previously described with slight modifications [23]. The Streamer ® Shear Stress Device (cat. no. STR-400, Dunn Labortechnik GmbH, Asbach, Germany) was installed in a 38 • C incubator with 5% CO 2 and prepared as follows: 400 mL of PBS followed by RPMI-1640 medium were pumped through the device for approximately 10 min each. Prior to each change of content, flow direction was reversed to empty the tubes from the previous liquid. After washing, medium was replaced by 400 mL of new medium. The system was checked for leaks and air bubbles were eliminated. After preparation of the streamer, flow direction was again reversed until the streamer was half-filled by medium. Tubes were released from the pump, the system was taken out of the incubator, Culture Slips ® with hPC were inserted in the Streamer ® , and the system was placed back into the incubator. All 6 slots of the Streamer ® were filled to allow consistent flow. Based on previous research, we applied FFSS at 2 dynes/cm 2 for 2 h [23]. At the end of each experiment, flow rate was reversed and cells were released from the device. Control cells were put in the same incubator with the same medium but were not exposed to FFSS.

Animals
The MWF rat served as a model for CKD with albuminuria, while the spontaneously hypertensive rat (SHR) served as a control strain. SHR rats develop hypertension early in life but are resistant to albuminuria development as reviewed in [47].
Male rats at 8 weeks of age were deployed from our MWF/Rkb (RRID:RGD_724569, laboratory code Rkb https://www.nationalacademies.org/ilar/lab-code-database) and SHR/Rkb (RRID:RGD_631696, laboratory code Rkb https://www.nationalacademies.org/ilar/lab-code-database) colonies at Charité-Universitätsmedizin Berlin, Germany. Rats were kept under standard conditions as described previously [44]. All experimental work in rat models was performed in accordance with the guidelines of the Charité-Universitätsmedizin Berlin and the local authority for animal protection (Landesamt für Gesundheit und Soziales, Berlin, Germany) for the use of laboratory animals. The registration numbers for the rat experiments are G 130/16 (approved 2 August 2016) and T 0189/02 (approved 31 August 2018). Anesthesia was achieved by ketamine-xylazine (87 and 13 mg/kg body weight, respectively). Kidneys were obtained, decapsulated, and sieved using a 125 µm steel sieve (Retsch GmbH, Haan, Germany) rinsed by PBS. The filtrate was then placed on a 71 µm steel sieve (Retsch GmbH, Haan, Germany) and washed with PBS. Glomeruli were kept on the sieve and were separated from the flow-through. Glomeruli were rinsed off the sieve with PBS, centrifuged, snap-frozen, and stored at −80 • C until further processing. Plasma was obtained by retrobulbary punction or punction of vena cava and collected in ethylenediaminetetraacetic acid (EDTA)-containing vials, centrifuged at 2 min at 4 • C, and stored subsequently at −80 • C.

Statistics
Statistical analysis was conducted using GraphPad Prism 8.4.0 (GraphPad Software, San Diego, CA, USA). Normal distribution was tested with the Shapiro-Wilk test. Normally distributed data were compared either by unpaired, two-tailed Student's t-test or one-way ANOVA with Tukey's or Dunnett's multiple comparisons test as indicated. Multiple comparisons tests after one-way ANOVA were used to compare every mean to every other mean (Tukey' follow up test) or to a control mean (Dunnett's follow up test). Results not normally distributed were analyzed by Mann-Whitney test or Kruskal-Wallis test with Dunn's multiple comparisons test as indicated. Significance level was set at p < 0.05. Statistical details for specific experiment can be found within figures and figure legends.

PGE 2 Leads to EP2-and EP4-Dependent Increased cAMP Levels in Differentiated hPC
Stimulation of hPC with 100 nM PGE 2 led to an immediate time-dependent increase in intracellular cAMP levels detected after 1 min onward and retained at least until 40 min of PGE 2 stimulation ( Figure 1). As intracellular cAMP levels remained comparably high until 20 min of PGE 2 stimulation, this incubation time was chosen for subsequent cAMP measurements.
Analysis of EP expression on hPC revealed the presence of PTGER1, PTGER2, and PTGER4 mRNA in differentiated hPC ( Figure S3), which encode for EP1, EP2, and EP4, respectively. As only EP2 and EP4 are reported to mediate an increase in intracellular cAMP (reviewed in [15][16][17]), we next investigated the effect of pharmacological inhibition of EP2 and EP4 signaling on PGE 2 -stimulated intracellular cAMP levels in hPC. Therefore, either the selective antagonist of EP2 (PF-04418948, 1 µM) or EP4 (ONO-AE3-208, 1 µM) were co-incubated with 100 nM PGE 2 individually and in combination ( Figure 2). Upon PGE 2 stimulation, antagonism of either EP2 (−92.5%) or EP4 (−63.7%) alone resulted in a marked albeit only partial decrease of intracellular cAMP levels compared to stimulated hPC without antagonists. In contrast, the PGE 2 stimulated intracellular cAMP increase was completely abrogated by combined EP2 and EP4 antagonism (Figure 2), suggesting that, in hPC, both EP2 and EP4 may mediate PGE 2 -dependent signaling. Analysis of EP expression on hPC revealed the presence of PTGER1, PTGER2, and PTGER4 mRNA in differentiated hPC ( Figure S3), which encode for EP1, EP2, and EP4, respectively. As only EP2 and EP4 are reported to mediate an increase in intracellular cAMP (reviewed in [15][16][17]), we next investigated the effect of pharmacological inhibition of EP2 and EP4 signaling on PGE2-stimulated intracellular cAMP levels in hPC. Therefore, either the selective antagonist of EP2 (PF-04418948, 1 µM) or EP4 (ONO-AE3-208, 1 µM) were co-incubated with 100 nM PGE2 individually and in combination ( Figure 2). Upon PGE2 stimulation, antagonism of either EP2 (−92.5%) or EP4 (−63.7%) alone resulted in a marked albeit only partial decrease of intracellular cAMP levels compared to stimulated hPC without antagonists. In contrast, the PGE2 stimulated intracellular cAMP increase was completely abrogated by combined EP2 and EP4 antagonism ( Figure 2), suggesting that, in hPC, both EP2 and EP4 may mediate PGE2-dependent signaling.  . Intracellular cAMP levels in hPC are increased by PGE2 stimulation (100 nM, pink triangles) in a time-dependent manner. For several controls (blue circles), cAMP levels fell below the lowest concentration of the recommended standard curve (0.78 pmol/mL) and were therefore set to zero. Each data point represents the mean ± SD of one experiment with n = 3-4 samples per time-point.
Analysis of EP expression on hPC revealed the presence of PTGER1, PTGER2, and PTGER4 mRNA in differentiated hPC ( Figure S3), which encode for EP1, EP2, and EP4, respectively. As only EP2 and EP4 are reported to mediate an increase in intracellular cAMP (reviewed in [15][16][17]), we next investigated the effect of pharmacological inhibition of EP2 and EP4 signaling on PGE2-stimulated intracellular cAMP levels in hPC. Therefore, either the selective antagonist of EP2 (PF-04418948, 1 µM) or EP4 (ONO-AE3-208, 1 µM) were co-incubated with 100 nM PGE2 individually and in combination ( Figure 2). Upon PGE2 stimulation, antagonism of either EP2 (−92.5%) or EP4 (−63.7%) alone resulted in a marked albeit only partial decrease of intracellular cAMP levels compared to stimulated hPC without antagonists. In contrast, the PGE2 stimulated intracellular cAMP increase was completely abrogated by combined EP2 and EP4 antagonism ( Figure 2), suggesting that, in hPC, both EP2 and EP4 may mediate PGE2-dependent signaling. Figure 2. Intracellular cAMP is increased by PGE 2 stimulation via EP2 and EP4 in hPC. Representative cAMP levels following PGE 2 stimulation for 20 min without concomitant EP2 or EP4 antagonist (100 nM, pink triangles) compared to controls without PGE 2 (pink circles), after co-incubation with either EP2 antagonist (PF-04418948, 1 µM, orange triangles) or EP4 antagonist (ONO-AE3-208, 1 µM, blue triangles) compared to controls without PGE 2 (orange circles for EP2 antagonist, blue circles for EP4 antagonist), and co-incubation of PGE 2 with both antagonists simultaneously (1 µM each, black triangles) compared to controls without PGE 2 (black circles). Each data point represents a single sample and plotted as mean ± SD (horizontal lines) per treatment group consisting of n = 6 samples. For several controls, cAMP levels fell below the lowest concentration of the recommended standard curve (0.78 pmol/mL) and were therefore set to zero. Experiments were done in duplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment, except for the separate EP4 inhibition, which was only performed once. Statistics: *, p < 0.01; $, p < 0.05; n.s., not significant, assessed by a Mann-Whitney test. + denotes addition of the respective EP antagonists.

PGE 2 Induces COX2 Gene Expression via EP2 and EP4 Signaling in Differentiated hPC
Stimulation of hPC with PGE 2 for 2 h revealed a dose-dependent upregulation of COX2 mRNA expression ( Figure 3a). In order to elucidate the role of EP2 and EP4 in PGE 2 -mediated COX2 upregulation, either the selective EP2 antagonist PF-04418948 (1 µM) or the selective EP4 antagonist ONO-AE3-208 (1 µM) were co-incubated with 100 nM PGE 2 individually or in combination (Figure 3b). Upon PGE 2 stimulation, antagonism of either EP2 or EP4 alone resulted in an increase in COX2 mRNA although lower compared to PGE 2 -stimulated hPC without antagonists (Figure 3b). Combined EP2 and EP4 antagonism completely inhibited PGE 2 -mediated COX2 upregulation (Figure 3b), suggesting that PGE 2 signals via both EP2 and EP4 to regulate COX2 levels in a positive feedback loop.
simultaneously (1 µM each, black triangles) compared to controls without PGE2 (black circles). Each data point represents a single sample and plotted as mean ± SD (horizontal lines) per treatment group consisting of n = 6 samples. For several controls, cAMP levels fell below the lowest concentration of the recommended standard curve (0.78 pmol/mL) and were therefore set to zero. Experiments were done in duplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment, except for the separate EP4 inhibition, which was only performed once. Statistics: *, p < 0.01; $, p < 0.05; n.s., not significant, assessed by a Mann-Whitney test. + denotes addition of the respective EP antagonists.

PGE2 Induces COX2 Gene Expression via EP2 and EP4 Signaling in Differentiated hPC
Stimulation of hPC with PGE2 for 2 h revealed a dose-dependent upregulation of COX2 mRNA expression ( Figure 3a). In order to elucidate the role of EP2 and EP4 in PGE2-mediated COX2 upregulation, either the selective EP2 antagonist PF-04418948 (1 µM) or the selective EP4 antagonist ONO-AE3-208 (1 µM) were co-incubated with 100 nM PGE2 individually or in combination ( Figure  3b). Upon PGE2 stimulation, antagonism of either EP2 or EP4 alone resulted in an increase in COX2 mRNA although lower compared to PGE2-stimulated hPC without antagonists (Figure 3b). Combined EP2 and EP4 antagonism completely inhibited PGE2-mediated COX2 upregulation (Figure 3b), suggesting that PGE2 signals via both EP2 and EP4 to regulate COX2 levels in a positive feedback loop.
(a) (b) Figure 3. COX2 gene expression is increased by PGE2 via EP2 and EP4 in hPC. qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group. (a) COX2 levels following PGE2 stimulation (pink triangles) for 2 h were upregulated in a dose-dependent manner compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). SD was not plotted when only two independent experiments were performed. Statistics: *, p < 0.01, assessed by one-way ANOVA with Dunnett's follow-up test; (b) Figure 3. COX2 gene expression is increased by PGE 2 via EP2 and EP4 in hPC. qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group. (a) COX2 levels following PGE 2 stimulation (pink triangles) for 2 h were upregulated in a dose-dependent manner compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). SD was not plotted when only two independent experiments were performed. Statistics: *, p < 0.01, assessed by one-way ANOVA with Dunnett's follow-up test; (b) COX2 levels following PGE 2 stimulation for 2 h without concomitant EP2 or EP4 antagonist (100 nM, pink triangles) compared to controls without PGE 2 (pink circles), after co-incubation with either EP2 antagonist (PF-04418948, 1 µM, orange triangles) or EP4 antagonist (ONO-AE3-208, 1 µM, blue triangles) compared to controls without PGE 2 (orange circles for EP2 antagonist, blue circles for EP4 antagonist), and co-incubation of PGE 2 with both antagonists simultaneously (1 µM each, black triangles) compared to controls without PGE 2 (black circles) obtained in three independent experiments. Each data point represents the mean of an independent experiment (performed at least in triplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). + indicates addition of the respective EP antagonists. Statistics: *, p < 0.01; n.s., not significant, assessed by two-tailed Student's t-test.

PGE 2 Reduces PTGER2 and PTGER4 Gene Expression Which Is Not Modified by EP2 or EP4 Antagonists in Differentiated hPC
Stimulation of hPC with rising concentrations of PGE 2 for 2 h revealed inconsistent changes of PTGER2 mRNA expression: 10 nM and 1 µM did not significantly change PTGER2 expression, whereas PGE 2 100 nM slightly reduced PTGER2 expression (Figure 4a). The weak downregulation by PTGER2 of 100 nM PGE 2 was not abrogated by co-incubation with the selective EP4 antagonist ONO-AE3-208 (1 µM) (Figure 4b).
Stimulation of hPC with PGE2 for 2 h revealed a dose-dependent reduction of PTGER4 mRNA expression (Figure 4c). In order to elucidate the role of EP2 in PGE2-mediated PTGER4 downregulation, the selective EP2 antagonist PF-04418948 (1 µM) was co-incubated with 100 nM PGE2 (Figure 4d). The PGE2-induced decrease in PTGER4 mRNA was not abolished by co-incubation with the EP2 antagonist (Figure 4d).  . PTGER2 and PTGER4 gene expression in hPC after PGE2 stimulation and following co-incubation with EP antagonists. qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group. (a) PTGER2 levels following PGE2 stimulation with 10 nM, 100 nM, and 1 µM (pink triangles) for 2 h compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). SD was not plotted when only two independent experiments were performed. Statistics: $, p < 0.05, assessed by one-way ANOVA with Dunnett's follow-up test; (b) PTGER2 levels following PGE2 stimulation for 2 h without concomitant EP4 antagonist (100 nM, pink triangles) compared to controls without PGE2 (pink circles), and after co-incubation with EP4 antagonist (ONO-AE3-208, 1 µM, blue triangles) compared to controls without PGE2 (blue circles) obtained in three independent experiments. Each data point represents the mean of an independent experiment (performed in triplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). + denotes addition of ONO-AE3-208. Statistics: n.s., not significant, assessed by a Mann-Whitney test; (c) PTGER4 levels following PGE2 stimulation with 10 nM, 100 nM, and 1 µM (pink triangles) for 2 h compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines   . PTGER2 and PTGER4 gene expression in hPC after PGE 2 stimulation and following co-incubation with EP antagonists. qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group. (a) PTGER2 levels following PGE 2 stimulation with 10 nM, 100 nM, and 1 µM (pink triangles) for 2 h compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). SD was not plotted when only two independent experiments were performed. Statistics: $, p < 0.05, assessed by one-way ANOVA with Dunnett's follow-up test; (b) PTGER2 levels following PGE 2 stimulation for 2 h without concomitant EP4 antagonist (100 nM, pink triangles) compared to controls without PGE 2 (pink circles), and after co-incubation with EP4 antagonist (ONO-AE3-208, 1 µM, blue triangles) compared to controls without PGE 2 (blue circles) obtained in three independent experiments. Each data point represents the mean of an independent experiment (performed in triplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). + denotes addition of ONO-AE3-208. Statistics: n.s., not significant, assessed by a Mann-Whitney test; (c) PTGER4 levels following PGE 2 stimulation with 10 nM, 100 nM, and 1 µM (pink triangles) for 2 h compared to untreated control (blue circles). Each data point represents the mean of an independent experiment (performed at least in duplicate on different cell passages and on different dates, each consisting of n = 3-8 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). SD was not plotted when only two independent experiments were performed. Statistics: $, p < 0.05, assessed by a Kruskal-Wallis test with Dunn's multiple comparisons test; (d) PTGER4 levels following PGE 2 stimulation for 2 h without concomitant EP2 antagonist (100 nM, pink triangles) compared to controls without PGE 2 (pink circles), after co-incubation with EP2 antagonist (PF-04418948, 1 µM, orange triangles) compared to controls without PGE 2 (orange circles) obtained in three independent experiments. Each data point represents the mean of an independent experiment (performed in triplicate on different cell passages and on different dates, each consisting of n = 5-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). + denotes addition of PF-04418948. Statistics: $, p < 0.05; n.s., not significant, assessed by a two-tailed Student's t-test.
Stimulation of hPC with PGE 2 for 2 h revealed a dose-dependent reduction of PTGER4 mRNA expression (Figure 4c). In order to elucidate the role of EP2 in PGE 2 -mediated PTGER4 downregulation, the selective EP2 antagonist PF-04418948 (1 µM) was co-incubated with 100 nM PGE 2 (Figure 4d). The PGE 2 -induced decrease in PTGER4 mRNA was not abolished by co-incubation with the EP2 antagonist (Figure 4d).

Cellular PGE 2 and Metabolite Profile in hPC after PGE 2 Stimulation: Effects of EP2 and EP4 Blockade
To investigate whether PGE 2 stimulation and subsequent COX2 induction lead to changes in cellular levels of PGE 2 and its downstream metabolites 15-keto-PGE 2 and 13,14-dihydro-15-keto-PGE 2 (Figure 5a), hPC were analyzed by LC/ESI-MS/MS. After stimulation with PGE 2 , the cellular PGE 2 -content was elevated (Figure 5b), while 15-keto-PGE 2 and 13,14-dihydro-15-keto-PGE 2 remained at control levels. Pharmacological inhibition of EP2 and EP4 reduced cellular PGE 2 significantly (Figure 5c). Our findings point towards an autocrine PGE 2 -EP2/EP4-COX2 signaling axis in hPC. and in untreated controls (circles). PGE2 levels were increased in PGE2-stimulated cells (pink triangles) vs. controls (pink circles). Each datapoint represents the mean of an independent experiment (performed in triplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). Statistics: *, p < 0.01, assessed by a two-tailed Student's t-test in each experiment; (c) elevated cellular PGE2 levels caused by PGE2 stimulation (pink triangles) were abrogated by simultaneous co-incubation with combined EP2 and EP4 antagonism (PF-04418948 and ONO-AE3-208, respectively, 1 µM each). + indicates addition of combined EP antagonists. Each datapoint represents a single sample and plotted as mean ± SD (horizontal lines) per treatment group consisting of n = 6 replicates obtained in a single experiment. Statistics: *, p < 0.01, assessed by a two-tailed Student's t-test.

Glomerular PGE2 and Metabolite Profile in Glomeruli and Plasma in the CKD MWF Model
Analysis of PGE2 and its subsequent metabolites 15-keto-PGE2 and 13,14-dihydro-15-keto-PGE2 in glomeruli and plasma of MWF and SHR at eight weeks of age revealed an increase of glomerular PGE2 and 15-keto-PGE2 levels in MWF compared to SHR. No difference was observed for 13,14-dihydro-15-keto-PGE2 (Figure 6a). In plasma, levels of PGE2 and 13,14-dihydro-15-keto-PGE2 Figure 5. PGE 2 and its metabolites were measured by LC/ESI-MS/MS in hPC. (a) structure of PGE 2 and its metabolites; (b) levels of cellular PGE 2 (pink), 15-keto-PGE 2 (green), and 13,14-dihydro-15-keto-PGE 2 (blue) were measured after PGE 2 stimulation for 2 h (100 nM, triangles) and in untreated controls (circles). PGE 2 levels were increased in PGE 2 -stimulated cells (pink triangles) vs. controls (pink circles). Each datapoint represents the mean of an independent experiment (performed in triplicate on different cell passages and on different dates, each consisting of n = 3-6 replicates per treatment) and plotted as combined mean ± SD (horizontal lines). Statistics: *, p < 0.01, assessed by a two-tailed Student's t-test in each experiment; (c) elevated cellular PGE 2 levels caused by PGE 2 stimulation (pink triangles) were abrogated by simultaneous co-incubation with combined EP2 and EP4 antagonism (PF-04418948 and ONO-AE3-208, respectively, 1 µM each). + indicates addition of combined EP antagonists. Each datapoint represents a single sample and plotted as mean ± SD (horizontal lines) per treatment group consisting of n = 6 replicates obtained in a single experiment. Statistics: *, p < 0.01, assessed by a two-tailed Student's t-test.

Glomerular PGE 2 and Metabolite Profile in Glomeruli and Plasma in the CKD MWF Model
Analysis of PGE 2 and its subsequent metabolites 15-keto-PGE 2 and 13,14-dihydro-15-keto-PGE 2 in glomeruli and plasma of MWF and SHR at eight weeks of age revealed an increase of glomerular PGE 2 and 15-keto-PGE 2 levels in MWF compared to SHR. No difference was observed for 13,14-dihydro-15-keto-PGE 2 (Figure 6a). In plasma, levels of PGE 2 and 13,14-dihydro-15-keto-PGE 2 did not differ between MWF and SHR (Figure 6b). The metabolite 15-keto-PGE 2 in plasma was below level of detection (data not shown).
Cells 2020, 9,1256 12 of 21 did not differ between MWF and SHR (Figure 6b). The metabolite 15-keto-PGE2 in plasma was below level of detection (data not shown).

FFSS Increases COX2 and PTGER2 Gene Expression in hPC
FFSS was previously shown to elevate intracellular PGE2 levels and Cox2 in murine podocytes [10,23]. We therefore investigated COX2 mRNA expression after FFSS in hPC. FFSS led to increased COX2 mRNA expression in hPC as shown in Figure 7. . qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group (blue circles). COX2 upregulation was quantified after 2 h of FFSS with 2 dynes/cm 2 . Statistics: $, p < 0.05, Figure 6. Levels of PGE 2 (pink), 15-keto-PGE 2 (green) and 13,14-dihydro-15-keto-PGE 2 (blue) were measured in glomeruli of MWF (triangles) and SHR (circles) at 8 weeks of age. (a) glomerular PGE 2 and 15-keto-PGE 2 levels were increased in MWF (pink and green triangles, respectively) compared to SHR (pink and green circles, respectively), whereas glomerular 13,14-dihydro-15-keto-PGE 2 (blue circles and triangles) did not differ between both strains. Each data point represents a single animal and plotted as mean ± SD (horizontal lines) per rat strain consisting of n = 9-10 animals each. Statistics: *, p < 0.01; n.s., not significant assessed by a two-tailed Student's t-test; (b) levels of PGE 2 (pink) and 13,14-dihydro-15-keto-PGE 2 (blue) in plasma did not differ between MWF (triangles) and SHR (circles). Each data point represents a single animal and plotted as mean ± SD (horizontal lines) per rat strain consisting of n = 8-9 animals, each. Statistics: n.s., not significant assessed by the Mann-Whitney test.

FFSS Increases COX2 and PTGER2 Gene Expression in hPC
FFSS was previously shown to elevate intracellular PGE 2 levels and Cox2 in murine podocytes [10,23]. We therefore investigated COX2 mRNA expression after FFSS in hPC. FFSS led to increased COX2 mRNA expression in hPC as shown in Figure 7.
Cells 2020, 9,1256 12 of 21 did not differ between MWF and SHR (Figure 6b). The metabolite 15-keto-PGE2 in plasma was below level of detection (data not shown).

FFSS Increases COX2 and PTGER2 Gene Expression in hPC
FFSS was previously shown to elevate intracellular PGE2 levels and Cox2 in murine podocytes [10,23]. We therefore investigated COX2 mRNA expression after FFSS in hPC. FFSS led to increased COX2 mRNA expression in hPC as shown in Figure 7.  . qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group (blue circles). COX2 upregulation was quantified after 2 h of FFSS with 2 dynes/cm 2 . Statistics: $, p < 0.05, assessed by a two-tailed Student's t-test. Each datapoint represents the mean of an independent experiment (performed in duplicate on different cell passages and on different dates, each consisting of n = 5-6 replicates per treatment) and plotted as combined mean (horizontal lines).
(a) (b) Figure 8. PTGER2 and PTGER4 gene expression in hPC subjected to FFSS (pink squares). qPCR results are presented as relative mRNA expression level normalized to GAPDH and referred to control group (blue circles). mRNA expression of PTGER2 (a) and PTGER4 (b) was quantified after 2 h of FFSS with 2 dynes/cm 2 . Statistics: $, p < 0.05, assessed by a two-tailed Student's t-test. Each datapoint represents a single sample and plotted as mean ± SD (horizontal lines) per treatment group consisting of n = 4-5 replicates obtained in a single experiment.

Discussion
Recent studies in murine models of hyperfiltration support a pathophysiological role of autocrine/paracrine COX2/PGE2 activation on podocyte damage, thus contributing to disturbances of the glomerular filtration barrier including the development of albuminuria [10,14,23] (reviewed in [2,5]). These findings suggest that induction of COX2 associates with podocyte damage, while selective or non-selective inhibition of COX2 reduces proteinuria in animal models as well as in patients (reviewed in [48]). So far, the mechanisms involved have not been investigated in detail.
In mouse podocytes, EP1, -2 and -4 expression was reported and EP2 and -4 were also detected on protein level [14]. In this study, we corroborate these findings in hPC, which also express EP1, -2 and -4. EP4 is known as a constitutively expressed protein reflected by abundant protein levels in untreated murine podocytes compared to EP2 [10]. Our results on apparently lower PTGER4 mRNA expression compared to PTGER2 in hPC should be interpreted with caution as they need to be confirmed on the protein level in future investigations. Stimulation of hPC with PGE2 led to an immediate intracellular cAMP increase starting at 1 min after PGE2 stimulation until at least 40 min of stimulation (Figure 1). Of note, the detected intracellular cAMP levels are net levels resulting from cAMP generation by adenylate cyclase and its concomitant degradation by phosphodiesterases. Phosphodiesterase activity was only blocked at the end of the stimulation experiments by adding HCl. Previous studies in immortalized murine podocytes revealed a similar time-course of cAMP increase occurring within the first 30 min after EP2 and/or EP4 stimulation [39,49,50]. In our experimental setting in hPC, this PGE2-stimulated intracellular cAMP increase was only completely abrogated by combined EP2 and EP4 antagonism pointing towards a comparable role of both receptors for cAMP induction in hPC (Figure 2). Multiple intracellular signaling pathways have been described for either EP2 and EP4 (reviewed in [51]). PGE2 stimulation of EP2 and EP4 activated the transcription factors T-cell factor (Tcf) and lymphocyte enhancer factor (Lef) signaling via PKA-and phosphatidylinositol 3-kinase/proteinkinase B (PI3K/Akt)-dependent phosphorylation of glycogen synthase kinase 3 (GSK3), thus promoting translocation of the transcription cofactor β-catenin into

Discussion
Recent studies in murine models of hyperfiltration support a pathophysiological role of autocrine/paracrine COX2/PGE 2 activation on podocyte damage, thus contributing to disturbances of the glomerular filtration barrier including the development of albuminuria [10,14,23] (reviewed in [2,5]). These findings suggest that induction of COX2 associates with podocyte damage, while selective or non-selective inhibition of COX2 reduces proteinuria in animal models as well as in patients (reviewed in [48]). So far, the mechanisms involved have not been investigated in detail.
In mouse podocytes, EP1, -2 and -4 expression was reported and EP2 and -4 were also detected on protein level [14]. In this study, we corroborate these findings in hPC, which also express EP1, -2 and -4. EP4 is known as a constitutively expressed protein reflected by abundant protein levels in untreated murine podocytes compared to EP2 [10]. Our results on apparently lower PTGER4 mRNA expression compared to PTGER2 in hPC should be interpreted with caution as they need to be confirmed on the protein level in future investigations. Stimulation of hPC with PGE 2 led to an immediate intracellular cAMP increase starting at 1 min after PGE 2 stimulation until at least 40 min of stimulation (Figure 1). Of note, the detected intracellular cAMP levels are net levels resulting from cAMP generation by adenylate cyclase and its concomitant degradation by phosphodiesterases. Phosphodiesterase activity was only blocked at the end of the stimulation experiments by adding HCl. Previous studies in immortalized murine podocytes revealed a similar time-course of cAMP increase occurring within the first 30 min after EP2 and/or EP4 stimulation [39,49,50]. In our experimental setting in hPC, this PGE 2 -stimulated intracellular cAMP increase was only completely abrogated by combined EP2 and EP4 antagonism pointing towards a comparable role of both receptors for cAMP induction in hPC ( Figure 2). Multiple intracellular signaling pathways have been described for either EP2 and EP4 (reviewed in [51]). PGE 2 stimulation of EP2 and EP4 activated the transcription factors T-cell factor (Tcf) and lymphocyte enhancer factor (Lef) signaling via PKA-and phosphatidylinositol 3-kinase/proteinkinase B (PI3K/Akt)-dependent phosphorylation of glycogen synthase kinase 3 (GSK3), thus promoting translocation of the transcription cofactor β-catenin into the nucleus where interaction with Tcf and Lef modulated gene expression, e.g., of COX2 [52,53]. However, participation of EP2 in PI3K/Akt signaling remains a matter of debate, as some investigators suggest that only EP4 but not EP2 are linked with PI3K/Akt [54]. Mechanotransduction in murine podocytes was previously suggested to be mediated by Akt-GSK3β-β-catenin, extracellular-signal regulated kinases (ERK)1/2, and p38MAPK, but not cAMP-PKA signaling upon FFSS [40]. The lack of cAMP elevation upon FFSS in that study might be explained, though by the experimental design as intracellular cAMP was measured at the earliest 2 h after applying FFSS. In contrast, the cAMP-PKA pathway was shown to be involved upon PGE 2 stimulation of murine podocytes [40], which better matches our setting of PGE 2 stimulation of hPC. Moreover, the cAMP-PKA pathway has been shown in intracellular signaling upon FFSS in osteocytes [55,56].
We detected upregulation of COX2 by PGE 2 in a dose-dependent fashion with both EP2 and EP4 being involved in hPC ( Figure 3). PGE 2 stimulation of EP2 and EP4 was reported to increase cAMP response element-binding protein (CREB), which was demonstrated to be PKA-dependent for EP2, whereas EP4-coupled PI3K signaling was suggested to counteract CREB formation [57][58][59][60]. Of note, transcription of COX2 can be modulated by CREB, as CRE is part of the COX2 promotor [61,62]. Therefore, subsequent PKA/CREB activation could play a role for the observed increase in COX2 expression following intracellular cAMP level elevation in hPC. To further investigate this aspect, experiments with PKA-inhibitors, e.g., H-89, will be performed as well as analysis of CREB phosphorylation status, and activation of the transcription factors Tcf and Lef. Functional analysis of COX2 protein activity might also be helpful. However, previous data in various cell types including murine podocytes already revealed that COX2 protein is indeed increased after 2 h stimulation with PGE 2 [19,58]. Taken together, our results on concerted EP2 and EP4 signaling being involved in upregulation of intracellular cAMP and COX2 levels represent a novel finding. To validate our results on the role of EP2 and EP4 on the intracellular cAMP increase and upregulation of COX2, effects of hPC stimulation with an EP2 and/or EP4 agonist without PGE 2 should be investigated in future experiments.
Elevated PGE 2 levels were previously associated with podocyte damage, suggesting that it might be a biomarker of progressive CKD [10,69]. LC-MS/MS based methods are beneficial to precisely study cellular prostaglandin metabolism with a maximum of selectivity and specificity. However, cell culture experiments are mainly restricted to small sample amounts that might hamper analysis by LC-MS/MS [26][27][28][29][30][31][32]. Here, we present a refined protocol for prostaglandin analysis in cells by LC/ESI/MS-MS. This might help to further elucidate cellular prostaglandin metabolism under pathophysiological conditions particularly in vitro but also in vivo. The observed increases in cellular PGE 2 content upon PGE 2 stimulation in hPC might be due to several mechanisms. One possibility is that extracellular PGE 2 enters the podocyte by simple diffusion or by uptake transport mediated by OATP2A1, which was previously reported to facilitate bidirectional transport of PGE 2 over membranes [66][67][68]. A second reason could be autocrine/paracrine mechanisms, i.e., extracellular PGE 2 activates EP2 and EP4 signaling, thus increasing COX2 transcription and translation. As COX2 delineates the rate-limiting step of PGE 2 synthesis (reviewed in [70]), its induction leads to higher cellular PGE 2 levels. Our results support the latter option, as combined inhibition of EP2 and EP4 signaling diminished the increase in cellular PGE 2 levels.
In our corresponding in vivo study, we employed the MWF model, which represents a suitable model with glomerular hyperfiltration and thus FFSS [47]. The MWF model was previously extensively characterized (reviewed [47]). Thus, the MWF model is a non-diabetic inbred, genetic model with an inherited nephron deficit of 30-50% depending on the comparator rat strain [71,72]. Consequently, male MWF rats are characterized by increased single nephron glomerular filtration rate but with normal mean glomerular capillary pressure [72,73]. In addition, MWF rats develop mild arterial hypertension and spontaneous progressive albuminuria [47]. Thus, male MWF rats develop spontaneous albuminuria at an early age between weeks 4 and 8 after birth and subsequently progressive proteinuria and glomerulosclerosis [74]. The latter was also demonstrated in the MWF strain from our own colony and thus in the animals used in the current study [47]. Early onset albuminuria in young MWF animals occurs at six weeks of age and is preceded by glomerular hypertrophy, accompanied by focal and segmental loss of podoplanin and followed by podocyte foot process effacement at 8 weeks of age, i.e., at onset of albuminuria [75]. For these reasons, we selected animals at this age for our analysis in the current study. As a comparator strain, we use the previously characterized spontaneously hypertensive rats (SHR) that are resistant to albuminuria development [71,76]. Taken together, we showed the feasibility of the LC/ESI-MS/MS methodology to characterize the PGE 2 pathway at the glomerular tissue level by using the MWF strain. The observed increases in both PGE 2 and 15-keto-PGE 2 in isolated glomeruli of MWF support the activation of this pathway in glomerular hyperfiltration. However, these results should be viewed against the background that the cellular origin of this finding was not determined, and thus the contribution of other cell types, e.g., glomerular endothelial cells or mesangial cells remains unclear. Up to now, direct isolation of podocytes from glomeruli was reported for transgenic mice [77][78][79][80]. In the rat, there seem to be more technical difficulties as transgenic implementation of fluorescent dyes was not yet accomplished. Antibody staining for podocyte markers and subsequent analysis by FACS is possible [81], but whether isolated primary rat podocytes can be subjected to transcriptomic, proteomic, or lipidomic analysis remains to be investigated. Urinary PGE 2 was suggested as a biomarker for adaptive hyperfiltration in human solitary kidney [69]. The analysis of the urinary PGE 2 and metabolite profile in our CKD MWF model is currently not established due to experimental challenges to establish robust LS/ESI-MS/MS analysis in rat urine. In contrast, the profile in plasma did not differ significantly between MWF and SHR ( Figure 6b). In plasma, dilution of prostaglandins might be a major problem as 15-keto-PGE 2 levels were below the limit of detection. Therefore, plasma levels provide rather a rough estimate, while analysis of glomeruli offers a closer insight into podocyte prostaglandin metabolism.
Besides mimicking hyperfiltration by exogenous supplementation with PGE 2 , we also applied FFSS on hPC. This model aimed to imitate intensified flow of the ultrafiltrate in Bowman's space, thus causing podocyte injury. Similar to PGE 2 treatment, FFSS leads to upregulation of COX2 (Figure 7). Our data corroborate the work by Srivastava and coworkers, who suggested Akt/GSK3β-βcatenin and the MAPK pathway to be involved in mechanotransduction on mouse podocytes [10,40]. We thus aim to investigate these signaling pathways in hPC upon FFSS in our future work. We corroborate recent findings that EP2 is upregulated by FFSS while EP4 expression is not changed [10] (Figure 8).

Conclusions
An autocrine/paracrine pathway between COX2 and PGE 2 exists also in hPC and is mediated by both EP2 and EP4. Distinct analysis of cellular PGE 2 and its metabolites was enabled by a modified protocol using LS/ESI-MS/MS. Elevated PGE 2 and 15-keto-PGE 2 levels were detected in glomeruli of MWF, a model for CKD, thereby strengthening the hypothesis that glomerular PGE 2 accumulation is associated with albuminuria due to podocyte damage. Understanding prostaglandin signaling in hPC may contribute to identifying novel target pathways to protect against maladaptive responses to hyperfiltration in podocytes.