Characterization of the Signaling Modalities of Prostaglandin E2 Receptors EP2 and EP4 Reveals Crosstalk and a Role for Microtubules

Abstract

Prostaglandin E2 (PGE2) is a lipid mediator that modulates the function of myeloid immune cells such as macrophages and dendritic cells (DCs) through the activation of the G protein-coupled receptors EP2 and EP4. While both EP2 and EP4 signaling leads to an elevation of intracellular cyclic adenosine monophosphate (cAMP) levels through the stimulating Gα_s protein, EP4 also couples to the inhibitory Gα_i protein to decrease the production of cAMP. The receptor-specific contributions to downstream immune modulatory functions are still poorly defined. Here, we employed quantitative imaging methods to characterize the early EP2 and EP4 signaling events in myeloid cells and their contribution to the dissolution of adhesion structures called podosomes, which is a first and essential step in DC maturation. We first show that podosome loss in DCs is primarily mediated by EP4. Next, we demonstrate that EP2 and EP4 signaling leads to distinct cAMP production profiles, with EP4 inducing a transient cAMP response and EP2 inducing a sustained cAMP response only at high PGE2 levels. We further find that simultaneous EP2 and EP4 stimulation attenuates cAMP production, suggesting a reciprocal control of EP2 and EP4 signaling. Finally, we demonstrate that efficient signaling of both EP2 and EP4 relies on an intact microtubule network. Together, these results enhance our understanding of early EP2 and EP4 signaling in myeloid cells. Considering that modulation of PGE2 signaling is regarded as an important therapeutic possibility in anti-tumor immunotherapy, our findings may facilitate the development of efficient and specific immune modulators of PGE2 receptors.

Keywords: G protein-coupled receptor; membrane receptor; myeloid cells; podosome; prostaglandin E2 signaling.

Figure 1

Prostaglandin E2 (PGE2)-induced podosome dissolution in human immature dendritic cells (iDCs) is mostly mediated by EP4. (A) Representative images of PBMC-derived iDCs that were left untreated or were treated with 1 μM EP4 agonist L-902688, 1 μM EP2 agonist (R)-Butaprost, both 1 μM L-902688 and 1 μM (R)-Butaprost, 1 μM PGE2 alone or 1 μM PGE2 after pretreatment with EP2 antagonist (ant.) AH6809, EP4 antagonist GW627368X or both AH6809 and GW627368X. Cells were stained for actin (green) and vinculin (magenta). Scale bar = 10 μm. (B) iDCs were treated with different concentrations of EP2 agonist (ago.) (R)-Butaprost, EP4 agonist L-902688 or both (R)-Butaprost and L-902688. Cells were stained for actin and vinculin and the number of podosomes per image was quantified and normalized to untreated control. Cells treated with 10 μM PGE2 were included as positive control. The error bars represent mean ± SD. Data presented are from two different donors. ns, not significant, *P<0.05; ^###P<0.001 versus untreated control, Welch ANOVA with Dunnett’s T3 multiple comparison test. (C) iDCs were treated with different concentrations of PGE2 with or without pretreatment with EP2 antagonist (ant.) AH6809, EP4 antagonist GW627368X or both AH6809 and GW627368X. Cells were stained for actin and vinculin and the number of podosomes per image was quantified and normalized to untreated control. The error bars represent mean ± SD. Data presented are from three different donors. ns, not significant, *P<0.05, **P<0.01, ***P<0.001; ^##P<0.01, ^###P<0.001 versus untreated control, Welch ANOVA with Dunnett’s T3 multiple comparison test.

Figure 2

EP2 and EP4 induce distinct cyclic adenosine monophosphate (cAMP) responses. (A) Schematic illustration of intramolecular cAMP FRET sensor t-Epac-vv. Binding of cAMP to t-Epac-vv reduces Förster Resonance Energy Transfer (FRET) between the mTurquoise donor and Venus acceptor fluorophores of t-Epac-vv, making a decreased ratio of the fluorescent intensities a direct measure of cAMP accumulation [adapted from (34)]. (B) The mTurquoise (cyan) and Venus (yellow) signal were acquired with widefield microscopy (WF panels). After background subtraction in each image and channel, the FRET ratio was calculated as the Venus intensity over the mTurquoise intensity for each timepoint and was normalized to the average of prestimulus values (FRET panels). Normalized FRET values range from 0 (red) to 1 (blue). Scale bar = 5 μm. (C) FRET ratios of t-Epac-vv before and after the addition of different prostaglandin E2 (PGE2) concentrations were measured in transiently transfected RAW macrophages. A control was performed with the addition of buffer only. The data presented are mean ± SD from ≥5 cells per condition. (D) Example FRET curve that illustrates the definition of the relative cAMP peak and cAMP production rate. The amplitude of the cAMP peak was defined as the maximal decrease in FRET ratio. The cAMP production rate was quantified by determining the slope between the final prestimulus timepoint and the timepoint at which minimal FRET ratios were observed using a linear fit over all included timepoints. (E) The cAMP production peak and the cAMP production rate were measured from the FRET curve of individuals cells from (C) and the average peak was plotted as a function of the average production rate per condition. The error bars represent SD for both parameters. (F, H, J) FRET ratios were measured after the addition of PGE2 in cells pretreated with EP4 antagonist (ant.) GW627368X (F), pretreated with EP2 antagonist AH6809 (H) or pretreated with both GW627368X and AH6809 (J). The data presented are mean ± SD from ≥4 cells per condition. (G, I, K) The relative cAMP production peak and the cAMP production rate were measured from (F, H, J), respectively. The error bars represent SD for both parameters.

Figure 3

EP4-coupled Ga_i dampens the prostaglandin E2 (PGE2)-induced cyclic adenosine monophosphate (cAMP) production. (A) RAW macrophages were transfected with Gγ-CFP only or with Gγ-CFP (donor), Gα_i-Citrine (acceptor) and Gβ wildtype together. The average lifetime of Gγ-CFP for individual cells were calculated from frequency-domain FLIM images and the distributions were fitted with a Gaussian profile (solid lines) to obtain the average lifetimes. Photobleaching of Gα_i-Citrine was used as a control for the occurrence of FRET. (B) The average donor lifetime in cells expressing both donor and acceptor is plotted before and after addition of 10 μM PGE2 in absence or presence of EP4 antagonist AH23848 or after addition of 10 μM EP2 agonist Butaprost. The data presented are mean ± SD from ≥2 cells. (C) FRET ratios of t-Epac-vv before and after addition of PGE2 were measured in transiently transfected RAW macrophages that were left untreated or were pretreated with Gα_i inhibitor pertussis toxin (PTx). Controls were performed with the addition of buffer only. The data presented are mean ± SD of measurements from ≥4 cells per condition. (D) The cAMP peak and the cAMP production rate were quantified as described in Figure 2D from (C) and the average peak was plotted as a function of the average production rate per condition. The error bars represent SD for both parameters. (E) iDCs were treated with different concentrations of PGE2 with or without PTx pretreatment. Cells were stained for actin and vinculin and the number of podosomes per image was quantified and normalized to untreated control. The error bars represent mean ± SD. Data presented are from three different donors. **P<0.01, ***P<0.001, Welch ANOVA with Dunnett’s T3 multiple comparison test.

Figure 4

Efficient signaling of EP2 and EP4 relies on microtubule integrity. (A, C, E) The Förster Resonance Energy Transfer (FRET) ratio of t-Epac-vv was measured in cells that were untreated or pretreated with nocodazole (Noc.) before and after addition of prostaglandin E2 (PGE2). Shown are the ratios obtained in cells in the absence of antagonists (A), in the presence of EP2 antagonist AH6809 (C) or EP4 antagonist GW627368X (E). Controls were performed with the addition of buffer only. The data are mean ± SD from ≥5 cells per condition. (B, D, F) The cyclic adenosine monophosphate (cAMP) production peak and the cAMP production rate were measured from the FRET curve of individual cells from (A, C, E), respectively, and the average peak was plotted as a function of the average production rate per condition. The error bars represent SD for both parameters.

Figure 5

Schematic overview of the cyclic adenosine monophosphate (cAMP) responses induced by EP2 and EP4. (A) When only EP4 is active, both Gα_s and Gα_i control AC activity. Gα_s induces a dose-dependent cAMP response that is dampened by Gα_i. The cAMP signal subsides over time and is attenuated by microtubule disruption. (B) When EP2 is activated selectively, only Gα_s modulates AC activity. The resulting cAMP response is either weak or strong, does not subside and completely relies on an intact microtubule network. (C) When both EP2 and EP4 are active, competition for Gα_s dampens the integrated cAMP response. Signaling crosstalk between EP2 and EP4 allows the cell to respond differently to prostaglandin E2 (PGE2) depending on the organization and expression of EP2 and EP4.