Characterization of the signalling modalities of prostaglandin E2 receptors EP2 and EP4 reveals crosstalk and a role for microtubules

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 signalling leads to an elevation of intracellular cyclic adenosine monophosphate (cAMP) levels through the stimulating Gαs protein, EP4 also couples to the inhibitory Gαi 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 signalling 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 signalling 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 signalling in myeloid cells. Considering that modulation of PGE2 signalling is regarded as an important therapeutic possibility in anti-tumour immunotherapy, our findings may facilitate the development of efficient and specific immune modulators of PGE2 receptors.

EP2 and EP4. While both EP2 and EP4 signalling leads to an elevation of intracellular cyclic adenosine 23 monophosphate (cAMP) levels through the stimulating Gαs protein, EP4 also couples to the inhibitory 24 Gαi protein to decrease the production of cAMP. The receptor-specific contributions to downstream 25 immune modulatory functions are still poorly defined. Here, we employed quantitative imaging 26 methods to characterize the early EP2 and EP4 signalling events in myeloid cells and their contribution 27 to the dissolution of adhesion structures called podosomes, which is a first and essential step in DC 28 maturation. We first show that podosome loss in DCs is primarily mediated by EP4. Next, we 29 demonstrate that EP2 and EP4 signalling leads to distinct cAMP production profiles, with EP4 inducing 30 a transient cAMP response and EP2 inducing a sustained cAMP response only at high PGE2 levels. 31 We further find that simultaneous EP2 and EP4 stimulation attenuates cAMP production, suggesting a 32 reciprocal control of EP2 and EP4 signaling. Finally, we demonstrate that efficient signaling of both 33 EP2 and EP4 relies on an intact microtubule network. Together, these results enhance our 34 understanding of early EP2 and EP4 signalling in myeloid cells. Considering that modulation of PGE2 35 signalling is regarded as an important therapeutic possibility in anti-tumour immunotherapy, our 36 findings may facilitate the development of efficient and specific immune modulators of PGE2 37 receptors. 38 Introduction with with 100 ng/ml pertussis toxin for 16 hrs or left untreated before adding 3 μM AH6809 for 1 h, 133 or 10 μM GW627368X for 1 h, with and without 5 μM nocodazole for 20 min,. Six mTurquoise and 134 Venus emission images were acquired followed by automated addition of PGE2 and subsequent 135 acquisition of another 20 mTurquoise and Venus emission images (tlag=10 s). The mean fluorescence 136 intensity of the Venus and mTurqoise signal in a cell was corrected by subtraction of the background 137 signal in each image and channel before dividing the Venus over mTurqoise mean fluorescence 138 intensity to obtain the FRET ratio. Butaprost. 155

158
To assess different contributions of EP2 and EP4 in mediating PGE2 signalling in DCs, we determined 159 the effect of receptor-specific inhibition or stimulation in podosome dissolution. We first treated 160 immature DCs (iDCs) with PGE2 or selective EP2 and EP4 agonists and quantified the number of 161 podosomes per cell (Figure 1A,B). In line with our previous observations, addition of PGE2 resulted 162 in an almost complete loss of podosomes in iDCs. Interestingly, both EP2-and EP4-specific 163 stimulation reduced the number of podosomes, with EP4 agonist stimulation being slightly more 164 efficient ( Figure 1B). These results indicate that individual EP2 and EP4 downstream signalling can 165 lead to podosome dissolution. 166 To better investigate the respective contribution of EP2 and EP4 signalling after the addition of 167 their natural ligand PGE2, we pretreated the cells with selective EP2 and EP4 antagonists before PGE2 168 addition and subsequently quantified podosome dissolution. Figure 1C shows that inhibition of EP4 169 attenuates podosome dissolution upon stimulation with 0.01-0. to t-Epac-vv reduces FRET between the mTurquoise donor and Venus acceptor fluorophores, a 188 decreased FRET ratio in the macrophages is a direct measure of cAMP production (Figure 2A,B).

189
After the addition of PGE2, cAMP was produced immediately and reached a maximum concentration 190 after about 40 seconds, subsiding to lower levels after 200 seconds ( Figure 2C). To compare the cAMP 191 kinetics across difference treatment conditions, we quantified the peak of cAMP production and the 192 production rate, as shown in Figure 2D. Both parameters scaled with increasing PGE2 concentrations, 193 indicating that the rate and the magnitude of the induced cAMP response is dose-dependent ( Figure  194 2E).

195
Compared to PGE2 only, EP2 inhibition led to higher cAMP levels at all tested PGE2 196 concentrations, while cAMP concentrations subsided to a similar extent ( Figure 2F). The PGE2-197 induced cAMP production rate and cAMP peak remained dose-dependent upon EP2 inhibition as both 198 parameters scaled with PGE2 concentration (Figure 2G). These results indicate that EP2 blockade 199 increases the signaling efficiency of EP4 in response to PGE2. Inhibition of EP4 led to dramatically 200 different cAMP production. In contrast to EP2 inhibition, robust cAMP production was not observed 201 until 1 M PGE2 when EP4 signaling was blocked (Figure 2H,I). Furthermore, this strong cAMP 202 response did not attenuate as observed in the absence of EP4 inhibition. Compared to PGE2 only, the 203   inhibition of EP2 and EP4 ( Figure 2J). Pretreatment with both antagonists effectively inhibited total 278 cAMP production at 0.1 and 1 μM PGE2, showing that both receptors are completely blocked at 279 physiological concentrations of PGE2 ( Figure 2J,K). 280 Our results demonstrate that the selective stimulation of EP2 and EP4 by PGE2 induces 281 kinetically distinct cAMP production profiles. While PGE2-EP4 signalling results in a fast and 282 transient cAMP production that linearly increases with increasing ligand concentrations, PGE2-EP2 283 signalling is induced only by PGE2 concentrations above 1 M and cAMP production and is more 284 prolonged. We also show that co-stimulation of EP2 and EP4 mutually dampens their signaling 285 efficiency, as both receptors induce higher cAMP production when they are individually triggered by 286 PGE2. 287 288 EP4-coupled Gαi finetunes the PGE2-induced cAMP production 289 Given that EP2 and EP4 differentially control cAMP dynamics, we sought to identify factors that 290 contribute to these differences. Since the inhibitory G protein Gαi has been shown to couple to EP4 291 [22], we hypothesized that Gαi dampens the PGE2-induced cAMP response in cells expressing EP4. 292 To demonstrate that EP4 selectively activates Gαi also in macrophages, we performed fluorescence 293 lifetime imaging (FLIM) to measure FRET between cyan fluorescent protein (CFP)-tagged Gγ (Gγ-294 CFP) and Citrine-tagged Gαi (Gαi-Citrine). The fluorescent lifetime of the FRET donor decreased upon 295 co-expression with the acceptor and was restored to control levels upon acceptor photobleaching 296 (Figure 3A), indicating that FRET occurred between Gγ-CFP and Gαi-Citrine. Since Gαi is known to 297 undergo conformational rearrangements upon activation [33] and FRET between Gγ-CFP and Gαi-298 Citrine is likely affected by such rearrangements, a shift in fluorescence lifetime is expected upon EP4 299 stimulation. Treatment with PGE2 induced a gradual reduction in the lifetime of the donor fluorophore, 300 whereas no shift in the lifetime phase was observed upon either inhibition of EP4 or selective 301 stimulation of EP2 ( Figure 3B). These findings confirm that PGE2 induces Gαi activation via EP4 302 only. 303 To determine the consequences of EP4-mediated Gαi activation on PGE2 signaling, we 304 measured cAMP elevation using t-Epac-vv upon inhibition of Gαi with pertussis toxin (PTx). Gαi 305 blockade significantly enhanced the cAMP peak concentrations and production induced by 0.1 μM 306 PGE2 and by 1 μM PGE2, albeit at a lower extent (Figure 3C), indicating that Gαi attenuates cAMP 307 production most strongly at lower PGE2 concentrations. The effect of Gαi inhibition on cAMP 308 production is more clearly depicted in Figure 3D, where a higher cAMP peak and an increased 309 production rate are observed after addition of PTx. 310 Next, to investigate whether EP4-mediated Gαi activation would enhance cAMP-dependent 311 processes such as podosome dissolution, we determined PGE2-mediated podosome loss in iDCs with 312 or without PTx treatment. We found that Gαi inhibition led to slightly increased podosome loss at all 313 PGE2 concentrations tested, with 1 M PGE2 being statistically significant while 0.01 and 0.1 M 314 PGE2 show a non-significant but clear trend ( Figure 3E). It should be considered that such low 315 concentrations of PGE2 are less powerful in inducing podosome dissolution, which means that PTx 316 effect is more difficult to assess. This result indicates that the Gαi-mediated dampening of cAMP 317 production also affects cellular decisions downstream of EP2 and EP4. 318 Together, these findings indicate that Gαi dampens the onset of cAMP production, suggesting 319 that the PGE2-EP4-Gαi axis might act as signalling gatekeeper when low PGE2 levels slightly 320 fluctuate. 321

EP2-and EP4-mediated signalling requires cortical microtubule integrity 339
Since the interplay between G proteins and tubulin is well documented as well as their localization 340 along microtubules [34; 35; 36], we investigated whether microtubule integrity is important for PGE2-341 induced cAMP production. We found that microtubule disruption deregulates PGE2-induced cAMP 342 elevation ( Figure 4A). More specifically, when both receptors are activated, attenuation of the cAMP 343 response by nocodazole was only observed at 1 μM PGE2 and not at 0.1 μM PGE2 (Figure 4A,B). 344 Upon EP2 inhibition, however, the cAMP production rate and the maximum cAMP levels induced by 345 PGE2-EP4 were reduced at all PGE2 concentrations tested (Figure 4C,D). Finally, EP4 inhibition 346 revealed that the PGE2-EP2 strong and sustained cAMP response is completely prevented by 347 microtubule disruption (Figure 4E,F). These results demonstrate that the Gαs-mediated cAMP 348 response to PGE2 relies on an intact microtubule network and that disruption of this network reduces 349 the signaling efficiency of both EP2 and EP4, with EP2 activity being significantly more sensitive to 350 microtubule integrity than EP4 activity. 351 352  are activated (Figure 5A). Active Gαs proteins modulate the activity of AC, resulting in a strong cAMP 474 response. Gαi functions to fine-tune the cAMP production at low PGE2 concentrations. As EP4 is 475 subjected to desensitization and internalization [23; 25], the elicited cAMP response subsides over 476 time. When EP2 is selectively stimulated, only Gαs controls AC activity ( Figure 5B). The resulting 477 cAMP response does not subside because EP2 is insensitive to receptor desensitization and 478