The Effects of Prostaglandin Analogues on Prostanoid EP1, EP2, and EP3 Receptor-Deficient Mice

Abstract

purpose. To determine the role of prostanoid EP receptors in the intraocular pressure (IOP)–lowering effect of prostaglandin analogues in EP receptor-deficient mice.

methods. Animals were bred and acclimatized in a 12-hour light-dark cycle. The diurnal IOP variation was measured by a microneedle method in EP1, EP2, and EP3 receptor-deficient mice (EP1KO, EP2KO and EP3KO) and in their wild-type (WT) background strain. IOP was measured in each mouse at night 3 hours after application of latanoprost, travoprost (0.004%), bimatoprost (0.03%), or unoprostone (0.12%). In WT and EP3KO mice, the effects of preapplication of diclofenac Na on drug-induced IOP reduction were examined.

results. Baseline IOPs were the same for all strains. Higher baseline IOPs were observed at night. Maximum IOP reduction occurred in WT mice 3 hours after latanoprost application during the day and night. Three hours after instillation at night, each of the four drugs lowered IOP significantly in WT, EP1KO, and EP2KO mice, whereas EP3KO a significantly lesser effect was induced by latanoprost, travoprost, and bimatoprost. Preapplication of diclofenac Na significantly attenuated drug-induced IOP reduction in WT but not in EP3KO mice.

conclusions. Deficiency of EP receptors had no effect on physiological IOP. EP1 and EP2 receptors are not involved in prostaglandin analogue-induced IOP reduction, whereas EP3 receptors may play a role.

Prostaglandin analogues (PG analogues) have been widely used as ocular hypotensive drugs for the treatment of glaucoma and ocular hypertension, because they have a greater effect on lowering intraocular pressure (IOP) and fewer systemic side effects than do β-blockers. ^{1 2 3} Currently, four different types of PG analogues (latanoprost, travoprost, bimatoprost, and isopropyl unoprostone [unoprostone]) are for the treatment of glaucoma and ocular hypertension. Latanoprost and travoprost are thought to lower IOP mainly via the FP receptor, ^{4 5} and there is evidence that bimatoprost and unoprostone also lower IOP through this receptor. ⁵  

It has been reported that the active forms of latanoprost and unoprostone facilitate release of PGE2 and that this effect can be inhibited by indomethacin. ^{6 7} Kashiwagi and Tsukahara. ⁸ reported that coadministration of a nonsteroidal anti-inflammatory drug (NSAID) inhibits latanoprost-induced IOP reduction in normal volunteers. Kaplan-Messas et al. ⁹ indicated that cholinergic and adrenergic drug-induced endogenous PGE2 production in the iris-ciliary body may have a role in the hypotensive effect of adrenergic and cholinergic drugs. ⁹ These reports suggest that not only the administered drug but also PG produces endogenously after drug application and non-FP prostanoid receptors may contribute to PG analogue-induced reduction of IOP. 

One of the prostanoid receptors possibly influencing IOP is the EP receptor, which consists of four subtypes: EP1, EP2, EP3, and EP4. PGE2 induces ocular inflammation in cats and rabbits. ^{10 11 12 13} Recently, it has been reported that topically applied 8-iso-PGE2, which acts on prostanoid EP and TP receptors, ^{14 15 16} lowers IOP in normal and glaucomatous monkey eyes in a different way than latanoprost. ^{17 18 19 20} Selective EP1 (17-phenyl trinor PGE2), ²¹ and EP2 (AH13205) ²² and relatively selective EP3 (sulprostone) ²³ receptor agonists lower IOP in cats, monkeys, and rabbits, respectively. Prostanoid EP receptors are reportedly concerned with relaxation of the trabecular meshwork and ciliary muscle ²⁴ and degradation of extracellular matrix around ciliary smooth muscle cells. ^{25 26 27} These previous studies strongly suggest the involvement of prostanoid EP receptors in the regulation of IOP. However, the contribution of prostanoid EP receptors to the IOP-lowering effect of clinically used PG analogues and to physiological IOP regulation has not been addressed. In this work, we studied the effect of prostanoid EP-receptor deficiency on the diurnal variation of IOP and on PG analogue-induced IOP reduction, in prostanoid EP1-, EP2-, and EP3-knockout mice, of which the wild-type (WT) has an ocular distribution of prostanoid EP receptors similar to humans. ²⁸  

Materials and Methods

Animals

All experiments were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mice (C57BL/6) were purchased from Japan SLC (Hamamatsu, Japan) at 5 weeks of age. The mouse genes encoding EP1, EP2, or EP3 receptors were disrupted by gene knockout methods using homologous recombination, as reported previously, ^{29 30} and EP1, EP2, or EP3 homozygous knockout (EP1KO, EP2KO, and EP3KO, respectively) mice were produced. C57BL/6 mice, which are the background species of the knockout mice, were used as the WT control. Mice were bred and housed in clear cages covered loosely with air filters. The cages contained white chip bedding. The environment was kept at 21°C with a 12-hour light (0600–1800 hours) and a 12-hour dark cycle. All mice were fed ad libitum and were acclimatized to the environment for at least 2 weeks before experiments. In all experiments, we used mice older than 8 weeks. 

Preparation and Instillation of Ophthalmic Solution

Latanoprost and bimatoprost were purchased from Cayman Chemical Co. (Ann Arbor, MI). Latanoprost (0.005%) was dissolved in its vehicle solution, as reported previously. ³¹ Bimatoprost (0.03%) was dissolved in phosphate-buffered saline (PBS). Travoprost (0.004%) and unoprostone (0.12%) ophthalmic solutions and vehicle solution for each were provided by Alcon, Inc. (Fort Worth, TX) and R-Tech Ueno, Ltd. (Hyogo, Japan), respectively. Diclofenac Na was purchased from Sigma-Aldrich Japan (Tokyo, Japan) and dissolved in PBS at a concentration of 0.1%. With a micropipette, 3 μL of each drug solution was applied topically to a randomly selected eye in a masked manner, whereas the other eye remained untreated to serve as the control. 

IOP Measurement

IOP was measured by a microneedle method in mice anesthetized with ketamine and xylazine, as described previously. ^{32 33} Briefly, a microneedle made of borosilicate glass (100-μm tip diameter and 1.0-mm outer diameter, World Precision Instruments [WPI], Sarasota, FL) was connected to a pressure transducer (model BLPR; WPI). The system pressure detected by the transducer was recorded by a data acquisition and analysis system (PowerLab; ADInstruments, Colorado Springs, CO). The microneedle was placed in the anterior chamber and the conducted pressure was recorded in both eyes during a 4- to 7-minute time window after anesthesia. Until the mouse was placed on the table for IOP measurement, room lighting was maintained similar to that in the vivarium. During the dark phase, all procedures were performed under red light illumination to eliminate the effect of lighting on IOP. The effect of each drug in each mouse was calculated as the ratio of IOP reduction (%), defined as 100 × (IOP of treated eye − IOP of contralateral untreated eye)/IOP of contralateral untreated eye. 

Diurnal Variation of IOP in the Mouse Strains and Effect of Latanoprost on IOP in WT Mice

Diurnal IOP variation in mice under general anesthesia was measured during the day (0900 hours) and at night (2100 hours) by the microneedle method. The time points of IOP measurement were determined according to previous reports where trough and peak IOP under the 12 hour light-dark cycle were observed at 0900 and 2100 hours in NIH Swiss ³⁴ and ddY mice. ³³  

To determine the best time for demonstrating the ocular hypotensive effect of PG analogues, we measured the time course of latanoprost’s effect on IOP in WT mice. Latanoprost (3 μL, 0.005%) was applied at 0600 or 1800 hours as described earlier, and the IOP-lowering effect was measured 1, 2, 3, and 6 hours after the instillation. The time when the ocular hypotensive effect was greatest was used for further studies to evaluate the effects of PG analogues. 

IOP-Lowering Effects of PG Analogues on Mice at Night

Three microliters of latanoprost (0.005%), travoprost (0.004%), bimatoprost (0.03%) or unoprostone (0.12%) was applied topically at 1800 hours. Three investigators instilled eye drops without knowing what was administered or what the other two investigators had administered. A fourth investigator, masked to the treatment, measured IOP 3 hours after drug instillation. Thus, all measurements were performed under masked conditions. 

Effect of Diclofenac Na on PG Analogue-Induced IOP Reduction in WT Mice

Diclofenac Na (3 μL, 0.1%) or PBS was applied topically 30 minutes before the application of PG analogues in WT and EP3KO mice. Each PG analogue was applied at 1800 hours and IOP measurement was made 3 hours after the PG analogue application. The effect of diclofenac Na on baseline IOP in WT mice was examined in the same way as a control. 

Statistical Analysis

The Wilcoxon signed-rank or rank-sum test was used as appropriate. The Kruskal-Wallis test was used for multiple comparison of baseline IOPs. The Steel and the Steel-Dwass tests were used for multiple comparison of IOP reduction. P < 0.05 was considered statistically significant. All data are presented as the mean ± SEM. 

Results

Diurnal Variation of IOP in WT, EP1KO, EP2KO, and EP3KO Mice

Mean IOPs of WT, EP1KO, EP2KO, and EP3KO mice during the day (at 0900 hours) and at night (2100 hours) are shown in Table 1 . The IOP observed at night in each genotype was higher than that during the day; however, there was no significant difference between IOPs during the day and night among the genotypes. 

Effect of Latanoprost on WT Mice during the Day and at Night

The IOP-lowering effect of latanoprost (0.005%) was examined 1, 2, 3, and 6 hours after instillation at 0600 (day) and 1800 (night; Fig. 1 ) hours. During the day, significant IOP reductions were observed at 2 (10.5% ± 5.0%; P = 0.0195) and 3 (12.9% ± 1.7%; P = 0.0020) hours after the instillation. At night, latanoprost significantly lowered IOP at 1 (10.1% ± 2.1%, P = 0.0039), 2 (19.1% ± 2.0%, P = 0.0020), 3 (20.0% ± 1.5%, P = 0.0020), and 6 (10.3% ± 2.1%, P = 0.0039) hours after instillation. There was no significant difference in mean IOPs of contralateral eyes among the four time points during the day (P = 0.2768) and at night (P = 0.1396). The maximum reduction in IOP was observed 3 hours after the instillation both during the day and at night, and a stronger IOP reduction was observed at night (P = 0.0041–0.0494, 1–6 hours after instillation). Therefore, comparison of the IOP-lowering effects of PG analogues in this study was performed 3 hours after the instillation at night. 

Effect of PG Analogues on IOP at Night

Three hours after the treatment with vehicle solution for each drug, the mean IOP reductions at night ranged from −3.0% to 2.1% in all strains. Latanoprost (0.005%) significantly lowered IOP in WT (22.1% ± 1.1%, P = 0.0039), EP1KO (18.6% ± 1.6%, P = 0.0010), EP2KO (20.3% ± 1.4%, P = 0.0020), and EP3KO (15.0% ± 1.9%, P = 0.0005) mice (Fig. 2) . IOP reduction induced by latanoprost in EP3KO mice was significantly less than that in WT mice (P = 0.0306), whereas the reductions in EP1KO and EP2KO mice were not significantly different from that in WT mice. Travoprost significantly lowered IOP in WT (26.1% ± 1.2%, P = 0.0020), EP1KO (25.8% ± 1.7%, P = 0.0020), EP2KO (25.2% ± 1.8%, P = 0.0020), and EP3KO (15.4% ± 1.5%, P = 0.0005) mice. The IOP reduction induced by travoprost in EP3KO mice was significantly less than that in WT mice (P = 0.0004), whereas the reductions in EP1KO and EP2KO mice were not significantly different from that in WT mice. Bimatoprost significantly lowered IOP in WT (20.3% ± 1.5%, P = 0.0010), EP1KO (18.1% ± 1.6%, P = 0.0020), EP2KO (17.9% ± 1.9, P = 0.0010), and EP3KO (12.0% ± 2.2, P = 0.0015) mice. The IOP reduction induced by bimatoprost in EP3KO mice was significantly less than that in WT mice (P = 0.0373), whereas the reductions in EP1KO and EP2KO mice were not significantly different from those in WT mice. Unoprostone significantly lowered IOP in WT (13.7% ± 1.9%, P = 0.0020), EP1KO (11.3% ± 1.5%, P = 0.0020), EP2KO (14.9% ± 2.0%, P = 0.0010), and EP3KO (10.8% ± 1.1%, P = 0.0020) mice. There were no significant differences in IOP reductions induced by unoprostone between WT mice and EP1KO (P = 0.5745), EP2KO (P = 0.9932), and EP3KO (P = 0.3116) mice. 

Effect of Diclofenac Na on PG Analogue-Induced IOP Reduction in WT Mice

To determine whether secondary PG production is involved with drug-induced IOP reduction in WT mice, we examined the effect of pretreatment by instillation at night of diclofenac Na (0.1%), an NSAID (Fig. 3) . Instillation of diclofenac Na did not affect IOP in WT mice (P = 0.7969). Latanoprost (21.2% ± 0.8%, P < 0.0001), travoprost (24.7% ± 1.4%, P < 0.0001), bimatoprost (20.1% ± 1.0%, P < 0.0001), and unoprostone (13.8% ± 1.2%, P = 0.0001) significantly lowered IOP in WT mice. In WT mice, the IOP reductions induced by latanoprost (16.4% ± 1.5%, P = 0.0350), travoprost (18.5% ± 1.4%, P = 0.0351), and bimatoprost (12.9% ± 1.1%, P = 0.0010) were attenuated by pretreatment with diclofenac Na compared with control animals pretreated with PBS. Unoprostone-induced IOP reductions were slightly attenuated by diclofenac Na pretreatment (10.8% ± 1.4%, P = 0.3213) compared with WT mice pretreated with PBS. In EP3KO mice, the IOP reductions induced by latanoprost, travoprost, and bimatoprost, each pretreated with PBS, were significantly less than those observed in WT mice pretreated with PBS. However, there was no significant difference in IOP reduction between WT and EP3KO mice treated with diclofenac Na plus latanoprost (P = 0.8610), travoprost (P = 0.0791), or bimatoprost- (P = 0.9792). In addition, there was no significant difference in IOP reductions between the diclofenac plus latanoprost-, travoprost-, or bimatoprost-treated groups in WT mice and the PBS plus each PG analogue in EP3KO mice (P = 0.9798, 0.4742, and 9977, respectively). 

Discussion

The expression of prostanoid EP receptors in the eyes of various animals and humans have been studied. ^{28 35 36 37 38 39} In mice, the expression of prostanoid EP1 (cornea, conjunctiva, trabecular meshwork, nonpigmented epithelium of ciliary body, and retina), EP2 (cornea, conjunctiva, trabecular meshwork, iris, ciliary body, and retina), EP3 (nonpigmented epithelium of ciliary body), and EP4 (cornea, conjunctiva, trabecular meshwork, nonpigmented epithelium of ciliary body, and retina) are similar to the distribution of EP receptors in humans. ²⁸ EP1–EP4 receptors are expressed in cornea, conjunctiva, trabecular meshwork, iris, ciliary body, and retina in human eyes. ^{37 40 41 42} The similarity of receptor distributions indicates that the mouse is a suitable animal model for investigating the involvement of prostanoid EP receptors in drug effects, although in this study, EP4KO mice could not be used, because the mutation is lethal for the embryo. ⁴³  

We measured baseline IOP at trough (0900 hours) and peak (2100 hours) times in our previous studies. ^{33 34} The results in this study revealed that baseline IOPs in WT, EP1KO, EP2KO, and EP3KO mice are the same and that baseline IOP in each genotype is higher at night than during the day (Table 1) , which is consistent with our previous work. ^{33 34} We conclude that the physiology of aqueous humor dynamics in EP1KO, EP2KO, and EP3KO mice is similar to that in WT mice under normal conditions. 

In WT, EP1KO, and EP2KO mice, all PG analogues used in the study significantly lowered IOP to a similar extent in each genotype (Fig. 2) , whereas, in EP3KO mice, latanoprost-, travoprost-, and bimatoprost-induced IOP reductions were significantly less than those in WT mice. Although the acid forms of latanoprost and bimatoprost have relatively high affinities for the EP1 receptor (EC₅₀ = 119 and 2.7 nM, respectively), ⁴⁴ the IOP reduction induced by latanoprost and bimatoprost in EP1KO mice was the same as in WT mice. Our previous study revealed that latanoprost- and bimatoprost-induced IOP reductions were abolished in FPKO mice after a single application, ⁵ suggesting that latanoprost and bimatoprost do not lower IOP through action at the EP1 receptor at short times after a single application. In EP2KO mice, all PG analogues lowered IOP to a similar extent compared with WT mice. Affinities of the acid forms of latanoprost, travoprost, bimatoprost, and unoprostone for the EP2 receptors are very low (EC₅₀ > 10 μM), ⁴⁴ which is consistent with results in this study that a deficiency of EP2 receptors did not affect the IOP-lowering efficacies of the PG analogues. 

Latanoprost acid and travoprost acid have low affinity for the EP3 receptor, ⁴⁴ and bimatoprost has some affinity for the EP3 receptor. ⁴⁴ However, IOP reduction induced by these PG analogues was abolished in FPKO mice. ⁵ These results and our finding of a reduced IOP-lowering effect of these three drugs in EP3KO mice do not support the hypothesis of direct stimulation of EP3 receptors by these PG analogues. A possible explanation is that there was an alteration of aqueous humor dynamics arising from histologic change in the anterior segment of EP3KO mice. Although it is possible that abnormalities in ultrastructure and composition of extracellular matrix of the anterior segment existed and that those abnormalities affected aqueous humor dynamics, EP1KO, EP2KO, and EP3KO mice had no significant tissue abnormalities in their paraffin-embedded and hematoxylin-eosin-stained sections of anterior segment observable with light microscopy (data not shown). These observations do not support the explanation. 

Another explanation for our finding is that there was a partial contribution of endogenous PGs, which were secondarily produced by the direct stimulation of the FP-receptor, to EP3 receptor stimulation. To clarify whether endogenously produced PGs play a role in the IOP-lowering effect of PG analogues currently used in mice, we examined the effect of pretreatment with diclofenac Na on PG analogue-induced IOP reduction in WT and EP3KO mice. Diclofenac Na significantly attenuated drug-induced IOP reduction only in WT mice (Fig. 2) . Although unoprostone showed a trend in IOP change similar to that of other PG analogues currently used, a substantially weaker potential to reduce IOP may result in the lack of power to detect a small difference in IOP reductions among strains. It is reported that latanoprost and unoprostone induce endogenous PGE2, which was inhibited by NSAIDs, and that the IOP-lowering effect of latanoprost was attenuated by coadministration of NSAIDs in normal volunteers and in rabbits. ^{6 7 45} Taken together, these results suggest that PG analogues in current use induce endogenous PGs that lower IOP through the EP3 receptor in mice. 

The EP3 and FP receptors have similar cellular distributions in the trabecular meshwork and ciliary body. ³⁷ In our earlier ⁵ and recent studies, ⁴⁶ latanoprost, bimatoprost, travoprost, and unoprostone showed no ocular hypotensive effect in FP-knockout mice. These data suggest that the stimulation of FP receptor may be primarily required for endogenously produced PGE2 and its EP3 receptor-mediated IOP-lowering effect. Costagliola et al. ⁴⁷ reported recently that diclofenac Na significantly enhances the hypotensive effect of latanoprost in patients with primary open-angle glaucoma. They hypothesized that continuous inhibition of COX by repeated NSAID administration induces depression of endogenous PG synthesis followed by upregulation of prostanoid receptors in glaucomatous patients, and they suggested that pathophysiological upregulation of prostanoid receptors is responsible for a more marked IOP reduction. In contrast, Kashiwagi and Tsukahara ⁸ reported that coadministration of an NSAID inhibited latanoprost-induced IOP reduction in normal volunteers. The effects of multiple applications of PG analogues, with or without NSAIDs, on mouse IOP have not been studied yet and should be the subject of future studies. Needless to say, it is difficult to compare the current results with a single use of NSAID in mouse eyes with the results in the human study. PGs secondarily produced after application of PG analogues should also be identified chemically, and further studies using agonists and antagonists specific for EP1, EP2, EP3, and EP4 receptor each would provide more information on EP receptor-mediated mechanisms of IOP reduction. 

In conclusion, the prostanoid EP1, EP2, and EP3 receptors do not affect diurnal IOP variation and structure of the anterior segment of the mouse eye. EP1 and EP2 receptors do not contribute significantly to PG analogue-induced IOP reduction. Finally, endogenous PGs produced by FP-receptor stimulation are partly involved in the IOP reduction after a single application of clinically used FP agonists in mice, and this effect is mediated via EP3 receptors. 

 Supported in part by Grant-in-Aid for Science Research A-14207049 from the Japanese Ministry of Education, Culture, Sports, Science, and Technology (MAr).

 Submitted for publication January 30, 2006; revised March 10, 2006; accepted May 25, 2006.

 Disclosure: T. Ota, None; M. Aihara, Alcon, Inc. (F), R-Tech Ueno, Ltd. (F); T. Saeki, None; S. Narumiya, None; M. Araie, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Baseline IOP in WT, EP1KO, EP2KO, and EP3KO Mice

	IOP (mm Hg)
		WT	EP1 KO	EP2 KO	EP3 KO
Daytime	14.3 ± 0.3 (22)	14.4 ± 0.2 (21)	14.7 ± 0.2 (23)	14.5 ± 0.2 (25)
Nighttime	19.3 ± 0.3 (24)	19.3 ± 0.4 (23)	19.6 ± 0.2 (22)	19.5 ± 0.4 (21)

 Data are expressed as the mean ± SEM (number of animals). There was no significant difference in IOP among genotypes during the day or night (Kruskal-Wallis test).

Figure 1.

 

Effect of latanoprost on IOP in WT mice during the day and at night. Data are expressed as the mean ± SEM (n = 10/time point). *P < 0.05 for treated vs. contralateral eyes (Wilcoxon signed-ranks test).

Figure 2.

 

Effect of latanoprost, travoprost, bimatoprost, and unoprostone on WT, EP1KO, EP2KO, and EP3KO mice at night. Each drug was applied at 1800 hours, and IOP was measured at 3 hours after application. Data are expressed as the mean ± SEM (n = 9–12) *P < 0.05 versus WT (Steel test).

Figure 3.

 

IOP-lowering effect of latanoprost, travoprost, bimatoprost, and unoprostone and the effect of diclofenac Na on PG-induced IOP reduction in WT and EP3KO mice at night. Diclofenac Na (0.1%) or PBS was instilled 30 minutes before the instillation of PG analogues. PG analogues were instilled at 1800 hours, and IOP was measured at 3 hours after application. Diclofenac Na did not affect baseline IOP in WT mice. Data are expressed as the mean ± SEM (n = 9–17). *P < 0.05 vs. WT treated with the PG analogues plus PBS (Steel-Dwass test).