Flow after Prostaglandin E₁ Is Mediated by Receptor-Coupled Adenylyl Cyclase in Human Anterior Segments

Abstract

purpose. To assess the effect of prostaglandin (PG) F_2α and PGE₁ on flow through the trabecular meshwork in organ preserved human anterior segments.

methods. Isolated human anterior segments were perfused under standard conditions at a constant pressure of 10 mm Hg, while flow was continuously monitored. After a stabilization period, 6 consecutive concentrations of PGs were administered. cAMP levels were determined in the perfusate at baseline conditions and at 10⁻⁶ M PG.

results. Perfusion with concentrations ranging from 10⁻¹⁰ to 10⁻⁵ M PGE₁ resulted in a dose-dependent increase in flow (P < 0.0001), reaching a plateau of a 26% increase at 10⁻⁷ M. Perfusion with PGF_2α or placebo (Eagle’s minimum essential medium) did not influence baseline flow. cAMP produced by human anterior segments increased from 4.8 ± 0.6 pmol · 30 min⁻¹ per anterior segment at baseline to 19.2 ± 4.8 pmol · 30 min⁻¹ per anterior segment after perfusion with 10⁻⁶ M PGE₁ (P < 0.005). Perfusion with 10⁻⁶ M PGF_2α did not influence baseline cAMP production. Perfusion with 10⁻⁵ M GDP–β–S, an inhibitor of G protein, before and in combination with 10⁻⁶ M PGE₁ completely inhibited the increase in flow and cAMP production as observed after PGE₁ alone. Perfusion with 10⁻⁵ M GDP–β–S alone did not affect baseline cAMP production.

conclusions. In organ preserved perfused human anterior segments, flow and cAMP production in the perfusate are not mediated by receptor-coupled adenylyl cyclase activity at baseline conditions. Perfusion with PGE₁ is suggested to increase flow through the trabecular meshwork by stimulation of prostanoid EP₂ receptor subtype, EP₄ receptor subtype, or both, coupled to G_(s) protein, inducing activation of the adenylyl cyclase catalytic unit. The results may indicate a physiological role for EP₂ receptor subtype, EP₄ receptor subtype, or both in the modulation of flow through the trabecular meshwork after stimulation.

There is ample evidence that prostaglandin (PG) F_2α analogues lower intraocular pressure (IOP) in glaucoma. ^{1 2 3 4} The main mechanism of action by which these analogues are thought to lower IOP is not by increasing trabecular flow but rather by stimulating aqueous humor flow through the uveoscleral channels as was shown in primates, including humans. ^{5 6 7 8}  

In primates PGF_2α, as well as PGE₁ and PGE₂, administered intracamerally, did not induce an increase of outflow facility, both in normal eyes and in eyes after disinsertion or retrodisplacement of the ciliary muscle. ⁹ It was then concluded that in general PGs do not lower IOP via a direct effect on the trabecular meshwork. 

Cultured human trabecular cells have been demonstrated to generate PGE₂ and to a lower extent PGF_2α and 6-keto PGF_1α, the stable end product of prostacyclin (PGI₂). ^{10 11} Incubations of rabbit sclera trabecular rings with PGE₁ revealed an increase of cAMP. ¹² In human trabecular meshwork membrane fractions, PGE₁ and PGE₂, but not PGF_2α, stimulated adenylyl cyclase activity. ¹³ Adenosine 3′–5′–monophosphate is known to increase outflow facility in rabbit and monkey eyes and 8-bromo–cAMP lowers IOP in rabbits. ^{14 15 16}  

Because in human trabecular meshwork membrane fractions, PGE₁ is linked to the adenylyl cyclase system and PGF_2α is not, and because cAMP increases outflow facility in rabbit and monkey eyes, it would be of interest to investigate the effect of these PGs on trabecular flow in the human eye. 

The present study reports the effect of a wide-range dose of PGE₁ and PGF_2α on the flow through the trabecular meshwork of organ preserved human anterior segments. 

Materials and Methods

Human eyes were obtained from the Cornea Bank at the Netherlands Ophthalmic Research Institute. The eyes were discarded because of corneal pathology. Under sterile conditions the eyes were dissected as reported previously. ¹⁷ The corneoscleral explants were mounted on a perfusion apparatus and connected to a real-time flow measurement system. ^{18 19} The pressure was computer-link adjusted and maintained at the constant level of 10 mm Hg. Flow through the trabecular meshwork was also monitored by the computer. Every 20 seconds the actual pressure and flow rate were sampled. Under standard conditions (5% CO₂ and at 35°C) the eyes were perfused with modified Eagle’s minimum essential medium (EMEM) containing 2% fetal bovine serum (final protein concentration 0.4%), 100 μg · ml⁻¹ penicillin, and 50 μg · ml⁻¹ streptomycin. 

After a stabilization period of 2 hours, 6 consecutive concentrations of PGE₁ (11,13–dihydroxy–9–oxo–(11α,13E,15S)–prost–13–en–1–oic acid), PGF_2α–TS (both from Sigma Chemical, St Louis, MO) or placebo (EMEM) were administered at 60-minute intervals. The concentrations of the prostaglandins were 10⁻¹⁰, 10⁻⁹, 10⁻⁸, 10⁻⁷, 10⁻⁶, and 10⁻⁵ M, respectively. In addition, the effects of 10⁻⁶ M PGE₁ after and in combination with 10⁻⁵ M GDP–β–S (Fluka Chemie AG, Buchs, Switzerland) were assessed. At baseline and after perfusion with 10⁻⁶ M PG, perfusate was collected and immediately frozen at −70°C for later cAMP detection. An enzyme Immunoassay Kit (Cayman Chemical, Ann Arbor, MI) was used to determine cAMP concentrations in the samples after acetylation and purification. The amount of cAMP was calculated as picomoles produced per anterior segment for 30 minutes after a drug administration. Baseline flow rate was defined as the mean flow during the last 15 minutes before drug administration. Mean flows for each interval were calculated from the equilibrated period of that interval. Difference in flow was expressed in percent by Flow Exp/Flow baseline × 100% − 100%. 

In addition, total lactate dehydrogenase activity (LDH) was measured hourly in 10 μl perfusate of 5 human anterior segments, perfused for 2 hours with plain medium and 1 hour with 10⁻⁶ M PGE₁. ²⁰ LDH activity was expressed in U · 30 min⁻¹ per anterior segment (U =μ mol · min⁻¹). 

Data were expressed as mean ± SD for age and postmortem times and with standard error of the mean for flow, cAMP, and LDH data. Statistical analyses of data were performed using a repeated-measures ANOVA, a Newman–Keuls test, and the two-sided paired Student’s t-test. 

Results

The mean age and the mean postmortem times of the donors that provided the eyes for organ preserved perfusion are listed in Table 1 . A typical flow rate tracing of one of the treated eyes is shown in Figure 1 . 

Baseline flow in the PGE₁ group (n = 7) was 4.1 ± 0.5 μl · min⁻¹ and increased in a dose-dependent fashion after perfusion with all concentrations of PGE₁ used (P < 0.0001, repeated-measures ANOVA, followed by Newman–Keuls test). Table 2 shows an overview of the data. The plateau of trabecular flow of 26% reached at 10⁻⁷ M PGE₁ (EC₅₀: 10⁻⁹ M) can be distinguished. In contrast, none of the concentrations of PGF_2α used affected flow through the trabecular meshwork. Baseline flow in the PGF_2α group (n = 7) was 3.6 ± 0.5 μl · min⁻¹ and showed a tendency only to decrease at the highest concentrations of PGF_2α (10⁻⁶ and 10⁻⁵ M). Baseline flow in the placebo group (n = 7) perfused with EMEM was 4.0 ± 1.0 μl · min⁻¹ and did not change significantly after consecutive administration of placebo at time intervals similar to those with the PGs. 

GDP–β–S, an inhibitor of G proteins, was perfused for 60 minutes in a concentration of 10⁻⁵ M before administration of the combination of 10⁻⁶ M PGE₁ and 10⁻⁵ M GDP–β–S in 8 human anterior segments. ²¹ Table 2 shows that the baseline flow was 2.8 ± 0.3 μl · min⁻¹ and 3.4±0.4 μl · min⁻¹ after 10⁻⁵ M GDP–β–S. Perfusion with 10⁻⁵ M GDP–β–S and 10⁻⁶ M PGE₁ together did result in a flow of 3.5 ± 0.5 μl · min⁻¹, which is not significantly different from the flow after GDP–β–S alone. 

The cAMP levels after perfusion with 10⁻⁶ M PGE₁ and 10⁻⁶ M PGF_2α are also listed in Table 2 . In the PGE₁ group baseline cAMP was 4.8 ± 0.6 pmol· 30 min⁻¹ per anterior segment and increased after 10⁻⁶ M PGE₁ to 19.2 ± 4.8 pmol · 30 min⁻¹ per anterior segment (P < 0.005); the average stimulation index was 3.9. In the PGF_2α group baseline cAMP in the perfusate was 4.2 ± 1.4 pmol · 30 min⁻¹ per anterior segment and did not differ after perfusion with 10⁻⁶ M PGF_2α (3.7 ± 0.7 pmol · 30 min⁻¹ per anterior segment). 

Baseline cAMP before perfusion with 10⁻⁵ M GDP–β–S was 3.4 ± 0.7 pmol · 30 min⁻¹ per anterior segment and was 3.8 ± 1.1 pmol · 30 min⁻¹ per anterior segment during perfusion with GDP–β–S. Perfusion with 10⁻⁵ M GDP–β–S together with 10⁻⁶ M PGE₁ resulted in a cAMP production of 3.8 ± 1.0 pmol · 30 min⁻¹ per anterior segment. 

There was no significant difference between the cAMP values after perfusion with GDP–β–S alone and the combination of GDP–β–S and PGE₁. 

Additional perfusion of 5 anterior segments with EMEM resulted in a LDH activity of 0.047 ± 0.015 U · 30 min⁻¹ during the first hour and 0.036 ± 0.007 U · 30 min⁻¹ during the second hour. Subsequent perfusion with 10⁻⁶ M PGE₁ resulted in a LDH activity of 0.026 ± 0.009 U · 30 min⁻¹. 

Discussion

In this study PGE₁, but not PGF_2α, increased flow through the trabecular meshwork. The amount of cAMP produced by human anterior segments increased approximately fourfold from baseline values after perfusion with 10⁻⁶ M PGE₁. 

In a previous study we observed that in bovine trabecular meshwork particulate fractions, adenylyl cyclase was stimulated by PGE₁ and PGE₂ but not by PGF_2α. ¹³ Furthermore, the stimulation of adenylyl cyclase by PGE₁ was dose dependent. It was shown that GDP–β–S, known to inhibit G proteins irreversibly, downregulated PGE₁-stimulated adenylyl cyclase activity. Also, in membrane preparations of human trabecular meshwork, PGE₁ and PGE₂ stimulated adenylyl cyclase activity. It was then concluded that bovine and human trabecular meshwork contained prostanoid EP receptor sites of the subtype bound to adenylyl cyclase. 

An increase in flow after perfusion with PGE₁, as observed in this study, may not originate from an increase in cAMP production, since both events may exist independently. Therefore, the relation of the increase in flow with a possible receptor-mediated activity of PGE₁ to stimulate adenylyl cyclase via G proteins was tested by perfusion with GDP–β–S in the presence and absence of PGE₁. Because GDP–β–S hardly has an influence on basal adenylyl cyclase activity, it is suggested that the baseline cAMP level is not originating from receptor-coupled adenylyl cyclase activity. 

The combination of GDP–β–S and PGE₁ completely inhibited the increase in flow through the meshwork and the increase of cAMP in perfusate as observed after PGE₁ alone. 

From this it is concluded that after perfusion of human anterior segments with PGE₁ the increase of flow through the trabecular meshwork is mediated by receptor-bound adenylyl cyclase activity. This indicates that EP receptors either of the subtype EP₂, subtype EP₄, or both not only are present in human trabecular meshwork but also may play a physiological role by promoting trabecular flow after stimulation. ^{13 22 23}  

cAMP is known to be a poorly permeable intracellular messenger. However, many studies show cAMP levels extracellularly in vitro of intact cells or whole tissue preparations and in vivo, for instance in the aqueous humor. ^{12 13 24 25 26 27 28 29 30 31 32} In those studies, agents upregulating intracellular cAMP show an increase in cAMP in supernatant or in the aqueous humor. Because cAMP is not easily permeable, studies can give no quantitative information on the upregulation of the catalytic unit intracellularly, and the amount is only a faint expression of the events occurring intracellularly. Therefore, data on extracellular cAMP should be considered only qualitatively as an up- or downregulation of intracellular cAMP. It can be argued that 10⁻⁶ M PGE₁ is toxic and induces lysis of cells lining the cornea and outflow pathways, resulting in a continuous release of intracellular cAMP in the perfusate. This could be the explanation of the increase of cAMP in perfusate after PGE₁, but not the absence of the increase after PGF_2α, unless the latter PG is not toxic. To investigate this, we determined LDH activity in perfusate in the presence and absence of PGE₁. LDH is an intracellular enzyme in the cytosol. The activity found in extracellular fluid is a measure for cell lysis. The observed LDH activity in perfusate should be considered relatively. Our data clearly show that there is no increase of LDH activity after perfusion with 10⁻⁶ M PGE₁. From this it is concluded that the increased levels of cAMP after PGE₁ cannot be the result of cell lysis. Moreover, if cAMP in perfusate was caused by cell lysis, one would expect also an increase of cAMP in perfusate after perfusion with GDP–β–S and PGE₁ together, which was not the case. 

Kaufman ⁹ reported hardly any increase in outflow facility in cynomolgus monkeys after intracameral infusion with PGE₁, PGE₂, and PGF_2α in final concentrations ranging from 3 × 10⁻⁷ to 3 × 10⁻⁴ M in vivo. However, a 30% to 50% increase of total facility was observed after topical PGE₂ in albino rabbits and in rhesus monkeys in vivo. ^{33 34}  

In our opinion the most fundamental differences between the study of Kaufman and the present study are monkey versus human eyes and perfusion of whole eyes in vivo (iris–ciliary body present) versus anterior segment perfusion. It was stated earlier that human and subhuman eyes differ fundamentally in the composition, cell biology, and physiology of outflow pathways. ³⁵ In particular, human eyes do not show evidence of a time-dependent increase of outflow facility (washout). ³⁶ In the study of Kaufman, ⁹ the perfusion data were individually decreased by 15% to correct for washout. It is questionable whether the correction for washout in that study obscured an effect of PGE₁ on outflow facility. Furthermore, in vivo PGs are rapidly removed from the aqueous by an active transport into the ciliary processes. ^{37 38 39} It is suggested that the uptake of PGs by the ciliary processes in cynomolgus monkeys in vivo may have leveled off the possible effect on outflow facility in that study, whereas in anterior segment perfusion the concentration of PGs in contact with the trabecular outflow tissues is approximately constant. Finally, the differences in results between the study of Kaufman ⁹ and the present study may be the result of species differences pertaining to the presence of functional EP-receptor sites in the trabecular meshwork. 

The present study indicates that PGE₁ increases flow through the trabecular meshwork in organ preserved human anterior segments by stimulation of EP₂, EP₄, and/or, receptor sites coupled to G_(s)-protein–stimulated adenylyl cyclase activity. It is suggested to design a prostaglandin prodrug stimulating prostanoid FP receptor subtype (FP) as well as EP₂ or EP₄ receptor subtypes or both. When a potential influence of this prodrug on conjunctival vessels and blood-aqueous barriers can be circumvented, a PG analogue will be available for glaucoma treatment with a dual mechanism on outflow pathways, i.e., on trabecular as well as on uveoscleral flow. 

 Submitted for publication May 7, 1998; revised April 23, 1999; accepted May 20, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Mean Age and Mean Postmortem Time of Donor Eyes

Group	Age, y	Postmortem Time
Placebo (n = 7)	77 ± 10	18 ± 7
PGE1 (n = 7)	71 ± 10	20 ± 7
PGF2α (n = 7)	74 ± 11	16 ± 4
GDP–β–S± PGE1 (n = 8)	68 ± 13	31 ± 10

 Values are mean ± SD.

Figure 1.

 

Typical flow rate tracing of a human anterior segment after addition of 6 consecutive concentrations of PGE₁.

Table 2.

 

Mean Flow ± SEM in Organ Preserved Human Anterior Segments at Baseline, after Increasing Doses of PGE₁ and PGF_2α, during EMEM (Placebo), and after 10⁻⁵ M GDP–β–S Alone, Followed by 10⁻⁵ M GDP–β–S in Combination with 10⁻⁶ M PGE₁

	Baseline			10−10 M	10−9 M	10−8 M	10−7 M	10−6 M			10−5 M
		Flow	cAMP	Flow	Flow	Flow	Flow	Flow	cAMP	Flow	cAMP
Placebo	3.96 ± 1.0		3.87 ± 1.0	3.80 ± 1.0	3.83 ± 1.0	3.83 ± 0.9	3.77 ± 0.9		3.76 ± 0.9
PGE1	4.05 ± 0.5	4.8 ± 0.6	4.42 ± 0.7	4.65 ± 0.7	4.90 ± 0.7	5.08 ± 0.7	5.02 ± 0.7	19.2 ± 4.8*	5.10 ± 0.6
PGF2α	3.59 ± 0.5	4.2 ± 1.4	3.65 ± 0.6	3.66 ± 0.6	3.65 ± 0.6	3.62 ± 0.6	3.50 ± 0.6	3.7 ± 0.7	3.41 ± 0.6
GDP–β–S	2.82 ± 0.3	3.4 ± 0.7							3.35 ± 0.4	3.8 ± 1.1
+ PGE1							3.46 ± 0.5	3.8 ± 1.0

 Values are μl · min⁻¹. Simultaneous determination of cAMP ± SEM (pmol · 30 min⁻¹ per anterior segment) in the perfusate.

 Effect with increasing concentrations: PGE₁ P < 0.0001, PGF_2α P = 0.15 (NS), placebo P = 0.909 (NS) repeated-measures ANOVA:

*  indicates P < 0.005, 2-sided paired t-test.

The authors thank Elisabeth Pels and coworkers of the Cornea Bank for providing the eyes and the culture medium.