- Email: [email protected]
Ocular Penetration and Bioconversion of Prostaglandin F2α Prodrugs in Rabbit Cornea and Conjunctiva
Ocular Penetration and Bioconversion of Prostaglandin F2r Prodrugs in Rabbit Cornea and Conjunctiva DU-SHIENG CHIEN*,‡, DIANE D.-S. TANG-LIU*X,
AND
DAVID. F. WOODWARD§
Received September 1, 1995, from the *Department of Pharmacokinetics and §Department of Biological Sciences, Allergan, Inc., 2525 Dupont Drive, Irvine, CA 92715, and ‡Pharmacokinetics and Drug Metabolism, Bayer Corporation, 400 Morgan Lane, West Haven, CT 06516. Final revised manuscript received March 3, 1997. Accepted for publication June 17, 1997X. Abstract 0 The objective of this study was to identify prostaglandin F2R (PGF2R) prodrugs that have an optimal ocular absorption profile and therefore could be potentially useful for the treatment of glaucoma. Rabbit cornea, conjunctiva, and iris/ciliary body were mounted in a flow-through chamber to evaluate the permeability and bioconversion of PGF2R and its prodrugs. The prodrugs tested were PGF2R 1-isopropyl, 1,11-lactone, 15-acetyl, 15-pivaloyl, 15-valeryl, and 11,15-dipivaloyl esters. After 4 h in the donor or acceptor compartments, the products and formation of PGF2R were analyzed by HPLC. Effects on intraocular pressure and ocular surface hyperemia were also determined. All prodrugs penetrated the rabbit cornea faster than PGF2R by 4- to 83-fold. All prodrugs penetrated conjunctiva faster than PGF2R, except the 15-acetyl ester prodrug, which was equally permeable. No direct correlation between drug lipophilicity and permeability across the cornea or conjunctiva was apparent. The most metabolically stable prodrug was the 1,11-lactone, followed by the 11,15-dipivaloyl, 15-pivaloyl, 15-acetyl, 1-isopropyl, and the 15-valeryl esters, the latter of which was extensively converted to PGF2R. A separation index for various prodrugs was calculated from the ratio of the bioavailable PGF2R for ocular hypotension to the bioavailable PGF2R for hyperemia. The highest separation index was observed for the 1,11-lactone prodrug (2.33), followed by the 11,15dipivaloyl ester prodrug (1.80). Thus the 1,11-lactone and 11,15-dipivaloyl ester prodrugs appeared to be superior to the others in providing bioavailable PGF2R for ocular hypotension, while minimizing hyperemia. The favorable separation index for these compounds appeared to be due to their metabolic stability at the corneal surface and conjunctiva combined with sufficient bioavailability for ocular hypotension.
Introduction Prostaglandins mediate a spectrum of biological responses in the eye. In primary, open-angle glaucoma, topical prostaglandins have been used to reduce intraocular pressure.1-3 However, because the cornea is not readily permeable to PGF2R, a relatively high concentration of PGF2R is necessary for effective intraocular pressure reduction, causing conjunctival hyperemia, ocular discomfort, headaches, and other side effects.2 A more lipophilic PGF2R prodrug, PGF2R 1-isopropyl ester, was synthesized to increase ocular bioavailability and thus reduce the effective dose in an attempt to minimize the side effects. Although studies in humans and several animal species showed that PGF2R 1-isopropyl ester was more potent than PGF2R in lowering intraocular pressure, it also caused ocular discomfort and hyperemia.4-6 To identify a PGF2R prodrug with increased ocular hypotensive efficacy and reduced ocular surface hyperemia, a series of PGF2R-monoesters, -diesters, and -lactone prodrugs were tested in rabbits. Penetration across the conjunctiva and cornea were evaluated as well as the sites of metabolism to nascent PGF2R. Bioavailability values for the hypotensive and X
Abstract published in Advance ACS Abstracts, August 1, 1997.
1180 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
hyperemic effects of PGF2R were calculated to predict the most promising prodrugs for the treatment of glaucoma in humans.
Experimental Procedures Section MaterialssProstaglandin F2R and the PGF2R prodrugs tested in this study are summarized as follows:
Compound
Perfusion Concentration (µg/mL)
PGF2R
251
PGF2R-15-acetyl ester
202
PGF2R-1-isopropyl ester
314
PGF2R-1,11-lactone
194
PGF2R-15-pivaloyl ester
106
PGF2R-15-valeryl ester
314
PGF2R-11,15-dipivaloyl ester
95
Source Lot # PF2041, Toray Industrial, Inc, New York, NY Chemical Sciences, Allergan, Inc, Irvine, CA Chemical Sciences, Allergan, Inc, Irvine, CA Lot #12070-GLB-96, Upjohn Co. (U-49653), Kalamazoo, MI Chemical Sciences, Allergan, Inc. Irvine, CA Chemical Sciences, Allergan, Inc. Irvine, CA Chemical Sciences, Allergan, Inc. Irvine, CA
The chemical structures, molecular weights, and apparent partition coefficients of these compounds are listed in Table 1. All other solvents and chemicals were of analytical grade and were used as received. MethodssAdult female New Zealand albino rabbits, weighing 2.0-2.5 kg, were used in this study. All procedures were performed in accordance with the ARVO resolution on the use of animals in research. The flow-through diffusion chamber7 was used to measure the permeability of the test compounds across rabbit cornea and conjunctiva. All ocular tissues were freshly dissected before use. The rabbit cornea or conjunctiva was secured between two chambers without any wrinkles or folding. PGF2R and all prodrug solutions were freshly prepared in glutathione bicarbonated Ringer's solution (GBR) containing 1% propylene glycol (v/v). After the ocular membrane was mounted, 2.5 mL of the drug-free GBR solution, preheated to 35 °C, was placed into the chamber facing the endothelial side (acceptor chamber). The drug solution was then slowly infused through the chamber facing the epithelium (donor chamber) at a flow rate of 25.8 µL/min. The volumes of the acceptor and donor chambers were 2.5 and 0.125 mL, respectively. The penetration areas of the cornea and conjunctiva were 1.09 and 0.95 cm2, respectively. In several corneal penetration experiments, the iris/ciliary body was dissected and placed in the acceptor chamber to mimic the intraocular condition in vivo. The experiments were conducted at 35 °C for 4 h in a waterjacketed chamber. Fifty microliter samples of acceptor fluid and anterior effluent were collected every 30 min throughout the perfusion period. The samples were mixed immediately with acetonitrile/0.02 M KH2PO4 (pH 3.2) to terminate any enzymatic reaction before analysis of the test compounds and their metabolic products by HPLC.
S0022-3549(95)00373-X CCC: $14.00
© 1997, American Chemical Society and American Pharmaceutical Association
Table 1sChemical Structure of PGF2r and Its Prodrugs MWa
log Pb
PGF2R 15-Acetyl ester 1-Isopropyl ester 1,11-Lactone
355 396 396 337
1.26 2.16 2.50 2.61
H CH3CO H
15-Pivaloyl ester 15-Valeryl ester 11,15-Dipivaloyl ester
439 439 523
3.50 3.75 5.0
C(CH3)3CO C4H9CO C(CH3)3CO
compound
RI(C-15)
R2(C-1)
H H CH2(CH3)2
H H H
R3(C-11)
H H H
H H C(CH3)3CO
a Molecular weight. b Log P was determined from the Pomona Med. Chem. Software system or measured by HPLC.
Scheme 1sCorneal penetration of PGF2R prodrugs. The chemical stability of each compound was monitored for 4 h in GBR solution at 35 °C. A reversed-phase HPLC analysis was performed with a Beckman ODS, 5 µm particle column with dimensions 4.6 mm by 15 cm. A WISP 710B injector (Waters Associates) and a Beckman 114M pump were used to apply the samples and deliver the mobile phase, respectively. The flow rate of the mobile phase (acetonitrile/0.02 M KH2PO4, pH 3.2, adjusted by phosphoric acid) was 1.0 to 1.5 mL/ min. Drug concentration was monitored by a Kratos 783 UV spectrophotometer at a wavelength of 200 nm. The cumulative amounts (µg) of PGF2R or prodrug in the acceptor chamber were plotted versus time over a 4-h period. The slope of the linear portion of the curve was calculated by linear regression analysis to determine the steady-state flux rate (µg/min). The apparent ocular permeability (Papp) of the test compound was calculated by dividing the steady-sate flux rate by the surface area of the membrane and the initial concentration of the test compound. The bioavailability of PGF2R for the hypotensive pharmacological effect is a function of the corneal permeability and conversion of the prodrug. The amount of PGF2R was measured at the end of each experiment to determine the enzymatic stability and site of bioconversion of the test prodrug in the ocular tissues. Both the iris/ciliary body and the cornea are possible sites for prodrug conversion to PGF2R. Thus, the bioavailable PGF2R in the ocular anterior fluid was determined from the formation of PGF2R in the acceptor fluid when prodrug was perfused with the cornea and the iris/ciliary body. The ocular absorption and disposition of the test prodrug are depicted in Scheme 1. The hyperemic effect of PGF2R is the result of PGF2R being enzymatically formed from its prodrug at the ocular surface. Because the PGF2R formed from the corneal surface would be readily absorbed into the conjunctiva, the bioavailability of PGF2R for the hyperemic effect was calculated from the total amount of PGF2R formed by the
Scheme 2sConjunctival penetration of PGF2R prodrugs. conjunctiva in the acceptor fluid and the corneal surface in the donor fluid. The absorption and disposition of PGF2R prodrug in the conjunctiva is presented in Scheme 2. The ratio of the bioavailable PGF2R for ocular hypotension to that for hyperemia is used as the separation index for the test prodrugs. Presumably, the larger the separation index of a prodrug, the better the separation between ocular hypotension and hyperemia that could be obtained after topical administration. Intraocular pressure (IOP) was measured with a pneumatonometer (Digilab) calibrated against the eyes of anesthetized rabbits by closed stopcock manometry. The correlation coefficient over the range 10-30 mmHg was 0.98. The animals were acclimated to pneumatonometry by taking unrecorded measurements before experimental determination of IOP. Corneal anesthesia for tonometry was provided by topical application of one drop of 0.05% proparacaine. All IOP studies were performed in a masked fashion. Animals were randomly assigned to each treatment group, and test and control eyes were also selected randomly. Solutions were instilled into the lower conjunctival sac in a 25 µL volume. Drug solution was administered to one eye and the contralateral eye received vehicle as a control. IOP was recorded immediately before topical dosing and then typically at 1, 2, 4, 6, 8, and 10 h after dosing. Changes in IOP were expressed as the change from baseline IOP (∆ IOP) in the treated eye minus the change from baseline IOP (∆ IOP) in the control eye and statistically analyzed by Student’s t test. Ocular surface hyperemia was scored according to the following graded scale: 1 ) slight with some vessels definitely injected above normal; 2 ) moderate ) eye is a more uniform, diffuse red; 3 ) severe ) a diffuse “beefy” red eye.
Results The effects of PGF2R and its prodrugs on IOP are depicted in Figure 1. All compounds were examined at a 0.1% dose level. PGF2R produced a modest, transient increase in IOP followed by a decrease (Figure 1a). No meaningful effect on IOP was apparent at 6 h post-administration for PGF2R. PGF2R 15-acetyl ester produced a similar biphasic IOP response (Figure 1b) but the onsets of both the ocular hypertensive and hypotensive phases were later than the onset following exposure to PGF2R. PGF2R 1-isopropyl ester caused a pronounced increase in IOP followed by a profound and wellmaintained decrease (Figure 1c). Even at 10 h post-dosing, the ocular hypotensive response was maximal. In contrast to PGF2R 1-isopropyl ester, PGF2R 1,11-lactone produced an increase in IOP of similar magnitude to PGF2R (Figure 1d). The hypotensive phase of the response to PGF2R 1,11-lactone was, however, relatively protracted, and the greatest decreases in IOP were recorded at 8 h post-dosing. The effects of PGF2R 15-pivaloyl ester and PGF2R 11,15-dipivaloyl ester on IOP are depicted in Figures 1e and 1f, respectively. Both prodrugs produced an initial marked increase in IOP followed by a pronounced decrease. The ocular hypotensive response was fully maintained until the 10 h final time point for PGF2R 11,15-dipivaloyl ester whereas in the case of the 15-monopivaloyl ester, it had partially resolved by this time. The effects of PGF2R and its prodrugs on ocular surface hyperemia are shown in Figure 2. PGF2R caused an ocular
Journal of Pharmaceutical Sciences / 1181 Vol. 86, No. 10, October 1997
Figure 1sEffects of PGF2R and ester prodrugs on rabbit IOP following administration of a single 0.1% dose. Data are expressed as the change from baseline IOP in the treated eye (∆IOP) − the change from baseline IOP (∆IOP) in the control eye. Points are mean ± SEM. The IOP response for the individual prodrugs is depicted as follows: PGF2R (panel a), PGF2R 15-acetyl ester (panel b), PGF2R 1-isopropyl ester (panel c), PGF2R 1,11-lactone (panel d), PGF2R 15pivaloyl ester (panel e), PGF2R 11,15-dipivaloyl ester (panel f). Key to significance according to the Student’s paired t test: (*) p < 0.05; (**) p < 0.01 (n ) 6−9).
surface hyperemic response that was maintained over the initial 4 h period but then markedly declined by 6 h (Figure 2a). PGF2R 1-isopropyl ester caused a more marked degree of ocular surface hyperemia that remained virtually unchanged throughout the entire 10-h experimental time course (Figure 2b). PGF2R 1,11-lactone exhibited essentially the same profile of activity as PGF2R (Figure 2c). PGF2R 15-pivaloyl ester caused a very pronounced initial ocular surface hyperemic effect, which was maintained at a lesser degree between the 2- and 6-h time points (Figure 2d). The ocular surface hyperemic response to PGF2R 15-pivaloyl ester had resolved by the 10-h time point. The ocular surface hyperemic response to PGF2R 11,15-dipivaloyl ester also exhibited an initial peak at the 1-h time point but then steadily declined until the 8-h time point when it resolved (Figure 2e). The ocular hypotensive and hyperemic effects occurred simultaneously after the prodrug administration, so an effective separation for test compounds was observed only when the hypotension notably persisted after the disappearance of hyperemia. The 11,15-dipivaloyl ester and 1,11-lactone prodrugs both produced a long-lasting hypotension that endured beyond the hyperemic effect. PGF2R 15-pivaloyl ester showed a smaller temporal separation. The duration of hypotension and hyperemia were similar for PGF2R and other test compounds. The penetration profiles of the test compounds across cornea and conjunctiva are shown in Figures 3 and 4, respectively. Except for the 15-acetyl ester, all other prodrugs penetrated the conjunctiva faster than PGF2R by 3- to 17-fold. The 1,11-lactone prodrug showed the highest conjunctival permeability (29.8 × 10-6 cm/s). The rank order of conjunctival permeability of the test compounds was: 1,11-lactone > 1-isopropyl > 15-pivaloyl > 1182 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
Figure 2sEffects of PGF2R and ester prodrugs on rabbit ocular surface hyperemia following administration of a single 0.1% dose. Points represent mean ± SEM scores (n ) 6−9). The ocular surface hyperemic response for the individual prodrugs is depicted as follows: PGF2R (panel a), PGF2R 1-isopropyl ester (panel b), PGF2R 1,11-lactone (panel c), PGF2R 15-pivaloyl ester (panel d), PGF2R 11,15-dipivaloyl ester (panel e).
15-valeryl > 11,15-dipivaloyl > PGF2R ∼ 15-acetyl. PGF2R and its 15-acetyl prodrug were equieffective in penetrating the conjunctiva. The conjunctiva was more permeable than the cornea in the case of PGF2R and the other prodrugs. Further metabolism of PGF2R was not observed throughout the experiments. The permeability (Papp) results of the test compounds are summarized in Table 2. The corneal and conjunctival Papp of PGF2R were 0.2 ( 0.1 and 1.7 ( 0.8 × 10-6 cm/s (mean ( SD, n ) 5), respectively. The corneal Papp of the 1-isopropyl ester prodrug was 19.1 ( 7.6 × 10-6 cm/s (n ) 3), which is 83-fold faster than that of PGF2R. The 1,11-lactone prodrug penetrated the cornea faster than PGF2R by 76-fold, followed by, in rank order, 15-pivaloyl ester (21-fold), 15-valeryl ester (20fold), 11,15-dipivaloyl ester (17-fold), and 15-acetyl ester (4fold). In general, the prodrugs penetrated the cornea and conjunctiva more readily than PGF2R. The Papp increased as lipophilicity increased for the prodrugs with low to moderate permeability coefficients, but for prodrugs with higher permeability coefficients, Papp was actually lower. The 1-isopropyl ester and 1,11-lactone prodrugs, with moderate lipophilicity, were the most permeable in the corneal and conjunctival tissue, respectively. The degree of PGF2R formation reflected the metabolic stability of the test prodrugs in ocular tissues according to ester hydrolysis. The rank order for total formation of PGF2R in corneal perfusion was 15-valeryl ester (20.2%) > 1-isopropyl (15.7%) > 15-acetyl (8.6%) > 15-pivaloyl (2.3%) > 11,15dipivaloyl (0.8%) > 1,11-lactone (0.5%). When prodrug was perfused in the conjunctiva, the rank order for total PGF2R formation was 15-valeryl (28.5%) > 1-isopropyl (10.9%) > 15pivaloyl (6.7%) ∼ 15-acetyl (5.6%) > 11,15-dipivaloyl (0.5%) ∼ 1,11-lactone (0.4%). Overall, 1,11-lactone was the most stable among the test prodrugs, followed by the 11,15dipivaloyl, 15-pivaloyl, 15-acetyl, 1-isopropyl, and 15-valeryl esters. The amounts of PGF2R formed by corneal and conjunctival tissues were similar. Following addition of the iris/
Figure 3sCorneal penetration of PGF2R and its prodrugs. At time ) 0 min, PGF2R or the indicated prodrug was placed in the donor compartment of a flow-through chamber mounted with the rabbit cornea. Samples were taken from the acceptor compartment at various times thereafter and were analyzed for metabolites by HPLC. Data points are the mean ± SD (n ) 2−6).
Figure 4sConjunctival penetration of PGF2R and its prodrugs. At time ) 0 min, PGF2R or the indicated prodrug was placed in the donor compartment of a flowthrough chamber mounted with rabbit conjunctiva. Samples were taken from the acceptor compartment at various times thereafter and were analyzed for metabolites by HPLC. Data points are the mean ± SD (n ) 2−6).
Journal of Pharmaceutical Sciences / 1183 Vol. 86, No. 10, October 1997
Table 2sOcular Pharmacologic Effects, Permeabilities, and Bioavailability of PGF2r and its Prodrugsa,b PGF2R
5.0
15-Acetyl
1-Isopropyl
>4
0.2 (0.1)
1.0 (1.2,0.9)
0.2 (0.1)
0.4 (0.5,0.2)
1.00 6 1.7 (0.8) 100
15-Pivaloyl
Hypotension Maximal IOP reduction (mmHg) 12.0 7.1 9.7 Duration of IOP effect (h) >6 >10 >10 Corneal Papp (×10-6 cm/s) 19.1 17.7 5.2 (7.6) (19.0,16.4) (4.7,5.6) Bioavailable PGF2R (%dose) for hypotension 5.4 0.7 1.9 (5.6,5.1) (0.7,0.7) (1.7,2.0) Hyperemia Maximal ocular surface hyperemic effect (score) 1.31 0.83 1.75 Duration of hyperemia (h) >6 8 8 Conjunctival Papp (x10-6 cm/s) 21.3 29.8 15.2 (2.8) (29.0,30.5) (12.1,18.3) Bioavailable PGF2R (%dose) for hyperemia 12.6 0.3 3.8 (5.9) (0.1,0.5) (3.4,4.1) Separation indexd 0.43 2.33 0.50
3.3
6
1,11-Lactone
1.5 (2.3,0.7) 6.7 (6.0,7.5)
0.002
15-Valerylc
11,15-Dipivaloyl
10.5 12 4.6 (4.8,4.4)
4.0 (4.1,4.0)
1.5 (1.8,1.2)
0.9 (0.8,1.0)
1.42 10 8.7 (8.6,8.7)
5.5 (5.5,5.5)
20.1 25.1,15.1)
0.5 0.6,0.4) 1.80
a Applied dose ) 0.1%, site of hypotension is ciliary body and the site of hyperemia is conjunctiva. b Permeability and bioavailability data are expressed as mean (SD), n ) 3−6, or mean individual observations, n ) 2. c No in vivo data obtained. d Index ) bioavailable PGF2R for hypotension)/(bioavailable PGF2R for hyperemia).
Table 3sPercent of Drug Present as PGF2r in Individual Chambers when Prodrugs were Perfused with Cornea, Conjunctiva, and Cornea/Iris-Ciliary Bodya In Vitrob Prodrug 15-Acetyl ester Donor Acceptor 1-Isopropyl ester Donor Acceptor 1,11-Lactone Donor Acceptor 15-Pivaloyl ester Donor Acceptor 15-Valeryl ester Donor Acceptor 11,15-Dipivaloyl ester Donor Acceptor
Cornea
Conjunctiva
6.4, 10.3 86.9, 77.1
5.2, 7.5 45.3, 64.1
9.7, 7.7, 23.9 55.6, 46.0, 99.6
3.5, 1.3, 17.5 89.5, 92.0, 80.3
Cornea/IrisCiliary body NDc ND 22.9, 14.6 97.4, 93.8
0.1, 0.4 4.7, 5.8
0.0, 0.7 1.5, 1.7
0.2, 0.2 12.6, 13.3
0.7, 1.1 100.0, 100.0
2.0, 2.5 91.4, 95.9
1.3, 1.3 100.0, 100.0
27.3, 14.0 99.2, 99.2
18.1, ND 100.0, 100.0
13.0, 11.8 100.0, 100.0
0.3, 0.2 45.4, 50.2
0.4, 0.3 15.6, 13.3
0.8, 0.5 63.2, 61.1
a Iris-ciliary body was immersed in the acceptor fluid. b 100 × PGF /(PGF 2R 2R + prodrug), individual observations. c ND, Not determined.
ciliary body in the corneal chamber, the formation of PGF2R increased about two-fold for 1,11-lactone and 11,15-dipivaloyl prodrug. The percent of drug in the individual chambers present as PGF2R when prodrugs were perfused with the ocular tissues is shown in Table 3. The 1,11-lactone penetrated ocular membranes mostly as the intact prodrug. The 11,15-dipivaloyl ester penetrated the cornea as both PGF2R (48%) and the 15-pivaloyl ester (50%), whereas it penetrated the conjunctiva mostly as the intact diester prodrug (56%) and the 15-pivaloyl ester (31%). The 11-pivaloyl ester was not de1184 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
tected in the perfusion fluids for all experiments. The bioconversion pattern of 11,15-dipivaloyl indicated that the C-11 ester linkage was more susceptible to hydrolysis than the C-15 ester. When iris/ciliary body tissue was placed in the acceptor chamber during corneal perfusion, the bioconversion of prodrugs was not significantly changed, except for the 1,11-lactone. The iris/ciliary body increased PGF2R formation for the 1,11-lactone prodrug two-fold. For the prodrugs that were extensively metabolized by the cornea, the iris/ ciliary body played only a minor role in their intraocular bioconversion to PGF2R. During corneal perfusion, the major portion (>80%) of the PGF2R formed from the 15-acetyl, 15-valeryl, and 1-isopropyl ester prodrugs was present in the precorneal (donor) fluid. In the cases of PGF2R 15-pivaloyl ester and PGF2R 1,11-lactone, less than one-third PGF2R was detected in the precorneal area. The disposition pattern of PGF2R suggested that the major metabolic site of the 15-pivaloyl, 11,15-dipivaloyl, and 1,11lactone prodrugs was in the inner layer of the cornea, whereas the 15-acetyl, 15-valeryl, and 1-isopropyl ester prodrugs were metabolized in the outer layer of the cornea. Ocular Bioavailabilitysthe bioavailability of PGF2R for ocular hypotension was determined from the formation of PGF2R (% of applied dose) in the acceptor fluid when a prodrug was perfused with cornea and iris/ciliary body The relationship between the maximal IOP reduction and the PGF2R bioavailability for the prodrugs tested was characterized by Michaelis-Menten kinetics (Figure 5a) and a correlation coefficient of 0.98 was obtained. The maximal ocular hypotension (IOPVmax) was 12.5 ( 2.0 mmHg, and the bioavailable PGF2R that produced one-half of IOPVmax was 0.46 ( 0.26% of the applied dose. Thus, the ocular hypotension appeared to be correlated to the formation of PGF2R in the ocular anterior segment fluid after prodrug application. PGF2R and its ester prodrugs are also known to cause an initial ocular hypertensive response in rabbits,7,8 so the relationship between maximal IOP increase and PGF2R bioavailability after prodrug administration was similarly char-
also subjected to Michaelis-Menten analysis. The maximal ocular surface hyperemic response was 11.41 (area under curve), and the bioavailable PGF2R that produced a half maximal response was 0.56 ( 0.20% It appeared that there is a correlation between PGF2R formation in the conjunctiva and ocular surface redness. The separation index of the test compound was defined as the ratio of the bioavailable PGF2R formed for the ocular hypotension to that for the hyperemia (Table 2). The 1,11lactone (2.33) and the 11,15-dipivaloyl (1.80) prodrugs showed the largest separation index among the test prodrugs and also most effectively separated the ocular hypotension from hyperemia in the animal model. The test compounds with indices less than that of the 15-pivaloyl monoester (0.50) did not show any clear pharmacologic separation. The chemical stability of each prodrug was monitored in GBR buffer at 35 °C for 4 h. No significant formation of PGF2R was measured during this time period. For all experiments, the test prodrugs were quantitatively recovered as unchanged species and PGF2R at the end of the experiments, except for the 1,11-lactone and 11,15-dipivaloyl ester. The recovery of 1,11-lactone and 11,15-dipivaloyl ester was ≈83-88% of the applied dose, which was likely due to the drug accumulation in the ocular tissues or adsorption to the perfusion chambers. Approximately 1.2% and 1.0% of the applied 1,11-lactone and 11,15-dipivaloyl ester prodrugs, respectively, were recovered from the corneal tissue. The substantial accumulation of prodrugs in the cornea may prolong the duration of the ocular hypotensive effect. Accumulation of other test compounds was not noted in these ocular tissues.
Discussion
Figure 5sPanel a: correlation of ocular hypotension with PGF2R bioavailability. For PGF2R and each prodrug, the maximal IOP decrease obtained was plotted versus bioavailable PGF2R. Panel b: correlation of ocular hypertension with PGF2Rbioavailability. For PGF2R and each prodrug, the maximal IOP increase observed was plotted versus bioavailable PGF2R. Panel c: correlation of ocular surface hyperemia with PGF2R bioavailability. For PGF2R and each prodrug, the maximal hyperemic increase observed was plotted versus bioavailable. The individual data for panels (a) and (c) are given in Table 2.
acterized by Michaelis-Menten kinetics (Figure 5b). A correlation coefficient of 0.95 was obtained. The maximal ocular hypertensive response (IOPVmax) was 16.70 ( 2.2 mmHg, and the bioavailable PGF2R that would produce onehalf of IOPv max was 1.36 ( 1.14% of the applied dose. Similar to ocular hypotension, increases in IOP appeared to correlate with the formation of PGF2R in the ocular anterior segment fluid following PGF2R ester prodrug administration. The relationship between bioavailable PGF2R and conjunctival hyperemia was also determined (Figure 5c). No data for PGF2R 15-acetyl ester and PGF2R 15-valeryl ester were obtained in this regard. Consideration of the data suggested that an “area under the curve” analysis of the ocular surface hyperemia data might be more appropriate. The relationship between this biological parameter and bioavailable PGF2R was
PGF2R lowers IOP in humans and animals1-8 and would be considered a promising treatment for glaucoma except for its side effects, which include conjunctival hyperemia, ocular discomfort, and headaches. In this study, PGF2R prodrugs with enhanced corneal and conjunctival permeability were tested in an attempt to minimize ocular surface bioavailability of nascent PGF2R and thereby limit its side effects while, at the same time, retain the highly efficacious ocular hypotensive properties typical of PGF2R. All the PGF2R prodrugs tested were more lipophilic than PGF2R and all penetrated the cornea and conjunctiva faster than PGF2R, except the 15-acetyl ester prodrug, which was equally permeable. However, a direct relationship was not observed between the degree of apparent permeability and drug lipophilicity. The two most lipophilic prodrugs, the 15valeryl and 11,15-dipivaloyl esters, were less permeable in the cornea and conjunctiva than the less lipophilic prodrugs PGF2R 1-isopropyl ester, 1,11-lactone, and 15-pivaloyl ester. A similar bell-shaped relationship was noted for the corneal penetration of a series of timolol prodrugs and n-alkyl paminobenzoate esters in rabbits.10,11 The greater permeability of the 1-isopropyl ester and 1,11-lactone prodrugs relative to PGF2R and the other prodrugs might have been affected by their state of ionization in the perfusion medium. Both compounds are present as the un-ionized form, and their permeability might have been enhanced relative to the ionized species. Once PGF2R was generated from the prodrug, it was not further metabolized, as illustrated by the linear increase in PGF2R formed with time. Cheng-Bennett et al.12 also showed that PGF2R was not metabolized in rabbit or human ocular tissue. The major objective of this study was to identify the PGF2R prodrugs that exhibited a favorable ratio or separation index between bioavailability for hypotension and bioavailability for hyperemia. An early ocular hypertensive response
Journal of Pharmaceutical Sciences / 1185 Vol. 86, No. 10, October 1997
occurs in rabbits, which complicates evaluating degree of efficacy versus hyperemia at early time points. Nevertheless, this relationship can be satisfactorily estimated at later time points. Thus, the pharmacological results showed that only the 11,15-dipivaloyl and 1,11-lactone prodrugs exhibited some temporal separation between ocular hypotension and hyperemia among the test prodrugs. The 15-pivaloyl ester also demonstrated a small separation between these pharmacologic effects. These compounds exhibited a favorable combination between the rate and site of bioconversion in the ocular tissues. In this study, the Michaelis-Menten results were calculated for the ocular hypotension, hypertension, and surface hyperemia from bioavailability-effect curves. This approach was to demonstrate the achievable maximal effect for these ester prodrugs of PGF2Rin the rabbit model. The km (the % bioavailability to reach half of the maximal effect) values were calcuated to compare the influence of the presence of PGF2R in the ocular tissues on pharmacological responses. For example, the km values for hypotension, hypertension, and hyperemia were ∼0.46%, 1.36%, and 0.56%, respectively, indicating that the induction of the IOP reduction and surface hyperemia by the formation of PGF2R was similar, whereas more PGF2R was required to exhibit ocular hypertension. As shown in rabbit and pig ocular tissues,13,14 prodrug metabolism occurred mainly in the inner corneal layer and/ or the iris/ciliary body, whereas some prodrugs can be hydrolyzed in the corneal surface layer. The 1,11-lactone prodrug underwent slow hydrolysis in the ocular tissues, probably due to the intracyclic ester linkage. The 15-pivaloyl prodrug was enzymatically more stable than the 1-isopropyl and other 15-monoester prodrugs because of its bulky pivaloyl moiety that possibly impeded access of esterases to hydrolyze the pivaloyl ester linkage. The 11,15-dipivaloyl prodrug contains two pivaloyl groups and was more enzymatically stable than the 15-pivaloyl ester. The results of this study indicate that the bioconversion site of the prodrugs was also important for the pharmacological
1186 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
response. Intraocular conversion was more important than the surface enzymatic hydrolysis of PGF2R prodrugs in separating pharmacological effects. A prodrug that was stable at the ocular surface would be metabolized intraocularly, whereas an enzymatically labile prodrug would be hydrolyzed in the surface ocular tissues. For the compounds tested, prodrugs that were stable at the ocular surface appeared to separate the pharmacological effects more effectively than the labile prodrugs. The development of a potent PGF2R prodrug with minimal side effects awaits definitive information on esterase distribution and activity in the ocular tissues.
References and Notes 1. Mishima, S.; Masuda, K. Metab. Pediat. Ophthalmol. 1979, 3, 179-186. 2. Giuffre, G. Graefe's Arch. Ophthalmol. 1985, 222, 139-141. 3. Alm, A.; Villumsen, J. Proc. Int. Soc. Eye Res. 1986, 4, 14-20. 4. Bito, L. Z. Exp. Eye Res. 1984, 38, 181-194. 5. Bito, L. Z.; Kirby, H. E.; Baroody, R. A.; Miranda, O. C. Invest. Ophthalmol. Vis. Sci. Suppl. 1986, 27, 178. 6. Villumsen, J.; Alm, A. Invest. Ophthalmol. Vis. Sci. Suppl. 1987, 28, 378. 7. Woodward, D. F.; Burke, J. A.; Williams, L. S.; Palmer, B. P.; Wheeler, L. A.; WoldeMussie, E.; Ruiz, G, Chen, J. Invest. Ophthalmol. Vis. Sci. 1989, 30, 1838. 8. Woodward, D. F.; Chan, M. F.; Burke, J. A.; Cheng-Bennett, A.; Chen, G.; Fairbairn, C. E.; Gac, T.; Garst, M. E.; Gluchowski, C.; Kaplan, L. J.; Lawrence, R. A.; Roof, M.; Sachs, G.; Shan, T.; Wheeler, L. A.; Williams, L. S. J. Ocular Pharmacol. 1994, 10, 177. 9. Richman, J. B.; Tang-Liu, D. D.-S. J. Pharm. Sci. 1990, 153157. 10. Chien, D.-S.; Bundgaard, H.; Lee, V. H. L. J. Ocular Pharmacol. 1988, 4, 137-146. 11. Mosher, G.; Mikkelson, T. Int. J. Pharm. 1979, 2, 239-243. 12. Cheng-Bennett, A.; Poyer, J.; Weinkam, R. J.; Woodward, D. F. Invest. Ophthalmol Vis. Sci. 1990, 31, 1389-1393. 13. Camber, O.; Edman, P.; Olsson, L.-I. Int. J. Pharm. 1986, 29, 259-266. 14. Bito, L. Z.; Baroody, R. A. Exp. Eye Res. 1987, 44, 217-226.
JS950373B
AND
DAVID. F. WOODWARD§
Received September 1, 1995, from the *Department of Pharmacokinetics and §Department of Biological Sciences, Allergan, Inc., 2525 Dupont Drive, Irvine, CA 92715, and ‡Pharmacokinetics and Drug Metabolism, Bayer Corporation, 400 Morgan Lane, West Haven, CT 06516. Final revised manuscript received March 3, 1997. Accepted for publication June 17, 1997X. Abstract 0 The objective of this study was to identify prostaglandin F2R (PGF2R) prodrugs that have an optimal ocular absorption profile and therefore could be potentially useful for the treatment of glaucoma. Rabbit cornea, conjunctiva, and iris/ciliary body were mounted in a flow-through chamber to evaluate the permeability and bioconversion of PGF2R and its prodrugs. The prodrugs tested were PGF2R 1-isopropyl, 1,11-lactone, 15-acetyl, 15-pivaloyl, 15-valeryl, and 11,15-dipivaloyl esters. After 4 h in the donor or acceptor compartments, the products and formation of PGF2R were analyzed by HPLC. Effects on intraocular pressure and ocular surface hyperemia were also determined. All prodrugs penetrated the rabbit cornea faster than PGF2R by 4- to 83-fold. All prodrugs penetrated conjunctiva faster than PGF2R, except the 15-acetyl ester prodrug, which was equally permeable. No direct correlation between drug lipophilicity and permeability across the cornea or conjunctiva was apparent. The most metabolically stable prodrug was the 1,11-lactone, followed by the 11,15-dipivaloyl, 15-pivaloyl, 15-acetyl, 1-isopropyl, and the 15-valeryl esters, the latter of which was extensively converted to PGF2R. A separation index for various prodrugs was calculated from the ratio of the bioavailable PGF2R for ocular hypotension to the bioavailable PGF2R for hyperemia. The highest separation index was observed for the 1,11-lactone prodrug (2.33), followed by the 11,15dipivaloyl ester prodrug (1.80). Thus the 1,11-lactone and 11,15-dipivaloyl ester prodrugs appeared to be superior to the others in providing bioavailable PGF2R for ocular hypotension, while minimizing hyperemia. The favorable separation index for these compounds appeared to be due to their metabolic stability at the corneal surface and conjunctiva combined with sufficient bioavailability for ocular hypotension.
Introduction Prostaglandins mediate a spectrum of biological responses in the eye. In primary, open-angle glaucoma, topical prostaglandins have been used to reduce intraocular pressure.1-3 However, because the cornea is not readily permeable to PGF2R, a relatively high concentration of PGF2R is necessary for effective intraocular pressure reduction, causing conjunctival hyperemia, ocular discomfort, headaches, and other side effects.2 A more lipophilic PGF2R prodrug, PGF2R 1-isopropyl ester, was synthesized to increase ocular bioavailability and thus reduce the effective dose in an attempt to minimize the side effects. Although studies in humans and several animal species showed that PGF2R 1-isopropyl ester was more potent than PGF2R in lowering intraocular pressure, it also caused ocular discomfort and hyperemia.4-6 To identify a PGF2R prodrug with increased ocular hypotensive efficacy and reduced ocular surface hyperemia, a series of PGF2R-monoesters, -diesters, and -lactone prodrugs were tested in rabbits. Penetration across the conjunctiva and cornea were evaluated as well as the sites of metabolism to nascent PGF2R. Bioavailability values for the hypotensive and X
Abstract published in Advance ACS Abstracts, August 1, 1997.
1180 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
hyperemic effects of PGF2R were calculated to predict the most promising prodrugs for the treatment of glaucoma in humans.
Experimental Procedures Section MaterialssProstaglandin F2R and the PGF2R prodrugs tested in this study are summarized as follows:
Compound
Perfusion Concentration (µg/mL)
PGF2R
251
PGF2R-15-acetyl ester
202
PGF2R-1-isopropyl ester
314
PGF2R-1,11-lactone
194
PGF2R-15-pivaloyl ester
106
PGF2R-15-valeryl ester
314
PGF2R-11,15-dipivaloyl ester
95
Source Lot # PF2041, Toray Industrial, Inc, New York, NY Chemical Sciences, Allergan, Inc, Irvine, CA Chemical Sciences, Allergan, Inc, Irvine, CA Lot #12070-GLB-96, Upjohn Co. (U-49653), Kalamazoo, MI Chemical Sciences, Allergan, Inc. Irvine, CA Chemical Sciences, Allergan, Inc. Irvine, CA Chemical Sciences, Allergan, Inc. Irvine, CA
The chemical structures, molecular weights, and apparent partition coefficients of these compounds are listed in Table 1. All other solvents and chemicals were of analytical grade and were used as received. MethodssAdult female New Zealand albino rabbits, weighing 2.0-2.5 kg, were used in this study. All procedures were performed in accordance with the ARVO resolution on the use of animals in research. The flow-through diffusion chamber7 was used to measure the permeability of the test compounds across rabbit cornea and conjunctiva. All ocular tissues were freshly dissected before use. The rabbit cornea or conjunctiva was secured between two chambers without any wrinkles or folding. PGF2R and all prodrug solutions were freshly prepared in glutathione bicarbonated Ringer's solution (GBR) containing 1% propylene glycol (v/v). After the ocular membrane was mounted, 2.5 mL of the drug-free GBR solution, preheated to 35 °C, was placed into the chamber facing the endothelial side (acceptor chamber). The drug solution was then slowly infused through the chamber facing the epithelium (donor chamber) at a flow rate of 25.8 µL/min. The volumes of the acceptor and donor chambers were 2.5 and 0.125 mL, respectively. The penetration areas of the cornea and conjunctiva were 1.09 and 0.95 cm2, respectively. In several corneal penetration experiments, the iris/ciliary body was dissected and placed in the acceptor chamber to mimic the intraocular condition in vivo. The experiments were conducted at 35 °C for 4 h in a waterjacketed chamber. Fifty microliter samples of acceptor fluid and anterior effluent were collected every 30 min throughout the perfusion period. The samples were mixed immediately with acetonitrile/0.02 M KH2PO4 (pH 3.2) to terminate any enzymatic reaction before analysis of the test compounds and their metabolic products by HPLC.
S0022-3549(95)00373-X CCC: $14.00
© 1997, American Chemical Society and American Pharmaceutical Association
Table 1sChemical Structure of PGF2r and Its Prodrugs MWa
log Pb
PGF2R 15-Acetyl ester 1-Isopropyl ester 1,11-Lactone
355 396 396 337
1.26 2.16 2.50 2.61
H CH3CO H
15-Pivaloyl ester 15-Valeryl ester 11,15-Dipivaloyl ester
439 439 523
3.50 3.75 5.0
C(CH3)3CO C4H9CO C(CH3)3CO
compound
RI(C-15)
R2(C-1)
H H CH2(CH3)2
H H H
R3(C-11)
H H H
H H C(CH3)3CO
a Molecular weight. b Log P was determined from the Pomona Med. Chem. Software system or measured by HPLC.
Scheme 1sCorneal penetration of PGF2R prodrugs. The chemical stability of each compound was monitored for 4 h in GBR solution at 35 °C. A reversed-phase HPLC analysis was performed with a Beckman ODS, 5 µm particle column with dimensions 4.6 mm by 15 cm. A WISP 710B injector (Waters Associates) and a Beckman 114M pump were used to apply the samples and deliver the mobile phase, respectively. The flow rate of the mobile phase (acetonitrile/0.02 M KH2PO4, pH 3.2, adjusted by phosphoric acid) was 1.0 to 1.5 mL/ min. Drug concentration was monitored by a Kratos 783 UV spectrophotometer at a wavelength of 200 nm. The cumulative amounts (µg) of PGF2R or prodrug in the acceptor chamber were plotted versus time over a 4-h period. The slope of the linear portion of the curve was calculated by linear regression analysis to determine the steady-state flux rate (µg/min). The apparent ocular permeability (Papp) of the test compound was calculated by dividing the steady-sate flux rate by the surface area of the membrane and the initial concentration of the test compound. The bioavailability of PGF2R for the hypotensive pharmacological effect is a function of the corneal permeability and conversion of the prodrug. The amount of PGF2R was measured at the end of each experiment to determine the enzymatic stability and site of bioconversion of the test prodrug in the ocular tissues. Both the iris/ciliary body and the cornea are possible sites for prodrug conversion to PGF2R. Thus, the bioavailable PGF2R in the ocular anterior fluid was determined from the formation of PGF2R in the acceptor fluid when prodrug was perfused with the cornea and the iris/ciliary body. The ocular absorption and disposition of the test prodrug are depicted in Scheme 1. The hyperemic effect of PGF2R is the result of PGF2R being enzymatically formed from its prodrug at the ocular surface. Because the PGF2R formed from the corneal surface would be readily absorbed into the conjunctiva, the bioavailability of PGF2R for the hyperemic effect was calculated from the total amount of PGF2R formed by the
Scheme 2sConjunctival penetration of PGF2R prodrugs. conjunctiva in the acceptor fluid and the corneal surface in the donor fluid. The absorption and disposition of PGF2R prodrug in the conjunctiva is presented in Scheme 2. The ratio of the bioavailable PGF2R for ocular hypotension to that for hyperemia is used as the separation index for the test prodrugs. Presumably, the larger the separation index of a prodrug, the better the separation between ocular hypotension and hyperemia that could be obtained after topical administration. Intraocular pressure (IOP) was measured with a pneumatonometer (Digilab) calibrated against the eyes of anesthetized rabbits by closed stopcock manometry. The correlation coefficient over the range 10-30 mmHg was 0.98. The animals were acclimated to pneumatonometry by taking unrecorded measurements before experimental determination of IOP. Corneal anesthesia for tonometry was provided by topical application of one drop of 0.05% proparacaine. All IOP studies were performed in a masked fashion. Animals were randomly assigned to each treatment group, and test and control eyes were also selected randomly. Solutions were instilled into the lower conjunctival sac in a 25 µL volume. Drug solution was administered to one eye and the contralateral eye received vehicle as a control. IOP was recorded immediately before topical dosing and then typically at 1, 2, 4, 6, 8, and 10 h after dosing. Changes in IOP were expressed as the change from baseline IOP (∆ IOP) in the treated eye minus the change from baseline IOP (∆ IOP) in the control eye and statistically analyzed by Student’s t test. Ocular surface hyperemia was scored according to the following graded scale: 1 ) slight with some vessels definitely injected above normal; 2 ) moderate ) eye is a more uniform, diffuse red; 3 ) severe ) a diffuse “beefy” red eye.
Results The effects of PGF2R and its prodrugs on IOP are depicted in Figure 1. All compounds were examined at a 0.1% dose level. PGF2R produced a modest, transient increase in IOP followed by a decrease (Figure 1a). No meaningful effect on IOP was apparent at 6 h post-administration for PGF2R. PGF2R 15-acetyl ester produced a similar biphasic IOP response (Figure 1b) but the onsets of both the ocular hypertensive and hypotensive phases were later than the onset following exposure to PGF2R. PGF2R 1-isopropyl ester caused a pronounced increase in IOP followed by a profound and wellmaintained decrease (Figure 1c). Even at 10 h post-dosing, the ocular hypotensive response was maximal. In contrast to PGF2R 1-isopropyl ester, PGF2R 1,11-lactone produced an increase in IOP of similar magnitude to PGF2R (Figure 1d). The hypotensive phase of the response to PGF2R 1,11-lactone was, however, relatively protracted, and the greatest decreases in IOP were recorded at 8 h post-dosing. The effects of PGF2R 15-pivaloyl ester and PGF2R 11,15-dipivaloyl ester on IOP are depicted in Figures 1e and 1f, respectively. Both prodrugs produced an initial marked increase in IOP followed by a pronounced decrease. The ocular hypotensive response was fully maintained until the 10 h final time point for PGF2R 11,15-dipivaloyl ester whereas in the case of the 15-monopivaloyl ester, it had partially resolved by this time. The effects of PGF2R and its prodrugs on ocular surface hyperemia are shown in Figure 2. PGF2R caused an ocular
Journal of Pharmaceutical Sciences / 1181 Vol. 86, No. 10, October 1997
Figure 1sEffects of PGF2R and ester prodrugs on rabbit IOP following administration of a single 0.1% dose. Data are expressed as the change from baseline IOP in the treated eye (∆IOP) − the change from baseline IOP (∆IOP) in the control eye. Points are mean ± SEM. The IOP response for the individual prodrugs is depicted as follows: PGF2R (panel a), PGF2R 15-acetyl ester (panel b), PGF2R 1-isopropyl ester (panel c), PGF2R 1,11-lactone (panel d), PGF2R 15pivaloyl ester (panel e), PGF2R 11,15-dipivaloyl ester (panel f). Key to significance according to the Student’s paired t test: (*) p < 0.05; (**) p < 0.01 (n ) 6−9).
surface hyperemic response that was maintained over the initial 4 h period but then markedly declined by 6 h (Figure 2a). PGF2R 1-isopropyl ester caused a more marked degree of ocular surface hyperemia that remained virtually unchanged throughout the entire 10-h experimental time course (Figure 2b). PGF2R 1,11-lactone exhibited essentially the same profile of activity as PGF2R (Figure 2c). PGF2R 15-pivaloyl ester caused a very pronounced initial ocular surface hyperemic effect, which was maintained at a lesser degree between the 2- and 6-h time points (Figure 2d). The ocular surface hyperemic response to PGF2R 15-pivaloyl ester had resolved by the 10-h time point. The ocular surface hyperemic response to PGF2R 11,15-dipivaloyl ester also exhibited an initial peak at the 1-h time point but then steadily declined until the 8-h time point when it resolved (Figure 2e). The ocular hypotensive and hyperemic effects occurred simultaneously after the prodrug administration, so an effective separation for test compounds was observed only when the hypotension notably persisted after the disappearance of hyperemia. The 11,15-dipivaloyl ester and 1,11-lactone prodrugs both produced a long-lasting hypotension that endured beyond the hyperemic effect. PGF2R 15-pivaloyl ester showed a smaller temporal separation. The duration of hypotension and hyperemia were similar for PGF2R and other test compounds. The penetration profiles of the test compounds across cornea and conjunctiva are shown in Figures 3 and 4, respectively. Except for the 15-acetyl ester, all other prodrugs penetrated the conjunctiva faster than PGF2R by 3- to 17-fold. The 1,11-lactone prodrug showed the highest conjunctival permeability (29.8 × 10-6 cm/s). The rank order of conjunctival permeability of the test compounds was: 1,11-lactone > 1-isopropyl > 15-pivaloyl > 1182 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
Figure 2sEffects of PGF2R and ester prodrugs on rabbit ocular surface hyperemia following administration of a single 0.1% dose. Points represent mean ± SEM scores (n ) 6−9). The ocular surface hyperemic response for the individual prodrugs is depicted as follows: PGF2R (panel a), PGF2R 1-isopropyl ester (panel b), PGF2R 1,11-lactone (panel c), PGF2R 15-pivaloyl ester (panel d), PGF2R 11,15-dipivaloyl ester (panel e).
15-valeryl > 11,15-dipivaloyl > PGF2R ∼ 15-acetyl. PGF2R and its 15-acetyl prodrug were equieffective in penetrating the conjunctiva. The conjunctiva was more permeable than the cornea in the case of PGF2R and the other prodrugs. Further metabolism of PGF2R was not observed throughout the experiments. The permeability (Papp) results of the test compounds are summarized in Table 2. The corneal and conjunctival Papp of PGF2R were 0.2 ( 0.1 and 1.7 ( 0.8 × 10-6 cm/s (mean ( SD, n ) 5), respectively. The corneal Papp of the 1-isopropyl ester prodrug was 19.1 ( 7.6 × 10-6 cm/s (n ) 3), which is 83-fold faster than that of PGF2R. The 1,11-lactone prodrug penetrated the cornea faster than PGF2R by 76-fold, followed by, in rank order, 15-pivaloyl ester (21-fold), 15-valeryl ester (20fold), 11,15-dipivaloyl ester (17-fold), and 15-acetyl ester (4fold). In general, the prodrugs penetrated the cornea and conjunctiva more readily than PGF2R. The Papp increased as lipophilicity increased for the prodrugs with low to moderate permeability coefficients, but for prodrugs with higher permeability coefficients, Papp was actually lower. The 1-isopropyl ester and 1,11-lactone prodrugs, with moderate lipophilicity, were the most permeable in the corneal and conjunctival tissue, respectively. The degree of PGF2R formation reflected the metabolic stability of the test prodrugs in ocular tissues according to ester hydrolysis. The rank order for total formation of PGF2R in corneal perfusion was 15-valeryl ester (20.2%) > 1-isopropyl (15.7%) > 15-acetyl (8.6%) > 15-pivaloyl (2.3%) > 11,15dipivaloyl (0.8%) > 1,11-lactone (0.5%). When prodrug was perfused in the conjunctiva, the rank order for total PGF2R formation was 15-valeryl (28.5%) > 1-isopropyl (10.9%) > 15pivaloyl (6.7%) ∼ 15-acetyl (5.6%) > 11,15-dipivaloyl (0.5%) ∼ 1,11-lactone (0.4%). Overall, 1,11-lactone was the most stable among the test prodrugs, followed by the 11,15dipivaloyl, 15-pivaloyl, 15-acetyl, 1-isopropyl, and 15-valeryl esters. The amounts of PGF2R formed by corneal and conjunctival tissues were similar. Following addition of the iris/
Figure 3sCorneal penetration of PGF2R and its prodrugs. At time ) 0 min, PGF2R or the indicated prodrug was placed in the donor compartment of a flow-through chamber mounted with the rabbit cornea. Samples were taken from the acceptor compartment at various times thereafter and were analyzed for metabolites by HPLC. Data points are the mean ± SD (n ) 2−6).
Figure 4sConjunctival penetration of PGF2R and its prodrugs. At time ) 0 min, PGF2R or the indicated prodrug was placed in the donor compartment of a flowthrough chamber mounted with rabbit conjunctiva. Samples were taken from the acceptor compartment at various times thereafter and were analyzed for metabolites by HPLC. Data points are the mean ± SD (n ) 2−6).
Journal of Pharmaceutical Sciences / 1183 Vol. 86, No. 10, October 1997
Table 2sOcular Pharmacologic Effects, Permeabilities, and Bioavailability of PGF2r and its Prodrugsa,b PGF2R
5.0
15-Acetyl
1-Isopropyl
>4
0.2 (0.1)
1.0 (1.2,0.9)
0.2 (0.1)
0.4 (0.5,0.2)
1.00 6 1.7 (0.8) 100
15-Pivaloyl
Hypotension Maximal IOP reduction (mmHg) 12.0 7.1 9.7 Duration of IOP effect (h) >6 >10 >10 Corneal Papp (×10-6 cm/s) 19.1 17.7 5.2 (7.6) (19.0,16.4) (4.7,5.6) Bioavailable PGF2R (%dose) for hypotension 5.4 0.7 1.9 (5.6,5.1) (0.7,0.7) (1.7,2.0) Hyperemia Maximal ocular surface hyperemic effect (score) 1.31 0.83 1.75 Duration of hyperemia (h) >6 8 8 Conjunctival Papp (x10-6 cm/s) 21.3 29.8 15.2 (2.8) (29.0,30.5) (12.1,18.3) Bioavailable PGF2R (%dose) for hyperemia 12.6 0.3 3.8 (5.9) (0.1,0.5) (3.4,4.1) Separation indexd 0.43 2.33 0.50
3.3
6
1,11-Lactone
1.5 (2.3,0.7) 6.7 (6.0,7.5)
0.002
15-Valerylc
11,15-Dipivaloyl
10.5 12 4.6 (4.8,4.4)
4.0 (4.1,4.0)
1.5 (1.8,1.2)
0.9 (0.8,1.0)
1.42 10 8.7 (8.6,8.7)
5.5 (5.5,5.5)
20.1 25.1,15.1)
0.5 0.6,0.4) 1.80
a Applied dose ) 0.1%, site of hypotension is ciliary body and the site of hyperemia is conjunctiva. b Permeability and bioavailability data are expressed as mean (SD), n ) 3−6, or mean individual observations, n ) 2. c No in vivo data obtained. d Index ) bioavailable PGF2R for hypotension)/(bioavailable PGF2R for hyperemia).
Table 3sPercent of Drug Present as PGF2r in Individual Chambers when Prodrugs were Perfused with Cornea, Conjunctiva, and Cornea/Iris-Ciliary Bodya In Vitrob Prodrug 15-Acetyl ester Donor Acceptor 1-Isopropyl ester Donor Acceptor 1,11-Lactone Donor Acceptor 15-Pivaloyl ester Donor Acceptor 15-Valeryl ester Donor Acceptor 11,15-Dipivaloyl ester Donor Acceptor
Cornea
Conjunctiva
6.4, 10.3 86.9, 77.1
5.2, 7.5 45.3, 64.1
9.7, 7.7, 23.9 55.6, 46.0, 99.6
3.5, 1.3, 17.5 89.5, 92.0, 80.3
Cornea/IrisCiliary body NDc ND 22.9, 14.6 97.4, 93.8
0.1, 0.4 4.7, 5.8
0.0, 0.7 1.5, 1.7
0.2, 0.2 12.6, 13.3
0.7, 1.1 100.0, 100.0
2.0, 2.5 91.4, 95.9
1.3, 1.3 100.0, 100.0
27.3, 14.0 99.2, 99.2
18.1, ND 100.0, 100.0
13.0, 11.8 100.0, 100.0
0.3, 0.2 45.4, 50.2
0.4, 0.3 15.6, 13.3
0.8, 0.5 63.2, 61.1
a Iris-ciliary body was immersed in the acceptor fluid. b 100 × PGF /(PGF 2R 2R + prodrug), individual observations. c ND, Not determined.
ciliary body in the corneal chamber, the formation of PGF2R increased about two-fold for 1,11-lactone and 11,15-dipivaloyl prodrug. The percent of drug in the individual chambers present as PGF2R when prodrugs were perfused with the ocular tissues is shown in Table 3. The 1,11-lactone penetrated ocular membranes mostly as the intact prodrug. The 11,15-dipivaloyl ester penetrated the cornea as both PGF2R (48%) and the 15-pivaloyl ester (50%), whereas it penetrated the conjunctiva mostly as the intact diester prodrug (56%) and the 15-pivaloyl ester (31%). The 11-pivaloyl ester was not de1184 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
tected in the perfusion fluids for all experiments. The bioconversion pattern of 11,15-dipivaloyl indicated that the C-11 ester linkage was more susceptible to hydrolysis than the C-15 ester. When iris/ciliary body tissue was placed in the acceptor chamber during corneal perfusion, the bioconversion of prodrugs was not significantly changed, except for the 1,11-lactone. The iris/ciliary body increased PGF2R formation for the 1,11-lactone prodrug two-fold. For the prodrugs that were extensively metabolized by the cornea, the iris/ ciliary body played only a minor role in their intraocular bioconversion to PGF2R. During corneal perfusion, the major portion (>80%) of the PGF2R formed from the 15-acetyl, 15-valeryl, and 1-isopropyl ester prodrugs was present in the precorneal (donor) fluid. In the cases of PGF2R 15-pivaloyl ester and PGF2R 1,11-lactone, less than one-third PGF2R was detected in the precorneal area. The disposition pattern of PGF2R suggested that the major metabolic site of the 15-pivaloyl, 11,15-dipivaloyl, and 1,11lactone prodrugs was in the inner layer of the cornea, whereas the 15-acetyl, 15-valeryl, and 1-isopropyl ester prodrugs were metabolized in the outer layer of the cornea. Ocular Bioavailabilitysthe bioavailability of PGF2R for ocular hypotension was determined from the formation of PGF2R (% of applied dose) in the acceptor fluid when a prodrug was perfused with cornea and iris/ciliary body The relationship between the maximal IOP reduction and the PGF2R bioavailability for the prodrugs tested was characterized by Michaelis-Menten kinetics (Figure 5a) and a correlation coefficient of 0.98 was obtained. The maximal ocular hypotension (IOPVmax) was 12.5 ( 2.0 mmHg, and the bioavailable PGF2R that produced one-half of IOPVmax was 0.46 ( 0.26% of the applied dose. Thus, the ocular hypotension appeared to be correlated to the formation of PGF2R in the ocular anterior segment fluid after prodrug application. PGF2R and its ester prodrugs are also known to cause an initial ocular hypertensive response in rabbits,7,8 so the relationship between maximal IOP increase and PGF2R bioavailability after prodrug administration was similarly char-
also subjected to Michaelis-Menten analysis. The maximal ocular surface hyperemic response was 11.41 (area under curve), and the bioavailable PGF2R that produced a half maximal response was 0.56 ( 0.20% It appeared that there is a correlation between PGF2R formation in the conjunctiva and ocular surface redness. The separation index of the test compound was defined as the ratio of the bioavailable PGF2R formed for the ocular hypotension to that for the hyperemia (Table 2). The 1,11lactone (2.33) and the 11,15-dipivaloyl (1.80) prodrugs showed the largest separation index among the test prodrugs and also most effectively separated the ocular hypotension from hyperemia in the animal model. The test compounds with indices less than that of the 15-pivaloyl monoester (0.50) did not show any clear pharmacologic separation. The chemical stability of each prodrug was monitored in GBR buffer at 35 °C for 4 h. No significant formation of PGF2R was measured during this time period. For all experiments, the test prodrugs were quantitatively recovered as unchanged species and PGF2R at the end of the experiments, except for the 1,11-lactone and 11,15-dipivaloyl ester. The recovery of 1,11-lactone and 11,15-dipivaloyl ester was ≈83-88% of the applied dose, which was likely due to the drug accumulation in the ocular tissues or adsorption to the perfusion chambers. Approximately 1.2% and 1.0% of the applied 1,11-lactone and 11,15-dipivaloyl ester prodrugs, respectively, were recovered from the corneal tissue. The substantial accumulation of prodrugs in the cornea may prolong the duration of the ocular hypotensive effect. Accumulation of other test compounds was not noted in these ocular tissues.
Discussion
Figure 5sPanel a: correlation of ocular hypotension with PGF2R bioavailability. For PGF2R and each prodrug, the maximal IOP decrease obtained was plotted versus bioavailable PGF2R. Panel b: correlation of ocular hypertension with PGF2Rbioavailability. For PGF2R and each prodrug, the maximal IOP increase observed was plotted versus bioavailable PGF2R. Panel c: correlation of ocular surface hyperemia with PGF2R bioavailability. For PGF2R and each prodrug, the maximal hyperemic increase observed was plotted versus bioavailable. The individual data for panels (a) and (c) are given in Table 2.
acterized by Michaelis-Menten kinetics (Figure 5b). A correlation coefficient of 0.95 was obtained. The maximal ocular hypertensive response (IOPVmax) was 16.70 ( 2.2 mmHg, and the bioavailable PGF2R that would produce onehalf of IOPv max was 1.36 ( 1.14% of the applied dose. Similar to ocular hypotension, increases in IOP appeared to correlate with the formation of PGF2R in the ocular anterior segment fluid following PGF2R ester prodrug administration. The relationship between bioavailable PGF2R and conjunctival hyperemia was also determined (Figure 5c). No data for PGF2R 15-acetyl ester and PGF2R 15-valeryl ester were obtained in this regard. Consideration of the data suggested that an “area under the curve” analysis of the ocular surface hyperemia data might be more appropriate. The relationship between this biological parameter and bioavailable PGF2R was
PGF2R lowers IOP in humans and animals1-8 and would be considered a promising treatment for glaucoma except for its side effects, which include conjunctival hyperemia, ocular discomfort, and headaches. In this study, PGF2R prodrugs with enhanced corneal and conjunctival permeability were tested in an attempt to minimize ocular surface bioavailability of nascent PGF2R and thereby limit its side effects while, at the same time, retain the highly efficacious ocular hypotensive properties typical of PGF2R. All the PGF2R prodrugs tested were more lipophilic than PGF2R and all penetrated the cornea and conjunctiva faster than PGF2R, except the 15-acetyl ester prodrug, which was equally permeable. However, a direct relationship was not observed between the degree of apparent permeability and drug lipophilicity. The two most lipophilic prodrugs, the 15valeryl and 11,15-dipivaloyl esters, were less permeable in the cornea and conjunctiva than the less lipophilic prodrugs PGF2R 1-isopropyl ester, 1,11-lactone, and 15-pivaloyl ester. A similar bell-shaped relationship was noted for the corneal penetration of a series of timolol prodrugs and n-alkyl paminobenzoate esters in rabbits.10,11 The greater permeability of the 1-isopropyl ester and 1,11-lactone prodrugs relative to PGF2R and the other prodrugs might have been affected by their state of ionization in the perfusion medium. Both compounds are present as the un-ionized form, and their permeability might have been enhanced relative to the ionized species. Once PGF2R was generated from the prodrug, it was not further metabolized, as illustrated by the linear increase in PGF2R formed with time. Cheng-Bennett et al.12 also showed that PGF2R was not metabolized in rabbit or human ocular tissue. The major objective of this study was to identify the PGF2R prodrugs that exhibited a favorable ratio or separation index between bioavailability for hypotension and bioavailability for hyperemia. An early ocular hypertensive response
Journal of Pharmaceutical Sciences / 1185 Vol. 86, No. 10, October 1997
occurs in rabbits, which complicates evaluating degree of efficacy versus hyperemia at early time points. Nevertheless, this relationship can be satisfactorily estimated at later time points. Thus, the pharmacological results showed that only the 11,15-dipivaloyl and 1,11-lactone prodrugs exhibited some temporal separation between ocular hypotension and hyperemia among the test prodrugs. The 15-pivaloyl ester also demonstrated a small separation between these pharmacologic effects. These compounds exhibited a favorable combination between the rate and site of bioconversion in the ocular tissues. In this study, the Michaelis-Menten results were calculated for the ocular hypotension, hypertension, and surface hyperemia from bioavailability-effect curves. This approach was to demonstrate the achievable maximal effect for these ester prodrugs of PGF2Rin the rabbit model. The km (the % bioavailability to reach half of the maximal effect) values were calcuated to compare the influence of the presence of PGF2R in the ocular tissues on pharmacological responses. For example, the km values for hypotension, hypertension, and hyperemia were ∼0.46%, 1.36%, and 0.56%, respectively, indicating that the induction of the IOP reduction and surface hyperemia by the formation of PGF2R was similar, whereas more PGF2R was required to exhibit ocular hypertension. As shown in rabbit and pig ocular tissues,13,14 prodrug metabolism occurred mainly in the inner corneal layer and/ or the iris/ciliary body, whereas some prodrugs can be hydrolyzed in the corneal surface layer. The 1,11-lactone prodrug underwent slow hydrolysis in the ocular tissues, probably due to the intracyclic ester linkage. The 15-pivaloyl prodrug was enzymatically more stable than the 1-isopropyl and other 15-monoester prodrugs because of its bulky pivaloyl moiety that possibly impeded access of esterases to hydrolyze the pivaloyl ester linkage. The 11,15-dipivaloyl prodrug contains two pivaloyl groups and was more enzymatically stable than the 15-pivaloyl ester. The results of this study indicate that the bioconversion site of the prodrugs was also important for the pharmacological
1186 / Journal of Pharmaceutical Sciences Vol. 86, No. 10, October 1997
response. Intraocular conversion was more important than the surface enzymatic hydrolysis of PGF2R prodrugs in separating pharmacological effects. A prodrug that was stable at the ocular surface would be metabolized intraocularly, whereas an enzymatically labile prodrug would be hydrolyzed in the surface ocular tissues. For the compounds tested, prodrugs that were stable at the ocular surface appeared to separate the pharmacological effects more effectively than the labile prodrugs. The development of a potent PGF2R prodrug with minimal side effects awaits definitive information on esterase distribution and activity in the ocular tissues.
References and Notes 1. Mishima, S.; Masuda, K. Metab. Pediat. Ophthalmol. 1979, 3, 179-186. 2. Giuffre, G. Graefe's Arch. Ophthalmol. 1985, 222, 139-141. 3. Alm, A.; Villumsen, J. Proc. Int. Soc. Eye Res. 1986, 4, 14-20. 4. Bito, L. Z. Exp. Eye Res. 1984, 38, 181-194. 5. Bito, L. Z.; Kirby, H. E.; Baroody, R. A.; Miranda, O. C. Invest. Ophthalmol. Vis. Sci. Suppl. 1986, 27, 178. 6. Villumsen, J.; Alm, A. Invest. Ophthalmol. Vis. Sci. Suppl. 1987, 28, 378. 7. Woodward, D. F.; Burke, J. A.; Williams, L. S.; Palmer, B. P.; Wheeler, L. A.; WoldeMussie, E.; Ruiz, G, Chen, J. Invest. Ophthalmol. Vis. Sci. 1989, 30, 1838. 8. Woodward, D. F.; Chan, M. F.; Burke, J. A.; Cheng-Bennett, A.; Chen, G.; Fairbairn, C. E.; Gac, T.; Garst, M. E.; Gluchowski, C.; Kaplan, L. J.; Lawrence, R. A.; Roof, M.; Sachs, G.; Shan, T.; Wheeler, L. A.; Williams, L. S. J. Ocular Pharmacol. 1994, 10, 177. 9. Richman, J. B.; Tang-Liu, D. D.-S. J. Pharm. Sci. 1990, 153157. 10. Chien, D.-S.; Bundgaard, H.; Lee, V. H. L. J. Ocular Pharmacol. 1988, 4, 137-146. 11. Mosher, G.; Mikkelson, T. Int. J. Pharm. 1979, 2, 239-243. 12. Cheng-Bennett, A.; Poyer, J.; Weinkam, R. J.; Woodward, D. F. Invest. Ophthalmol Vis. Sci. 1990, 31, 1389-1393. 13. Camber, O.; Edman, P.; Olsson, L.-I. Int. J. Pharm. 1986, 29, 259-266. 14. Bito, L. Z.; Baroody, R. A. Exp. Eye Res. 1987, 44, 217-226.
JS950373B