Comparison of Human Ocular Distribution of Bimatoprost and Latanoprost

Abstract

Purpose:

This study investigated the ocular distribution of bimatoprost, latanoprost, and their acid hydrolysis products in the aqueous humor, cornea, sclera, iris, and ciliary body of patients treated with a single topical dose of 0.03% bimatoprost or 0.005% latanoprost for understanding concentration–activity relationships.

Methods:

Thirty-one patients undergoing enucleation for an intraocular tumor not affecting the anterior part of the globe were randomized to treatment with bimatoprost or latanoprost at 1, 3, 6 or 12 h prior to surgery. Concentrations of bimatoprost, bimatoprost acid, latanoprost, and latanoprost acid in the human aqueous and ocular tissues were measured using liquid chromatography tandem mass spectrometry.

Results:

Following topical administration, intact bimatoprost was distributed in human eyes with a rank order of cornea/sclera >iris/ciliary body >aqueous humor. Bimatoprost acid was also detected in these tissues, where its low levels in the cornea relative to that of latanoprost acid indicated that bimatoprost hydrolysis was limited. Latanoprost behaved as a prodrug that entered eyes predominantly via the corneal route. Levels of latanoprost acid were distributed as cornea >>aqueous humor>iris>sclera>ciliary body.

Conclusions:

Our study provided experimental evidence that levels of bimatoprost in relevant ocular tissues, and not only aqueous humor, are needed to understand the mechanisms by which bimatoprost lowers intraocular pressure (IOP) in human subjects. The data suggest that bimatoprost reached the target tissues favoring the conjunctival/scleral absorption route. Findings of intact bimatoprost in the target ciliary body indicated its direct involvement in reducing IOP. However, bimatoprost acid may have only a limited contribution on the basis that bimatoprost has greater/similar IOP-lowering efficacy than latanoprost, yet bimatoprost acid levels were a fraction of latanoprost acid levels in the aqueous humor and cornea and only sporadically detectable in the ciliary body. In this report, human ocular tissues were examined concurrently with aqueous humor for the in vivo distribution of bimatoprost, bimatoprost acid, latanoprost, and latanoprost acid.

Introduction

Lumigan^® (0.03%, bimatoprost ophthalmic solution) and 0.005% Xalatan^® (latanoprost ophthalmic solution) are efficacious drugs for reducing elevated intraocular pressure (IOP) in patients with ocular hypertension or glaucoma. These agents are given once daily in glaucoma treatment because the IOP-lowering activity persists in the 24 h following dosing.^1–3 Both drugs lower IOP of humans by a mechanism of increasing aqueous humor outflow via the uveoscleral (predominant) and trabecular pathways.^2–5 It has been demonstrated that the ciliary body and trabecular meshwork are relevant target tissues; however, previous in vivo studies of the distribution of these drugs into human ocular tissues and fluids have only focused on sampling aqueous humor.^6–10 This may be due to the difficulty and time-consuming nature of conducting in vivo studies that involve human ocular tissues. Investigators have attempted to correlate the IOP reduction with aqueous humor concentration of intact bimatoprost (17-phenyl trinor prostaglandin F_2α-1-ethylamide) or its acid hydrolysis product 17-phenyl trinor prostaglandin F_2α (bimatoprost acid).^8–10 This approach has limitations, because aqueous humor levels of bimatoprost and bimatoprost acid may not be indicative of their respective levels in the target outflow tissues. Laboratory animal studies show that the rank order of drug concentrations in the ocular tissues differs, depending on the contributions of drug absorption via the corneal and conjunctival/scleral routes.^11,12 Therefore, our study obtained in vivo experimental evidence from relevant human ocular tissues, in addition to aqueous humor, to establish pharmacokinetics/pharmacodynamic correlations and answer the question of whether bimatoprost works directly to lower IOP or behaves as a prodrug like latanoprost (13,14-dihydro-17-phenyl trinor prostaglandin F_2α-1-isopropyl ester).

The aim of the present study was to determine the in vivo distribution of bimatoprost (C1-ethylamide) compared with latanoprost (C1-isopropyl ester) and their C1-acid hydrolysis products in ocular tissues and aqueous humor of subjects given a single topical dose. Our results showed bimatoprost and latanoprost have different drug distribution and metabolism profiles and suggest that bimatoprost as the intact molecule has direct actions on IOP reduction in humans.

Methods

Experimental protocol

This was a randomized, vehicle-controlled, prospective study conducted in compliance with the Declaration of Helsinki. All patients provided written informed consent before study entry. Patients in need of enucleation for eye tumors were enrolled from 2 centers: Wills Eye Institute, Philadelphia, Pennsylvania, and Semmelweis University, Budapest, Hungary. The study was approved by the Institution Review Boards of both centers (IRB 03-599; TUKEB 103/2004). The tumors in enrolled patients involved the posterior segment but not the ciliary body. Patients were excluded if they had tumors involving the ciliary body, iris, anterior chamber, or cornea. Eyes with previous radiation plaque procedures and previous history of antiglaucoma medications were not excluded. None of the patients was treated with prostaglandin analogs.

The sampling “time point” was the time interval between dosing and enucleation. Study eyes were randomized to treatment with a single drop of 0.03% bimatoprost (Lumigan^®, Allergan Inc., Irvine, CA), 0.005% latanoprost (Xalatan^®, Pfizer Inc., New York, NY), or no drops at 1, 3, 6, or 12 h before surgery (according to their designated “time point”). The patient was instructed to self-administer 1 drop from the assigned bottle at 1, 3, 6, or 12 h prior to scheduled enucleation. Two eyes did not receive any drops and served as controls for the assay methodology. In total, 31 eyes from 31 patients were included in this study.

After enucleation, an anterior chamber paracentesis of 0.1 cc (100 μL) of aqueous humor was immediately performed. The globes and aqueous humor samples were sent in a container on ice to the Pathology Service at each study center. At the Pathology Service, a section of the whole anterior chamber was performed, including the critical components of cornea, sclera, iris, and ciliary body. The time duration between enucleation and dissection was about 5–10 min for each eye. To avoid delay in freezing, minimal handling was done on the fresh, unfrozen tissue. The samples were then placed in 1.5-mL chemically resistant conical microcentrifuge tubes (Cole-Palmer, Vernon Hills, IL) to be stored at temperatures down to −70°C. The sample volume comprised of approximately 40–80 mg of ocular tissue and 100 μL of of aqueous humor. The samples were shipped under dry ice (−70°C) to independent laboratories where the ocular tissues were removed from the container, thawed, and weighed. Levels of bimatoprost, latanoprost, and their acid hydrolysis products (17-phenyl trinor PGF_2α and 13,14-dihydro-17-phenyl trinor PGF_2α, respectively) in different parts of the globe were determined using a validated high-performance liquid chromatography–tandem mass spectrometry (LC-MS/MS) method.

Bioanalytical procedures

Assay methodology

Aqueous humor, cornea, sclera, iris, and ciliary body concentrations of bimatoprost, bimatoprost acid metabolite, latanoprost, and latanoprost acid metabolite were determined using tetradeuterated (d₄) internal standards and LC-MS/MS at two contract research laboratories. The human ocular samples were received frozen from the study center and stored at −70°C until analysis at the contract laboratory.

Materials

Bimatoprost and bimatoprost acid reference standards were obtained from Allergan (Irvine, CA). Latanoprost and latanoprost acid reference standards were obtained from Cayman Chemical Company (Ann Arbor, MI). The reference standards for the bimatoprost-d_4, bimatoprost acid-d₄, latanoprost-d₄, and latanoprost acid-d₄ internal standards were obtained from Cayman Chemical Company. Blank rabbit aqueous humor, cornea, sclera, iris, and ciliary body from PelFreeze (Rogers, AK) were used as proxy matrix for the corresponding human ocular tissue or fluid.

Calibration standards and quality control (QC) samples for this study were prepared by spiking blank rabbit ocular matrix with bimatoprost and bimatoprost acid reference standards or latanoprost and latanoprost acid reference standards. Assay specificity was determined by analyzing individual blank rabbit ocular matrices and blank human ocular matrices of human control subjects who did not receive bimatoprost or latanoprost.

Aqueous humor sample preparation

Study samples were prepared by adding internal standard solution (bimatoprost-d₄ and bimatoprost acid-d₄ or latanoprost-d₄ and latanoprost acid-d₄) to 50 μL of aqueous humor samples in silanized glass tubes. The analytes were extracted from aqueous humor under acidic pH with methyl tert-butyl ether (MTBE). The organic layer was evaporated to dryness. Extracts were reconstituted with 100 μL of acetonitrile for analysis of bimatoprost and acid metabolite. Extracts were reconstituted with 100 μL of methanol/water (1:1) with 0.2% acetic acid for analysis of latanoprost and acid metabolite.

Ocular tissue sample preparation

Ocular tissue samples were analyzed by either solvent extraction or tissue homogenization. Cornea, sclera, iris, and ciliary body samples for most of the subjects from both sites were analyzed by a validated solvent extraction method (Analytical Division of Drug Studies Unit, University of California San Francisco, CA). Cornea, sclera, iris, and ciliary body samples from 5 subjects were analyzed by a validated method involving tissue homogenization (Prevalere Life Sciences, Whitesboro, NY). Aqueous humor samples from both sites were analyzed at both analytical contract labs using the same method. All samples were analyzed using the same LC-MS/MS method.

For the validated solvent extraction method, ocular tissues were soaked overnight in 4 mL of 50% methanol/50% water. After centrifugation, internal standards were added to a 1-mL aliquot of the supernatant, and analytes were extracted under acidic conditions with MTBE. For the validated tissue homogenization method, ocular tissues were homogenized in 1 mL of 50% methanol/50% water at 1-min intervals using a Bead Beater, followed by microcentrifugation for 5 min at 13,000 rpm. After addition of internal standards to 250 μL aliquot of the supernatant, the analytes were extracted by liquid–liquid extraction under acidic pH using MTBE. The organic layer was removed, dried, and reconstituted in 100 μ of acetonitrile for bimatoprost and metabolite analysis or 100 μL of methanol/water (1:1) with 0.2% acetic acid for analysis of latanoprost and metabolite.

LC-MS/MS analysis

Aliquots of the reconstituted extract were analyzed by LC-MS/MS using a triple quadrupole mass spectrometer coupled to an electrospray ionization source. Normal-phase (polar organic) high-performance liquid chromatography (HPLC) was performed for separation of bimatoprost and bimatoprost acid on a Thermo Hypersil APS-2 column (2.1×150 mm, 3 μm) using a gradient elution, with 0.5% formic acid in acetonitrile (A) and 0.5% formic acid in methanol (B), at a flow rate of 0.3 mL/min. Quantitation of bimatoprost was performed in positive ion multiple reaction monitoring (MRM) mode, while bimatoprost acid was quantitated in the negative ion MRM mode with positive-to negative switch at about 3.5 min. The specific precursor-product ion pairs used in MRM analysis were: m/z 398.4→362.4 or 416.5→362.4 (bimatoprost); m/z 387.3→193.3 (bimatoprost acid); m/z 402.4→366.4 or 420.5→366.4 (bimatoprost-d₄); and m/z 391.3→197.3 (bimatoprost acid-d₄). The total analysis time was 6.5 min.

LC-MS/MS analysis of latanoprost and latanoprost acid was performed using a triple quadrupole mass spectrometer coupled to an electrospray ionization source. Chromatographic separations of latanoprost and latanoprost acid were achieved using a Zorbax XDB C8 column (2.1×50 mm, 3.5 μm) with gradient elution, with 0.2% acetic acid in water (A) and 0.2% acetic acid in methanol (B), at a flow rate of 0.3 mL/min. Quantitation of latanoprost was done in positive ion MRM mode, while latanoprost acid was quantitated in the negative ion MRM mode with positive-to negative switch at about 4 min. The specific precursor–product ion pairs used in MRM analysis were: m/z 433.5→379.5 (latanoprost); m/z 389.3→345.5 (latanoprost acid); m/z 437.5→383.5 (latanoprost-d₄); and m/z 393.3→349.5 (latanoprost acid-d₄). The total analysis time was 8 min.

Using linear regression analysis, calibration curves were generated from relating peak area ratios of analyte/internal standard to amount or concentration of analyte in matrix. The assay ranges were as follows: 0.2–100 ng/mL for aqueous humor and 0.1–50 ng/tissue for the ocular tissues. Each analytical run for human sample analysis was accepted if at least two-thirds of all QCs (low, medium, and high) showed accuracy within 85%–115% of nominal values. The ocular tissue concentrations (ng/g) for human study samples were calculated for data analysis. Samples that had amounts or concentrations below the lower limit of quantitation are designated as below limit of quantitation (BLQ). The tissue extraction efficiencies were approximately 100% for the four analytes using rabbit iris–ciliary body as proxy matrix. The extraction efficiencies for bimatoprost, bimatoprost acid, latanoprost, and latanoprost in aqueous humor were 60%, 77%, 80%, and 88% using rabbit aqueous humor as proxy matrix. Accurate, reproducible, and reliable assays were employed for the quantitation of bimatoprost, bimatoprost acid, latanoprost, and latanoprost acid concentrations in human aqueous humor and ocular tissues using highly sensitive and selective LC-MS/MS methods.

The area under the concentration–time curve (AUC_0-t) was used for graphical visualization of bimatoprost, bimatoprost acid, and latanoprost acid average concentrations [nM] over the total time intervals (0–12 h, except 0–6 h for latanoprost acid in aqueous humor) in ocular tissues and aqueous humor. AUC was calculated by the Linear-Log Trapezoidal Method, and an average concentration of compound in the tissues and fluid was determined by dividing the total exposure AUC by the total time interval. The mean concentration (C) was calculated for each time interval (t₁−t₂). If C₁<C₂, the linear rule was applied and AUC=(C₁+C₂)/2 * (t₂−t₁). If C₁>C₂, the logarithmic rule was applied and AUC=[C₁−C₂)/(ln(C₁) – ln(C₂)] _* (t₂−t₁). No statistical analysis of the sample data was performed due to limitations of low sample sizes and high variability in some of the measured levels since each sample was obtained from different patients.

Results

Aqueous and tissue samples from all 31 randomized eyes were analyzed for bimatoprost, latanoprost, or their acid metabolites. Out of 31 eyes, 20 eyes received bimatoprost (5 at 1 h, 5 at 3 h, 6 at 6 h, and 4 at 12 h), 9 eyes received latanoprost (3 at 1 h, 2 at 3 h, 2 at 6 h, and 2 at 12 h), and 2 eyes were controls. Mean weight of tissue samples of cornea, sclera, iris, and ciliary body was 0.029 g, 0.051 g, 0.016 g, and 0.043 g, respectively.

Figures 1 and 2 display MRM chromatograms of bimatoprost and bimatoprost acid in aqueous humor and iris–ciliary body calibration standards, respectively. Bimatoprost and bimatoprost-d₄ had retention times of ≈2.7 min, as compared to retention times of ≈4.3 min for bimatoprost acid and bimatoprost acid-d₄. Figures 3 and 4 show MRM chromatograms of latanoprost and latanoprost acid in aqueous humor and iris–ciliary body calibration standards, respectively. The retention time of latanoprost acid and its internal standard was ≈2.6 min, as compared to ≈5.4 min for latanoprost and its deuterated internal standard. No significant interference was observed from analysis of blank rabbit and human ocular matrices, indicating assay selectivity for the analytes and internal standards.

FIG. 1.

Representative chromatogram of bimatoprost and bimatoprost acid in aqueous humor, 0.200 ng/mL calibration standard.

FIG. 2.

Representative chromatogram of bimatoprost and bimatoprost acid in iris–ciliary body, 0.100 ng/tissue calibration standard [lower limit of quantification (LLOQ)].

FIG. 3.

Representative chromatogram of latanoprost acid and latanoprost in aqueous humor, 0.200 ng/mL calibration standard [lower limit of quantification (LLOQ)].

FIG. 4.

Representative chromatogram of latanoprost acid and latanoprost in iris–ciliary body, 0.100 ng/tissue calibration standard [lower limit of quantification (LLOQ)].

The lower limit of quantitation for aqueous humor assay was 0.48 nM for bimatoprost, 0.51 nM for bimatoprost acid, 0.46 nM for latanoprost, and 0.51 nM for latanoprost acid. On the basis of an overall mean tissue weight of 0.03 g and assay lower limit of 0.1 ng for tissues, the lower limit of quantitation for assay of the ocular tissues was 0.80 nM for bimatoprost, 0.86 nM for bimatoprost acid, 0.77 nM for latanoprost, and 0.85 nM for latanoprost acid. Samples with parent or acid metabolite below limit of quantitation were assigned a value of 0 nM. A limitation of quantitation of these drugs and their hydrolysis products is the small sample size and assay sensitivity.

Aqueous samples, taken from the enrolled patients at 1, 3, 6, or 12 h after a single dose of bimatoprost 0.03%, contained low or undetectable levels of bimatoprost with mean levels ranging from 1.5 nM at 1 h to BLQ at 12 h. Mean levels of bimatoprost acid were 4.4 nM at 1 h, 5.4 nM at 3 h, 4.0 nM at 6 h, and 0.4 nM at 12 h. Bimatoprost acid was BLQ in 1 of 6 samples at 6 h and 2 of 3 samples at 12 h (Table 1A).

Table 1.

Bimatoprost and Bimatoprost Acid Levels [nM] in Aqueous Humor after Lumigan ^® Treatment and Latanoprost and Latanoprost Acid levels [nM] in Aqueous Humor after Xalatan ^® Treatment

	1 h (n=5)	3 h (n=5)	6 h (n=5)	12 h (n=3)
A. Bimatoprost
Mean (SD)	1.5 (1.5)	0.9 (1.4)	0.2 (0.4)	0
Individual values	0, 0.6, 1.4, 1.8, 3.9	0, 0, 0, 1.4, 3.1	0, 0, 0, 0, 0.8	0, 0, 0
Bimatoprost acid
Mean (SD)	4.4 (3.0)	5.4 (5.5)	4.0 (4.9)	0.4 (0.7)
Individual values	0.8, 3.0, 3.8, 5.4, 8.8	1.4, 1.5, 3.9, 5.4, 14.8	0, 0.9, 2.0, 5.0, 12.0	0, 0, 1.1

	1 h (n=3)	3 h (n=2)	6 h (n=2)	12 h (n=0)
B. Latanoprost
Mean (SD)	0	0	0	—
Individual values	0, 0, 0	0, 0	0, 0	—
Latanoprost acid
Mean (SD)	18.9 (13.1)	41.1 (53.2)	14.0 (3.0)	—
Individual values	4.0, 24.4, 28.3	3.4, 78.7	11.9, 16.2	—

Samples with below the limit of quantitation (BLQ) of 0.48 nM for bimatoprost and 0.51 nM for bimatoprost acid were assigned a value of 0 nM.

Samples with below the limit of quantitation (BLQ) of 0.46 nM for latanoprost and 0.51 nM for latanoprost acid were assigned a value of 0 nM.

SD, Standard deviation.

None of the aqueous samples from patients treated with topical latanoprost 0.005% contained quantifiable levels of intact latanoprost (Table 1B), consistent with previous findings that latanoprost is rapidly and efficiently hydrolyzed to its acid metabolites.^6–9 On the contrary, 1 h, 3 h, and 6 h aqueous samples from patients treated with topical bimatoprost contained measurable levels of nonmetabolized drug (Table 1A). Latanoprost acid levels were higher than bimatoprost acid levels in the aqueous humor at 1 h, 3 h, and 6 h time points. The maximum mean level of latanoprost acid detected in aqueous (41.1 nM at 3 h) was 7.6 times higher than the maximum mean level of bimatoprost acid detected (5.4 nM at 3 h). The mean aqueous humor level ratios (×100) of bimatoprost acid/latanoprost acid were 23% (4.4 nM/18.9 nM) at 1 h, 13% (5.4 nM/41.1 nM) at 3 h, and 28.5% (4.0 nM/14.0 nM) at 6 h.

The mean concentrations of bimatoprost, bimatoprost acid, and latanoprost acid were highest in cornea at 1 h after a single topical dose (Tables 1 –5). Bimatoprost AUC in the ocular tissues and fluid, averaged over the 0–12 h time period, showed a rank order of ocular distribution of cornea ⩰ sclera>iris ⩰ ciliary body>aqueous humor (Fig. 5A,B). Bimatoprost acid AUC over the same time period showed a different distribution pattern of cornea>sclera>aqueous humor>ciliary body ⩰ iris (Fig. 5A,C). In the iris, bimatoprost was BLQ in 15 of 16 samples. Bimatoprost and its acid metabolite could not be determined in two iris tissue samples due to sample extraction error. In the ciliary body, bimatoprost acid was BLQ in most of the samples and only sporadically detected in 2 of 5 samples at 1 h and 1 of 5 samples at 3 h.

FIG. 5.

Bimatoprost (A,B), bimatoprost acid (A,C), and latanoprost acid (A,D) average concentrations [nM] in cornea, sclera, iris, ciliary body (cil body), and aqueous humor (aq h) of patients treated with a single topical dose of 0.03% bimatoprost or 0.005% latanoprost over the study total time intervals (0–12 h except 0–6 h for latanoprost acid in aq h). Concentrations were analyzed using the linear-log trapezoidal method for area under the concentration–time curve (AUC_0-t).

Table 2.

Bimatoprost and Bimatoprost Acid levels [nM] in Cornea after Lumigan ^® Treatment and Latanoprost and Latanoprost Acid Levels [nM] in Cornea after Xalatan ^® Treatment

	1 h (n=4)	3 h (n=5)	6 h (n=5)	12 h (n=3)
A. Bimatoprost
Mean (SD)	36.2 (29.3)	14.6 (28.1)	0	0
Individual values	0, 25.7, 64.9, 54.3	0, 0, 0, 8.4, 64.3	0, 0, 0, 0, 0	0, 0, 0
Bimatoprost acid
Mean (SD)	51.6 (43.4)	33.3 (38.1)	12.2 (27.2)	0
Individual values	0, 34.6, 72.7, 99.2	0, 12.2, 25.9, 30.1, 98.1	0, 0, 0, 0, 60.8	0, 0, 0

	1 h (n=3)	3 h (n=2)	6 h (n=2)	12 h (n=2)
B. Latanoprost
Mean (SD)	0	0	0	0
Individual values	0, 0, 0	0, 0	0, 0	0, 0
Latanoprost acid
Mean (SD)	290.5 (295.5)	227.3 (272.2)	100.7 (83.7)	0
Individual values	84.0, 158.5, 628.9	34.8, 419.8	41.5, 160.0	0, 0

Samples with below the limit of quantitation (BLQ) of 0.80 nM for bimatoprost and 0.86 nM for bimatoprost acid were assigned a value of 0 nM.

Samples with below the limit of quantitation (BLQ) of 0.77 nM for latanoprost and 0.85 nM for latanoprost acid were assigned a value of 0 nM.

SD, Standard deviation.

Table 3.

Bimatoprost and Bimatoprost Acid Levels [nM] in Sclera after Lumigan ^® Treatment and Latanoprost and Latanoprost Acid levels [nM] in Sclera after Xalatan ^® Treatment

	1 h (n=5)	3 h (n=5)	6 h (n=6)	12 h (n=4)
A. Bimatoprost
Mean (SD)	13.0 (13.0)	20.2 (23.0)	0	14.4 (14.7)
Individual values	0, 0, 13.3, 23.7, 27.8	0, 5.3, 10.7, 28.6, 56.6	0, 0, 0, 0, 0, 0	0, 3.4, 25.7, 28.3
Bimatoprost acid
Mean (SD)	6.1 (6.7)	44.0 (56.1)	5.9 (13.5)	0
Individual values	0, 0, 4.1, 12.2, 14.0	8.0, 15.4, 108.6	0, 0, 0, 0, 2.2, 33.5	0, 0, 0, 0

	1 h (n=2)	3 h (n=0)	6 h (n=1)	12 h (n=2)
B. Latanoprost
Mean (SD)	0	—	0	0
Individual values	0, 0	—	0	0, 0
Latanoprost acid
Mean (SD)	0	38.4	5.3	0
Individual balues	0, 0	38.4	5.3	0, 0

Samples with below the limit of quantitation (BLQ) of 0.80 nM for bimatoprost and 0.86 nM for bimatoprost acid were assigned a value of 0 nM.

Samples with below the limit of quantitation (BLQ) of 0.77 nM for latanoprost and 0.85 nM for latanoprost acid were assigned a value of 0 nM.

SD, Standard deviation.

Table 4.

Bimatoprost and Bimatoprost Acid Levels [nM] in Iris after Lumigan ^® and Latanoprost Treatment and Latanoprost Acid Levels [nM] in Iris after Xalatan ^® Treatment

	1 h (n=5)	3 h (n=4)	6 h (n=4)	12 h (n=3)
A. Bimatoprost
Mean (SD)	13.7 (17.6)	5.7 (11.4)	1.7 (3.4)	0
Individual values	0, 5.8, 7.7, 10.8, 44.5	0, 0, 0, 22.8	0, 0, 0, 6.8	0, 0, 0
Bimatoprost acid
Mean (SD)	0	5.4 (10.7)	0	0
Individual values	0, 0, 0, 0, 0	0, 0, 0, 21.4	0, 0, 0, 0	0, 0, 0

	1 h (n=3)	3 h (n=2)	6 h (n=2)	12 h (n=2)
B. Latanoprost
Mean (SD)	4.9 (8.5)	0	0	0
Individual values	0, 0, 14.7	0, 0	0, 0	0, 0
Latanoprost acid
Mean (SD)	24.3 (42.1)	30.6 (43.2)	4.9 (6.9)	0
Individual values	0, 0, 73.0	0, 61.1	0, 9.8	0, 0

Samples with below the limit of quantitation (BLQ) of 0.80 nM for bimatoprost and 0.86 nM for bimatoprost acid were assigned a value of 0 nM.

Samples with below the limit of quantitation (BLQ) of 0.77 nM for latanoprost and 0.85 nM for latanoprost acid were assigned a value of 0 nM.

SD, Standard deviation.

Table 5.

Bimatoprost and Bimatoprost Acid Levels [nM] in Ciliary Body after Lumigan ^® Treatment and Latanoprost and Latanoprost Acid Levels [nM] in Ciliary Body after Xalatan ^® Treatment

	1 h (n=5)	3 h (n=5)	6 h (n=6)	12 h (n=4)
A. Bimatoprost
Mean (SD)	8.1 (13.1)	3.6 (5.1)	1.2 (3.0)	0
Individual values	0, 0, 3.4, 5.9, 31.1	0, 0, 0.8, 5.4, 11.7	0, 0, 0, 0, 0, 7.3	0, 0, 0, 0
Bimatoprost acid
Mean (SD)	3.5 (6.7)	6.2 (13.9)	0	0
Individual values	0, 0, 0, 2.3, 15.4	0, 0, 0, 0, 31.0	0, 0, 0, 0, 0	0, 0, 0

	1 h (n=3)	3 h (n=2)	6 h (n=2)	12 h (n=2)
B. Latanoprost
Mean (SD)	0	0	0	0
Individual values	0, 0, 0	0, 0	0, 0	0, 0
B. Latanoprost
Mean (SD)	3.1 (5.3)	10.8 (15.3)	0	0
Individual values	0, 0, 9.2	0, 21.6	0, 0	0, 0

Samples with below the limit of quantitation (BLQ) of 0.80 nM for bimatoprost and 0.86 nM for bimatoprost acid were assigned a value of 0 nM.

Samples with below the limit of quantitation (BLQ) of 0.77 nM for latanoprost and 0.85 nM for latanoprost acid were assigned a value of 0 nM.

SD, Standard deviation.

Intact latanoprost was detected only in one iris sample at 1 h, whereas it was BLQ in all the samples of aqueous humor, cornea, sclera, and ciliary body (Tables 1B –5B). Latanoprost acid AUC, averaged over the 0–12 h (tissues) or 0–6 h (aqueous humor) time periods, showed a distribution pattern of cornea>> aqueous humor>iris>sclera>ciliary body (Fig. 5A,D). In cornea, latanoprost acid levels were higher than bimatoprost acid levels at the 1 h, 3 h, and 6 h time points. The maximum mean level of latanoprost acid detected in cornea (290.5 nM at 1 h) was 5.6 times higher than the maximum mean level of bimatoprost acid detected (51.6 nM at 1 h). The mean cornea level ratios (×100) of bimatoprost acid/latanoprost acid were 18% (51.6 nM/290.5 nM) at 1 h, 15% (33.3 nM/227.3 nM) at 3 h, and 12% (12.2 nM/100.7 nM) at 6 h. A limitation of quantifying latanoprost and latanoprost acid in this study was the low sample sizes for aqueous humor and the ocular tissues and the assay sensitivity. In addition, each sample was obtained from different patients which resulted in high variability in the measured latanoprost acid levels.

Discussion

The pharmacological activity of lowering IOP is dependent upon drug absorption into intraocular tissues, distribution to the target tissues, potency, and sufficient concentration at the active sites of ciliary body and/or trabecular region of the anterior chamber. For topically applied drugs to reach the intraocular tissues, the drug first must be absorbed or penetrate across the ocular tissue barriers of cornea, conjunctiva, and sclera. The roles of the two main absorption pathways, cornea and conjunctiva/sclera, depend mainly on a drug's physiochemical properties of lipophilicity and molecular size.^13,14 Another factor is the ionic state of drugs, because their ionization results in decreased lipophilicity relative to the neutral state. For comparison of latanoprost (C1-isopropyl ester) and bimatoprost, the molecular size and ionic state may be removed from consideration because both drugs are small molecules of similar size and neutral compounds (partition coefficient Log P=distribution coefficient Log D; calculated using ACD/Labs' ACD/PhysChem Suite software). Thus, it is the lipophilic properties of these drugs that become important for absorption into the eye after topical application.

Our results suggest that the corneal route is a major pathway of ocular penetration in humans for latanoprost. Drug absorption via the corneal route is characterized by a rank order of drug concentration of cornea (with some lateral diffusion to sclera) >aqueous humor>iris/ciliary body, between which lateral diffusion also occurs.^11,12 The corneal route favors the more lipophilic latanoprost (Log D=3.65 at pH 7.4), because a general trend is that corneal permeability increases with increased lipophilicity.¹³ Among the cornea tissue layers, the epithelium is a lipophilic tissue and the primary barrier to permeability of hydrophilic drugs.¹³ Lipophilic prostaglandin prodrugs such as latanoprost can readily penetrate the corneal epithelium, where it is mainly hydrolyzed by esterases to the more hydrophilic latanoprost acid.¹⁵ In the present study, latanoprost was rapidly and almost completely hydrolyzed by the 1 h time point to latanoprost acid, its pharmacologically active free acid metabolite. The high levels of latanoprost acid in the cornea suggest that the corneal stroma, which does not have dependency on lipophilicity,¹³ served as a temporary depot for latanoprost acid before it entered the aqueous humor. Regardless of our low sample sizes, this hydrolysis profile and levels of latanoprost acid in the aqueous humor are consistent with results of previous studies in humans given topical latanoprost.^6,7,9,10,16

Compared to latanoprost, bimatoprost has decreased lipophilicity (Log D=1.98 at pH 7.4) and would likely be more dependent on the conjunctival/scleral penetration pathway for absorption, because these tissues have no clear dependency on lipophilicity or distribution coefficient.¹³ The finding that bimatoprost had a 4.5 times higher rate of penetration in the sclera than in the cornea of human isolated tissues, suggested that it favored the scleral route.¹⁷ The rank order of drug concentration for the conjunctival/scleral route is generally cornea/sclera and conjunctiva>iris/ciliary body, between which lateral diffusion also occurs>aqueous humor.^11,12 Low levels of both bimatoprost and bimatoprost acid were detected in the aqueous humor of patients in the present study. Thus, the corneal epithelium appeared to be more of a barrier to bimatoprost's penetration because it is less lipophilic than latanoprost. In addition, the presence of an ethylamide moiety instead of isopropyl ester at position C1 limited the hydrolysis of bimatoprost to bimatoprost acid. Our results are consistent with studies in living non-human primates in which topically applied bimatoprost 0.1% was rapidly distributed in eyes with levels in iris/ciliary body>> aqueous humor.¹⁸ This ocular pathway indicated that the levels of bimatoprost in the aqueous humor do not entirely reflect its levels at the active site, as bimatoprost reaches the ciliary body mostly through the conjunctiva and sclera. Importantly, the mechanisms by which bimatoprost lowers IOP cannot be adequately addressed by using its aqueous humor levels.

A prodrug is commonly defined as a drug that is inactive or weakly active when administered and then subsequently metabolized in vivo into an active metabolite. Latanoprost is a prodrug that enters the eye through the cornea, where it is rapidly and completely hydrolyzed by esterases to latanoprost acid, its acid hydrolysis product.⁶ Latanoprost acid is a potent prostanoid FP receptor agonist that is responsible for the IOP-lowering activity of latanoprost. Our results are consistent with latanoprost behaving as a prodrug that is rapidly and efficiently metabolized in human cornea. We found high concentrations of latanoprost acid in the cornea that were 290 nM, 227 nM, and 101 nM at 1 h, 3h, and 6 h postdose, respectively. These levels support the ocular pharmacokinetics of latanoprost, where the cornea appears to act as a slow-release repository for latanoprost acid in the aqueous humor.⁶ For prodrugs like latanoprost, the levels of latanoprost acid in the aqueous humor, which exits the eye via the outflow pathways, may be a useful indicator for the IOP-lowering effects.

By contrast, bimatoprost does not have characteristics of a prodrug as defined herein, because it has potent intrinsic activity in prostanoid test systems of well-coupled isolated smooth muscle preparations of cat and rabbit.^17,19,20 Bimatoprost also exhibited intrinsic activity that was separate from FP receptor stimulation in human cultured T cells.²¹ In a recent study, bimatoprost produced immediate and concentration-dependent relaxations of human cultured primary ocular cells of ciliary smooth muscle [potency half-maximal effective concentration (EC₅₀)=1.7 nM], trabecular meshwork (EC₅₀=4.3 nM), and Schlemm's canal (EC₅₀=1.2 nM) using an impedance assay.²² In these ocular cells, the selective prostanoid FP receptor agonists, fluprostenol and 17-phenyl prostaglandin F_2α (bimatoprost acid), had EC₅₀ values that ranged from 1.2 nM to 51 nM.²² The data in our study show intact bimatoprost in the target ciliary body at levels [mean±standard error of the mean (SEM)] of 8.1±5.9 nM at 1 h, 3.6±2.3 nM at 3 h, and 1.2±1.2 nM at 6 h, which implicated its involvement in IOP reduction in humans. A 2008 study found bimatoprost IOP lowering may potentially involve the FP-altFP receptor heterodimer.²³ In addition, a selective prostamide antagonist was shown to block ocular hypotension induced by bimatoprost but not those of latanoprost in dogs.²⁴

Unlike bimatoprost, the levels of bimatoprost acid were distributed as cornea>sclera>aqueous humor>iris/ciliary body. Bimatoprost appeared to undergo limited hydrolysis to bimatoprost acid in the human eyes of our study, based on comparison to latanoprost acid levels in the cornea (high) and aqueous humor. This hydrolysis activity may occur in the cornea, sclera, and possibly iris/ciliary body tissues, as previously reported in human ocular tissue in vitro studies.^25,26 Bimatoprost acid contributions to the reductions in IOP could be generally estimated by using the levels of latanoprost acid as a comparator. On the basis that the IOP-lowering efficacy of bimatoprost is greater than or similar to that of latanoprost in numerous clinical studies,^16,27 bimatoprost acid can account for only a fraction of the IOP-lowering effects of bimatoprost according to ocular levels of the corneal pathway. In aqueous humor, the levels of bimatoprost acid were 23% that of latanoprost acid at 1 h, 13% at 3 h, and 28.5% at 6 h. Similarly, in cornea the levels of bimatoprost acid were 18% that of latanoprost acid at 1 h, 15% at 3 h, and 12% at 6 h.

We were interested in observing whether the pharmacological effect in lowering IOP correlates temporally with the drug concentration at the active sites and that the levels are sufficiently high, on the basis of its potency at prostanoid FP receptors, to activate the cellular target. Agonist functional potency values of latanoprost acid and bimatoprost acid at endogenous prostanoid FP receptors in human cultured ocular cells of aqueous humor outflow pathways have been collated by investigators. The EC₅₀ values for latanoprost acid are 35 nM (trabecular meshwork cells) and 124 nM, 198 nM (ciliary muscle cells).^9,28 In our study, latanoprost acid levels were maximum 41 nM (n=2) in the aqueous humor at 3 h following topical administration. Its levels at 6 h, a time associated with maximal ocular hypotensive activity, was 14 nM (n=2) in the aqueous humor.

Thus, the ocular levels of latanoprost acid in the aqueous humor reached its agonist potency EC₅₀ values at prostanoid FP receptors in the trabecular meshwork cells (35 nM), and this was likely due to corneal depot. Bimatoprost acid had EC₅₀ values of 26 nM, 51 nM, 112 nM (trabecular meshwork cells), and 2.8 nM, 4 nM (ciliary muscle cells).^9,22,28 For bimatoprost acid, its maximum level of 5.4±2.5 nM (mean±SEM, n=5) in the aqueous humor at 3 h postdose was below its potency EC₅₀ values (26–112 nM) in the trabecular meshwork cells.

In the ciliary body, bimatoprost acid was below the limit of quantitation in a majority of the samples and only sporadically detected in 3 of 18 samples over the 12 h study. These bimatoprost acid levels do not appear to relate temporally with the effects of bimatoprost on lowering IOP in humans that persisted for 24 h and longer.^1,2 Assuming that bimatoprost acid does contribute to the IOP-lowering effects of bimatoprost, then perhaps its ocular levels are sufficient but lower than its potency values and the detection limits of our assay. In this case, prostaglandin transporters may be possibly involved in ocular hypotension.²⁹ The results of our human ocular distribution study, placed in the context of potency and time distribution, indicated that the mode of IOP reductions by bimatoprost and latanoprost are complex.

In summary, latanoprost and bimatoprost were shown to exhibit differences in human ocular distribution and metabolism. Latanoprost acted as a prodrug and was rapidly and efficiently metabolized predominantly in the cornea to latanoprost acid. This single-dose, in vivo tissue distribution study provided experimental pharmacokinetics evidence that intact bimatoprost reached the target ciliary muscle, favoring the conjunctival/scleral absorption route and in sufficient concentration to account for IOP reductions in human subjects. Our results demonstrated the need to measure levels of bimatoprost and other drugs in relevant ocular tissues, and not just aqueous humor, to establish pharmacokinetics/pharmacodynamic correlations that further advance our understanding of the IOP lowering mechanisms.

Footnotes

Acknowledgment

This study was supported by an unrestricted research grant from Allergan. A. Acheampong, J. Chen, and L.A. Wheeler are employees of Allergan.

Author Disclosure Statement

No competing financial interests exist.

References

Frenkel

R.E.

, Noecker

R.J.

, Craven

E.R.

Evaluation of circadian control of intraocular pressure after a single drop of bimatoprost 0.03% or travoprost 0.004%

Curr. Med. Res. Opin., 24:919–923. 2008.

Brubaker

R.F.

Mechanism of action of bimatoprost (Lumigan)

Surv. Ophthalmol., 45,Suppl 4:S347–S351. 2001.

Ziai

, Dolan

J.W.

, Kacere

R.D.

, Brubaker

R.F

. The effects on aqueous dynamics of PhXA41, a new prostaglandin F_2α analogue, after topical application in normal and ocular hypertensive human eyes. Arch. Ophthalmol., 111:1351–1358. 1993.

Christiansen

G.A.

, Nau

C.B.

, McLaren

J.W.

, Johnson

D.H.

Mechanism of ocular hypotensive action of bimatoprost (Lumigan) in patients with ocular hypertension or glaucoma. Ophthalmology, 111:1658–1662. 2004.

Lim

K.S.

, Nau

C.B.

, O'Byrne

M.M.

et al. Mechanism of action of bimatoprost, latanoprost, and travoprost in healthy subjects. A crossover study. Ophthalmology, 2008; 115:790–795.e4.

Sjöquist

, Stjernschantz

Ocular and systemic pharmacokinetics of latanoprost in humans. Surv. Ophthalmol., 47,Suppl 1:S6–S12. 2002.

Calissendorff

, Sjöquist

, Högberg

, Grunge-Lowerud

Bioavailability in the human eye of a fixed combination of latanoprost and timolol compared to monotherapy. J Ocul Pharmacol Ther., 18:127–131. 2002.

Camras

C.B.

, Toris

C.B.

, Sjoquist

et al. Detection of the free acid of bimatoprost in aqueous humor samples from human eyes treated with bimatoprost before cataract surgery. Ophthalmology, 111:2193–2198. 2004.

Cantor

L.B.

, Hoop

, Wudunn

et al. Levels of bimatoprost acid in the aqueous humour after bimatoprost treatment of patients with cataract. Br. J. Ophthalmol., 91:629–632. 2007.

10.

Faulkner

, Sharif

N.A.

, Orr

et al. Aqueous humor concentrations of bimatoprost free acid, bimatoprost and travoprost free acid in cataract surgical patients administered multiple topical ocular doses of LUMIGAN or TRAVATAN. J. Ocul. Pharmacol. Ther., 26:147–156. 2010.

11.

Chien

D.S.

, Homsy

J.J.

, Gluchowski

, Tang-Liu

D.D.

Corneal and conjunctival/sclera penetration of p-aminoclonidine, AGN 190342, and clonidine in rabbit eyes. Curr. Eye Res., 9:1051–1059. 1990.

12.

Schoenwald

R.D.

, Deshpande

G.S.

, Rethwisch

D.G.

, Barfknecht

C.F.

Penetration into the anterior chamber via the conjunctival/sclera pathway. J. Ocul. Pharmacol. Ther., 13:41–59. 1997.

13.

Lawrence

M.S.

, Miller

J.W.

Ocular tissue permeabilities. Int. Ophthalmol. Clin., 44:53–61. 2004.

14.

Bhal

S.K.

, Kassam

, Peirson

I.G.

, Pearl

G.M.

The rule of five revisited: Applying Log D in place of Log P in drug-likeness filters. Mol. Pharm., 4:556–560. 2007.

15.

Bito

L.A.

, Baroody

R.A.

The ocular pharmacokinetics of eicosanoids and their derivatives. 1. Comparison of ocular eicosanoid penetration and distribution following the topical application of PGF_2α, PGF_2α-1-methyl ester, and PGF_2α-1-isopropyl ester. Exp. Eye Res., 44:217–226. 1987.

16.

Bean

G.W.

, Camras

C.B.

Commercially available prostaglandin analogs for the reduction of intraocular pressure: Similarities and differences. Surv. Ophthalmol., 53,Suppl 1:S69–S84. 2008.

17.

Woodward

D.F.

, Krauss

A.H.

, Chen

et al.

The pharmacology of bimatoprost (Lumigan)

Surv. Ophthalmol., 45,Suppl 4:S337–S345. 2001.

18.

Woodward

D.F.

, Phelps

R.L.

, Krauss

A.H.

et al. Bimatoprost: A novel antiglaucoma agent. Cardiovasc. Drug Rev., 22:103–120. 2004.

19.

Woodward

D.F.

, Krauss

A.H.

, Chen

et al.

Pharmacological characterization of a novel antiglaucoma agent, Bimatoprost (AGN 192024)

J. Pharmacol. Exp. Ther., 305:772–785. 2003.

20.

Chen

, Senior

, Marshall

et al. Studies using isolated uterine and other preparations show bimatoprost and prostanoid FP agonists have different activity profiles. Br. J. Pharmacol., 144:493–501. 2005.

21.

Chen

, Lu

R.T.

, Lai

et al. Bimatoprost-induced calcium signaling in human T-cells does not involve prostanoid FP or TP receptors. Curr. Eye Res., 34:184–195. 2009.

22.

Stamer

W.D.

, Piwnica

, Jolas

et al. Cellular basis for bimatoprost effects on human conventional outflow. Invest. Ophthalmol. Vis. Sci., 51:5176–5181. 2010.

23.

Liang

, Woodward

D.F.

, Guzman

V.M.

et al. Identification and pharmacological characterization of the prostaglandin FP receptor and FP receptor variant complexes. Br. J. Pharmacol., 154:1079–1093. 2008.

24.

Jones

R.L.

, Giembycz

M.A.

, Woodward

D.F.

Prostanoid receptor antagonists: development strategies and therapeutic applications. Br. J. Pharmacol., 158,1:104–145. 2009.

25.

Davies

S.S.

, Ju

W.K.

, Neufeld

A.H.

, Abran

, Chemtob

, Roberts

L.J.

2nd . Hydrolysis of bimatoprost (Lumigan) to its free acid by ocular tissue in vitro. J. Ocul. Pharmacol. Ther., 19:45–54. 2003.

26.

Hellberg

M.R.

, Ke

T.L.

, Haggard

et al. The hydrolysis of the prostaglandin analog prodrug bimatoprost to 17-phenyl-trinor PGF_2α by human and rabbit ocular tissue. J. Ocul. Pharmacol. Ther., 19:97–103. 2003.

27.

Holmstrom

, Buchholz

, Walt

, Wickstrøm

, Aagren

Analytic review of bimatoprost, latanoprost and travoprost in primary open angle glaucoma. Curr. Med. Res. Opin., 21:1875–1883. 2005.

28.

Sharif

N.A.

, Klimko

Update and commentary on the pro-drug bimatoprost and a putative ‘prostamide receptor’

Expert Rev. Ophthalmol., 4:477–489. 2009.

29.

Schuster

V.L.

, Itoh

, Andrews

S.W.

et al. Synthetic modification of prostaglandin F_2α indicates different structural determinants for binding to the prostaglandin F receptor versus the prostaglandin transporter. Mol. Pharmacol., 58:1511–1516. 2000.