- Email: [email protected]
Drug delivery through the sclera: effects of thickness, hydration, and sustained release systems
Experimental Eye Research 78 (2004) 599–607 www.elsevier.com/locate/yexer
Drug delivery through the sclera: effects of thickness, hydration, and sustained release systems Sang-Bumm Leea,c, Dayle H. Geroskia, Mark R. Prausnitzb, Henry F. Edelhausera,* b
a Emory Eye Center, Emory University, Suite B2600, 1365B Clifton Road, NE, Atlanta, GA, USA Schools of Chemical Engineering and Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, USA c Department of Ophthalmology, Yeungnam University Medical Center, Taegu, South Korea
Received 25 June 2003; accepted 26 June 2003
Abstract The purpose of this study was to determine whether trans-scleral pressure affects scleral solute permeability by altering scleral thickness or hydration, and to investigate the sustained release delivery of dexamethasone. Scleral sections from donor human globes were mounted for in vitro flux studies. Scleral thickness and hydration were measured as functions of trans-scleral pressure. For the sustained release studies, 3Hdexamethasone in pluronic F-127 gel or in fibrin sealant was added to the episcleral side of the tissue and flux studies were performed. While scleral thickness showed a tendency to decrease with increasing pressure, a significant decrease in thickness was measured only at a transscleral pressure of 60 mmHg. No significant changes in scleral hydration were measured over the range of trans-scleral pressures studied. The apparent permeability constants ðKtrans Þ of human sclera for 3H-dexamethasone in BSS plus, fibrin sealant and F-127 gel were 11·5 £ 1026, 7·3 £ 1026, and 1·5 £ 1026 cm sec21, respectively. Human scleral permeability to dexamethasone differed significantly among the three vehicles ðp , 0·0001Þ: Cumulative delivery of dexamethasone from BSS plus, F-127 gel, and fibrin sealant were 85·0, 29·3, and 67·9% at 20 hr, respectively. Scleral hydration was unaffected by trans-scleral pressures. Scleral thinning was only observed at 60 mmHg. Transscleral pressures below 60 mmHg would not be expected to significantly affect the permeability of the tissue to solutes in the size range of conventional drugs. F-127 gel and fibrin sealant provided a slow, relatively uniform sustained release through a 24 hr period. These systems might be employed to achieve sustained therapeutic levels of drugs to the posterior segment of eye. q 2003 Elsevier Ltd. All rights reserved. Keywords: scleral drug delivery; scleral permeability; scleral thickness; hydration; sustained release; fibrin sealant; pluronic F-127 gel
1. Introduction In David Maurice’s insight and wisdom he was the first person to address the sclera as a route of drug penetration to the posterior segment. His seminal paper dealt with the bovine scleral permeability (Maurice and Polgar, 1977) and this basic drug permeability information remained dormant for a number of years until a major interest developed treating retinal diseases with therapeutic drugs, growth factors proteins and oligonucleotides. In David’s last published paper he addresses the issue of drug delivery to the Posterior Segment from Drops (Maurice, 2002). He stated that ‘he found it difficult to conceive that any * Corresponding author. Dr Henry F. Edelhauser, Emory Eye Center, Emory University, Suite B2600, 1365B Clifton Road, NE, Atlanta, GA, USA. E-mail address: [email protected] (H.F. Edelhauser). 0014-4835/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. DOI:10.1016/S0014-4835(03)00211-2
biological purpose could be served by a system that directs a solute from the tear film, and presents to the retina at about one-millionth of its original concentration’. Therefore there must be another route for a drug to obtain entry to the eye, and he reviewed the importance of the conjunctival cul-desac route to the posterior sclera and an increased residence time. Our laboratory has expanded on David’s initial scleral permeability studies and this paper addresses some of the issues of trans-scleral drug delivery. Topical and systemic treatment of both the anterior and posterior segments of the eye are often difficult because of low ocular tissue permeabilities, diffusion of tear fluid away from the cornea, unwanted counterdirectional convection of drugs, and long diffusional path lengths (Chrai et al., 1973; Lang, 1995; Watsky et al., 1995; Schoenwald et al., 1997). Drugs that are topically applied are thus poorly absorbed,
600
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
with typically less than 5% of the drug reaching anterior intraocular tissues (Lang, 1995). Although systemic administration can deliver drugs to the posterior eye, the large systemic doses necessary are often associated with side effects (Rudnick et al., 1999). Intravitreal injection also delivers drug directly into the posterior chamber. It is, however, an invasive technique and thus has inherent risks for ocular infections and tissue damage (Freeman, 1989). Delivering drugs across the sclera offers another vector to obtain therapeutic vitreous and retinal drug concentrations (Chrai et al., 1973, Maurice and Polgar, 1977; Olsen et al., 1995; Barza et al., 1977, 1978; Unlu and Robinson, 1998). This approach is less invasive and thus should be safer than intravitreal injection. Trans-scleral delivery could take advantage of the large surface area of the sclera and would also offer the potential for localized, sustained release delivery. The relatively high scleral permeability (Ha¨ma¨la¨inen et al., 1977; Ahmed and Patton, 1985; Edelhauser and Maren, 1988; Schoenwald et al., 1997), as compared to the cornea, has motivated research in trans-scleral drug delivery, especially for compounds that need to be administered to the posterior part of the eye (Olsen et al., 1998; Edwards and Prausnitz, 1998), and since the human sclera has a large surface area compared with the corneal surface area – it comprises 95% of the surface area of the eye (Olsen et al., 1998) it provides a promising alternative as a drug delivery route to the posterior segment. Once the scleral permeability is known for a given drug, new delivery systems can be evaluated for sustained release. A variety of sustained release drug delivery systems exists, including gels, erodible polymers, microspheres, liposomes, inserts, miniosmotic pumps and combinations of these technologies. In recent years, pluronic F-127 and fibrin sealant have been widely used in medical and pharmaceutical systems (Miyazaki et al., 1984; Yu et al., 1996). These compounds have good tissue compatibility and show exciting promise for trans-scleral application. An approach to trans-scleral sustained release drug delivery could be based on the injection of a small, concentrated depot into a solid or semisolid sustained release delivery system, such as fibrin sealant or pluronic F-127, located on or within the sclera. For this study, we used a chamber that emulates depot delivery from the scleral surface (Rudnick et al., 1999). The receiver chamber, representing the uveal tissues, is perfused at a slow rate. The donor chamber is held static, similar to a drug depot added to Tenon’s space and directly exposed to the sclera. We studied the potential effects of trans-scleral pressure on the solute permeability of the human sclera. Additionally, we studied two sustained release systems (pluronic F-127 gel and fibrin sealant). The diffusion of dexamethasone across human sclera was measured following a depot application with both sustained release systems to the episclera.
2. Materials and methods Human scleral tissue was obtained from 43 human donor eyes (Georgia Eye Bank, Atlanta, GA, USA) that had been stored in moist chambers for an average (^ S.D. ) of 5·0 ^ 1·2 days. Mean age (^ S.D. ) at time of death was 50·2 ^ 18·2 years. All scleral tissue was taken from adult eyes. Past studies have shown that human scleral water content does not change significantly following moist chamber storage for up to 10 days and that BSS Plus (Alcon Laboratories, Ft. Worth, TX, USA) will maintain tissue hydration during in vitro diffusion studies (Olsen et al., 1995). A razor blade was used to create a full-thickness incision through the sclera near the limbus. Curved scissors were then used to remove a circular piece of sclera. Adherent tissue associated with the retina, choroid, episclera, or Tenon’s capsule was gently removed with a cotton-tip applicator to isolate the bare sclera. Full-thickness scleral disks of 15– 20 mm in diameter were excised from the superior temporal quadrant of the globe, from just posterior to the limbus to 12 – 15 mm posterior to the limbus. This area avoids the anterior ciliary perforating vessels, the short posterior ciliary perforating vessels, and the vortex veins. For the permeability studies, the excised sclera was mounted vitreous side down in a specially designed lucite perfusion chamber in which the sclera is mounted horizontally (Fig. 1). The sclera was clamped between two 2·5 mm wide and 1 mm thick cylindrical Sylgard rings (Dow Corning, Inc., Midland, MI, USA) to prevent lateral leakage and scleral edge damage. The donor (upper) chamber, which serves as a depot, holds 100 –200 ml of a drug in solution or gel. The receiver (lower) chamber, which is a flow-through chamber, holds 500 ml and is continually stirred with a magnetic bar. BSS Plus, containing 1% antibiotic-antimycotic solution (Sigma, St Louis, MO, USA) was perfused through the receiver chamber at a rate of 0·03 ml min21 for experiments less than 8 hr in duration, and at a rate of 0·01 ml min21 for longer experiments. A 7 mm diameter perfusion chamber was used for all experiments. Additional experiments using a 10 mm diameter perfusion chamber were performed for the scleral thickness studies. The larger diameter chamber facilitated the placement of the ultrasonic pachymeter probe on the exposed scleral surface. The tissue was perfused for 15 –30 min to verify that no leaks were present before adding a test compound to the donor chamber. Experiments where leakage occurred were not included in the results. Except as noted, experiments were performed with 15 mmHg applied to the vitreous side of the sclera. Pressure was determined by measuring the vertical distance between the tissue and the outflow tube as it flowed into the collector receptacle (e.g. 15 mmHg is equivalent to a 22 cm water column). The pressure across the tissue was verified using a Statham pressure transducer (Oxnard, CA, USA) connected to the receiver chamber. The temperature of the water-jacketed perfusion chamber was maintained at 37 8C by a circulating water bath.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
601
Fig. 1. Schematic diagram of the perfusion chamber used for measuring scleral permeability under simulated intraocular pressure. Sclera is mounted in a horizontal perfusion setup (episcleral side up) to allow for an intraocular pressure to be simulated by creating a positive water column with the outflow tube. The compound to be tested is added to the donor (upper), episcleral hemichamber. The receiver (lower), uveal hemichamber has a continual flow and is mixed by a magnetic stir bar with the perfusion chamber placed on a magnetic stir plate.
2.1. Scleral thickness and hydration as functions of pressure
2.2. Sustained release permeability studies
These experiments were performed to determine whether trans-scleral pressure affected either the thickness or the hydration of the tissue, which in turn might alter the tissue’s permeability. Scleral tissue was mounted in the perfusion chamber as previously described. Initial (control) scleral thickness was measured using a DGH 500 Pachette ultrasonic pachymeter (DGH Technology, Frazer, PA, USA) just after tissue mounting and prior to any pressure application. The pressure was increased to 15 mmHg and scleral thickness was measured at 15 min intervals. After two hours, the pressure was raised to 30 mmHg and the thickness measurements were continued at 15 min intervals for 2 hr. Pressure was raised once more to 60 mmHg and scleral thickness was measured for the final two hours of the experiment. At each 15 min interval, five scleral thickness readings were taken and the readings averaged to establish scleral thickness at each time point. For the hydration studies, the sclera was mounted in the perfusion system and a constant pressure was maintained for 4 hr. The tissue was then removed, blotted dry and weighed. The sclera was then dried to constant weight at 120 8C for 24 hr. The dried tissue was immediately placed in a tissue desiccator and allowed to cool for 30 min. Samples were re-weighed. The hydration of each piece of sclera was then calculated as: mg H2O mg21 wet tissue (%) [(wet weight 2 dry weight)/wet weight £ 100%]. Five individual experiments were performed at each transscleral pressure (15, 30, and 60 mmHg). To determine control hydration, an additional piece of fresh scleral tissue was excised in each experiment and hydration was measured in each of these unperfused (control) specimens.
For the preparation of fibrin sealant (Tisseelw VH Kit 2·0 ml: Baxter Healthcare Corporation, Glendale, CA, USA) formulations, 80 mCi in 80 ml of 3H-dexamethasone was mixed into 2 ml CaCl2 solution (40 mmol ml21). The 3 H-dexamethasone and CaCl2 solution was injected into the vial containing the freeze-dried human thrombin (500 I.U. ml21). The thrombin solution thus prepared was kept at 37 8C until used. Two milliliters of fibrinolysis inhibitor solution (bovine, 3000 KIU of aprotinin ml21) was heated at 37 8C for 10 min and transferred into the vial containing the freeze-dried human sealer protein concentrate (100 –130 mg ml21 of total protein, 75 –115 mg ml21 of fibrinogen). The resulting thrombin solution and sealer protein solution were then combined to form the fibrin sealant by delivering equal volumes of each with a Duplojectw syringe. For the preparation of the F-127 gel formulations, 25 g of F-127 poloxamer was dissolved into 75 ml of chilled BSS plus (5 – 10 8C). The volume was brought to 100 ml by adding chilled BSS plus with constant mixing. The solution was left overnight in the refrigerator to ensure complete dissolution. The clear, viscous solution thus prepared remains fluid at refrigerator temperature (4 –5 8C). Fifty microcurie in 50 ml of 3H-dexamethasone was added to 450 ml of the F-127 solution. The dexamethasone in fibrin sealant or in the F-127 gel (100 ml total volume) was added to the episcleral surface of the sclera 15– 30 min after the sclera had been mounted in the perfusion chamber. The donor hemichamber containing the test compound was sealed with parafilm and silicone grease (Dow Corning) along the edges of the exposed area of the chamber to prevent evaporation. This provided a flexible seal that did not alter trans-scleral
602
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
pressure. At the beginning and end of each experiment, 10 ml aliquots of F-127 and a piece of fibrin sealant were taken from the episcleral side to calculate 3H-dexamethasone concentration in the sustained release systems. Permeability to two sustained release dexamethasone formulations were evaluated: 3H-dexamethasone-acetate in fibrin sealant, and 3H-dexamethasone-acetate in pluronic F127 gel. The ‘control’ for these sustained release studies was 3 H-dexamethasone-acetate in BSS plus (10 mCi in 100 ml). Samples of the perfusate were collected by a fraction collector (Isco, Lincoln, NE) at 30 – 60 min intervals. At the completion of the perfusion, 100 ml of each fraction was added to 10 ml of Aquasol (Packard, Meriden, CT, USA), and tritium disintegrations were measured using a Beckman LS 5801 liquid scintillation spectrometer (Irvine, CA, USA). Disintegrations per minute (DPMs) were calculated based on quenched standards of 3H. For the 3H-dexamethasone-acetate in BSS plus experiments, samples (10 ml) were taken from the donor chamber at the beginning and end of the experiment to verify the initial donor concentration and to measure donor drug depletion over the course of the experiment. For the fibrin sealant experiments, small pieces of the glue were taken from donor chamber at the beginning and end of each experiment. The pieces were weighed and then were solubilized in 2 ml of Solvable (Packard). After the glue dissolved, 10 ml of UltimaGold scintillation cocktail (Packard) was added to each sample and tritium was measured using scintillation spectroscopy. The concentration (dpm ml21) of 3H-dexamethasone in the initial and final samples of fibrin sealant from the donor chamber were calculated from the density of fibrin sealant (770 mg ml21 as determined by gravimetric analysis) and sample weight. For the F-127 experiments, 10 ml of pre- and post-perfusion donor chamber gel samples were taken in the cold room at 4 8C. The steady-state permeability constant ðKtrans Þ was calculated as Ktrans ¼
Rtotal 1 ; ðtÞðAÞ ½D
where Rtotal is the total amount of drug that diffused into the receiver chamber over a given interval of time, t: A is the area of exposed sclera, 0·385 cm2 for the 7 mm chamber. ½D is the concentration of drug in the donor chamber at
the time of the measurement, which was estimated using a running mass balance, where the cumulative amount of drug collected in the receiver chamber from the beginning of the experiment through the time of the measurement of interest was subtracted from the initial amount of drug in the donor chamber. Permeability thus represents the flux normalized by donor concentration. This calculation expresses the rate (cm sec21) at which test drugs traverse the sclera at steadystate. Each experiment was given enough time for the initial transient to end. Permeability values reported here represent the average of permeabilities calculated at all time points after this transient. The rate of delivery for each drug was calculated by dividing the measured flux [nmole (cm2 hr)21] by the actual amount (nmole) of drug in the donor chamber and multiplying by the area (cm2) of exposed sclera. The average rate of delivery represents the average of the rates of delivery calculated at all time points after the initial transient described above. Cumulative percent release of each drug was calculated from the total amount of drug collected from the receiver chamber divided by initial amount of drug in donor chamber. Mean permeability values (^ S.D. ) were calculated from 5 to 7 experiments performed for each compound. Analysis of variance (ANOVA) was calculated to compare the Ktrans values among experimental groups.
3. Results 3.1. Scleral thickness as a function of pressure Control scleral thickness was measured as 600 ^ 49 mm (mean ^ S.E. ) using the standard 7 mm perfusion chamber ðn ¼ 6Þ: As trans-scleral pressure was increased, no significant change in thickness was observed until a pressure of 60 mmHg was reached. The percent change in scleral thickness from initial control values was 0·38 ^ 2·1, 0·26 ^ 3·0, and 2 3·9 ^ 3·4% at 15, 30, and 60 mmHg, respectively. Only the thickness change at 60 mmHg was found to be statistically significant from control thickness ðp , 0·05Þ; repeated measures ANOVA (Table 1 and Fig. 2). Control scleral thickness was measured as 634 ^ 38 mm for the 10 mm chamber experiments ðn ¼ 5Þ: As
Table 1 Scleral thickness (mm) as a function of trans-scleral pressure Pressure (mmHg)
0 15 30 60
Duration (hr)
Initial 2 2 2
Chamber size: 7 mm ðN ¼ 6Þ
Chamber size: 10 mm ðN ¼ 5Þ
Scleral thickness
Percent change
Scleral thickness
Percent change
600 ^ 49 595 ^ 46 599 ^ 47 573 ^ 43
Control value 20·38 ^ 2·1 0·26 ^ 3·0 23·9 ^ 3·4*
634 ^ 38 630 ^ 38 627 ^ 35 616 ^ 31
Control value 20·56 ^ 1·3 21·0 ^ 1·6 22·7 ^ 1·6
Values shown represent the mean ^ S.E . *P , 0·05 compared to control value at 0 mmHg.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
Fig. 2. Percent change in scleral thickness as a function of trans-scleral pressure. Values shown represent the mean ^ S.E . For the 7 mm chamber, there was a statistically significant decrease in scleral thickness from 30 to 60 mmHgp (* p , 0·05; Repeated measures ANOVA). For the 10 mm chamber, no statistically significant changes in scleral thickness were observed through the range of trans-scleral pressures studied.
trans-scleral pressure was increased, no significant change in thickness was observed (p . 0·05; Repeated measures ANOVA). The percent change in scleral thickness from initial control thickness was 2 0·56 ^ 1·3, 2 1·0 ^ 1·6, and 2 2·7 ^ 1·6% under 15, 30, and 60 mmHg, respectively. (Table 1 and Fig. 2). 3.2. Scleral hydration as a function of pressure The mean hydration of 15 control scleral specimens at 0 mmHg was measured as 71·1 ^ 2·3%. Changes in scleral hydration from its initial control value (at 0 mmHg) using the 7 mm perfusion chamber were 2 0·79 ^ 1·06% ðn ¼ 5Þ; 0·66 ^ 1·40% ðn ¼ 5Þ; and 0·05 ^ 1·30% ðn ¼ 5Þ at 15, 30, and 60 mmHg, respectively. No statistically significant changes in scleral hydration were observed over a transscleral pressure range of 0 – 60 mmHg (p . 0·10; One-way ANOVA) (Table 2 and Fig. 3). 3.3. Permeability studies The permeability constant ðKtrans Þ of the human sclera for H-dexamethasone in BSS plus (control value) was 11·5(^ 0·1) £ 1026 cm sec21 ðn ¼ 5Þ: For convenience of
3
Table 2 Scleral hydration as a function of trans-scleral pressure (7 mm chamber) compared to control hydration at 0 mmHg Trans-scleral pressure (mmHg)
15 30 60
Duration (hr)
4 4 4
N
5 5 5
Scleral hydration (%)
Initial control value
After experiment
% change
70·7 ^ 1·3 70·5 ^ 1·0 72·0 ^ 0·9
70·1 ^ 1·0 71·0 ^ 1·0 72·2 ^ 1·8
20·79 ^ 1·1 0·66 ^ 1·4 0·05 ^ 1·3
Values shown represent the mean ^ S.E .
603
Fig. 3. Percent change in scleral hydration as a function of trans-scleral pressure. Values shown represent the mean ^ S.E . No significant (p . 0·10; One-way ANOVA) change in scleral hydration was observed through the range of trans-scleral pressures studied.
comparison, an apparent permeability constant can be calculated to quantify transport of 3H-dexamethasone from F-127 gel and fibrin sealant across the sclera. Although the presence of these controlled release systems would not be expected to affect the intrinsic permeability of the sclera, the apparent permeability constants reported here account for the barrier to diffusion imposed by the combination of the sclera and the controlled release system. The Ktrans for dexamethasone in F-127 gel and fibrin sealant were 1·5(^ 0·3) £ 1026 cm sec21 ðn ¼ 5Þ and 7·3(^ 1·0) £ 1026 cm sec21 ðn ¼ 7Þ; respectively. The apparent permeabilities of dexamethasone in F-127 and fibrin sealant were 13 and 64% of the BSS plus control value, respectively. Human scleral permeability to dexamethasone ˚ ) in the (molecular weight: 392 Da, molecular radius: 5·2 A different vehicles are shown in Table 3. ANOVA showed that human scleral permeability to dexamethasone differed significantly among the three vehicles ðp , 0·0001Þ: The average steady-state rate of delivery of 3H-dexamethasone across human sclera from BSS plus (control value) was calculated to be 8·1(^ 0·7) £ 1022 hr21 ðn ¼ 5Þ: The value in pluronic F-127 gel and fibrin sealant were 1·8(^ 0·03) £ 1022 hr21 ðn ¼ 5Þ and 4·3(^ 0·1) £ 1022 hr21 ðn ¼ 4Þ; respectively. The rates of delivery in F-127 and fibrin sealant were 22 and 54% of control value, respectively. These rates of steady-state delivery were significantly less than control delivery in BSS plus ðp , 0·0001Þ; one-way ANOVA (Table 3 and Fig. 4). As shown in Fig. 4, pluronic F-127 gel and fibrin sealant did provide a slow, uniform sustained release through a 24 hr period. Under the conditions used, the F-127 provided more uniform sustained release compared to the fibrin sealant. The cumulative release of 3H-dexamethasone through human sclera from BSS plus (control value) was 84·0 ^ 1·5% (n ¼ 5; 20 hr). Cumulative release from pluronic F-127 gel and fibrin sealant were 29·3 ^ 5·8% (n ¼ 5; 20 hr) and 61·4 ^ 5·9% (n ¼ 4; 20 hr), respectively. One-way Analysis of Variance showed these values to be significantly different ðp , 0·0001Þ: The F-127 and fibrin
604
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
Table 3 Apparent scleral permeability ðKtrans Þ; average rate of delivery, and cumulative delivery of dexamethasone in BSS plus and in sustained release systems
Pluronic F-127 gel Fibrin sealant BSS plus (control)
Ktrans ( £ 1026 cm sec21)
Average rate of delivery ( £ 1022 hr21)
Cumulative percent delivery (%)
1·5 ^ 0·3 ðn ¼ 5Þ 7·3 ^ 1·0 ðn ¼ 7Þ 11·5 ^ 1·0 ðn ¼ 5Þ
1·8 ^ 0·03 ðn ¼ 5Þ 4·3 ^ 0·1 ðn ¼ 4Þ 8·1 ^ 0·7 ðn ¼ 5Þ
29·3 ^ 5·8 (20 hr, n ¼ 5) 61·4 ^ 5·9 (20 hr, n ¼ 4) 84·0 ^ 1·5 (20 hr, n ¼ 5)
Values shown represent the mean ^ S.E .
sealant cumulative releases were 34 and 80% of the control BSS plus value, respectively (Table 3 and Fig. 5).
Several studies have shown the sclera to be permeable to a range of drugs and solutes (Ahmed and Patton, 1985, 1987; Schoenwald et al., 1997). Maurice and Polgar (1977) have reported that the sclera generally offers less resistance to solute diffusion than does the cornea. The permeability of the sclera to carbonic anhydrase inhibitors is similar to the permeability of the corneal stroma (Edelhauser and Maren, 1988). Hydrocortisone permeability is five times greater in sclera than in corneas with intact epithelia (Unlu and Robinson, 1998). Ahmed et al., 1987 compared conjunctival, scleral and corneal permeation of adrenergic blocking agents and found the scleral permeabilities for all drugs tested to be significantly higher than their corneal permeabilities. The sclera was also found to be more permeable than the cornea to polyethylene glycol (Ha¨ma¨la¨inen et al., 1977). Olsen et al. (1995) also investigated the permeability of the human sclera to solutes ranging in molecular weight from 350 to 70 000 Da, finding an
inverse relationship between scleral permeability and solute molecular weight. More recently, Ambati et al. (2000a,b) have shown the sclera to be permeable to higher molecular weight dextrans, as well as to the proteins IgG and bovine serum albumin. This group has also demonstrated that in vivo trans-scleral delivery is capable of maintaining significant levels of biologically active protein in the choroid and retina of the rabbit eye (Ambati et al., 2000a,b). The sclera is an elastic and microporous tissue composed of proteoglycans and closely packed collagen fibrils, (Foster and Sainz de la Maza, 1994; Newell, 1996) containing approximately 70% water. Solutes traverse the tissue through the interfibrillar aqueous media of the gel-like proteoglycans (Maurice and Mishima, 1984). While the cornea is relatively impermeable to solutes having a molecular size over 1 kDa, dextran (40 kDa) and serum albumin (69 kDa) can readily penetrate across the scleral tissue (Bill, 1965; Olsen et al., 1995). Molecular size is therefore more of a limiting factor in the diffusion of drugs across the sclera than the lipophilicity of the solute. In a previous study, we observed that elevated transscleral pressure reduced human scleral permeability to dexamethasone by as much as four-fold at 60 mmHg
Fig. 4. Rate of delivery of dexamethasone in BSS Plus (average steady-state value of 8·06 ^ 0·67 £ 1022 hr21) and in the two sustained release delivery systems—F-127 (average steady-state value of 1·80 ^ 0·03 £ 1022 hr21) and fibrin sealant (average steady-state value of 4·33 ^ 0·11 £ 1022 hr21). Values shown represent the mean ^ S.E . The average steady-state rates of delivery in F-127 and fibrin sealant were 22 and 54% of the BSS plus control value, respectively. These calculated steady-state rates of delivery were significantly different (p , 0·0001; One-way ANOVA).
Fig. 5. Percent cumulative release of 3H-dexamethasone through human sclera from BSS Plus and the two sustained release systems—F-127 and fibrin sealant. Values shown represent the mean ^ S.E . The cumulative release of 3H-dexamethasone through human sclera from BSS plus (control value) was 84·0 ^ 1·5% (n ¼ 5; 20 hr). The values in pluronic F-127 gel and fibrin sealant were 29·3 ^ 5·9% (n ¼ 5; 20 hr) and 61·4 ^ 5·9% (n ¼ 4; 20 hr), respectively. One-way Analysis of Variance showed these values to be significantly different ðp , 0·0001Þ:
4. Discussion
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
(Rudnick et al., 1999). We hypothesized that this lower permeability was due to compression of the sclera, which reduced its water content and thereby reduced the spaces between extracellular matrix molecules, which are the pathways through which dexamethasone molecules diffuse across the sclera. In the present study we measured changes in scleral thickness and hydration. We found that transscleral pressure had no significant effect on the hydration of the tissue over the range of pressures studied, up to 60 mmHg. In this study, control scleral hydration was 71·1 ^ 2·3% or 2·5 ^ 0·3 mg H2O mg21 dry weight. This value is slightly lower than the 2·9 ^ 0·2 mg H2O mg21 dry weight for scleral tissue obtained from two-day moist chamber-stored globes in Olsen’s report (1995). Our value agrees well with the hydration value of 71·8% in Madhu’s report (1998). Though our results do indicate that intraocular pressure can affect scleral thickness, we found that the effect of pressure on thickness was minimal at trans-scleral pressures up to 60 mmHg. The modest changes in thickness that do occur at higher pressures most likely occur by changes in the tissue’s microanatomy. Elevated pressure compressing the tissue, which in turn would reduce the spaces between the collagen fibers and extracellular matrix molecules that define the pathways for diffusion. However, the modest changes in thickness at high trans-scleral pressures are probably not sufficient to significantly affect rates of diffusion across the tissue, especially for small solutes such as dexamethasone. Pluronic F-127, a polyoxyethylene-polyoxypropylene surface-active block copolymer, with an average molecular weight of 11 500 Da (Miyazaki et al., 1984) has good potential for use as a sustained release delivery system since it has good drug release characteristics, low toxicity, and exhibits reverse thermal gelation (Schmolka, 1972). It remains in the liquid state at refrigerator temperatures and gels upon warming to physiological temperatures (Miyazaki et al., 1984). Thus, if the liquid form, containing drug, were injected into subconjunctival space, it would form a solid sustained release depot. Since F-127 gels are viscous isotropic liquid crystals consisting of micelles, (Chen-Chow and Frank, 1981) incorporated drugs would be released by diffusion through the extramicellar water channels of the gel matrix. Hence, the rate of drug release would be determined by the micro-viscosity of the extramicellar fluid, the size and tortuosity of the water channels, and the drug partitioning between the micellar phase and the external water phase. Chen-Chow and Frank (1981) and Miyazaki et al. (1984) studied the effect of F-127 concentration and temperature on the release of drug and found that the release rate decreased with increasing concentration of F-127. This was the result of a reduction in the size and increase in tortuosity of the water channels and an increase in the micro-viscosity of the water channels of the gel. Release rate increased with increasing temperature due to decreased
605
viscosity of water channels (Chen-Chow and Frank, 1981; Miyazaki et al., 1984). Fibrin sealant is a water-insoluble gel matrix formed when fibrinogen is activated by thrombin in the presence of Ca2þ ion and factor XIII. In recent years, fibrin sealant has become a well-known biomaterial because of its hemostatic, adhesive, and biocompatible properties (Yu et al., 1996). It has been successfully used in a wide range of surgical fields such as skin graft fixation, nerve repair, cartilage reattachment and microvascular anastomoses (Tashiro et al., 1987; Wood and Harner, 1988; Sierra et al., 1990; Byrne et al., 1991; Yu et al., 1996). Fibrin sealant also has the additional advantages of being biodegradable and commercially available. The results of this study indicate that fibrin sealant has potential for use as a sustained release drug delivery system. One approach to trans-scleral drug delivery would be to administer a drug, in a solution or suspension, by a periorbital, retrobulbar, or intrascleral injection. In this case, the rate at which the drug would diffuse into the eye would be determined principally by scleral permeability. Another approach would be to utilize a sustained release delivery system formed by a gel, a solid polymer, or another system that could be placed in contact with the outer surface of the sclera. Ideally, the drug could be released from the system at a controlled rate. In this case, the maximum drug delivery rate into the eye would be constrained by scleral permeability, but the device itself could control the rate of delivery if the barrier to diffusion imposed by the device were less than that of the sclera. Dexamethasone was chosen as a model compound since it is similar in molecular size and diffusivity to most conventional ocular drugs and it represents a reasonably hydrophilic drug that can be used to test the pathway of drug movement (Prausnitz and Noonan, 1988). The permeability of human sclera to dexamethasone in BSS plus measured in this study was 11·5(^ 1·1) £ 1026 cm sec21. This value is comparable to that reported by Rudnick et al. (1999) but is lower than that reported by Olsen et al. (1995) 23·5(^ 7·7) £ 1026 cm sec21. Different chambers, however, were used in these studies. It is possible that Olsen’s chamber design with constant chamber mixing may yield a higher scleral permeability measurement than that determined in this study which employed an unmixed depot on the surface of the sclera in which static boundary layers could form. The apparent permeability values of dexamethasone in F127 and fibrin sealant were 13 and 64% of BSS plus control value, respectively. These differences might be due to smaller pore sizes within the F-127 gel compared to that of the fibrin sealant. The F-127 gel was found to have a lower release rate and higher drug retention, resulting in a longer sustained release delivery by F-127 compared to fibrin sealant. There is an initial lag period in dexamethasone flux during which the flux sharply increases and then reaches a peak after approximately 3 hr for BSS plus, 4 hr for fibrin
606
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
sealant, and 7 hr for F-127, Fig. 4. This lag time is due to both diffusion through and possible binding within the sclera. After reaching its peak, the flux slowly decreases. This is because the donor source of solute is finite and becomes depleted over time, thereby reducing the concentration gradient across the sclera. These lag times are longer than the residence time at the scleral surface for drug introduced by peribulbar injection. Therefore, the use of F127 and fibrin sealant should significantly extend the time scale of trans-scleral delivery and possibly increase the total amount of drug delivered to the eye. To summarize, we have established the initial feasibility of trans-scleral drug delivery using the sustained release systems of F-127 and fibrin sealant. These systems provided a slow sustained release through a 24 hr period. Additionally, scleral hydration was found not to change significantly over the range of trans-scleral pressures studied (0 – 60 mmHg). Scleral thickness did show a tendency to decrease as trans-scleral pressure increased to 60 mmHg. The results of this study suggest that sustained release drug delivery using systems such as F127 or fibrin sealant show promise as systems to deliver therapeutic levels of bioactive drugs to the posterior segment of eye. These compounds may be employed in the future development of a practical system or device for drug delivery across the sclera for the treatment of a variety of chorioretinal disorders.
Acknowledgements Funded in part by a grant from The Foundation Fighting Blindness, a grant from the Knights Templar Educational Foundation, by NEI P30 EY06360, and by an unrestricted grant from RPB, Inc.
References Ahmed, I., Gokhale, R.D., Shah, M.V., Patton, T.F., 1987. Physicochemical determinants of drug diffusion across the conjunctiva, sclera, and cornea. J. Pharm. Sci. 76, 583–586. Ahmed, I., Patton, T.F., 1985. Importance of the noncorneal absorption route in topical ophthalmic drug delivery. Invest. Ophthalmol. Vis. Sci. 26, 584 –587. Ahmed, I., Patton, T.F., 1987. Disposition of timolol and inulin in the rabbit eye following corneal versus non-corneal absorption. Int. J. Pharm. 38, 9–21. Ambati, J., Canakis, C.S., Miller, J.W., Gragoudas, E.S., Edwards, A., Weissgold, D.J., Kim, I., Delori, F.C., Adamis, A.P., 2000. Diffusion of high molecular weight compounds through sclera. Invest. Ophthalmol. Vis. Sci. 41, 1181–1185. Ambati, J., Gragoudas, E.S., Miller, J.W., You, T.T., Miyamoto, K., Delori, F.C., Adamis, A.P., 2000. Transscleral delivery of bioactive protein to the choroid and retina. Invest. Ophthalmol. Vis. Sci. 41, 1186–1191. Barza, M., Kane, A., Baum, J.L., 1977. Regional differences in ocular concentrations of gentamicin after subconjunctival and retrobulbar injection in the rabbit. Am. J. Ophthalmol. 83, 407–413.
Barza, M., Kane, A., Baum, J.L., 1978. Intraocular penetration of gentamicin after subconjunctival and retrobulbar injection. Am. J. Ophthalmol. 85, 541–547. Bill, A., 1965. Movement of albumin and dextran through the sclera. Arch. Ophthalmol. 74, 248–252. Byrne, D.J., Hardy, J., Wood, R.A.B., McIntosh, R., Cuschieri, A., 1991. Effect of fibrin glues on the mechanical properties of healing wounds. Br. J. Surg. 78, 841–843. Chen-Chow, P., Frank, S.G., 1981. Comparison of lidocaine release from pluronic F-127 gels and other formulations. Acta Pharm. 18, 239 –244. Chrai, S.S., Patton, T.F., Mehta, A., Robinson, J.F., 1973. Lacrimal and instilled fluid dynamics in rabbit eyes. J. Pharm. Sci. 62, 1112–1121. Edelhauser, H.F., Maren, T.H., 1988. Permeability of human cornea and sclera to sulfonamide carbonic anhydrase inhibitors. Arch. Ophthalmol. 106, 1110–1115. Edwards, A., Prausnitz, M.R., 1998. A fiber matrix model of sclera and corneal stroma for drug delivery to the eye. AIChE J. 44, 214–225. Foster, C.S., Sainz de la Maza, M., 1994. The Sclera, Springer, New York, NY, pp. 1 –32. Freeman, W.R., 1989. Intraocular antiviral therapy. Arch. Ophthalmol. 107, 1737–1739. Ha¨ma¨la¨inen, K.M., Kananen, K., Auriola, S., Kontturi, K., Urtti, A., 1977. Caracterization of paracellular and aqueous penetration routes in cornea, conjunctiva, and sclera. Invest. Ophthalmol. Vis. Sci. 38, 627 –634. Lang, J.C., 1995. Ocular drug delivery: conventional ocular formulations. Adv. Drug Deliv. Rev. 16, 39–43. Madhu, C., Rix, P., Nguyen, T., Chein, D., Woodward, D.F., Diane, D.S., Tang-Liu, D.D.S., 1998. Penetration of natural prostaglandins and their ester prodrugs and analogs across human ocular tissues in vitro. Ocul. Pharmacol. Ther. 14, 389–399. Maurice, D.M., 2002. Drug delivery to the posterior segment from drops. Surv. Ophthalmol. 47, S41–S51. Maurice, D.M., Mishima, S., 1984. Ocular pharmacokinetics. In: Sears, M.L. (Ed.), Pharmacology of the Eye, Springer, New York, NY, pp. 19 –116. Maurice, D.M., Polgar, J., 1977. Diffusion across the sclera. Exp. Eye Res. 25, 577–582. Miyazaki, S., Takeuchi, S., Yokouchi, C., Takada, M., 1984. Pluronic F127 gels as a vehicle for topical administration of anticancer agents. Chem. Pharm. Bull. 32, 4205–4208. Newell, F.V., 1996. Ophthalmology, Principles and Concepts, Sixth Ed, CV Mosby Co, St Louis, MO. Olsen, T.W., Aaberg, S., Geroski, D.H., Edelhauser, H.F., 1998. The human sclera: thickness and surface area. Am. J. Ophthalmol. 125, 237– 241. Olsen, T.W., Edelhauser, H.F., Lim, J.I., Geroski, D.H., 1995. Human scleral permeability: effects of age, cryotherapy, trans-scleral diode laser, and surgical thinning. Invest. Ophthalmol. Vis. Sci. 36, 1893–1903. Prausnitz, M.R., Noonan, J.S., 1998. Permeability of cornea, sclera, and conjunctiva: a literature analysis for drug delivery to the eye. J. Pharm. Sci. 87, 1479–1488. Rudnick, D.E., Noonan, J.S., Geroski, D.H., Prausnitz, M.R., Edelhauser, H.F., 1999. The effect of intraocular pressure on human and rabbit scleral permeability. Invest. Ophthalmol. Vis. Sci. 40, 3054–3058. Schmolka, I.R., 1972. Artificial skin. I. Preparation and properties of pluronic F-127 gels for treatment of burns. J. Biomed. Mater. Res. 6, 571 –582. Schoenwald, R.D., Deshpande, G.S., Rethwisch, D.G., Barfknecht, C.F., 1997. Penetration into the anterior chamber via the conjunctival/scleral pathway. J. Ocul. Pharmacol. 13, 41 –59. Sierra, D.H., Nissen, A.J., Welch, J., 1990. The use of fibrin glue in intracranial procedures: preliminery results. Laryngoscope 100, 360 –363. Tashiro, S., Murata, E., Hiraoka, K., Nakakuma, K., Watanabe, E., Miyauchi, Y., 1987. New technique for pancraticojejunostomy using a biological adhesive. Br. J. Surg. 74, 392– 394.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607 Unlu, N., Robinson, J.R., 1998. Scleral permeability to hydrocortisone and mannitol in the albino rabbit eye. J. Ocul. Pharmacol. 14, 273–281. Watsky, M.A., Olsen, T.W., Edelhauser, H.F., 1995. Cornea and sclera. In: Tasman, W., Jaeger, E.A. (Eds.), Cornea and Sclera, Lippincott/Raven, Philadelphia, PA, pp. 1–29.
607
Wood, A.P., Harner, S.G., 1988. The effect of fibrin tissue adhesive on the middle and inner ears of chinchillas. Otolaryngol. Head Neck Surg. 98, 104– 110. Yu, B.G., Kwon, I.C., Kim, Y.H., Han, D.K., Park, K.D., Han, K., Jeong, S.Y., 1996. Development of a local antibiotic delivery system using fibrin glue. J. Control. Release 39, 65–70.
Drug delivery through the sclera: effects of thickness, hydration, and sustained release systems Sang-Bumm Leea,c, Dayle H. Geroskia, Mark R. Prausnitzb, Henry F. Edelhausera,* b
a Emory Eye Center, Emory University, Suite B2600, 1365B Clifton Road, NE, Atlanta, GA, USA Schools of Chemical Engineering and Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, USA c Department of Ophthalmology, Yeungnam University Medical Center, Taegu, South Korea
Received 25 June 2003; accepted 26 June 2003
Abstract The purpose of this study was to determine whether trans-scleral pressure affects scleral solute permeability by altering scleral thickness or hydration, and to investigate the sustained release delivery of dexamethasone. Scleral sections from donor human globes were mounted for in vitro flux studies. Scleral thickness and hydration were measured as functions of trans-scleral pressure. For the sustained release studies, 3Hdexamethasone in pluronic F-127 gel or in fibrin sealant was added to the episcleral side of the tissue and flux studies were performed. While scleral thickness showed a tendency to decrease with increasing pressure, a significant decrease in thickness was measured only at a transscleral pressure of 60 mmHg. No significant changes in scleral hydration were measured over the range of trans-scleral pressures studied. The apparent permeability constants ðKtrans Þ of human sclera for 3H-dexamethasone in BSS plus, fibrin sealant and F-127 gel were 11·5 £ 1026, 7·3 £ 1026, and 1·5 £ 1026 cm sec21, respectively. Human scleral permeability to dexamethasone differed significantly among the three vehicles ðp , 0·0001Þ: Cumulative delivery of dexamethasone from BSS plus, F-127 gel, and fibrin sealant were 85·0, 29·3, and 67·9% at 20 hr, respectively. Scleral hydration was unaffected by trans-scleral pressures. Scleral thinning was only observed at 60 mmHg. Transscleral pressures below 60 mmHg would not be expected to significantly affect the permeability of the tissue to solutes in the size range of conventional drugs. F-127 gel and fibrin sealant provided a slow, relatively uniform sustained release through a 24 hr period. These systems might be employed to achieve sustained therapeutic levels of drugs to the posterior segment of eye. q 2003 Elsevier Ltd. All rights reserved. Keywords: scleral drug delivery; scleral permeability; scleral thickness; hydration; sustained release; fibrin sealant; pluronic F-127 gel
1. Introduction In David Maurice’s insight and wisdom he was the first person to address the sclera as a route of drug penetration to the posterior segment. His seminal paper dealt with the bovine scleral permeability (Maurice and Polgar, 1977) and this basic drug permeability information remained dormant for a number of years until a major interest developed treating retinal diseases with therapeutic drugs, growth factors proteins and oligonucleotides. In David’s last published paper he addresses the issue of drug delivery to the Posterior Segment from Drops (Maurice, 2002). He stated that ‘he found it difficult to conceive that any * Corresponding author. Dr Henry F. Edelhauser, Emory Eye Center, Emory University, Suite B2600, 1365B Clifton Road, NE, Atlanta, GA, USA. E-mail address: [email protected] (H.F. Edelhauser). 0014-4835/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. DOI:10.1016/S0014-4835(03)00211-2
biological purpose could be served by a system that directs a solute from the tear film, and presents to the retina at about one-millionth of its original concentration’. Therefore there must be another route for a drug to obtain entry to the eye, and he reviewed the importance of the conjunctival cul-desac route to the posterior sclera and an increased residence time. Our laboratory has expanded on David’s initial scleral permeability studies and this paper addresses some of the issues of trans-scleral drug delivery. Topical and systemic treatment of both the anterior and posterior segments of the eye are often difficult because of low ocular tissue permeabilities, diffusion of tear fluid away from the cornea, unwanted counterdirectional convection of drugs, and long diffusional path lengths (Chrai et al., 1973; Lang, 1995; Watsky et al., 1995; Schoenwald et al., 1997). Drugs that are topically applied are thus poorly absorbed,
600
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
with typically less than 5% of the drug reaching anterior intraocular tissues (Lang, 1995). Although systemic administration can deliver drugs to the posterior eye, the large systemic doses necessary are often associated with side effects (Rudnick et al., 1999). Intravitreal injection also delivers drug directly into the posterior chamber. It is, however, an invasive technique and thus has inherent risks for ocular infections and tissue damage (Freeman, 1989). Delivering drugs across the sclera offers another vector to obtain therapeutic vitreous and retinal drug concentrations (Chrai et al., 1973, Maurice and Polgar, 1977; Olsen et al., 1995; Barza et al., 1977, 1978; Unlu and Robinson, 1998). This approach is less invasive and thus should be safer than intravitreal injection. Trans-scleral delivery could take advantage of the large surface area of the sclera and would also offer the potential for localized, sustained release delivery. The relatively high scleral permeability (Ha¨ma¨la¨inen et al., 1977; Ahmed and Patton, 1985; Edelhauser and Maren, 1988; Schoenwald et al., 1997), as compared to the cornea, has motivated research in trans-scleral drug delivery, especially for compounds that need to be administered to the posterior part of the eye (Olsen et al., 1998; Edwards and Prausnitz, 1998), and since the human sclera has a large surface area compared with the corneal surface area – it comprises 95% of the surface area of the eye (Olsen et al., 1998) it provides a promising alternative as a drug delivery route to the posterior segment. Once the scleral permeability is known for a given drug, new delivery systems can be evaluated for sustained release. A variety of sustained release drug delivery systems exists, including gels, erodible polymers, microspheres, liposomes, inserts, miniosmotic pumps and combinations of these technologies. In recent years, pluronic F-127 and fibrin sealant have been widely used in medical and pharmaceutical systems (Miyazaki et al., 1984; Yu et al., 1996). These compounds have good tissue compatibility and show exciting promise for trans-scleral application. An approach to trans-scleral sustained release drug delivery could be based on the injection of a small, concentrated depot into a solid or semisolid sustained release delivery system, such as fibrin sealant or pluronic F-127, located on or within the sclera. For this study, we used a chamber that emulates depot delivery from the scleral surface (Rudnick et al., 1999). The receiver chamber, representing the uveal tissues, is perfused at a slow rate. The donor chamber is held static, similar to a drug depot added to Tenon’s space and directly exposed to the sclera. We studied the potential effects of trans-scleral pressure on the solute permeability of the human sclera. Additionally, we studied two sustained release systems (pluronic F-127 gel and fibrin sealant). The diffusion of dexamethasone across human sclera was measured following a depot application with both sustained release systems to the episclera.
2. Materials and methods Human scleral tissue was obtained from 43 human donor eyes (Georgia Eye Bank, Atlanta, GA, USA) that had been stored in moist chambers for an average (^ S.D. ) of 5·0 ^ 1·2 days. Mean age (^ S.D. ) at time of death was 50·2 ^ 18·2 years. All scleral tissue was taken from adult eyes. Past studies have shown that human scleral water content does not change significantly following moist chamber storage for up to 10 days and that BSS Plus (Alcon Laboratories, Ft. Worth, TX, USA) will maintain tissue hydration during in vitro diffusion studies (Olsen et al., 1995). A razor blade was used to create a full-thickness incision through the sclera near the limbus. Curved scissors were then used to remove a circular piece of sclera. Adherent tissue associated with the retina, choroid, episclera, or Tenon’s capsule was gently removed with a cotton-tip applicator to isolate the bare sclera. Full-thickness scleral disks of 15– 20 mm in diameter were excised from the superior temporal quadrant of the globe, from just posterior to the limbus to 12 – 15 mm posterior to the limbus. This area avoids the anterior ciliary perforating vessels, the short posterior ciliary perforating vessels, and the vortex veins. For the permeability studies, the excised sclera was mounted vitreous side down in a specially designed lucite perfusion chamber in which the sclera is mounted horizontally (Fig. 1). The sclera was clamped between two 2·5 mm wide and 1 mm thick cylindrical Sylgard rings (Dow Corning, Inc., Midland, MI, USA) to prevent lateral leakage and scleral edge damage. The donor (upper) chamber, which serves as a depot, holds 100 –200 ml of a drug in solution or gel. The receiver (lower) chamber, which is a flow-through chamber, holds 500 ml and is continually stirred with a magnetic bar. BSS Plus, containing 1% antibiotic-antimycotic solution (Sigma, St Louis, MO, USA) was perfused through the receiver chamber at a rate of 0·03 ml min21 for experiments less than 8 hr in duration, and at a rate of 0·01 ml min21 for longer experiments. A 7 mm diameter perfusion chamber was used for all experiments. Additional experiments using a 10 mm diameter perfusion chamber were performed for the scleral thickness studies. The larger diameter chamber facilitated the placement of the ultrasonic pachymeter probe on the exposed scleral surface. The tissue was perfused for 15 –30 min to verify that no leaks were present before adding a test compound to the donor chamber. Experiments where leakage occurred were not included in the results. Except as noted, experiments were performed with 15 mmHg applied to the vitreous side of the sclera. Pressure was determined by measuring the vertical distance between the tissue and the outflow tube as it flowed into the collector receptacle (e.g. 15 mmHg is equivalent to a 22 cm water column). The pressure across the tissue was verified using a Statham pressure transducer (Oxnard, CA, USA) connected to the receiver chamber. The temperature of the water-jacketed perfusion chamber was maintained at 37 8C by a circulating water bath.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
601
Fig. 1. Schematic diagram of the perfusion chamber used for measuring scleral permeability under simulated intraocular pressure. Sclera is mounted in a horizontal perfusion setup (episcleral side up) to allow for an intraocular pressure to be simulated by creating a positive water column with the outflow tube. The compound to be tested is added to the donor (upper), episcleral hemichamber. The receiver (lower), uveal hemichamber has a continual flow and is mixed by a magnetic stir bar with the perfusion chamber placed on a magnetic stir plate.
2.1. Scleral thickness and hydration as functions of pressure
2.2. Sustained release permeability studies
These experiments were performed to determine whether trans-scleral pressure affected either the thickness or the hydration of the tissue, which in turn might alter the tissue’s permeability. Scleral tissue was mounted in the perfusion chamber as previously described. Initial (control) scleral thickness was measured using a DGH 500 Pachette ultrasonic pachymeter (DGH Technology, Frazer, PA, USA) just after tissue mounting and prior to any pressure application. The pressure was increased to 15 mmHg and scleral thickness was measured at 15 min intervals. After two hours, the pressure was raised to 30 mmHg and the thickness measurements were continued at 15 min intervals for 2 hr. Pressure was raised once more to 60 mmHg and scleral thickness was measured for the final two hours of the experiment. At each 15 min interval, five scleral thickness readings were taken and the readings averaged to establish scleral thickness at each time point. For the hydration studies, the sclera was mounted in the perfusion system and a constant pressure was maintained for 4 hr. The tissue was then removed, blotted dry and weighed. The sclera was then dried to constant weight at 120 8C for 24 hr. The dried tissue was immediately placed in a tissue desiccator and allowed to cool for 30 min. Samples were re-weighed. The hydration of each piece of sclera was then calculated as: mg H2O mg21 wet tissue (%) [(wet weight 2 dry weight)/wet weight £ 100%]. Five individual experiments were performed at each transscleral pressure (15, 30, and 60 mmHg). To determine control hydration, an additional piece of fresh scleral tissue was excised in each experiment and hydration was measured in each of these unperfused (control) specimens.
For the preparation of fibrin sealant (Tisseelw VH Kit 2·0 ml: Baxter Healthcare Corporation, Glendale, CA, USA) formulations, 80 mCi in 80 ml of 3H-dexamethasone was mixed into 2 ml CaCl2 solution (40 mmol ml21). The 3 H-dexamethasone and CaCl2 solution was injected into the vial containing the freeze-dried human thrombin (500 I.U. ml21). The thrombin solution thus prepared was kept at 37 8C until used. Two milliliters of fibrinolysis inhibitor solution (bovine, 3000 KIU of aprotinin ml21) was heated at 37 8C for 10 min and transferred into the vial containing the freeze-dried human sealer protein concentrate (100 –130 mg ml21 of total protein, 75 –115 mg ml21 of fibrinogen). The resulting thrombin solution and sealer protein solution were then combined to form the fibrin sealant by delivering equal volumes of each with a Duplojectw syringe. For the preparation of the F-127 gel formulations, 25 g of F-127 poloxamer was dissolved into 75 ml of chilled BSS plus (5 – 10 8C). The volume was brought to 100 ml by adding chilled BSS plus with constant mixing. The solution was left overnight in the refrigerator to ensure complete dissolution. The clear, viscous solution thus prepared remains fluid at refrigerator temperature (4 –5 8C). Fifty microcurie in 50 ml of 3H-dexamethasone was added to 450 ml of the F-127 solution. The dexamethasone in fibrin sealant or in the F-127 gel (100 ml total volume) was added to the episcleral surface of the sclera 15– 30 min after the sclera had been mounted in the perfusion chamber. The donor hemichamber containing the test compound was sealed with parafilm and silicone grease (Dow Corning) along the edges of the exposed area of the chamber to prevent evaporation. This provided a flexible seal that did not alter trans-scleral
602
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
pressure. At the beginning and end of each experiment, 10 ml aliquots of F-127 and a piece of fibrin sealant were taken from the episcleral side to calculate 3H-dexamethasone concentration in the sustained release systems. Permeability to two sustained release dexamethasone formulations were evaluated: 3H-dexamethasone-acetate in fibrin sealant, and 3H-dexamethasone-acetate in pluronic F127 gel. The ‘control’ for these sustained release studies was 3 H-dexamethasone-acetate in BSS plus (10 mCi in 100 ml). Samples of the perfusate were collected by a fraction collector (Isco, Lincoln, NE) at 30 – 60 min intervals. At the completion of the perfusion, 100 ml of each fraction was added to 10 ml of Aquasol (Packard, Meriden, CT, USA), and tritium disintegrations were measured using a Beckman LS 5801 liquid scintillation spectrometer (Irvine, CA, USA). Disintegrations per minute (DPMs) were calculated based on quenched standards of 3H. For the 3H-dexamethasone-acetate in BSS plus experiments, samples (10 ml) were taken from the donor chamber at the beginning and end of the experiment to verify the initial donor concentration and to measure donor drug depletion over the course of the experiment. For the fibrin sealant experiments, small pieces of the glue were taken from donor chamber at the beginning and end of each experiment. The pieces were weighed and then were solubilized in 2 ml of Solvable (Packard). After the glue dissolved, 10 ml of UltimaGold scintillation cocktail (Packard) was added to each sample and tritium was measured using scintillation spectroscopy. The concentration (dpm ml21) of 3H-dexamethasone in the initial and final samples of fibrin sealant from the donor chamber were calculated from the density of fibrin sealant (770 mg ml21 as determined by gravimetric analysis) and sample weight. For the F-127 experiments, 10 ml of pre- and post-perfusion donor chamber gel samples were taken in the cold room at 4 8C. The steady-state permeability constant ðKtrans Þ was calculated as Ktrans ¼
Rtotal 1 ; ðtÞðAÞ ½D
where Rtotal is the total amount of drug that diffused into the receiver chamber over a given interval of time, t: A is the area of exposed sclera, 0·385 cm2 for the 7 mm chamber. ½D is the concentration of drug in the donor chamber at
the time of the measurement, which was estimated using a running mass balance, where the cumulative amount of drug collected in the receiver chamber from the beginning of the experiment through the time of the measurement of interest was subtracted from the initial amount of drug in the donor chamber. Permeability thus represents the flux normalized by donor concentration. This calculation expresses the rate (cm sec21) at which test drugs traverse the sclera at steadystate. Each experiment was given enough time for the initial transient to end. Permeability values reported here represent the average of permeabilities calculated at all time points after this transient. The rate of delivery for each drug was calculated by dividing the measured flux [nmole (cm2 hr)21] by the actual amount (nmole) of drug in the donor chamber and multiplying by the area (cm2) of exposed sclera. The average rate of delivery represents the average of the rates of delivery calculated at all time points after the initial transient described above. Cumulative percent release of each drug was calculated from the total amount of drug collected from the receiver chamber divided by initial amount of drug in donor chamber. Mean permeability values (^ S.D. ) were calculated from 5 to 7 experiments performed for each compound. Analysis of variance (ANOVA) was calculated to compare the Ktrans values among experimental groups.
3. Results 3.1. Scleral thickness as a function of pressure Control scleral thickness was measured as 600 ^ 49 mm (mean ^ S.E. ) using the standard 7 mm perfusion chamber ðn ¼ 6Þ: As trans-scleral pressure was increased, no significant change in thickness was observed until a pressure of 60 mmHg was reached. The percent change in scleral thickness from initial control values was 0·38 ^ 2·1, 0·26 ^ 3·0, and 2 3·9 ^ 3·4% at 15, 30, and 60 mmHg, respectively. Only the thickness change at 60 mmHg was found to be statistically significant from control thickness ðp , 0·05Þ; repeated measures ANOVA (Table 1 and Fig. 2). Control scleral thickness was measured as 634 ^ 38 mm for the 10 mm chamber experiments ðn ¼ 5Þ: As
Table 1 Scleral thickness (mm) as a function of trans-scleral pressure Pressure (mmHg)
0 15 30 60
Duration (hr)
Initial 2 2 2
Chamber size: 7 mm ðN ¼ 6Þ
Chamber size: 10 mm ðN ¼ 5Þ
Scleral thickness
Percent change
Scleral thickness
Percent change
600 ^ 49 595 ^ 46 599 ^ 47 573 ^ 43
Control value 20·38 ^ 2·1 0·26 ^ 3·0 23·9 ^ 3·4*
634 ^ 38 630 ^ 38 627 ^ 35 616 ^ 31
Control value 20·56 ^ 1·3 21·0 ^ 1·6 22·7 ^ 1·6
Values shown represent the mean ^ S.E . *P , 0·05 compared to control value at 0 mmHg.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
Fig. 2. Percent change in scleral thickness as a function of trans-scleral pressure. Values shown represent the mean ^ S.E . For the 7 mm chamber, there was a statistically significant decrease in scleral thickness from 30 to 60 mmHgp (* p , 0·05; Repeated measures ANOVA). For the 10 mm chamber, no statistically significant changes in scleral thickness were observed through the range of trans-scleral pressures studied.
trans-scleral pressure was increased, no significant change in thickness was observed (p . 0·05; Repeated measures ANOVA). The percent change in scleral thickness from initial control thickness was 2 0·56 ^ 1·3, 2 1·0 ^ 1·6, and 2 2·7 ^ 1·6% under 15, 30, and 60 mmHg, respectively. (Table 1 and Fig. 2). 3.2. Scleral hydration as a function of pressure The mean hydration of 15 control scleral specimens at 0 mmHg was measured as 71·1 ^ 2·3%. Changes in scleral hydration from its initial control value (at 0 mmHg) using the 7 mm perfusion chamber were 2 0·79 ^ 1·06% ðn ¼ 5Þ; 0·66 ^ 1·40% ðn ¼ 5Þ; and 0·05 ^ 1·30% ðn ¼ 5Þ at 15, 30, and 60 mmHg, respectively. No statistically significant changes in scleral hydration were observed over a transscleral pressure range of 0 – 60 mmHg (p . 0·10; One-way ANOVA) (Table 2 and Fig. 3). 3.3. Permeability studies The permeability constant ðKtrans Þ of the human sclera for H-dexamethasone in BSS plus (control value) was 11·5(^ 0·1) £ 1026 cm sec21 ðn ¼ 5Þ: For convenience of
3
Table 2 Scleral hydration as a function of trans-scleral pressure (7 mm chamber) compared to control hydration at 0 mmHg Trans-scleral pressure (mmHg)
15 30 60
Duration (hr)
4 4 4
N
5 5 5
Scleral hydration (%)
Initial control value
After experiment
% change
70·7 ^ 1·3 70·5 ^ 1·0 72·0 ^ 0·9
70·1 ^ 1·0 71·0 ^ 1·0 72·2 ^ 1·8
20·79 ^ 1·1 0·66 ^ 1·4 0·05 ^ 1·3
Values shown represent the mean ^ S.E .
603
Fig. 3. Percent change in scleral hydration as a function of trans-scleral pressure. Values shown represent the mean ^ S.E . No significant (p . 0·10; One-way ANOVA) change in scleral hydration was observed through the range of trans-scleral pressures studied.
comparison, an apparent permeability constant can be calculated to quantify transport of 3H-dexamethasone from F-127 gel and fibrin sealant across the sclera. Although the presence of these controlled release systems would not be expected to affect the intrinsic permeability of the sclera, the apparent permeability constants reported here account for the barrier to diffusion imposed by the combination of the sclera and the controlled release system. The Ktrans for dexamethasone in F-127 gel and fibrin sealant were 1·5(^ 0·3) £ 1026 cm sec21 ðn ¼ 5Þ and 7·3(^ 1·0) £ 1026 cm sec21 ðn ¼ 7Þ; respectively. The apparent permeabilities of dexamethasone in F-127 and fibrin sealant were 13 and 64% of the BSS plus control value, respectively. Human scleral permeability to dexamethasone ˚ ) in the (molecular weight: 392 Da, molecular radius: 5·2 A different vehicles are shown in Table 3. ANOVA showed that human scleral permeability to dexamethasone differed significantly among the three vehicles ðp , 0·0001Þ: The average steady-state rate of delivery of 3H-dexamethasone across human sclera from BSS plus (control value) was calculated to be 8·1(^ 0·7) £ 1022 hr21 ðn ¼ 5Þ: The value in pluronic F-127 gel and fibrin sealant were 1·8(^ 0·03) £ 1022 hr21 ðn ¼ 5Þ and 4·3(^ 0·1) £ 1022 hr21 ðn ¼ 4Þ; respectively. The rates of delivery in F-127 and fibrin sealant were 22 and 54% of control value, respectively. These rates of steady-state delivery were significantly less than control delivery in BSS plus ðp , 0·0001Þ; one-way ANOVA (Table 3 and Fig. 4). As shown in Fig. 4, pluronic F-127 gel and fibrin sealant did provide a slow, uniform sustained release through a 24 hr period. Under the conditions used, the F-127 provided more uniform sustained release compared to the fibrin sealant. The cumulative release of 3H-dexamethasone through human sclera from BSS plus (control value) was 84·0 ^ 1·5% (n ¼ 5; 20 hr). Cumulative release from pluronic F-127 gel and fibrin sealant were 29·3 ^ 5·8% (n ¼ 5; 20 hr) and 61·4 ^ 5·9% (n ¼ 4; 20 hr), respectively. One-way Analysis of Variance showed these values to be significantly different ðp , 0·0001Þ: The F-127 and fibrin
604
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
Table 3 Apparent scleral permeability ðKtrans Þ; average rate of delivery, and cumulative delivery of dexamethasone in BSS plus and in sustained release systems
Pluronic F-127 gel Fibrin sealant BSS plus (control)
Ktrans ( £ 1026 cm sec21)
Average rate of delivery ( £ 1022 hr21)
Cumulative percent delivery (%)
1·5 ^ 0·3 ðn ¼ 5Þ 7·3 ^ 1·0 ðn ¼ 7Þ 11·5 ^ 1·0 ðn ¼ 5Þ
1·8 ^ 0·03 ðn ¼ 5Þ 4·3 ^ 0·1 ðn ¼ 4Þ 8·1 ^ 0·7 ðn ¼ 5Þ
29·3 ^ 5·8 (20 hr, n ¼ 5) 61·4 ^ 5·9 (20 hr, n ¼ 4) 84·0 ^ 1·5 (20 hr, n ¼ 5)
Values shown represent the mean ^ S.E .
sealant cumulative releases were 34 and 80% of the control BSS plus value, respectively (Table 3 and Fig. 5).
Several studies have shown the sclera to be permeable to a range of drugs and solutes (Ahmed and Patton, 1985, 1987; Schoenwald et al., 1997). Maurice and Polgar (1977) have reported that the sclera generally offers less resistance to solute diffusion than does the cornea. The permeability of the sclera to carbonic anhydrase inhibitors is similar to the permeability of the corneal stroma (Edelhauser and Maren, 1988). Hydrocortisone permeability is five times greater in sclera than in corneas with intact epithelia (Unlu and Robinson, 1998). Ahmed et al., 1987 compared conjunctival, scleral and corneal permeation of adrenergic blocking agents and found the scleral permeabilities for all drugs tested to be significantly higher than their corneal permeabilities. The sclera was also found to be more permeable than the cornea to polyethylene glycol (Ha¨ma¨la¨inen et al., 1977). Olsen et al. (1995) also investigated the permeability of the human sclera to solutes ranging in molecular weight from 350 to 70 000 Da, finding an
inverse relationship between scleral permeability and solute molecular weight. More recently, Ambati et al. (2000a,b) have shown the sclera to be permeable to higher molecular weight dextrans, as well as to the proteins IgG and bovine serum albumin. This group has also demonstrated that in vivo trans-scleral delivery is capable of maintaining significant levels of biologically active protein in the choroid and retina of the rabbit eye (Ambati et al., 2000a,b). The sclera is an elastic and microporous tissue composed of proteoglycans and closely packed collagen fibrils, (Foster and Sainz de la Maza, 1994; Newell, 1996) containing approximately 70% water. Solutes traverse the tissue through the interfibrillar aqueous media of the gel-like proteoglycans (Maurice and Mishima, 1984). While the cornea is relatively impermeable to solutes having a molecular size over 1 kDa, dextran (40 kDa) and serum albumin (69 kDa) can readily penetrate across the scleral tissue (Bill, 1965; Olsen et al., 1995). Molecular size is therefore more of a limiting factor in the diffusion of drugs across the sclera than the lipophilicity of the solute. In a previous study, we observed that elevated transscleral pressure reduced human scleral permeability to dexamethasone by as much as four-fold at 60 mmHg
Fig. 4. Rate of delivery of dexamethasone in BSS Plus (average steady-state value of 8·06 ^ 0·67 £ 1022 hr21) and in the two sustained release delivery systems—F-127 (average steady-state value of 1·80 ^ 0·03 £ 1022 hr21) and fibrin sealant (average steady-state value of 4·33 ^ 0·11 £ 1022 hr21). Values shown represent the mean ^ S.E . The average steady-state rates of delivery in F-127 and fibrin sealant were 22 and 54% of the BSS plus control value, respectively. These calculated steady-state rates of delivery were significantly different (p , 0·0001; One-way ANOVA).
Fig. 5. Percent cumulative release of 3H-dexamethasone through human sclera from BSS Plus and the two sustained release systems—F-127 and fibrin sealant. Values shown represent the mean ^ S.E . The cumulative release of 3H-dexamethasone through human sclera from BSS plus (control value) was 84·0 ^ 1·5% (n ¼ 5; 20 hr). The values in pluronic F-127 gel and fibrin sealant were 29·3 ^ 5·9% (n ¼ 5; 20 hr) and 61·4 ^ 5·9% (n ¼ 4; 20 hr), respectively. One-way Analysis of Variance showed these values to be significantly different ðp , 0·0001Þ:
4. Discussion
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
(Rudnick et al., 1999). We hypothesized that this lower permeability was due to compression of the sclera, which reduced its water content and thereby reduced the spaces between extracellular matrix molecules, which are the pathways through which dexamethasone molecules diffuse across the sclera. In the present study we measured changes in scleral thickness and hydration. We found that transscleral pressure had no significant effect on the hydration of the tissue over the range of pressures studied, up to 60 mmHg. In this study, control scleral hydration was 71·1 ^ 2·3% or 2·5 ^ 0·3 mg H2O mg21 dry weight. This value is slightly lower than the 2·9 ^ 0·2 mg H2O mg21 dry weight for scleral tissue obtained from two-day moist chamber-stored globes in Olsen’s report (1995). Our value agrees well with the hydration value of 71·8% in Madhu’s report (1998). Though our results do indicate that intraocular pressure can affect scleral thickness, we found that the effect of pressure on thickness was minimal at trans-scleral pressures up to 60 mmHg. The modest changes in thickness that do occur at higher pressures most likely occur by changes in the tissue’s microanatomy. Elevated pressure compressing the tissue, which in turn would reduce the spaces between the collagen fibers and extracellular matrix molecules that define the pathways for diffusion. However, the modest changes in thickness at high trans-scleral pressures are probably not sufficient to significantly affect rates of diffusion across the tissue, especially for small solutes such as dexamethasone. Pluronic F-127, a polyoxyethylene-polyoxypropylene surface-active block copolymer, with an average molecular weight of 11 500 Da (Miyazaki et al., 1984) has good potential for use as a sustained release delivery system since it has good drug release characteristics, low toxicity, and exhibits reverse thermal gelation (Schmolka, 1972). It remains in the liquid state at refrigerator temperatures and gels upon warming to physiological temperatures (Miyazaki et al., 1984). Thus, if the liquid form, containing drug, were injected into subconjunctival space, it would form a solid sustained release depot. Since F-127 gels are viscous isotropic liquid crystals consisting of micelles, (Chen-Chow and Frank, 1981) incorporated drugs would be released by diffusion through the extramicellar water channels of the gel matrix. Hence, the rate of drug release would be determined by the micro-viscosity of the extramicellar fluid, the size and tortuosity of the water channels, and the drug partitioning between the micellar phase and the external water phase. Chen-Chow and Frank (1981) and Miyazaki et al. (1984) studied the effect of F-127 concentration and temperature on the release of drug and found that the release rate decreased with increasing concentration of F-127. This was the result of a reduction in the size and increase in tortuosity of the water channels and an increase in the micro-viscosity of the water channels of the gel. Release rate increased with increasing temperature due to decreased
605
viscosity of water channels (Chen-Chow and Frank, 1981; Miyazaki et al., 1984). Fibrin sealant is a water-insoluble gel matrix formed when fibrinogen is activated by thrombin in the presence of Ca2þ ion and factor XIII. In recent years, fibrin sealant has become a well-known biomaterial because of its hemostatic, adhesive, and biocompatible properties (Yu et al., 1996). It has been successfully used in a wide range of surgical fields such as skin graft fixation, nerve repair, cartilage reattachment and microvascular anastomoses (Tashiro et al., 1987; Wood and Harner, 1988; Sierra et al., 1990; Byrne et al., 1991; Yu et al., 1996). Fibrin sealant also has the additional advantages of being biodegradable and commercially available. The results of this study indicate that fibrin sealant has potential for use as a sustained release drug delivery system. One approach to trans-scleral drug delivery would be to administer a drug, in a solution or suspension, by a periorbital, retrobulbar, or intrascleral injection. In this case, the rate at which the drug would diffuse into the eye would be determined principally by scleral permeability. Another approach would be to utilize a sustained release delivery system formed by a gel, a solid polymer, or another system that could be placed in contact with the outer surface of the sclera. Ideally, the drug could be released from the system at a controlled rate. In this case, the maximum drug delivery rate into the eye would be constrained by scleral permeability, but the device itself could control the rate of delivery if the barrier to diffusion imposed by the device were less than that of the sclera. Dexamethasone was chosen as a model compound since it is similar in molecular size and diffusivity to most conventional ocular drugs and it represents a reasonably hydrophilic drug that can be used to test the pathway of drug movement (Prausnitz and Noonan, 1988). The permeability of human sclera to dexamethasone in BSS plus measured in this study was 11·5(^ 1·1) £ 1026 cm sec21. This value is comparable to that reported by Rudnick et al. (1999) but is lower than that reported by Olsen et al. (1995) 23·5(^ 7·7) £ 1026 cm sec21. Different chambers, however, were used in these studies. It is possible that Olsen’s chamber design with constant chamber mixing may yield a higher scleral permeability measurement than that determined in this study which employed an unmixed depot on the surface of the sclera in which static boundary layers could form. The apparent permeability values of dexamethasone in F127 and fibrin sealant were 13 and 64% of BSS plus control value, respectively. These differences might be due to smaller pore sizes within the F-127 gel compared to that of the fibrin sealant. The F-127 gel was found to have a lower release rate and higher drug retention, resulting in a longer sustained release delivery by F-127 compared to fibrin sealant. There is an initial lag period in dexamethasone flux during which the flux sharply increases and then reaches a peak after approximately 3 hr for BSS plus, 4 hr for fibrin
606
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607
sealant, and 7 hr for F-127, Fig. 4. This lag time is due to both diffusion through and possible binding within the sclera. After reaching its peak, the flux slowly decreases. This is because the donor source of solute is finite and becomes depleted over time, thereby reducing the concentration gradient across the sclera. These lag times are longer than the residence time at the scleral surface for drug introduced by peribulbar injection. Therefore, the use of F127 and fibrin sealant should significantly extend the time scale of trans-scleral delivery and possibly increase the total amount of drug delivered to the eye. To summarize, we have established the initial feasibility of trans-scleral drug delivery using the sustained release systems of F-127 and fibrin sealant. These systems provided a slow sustained release through a 24 hr period. Additionally, scleral hydration was found not to change significantly over the range of trans-scleral pressures studied (0 – 60 mmHg). Scleral thickness did show a tendency to decrease as trans-scleral pressure increased to 60 mmHg. The results of this study suggest that sustained release drug delivery using systems such as F127 or fibrin sealant show promise as systems to deliver therapeutic levels of bioactive drugs to the posterior segment of eye. These compounds may be employed in the future development of a practical system or device for drug delivery across the sclera for the treatment of a variety of chorioretinal disorders.
Acknowledgements Funded in part by a grant from The Foundation Fighting Blindness, a grant from the Knights Templar Educational Foundation, by NEI P30 EY06360, and by an unrestricted grant from RPB, Inc.
References Ahmed, I., Gokhale, R.D., Shah, M.V., Patton, T.F., 1987. Physicochemical determinants of drug diffusion across the conjunctiva, sclera, and cornea. J. Pharm. Sci. 76, 583–586. Ahmed, I., Patton, T.F., 1985. Importance of the noncorneal absorption route in topical ophthalmic drug delivery. Invest. Ophthalmol. Vis. Sci. 26, 584 –587. Ahmed, I., Patton, T.F., 1987. Disposition of timolol and inulin in the rabbit eye following corneal versus non-corneal absorption. Int. J. Pharm. 38, 9–21. Ambati, J., Canakis, C.S., Miller, J.W., Gragoudas, E.S., Edwards, A., Weissgold, D.J., Kim, I., Delori, F.C., Adamis, A.P., 2000. Diffusion of high molecular weight compounds through sclera. Invest. Ophthalmol. Vis. Sci. 41, 1181–1185. Ambati, J., Gragoudas, E.S., Miller, J.W., You, T.T., Miyamoto, K., Delori, F.C., Adamis, A.P., 2000. Transscleral delivery of bioactive protein to the choroid and retina. Invest. Ophthalmol. Vis. Sci. 41, 1186–1191. Barza, M., Kane, A., Baum, J.L., 1977. Regional differences in ocular concentrations of gentamicin after subconjunctival and retrobulbar injection in the rabbit. Am. J. Ophthalmol. 83, 407–413.
Barza, M., Kane, A., Baum, J.L., 1978. Intraocular penetration of gentamicin after subconjunctival and retrobulbar injection. Am. J. Ophthalmol. 85, 541–547. Bill, A., 1965. Movement of albumin and dextran through the sclera. Arch. Ophthalmol. 74, 248–252. Byrne, D.J., Hardy, J., Wood, R.A.B., McIntosh, R., Cuschieri, A., 1991. Effect of fibrin glues on the mechanical properties of healing wounds. Br. J. Surg. 78, 841–843. Chen-Chow, P., Frank, S.G., 1981. Comparison of lidocaine release from pluronic F-127 gels and other formulations. Acta Pharm. 18, 239 –244. Chrai, S.S., Patton, T.F., Mehta, A., Robinson, J.F., 1973. Lacrimal and instilled fluid dynamics in rabbit eyes. J. Pharm. Sci. 62, 1112–1121. Edelhauser, H.F., Maren, T.H., 1988. Permeability of human cornea and sclera to sulfonamide carbonic anhydrase inhibitors. Arch. Ophthalmol. 106, 1110–1115. Edwards, A., Prausnitz, M.R., 1998. A fiber matrix model of sclera and corneal stroma for drug delivery to the eye. AIChE J. 44, 214–225. Foster, C.S., Sainz de la Maza, M., 1994. The Sclera, Springer, New York, NY, pp. 1 –32. Freeman, W.R., 1989. Intraocular antiviral therapy. Arch. Ophthalmol. 107, 1737–1739. Ha¨ma¨la¨inen, K.M., Kananen, K., Auriola, S., Kontturi, K., Urtti, A., 1977. Caracterization of paracellular and aqueous penetration routes in cornea, conjunctiva, and sclera. Invest. Ophthalmol. Vis. Sci. 38, 627 –634. Lang, J.C., 1995. Ocular drug delivery: conventional ocular formulations. Adv. Drug Deliv. Rev. 16, 39–43. Madhu, C., Rix, P., Nguyen, T., Chein, D., Woodward, D.F., Diane, D.S., Tang-Liu, D.D.S., 1998. Penetration of natural prostaglandins and their ester prodrugs and analogs across human ocular tissues in vitro. Ocul. Pharmacol. Ther. 14, 389–399. Maurice, D.M., 2002. Drug delivery to the posterior segment from drops. Surv. Ophthalmol. 47, S41–S51. Maurice, D.M., Mishima, S., 1984. Ocular pharmacokinetics. In: Sears, M.L. (Ed.), Pharmacology of the Eye, Springer, New York, NY, pp. 19 –116. Maurice, D.M., Polgar, J., 1977. Diffusion across the sclera. Exp. Eye Res. 25, 577–582. Miyazaki, S., Takeuchi, S., Yokouchi, C., Takada, M., 1984. Pluronic F127 gels as a vehicle for topical administration of anticancer agents. Chem. Pharm. Bull. 32, 4205–4208. Newell, F.V., 1996. Ophthalmology, Principles and Concepts, Sixth Ed, CV Mosby Co, St Louis, MO. Olsen, T.W., Aaberg, S., Geroski, D.H., Edelhauser, H.F., 1998. The human sclera: thickness and surface area. Am. J. Ophthalmol. 125, 237– 241. Olsen, T.W., Edelhauser, H.F., Lim, J.I., Geroski, D.H., 1995. Human scleral permeability: effects of age, cryotherapy, trans-scleral diode laser, and surgical thinning. Invest. Ophthalmol. Vis. Sci. 36, 1893–1903. Prausnitz, M.R., Noonan, J.S., 1998. Permeability of cornea, sclera, and conjunctiva: a literature analysis for drug delivery to the eye. J. Pharm. Sci. 87, 1479–1488. Rudnick, D.E., Noonan, J.S., Geroski, D.H., Prausnitz, M.R., Edelhauser, H.F., 1999. The effect of intraocular pressure on human and rabbit scleral permeability. Invest. Ophthalmol. Vis. Sci. 40, 3054–3058. Schmolka, I.R., 1972. Artificial skin. I. Preparation and properties of pluronic F-127 gels for treatment of burns. J. Biomed. Mater. Res. 6, 571 –582. Schoenwald, R.D., Deshpande, G.S., Rethwisch, D.G., Barfknecht, C.F., 1997. Penetration into the anterior chamber via the conjunctival/scleral pathway. J. Ocul. Pharmacol. 13, 41 –59. Sierra, D.H., Nissen, A.J., Welch, J., 1990. The use of fibrin glue in intracranial procedures: preliminery results. Laryngoscope 100, 360 –363. Tashiro, S., Murata, E., Hiraoka, K., Nakakuma, K., Watanabe, E., Miyauchi, Y., 1987. New technique for pancraticojejunostomy using a biological adhesive. Br. J. Surg. 74, 392– 394.
S.-B. Lee et al. / Experimental Eye Research 78 (2004) 599–607 Unlu, N., Robinson, J.R., 1998. Scleral permeability to hydrocortisone and mannitol in the albino rabbit eye. J. Ocul. Pharmacol. 14, 273–281. Watsky, M.A., Olsen, T.W., Edelhauser, H.F., 1995. Cornea and sclera. In: Tasman, W., Jaeger, E.A. (Eds.), Cornea and Sclera, Lippincott/Raven, Philadelphia, PA, pp. 1–29.
607
Wood, A.P., Harner, S.G., 1988. The effect of fibrin tissue adhesive on the middle and inner ears of chinchillas. Otolaryngol. Head Neck Surg. 98, 104– 110. Yu, B.G., Kwon, I.C., Kim, Y.H., Han, D.K., Park, K.D., Han, K., Jeong, S.Y., 1996. Development of a local antibiotic delivery system using fibrin glue. J. Control. Release 39, 65–70.