The Influence of Intraocular Pressure on the Transscleral Diffusion of High-Molecular-Weight Compounds

Abstract

purpose. To evaluate the effects of intraocular pressure on the permeability of the human sclera to high-molecular-weight compounds.

methods. Human transscleral permeability to FITC-albumin (70 kDa) and 70-kDa and 150-kDa FITC-dextran was determined at transscleral pressures from 0 to 60 mm Hg. For each compound at each pressure, six to eight experiments were performed. Scleral sections were mounted in a two-compartment perfusion chamber. Temperature was maintained at 37°C. Fractions of choroidal perfusate were collected, and fluorescence was measured with a spectrofluorometer. From these data, scleral permeability K _trans (in centimeters per second) was calculated.

results. Permeability to FITC-albumin was decreased by approximately one half when pressure was elevated from 0 to 30 mm Hg (P < 0.05). No significant differences in permeability to 70-kDa FITC-dextran were observed at pressures from 0 to 60 mm Hg. Permeability to 150-kDa FITC-dextran decreased by a little more than one half when transscleral pressure was raised from 0 to 15 mm Hg and was approximately 10 times lower at 60 mm Hg than at 0 mm Hg (P < 0.01).

conclusions. Human sclera was permeable to compounds with a molecular weight of up to 150 kDa at transscleral pressures ranging from 0 to 60 mm Hg. Transscleral diffusion was relatively unaffected by the pressure gradient, although for 150-kDa FITC-dextran at 60 mm Hg a 10-fold decrease was observed compared with that at 0 mm Hg. These experiments suggest that high-molecular-weight compounds (e.g., immunoglobulins and oligonucleotides) could be effectively delivered transsclerally to the intraocular tissues under circumstances of physiological or elevated intraocular pressure.

Finding alternative routes for drug delivery into the eye has been the focus of interest in recent years. ^{1 2 3 4 5} New and promising therapies are being developed for a variety of posterior segment diseases, including neuroprotective agents and antiangiogenic drugs. ^{6 7} Current ocular drug delivery methods all come with limitations for posterior segment drug delivery. ¹ Topical drops bring only a small amount of drug into the eye and are insufficient for drug delivery to the posterior segment. ^{8 9} Systemic treatment of ocular disease is limited by dose toxicity and systemic side effects. In addition, the blood–retina barrier provides an anatomic and physiological protection to the visual pathway from exogenous substances, including many drugs. Intraocular injections and intravitreal sustained release devices may lead to complications associated with intraocular surgery (retinal detachment, hemorrhage, endophthalmitis, and cataract), and the injections often must be repeated. Furthermore, the internal limiting membrane of the retina seems to provide a permeability barrier to solutes larger than 4 to 5 nm. Depending on the molecule’s shape, this size corresponds to solutes with molecular weights in the 40,000 to 70,000 range. ^{10 11 12 13} Many of the newly developed agents to treat chorioretinal disorders are high-molecular-mass compounds, with molecular masses ranging from 17 kDa for certain neuroprotective growth factors to 150 kDa for anti-VEGF antibodies to inhibit angiogenesis. ¹⁴  

Drug delivery across the sclera would avoid the risks inherent in invasive intraocular alternatives and also would offer the potential for localized sustained-release drug delivery. An additional benefit of the transscleral route is the hypocellularity of the human sclera, which provides an environment with few proteolytic enzymes and protein-binding sites that could interfere with drug diffusion. Furthermore, drug uptake by the sclera itself could be an additional factor in prolonging intraocular drug release over an extended period. 

Previous studies ^{4 15 16} have shown scleral permeability to be dependent on solute size, solubility, and molecular conformation. Human scleral tissue has been shown to be permeable to 70-kDa dextrans ⁴ and rabbit sclera is permeable to 150-kDa dextrans and IgG of the same molecular weight. ⁵  

Transscleral (intraocular) pressure may also affect scleral solute permeability. Rudnick et al. ³ demonstrated that the scleral permeability to small molecules (molecular weight, 18–392) at transscleral pressures ranging from 15 to 60 mm Hg was decreased by approximately one half when compared with scleral tissue with no pressure applied. The permeability of the sclera to larger molecules as a function of pressure has not been reported. 

In this study, we evaluated the effect of intraocular pressure on the transscleral permeability of FITC-albumin, a molecule with a globular structure and a molecular weight of 70,000, and two FITC-dextrans with linear molecular configurations and molecular weights of 70,000 and 150,000, respectively. 

Methods

Human transscleral diffusion of a 1.0 × 10⁻⁴-M solution (in physiologic saline; BSS, Alcon Laboratories, Inc. Fort Worth, TX) of FITC-albumin (70 kDa) or 70-kDa or 150-kDa FITC-dextran (Sigma-Aldrich, St. Louis, MO) was evaluated in vitro at transscleral pressures of 0, 15, 30, and 60 mm Hg. 

Tissue Preparation

Human donor globes were obtained from the Georgia Eye Bank (Atlanta, GA). Scleral sections were excised from moist-chamber–stored human globes. For the tissue used in this study, mean storage time (±SE) was 5.4 ± 0.2 days, mean donor age at time of death (±SE) was 57.9 ± 1.8 years. Past studies have shown that human scleral water content does not change significantly after moist-chamber storage for up to 10 days and that balanced saline (BSS; Alcon laboratories, Fort Worth, TX) will maintain tissue hydration during in vitro studies. ⁴ The superior–temporal area of the sclera, posterior to the equatorial region, was used to avoid the ciliary vessels and the vortex veins. Adherent tissue (retina, choroid, episclera, Tenon’s capsule) was gently removed with a cotton-tip applicator to isolate the bare sclera. 

Experimental Design

The scleral sections were mounted horizontally, vitreous side down, in an Ussing-type perfusion chamber that permits the imposition of a transscleral pressure that simulates intraocular pressure (Fig. 1) . The sclera divided the perfusion chamber into an upper (donor) and a lower (receiver) compartment. The volume of the donor chamber, which serves as a depot, was approximately 600 μL. Receiver chamber was 500 μL. The sclera was clamped between two 2.5-mm wide and approximately 1-mm thick cylindrical silicone elastomer rings (Dow Corning, Inc., Midland, MI) to prevent lateral leakage and scleral edge damage. The temperature of the water-jacketed perfusion chamber was kept at 37°C by a circulating water bath. The transscleral pressure was varied from 0 to 60 mm Hg by adjusting the height of the outflow tubing above the chamber. A magnetic stirrer was placed in the receiver chamber to ensure adequate mixing. 

Fraction Collection and Measurements

A small depot of each compound (500 μL of a 1.0 × 10⁻⁴-M solution in balanced saline) was added to the episcleral surface. For each solute pressure studied, six to eight experiments were performed. The saline solution was perfused slowly through the receiver chamber at a rate of 0.03 mL/min. The donor chamber was held static, modeling the situation in which a drug added to Tenon’s space would be exposed directly to the sclera. The perfusate was collected as 1-hour fractions using a fraction collector (Isco, Lincoln, NE), and the fluorescence was measured in each fraction with a spectrofluorometer (Fluoro IV; Gilford, Oberlin, OH). Wavelengths for excitation and emission were 493 and 520 nm, respectively. Transscleral drug diffusion (in moles) was calculated from the fluorescence measurements. 

Fluorescent Compounds

The molecular weight of FITC-albumin was 70 kDa with a molecular radius of 3.62 nm. ⁵ The FITC-dextrans used in this study had molecular masses of 70 and 150 kDa, with molecular radii of 6.4 and 8.25 nm, respectively. ⁵ Maximum solubility (as provided by the manufacturer; Sigma-Aldrich) for FITC-albumin was 10 mg/mL, whereas solubility was more than 25 mg/mL for 70- and 150-kDa FITC-dextran. 

Scleral Permeability

The scleral permeability was expressed as K _trans, which is specific for a certain solute diffusing through the sclera. Mean steady state permeabilities (K _trans± SE) were calculated from the experimental data as  

\[K_{\mathrm{trans}}\ {=}\ \frac{R_{\mathrm{total}}}{(t)(A)}\ {\times}\ \frac{1}{D},\]

where R _total is the total amount (in moles) of drug that diffused through the sclera over a given interval of time t (in seconds). A is the area of exposed sclera (square centimeters), 0.385 cm² in these experiments. R _total /(t)(A) is equal to the flux across the tissue. D is the concentration of drug in the donor chamber. Permeability was calculated over the time during which solute flux across the sclera remained at its maximum, which typically occurred during the 3- to 5-hour period, before any significant depletion of the donor solution could occur. K _trans, the permeability constant, thus represents the steady state flux normalized by donor concentration. 

Statistics

Statistical analysis was performed (ANOVA; Tukey-Kramer multiple comparisons test) to compare the K _trans level of each compound at the different transscleral pressures. An unpaired t-test was performed to compare different solutes at equal pressures. 

Results

Figure 2Ashows the transscleral permeability to 70-kDa FITC-albumin with transscleral pressures applied of 0, 15, 30, and 60 mm Hg. Mean (±SE, × 10⁻⁶ cm/s) permeabilities were 2.38 ± 0.45, 1.43 ± 0.19, 1.27 ± 0.13, and 1.25 ± 0.20, respectively. Diffusion of FITC-albumin was highest when no pressure was applied (0 mm Hg). Raising the transscleral pressure resulted in lower transscleral diffusion at each 15-mm Hg pressure increase. Scleral permeability to this molecule, however, was only significantly (ANOVA, P < 0.05) lower at 30 and 60 mm Hg, compared with 0 mm Hg. No significant (P > 0.05) differences were observed when the pressure was raised from 0 to 15 mm Hg and from 15 to either 30 or 60 mm Hg. 

For 70-kDa FITC-dextran (Fig. 2B) , permeability appeared to decrease as pressure was raised from 0 to 15 mm Hg. Changes in permeability with pressure, however, were not statistically significant (ANOVA P > 0.05) for this solute. Mean (±SE, × 10⁻⁸ cm/sec) permeabilities were 8.17 ± 1.20, 5.55 ± 0.64, 4.65 ± 0.79, and 5.02 ± 0.79 at pressures of 0, 15, 30, and 60 mm Hg, respectively. 

Figure 2Cshows the permeability to 150-kDa FITC-dextran at each pressure studied. For pressures of 0, 15, 30, and 60 mm Hg the permeability (mean ± SE, × 10⁻⁸ cm/s) was measured as 9.66 ± 2.17, 3.66 ± 1.23, 2.11 ± 0.56, and 0.96 ± 0.19, respectively. For the 150 kDa FITC-dextran transscleral diffusion was significantly higher at 0 mm Hg compared to 15 (P < 0.05), 30 (P < 0.01), and 60 (P < 0.01) mm Hg. For this solute, transscleral permeability decreased significantly at 15 mm Hg of pressure, no further significant decrease in permeability was observed at the higher pressures. 

After the completion of the experiments, 5% to 13% of FITC-albumin had diffused across the sclera. For both 70-kDa and 150-kDa FITC-dextran, <1% diffused in a 24-hour period. Human scleral permeability to 70-kDa FITC-albumin was significantly higher than to 70-kDa FITC-dextran (P < 0.001) at each applied transscleral pressure. On average this difference between permeability to the two compounds was approximately a factor of 25. Transscleral permeability of 70-kDa FITC-dextran was comparable to that of 150-kDa FITC-dextran at pressures of 0 and 15 mm Hg. However, at pressures of 30 and 60 mm Hg, the permeability was two- (P < 0.05) to fivefold (P < 0.001) lower for 150-kDa FITC-dextran. 

Discussion

This study demonstrated that human sclera is permeable to solutes with a molecular weight of up to 150,000. Previously, rabbit sclera was also shown to be permeable to 150-kDa FITC-dextran. ⁵ In addition, we have shown that compounds with a molecular masses of 70 to 150 kDa permeate the human sclera at transscleral pressures ranging from 0 to 60 mm Hg. The permeability of human sclera to 70-kDa FITC-albumin and 70- and 150-kDa FITC-dextran at 0 mm Hg were 2.38 ± 0.45 × 10⁻⁶, 8.17 ± 1.20 × 10⁻⁸, and 9.66 ± 2.17 × 10⁻⁸ cm/s, respectively. These values are lower than those reported by Ambati et al. ⁵ (5.49 ± 2.12 × 10⁻⁶, 1.39 ± 0.88 × 10^-6, 1.34 ± 0.88 × 10⁻⁶, respectively) and Olsen et al. ⁴ (1.9 ± 0.4 × 10⁻⁶ for 70-kDa FITC-dextran). The lower permeability values found in the present study might reflect species-related variables in scleral thickness and structure, ¹⁷ as well as differences in the experimental design (perfusion apparatus, applied transscleral pressure, stirred receiver chamber, etc.). In addition, a perfusion chamber (as used by Olsen et al. ⁴ ) with constant mixing of the donor compartment may yield a higher permeability measurement than when, as in this study, an unmixed depot was placed on the surface of the sclera in which static boundary layers could form. 

In this study, no further measurements were performed to determine whether any separation of the fluorescent tags to the molecules occurred. However, several studies of fluorescence-labeled dextrans indicate that such separation is virtually nonexistent. ¹⁸ Also, in studies of a similar design protein precipitation of samples containing FITC-albumin showed no dissociation of the FITC conjugate. ⁵  

To compare the permeability data of the solutes tested at different transscleral pressure gradients, we calculated the permeability, K _trans (mean of a series of experiments of each compound at each pressure). K _trans, a drug-specific measure of diffusion through the sclera is related to the structure, surface, and thickness of the sclera and the diffusional characteristics of the drug itself. These results showed that the intraocular pressure moderately affected the transscleral permeability of the compounds tested. Only when pressure was raised from 0 to 30 and from 0 to 60 mm Hg was human scleral permeability to 70-kDa FITC-albumin significantly lower. The differences in permeability for this compound between 0 and 30 and 0 and 60 mm Hg were roughly a factor of two, comparable to the study of Rudnick et al. ³ for low-molecular-weight compounds. Transscleral pressures between 0 and 60 mm Hg did not significantly alter the scleral permeability to 70-kDa FITC-dextran. Permeability to 150-kDa FITC-dextran decreased by a little more than one half when transscleral pressure was raised from 0 to 15 mm Hg and was about 10 times lower at 60 mm Hg than at 0 mm Hg. These results showed that the permeability of human sclera to high-molecular-weight compounds was affected by the transscleral pressure gradient, but this effect was rather small. Although the maximum difference between 0 and 60 mm Hg for the largest compound, 150-kDa FITC-dextran, was roughly a factor of 10, for the other compounds and pressures this was closer to a factor of 2 or less. 

Factors that can influence the permeability of the sclera include molecular size and radius of the solute and the intraocular pressure. ^{3 4 5} Transscleral diffusion presumably occurs by diffusion through an interfibrillar aqueous media of gellike proteoglycans rather than by diffusion through cellular membranes or pores. ¹⁹ The sclera is an elastic and microporous tissue, composed of proteoglycans and closely packed collagen fibrils, and is composed of approximately 70% water. All compounds that were tested in this study were hydrophilic molecules, with a higher water solubility for the FITC-dextran compounds. 

Maurice and Polgar ¹⁵ and Olsen et al. ⁴ showed that the transscleral diffusion of ions and solutes was inversely related to their molecular weight. Similar to rabbit sclera, ⁵ human sclera was more permeable to the globular FITC-albumin than to the linear FITC-dextran molecule of the same molecular weight, despite the higher water solubility of FITC-dextran. This supports the assumption that the molecular radius has a greater role in determining scleral permeability than molecular weight or charge, similar to diffusion through extracellular tissue in the brain. ²⁰ It has also been reported that scleral permeability declines roughly exponentially with molecular radius. ^{5 21 22}  

Possible pathways for transscleral diffusion are through the loose connective tissue around the blood vessels and nerves as well as through the scleral stroma itself. ^{4 23} The diffusion through the scleral stroma is most likely through the interfibrillar aqueous media of the gellike proteoglycans. ¹⁹ Rudnick et al. ³ demonstrated, using the Peclet number analysis, that the elevated pressure reduces the transscleral diffusion by altering the microanatomy of the sclera, rather than by counteraction hydrostatic flow. Narrowing of the intracollagen pathways by compression of the scleral fibers due to an increase of intraocular pressure is thought to be responsible for the decrease in permeability. 

The measured permeability of the human sclera to 150-kDa FITC-dextran was not significantly lower than the permeability to 70-kDa FITC-dextran at 0 and 15 mm Hg, despite the difference in molecular weight. Although the molecular weight difference is more than a factor of two, the radius of 150-kDa FITC-dextran is <30% larger than that of 70-kDa FITC-dextran. It has also been suggested by Ambati et al. ⁵ that this is because the hydrodynamic radii of these molecules within the scleral tissue is not identical with their radii in aqueous solution. These authors compare this phenomenon with a similar situation that exists in the brain tissue, where the diffusion of 40- and 70-kDa FITC-dextran was not significantly different. ²⁴  

Transscleral diffusion is also influenced by the thickness of the sclera, which ranges from 0.53 ± 0.14 mm (mean ± SD) near the limbus to 0.39 ± 0.17 mm at the equator in human eyes. ²⁵ Its thickness increases to 0.9 to 1.0 mm near the optic nerve. Thinning of the sclera increases the permeability ⁴ ; thus, transscleral delivery of a drug would be facilitated by placing a depot of drug near the equator of the eye where the sclera is the thinnest. In the present study, scleral sections were all taken from the superotemporal area of the globe, just posterior to the equator, to avoid any obvious differences in scleral thickness in the experiments. It is also reported that the physicochemical parameters of the sclera, like thickness, might be of greater significance for larger molecules than for smaller compounds. ²⁵ Lee et al. ²⁶ demonstrated a decrease of scleral thickness with increasing pressure, with a significant decrease at 60 mm Hg. However, this effect of pressure on the scleral thickness appeared to be low. Scleral sections in this study were taken just posteriorly from the equator, as described by Lee et al. Scleral thickness (micrometers ± SE) was measured to be 600 ± 49 at 0 mm Hg, 595 ± 46 at 15 mm Hg, 599 ± 47 at 30 mm Hg, and 573 ± 43 at 60 mm Hg, with the same perfusion apparatus used as in the present study. Modest changes in thickness at high transscleral pressures were not assumed to have a significant effect on rates of diffusion across the tissue for solutes with a low molecular weight. In addition, the hydration of the scleral tissue did not appear to be significantly affected at transscleral pressures from 0 to 60 mm Hg. Indeed, in our study we found a tendency of lower permeability values at higher pressure rates, making it feasible that compression of the diffusional pathways in the sclera would be of more importance than thinning of the sclera and thereby shortening the diffusional distance. An increase in pressure affected the scleral permeability to 150-kDa FITC-dextran more than the transscleral diffusion of 70-kDa FITC-dextran. It is possible that the molecular radius of the former, being closer to the critical limitation of the diffusional pathway, would be more affected by an increase in pressure and consequent narrowing of the pathway. 

Drug diffusion through the sclera holds potential for drug delivery to the posterior segment. It can be performed as a minimally invasive procedure and thereby minimizes the risk of complications related to intraocular alternatives. The sclera accounts for 95% of the total surface area of the globe ²⁵ with an average surface area of 16.3 cm², making it a suitable entry route for localized drug delivery away from the optical axis. High-molecular-weight compounds that would not be able to reach the chorioretinal tissues after intravitreal administration due to the barrier provided by the internal limiting membrane diffuse readily through human scleral tissue. The transscleral pathway is hypocellular and shows few hindrances to diffusion. In addition, scleral permeability does not decrease with age, and scleral and surrounding tissues have a notable tolerance to foreign bodies (i.e., scleral buckles) making it a feasible site for depot drug delivery. 

These scleral diffusion studies were performed with simulated intraocular pressures from 0 to 60 mm Hg and at a temperature of 37°C. Other determining factors such as orbital clearance, choroidal blood flow, and the blood–retina barriers are likely to influence transscleral drug delivery. ²⁷ Recently, Pitkanen et al. ²⁸ demonstrated the retinal pigment epithelium to be a large barrier to solute flux to the posterior segment. However, several in vivo studies, showed that the transscleral route might be a viable pathway for drug delivery to the eye. ^{14 29 30 31}  

In summary, these experiments suggest that transscleral delivery of high-molecular-weight compounds to the intraocular tissues is relatively unaffected by the pressure gradient. Transscleral drug delivery of compounds with a high molecular weight appears to be a promising pathway for intraocular drug delivery under circumstances of physiological or elevated intraocular pressure. 

 Supported in part by a grant from Foundation Fighting Blindness, National Eye Institute Grant P30 EY06360, Research to Prevent Blindness, and Landelijke Stichting voor Blinden en Slechtzienden, The Netherlands.

 Submitted for publication December 3, 2004; revised April 29 and June 2, 2005; accepted August 2, 2005.

 Disclosure: L.P.J. Cruysberg, None; R.M.M.A. Nuijts, None; D.H. Geroski, None; J.A. Gilbert, None; F. Hendrikse, None; H.F. Edelhauser, None

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

 

Figure 1.

 

Schematic diagram of the perfusion chamber. The sclera was mounted horizontally between two compartments. The compound to be tested was added to the episcleral (top) compartment. The uveal (bottom) compartment had a continual flow and was mixed by a magnetic stir bar. Intraocular pressure was simulated by creating a positive water column with the outflow tube.

Figure 2.

 

Transscleral diffusion of (A) FITC-albumin (n = 6 for all pressures), (B) 70-kDa FITC-dextran (n = 8 for experiments performed at 0 mm Hg; n = 6 for experiments at 15, 30, and 60 mm Hg), and (C) 150-kDa FITC-dextran (n = 8 for experiments performed at 0, 15, and 30 mm Hg; n = 6 for experiments at 60 mm Hg). K _trans as a function of transscleral pressure. Data are the mean ± SE. Permeability to FITC-albumin decreased by roughly 50%, when pressure was raised from 0 to 60 mm Hg. Statistical analysis showed no significant differences in permeability to 70-kDa FITC-dextran at the different pressures. Transscleral diffusion of 150-kDa FITC-dextran decreased by >50% when pressure was raised from 0 to 15 mm Hg and was approximately 10 times lower at 60 mm Hg than when no pressure was applied.