Enhanced FGF-2 Movement through Human Sclera after Exposure to Latanoprost

Abstract

purpose. To determine whether exposure of sclera to latanoprost acid alters transscleral permeation by FGF-2.

methods. Pieces of human sclera were isolated from donor eyes after death, placed in organ culture, and exposed to 50 to 200 nM latanoprost acid or vehicle for 3 days. Transscleral permeability was then assessed by placing each scleral piece into a Ussing apparatus and measuring the amount of FGF-2 that moves from the orbital side to the uveal side of the scleral piece. Transscleral permeation by 10-kDa tetramethylrhodamine-dextran also was determined, for comparison.

results. Transscleral permeation by FGF-2 through sclera that had been incubated with vehicle was 1.53 ± 0.86 × 10⁻⁸ cm/sec. Transscleral permeation by 10-kDa tetramethylrhodamine-dextran was 1.04 ± 0.39 × 10⁻⁶ cm/sec. FGF-2 permeation of sclera exposed to 50, 100, and 200 nM latanoprost acid was increased by an average of 48% ± 62%, 100% ± 108%, and 108% ± 79%, respectively, compared with sclera exposed to vehicle (n= 13; P < 0.05). Scleral permeation by 10-kDa dextran after exposure to 50, 100, or 200 nM latanoprost acid was significantly increased by 42% ± 36%, 59% ± 51%, and 65% ± 49%, respectively (n = 14; P < 0.05). The ratio of dextran to FGF-2 permeation was approximately 90 and did not vary with 50, 100, or 200 nM latanoprost acid (P= 0.93, ANOVA).

conclusions. Exposure of sclera to latanoprost acid increases transscleral permeation by FGF-2 in human scleral organ cultures. Because this increase parallels the increased scleral permeability caused by dextran, it may reflect a general enhancement of permeability, a possibility that future in vivo studies should explore.

Recent investigations have demonstrated the potential for treating several diseases of the posterior segment with macromolecules. For example, intravitreal injection of brain-derived neurotrophic factor (BDNF) or ciliary neurotrophic factor (CNTF) prolongs retinal ganglion cell survival after optic nerve transsection. ^{1 2} Intravitreal injection of fibroblast growth factor-2 (also known as basic fibroblast growth factor [bFGF], or FGF-2) reduced photoreceptor loss that normally occurs with aging in Fisher 344 rats. ³ In addition, intravitreal injection of a chimeric protein containing VEGF receptor was shown to inhibit retinal neovascularization that occurs after retinal ischemia. ^{4 5} However, repeated intravitreal injections are not practical for clinical treatment, and transcorneal penetration after topical treatment is limited for molecules larger than 10 kDa. ⁶  

Previous studies have shown that there is only limited permeation by macromolecules through isolated human sclera obtained after death. ⁷ The rate of scleral permeation by labeled dextrans is dependent on size (molecular weight). ⁷ Recently, transscleral permeation by immunoglobulin (150 kDa) also has been demonstrated. ^{8 9} However, it is unknown whether the degree of scleral permeability would permit transscleral delivery of sufficient amounts of externally applied growth factors to enhance neuronal survival in the retina or to exert other desired biological effects. Therefore, a method to enhance the transscleral delivery of growth factors could be beneficial. 

The possibility that prostaglandins (PGs) or PG analogues could enhance transscleral permeability is suggested by several observations. First, normal human sclera contains detectable amounts of matrix metalloproteinase (MMP)-1, ¹⁰ a neutral extracellular proteinase that can initiate the degradation of fibrillar collagen. ¹¹ This suggests that collagen turnover is an ongoing process in sclera. Next, topical treatment of monkey eyes with PGF_2α-isopropyl ester (PGF_2α-IE) has been shown to increase uveoscleral outflow (which passes through the ciliary muscle and sclera) and to reduce the amount of collagens present in both the ciliary muscle and sclera. ¹² In addition, exposure to PGs increased MMP biosynthesis and release in cultures of human ciliary muscle cells and scleral tissue culture. ^{13 14 15} Because fibrillar collagens constitute a major portion of scleral structure, ^{16 17 18 19 20 21} it is possible that reduction of scleral fibrillar collagen induces increased scleral permeability. Recently, we found that PGs and their analogues enhance transscleral permeation by dextran. ¹⁴ These observations suggest that PGs may be used not only as anti-glaucoma agents to reduce intraocular pressure, but also to facilitate delivery of other drugs to the posterior pole by increasing transscleral permeability. In the present study we investigated whether latanoprost, widely used as an antiglaucoma drug to reduce intraocular pressure, also can increase transscleral permeation by a growth factor known to promote the survival of retinal neurons. 

A good candidate for such a survival-promoting growth factor is FGF-2. Addition of FGF-2 to retinal neuron cultures can support the survival of the neurons. ²² Recently, protective effects of FGF-2 for retinal neurons and photoreceptors in vivo have been reported after retinal ischemia, optic nerve transsection, and phototoxic insult and in animals with inherited retinal degenerations. ^{23 24 25 26 27} The present study was undertaken to investigate the possibility that latanoprost can facilitate the transscleral permeation by FGF-2. To evaluate and compare the mechanisms of permeability changes, the transscleral movement of labeled dextran was assessed at the same time. 

Materials and Methods

Human Scleral Tissue Culture

Nine pairs of human eyes were obtained from the San Diego Eye Bank. Donors had no history of glaucoma or other ocular diseases. The mean age was 66.8 ± 13 years (mean ± SD; Table 1 ). Of the nine donors, seven were white, one was Hispanic, and one was of unspecified race. Each pair of eyes was enucleated within 5 hours after death and immediately preserved in a moist chamber at 4°C. Eyes that were apparently intact were selected, and any eye showing scleral damage or thin sclera, such as that caused by posterior staphyloma, was not used. Within 24 hours after preservation, the sclera was dissected and placed into organ culture. ¹⁴ Briefly, the eyes were first incubated in Hank’s buffered saline solution (HBSS) containing 50 U/ml penicillin and 50 U/ml streptomycin for 30 minutes. Next, residual extraocular muscles and orbital connective tissues were removed, and the sclera was dissected into four pieces, excluding the long ciliary nerve and artery, insertion of muscles, and vortex veins in each center area. Uveal tissue and retina were gently removed from the sclera with a cotton-tipped applicator. Scleral pieces were placed into 12-well plates containing DMEM-F-12 supplemented with 1% FCS and 1 ng/ml human recombinant FGF-2 (17-kDa; Gibco BRL, Grand Island, NY). The low concentration of serum was used to minimize a nonspecific increase in MMP, because serum contains various factors that stimulate MMP synthesis. The explants were incubated at 37°C in a humidified atmosphere of 95% air and 5% CO₂. 

Latanoprost Treatment

To investigate the effect of latanoprost on scleral permeability, the culture medium was changed to fresh medium supplemented with latanoprost acid (Cayman Chemical Co., Ann Arbor, MI). Latanoprost acid is the biologically active form of the prodrug latanoprost. After topical application, latanoprost is hydrolyzed to latanoprost acid as it passes into the anterior chamber. ²⁸ Tested concentrations of latanoprost acid included 50, 100, and 200 nM, because the peak concentration in human aqueous humor after topical application of a clinical dose of latanoprost is approximately 100 nM. ²⁹ Treatment duration was 3 days. 

Previous analysis of cell viability in control human scleral organ cultures with ethidium homodimer showed that viability was 83% on day 1, 81% on day 2, and 80% on day 3. ¹⁴ Moreover, the addition of 500 nM latanoprost acid altered these viability results by less than 1%. ¹⁴  

Scleral Permeability Analysis

After 3 days’ incubation, the scleral tissue was clamped into a Ussing apparatus (model CHM2; World Precision Instruments, Inc., Sarasota, FL). The Ussing apparatus consisted of two chambers with 0.75-ml capacity. The two conical chambers had 9-mm diameter openings facing each other and were held together by a screw clamp. The scleral tissue was washed three times in phenol red–free HBSS (PF-HBSS; 16 kDa; R&D Systems, Minneapolis, MN) to remove culture medium and was carefully sandwiched by the chambers to avoid vortex veins at the openings. Each chamber had three ports through which to fill and drain samples. The orbital-side chamber was filled with 20 nM FGF-2 dissolved in PF-HBSS and the uveal-side chamber was filled with plain PF-HBSS. After filling, the Ussing apparatus was incubated at 37°C. In the first group of experiments, 100 μl was removed from the uveal-side chamber and 100 μl of PF-HBSS was replaced every 20 minutes. In the second group of experiments, the Ussing apparatus was incubated undisturbed for 2 hours before the entire contents of the uveal-side chamber were collected. 

After completion of these 2-hour FGF-2 incubations, the chambers were rinsed with PF-HBSS, and the orbital-side chamber was refilled with 0.8 mg/ml 10-kDa tetramethylrhodamine-dextran (Molecular Probes, Eugene, OR), dissolved in PF-HBSS, and the uveal-side chamber was refilled with plain PF-HBSS. The refilled Ussing apparatus was again incubated for 2 hours at 37°C and the uveal-side fluid was collected. This subsequent analysis was included to determine whether changes in FGF-2 permeation were similar to changes in dextran permeation as well as to compare the results of the present studies with those of previous investigations of transscleral dextran permeation. ^{7 14}  

After draining FGF-2 samples from the uveal-side chamber, samples were stored at −20°C for later determinations of FGF-2 concentration. The dextran samples were drained from the uveal chamber and placed in a light-protected box, and the concentration of dextran in these samples was immediately determined. To evaluate the integrity of the scleral extracellular matrix, both vehicle-treated and latanoprost-treated cultures from three donors were fixed in 4% formaldehyde, embedded in paraffin, sectioned, and stained using Masson’s trichrome stain and Alcian blue (pH 2.5). Microscopic evaluation showed well-preserved lamellar scleral structure and similar staining intensity for both stains in both the vehicle-treated and latanoprost-treated cultures (data not shown). 

Measurement of Dextran

Tetramethylrhodamine-dextran concentration in the PF-HBSS collected from the uveal-side chambers was determined using a spectrofluorometer (Kontron Elektronik, GmBH, Zürich, Switzerland). The excitation and emission wavelengths were 550 and 580 nm, respectively. Standard curves of fluorescence versus concentrations were obtained by serial dilution of tetramethylrhodamine-dextran dissolved in PF-HBSS. Each sample was measured eight times and the measurements were averaged. 

Measurement of FGF-2

FGF-2 concentration in the medium collected from the Ussing chamber was measured using a sandwich enzyme immunosorbent assay (R&D Systems). The samples were placed in a 96-well microplate, precoated with a monoclonal antibody against FGF-2, for 2 hours at room temperature. Any FGF-2 present was bound by the immobilized antibody. After unbound substances were washed away, a horseradish peroxidase–linked monoclonal antibody against FGF-2 was added to the wells for 2 hours at room temperature. After a wash to remove any unbound reagent, a substrate solution containing hydrogen peroxide and tetramethylbenzidine was added to the wells for 30 minutes at room temperature. The color developed in proportion to the amount of FGF-2 in the sample. The reaction was stabilized by the addition of 1 N H₂SO₄. Optical density was measured at 450 nm and 540 nm using a microtiter plate reader (SpectraMax 250; Molecular Devices, Sunnyvale, CA). To correct for nonspecific variation, the absorbance at 540 nm was subtracted from that at 450 nm. Standard curves of absorbance versus concentration were obtained by serial dilution of standard purified FGF-2. For the time-course experiments, calculations were made to correct for the effect of removing the 100-μl samples from uveal-side chambers. 

Permeability Coefficient Determination

Diffusion from the orbital chamber to the uveal chamber was characterized by determination of a permeability coefficient (P _c), which is the ratio of steady state flux to the concentration gradient. ³⁰ In this study, the concentration of agents in the uveal-side chamber (C _U) was less than 1% of their concentration in the orbital-side chamber, C _O. Thus the change of C _O was assumed to be under the limit of detection, and the permeability coefficient was therefore calculated as follows  

\[P_{\mathrm{c}}(\mathrm{cm}/\mathrm{sec}){=}(C_{\mathrm{Ut}}-C_{\mathrm{U0}})V/C_{\mathrm{O}}tS\]

where C _U0 and C _Ut are the concentrations in the Ussing chamber at 0 and at t hours, respectively, C _O is the initial drug concentration in the orbital chamber, V is the volume of the each chamber (0. 75 ml), t is the duration time of steady state flux converted the unit from hour to second, and S is the surface area of exposed sclera (0.65 cm). ²  

Statistical Evaluation

At least 13 experiments were performed on FGF-2 and dextran at each concentration of latanoprost. Results from each experimental group were compared with a common vehicle control group using the Student-Newman-Keuls t-test. P < 0.05 was considered statistically significant. Data are presented as mean ± SD. 

Results

Time Course Analysis

The time course of FGF-2 penetration of sclera within the Ussing chamber was assessed at 20-minute intervals for 120 minutes. As shown in Figure 1 , the concentration increased linearly for the entire 120 minutes. The concentration increase rate was 0.27 ± 0.09 pg/ml · min when calculated according to the slope of the increase shown in Figure 1 (n = 6). This value is equal to (C_Ut − C_U0)/t and is similar to the calculated value from the concentration observed after 2 hours of incubation (0.28 ± 0.11 pg/ml · min; P = 0.90). When converted to permeability, these measurements were 1.63 ± 0.56 × 10⁻⁸ cm/sec and 1.66 ± 0.68 × 10⁻⁸ cm/sec, respectively. 

Dose-Response Analysis of Transscleral Flux

In these experiments, transscleral FGF-2 and 10-kDa dextran permeation in the vehicle-treated cultures were 1.52 ± 0.86 × 10⁻⁸ cm/sec (n = 13) and 1.04 ± 0.39 × 10⁻⁶ cm/sec (n = 14), respectively. Increasing the concentration of latanoprost acid in the cultures increased scleral permeation by both FGF-2 and 10-kDa dextran (Figs. 2 and 3) . FGF-2 permeation in sclera exposed to 50, 100, and 200 nM latanoprost acid was increased by 48% ± 62%, 100% ± 108%, and 108% ± 79%, respectively compared with scleral cultures exposed to vehicle (P < 0.05 for the 100- and 200-nM experiments). Similarly, when compared with vehicle-treated cultures, 10-kDa dextran permeation in cultures exposed to 50, 100, or 200 nM latanoprost acid was increased by 42% ± 36%, 59% ± 51%, and 65% ± 49%, respectively (n = 14, P < 0.05 for all three test concentrations). 

Comparison of FGF-2 and Dextran Movement

The increase of FGF-2 and 10-kDa dextran permeation were compared in the same scleral tissues exposed to various concentrations of latanoprost acid (Table 2) . In these experiments, transscleral permeability in the vehicle-treated scleral cultures was 1.20 ± 0.77 × 10⁻⁸ cm/sec. Because of Ussing chamber leakage in several experiments, the number of experiments with results from both the FGF-2 and dextran permeation assessment in the same scleral culture varied from 9 to 13. The ratio of dextran versus FGF-2 permeation was approximately 90 and did not vary among scleral cultures exposed to vehicle or 50, 100, or 200 nM latanoprost acid (P= 0.93, ANOVA). 

Discussion

The present results indicate that exposure of human sclera to latanoprost acid increased FGF-2 permeation in a dose-dependent manner. This may reflect a general enhancement of scleral permeability, because it paralleled increased scleral permeation by10-kDa tetramethylrhodamine-dextran. There was no significant relationship between age, race, and scleral permeability coefficient. 

There was a linear increase of the concentration of FGF-2 over 120 minutes in the uveal-side Ussing chamber (Fig. 1) . This indicated that FGF-2 can pass through the human sclera with a constant rate in vitro and suggests that its movement reflects molecular diffusion. ³¹ The permeation rate of FGF-2 was approximately 90 times lower than that of 10-kDa dextran. The greater permeation rate of 10-kDa dextran may be related to binding of FGF-2 to molecules within the sclera. These molecules include collagen types I, III, V, VI, and VIII and the glycosaminoglycans (GAGs) chondroitin sulfate, dermatan sulfate, keratan sulfate, heparan sulfate, and hyaluronan. ^{32 33} Much of the chondroitin sulfate, dermatan sulfate, and keratan sulfate may be covalently linked to the proteoglycans decorin, biglycan, and aggrecan. ³² Immunoreactivity of each of these proteoglycans is distributed throughout the thickness of sclera. It is well established that FGF-2 strongly binds to heparan sulfate (K _d = 0.34μ M). ³⁴ Recently, it has been shown that FGF-2 also can bind to dermatan sulfate (K _d = 2.5μ M). ³⁵ Because each of these GAGs is present within sclera, it is possible that binding of FGF-2 to these extracellular matrix components could impede the movement of FGF-2 through the sclera. 

Another possible explanation of the substantially lesser permeation rate of FGF-2 may be due to the formation of multimers. The molecular radius of 16-kDa FGF-2 monomer has been determined to be 3 nm and is almost the same as that of 10-kDa tetramethylrhodamine-dextran. ³⁶ However FGF-2 readily forms dimers in solution and can be multimerized by oxidation. In pentamers, the molecular radius of FGF-2 can be 6 nm larger than the dimer (13 nm). The permeation rate of 70-kDa tetramethylrhodamine-dextran, whose molecular radius is 6.4 nm, similar to FGF-2, was 20-fold less than that of 10-kDa dextran, ¹⁴ and four times higher than FGF-2. Hence, the permeation rates of bioactive proteins such as FGF-2 may depend on molecular weight and molecular radius, including the formation of multimers, and may be modulated by binding and degradation in vital tissues. 

Increased transscleral permeation by FGF-2 after PG treatments suggests that cotreatment with PGs may facilitate the use of FGF-2 to enhance survival of retinal neurons in glaucoma and other eye diseases. Previous studies have shown that FGF-2 can promote neuronal survival in vitro and in vivo. ^{37 38} Beneficial effects were observed with concentrations as low as 20 pg/ml. Moreover, intraventricular infusion of FGF-2 can promote neuronal survival after experimental axotomy, ischemia, neurotoxin treatment, or contusion of brain or spinal cord tissue. ^{39 40 41}  

It should be noted that infusion of FGF-2 may stimulate responses in many other tissues, besides neural tissues, that may be either beneficial or detrimental to the desired neural tissue response. There also may be specific requirements for additional factors in the case of retinal ganglion cells. ⁴² Except in retina and neural tissues, FGF receptors are present in cornea, trabecular meshwork, lens epithelial cell, vitreous fluid, vascular endothelial cells, and the membrane of eyes with proliferative vitreoretinopathy or diabetic retinopathy. ^{43 44 45 46 47 48 49} Therefore, enhanced transscleral delivery of FGF-2 may influence a number of ocular tissues. Other receptors for growth factors and neurotrophic factors also are expressed in non-neural posterior pole tissues. Thus, realizing the benefits of increased transscleral permeability may require simultaneous suppression of undesired side effects. ⁵⁰  

After exposure of sclera to PGs, there is evidence of biochemical changes within the sclera, including increased MMPs and reduced collagens. ^{12 14} However, laboratory studies ¹² and clinical trials ⁵¹ have not revealed any morphologic or functional changes in sclera. Although there was reduced collagen in the sclera of monkey eyes receiving topical PGF_2α-IE, for example, there was no significant change in the structural organization of the sclera or associated structures. ¹² Moreover, neither topical application of latanoprost nor subconjunctival injection of latanoprost acid altered the progression of form-deprivation myopia in chicks. ⁵² This suggests that endogenous control of scleral growth is not altered by FP-receptor activation. These findings collectively suggest that long-term clinical use of latanoprost is not detrimental to the sclera. 

In conclusion, the increase in the permeation rate of FGF-2 across the sclera after exposure to latanoprost suggests that latanoprost cotreatment may facilitate delivery of FGF-2 to posterior pole tissues. Because this study evaluated organ-cultured sclera, in vivo studies should be undertaken to confirm the response. This response may be particularly useful for drug delivery to the choroid and optic nerve, because the barrier for macromolecule diffusion between these tissue compartments is minimal. ^{53 54} It also may facilitate delivery of FGF-2 to the retina, especially if permeability of the blood–retinal barrier can be concomitantly increased by opening retinal pigment epithelium tight junctions. A prolonged increase in blood–retinal barrier permeability may lead to certain adverse effects, however. If this occurs, the adverse effects would have to be weighed against the potential benefits of FGF-2 delivery to the retina. Limiting the duration of opening of the blood–retinal barrier might be one approach to avoiding such adverse effects. Finally, in addition, cotreatment with latanoprost or other similar prostaglandin analogues may facilitate the transscleral delivery to the posterior segment of other peptide growth factors or macromolecules in addition to FGF-2. 

 Supported in part by the National Eye Institute, EY05990 (RNW), the Foundation for Eye Research (MA), and the Joseph Drown Foundation (JDL). MA was supported by the Department of Ophthalmology, Faculty of Medicine, University of Tokyo, Japan.

 Submitted for publication March 19, 2001; revised May 23, 2001; accepted June 13, 2001.

 Commercial relationships policy: N.

 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.

 

Table 1.

 

Characteristics of Donors Including Age, Sex, Race, and Cause of Death

Age	Sex	Race	Cause of Death	Usage
								FGF-2	Dextran
77	F	White	Cerebral vascular accident		*
62	F	White	Pancreatic cancer	*	*
78	M	White	Lung cancer	*	*
70	M	White	Cerebral vascular accident	*
70	M	White	Asphyxiation	*	*
36	F	White	Gastric bleeding	*	*
68	F	Hispanic	Cerebral vascular accident	*	*
78	F	White	Ventricular arrhythmia	*	*
72	M	Not specified	Cardiovascular disease	*	*

*  Usage of donor eye in each experiment. The mean donor age was 66.8 ± 13 (mean ± SD) years.

Figure 1.

 

Time course of FGF-2 concentration in the receiving Ussing chamber fitted with human sclera previously cultured for 3 days in control medium. Data are presented as mean ± SD. Increase in concentration with time was linear (R ² = 0.91). Based on this, the permeation rate was determined to be 1.63 × 10⁻⁸ cm/sec (n = 6 donor eyes).

Figure 2.

 

Scleral permeability to FGF-2 after exposure to various concentrations of latanoprost acid for 3 days. Data indicate the percentage of permeability of nontreated sclera (mean ± SD).* P < 0.05, Student-Newman-Keuls test; n= 13 donor eyes.

Figure 3.

 

Scleral permeation by 10-kDa tetramethylrhodamine-dextran after exposure to various concentrations of latanoprost acid for 3 days. Data are presented as mean ± SD. *P < 0.05, Student-Newman-Keuls test; n = 14 donor eyes.

Table 2.

 

Comparison of Scleral Permeability of FGF-2 and Dextran, with or without Latanoprost Treatment

Latanoprost	FGF-2*	Dextran, †	Dex/FGF-2	n
0 nM	1.20 ± 0.77	0.70 ± 0.35	93.1 ± 53.7	9
50 nM	1.64 ± 0.72	1.29 ± 0.39	88.5 ± 40.0	13
100 nM	1.88 ± 0.72	1.37 ± 0.33	83.1 ± 38.9	13
200 nM	2.35 ± 1.37	1.45 ± 0.35	86.8 ± 61.3	10

*  ×10⁻⁸ cm/sec (mean ± SD).

^†  ×10⁻⁶ cm/sec.

The authors thank the San Diego Eye Bank for providing the human donor eyes studied in this investigation. This work was prepared in partial fulfillment of the requirements for membership for Robert N. Weinreb in the American Ophthalmological Society.