Presentation Transcript

  1. Weinreb, Thesis 11/9/01 11:30 AM Page 319 ENHANCEMENT OF SCLERAL MACROMOLECULAR PERMEABILITY WITH PROSTAGLANDINS* BY Robert N. Weinreb, MD ABSTRACT Purpose: It is proposed that the sclera is a metabolically active and pharmacologically responsive tissue. These studies were undertaken to determine whether prostaglandin exposure can enhance scleral permeability to high-molecular- weight substances. Methods: Topical prostaglandin F2?(PGF2?) was administered to monkeys to determine if this altered the amount of scle- ral matrix metalloproteinases (MMPs). Experiments also were performed to determine whether the prostaglandin F (FP) receptor and gene transcripts are expressed in normal human sclera. Permeability of organ-cultured human scle- ra following prostaglandin exposure then was studied and the amount of MMP released into the medium measured. Finally, the permeability of human sclera to basic fibroblast growth factor (FGF-2) was determined following prostaglandin exposure. Results: Topical prostaglandin administration that reduced scleral collagen also increased scleral MMP-1, MMP-2, and MMP-3 by 63 ± 35%, 267 ± 210%, and 729 ± 500%, respectively. FP receptor protein was localized in scleral fibrob- lasts, and FP receptor gene transcript was identified in sclera. Exposure to prostaglandin F2?, 17-phenyltrinor, PGF2?, or latanoprost acid increased scleral permeability by up to 124%, 183%, or 213%, respectively. In these cultures, MMP- 1, MMP-2, and MMP-3 were increased by up to 37%, 267%, and 96%, respectively. Finally, transscleral absorption of FGF-2 was increased by up to 126% with scleral exposure to latanoprost. Conclusions: These studies demonstrate that the sclera is metabolically active and pharmacologically responsive to prostaglandins. Further, they demonstrate the feasibility of cotreatment with prostaglandin to enhance transscleral deliv- ery of peptides, such as growth factors and high-molecular-weight substances, to the posterior segment of the eye. Tr Am Ophth Soc 2001;99:319-343 It is proposed here that enhancement of scleral macromolecular permeability with prostaglandins may be such a method. The hypothesis investigated in this thesis is that the sclera is metabolically active and that prostaglandin cotreatment activates enzymes that enhance scleral macromolecular permeability. INTRODUCTION Peptides, including growth factors and other high-molec- ular-weight substances, are potential therapeutic agents for delivery to the posterior segment in glaucoma, age- related macular degeneration, and other ocular disor- ders.1-4The potential benefit of such treatments is sug- gested by enhanced survival and differentiation of retinal neurons in cultures5-7and improved neuronal viability in experimental models8-11to which certain growth factors or neurotrophins have been added. However, targeted deliv- ery of even low-molecular-weight drugs to the optic nerve, retina, and choroid has been problematic. Delivery of high-molecular-weight drugs is even more challenging. Methods for simple and safe drug delivery to the posteri- or segment clearly are needed if such neuroprotection strategies are to be effective. CURRENT APPLICABILITY TO THE POSTERIOR SEGMENT METHODS OF DRUG DELIVERY AND THEIR The applicability of current methods of drug delivery to the posterior segment is limited by poor drug absorption, particularly of high-molecular-weight substances. Topical Application Topical application is the primary route of drug delivery to the anterior segment of the eye. This route of administra- tion is noninvasive, simple for the patient to use, and rela- tively free of systemic side effects, particularly when applied with punctal occlusion or gentle lid closure to min- imize systemic drug absorption. However, topical applica- tion requires rigorous patient compliance over an extended time to effectively treat chronic disease, and it is largely *From the Glaucoma Center, University of California, San Diego, School of Medicine, La Jolla. Supported in part by grant EY-05990 from the National Eye Institute; a Senior Scientist Award from Research to Prevent Blindness, Inc, New York, New York; and the Joseph Drown Foundation, Los Angeles, California. Tr. Am. Ophth. Soc. Vol. 99, 2001 319

  2. Weinreb, Thesis 11/9/01 11:30 AM Page 320 Weinreb ineffective for drug delivery to the posterior segment. Many drugs are capable of penetrating intact corneal epithelium to achieve significant levels in the cornea, anterior chamber, iris, and ciliary body. Topically applied drugs also may enter the eye by crossing the conjunctiva and diffusing through the sclera, but do so only to a minor extent. Although the transcorneal penetration of drugs is essential to achieve therapeutic drug levels in anterior segment tissues, drugs applied to the cul-de-sac typically do not achieve pharmacologically active concentrations in posterior segment tissues following topical administration. An important factor that limits topical application as a means of drug delivery to the posterior segment is the loss of drug from the precorneal area. Induced lacrimation because of an instilled volume into the conjunctival sac, blinking, physiologic tear turnover, drug-protein binding, and enzymatic degradation of drug in tear fluid allow only a small amount of an applied dose to pass into the aque- ous humor and surrounding tissues in the anterior cham- ber. Edelhauser and Maren found that lower corneal per- meability in humans than rabbits may result, in part, from our fourfold greater blinking rate and twofold greater tear turnover.12 The corneal epithelium is another contributing factor that limits topical application, as it is a barrier to drug absorption, particularly for high-molecular-weight sub- stances such as growth factors. As a result of these hur- dles, the absorption of drugs applied topically to the eye is quite poor compared with the systemic route of adminis- tration. The extent of absorption, as measured by taking the ratio of the total amount of drug that has entered the eye divided by the instilled dose, ranges from 1% to 7% for ophthalmic drugs.13,14In contrast, the extent of absorp- tion of systemic drugs is usually greater than 75%. Within the anterior chamber, drug dilution and removal also reduce the amount of drug available for dif- fusion posteriorly to the optic nerve, retina, and choroid. Aqueous secretion within the anterior segment dilutes the aqueous humor drug concentration, and normal aqueous drainage removes drug that has penetrated the corneal tis- sues. Drug also may diffuse into blood vessels within the anterior segment and then be removed from ocular tissues. A drug that is absorbed into the anterior chamber also must redistribute from the aqueous to the vitreous humor. However, a drug topically applied to the eye in general does not enter the vitreous in significant concentrations. Two main factors explain the relative lack of penetration in the vitreous cavity. One factor is that there is only a minute space between the ciliary processes and the lens, and drugs must diffuse against an aqueous humor flow gradient. Another factor is the relatively slow diffusion of drugs in the vitreous. Molecular charge and lipophilicity have little effect on the diffusion of drugs of ocular interest.13-15Drug diffusion depends on molecular move- ment through an aqueous environment. Although it is not restricted by the presence of collagen in the vitreous, drug diffusion is too slow to allow significant drug accumulation within the vitreous. Finally, cell junction barriers, as dis- cussed subsequently, can limit the diffusion of drugs with- in the vitreous into the optic nerve, retina, and choroid.15 Systemic Administration Despite the excellent absorption of systemic drugs, sys- temic administration of biologically active agents is inef- fective at achieving therapeutic levels in the posterior seg- ment of the eye. In the intact eye, systemic routes, such as oral or parenteral, may not produce a high enough con- centration of drug because of resistance to entry from blood-ocular barriers, metabolism of drug to an inactive species, or significant uptake into another organ or tissue. Drug absorption into the eye is increased during ocular inflammation, which is associated with a disruption of the blood-aqueous barrier, the blood-retinal barrier, or both. Even in this case, however, systemic drug administration at doses sufficient to be absorbed into the eye may cause unacceptable systemic side effects, since the drug actions are unlikely to be localized to the eye. As an example, neurotrophins, which are high-molecular-weight drug candidates for optic nerve and retinal neuroprotection, have many properties aside from their roles in neuronal survival and axonal growth associated with retrograde transport. Neurotrophins also are anterogradely trans- ported and released from presynaptic to postsynaptic tar- gets.16 Further, they modulate membrane excitability, induce neuronal hypertrophy, and affect cell differentia- tion.2If cells throughout the body were exposed to exoge- nously administered neurotrophins, the systemic side effects would likely be ubiquitous and deleterious. Intravitreal Injection Intravitreal injection, most often via the pars plana, offers a direct route to the posterior segment and often can pro- vide adequate tissue drug levels. For some infectious or inflammatory diseases of the posterior segment, intravit- real injections are an effective and essential component of treatment. Local delivery of drugs to the eye via intravit- real injection offers several advantages over other routes of administration. First, it avoids many of the side effects associated with systemic therapy. This is particularly of benefit in the case of medications that may be too toxic for systemic administration but are well tolerated by the eye. Second, it bypasses the blood-ocular barrier, allowing higher intraocular drug levels than might otherwise be achieved. This may be particularly advantageous for high- molecular-weight substances. Even if a drug can be delivered intravitreally, 320

  3. Weinreb, Thesis 11/9/01 11:30 AM Page 321 Enhancement of Scleral Macromolecular Permeability with Prostaglandins however, the barrier presented by the internal limiting membrane is an important factor that may preclude intravitreal delivery of many peptides, including growth factors and other high-molecular-weight substances, to the retina. In rabbits with experimental subretinal detachments, Takeuchi and associates17observed that flu- orescein isothiocyanate albumin (67 kDa) injected intrav- itreally can diffuse, but only slowly, across the sensory reti- na into the subretinal space. Kamei and associates18inject- ed tissue plasminogen activator (70 kDa) labeled with flu- orescein isothiocyanate and rhodamine B isothio- cyanate–labeled dextran (20 kDa) into the midvitreous cavity of normal rabbits and those with experimental sub- retinal hemorrhage. The smaller-molecular-weight dex- tran diffused slowly throughout the neural retina in each of the eyes. Intravitreal tissue plasminogen activator did not diffuse through the intact neural retina in any of them. Distribution of an intravitreal drug also may limit drug delivery to the retina. A drug with a rapid rate of clear- ance from the vitreous may require large boluses and fre- quent injections to ensure therapeutic levels over an extended period. Intravitreal injections have the inherent potential side effects of retinal detachment, endophthalmitis, cataract formation, and vitreous hemorrhage. The benefits of treatment must supersede these risks. In chronic dis- eases, long-term intravitreal treatment also might need frequent injections. Repeated injections have incremen- tal risks, and they generally would not be well tolerated by the patient. Therefore, sustained-release intravitreal drug delivery, which would require fewer injections, may be particularly advantageous for treatment of chronic eye diseases.19 The use of sustained-release drug delivery systems that are placed within the vitreous is being investigated for a number of eye diseases, including cytomegalovirus retinitis,20-23uveitis,24-26proliferative vitreoretinopathy,27-30 and choroidal neovascularization. In addition to the advantages of intravitreal injection, sustained drug deliv- ery offers the promise of relatively constant drug levels in the vitreous. On the other hand, drugs that may be safe to the eye when used for a short time may prove to be toxic with sustained intraocular levels. Further, once placed intraocularly, they would need to be surgically removed if there were untoward side effects. The devices, too, have risks and complications associated with their placement that may preclude their routine use in eyes with glaucoma or age-related macular degeneration. These limitations have delayed their introduction into clinical use, particularly for treatment of chronic diseases. the posterior segment by anterior or posterior sub- Tenon’s, subconjunctival, or retrobulbar injection is another method of delivering drugs to the posterior seg- ment of the eye. For many reasons, this is an attractive route for delivering drugs to the optic nerve, retina, or choroid. This route has a major advantage of bypassing the epithelial barriers of the cornea and conjunctiva, which limit drug absorption with topical application. Despite the frequent clinical use of periocular injec- tion for a plethora of ocular disorders, the mode of drug transfer from the periocular location into the ocular tis- sues is not clearly understood. The drug may leak from the conjunctival injection site and penetrate the cornea or may enter the eye in part via systemic absorption. Also, the drug may enter the eye through intrascleral vascular channels or through perivascular and perineural spaces surrounding penetrating blood vessels and nerves. Perhaps most significantly, the drug also may enter the eye after penetrating directly through the sclera.31,32 McCartney and associates33documented the direct penetration through underlying sclera of subconjunctival tritium-labeled hydrocortisone (molecular weight [MW], 362 Da), a low-molecular-weight drug, through the ocular tissues in the rabbit eye. In their study, the percentage of the total dose that penetrated into the eye appeared to be small (~1% to 2%). Studies in normal squirrel monkeys by Barza and associates34demonstrated that gentamicin (MW, 450 Da to 477 Da), a mixture of 3 similar low- molecular-weight compounds, also can penetrate directly through underlying sclera when administered by subcon- junctival or retrobulbar injection. They also found the highest drug concentrations in the superior and inferior segments of the rabbit sclera, but no detectable drug in the nasal and temporal areas.35On the basis of these data, they hypothesized that the drug solution tends to remain localized and spreads over the superior segment nearest the injection site for a period long enough to be absorbed by the sclera. It then moves to the inferior segment, pos- sibly as a result of gravity. Depending on the position of the subject, it is absorbed through the superior and infe- rior scleral surfaces rather than nasal and temporal areas. They observed significant levels in the retina and choroid, but not the vitreous. Lim and associates36detected tissue plasminogen acti- vator (MW, 70 kDa) in rabbit vitreous after subconjuncti- val injection, but the concentrations were very low. Subsequently, Lincoff and associates37observed that recombinant human interferon ?-2a (MW, 20 kDa) dif- fused into the rabbit choroid, but only in small amounts, after retrobulbar injection. The total choroidal concentra- tion was only 3% of the amount injected in the retrobulbar space, and the serum concentration was less than 1% of the choroidal concentration. Weijtens and associates38,39 Periocular Injection Directly introducing a drug into the tissues surrounding 321

  4. Weinreb, Thesis 11/9/01 11:30 AM Page 322 Weinreb studied dexamethasone concentrations in subretinal fluid of patients with a rhegmatogenous retinal detachment undergoing a scleral buckling procedure with drainage. They found that a subconjunctival injection, preceded by topical cocaine to disrupt the corneal and conjunctival epithelial barrier, resulted in a higher maximal vitreous concentration of dexamethasone (MW, 393 Da) than a sub-Tenon’s or retrobulbar injection. However, even the highest vitreous concentrations in their study were lower than those needed for a therapeutic effect based on in vitro testing of human retinal pigment epithelial cell pro- liferation.39 Although periocular injection can deliver some low- molecular-weight substances in clinically relevant concen- trations to the posterior segment, the rate of delivery is not as effective with high-molecular-weight substances. Improvement in the extent of absorption can come from better retention at the site of absorption through the use of gels,40-42 nanospheres,43 biodegradable carriers. Improvement in the extent of absorption also can be achieved with improved penetra- bility. Methods to improve the penetrability of the sclera have been sparsely investigated. significantly less difference in average collagen fibril diameter between the inner and outer portions of anteri- or compared with more posterior human sclera. According to them, Purnell and McPherson61suggested that the larger fibril diameter, which they extrapolated to looser fibril arrangement, might provide less resistance to aqueous flow through the intervening ground substance. Bundles of thinner fibers, possibly precollagen, are found near scleral fibroblasts. The length of the bundles is not known. The bundles have a slightly fusiform shape with tapering ends and dichotomous branches. The turnover rate of scleral collagen is also unknown. The flat and elon- gated cells, the scleral fibroblasts, are few in number and are separated by collagen. The long axis of the cell and nucleus is parallel to the surface. Long, thin cytoplasmic extensions from the cells are attenuated to a diameter one-third to one-half the size of the collagen bundles. Experimental studies in an avian model suggest that the sclera is derived from 2 sources, the ectodermal neural crest and mesoderm.62 The mean total scleral surface area is approximately 17.0 ± 1.5 cm2.63Mean scleral thickness ± SD is 0.53 ± 0.14 mm at the corneoscleral limbus, significantly decreasing to 0.39 ± 0.17 mm near the equator and increasing to 0.9 to 1.0 mm near the optic nerve.64The sclera thins with age. The large surface area and thinness of the sclera are desirable features of a route for targeted drug delivery. In a normal eye, blood vessels only traverse the scle- ra and do not supply it directly. Therefore, the stroma of the sclera derives its nutrition from a distance and not through intimate contact with the capillary bed. This implies that the sclera must be permeable to fluids and metabolites, as reported first by Bill.65The external move- ment of substances through the sclera results from a pres- sure difference between the suprachoroidal space, where the pressure is 1 to 2 mm Hg lower than intraocular pres- sure, and the episcleral tissue, where the pressure is near 0 mm Hg. Transscleral movement should theoretically be slowed by reducing intraocular pressure. Whether this flow forms part of the normal flow of fluid from the supra- choroidal space is uncertain.66 liposomes,44-49 or other SCLERA Basics The sclera is a densely collagenous, hypocellular and elas- tic tissue that is composed of a proteoglycan matrix and closely packed collagen fibrils.50-53It has been thought to be relatively inactive metabolically, having no intrinsic capillary bed and few fibroblasts. Duke-Elder54described the sclera as “inert and purely supportive in function.” Watson and Hazleman55stated that it is “metabolically rel- atively inert.” The outer surface of the sclera is covered by the loosely organized episclera and its inner surface by the lamina fusca and suprachoroidal space. The sclera is per- forated by the emissarial canals for arteries, veins, and nerves. Unlike corneal collagen, scleral collagen bundles do not lie in orderly, regular lamellae but are interlaced in an irregular fashion, which accounts for the lack of trans- parency. Collagen forms 75% of the dry weight of the sclera;56the remainder is made up of noncollagenous pro- teins and mucopolysacharides.57,58Approximately 70% of the weight of intact sclera is water. Spitznas59studied the ultrastructure of human scleral collagen posterior to the ora serrata and reported that the diameter of the fibrils in the outer layers is significantly larger than that of the inner layers. There is a ratio of 1:2 between the diameter of collagen fibrils in the innermost and outermost layers of human posterior sclera, respec- tively.59,60Shields and associates61observed that there is Transscleral Fluid Movement Drug penetration across the sclera is a route of entry into the eye for some ocular drugs, particularly those with low molecular weight. However, the details of transscleral fluid movement are poorly understood. In addition to being the outer surface of the globe, the sclera is the distal component in the uveoscleral out- flow pathway. Histologic analysis of sclera following injec- tion of various tracers into the anterior chamber indicated the presence of transscleral fluid flux through the scleral 322

  5. Weinreb, Thesis 11/9/01 11:30 AM Page 323 Enhancement of Scleral Macromolecular Permeability with Prostaglandins stroma, as well as possibly through narrow spaces around penetrating nerves and blood vessels.65,67-69However, there is little information regarding the character of this fluid movement. What factors influence scleral permeability? Does the sclera allow only unidirectional flow from within the eye to the outside? Or might the flow be bidirection- al, and also from outside to within? Is it always passive, as might occur with porous diffusion through a fiber matrix? Or might it also be regulated by endogenous signals? Assessment of drug diffusion through the sclera by in vitro permeability studies is a useful approach to estimate drug movement for in vivo conditions. Maurice and Polgar70examined diffusion across bovine sclera with a broad range of different molecular weight solutes and ions, including methylene blue (MW, 320 Da) and serum albumin (MW, 69 kDa). They showed first that the diffu- sion of the ions and solutes was inversely related to molecular weight. Shields and associates71performed an in vitro study with postmortem human eyes to assess the permeability of outer anterior sclera following trabeculec- tomy. They recognized the permeability of sclera to fer- ritin or india ink, and they speculated that the route of flow was through vessels in the flap or through the “colla- gen ground substance between individual fibrils.” Ahmed and Patton72recognized that under certain conditions, even some topically applied drugs-timolol maleate (MW, 433 Da) and inulin (MW, 5 kDa)–can enter the eye via the transscleral route and bypass the anterior chamber. Edelhauser and Maren12reported that scleral permeabili- ty was greater than corneal permeability except for highly lipid soluble compounds, for which scleral and corneal permeability were the same. Olsen and associates64examined the diffusion across human sclera by using a range of different molecular weight solutes as high as dextran (MW, 70 kDa). They confirmed in human sclera that the diffusion of the ions and solutes was inversely related to molecular weight. Moreover, they observed that the absorption of molecular weight substances greater than 5 kDa was poor. As they studied only tissue without macroscopically visible perfo- rating channels or light and electron microscopic perfo- rating vascular channels, their studies were most consis- tent with direct penetration through the sclera as an important vector for transscleral drug transfer into the eye. More recently, Ambati and associates73measured the permeability of rabbit sclera to a series of fluorescein- labeled hydrophilic compounds with a wide range of molecular weights as high as 150 kDa. Scleral permeabil- ity decreased with increasing molecular weight, a finding consistent with previous human and bovine data. They also suggested that molecular radius might be a better predictor of scleral permeability than molecular weight. These studies all are consistent with the sclera being permeable, and the permeability being inversely related to the molecular weight of the permeating substance. Influencing Drug Penetration and Scleral Permeability The rate and extent of absorption are determined by both the physicochemical behavior of a substance and the per- meability of the sclera. Absorption can be enhanced for a particular substance by increasing drug penetration or by increasing scleral permeability. Few methods have been tested to enhance either of these, particularly for molecu- lar weight substances higher than 5 kDa. Drug penetration into the eye may be increased with iontophoresis, the process of moving a charged molecule by an electric current across the cornea or sclera. Transcorneal iontophoresis has been shown to result in significantly higher drug levels than those found after multiple drop treatments.74However, it is questionable whether significant drug concentrations within the optic nerve, retina, or choroid can be achieved with it. Experimental studies by Lam and associates75have shown that toxic tissue-damaging effects might accompany drug delivery by transscleral iontophoresis. They observed local retinal and choroidal lesions following transscleral ion- tophoresis of various drugs. Current density and duration of application affected the size and severity of the lesions. This technique also is not practical. Few studies have investigated the factors that con- tribute to scleral permeability. Olsen and associates64did not find any significant correlation between scleral per- meability to inulin and age. They also observed that cryotherapy did not significantly affect scleral permeabili- ty to 5-fluorouracil (MW, 130), inulin, or dextran (MW, 40 kDa).64 Further, scleral permeability to sucrose (MW, 342), inulin, and dextran (10 kDa) was unaffected by transscleral diode laser retinopexy. Scleral permeability does appear to be related to scle- ral thickness. Shields and associates71found that increased outflow was inversely related to the scleral flap thickness in a small number (N=6) of postmortem human eyes that underwent experimental trabeculectomy. Scleral permeability to the low-molecular-weight substances dex- amethasone and methotrexate (MW, 455) also was increased significantly with one half surgical thinning of the sclera.64Dan and Yaron76found increased transscleral flow of saline through bovine sclera with application of clostridial collagenase. Interestingly, they also observed that collagenase applied directly to rabbit sclera after a fornix-based peritomy resulted in scleral thinning and lower intraocular pressure.76 microapplicator, this response could not be precisely con- trolled. In contrast to increased permeability with scleral thinning, abnormal and thickened sclera may be associat- ed with reduced permeabilty. Trelstad and associates77 Even with the use of a 323

  6. Weinreb, Thesis 11/9/01 11:30 AM Page 324 Weinreb found that sclera from 2 nanophthalmic eyes was thicker than normal and contained unusually disordered collagen fibrils. Yue and associates78found that collagen fibers were twisted and more closely packed in nanophthalmic eyes, changes consistent with reduced scleral permeability. Interestingly, Gass79first suggested posterior sclerotomy, which increases transscleral flow, as an effective treatment before entering the anterior segment to prevent choroidal effusion in these eyes. As recommended by Brockhurst,80 vortex vein decompression also may be effective. Scleral permeability also may be influenced by intraocular pressure. Rudnick and associates81recently evaluated the permeability of human sclera to 3 low- molecular-weight compounds (carboxyfluorescein, dexam- ethasone, and water) and found a small effect of intraocu- lar pressure. They suggested that pressure-related com- pression of collagen and narrowing of intracollagen path- ways within the sclera slow diffusion of small molecules, yet may completely block transport of macromolecules.81 By enhancing scleral permeability, one might be able to more effectively deliver drugs to the posterior segment. Enhanced scleral permeability also might lower uveoscle- ral outflow resistance and lower intraocular pressure. initiation of collagen degradation in the ciliary muscle extracellular space.86There is immunohistochemical evi- dence that MMP-1, which can mediate normal turnover of fibrillar collagens, such as collagen type I and collagen type III, is present in normal human sclera.87Analysis with PG receptor agonists suggests that these PG respons- es are receptor-mediated.87 The biologic activity of a drug, whether it be thera- peutic or toxic, is often proportional to the concentration of that drug at the receptor site. Moreover, the persist- ence of its effects is directly related to the residence time of the drug at the receptor. The PG receptor type that most specifically recognizes F family prostaglandins is the FP prostanoid receptor, a G-protein–coupled cell mem- brane receptor.88,89In situ hybridization and immunohis- tochemical studies have demonstrated FP receptor tran- scripts and protein in several anterior segment tissues of monkey eyes.90In the sclera, however, only FP receptor immunoreactivity has been observed,90and no evidence of FP receptor transcripts has been detected. It is possible that the sensitivity of the in situ hybridization technique used was insufficiently sensitive to detect small amounts of FP receptor mRNA. In view of the potential respon- siveness of sclera to prostaglandins, direct assessment of FP receptor gene transcription and protein expression in human sclera clearly is needed. If prostaglandin exposure does enhance scleral macromolecular permeability, the assertions that might be valid include the following: • Topical prostaglandin administration reduces scler- al collagen by increasing scleral metalloproteinase (MMP) activity. • This effect is FP (prostaglandin F) receptor-medi- ated, and both gene transcription and protein expression are present in the sclera. • Exposure of isolated sclera to specific FP receptor agonists increases scleral permeability in association with increased MMP expression. • Transscleral absorption of a high-molecular-weight substance, basic fibroblast growth factor (FGF-2), increases with scleral exposure to prostaglandin. Can Transscleral Fluid Movement be Enhanced by Prostaglandins? The possibility that various prostaglandins (PGs) could modulate transscleral fluid movement and enhance scler- al macromolecular permeability is suggested by several observations. First, topical treatment of monkey eyes with PGF2?-isopropyl ester (IE) for 5 days is known to enhance uveoscleral outflow and to reduce collagen type I and collagen type III immunoreactivity within sclera by 43% and 45%, respectively.82Second, scleral collagen is predominantly type I collagen and accounts for about one half of the total dry weight of sclera.56Finally, evidence that compaction of extracellular matrix affects transscleral permeability suggests that collagen density within sclera is an important determinant of permeability.83Hence, it is plausible that PG-mediated reduction of scleral collagens could significantly alter permeability. Increased local biosynthesis of matrix metalloproteinases (MMPs), a fam- ily of secreted neutral proteinases that can initiate specif- ic degradation of key extracellular matrix components, may be regulating the reduction of scleral collagen fol- lowing topical PG.84,85 PGs induced substantial remodeling of ciliary muscle extracellular matrix in situ that reflected MMP-mediated collagen reduction.82 The sequence of cellular events underlying this response includes PG transduction at cell surface PG receptors, induction of MMP gene transcrip- tion, translation and secretion of proMMPs, activation of MMPs by proteolytic truncation, and MMP-mediated METHODS 1. MEASUREMENT TEINASES AFTER TOPICAL PROSTAGLANDIN METALLOPRO- OF SCLERAL MATRIX PGF2?-IE–Treated Monkey Eye Tissue Young adult cynomolgus monkeys were evaluated by slit- lamp biomicroscopy on 2 occasions prior to initiation of treatment to confirm the absence of signs of ocular inflammation. In addition, integrity of the blood-aqueous barrier was confirmed by measuring the appearance and 324

  7. Weinreb, Thesis 11/9/01 11:30 AM Page 325 Enhancement of Scleral Macromolecular Permeability with Prostaglandins disappearance of fluorescence in the anterior chamber (AC) following intravenous fluorescein administration. To qualify for further study, AC fluorescence levels and time course of appearance and decay in both eyes had to be sim- ilar and within the range of values obtained for control eyes in previous studies.91Monkeys meeting these conditions were presumed to have an intact blood-aqueous barrier.91 The following week each qualifying monkey received 2?g PGF2?-IE (in 5 ?L) twice daily (morning and after- noon, approximately 7 hours apart) in 1 eye and 5 ?L of vehicle in the other eye for 5 days, as previously described. On the fifth day of treatment, slit-lamp biomi- croscopy was performed. Eight eyes of 4 monkeys were evaluated. AC cells or flare was not observed during treatment of these monkeys. The animals were sacrificed on day 5.82The vascular bed was perfused with lactated Ringer’s solution to remove circulating MMPs from the ocular tissues. The anterior segments were dissected and immediately fixed in methacarn (60% methanol, 30% chloroform, 10% gla- cial acetic acid) for 3 hours. Increased sensitivity of immunohistochemical staining of various antibodies has been demonstrated for many antigens after methacarn fix- ation.92,93Fixed anterior segments were transferred to cold 100% ethanol. The tissues were embedded in paraffin and sections were collected from the midsagittal region of each eye on Vectabond coated slides (Vector Laboratories, Burlingame, Calif). For histopathologic analysis, 3 or 4 sections from each eye were stained with hematoxylin and eosin. All procedures were conducted in accordance with the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. Tissue sections ana- lyzed in the present study were cut from the same tissue blocks as sections analyzed in a previous study.82 washed in 3 xylene changes to remove paraffin, and rehy- drated through graded ethanols. The sections were treat- ed with antigen retrieval solution (AR-10, Biogenex, San Ramon, Calif) at 95°C for 5 minutes. After cooling, the sections were exposed to 3% H2O2for 10 minutes to sup- press endogenous peroxidase activity. To remove intrinsic melanin, sections were treated successively with aqueous potassium permanganate (2.5 g/L) for 10 minutes and oxalic acid (5 g/L) for 3 minutes.95-97After rinsing, the sec- tions were blocked for 30 minutes with 0.1% bovine serum albumin (Sigma Chemical Co, St Louis, Mo) and incubated for 2 hours with affinity-purified polyclonal sheep anti-porcine MMP-1 (dilution 1:25, AB772, Chemicon, Temecula, Calif), polyclonal rabbit anti- human MMP-2 (dilution 1:100, AB809, Chemicon), or rabbit anti-human MMP-3 (dilution 1:1,000, AB810, Chemicon). These concentrations had been optimized in pilot studies for quantitative analysis. Specificity of these antibodies has been previously confirmed.87,98After rins- ing, the sections treated with antibody to MMP-1 were exposed to biotinylated donkey anti-sheep IgG for 30 min- utes (Biotin-SP-Conjugated Immunoresearch Laboratories, Inc, West Grove, Pa, diluted 1:500). The sections exposed to antibodies to MMP-2 or MMP-3 were exposed to biotinylated goat anti-rabbit immunoglobulin (Biogenex) for 20 minutes. After rinsing, the sections were exposed to horseradish peroxidase-conjugated streptavidin for 20 minutes. Consecutively, each section was rinsed and incubated with 3,3-diaminobenzidine chromogen for 10 minutes (HRP- DAB Super Sensitive Immunodetection System, Biogenex). To facilitate comparability, sections from the control vehicle-treated and PG-treated eye of each mon- key were immunostained at the same time.99,100To serve as controls for nonspecific staining, sections from each eye were simultaneously processed by the same protocol but without the primary antibody. Affinipure, Jackson Immunohistochemistry Sclera was immunostained by a standardized protocol. Each step of the protocol was optimized as previously described.94The concentration of each solution contain- ing antibodies or horseradish peroxidase-conjugated streptavidin was optimized to obtain submaximal (nonsat- urating) staining intensity as determined by imaging den- sitometry (described in the next section). Finally, the incubation time with diaminobenzidine was optimized for each primary antibody to obtain the strongest signal (staining intensity) that still was increasing linearly with time. The elimination of saturating binding or develop- ment parameters from the protocol supports the position that the observed changes in staining intensity reflected differences in tissue content of target antigen. Sections from the treated and control eyes were stained at the same time. Five sections, 10 ?m thickness, from each eye were heated to 56°C for 20 minutes, Densitometric Analysis Immunohistochemical staining intensity was directly measured with a high-resolution imaging densitometer.94 Measurements from multiple sections stained at the same time facilitated assessment of measurement precision and permitted statistical comparison of differences among control and experimental eyes. Immunostained sections were scanned by placing the slides directly on the platen of an imaging densitometer (model GS-700, Bio-Rad, Hercules, Calif). Resolution of the scans was set to 1,200 dpi (50 ?m-wide pixels), and the scanning mode was set to transillumination. The optical density measurements of the immunostained sections all were less than 1.00 optical density units. Because the densitometer can accurately measure optical densities greater than 3.0 units (Bio-Rad 325

  8. Weinreb, Thesis 11/9/01 11:30 AM Page 326 Weinreb specifications), these measurements were well within the appropriate range for accurate determinations. The scanned digital data were displayed in a masked fashion and analyzed by using an image analysis program (Molecular Analyst : version 2.1, Bio-Rad). The optical density along 2 line segments positioned over the sclera was measured in each section by using two-dimensional imaging densitometry. A similar line segment was posi- tioned perpendicular to the sclera adjacent to the ciliary body to assess background optical density. Mean optical density scores were determined from the optical density volume scores (optical density x mm) and the corresponding line segments (mm).94For each eye, 10 scores were obtained from 5 midsagittal sections. Background optical density was defined as baseline and subtracted from the original optical density scores. The specific optical density scores along each line segment over the sclera were calculated by dividing the optical density area score (optical density ( mm, provided by the densitometer) by the length of the line segment (mm) for that score. Mean specific optical density scores from the PG-treated eye of each monkey were compared to corre- sponding scores from the contralateral control eyes using the paired Student’s t test. The unpaired Student’s t test was used to compare the mean of mean optical density scores of all PG-treated and all control eyes. In each case, a P value less than 0.05 was considered significant. TRIzol reagent (Gibco, BRL, Life Technologies, Grand Island, NY) using a homogenizer (Polytron P-10; Brinkmann, NY). Homogenized sample was transferred to sterile 1.5-mL tubes in 1-mL aliquots and incubated for 5 minutes at 25ºC. Chloroform (200 ?L) was added to each tube and mixed by brief vortex and incubated for 3 minutes at 25ºC. Samples were centrifuged (12,000 ? 3) for 15 minutes at 4ºC. The aqueous phase was transferred to fresh sterile 1.5 mL tubes. Isopropanol (500 ?L/tube) was added and allowed to incubate for 10 minutes at 25ºC. Samples were centrifuged (12,000 (? g) for 10 minutes at 4ºC. Supernatant was removed, and the RNA pellet was washed with 75% ethanol/diethylpyrocarbonate water and air dried. RNA was resuspended in a total volume of 50 ?L diethylpyrocarbonate water, and quality was checked by gel electrophoresis. Reverse Transcription-Polymerase Chain Reaction (PCR) Primers were chosen to amplify 1,186-nucleotides of the human FP receptor. The sense primer (nucleotides 170 to 193) corresponds to a position 61 nucleotides upstream of the translation start site, and the antisense primer (nucleotides 1333 to 1356) corresponds to a position 39 nucleotides downstream of the stop codon in the human FP sequence. Both PCR primers were 100% homologous with the reported cloned sequence of the human FP receptor. The sense and antisense primers were used for reverse transcription (RT)-PCR as previously described with total RNA isolated from human sclera tissue.101The PCR (final volume, 50 ?L) contained 5 ?L of the RT reaction, 5 ?L of 10X PCR buffer, 1 ?L of 10 mM dNTP mixture, 1.5 ?L of 50 mM MgCl2, 2.5 ?L of the sense and antisense primers (20 ?M), and 0.5 ?L taq polymerase (all reagents from Gibco BRL, Grand Island, NY). The PCR program consisted of an initial step at 95ºC for 3 minutes, followed by 30 cycles at 95ºC for 1 minute, 55ºC for 1 minute, 72ºC for 1 minute, and a final step at 72ºC for 7 minutes. Products were analyzed by electrophoresis in a 1% agarose gel. 2. EXPRESSION IN NORMAL HUMAN SCLERA These experiments were undertaken to determine whether the FP receptor is expressed in normal human sclera. FP RECEPTOR GENE TRANSCRIPTION AND PROTEIN Tissue Preparation Postmortem human eyes from a 76-year-old donor were obtained from within 24 hours of death. Eyes were placed in chilled Hepes-buffered saline solution (HBSS) and maintained on ice. Eyes were surgically cleaned of con- nective tissue, blood vessels, muscle, and conjunctiva and rinsed once in HBSS. The anterior chamber was removed by a circumferential incision approximately 4 to 5 mm behind the limbus and snap-frozen, as described below. A circumferential incision was made in front of the optic nerve head, and ocular contents, including retina and choroid, were removed. The sclera was cut into 10 mm- square pieces. Residual pigmented tissue was removed with a cotton-tipped applicator. Sclera tissue was further cut into 2 mm squares and placed in a sterile 50 mL con- ical tube on ice. FP Receptor Protein Localization in Sclera Antibodies to the human FP receptor were generated in rabbits by using a recombinant fusion protein consisting of glutathione-S-transferase and a portion of the carboxyl ter- minus of the receptor consisting of amino acids 317 to 362. Preparations of the fusion protein and antibody purifica- tion were done as described.102Initial characterization of the antibodies was done as previously described103using COS-7 (African green monkey kidney) cells transfected with plasmid DNA encoding the human FP receptor. For labeling of human tissues, pieces of sclera (8 to 10 mm square) were snap-frozen in embedding medium (OCT, Tissue-Tek, Miles Inc, Elkhart, Indiana), sectioned on a Isolation of Total RNA The scleral tissue squares were homogenized in 8 mL of 326

  9. Weinreb, Thesis 11/9/01 11:30 AM Page 327 Enhancement of Scleral Macromolecular Permeability with Prostaglandins cyrostat (8 to 10-?m sections), mounted on glass cover- slips, and postfixed in 4% paraformaldehyde. Tissue sec- tions were washed twice with phosphate buffered saline (PBS) and then placed in 30 mM sodium chloride/300 mM sodium citrate for 20 minutes. Sections were then incu- bated in 100 mM glycine solution for 20 minutes to block nonspecific binding, and then washed twice in sodium cit- rate buffer for 10 minutes. Sections were permeabilized with 30 mM sodium chloride/300 mM sodium citrate con- taining 0.1% triton-X100 for 1 hour. After an overnight incubation at 4°C with the primary antibody (0.5 to 1.0 ?g/mL), the cells were washed with sodium citrate con- taining 0.05% triton-X100 and incubated for 1 hour at room temperature with secondary antibody (rhodamine red-goat-anti-rabbit, Molecular Probes, Eugene, Ore) at a dilution of 1:200. Coverslips were washed and mounted on glass slides for viewing. supplemented with 1% fetal bovine serum and 1 ng/mL recombinant human FGF-2. As serum contains agents known to stimulate MMP biosynthesis,104low serum con- centration was used to minimize nonspecific induction of MMPs. The cultures were incubated at 37ºC in a humid- ified atmosphere of 95% air and 5% CO2. Prostaglandin Treatments The culture medium was changed to fresh medium sup- plemented with PGF2?, 17-phenyltrinor-PGF2?, PhXA85 (latanoprost acid) (Cayman Chemical Co, Grand Rapids, Mich), or vehicle control. 17-phenyltrinor-PGF2?and PhXA85 bind with greater specificity to the FP receptor (the endogenous PG-receptor that preferentially recog- nizes F-type prostaglandins) than PGF2?. Each PG was tested at concentrations of 100 nM, 200 nM, and 500 nM. PG concentrations were chosen on the basis of their receptor binding profiles as well as the observation that the peak concentration of PhXA85 observed in aqueous humor following topical application of a clinical dose of latanoprost to human eyes is approximately 100 nM.105,106 Exposure durations of 24, 48, and 72 hours were chosen on the basis of previous experiments that found increased MMPs in ciliary smooth muscle cells exposed to PGs for 24 hours to 72 hours.107,108Experimental treatment was initiated by addition of the test PGs prepared from 10 mM stock solutions in ethanol and appropriately diluted with DMEM-12 nutrient mixture. 3. AND MMPS WITH PROSTAGLANDIN EXPOSURE If there is increased MMP-1, MMP-2, and MMP-3 immunoreactivity in the sclera of monkey eyes that have received topical PGF2?-IE treatment, it would be unclear whether it is a direct response, reflecting increased pro- duction within sclera, or an indirect consequence of increased MMP release into the suprachoroidal space of the uveoscleral outflow pathway by ciliary muscle cells. The following experiments were undertaken to investigate this question by determining whether exposure of organ cultures of human sclera to various PGs increases scleral permeability and whether this is associated with increased release of MMPs. MEASUREMENT OF HUMAN SCLERAL PERMEABILITY Permeability Analysis Following 1- to 3-day incubation with test PG or vehicle control, the scleral tissue was clamped into the in vitro perfusion apparatus (Ussing apparatus, model CHM2; World Precision Instruments Inc, Sarasota, Fla). The 2 chambers, each with a 9 mm-diameter opening, sand- wiched a 14 mm-diameter piece of scleral tissue. This assembly was held together with a clamp. Each unstirred chamber contained 0.75 mL and could be filled, drained, and purged through 3 ports. Three rhodamine-dextran polymers (Molecular Probes) (MW, 10,000, 40,000, and 70,000 kDa) were diluted in phenol red-free HBSS (250 µg/mL). The “uveal-side” chamber was filled with phenol- free HBSS, and the “orbital-side” chamber was filled with rhodamine-dextran diluted in phenol red-free HBSS. Permeability was assessed in this direction because the orbital side was smoother than the uveal side, and thus the potential for measurement-altering small leaks around the edge of the tissue piece was less. Solutions were freshly prepared and warmed to 37ºC prior to use. After assem- bly and filling, the system was placed in the 37ºC incuba- tor. The apparatus was checked after 30 minutes to verify that no leaks were present. Any leaks of the dextran solution were readily apparent owing to the dark red color Human Scleral Organ Cultures Twenty-three pairs of human eyes from donors 45 to 80 years old were obtained within 24 hours after death. Enucleation was completed within 6 hours postmortem, and the eyes were stored in a moist chamber at 4ºC for less than 24 hours prior to generation of the organ cultures. Donors had no known history of glaucoma or other eye dis- ease. The eyes were placed in Dulbecco’s modified Eagle’s medium and Ham’s F12 nutrient mixture (DMEM-F12) medium containing 50 U/mL penicillin and 50 µg/mL streptomycin for 15 minutes. This was repeated twice prior to dissection. Tenon’s capsule and episclera were removed from the surface of the sclera using a sterile cot- ton-tip applicator. Curved scissors were used to excise cir- cular pieces of sclera. The chosen areas were selected to avoid the perforating anterior ciliary vessels and the vortex veins. The uveal tissues and retina were gently removed from the vitreous side of the sclera with a cotton-tipped applicator. The circular pieces of scleral tissue were placed into 12-well culture plates containing DMEM-F12 327

  10. Weinreb, Thesis 11/9/01 11:30 AM Page 328 Weinreb of the solution Leaks of the phenol red-free Hanks from the uveal-side chamber were recognized by reduction of the level of the fluid visible through the clear walls of the chamber. Four hours later, a 750 µL sample was removed through a valved port connected to the uveal-side cham- ber and stored at -80ºC. Samples were protected from light at all times before fluorescence measurement. minutes. This stain cannot pass through the plasma mem- brane of living cells but readily stains DNA within dead cells. The cultures were rinsed twice with (PBS). The cul- tures were then homogenized in PBS using a homogeniz- er. The homogenates were centrifuged and the super- natants were collected. Cell viability in these samples was determined by first measuring Sytox green fluorescence using a spectrofluo- rimeter (model SFM 25, Kontron, Zürich, Switzerland) with the excitation and emission wavelengths set at 500 and 525 nm, respectively. The amount of ethidium homod- imer was then measured using a 550 nm excitation wave- length. This wavelength excited ethidium homodimer at 83% of maximal efficiency, but minimally excited Sytox green. The emission wavelength analyzed was 650 nm because it retained 71% of maximal efficiency for ethidium homodimer and eliminated greater than 99% of the cross- talk signal coming from Sytox green. The photomultiplier voltage was optimized to 480 V to obtain all readings on 1 setting. The signals from the ethidium homodimer were normalized with signals from the Sytox green by dividing the ethidium homodimer results by the Sytox green results. Positive (live) controls were fresh cultures not exposed to any treatment, and negative (dead) controls were cultures first treated with 2% paraformaldehyde for 10 minutes and permeabilized with graded methanols before evaluation. The viability of each sample was deter- mined by interpolation from a standard curve that was generated by plotting positive and negative control values. Scleral Permeability Coefficient Diffusion from the “orbital” chamber to the “uveal” cham- ber was characterized by means of a permeability coeffi- cient (Pc), which is the ratio of steady-state flux (the mass of solute crossing a planar unit surface normal to the direction of transport per unit time) to the concentration gradient.64In this study, the concentration of “uveal-side” chamber, CU, was less than 1% of the concentration in the “orbital” chamber, Cowhich did not change measurably over the course of the experiment. Hence, the permeabil- ity coefficient was calculated as follows: Pc (cm/sec) = (CUt-CU0.5)V / AtCo where CU0.5and CUtare the concentration in the “uveal” chamber at 0.5 hour and at t hours, respectively. Cois the initial drug concentration (0.25 mg/mL), A is the surface area of exposed sclera (0.65 cm2). V is volume of the each chamber (0.75 ml), and t is duration of steady-state flux converted from hours to seconds. The term (CUt- CU0.5) / t is the permeation rate of dextran across each excised scleral piece (µg/hr). The fluorescence of rhodamine-dextran was meas- ured with a spectrofluorimeter at room temperature. The excitation and emission wavelengths were 550 and 580 nm, respectively. Standard curves of fluorescence versus concentration were obtained by serial dilution of rho- damine-dextran dissolved in diffusion medium (phenol red–free Hank’s buffered saline solution). Scleral Hydration Analysis Thirty scleral specimens were obtained from human eye bank eyes for the determination of scleral hydration. These studies were performed to ensure that maintaining sclera in the Ussing perfusion system did not hydrate the sclera, which may alter scleral permeability. Ten circular scleral pieces from 3-day-old preparations were incubated in DMEM/F-12 media only, or with media for 3 days fol- lowed by HBSS for an additional 4 hours. The prepara- tions were then weighed by using an analytical balance (accuracy, 0.0001 g, Mettler, Geissen, Germany), dried to constant weight at 100°C for 24 hours, placed immediate- ly in a tissue desiccator to cool for 30 minutes, and reweighed. Another 20 circular scleral pieces from fresh and 3-day-old moist chamber-stored globes perfused with HBSS and without perfusion were used to evaluate poten- tial effects of storage. The level of hydration in each piece of sclera was calculated by the following equation: Viability After Prolonged Exposure to Prostaglandins To assess viability in vitro, scleral organ cultures were incubated with 500 nM of each PG, the highest dose in this study, for 1, 2, or 3 days. Ethidium-homodimer was then added to the cultures to a final concentration of 1 µM and the cultures were returned to the incubator for 30 minutes (Molecular Probes). As this dye cannot penetrate living cells, it is only bound to the DNA of dead cells in the cultures. The scleral cultures were rinsed with phos- phate buffered saline without phenol red and then exposed to 2% paraformaldehyde in phosphate buffered saline for 10 minutes. The cultures were then permeabi- lized by passage through graded methanols (50%, 70%, 90%, 95%, and 100%), rehydrated, rinsed in phosphate buffered saline, and exposed to 5 mM Sytox green for 15 mg H2O / mg tissue = (wet weight - dry weight) / dry weight. MMP Immunosorbant Assays At the conclusion of the 1- to 3-day incubations with PGs 328

  11. Weinreb, Thesis 11/9/01 11:30 AM Page 329 Enhancement of Scleral Macromolecular Permeability with Prostaglandins or vehicle, media samples were collected from the scleral cultures for enzyme-linked immunosorbant assay (ELISA) analysis. Measurements of MMP-1, -2, and -3 concentration were performed with commercially avail- able ELISA kits (Biotrak, Amersham Pharmacia Biotech Inc, Piscataway, NJ). These assays are based on a two-site ELISA “sandwich” format, and detected both latent and active MMPs. For the MMP-1 assay, purified MMP stan- dards and samples were incubated in microtiter wells pre- coated with anti-MMP-1 antibody. The wells were then washed, incubated with second polyclonal antibody to MMP-1, washed, incubated with anti-rabbit horseradish peroxidase, washed, and developed by tetramethyl benzi- dine. After development at room temperature, the absorbency was measured at 630 nm using a microtiter plate reader (SpectraMax 250, Molecular Devices, Sunnyvale, Calif.). The procedures for the MMP-2 and MMP-3 assays were the same except the antibodies were to MMP-2 and MMP-3, respectively. nerve and artery, insertion of muscles, or vortex veins in each center area. Uveal tissue and retina were gently removed with a cotton-tipped applicator. Scleral pieces were placed into 12-well plates containing DMEM/F-12 supplemented with 1% fetal calf serum and 1 ng/mL human recombinant FGF-2. The low concentration of serum was used to minimize nonspecific increase of MMP because serum contains various stimulating factors of MMP synthesis. The explants were incubated at 37°C in a humidified atmosphere of 95% air and 5% CO2. Latanoprost Acid Pretreatment To investigate the effect of latanoprost acid on the scleral permeability, the culture medium was changed to fresh medium supplemented with latanoprost acid (Cayman Chemical Co). Tested concentrations included 50, 100, and 200 nM, because the peak concentration in human aqueous humor following topical application of a clinical dose is ~100 nM.106After 3 days exposure, the permeabil- ity assay was performed. Statistical Evaluation Experimental differences between control culture results and a single treatment group were evaluated using the Student t test. When results from several treatment groups were compared to those of a single control, signif- icance was evaluated using analysis of variance and the Student Newman Keuls test. A P value less than .05 was considered statistically significant. Scleral Permeability Analysis After 3 days’ incubation, the scleral tissue was clamped into the in vitro Ussing perfusion apparatus. Each cham- ber contained 0.75 mL of fluid. The 2 chambers were fac- ing an opening of 9 mm in diameter and were held together by a screw clamp. The scleral tissue was washed twice in phenol red-free HBSS to remove culture medium and carefully sandwiched to avoid vortex veins between both chambers. Both chambers were tightly clamped to avoid leakage of the medium. Each chamber had 3 ports to fill and drain samples. Tested molecules included human recombinant FGF-2 (16kD, R & D Systems, Minneapolis, Minn) and 10 kDa rhodamine-dextran poly- mer (Molecular Probes). This dextran was included in the analysis because it is stable in tissue, it has no physiologic activity, and its transscleral movement has been previous- ly characterized in normal sclera.64Rhodamine-dextran or FGF-2 was diluted in phenol red-free HBSS and applied in the orbital side chamber. After checking that there were no leaks in the uveal side chamber, phenol red-free HBSS was filled in this side. After assembly of Ussing chambers, the system was incubated at 37°C in a humidi- fied atmosphere of 95% air and 5% CO2. The assay of FGF-2 and dextran was dependently performed in the same scleral tissue. After the intended time, each sample was drained from the uveal side chamber and stored in a light-protected box. 4. PE RME ATION PROSTAGLANDIN EXPOSURE These experiments were undertaken to determine whether exposure of scleral explants to the PG analogue latanoprost acid increases permeability to fibroblast growth factor-2 (FGF-2) (also known as basic fibroblast growth factor). MEASUREMENT OF FIBROBLAST GROWTH FACTOR-2 THROUGH HUMAN SCL E RA WITH Human Scleral Tissue Explant Eight pairs of human eyes from donors were obtained from the San Diego Eye Bank. Donors had no history of glaucoma or other ocular diseases. The mean age was 70 ± 6 (mean ± SD) years old. Each pair of eyes was enu- cleated within 5 hours after death and immediately pre- served in a moist chamber at 4°C. Apparently intact eyes were selected and any eye showing scleral damage or thin sclera (posterior staphyloma) was not used. Within 24 hours after preservation, the sclera was dissected and placed into organ culture. Briefly, after incubation in HBSS medium containing 50 U/mL of penicillin and 50 U/mL of streptomycin for 30 minutes, residual extraocular muscles and orbital connective tissues were removed. Sclera was dissected into 4 pieces to exclude the long ciliary Measurement of Dextran Rhodamine-dextran concentration in the HBSS collected from the uveal side chambers was determined using a spectrofluorimeter. The excitation and emission 329

  12. Weinreb, Thesis 11/9/01 11:30 AM Page 330 Weinreb wavelengths were 550 and 580 nm, respectively. Standard curves of fluorescence versus concentrations were obtained by serial dilution of rhodamine-dextran dissolved in phenol red-free HBSS. Each sample was immediately measured 8 times, and the measurements were averaged. (Fig 1). Minimal MMP-3 immunoreactivity was present in the vehicle-treated eyes. In the eyes treated with PGF2?-IE, there was increased MMP-1 and MMP-2 immunoreactivity in the sclera when compared with the corresponding vehicle- treated eye (Fig 1). Compared to the vehicle-treated eyes, moderate MMP-3 staining also was observed in scle- ra of the treated eyes. Measurement of FGF-2 Concentration FGF-2 concentration in the medium collected from the Ussing chamber was measured using a sandwich enzyme immunosorbent assay (R & D Systems). Optical density was measured at 450 nm and 540 nm using a microtiter plate reader (SpectraMax 250, Molecular Devices). To correct for nonspecific variation, the absorbence value at 540 nm was subtracted from that of 450 nm. Standard curves of absorbency versus concentrations were obtained by serial dilution of standard purified FGF-2. Densitometric Analysis The intensity of immunostaining was assessed along 2 lines placed over the image of the sclera observed with an imaging densitometer (Fig 2). Compared with the vehi- cle-treated eyes, there was increased MMP-1 immunore- activity in all treated eyes (Table I). Table II shows the combined scores for MMP-1, MMP-2, and MMP-3. Overall, the optical density score for MMP-1 in the sclera in the treated eyes was increased by 63 ± 35% (mean ± SD). Similarly, the optical density score for MMP-2 was increased by 267 ± 210%, and the MMP-3 optical density score in the treated eyes was increased by 726 ± 500%. In Permeability Coefficient Determination Diffusion from the orbital chamber to the uveal chamber was characterized by determination of a permeability coef- ficient (Pc), which is the ratio of steady-state flux to the concentration gradient.64In this study, the concentration of agents in the uveal side chamber, CU, was less than 1% of it in the orbital chamber, Co, thus the change of Cowas assumed to be under the limit of detection. Hence, the permeability coefficient was calculated as follows: Pc (cm/sec) = (CUt-CU0)V/CotS where CU0and CUtare the concentration in the Ussing chamber at 0 hour and at t hours, respectively. COis the initial drug concentration in the orbital chamber. V is a volume of each chamber (0.75 mL), and t is a duration time of steady-state flux converted the unit from hour to second. S is the surface area of exposed sclera (0.65 cm2). Statistical Evaluation At least 7 experiments were performed on FGF-2 and dextran at each concentration of latanoprost acid. Each group was compared by using a Student’s t test. A P value less than 0.05 was considered statistically significant. All data are presented as mean ± SD. RESULTS 1. (MMPS) AFTER TOPICAL PROSTAGLANDIN MEASUREMENT OF SCLERAL METALLOPROTEINASES FIGURE 1 Anterior-segment tissues comparing hematoxylin and eosin staining (A) with MMP-1 immunoreactivity (B) containing sclera (s), ciliary body (cb), Schlemm's canal (sc), and iris (i). Dark areas in panel B indicate MMP-1 immunoreactivity. Box in panel B indicates scleral region shown at high magnification in panel C. D, Comparable area of sclera from eye that received topical PGF2?-IE shows increased MMP-1 immunoreactivity (magnification: A ? 30; B ? 39; C and D ? 116). Immunohistochemistry In the vehicle-treated monkey eyes, moderate immunore- activity for MMP-1 was observed in the sclera (Fig 1). The distribution of MMP-2 immunoreactivity in the vehi- cle-treated eyes was similar with diffuse light staining 330

  13. Weinreb, Thesis 11/9/01 11:30 AM Page 331 Enhancement of Scleral Macromolecular Permeability with Prostaglandins FP-specific product of 1,186 base pair (bp). Figure 3 shows an ethidium-stained agarose gel with the PCR products obtained from cDNA prepared using 3 different primer FIGURE 2 Comparison of bright field image of anterior segment (A) with densitome- try image (B) showing placement of 2 measurement lines over sclera. Optical density along these lines was integrated and then mean optical den- sity was determined by dividing the integrated score by the length of the measurement line (magnification ? 56). TABLE I: INCREASE OF MMP-1 IMMUNOREACTIVITY IN SCLERA OF 4 MONKEYS FOLLOWING 5 DAYS OF TOPICAL PGF2?-IE TREATMENT CONTROL EYE* TREATED EYE* % INCREASE MONKEY FIGURE 3 Reverse transcription–polymerase chain reaction (RT-PCR) of total RNA isolated from human sclera tissue and amplified with specific primers for human prostanoid FP receptor and for glyceraldehyde-3-phosphate dehy- drogenase (GAPDH). The products were separated by electrophoresis on a 1% DNA agarose gel. Standards (lanes 1 and 2) are 1 kilobase (kb) and 100 base pair (bp) DNA ladder (Gibco, BRL), respectively. Each reaction con- dition is represented. Lanes 3 (oligo DT alone), 4 (random primer alone), and 5 (oligo DT+random primer) yielded the predicted product (1,186 bp) after RT-PCR using total RNA isolated from donor sclera tissue, respective- ly. Lanes 6 (oligo DT alone), 7 (random primer alone), and 8 (oligo DT+ran- dom primer) represent the products obtained using specific primers for GAPDH from the identical RNA sample. The predicted product (299 bp) was obtained in each condition. The control (Gibco, BRL) for the cDNA synthesis reaction also yielded the predicted product size (500 bp, lane 9). 1 2 3 4 0.0193 0.0219 0.0151 0.0229 0.0240 0.0458 0.0230 0.0386 24 109 51 68 Mean ± SD 0.0198 ± 0.0034 0.0328 ± 0.0112 63 ± 35 IE, isopropyl ester. *Mean optical density. TABLE II: MEAN INCREASE OF MMP IMMUNOREACTIVITY IN MONKEY SCLERA FOLLOWING 5 DAYS OF TOPICAL PGF2?-IE TREATMENT % INCREASE (± SD) P VALUE* conditions from total RNA isolated from a single donor eye (lanes 1 through 3). Additionally, PCR products using primers specific for GAPDH that were predicted to yield a product of 299 bp were analyzed from the same cDNA samples used for amplification of the FP receptor transcript (lanes 4 through 6). The RT-PCR kit control was in vitro transcribed RNA from the chloramphenicol acetyltrans- ferase (CAT) gene that was engineered to contain a 3? poly(A) tail. Gene-specific primers used were predicted to yield a CAT-specific product of 500 bp (lane 9). As shown, products of the expected sizes were obtained in all conditions with the RNA isolated from the donor sclera. Because the primers used for the amplification of the FP receptor mRNA were chosen to span an intron within the FP receptor gene, the PCR products did not result from the amplification of genomic DNA and are consistent with the presence of mRNA encoding a human prostanoid FP receptor. MMP TYPE ALTERNATE NAME MMP-1 MMP-2 MMP-3 Interstitial collagenase Gelatinase A Stromelysin-1 63 ± 35 267 ± 210 729 ± 500 0.01 0.005 0.02 IE, isopropyl ester. * P value by Student's t test; N=4. each case, the increases in the treated eyes were statisti- cally significant when compared to the vehicle-treated eyes (Table II). 2. EXPRESSION IN HUMAN SCLERA FP Receptor Transcripts in Human Sclera To confirm that human sclera tissue contained mRNA that encodes the prostanoid FP receptor, RT-PCR was per- formed with primers that were predicted to yield a FP RECEPTOR GENE TRANSCRIPTION AND PROTEIN 331

  14. Weinreb, Thesis 11/9/01 11:30 AM Page 332 Weinreb FP Receptor Protein in Human Sclera Immunoreactivity for the FP receptor was observed with- in the cytoplasm of the scleral fibroblasts (Fig 4). The intensity of this granular staining was similar throughout the fibroblast processes that extended between the scler- al collagen bundles. No staining of these collagen bundles was observed. exposure time; however, these increases ranged from 7% to 21%. These permeability increases were statistically significant only on day 3 in the case of 100 nM PGF2?, but were significant for 200 nM or 500 nM PGF2? on all 3 days. Similar to the 40 kDa dextran, the flux of 70 kDa dextran increased with PGF2?dose and exposure time by 5% to 28%. These increases were significant with longer treatments at 100 nM or 200 nM, and were significant with 500 nM PGF2? on all 3 days. Incubation of scleral cultures with 17-phenyltrinor- PGF2?also increased permeability of the scleral organ cul- tures to the labeled dextrans. Permeability to the 10 kDa tracer increased in a dose-dependent and time-dependent manner from 5% to 183% (Fig 6). These increases were significant for all conditions except 100 nM 17-phenyltri- nor-PGF2?exposure for 1 day. Permeability to the 40 kDa tracer increased in a dose- and time-dependent manner from 4% to 31%. These increases were significant at all concentrations tested on days 2 and 3. Permeability to the 70 kDa tracer increased from 9% to 24%. These increases were significant at all concentrations and times measured. Overall, the increases observed with 17-phenyltrinor- PGF2?were similar to the increases observed with PGF2?. The exception to this was the larger permeability increase observed at 3 days with 100 nM 17-phenyltrinor-PGF2? than with 100 nM PGF2?. PhXA85 generally induced mod- erately larger increases in scleral permeability than PGF2? or 17-phenyltrinor-PGF2?(Fig 7). Flux of the 10 kDa trac- er was increased by 45% to 213%. These increases were dose-dependent, became larger as exposure time increased up to 3 days, and were significant for all conditions. The flux of 40 kDa dextran also increased with increasing PhXA85 and exposure time; however, these increases ranged from 6% to 41%. These increases were significant for all conditions except 100 nM PhXA85 exposure for 1 day. Similar to 40 kDa dextran, the flux of 70 kDa dextran increased with PhXA85 dose and exposure time by 13% to 48%. Also, these increases were significant for all condi- tions except 100 nM PhXA85 exposure for 1 day. 3. AND MMPS WITH PROSTAGLANDIN EXPOSURE MEASUREMENT OF HUMAN SCLERAL PERMEABILITY Scleral Permeability Scleral permeability was measured by assessing the flux of labeled dextrans across the scleral cultures in a Ussing chamber. Dextrans of different sizes were evaluated to model the potential differences among aqueous proteins of different sizes. As shown in Fig 5, flux across the scle- ral cultures incubated without PGs was 1.5 x 10-6cm/sec- ond for 10 kDa dextran, 0.7 x 10-6cm/second for 40 kDa dextran, and 0.4 x 10-6cm/second for 70 kDa dextran. Moreover, these fluxes did not change among cultures incubated without PGs for 1, 2, or 3 days. In contrast, incubation with PGF2?significantly increased the flux of the 10 kDa tracer. These increases ranged from 21% to 124%, were dose-dependent, became larger as exposure time increased up to 3 days, and were significant for all concentration and tested time points (P<.05). The flux of 40 kDa dextran also increased with increasing PGF2?and Viability of Sclera Survival of cells in the organ culture was assessed by measuring the exclusion of ethidium homodimer, a vital stain that binds to DNA. The standard for maximal viabil- ity was freshly obtained donor sclera, and the standard for complete loss of viability was donor sclera that had been exposed to 2% paraformaldehyde prior to ethidium homodimer exposure. As shown in Fig 8, viability for all cultures was about 83% on day 1, 81% on day 2, and 80% on day 3. Differences of viability among cultures exposed to 500 nM PGF2?, or 17-phenyltrinor-PGF2?were less than 1% on all 3 days. This suggests that incubation with FIGURE 4 FP receptor immunoreactivity within human scleral fibroblasts (A, B) and staining control (C, D). For each specimen, the bright-field image (A, C) and the fluorescence image (B, D) are shown. Staining was similar through- out fibroblast processes that extended between the scleral collagen bundles. Control sections were processed without exposure to primary antibody (magnification, x280). 332

  15. Weinreb, Thesis 11/9/01 11:30 AM Page 333 Enhancement of Scleral Macromolecular Permeability with Prostaglandins FIGURE 5 Scleral permeability after PGF2?exposure. Permeability determined by the transscleral movement of 10 kDa (A), 40 kDa (B), or 70 kDa (C) dextrans across treated sclera. Values represent mean (( 10-6 cm/sec) ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=4). FIGURE 6 Scleral permeability after 17-phenyltrinor-PGF2?exposure. Permeability determined by the transscleral movement of 10 kDa (A), 40 kDa (B), or 70 kDa (C) dextrans across treated sclera. Values represent mean (( 10-6 cm/sec) ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=4). FIGURE 7 Scleral permeability after PhXA85 exposure. Permeability determined by transscleral movement of 10 kDa (A), 40 kDa (B), or 70 kDa (C) dextrans across treated sclera. Values represent mean (( 10-6 cm/sec) ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=4). 333

  16. Weinreb, Thesis 11/9/01 11:30 AM Page 334 Weinreb exposure, and were significant only for the higher concen- trations and longer incubation times examined. Overall, there were slight increases of MMP-1 with increasing dose, and the effects of the different PGs tested were similar. In contrast to MMP-1, increases in MMP-2 were much larger and ranged from 124% to 267% (Fig 11). These increases were significant in every condition exam- ined and showed marked increases with increasing time of exposure. Overall, there were slight increases of MMP-2 with increasing PG concentration. The magnitude of the effects was least with 17-phenyltrinor-PGF2?, intermedi- ate with PGF2?, and greatest with PhXA85. MMP-3 concentration also increased in the media of cultures exposed to PGF2?, 17-phenyltrinor-PGF2? or PhXA85 (Fig 12). These increases ranged up to 96% and were larger than seen with MMP-1 but smaller than seen with MMP-2. These increases were clearly time-depend- ent, being generally insignificant on day 1 and significant on days 2 and 3. Dose dependence was clearly present with PhXA85 at every time point and less clear with PGF2?or 17-phenyltrinor-PGF2?. FIGURE 8 Cell viability within organ-cultured human sclera incubated for 3 days with vehicle, PGF2?, 17-phenyltrinor-PGF2?, or PhXA85. Concentration of each prostaglandin was 500 nM. Values represent mean ± SD (N=4). these PGs for 3 days had minimal influence on cell sur- vival in the scleral cultures. 4. PE RME ATION PROSTAGLANDIN EXPOSURE MEASUREMENT OF FIBROBLAST GROWTH FACTOR-2 THROUGH HUMAN Scleral Hydration To evaluate whether changes in scleral hydration occur with the culture conditions, the water content in the scle- ral cultures was determined in scleral cultures exposed to HBSS for 4 hours at room temperature, to complete cul- ture medium for 3 days at 37°C, or to complete culture medium for 3 days followed by 4 hours in HBSS. The mean water content, or scleral hydration, of fresh sclera was 3.05 ± 0.11 mg water/mg dry weight (N=5). As shown in Fig 9, scleral cultures incubated 4 hours in HBSS alone, in medium for 3 days, or in medium for 3 days followed by 4 hours in HBSS were insignificantly different from the fresh cultures (P<.05, Student Newman Keuls test). This indicated that these culture conditions did not alter hydra- tion within the scleral cultures. SCL E RA WITH Time Course Analysis The time course of FGF-2 penetration of sclera within the Ussing chamber was assessed by withdrawing a 40 µL sample from the test side at 30-minute intervals. As shown in Fig 13, the concentration increased linearly for the duration of the experiment. MMP Release Induced by Prostaglandin Treatments One possible explanation for the observed increases in scleral permeability following exposure to the PGs is reduction in collagen content by MMPs. Hence, the media of scleral cultures incubated with PGF2?, 17- phenyltrinor-PGF2?or PhXA85 were assayed for changes in the concentration of MMP-1, MMP-2, and MMP-3. Among cultures incubated in control medium for 1, 2, or 3 days, there were no significant changes in the concentra- tion of MMP-1, MMP-2, or MMP-3 (Figs 10 through 12). Evaluation of MMP-1 in the media of the treated cul- tures showed moderate increases in cultures exposed to PGF2?, 17-phenyltrinor-PGF2?or PhXA85 (Fig 10). These increases ranged up to 37%, increased with time of FIGURE 9 Evaluation of scleral hydration in fresh sclera, scleral organ cultures follow- ing incubation with Hank's balanced salt solution (BSS) for 4 hours (4hrs- BSS), tissue culture medium for 3 days (3days-Medium), or in culture incu- bated in medium for 3 days followed by 4 hours in Hank’s balanced salt solu- tion (3 days Medium + 4 hrs BSS). Values represent mean ± SD (N=5). 334

  17. Weinreb, Thesis 11/9/01 11:30 AM Page 335 Enhancement of Scleral Macromolecular Permeability with Prostaglandins FIGURE 10 Concentration of MMP-1 in the medium of scleral organ culture exposed to PhXA85 (A), PGF2?(B), or 17-phenyltrinor- PGF2?(C) for 1 to 3 days as deter- mined by ELISA. Cultures analyzed were from cultures generated from a 76-year-old male donor, a 66-year-old male donor, and a 45-year-old female donor (N=3). Values represent mean ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=4). FIGURE 11 Concentration of MMP-2 in the medium of scleral organ culture exposed to PhXA85 (A), PGF2?(B), or 17-phenyltrinor-PGF2?(C) for 1 to 3 days as deter- mined by ELISA. Cultures analyzed were from cultures generated from a 76-year-old male donor, an 80-year-old male donor, and a 66-year-old male donor Values represent mean ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=3). FIGURE 12 Concentration of MMP-3 in the medium of scleral organ culture exposed to PhXA85 (A), PGF2?(B), or 17-phenyltrinor- PGF2?(C) for 1 to 3 days as deter- mined by ELISA. Cultures analyzed were from cultures generated from a 76-year-old male donor, a 66-year-old male donor, and a 45-year-old female donor. Values represent mean ± SD. Asterisk indicates P<.05 by the Student Newman Keuls test (N=3). 335

  18. Weinreb, Thesis 11/9/01 11:30 AM Page 336 Weinreb DISCUSSION The limits of the possible are enlarged. Ralph Waldo Emerson In contrast to the sclera being “inert and purely support- ive in function,”54these studies clearly demonstrate that it has the potential to be metabolically active, to be pharma- cologically responsive, and to have other functions in addi- tion to structural support. Moreover, PGs can directly induce sclera to undergo structural modifications that enhance transscleral permeability, a response that is likely mediated by FP receptors on scleral fibroblasts. These results may have important implications for the facilita- tion of macromolecule delivery to posterior segment tis- sues. FIGURE 13 Time course of FGF-2 concentration in receiving Ussing chamber fitted with human sclera previously cultured for 3 days in control medium. Data presented as mean ± SD (( 10-8 cm/sec). Increase in concentration with time was linear (R2=0.91). Based on this, permeability was determined to be 1.68 x 10-8cm/sec. N=4 pairs of donor eyes. Mean age of donors was 67.5 ± 2.9 (SD) years. 1. SCLERAL COLLAGEN BY INCREASING SCLERAL METALLO- PROTEINASES TOPICAL PROSTAGLANDIN ADMINISTRATION REDUCES The quantitative immunohistochemistry results show that topical treatment of monkey eyes with PGF2?-IE increased expression of MMP-1, MMP-2, and MMP-3 in the sclera adjacent to the ciliary muscle. Increased MMP biosynthesis could then reduce scleral collagens and other extracellular matrix molecules after treatment with topical PGF2?-IE. This reduction has been previously confirmed for collagen type I, collagen type III, and collagen type IV.82MMP-1 is known to hydrolyze a specific site found in collagen types I and III.109,110 known to hydrolyze specific sites found in collagen type IV as well as in fibronectin. MMP-3 is known to hydrolyze specific sites found in collagen types III and IV, as well as Likewise, MMP-2 is FIGURE 14 Scleral permeability to FGF-2 after exposure to various concentrations of PhXA85 for 3 days. Data presented as mean ± SD (x 10-8cm/sec). Asterisk indicates P<.05 by the Student Newman Keuls test. N=8 pairs of donor eyes. Mean age of donors was 71.8 ± 6.9 (SD) years. Dose Response Analysis Increasing the concentration of latanoprost acid in scleral explant cultures maintained for 3 days increased the per- meability of both FGF-2 and 10 kDa dextran (Figs 14 and 15). FGF-2 permeability following 50 or 100 nM latanoprost acid was increased by an average of 56 ± 77% and 126 ± 120%, respectively (mean, ± SD, N=8). FGF- 2 permeability in sclera incubated with 200 nM latanoprost acid was similar to sclera incubated with 100 nM latanoprost acid. In contrast, 10-kDa dextran perme- ability following 50, 100, or 200 nM latanoprost acid was increased by an average of 50 ± 24%, 39 ± 19%, and 48 ± 24%, respectively. The ratio of FGF-2 to 10 kDa dextran permeability ranged from 40-fold to 90-fold; however, there was no clear relationship between the magnitude of the ratio and the latanoprost acid dose. FIGURE 15 Scleral permeability to 10 kDa dextran after exposure to various concentra- tions of PhXA85 for 3 days. Data presented as mean ± SD (x 10-6cm/sec). Asterisk indicates P<.05 by the Student Newman Keuls test. N=8 pairs of donor eyes. Mean age of donors was 71.8 ± 6.9 (SD) years. 336

  19. Weinreb, Thesis 11/9/01 11:30 AM Page 337 Enhancement of Scleral Macromolecular Permeability with Prostaglandins in fibronectin and laminin. Hence, the observed increas- es in MMPs-1, -2, and -3 suggest a concerted response leading to reduced scleral collagen. Other extracellular hydrolases also are likely to participate in the reduction of extracellular matrix. It should be noted that MMPs are secreted as inactive pro-enzymes that are subsequently activated by proteolytic truncation.111,112Also, MMP activ- ity is regulated by the presence of tissue inhibitor of matrix metalloproteinases (TIMPs).113,114Each of the anti- bodies used can recognize both the proenzyme and the active enzyme. Thus, the magnitude of the increased MMP activity may be less than the magnitude of the increased immunoreactivity. which the EC50is 100 nM, and not activation of EP1 or other PG receptors. The EC50concentrations for PhXA85 activation of PG receptors other than the FP receptor are at least tenfold higher than the highest PhXA85 concen- tration tested.105 Hence, it is likely that the increased MMPs observed in the PG-treated scleral cultures were released by FP-receptor-mediated activation of scleral fibroblasts. These MMPs would be well positioned to ini- tiate collagen remodeling within the scleral stroma that enlarged intrascleral supramolecular passages and thereby facilitated transscleral protein permeability. As the MMPs in the present experiments could accumulate in the closed culture system, whereas they might dissipate upon secretion in situ, the concentration of the MMPs measured may be greater than the concentrations that might occur in scleral interstitial fluid in situ. The studies of dextran permeability indicate that PGs directly increase the permeability of human sclera in organ culture. This increase in permeability is accompa- nied by increased release of MMPs from scleral tissue. These changes are consistent with the reduced collagens observed in monkey sclera following topical PG treatment and suggest that remodeling of the scleral extracellular matrix may explain the increased permeability. Hydration analysis indicates that this response does not reflect any alteration of scleral hydration. Viability analysis indicates that this response is not associated with altered cell sur- vival in the experimental system, nor is there any evidence of toxicity due to the PG treatments. These permeability changes are likely to be normal physiologic responses, as they are both dose- and time-dependent. That the PG treatments also increased release of MMP-1, MMP-2, and MMP-3 in these cultures suggests that the permeability changes may reflect a direct response of scleral tissue to PG exposure and that the mechanism of increased trans- scleral permeability likely reflects intrascleral collagen remodeling. This proposed model is well supported by the findings of Dan and Yaron,76who observed increased flow of saline across bovine sclera and thinning of rabbit sclera in response to focal application of clostridial colla- genase, an MMP-1 analogue. The permeability relationships of the various sizes of labeled dextran observed in the present control scleral cultures is similar to the permeability relationships64of these tracers previously observed in sclera freshly dissect- ed from donor eyes. For example, the present study found that permeability of the 40 kDa dextran through the scleral organ cultures was 1.7-fold less than that of 10 kDa dextran. This is similar to the previous observation that 40 kDa dextran permeability is 1.4-fold to 3.8-fold less than 10 kDa dextran in freshly dissected sclera.64 Likewise, the present observation that permeability of 70 kDa dex- tran in the scleral organ cultures was 3.7-fold less than 2. PRESENT IN HUMAN SCLERA FP RECEPTOR GENE TRANSCRIPTS AND PROTEIN ARE As the biologic activity of a drug is often mediated by a specific receptor, the observation of FP receptor tran- scripts and protein within human sclera suggests that FP receptor agonists can directly activate these receptors and initiate physiologic and pharmacologic responses. The predominant cell type in the sclera is the scleral fibroblast. Also present are vascular endothelial cells within the pen- etrating blood vessels. Scleral fibroblasts are interspersed among the collagen layers that make up the scleral stroma and biosynthesize scleral collagen. The immunohisto- chemical results in these studies confirm that FP receptor in sclera is present on the scleral fibroblasts. Cultured fibroblasts from other tissues are known to increase their production of MMPs following stimulation with certain peptides.115-117 Hence, it is highly plausible that exposure of scleral fibroblasts to FP agonists may promote their biosynthesis of MMPs. The second cell type is the vascu- lar endothelial cell of blood vessels penetrating the sclera. FP receptors have been detected in association with other ocular blood vessels.90However, previous studies indicate that specific FP receptor agonists have minimal effects on the permeability of intraocular blood vessels.118,119 3. RAL PERMEABILITY AND MMPS PROSTAGLANDIN EXPOSURE INCREASES HUMAN SCLE- Pharmacologic considerations of the permeability and MMP changes observed with the tested PGs further sup- port involvement of FP receptor activation. The concen- trations of PGF2?and 17-phenyltrinor-PGF2?tested were greater than the EC50for activation of the FP receptor.105 It is possible that, if present, EP1 receptors also may have been activated by the PGF2?or 17-phenyltrinor-PGF2? treatments in this study as EC50’s for these agonists are 320 nM and 650 nM, respectively.105However, the response to PhXA85 is likely to reflect FP receptor activation, for 337

  20. Weinreb, Thesis 11/9/01 11:30 AM Page 338 Weinreb that of 10 kDa dextran is similar to the previous observa- tion that 70 kDa dextran permeability was 2.6-fold to 4.2- fold less than 10 kDa dextran permeability in freshly dis- sected donor eye sclera.64These similarities suggest that hydrodynamic constraints to macromolecule movement through the scleral organ cultures were similar to freshly dissected donor sclera. Hence, the scleral cultures repre- sent a reasonable model system in which to study modu- lation of transscleral macromolecule movement by PGs. The greater increase in 10 kDa dextran permeability through PG-treated scleral cultures than was observed with 40 kDa or 70 kDa dextran suggests that PGs may alter the size of intrascleral supramolecular passages. Scleral collagen fibrils are organized into bundles that vary in their organization according to position near the outer or inner wall of the sclera.50-52Overall, the bundles vary in width and thickness, often give off branches, and intertwine with each other. At the outermost layers, there is substantial irregular intermingling of collagen fibrils in adjacent bundles. Like sclera, synthetic hydrogels contain substantial water content and long polymer units charac- terized by chemical cross-links and polymer entangle- ments.120,121 Within pH-sensitive hydrogels, lower pH increases the size of pore channels through the matrix, while higher pH causes the gel network to swell with a resulting increase in the size of pore channels. Analysis of a pH-sensitive hydrogel confirmed that protein perme- ability is enhanced under conditions that increase the size of the pore channels.121Moreover, the magnitude of the permeability increase was greater with lower-molecular- weight proteins than with higher-molecular-weight pro- teins. This relationship among protein size, macromole- cule permeability, and pore size also has been seen in hydrogels in which pore size was altered by changing the size of polymer subunits used to synthesize the hydro- gel.121Hence, the greater permeability increases with the smaller dextran tracers that was observed in the PG-treat- ed scleral cultures is consistent with enlargement of the intrascleral supramolecular passages. The mechanism of increased permeability within the PG-treated scleral cultures is suggested by the increased amounts of MMP-1, MMP-2, and MMP-3 detected with- in the medium of the treated scleral cultures. Sclera con- tains collagen types I, III, VI, VIII, XII, possibly a small amount of collagen type V, as well as fibronectin.56,122-126Of these extracellular matrix components, MMP-1, MMP-2, and MMP-3 are known to cleave sites within collagen types I, III, V, and fibronectin.85,127 MMP-2, MMP-3, and MMP-9 has been found in cultures of human ciliary smooth muscle cells exposed to PGF2?, 17-phenyltrinor-PGF2?, and PhXA85.107,108 ments also induce reorganization of collagen type I, colla- gen type III, laminin, and collagen type IV within the human ciliary muscle cultures.128,129 MMPs within sclera are likely to mediate reduction of scleral collagens. Hence, increased 4. FACTOR-2 IS INCREASED WITH PROSTAGLANDIN EXPOSURE TRANSSCLERAL PERMEATION OF FIBROBLAST GROWTH The present results also indicate that exposure of human sclera to latanoprost acid, a prostaglandin analogue, increases FGF-2 permeability. This increase is dose dependent and increases with increasing exposure times. It is likely to reflect increased general permeability as it parallels increased permeability to 10 kDa rhodamine- labeled dextrans. The greater permeability of 10 kDa dextran may be related to binding of FGF-2 to molecules within the scle- ra. 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.57,58Much of the chondroitin sulfate, der- matan sulfate, and keratan sulfate may be covalently linked to the proteoglycans decorin, biglycan, and aggre- can.58Immunoreactivity of each of these proteoglycans is distributed throughout the thickness of sclera.58It is well established that FGF-2 strongly binds to heparan sulfate (Kd=0.34 µM).130Recently, it has been shown that FGF- 2 also can bind to dermatan sulfate (Kd=2.5 ?M).131As each of these GAGs is present within sclera, it is possible that binding of FGF-2 to these extracellular matrix com- ponents could impede the movement of FGF-2 through the sclera. Increased transscleral permeability to FGF-2 follow- ing PG treatments suggests that cotreatment with PGs could 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.132,133 effects were observed with concentrations as low as 20 pg/mL. Moreover, intraventricular infusion of FGF-2 can promote neuronal survival following experimental axoto- my, ischemia, neurotoxin treatment, or contusion of brain or spinal cord tissue.134-136However, infusion of FGF-2 can 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 specif- ic requirements for additional factors in the case of retinal ganglion cells.137Further, systemic infusion of FGF-2 also can stimulate responses in many nonneural tissues that may be either beneficial or detrimental to the desired neural tissue response.138Hence, the ability to enhance scleral penetration of FGF-2 using PGs may allow a small- er concentration of FGF-2 to be delivered directly to the eye with consequent reduction of systemic absorption. Beneficial Increased MMP-1, These treat- 338

  21. Weinreb, Thesis 11/9/01 11:30 AM Page 339 Enhancement of Scleral Macromolecular Permeability with Prostaglandins lial cells and transiently facilitate macromolecular move- ment from the suprachoroidal space to the retina.17,145 While the blood vessels in the optic nerve head, by virtue of tight junctions, have a blood-optic nerve barrier, it is significant that the optic nerve head itself is not thought to possess a blood-ocular barrier.147,148The border tissue of Elschnig (separating the peripapillary choroid and optic nerve head) allows choroidal interstitial tissue fluid to leak into the optic nerve head from the peripapil- lary choroid. Horseradish peroxidase (42 kDa) can enter the monkey optic nerve head from blood.148Glial cells at the edge of the optic disc form a barrier that prevents the spread of peroxidase into the retina. Despite these possible limitations, the prospect of increased transscleral permeability by PG cotreatment may allow sufficient transscleral transport to provide delivery of therapeutics to posterior segment tissues in concentrations not otherwise possible. This may be par- ticularly important for glaucomatous eyes, as elevated intraocular pressure may reduce scleral permeability, par- ticularly for macromolecules.81Besides being important for macromolecule delivery to the posterior segment, transscleral fluid movement through scleral stroma may be important for uveoscleral outflow. Further study is needed to determine if PG-induced increases in trans- scleral permeability contribute to increased uveoscleral outflow facility and decreased intraocular pressure observed following topical PG treatments. SIGNIF ICANCE INTRAOCULAR DRUG DELIVERY OF INCRE ASE D PE RME ABIL ITY F OR Assessment of drug diffusion in vitro permeability studies is a useful approach in the field of ocular pharmacokinet- ics to estimate drug movement for in vivo conditions. Intraocular absorption of a drug is directly related to the transport characteristics of absorptive tissues of the eye, such as the sclera. The increased scleral permeability fol- lowing PG exposure may have implications for facilitating delivery of therapeutics to the posterior segment of the eye. For example, growth factors that may facilitate reti- nal neuron survival range from 10 kDa to 40 kDa (Table III).1,16,139Because of their size, these molecules cannot readily cross the cornea. Hence, a noncorneal absorption route through sclera may facilitate usefulness of such therapeutics. It may not be sufficient only to increase scleral per- meability to large molecules, as there also may be other limiting factors for drug absorption in the posterior seg- TABLE III: MOLECULAR WEIGHT AND SIZE OF SEVERAL NEUROTROPHIC PROTEINS MOLECULAR WEIGHT (DA) MOLECULAR RADIUS (NM)* MOLECULE Fibroblast growth factor–2 (FGF-2) Ciliary neurotrophic factor Nerve growth factor-? Brain-derived neurotrophic factor (dimer) Neurotrophin-3 (dimer) Neurotrophin-4 (dimer) 18,000 22,800 26,000 27,200 27,200 28,000 2.1 2.3 2.5 2.5 2.5 2.5 CONCLUSION Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning. Winston Churchill *Estimated. The sclera is not inert, but metabolically active. PG treat- ment can increase scleral MMPs and facilitate transscler- al permeability. This can promote the transscleral move- ment of molecules as large as 70 kDa and may enhance transscleral delivery of a number of different ocular drugs. Several findings reported here provide support for a series of cellular and molecular events that could explain the mechanism of this response. These findings include observation of a concomitant increase in the amount of MMPs within the sclera of PG-treated monkey eyes, the presence of FP receptor and protein within human sclera, increased MMPs within the medium of human scleral cul- tures incubated with PGs, and increases in the amount of FGF-2 permeation within human scleral cultures exposed to increasing concentrations of latanoprost acid. Decreased collagen type I and III has been observed pre- viously in monkey sclera and ciliary muscle following repeated topical PG treatment.82Moreover, induction of ment. Orbital clearance, intraocular pressure, uveoscler- al outflow, and choroidal blood flow might each limit drug access to the target tissues. Blood-ocular barriers, espe- cially the blood-retinal barrier, also may be important.140-144 However, the integrity of the blood-ocular barriers can be disrupted under certain conditions.17,145,146As an example, fluorescein (MW, 376) particles that are not bound to albumin can pass through the spaces between the endothelial cells of the capillaries of the choriocapillaris, but normally they cannot leak through the retinal pigment epithelium and zonula occludens between adjacent retinal pigment epithelial cells. Fluorescein in the choroid can- not enter the neurosensory retina unless there is a defect in the retinal pigment epithelium. Therefore, if the tight junctions of the blood-retinal barrier preclude retinal drug absorption of a transsclerally delivered drug, it may be possible to minimally damage retinal pigment epithe- 339

  22. Weinreb, Thesis 11/9/01 11:30 AM Page 340 Weinreb c-Fos, MMP synthesis, and collagen reduction in ciliary muscle cell cultures exposed to PGs also have been described.128,129,149Together, these observations suggest a cellular mechanism for the increased transscleral perme- ability occurring with PG exposure. As summarized in Fig 16, a PG diffuses into the scleral stroma following topical or periocular treatment. Within the sclera, it then binds to FP receptors on scleral fibroblasts. This triggers a cas- cade of molecular events that increase MMP gene tran- the postmortem human eyes. Finally, I would like to acknowledge Hilda Krestyn, whose administrative and organization support are unparalleled. REFERENCES 1. Lindsay RM. Therapeutic potential of the neurotrophins and neu- rotrophin-CNTF combinations in peripheral neuropathies and motor neuron diseases. Ciba Found Symp 1996;196:39-48. Horner PJ, Gage FH. Regenerating the damaged central nervous system. Nature2000;407:963-970. Weinreb RN, Levin LA. Is neuroprotection a viable therapy for glaucoma? Arch Ophthalmol 1999;117:1540-1544. LaVail MM, Yasumura D, Matthes MT, et al. Protection of mouse photoreceptors by survival factors in retinal degenerations. Invest Ophthalmol Vis Sci 1998;39:592-602. Fontaine V, Kinkl N, Sahel J, et al. Survival of purified rat photore- ceptors in vitro is stimulated directly by fibroblast growth factor-2. J Neurosci 1998;18:9662-9672. Hicks D, Courtois Y. Fibroblast growth factor stimulates photore- ceptor differentiation in vitro. J Neurosci 1992;12:2022-2033. Lindsey JD, Weinreb RN. Survival and differentiation of purified retinal ganglion cells in a chemically defined microenvironment. Invest Ophthalmol Vis Sci 1994;35:3640-3648. Pease ME, McKinnon SJ, Quigley HA, et al. Obstructed axonal transport of BDNF and its receptor TrkB in experimental glauco- ma. Invest Ophthalmol Vis Sci 2000;41:764-774. Chong NH, Alexander RA, Waters L, et al. Repeated injections of a ciliary neurotrophic factor analogue leading to long-term pho- toreceptor survival in hereditary retinal degeneration. Invest Ophthalmol Vis Sci 1999;40:1298-1305. Hackett SF, Schoenfeld CL, Freund J, et al. Neurotrophic factors, cytokines and stress increase expression of basic fibroblast growth factor in retinal pigmented epithelial cells. Exp Eye Res 1997;64:865-873. Caffe AR, Soderpalm AK, Holmqvist I, et al. A combination of CNTF and BDNF rescues rd photoreceptors but changes rod dif- ferentiation in the presence of RPE in retinal explants. Invest Ophthalmol Vis Sci 2001;42:275-282. Edelhauser HF, Maren TH. Permeability of human cornea and sclera to sulfonamide carbonic anhydrase inhibitors. Arch Ophthalmol 1988;106:1110-1115. Schoenwald RD. Ocular drug delivery. Pharmacokinetic consider- ations. Clin Pharmacokinet 1990;18:255-269. Lee VH, Robinson JR. Topical ocular drug delivery: recent devel- opments and future challenges. J Ocul Pharmacol 1986;2:67-108. Schoenwald R. Ocular Pharmacokinetics / Pharmachodynamics. In: Mitra A, ed. Ophthalmic Drug Delivery Systems. New York, NY: Marcel Dekker; 1993:83-110. Ip NY. The neurotrophins and neuropoietic cytokines: Two families of growth factors acting on neural and hematopoietic cells. Ann N Y Acad Sci 1998;840:97-106. Takeuchi A, Kricorian G, Yao XY, et al. The rate and source of albumin entry into saline-filled experimental retinal detachments. Invest Ophthalmol Vis Sci 1994;35:3792-3798. Kamei M, Misono K, Lewis H. A study of the ability of tissue plas- minogen activator to diffuse into the subretinal space after intrav- itreal injection in rabbits. Am J Ophthalmol 1999;128:739-746. Velez G, Whitcup SM. New developments in sustained release drug delivery for the treatment of intraocular disease. Br J Ophthalmol 1999;83:1225-1229. Smith TJ, Pearson PA, Blandford DL, et al. Intravitreal sustained- release ganciclovir. Arch Ophthalmol 1992;110:255-258. 2. 3. 4. 5. 6. 7. 8. FIGURE 16 9. Model of FP receptor agonist action on sclera that results in increased transscleral permeability. FP receptor agonists (indicated by PG) can gain access to sclera either via topical drops (in which the PG is linked to an ester group to enhance transcorneal permeability), and then travel to sclera by uveoscleral outflow, or directly by periocular placement, as shown. Topical PG-ester is hydrolyzed as it passes through the cornea (1) into the anterior chamber (2). PG arriving at the sclera binds to FP receptors on scleral fibroblasts (3) and triggers signals that lead to increased MMP gene tran- scription (4) and proMMP biosynthesis (5) and secretion (6). Upon extra- cellular activation (7), the MMPs initiate collagen degradation (8) that results in increased transscleral permeability (9). 10. 11. 12. scription and lead to increased proMMP biosynthesis and secretion. Upon activation, the MMPs alter scleral colla- gen, which increases scleral permeability.69In this way, cotreatment with PGs may induce pharmacologic alter- ations of the sclera that promote transscleral delivery of peptide therapeutics. This could be useful for ameliorat- ing some diseases of the posterior segment of the eye. 13. 14. 15. 16. ACKNOWLEDGEMENTS Makoto Aihara, MD., PhD, Todd Anthony, PhD, Dan Gaton, MD, PhD, Jae-Woo Kim, MD, PhD, Takeshi Sagara, MD, PhD, postdoctoral Fellows in Glaucoma, each participated in different phases of the experimenta- tion. James D. Lindsey, PhD, supported all phases of these studies and critically reviewed the manuscript. Paul Kaufman, MD, contributed to the design of the first part of these studies and performed the animal testing. The San Diego Eye Bank of San Diego, California, provided 17. 18. 19. 20. 340

  23. Weinreb, Thesis 11/9/01 11:30 AM Page 341 Enhancement of Scleral Macromolecular Permeability with Prostaglandins 40. Roziere A, Mazuel C, Grove J, et al. Gelrite: A novel, ion-activat- ed, in-situ gelling polymer for ophthalmic vehicles: effect on bioavailability of Timolol. Int J Pharm1989;57:163-168. Miller S, Donovan M. Effect of poloxamer 407 gel on the motic activ- ity of pilocarpine nitrate in rabbits. Int J Pharm1982;12:147-152. Kumar S, Haglund B, Himmelstein K. In situ-forming gels of oph- thalmic drug delivery. J Ocul Pharmacol Ther 1994;19:47-56. Gurny R, Boye T, Ibrahim H. Ocular therapy with nanoparticulate sys- tems for controlled drug delivery. J Control Release1985;2:353-361. Ahmed I, Patton T. Selective intraocular delivery of liposome encapsulated inulin via the non-corneal absorption route. Int J Pharm1986;34:163-167. Assil KK, Weinreb RN. Multivesicular liposomes. Sustained release of the antimetabolite cytarabine in the eye. Arch Ophthalmol 1987;105:400-403. Skuta GL, Assil K, Parrish RK, et al. Filtering surgery in owl mon- keys treated with the antimetabolite 5- fluorouridine 5’-monophos- phate entrapped in multivesicular liposomes. Am J Ophthalmol 1987;103:714-716. Assil KK, Lane J, Weinreb RN. Sustained release of the antimetabolite 5-fluorouridine-5'- monophosphate by multivesicu- lar liposomes. Ophthalmic Surg 1988;19:408-413. Assil KK, Hartzer M, Weinreb RN, et al. Liposome suppression of proliferative vitreoretinopathy. Rabbit model using antimetabolite encapsulated liposomes. Invest Ophthalmol Vis Sci 1991;32:2891- 2897. Assil KK, Frucht-Perry J, Ziegler E, et al. Tobramycin liposomes. Single subconjunctival therapy of pseudomonal keratitis. Invest Ophthalmol Vis Sci 1991;32:3216-3220. Komai Y, Ushiki T. The three-dimensional organization of collagen fibrils in the human cornea and sclera. Invest Ophthalmol Vis Sci 1991;32:2244-2258. Thale A, Tillmann B, Rochels R. Scanning electron-microscopic studies of the collagen architecture of the human sclera--normal and pathological findings. Ophthalmologica 1996;210:137-141. Yamabayashi S, Ohno S, Aguilar RN, et al. Ultrastructural studies of collagen fibers of the cornea and sclera by a quick-freezing and deep-etching method. Ophthalmic Res1991;23:320-329. Thale A, Tillmann B. The collagen architecture of the sclera-SEM and immunohistochemical studies. Anat Anz 1993;175:215-220. Duke-Elder S. System of Ophthalmology. Vol 8, Part II. London, England: Henry Kimpton; 1963;995. Watson PG, Hazleman BL. The Sclera and Systemic Disorders. Philadelphia, Pa: WB Saunders; 1976;10. Keeley FW, Morin JD, Vesely S. Characterization of collagen from normal human sclera. Exp Eye Res1984;39:533-542. Trier K, Olsen EB, Ammitzboll T. Regional glycosaminoglycans composition of the human sclera. Acta Ophthalmol (Copenh) 1990;68:304-306. Rada JA, Achen VR, Perry CA, et al. Proteoglycans in the human sclera. Evidence for the presence of aggrecan. Invest Ophthalmol Vis Sci 1997;38:1740-1751. Spitznas M. The fine structure of human scleral collagen. Am J Ophthalmol 1971;1:68. Virgilio LA, Williams RJ, Klintworth GK. An unusually large human eye with abnormal scleral collagen. Arch Ophthalmol 1976;94:101-105. Shields MB, Shelburne JD, Bell SW. The ultrastructure of human limbal collagen. Invest Ophthalmol Vis Sci 1977;16:864-866. Johnston MC, Noden DM, Hazelton RD, et al. Origins of avian ocular and periocular tissues. Exp Eye Res1979;29:27-43. Olsen TW, Aaberg SY, Geroski DH, et al. Human sclera: Thickness and surface area. Am J Ophthalmol 1998;125:237-241. Olsen TW, Edelhauser HF, Lim JI, et al. Human scleral perme- ability. Effects of age, cryotherapy, transscleral diode laser, and sur- gical thinning. Invest Ophthalmol Vis Sci 1995;36:1893-1903. 21. Martin DF, Parks DJ, Mellow SD, et al. Treatment of cytomegalovirus retinitis with an intraocular sustained- release gan- ciclovir implant. A randomized controlled clinical trial. Arch Ophthalmol 1994;112:1531-1539. Sanborn GE, Anand R, Torti RE, et al. Sustained-release ganci- clovir therapy for treatment of cytomegalovirus retinitis. Use of an intravitreal device. Arch Ophthalmol 1992;110:188-195. Anand R, Nightingale SD, Fish RH, et al. Control of cytomegalovirus retinitis using sustained release of intraocular gan- ciclovir. Arch Ophthalmol 1993;111:223-227. Pearson PA, Jaffe GJ, Martin DF, et al. Evaluation of a delivery sys- tem providing long-term release of cyclosporine. Arch Ophthalmol 1996;114:311-317. Jaffe GJ, Yang CS, Wang XC, et al. Intravitreal sustained-release cyclosporine in the treatment of experimental uveitis. Ophthalmology 1998;105:46-56. Cheng CK, Berger AS, Pearson PA, et al. Intravitreal sustained- release dexamethasone device in the treatment of experimental uveitis. Invest Ophthalmol Vis Sci 1995;36:442-453. Yang CS, Khawly JA, Hainsworth DP, et al. An intravitreal sus- tained-release triamcinolone and 5-fluorouracil codrug in the treatment of experimental proliferative vitreoretinopathy. Arch Ophthalmol 1998;116:69-77. Jaffe GJ, Yang CH, Guo H, et al. Safety and pharmacokinetics of an intraocular fluocinolone acetonide sustained delivery device. Invest Ophthalmol Vis Sci 2000;41:3569-3575. Berger AS, Cheng CK, Pearson PA, et al. Intravitreal sustained release corticosteroid-5-fluoruracil conjugate in the treatment of experimental proliferative vitreoretinopathy. Invest Ophthalmol Vis Sci 1996;37:2318-2325. Zhou T, Lewis H, Foster RE, et al. Development of a multiple- drug delivery implant for intraocular management of proliferative vitreoretinopathy. J Control Release1998;55:281-295. Ahmed I, Patton TF. Importance of the noncorneal absorption route in topical ophthalmic drug delivery. Invest Ophthalmol Vis Sci 1985;26:584-587. Ahmed I, Gokhale RD, Shah MV, et al. Physicochemical determi- nants of drug diffusion across the conjunctiva, sclera, and cornea. J Pharm Sci 1987;76:583-586. McCartney H, Drysdale I, Gornall A, et al. An autoradiographic study of the penetration of subconjunctivally injected hydrocorti- sone into the normal and inflamed rabbit eye. Invest Ophthalmol 1965;4:297-299. Barza M, Kane A, Baum J. Intraocular penetration of gentamicin after subconjunctival and retrobulbar injection. Am J Ophthalmol 1978;85:541-547. Barza M, Kane A, Baum JL. Regional differences in ocular con- centration of gentamicin after subconjunctival and retrobulbar injection in the rabbit. Am J Ophthalmol 1977;83:407-413. Lim JI, Maguire AM, John G, et al. Intraocular tissue plasminogen activator concentrations after Ophthalmology 1993;100:373-376. Lincoff H, Stanga P, Movshovich A, et al. Choroidal concentration of interferon after retrobulbar injection. Invest Ophthalmol Vis Sci 1996;37:2768-2771. Weijtens O, van der Sluijs FA, Schoemaker RC, et al. Peribulbar corticosteroid injection: vitreal and serum concentrations after dex- amethasone disodium phosphate injection. Am J Ophthalmol 1997;123:358-363. Weijtens O, Schoemaker RC, Lentjes EG, et al. Dexamethasone concentration in the subretinal fluid after a subconjunctival injec- tion, a peribulbar injection, or an oral dose. Ophthalmology 2000;107:1932-1938. 41. 22. 42. 43. 23. 44. 24. 45. 25. 46. 26. 47. 27. 48. 28. 49. 29. 50. 30. 51. 31. 52. 32. 53. 54. 33. 55. 56. 34. 57. 35. 58. 36. 59. subconjunctival delivery. 60. 37. 61. 38. 62. 63. 39. 64. 341

  24. Weinreb, Thesis 11/9/01 11:30 AM Page 342 Weinreb 65. Bill A. Movement of albumin and dextran through the sclera. Arch Ophthalmol 1965;74:248-252. Foulds WS. Clinical significance of trans-scleral fluid transfer. Trans Ophthalmol Soc U K 1976;96:290-308. Inomata H, Bill A. Exit sites of uveoscleral flow of aqueous humor in cynomolgus monkey eyes. Exp Eye Res1977;25:113-118. Unlu N, Robinson JR. Scleral permeability to hydrocortisone and mannitol in the albino rabbit eye. J Ocul Pharmacol Ther 1998;14:273-281. Wood RL, Koseki T, Kelly DE. Uveoscleral permeability to intra- camerally infused ferritin in eyes of rabbits and monkeys. Cell Tissue Res1992;270:559-567. Maurice DM, Polgar J. Diffusion across the sclera. Exp Eye Res 1977;25:577-582. Shields MB, Bradbury MJ, Shelburne JD, et al. The permeability of the outer layers of limbus and anterior sclera. Invest Ophthalmol Vis Sci 1977;16:866-869. Ahmed I, Patton T. Disposition of timolol and inulin in the rabbit eye following corneal versus non-corneal absorption. Int J Pharm 1987;38:9-21. Ambati J, Canakis CS, Miller JW, et al. Diffusion of high-molecu- lar-weight compounds through sclera. Invest Ophthalmol Vis Sci 2000;41:1181-1185. Sarraf D, Lee DA. The role of iontophoresis in ocular drug deliv- ery. J Ocul Pharmacol 1994;10:69-81. Lam TT, Fu J, Tso MO. A histopathologic study of retinal lesions inflicted by transscleral iontophoresis. Graefes Arch Clin Exp Ophthalmol 1991;229:389-394. Dan JA, Yaron A. Focal scleral permeability enhanced by collage- nase digestion. Experimental model of enzymatic sclerostomy. Ophthalmology 1994;101:461-472. Trelstad RL, Silbermann NN, Brockhurst RJ. Nanophthalmic scle- ra. Ultrastructural, histochemical, and biochemical observations. Arch Ophthalmol 1982;100:1935-1938. Yue BY, Duvall J, Goldberg MF, et al. Nanophthalmic sclera. Morphologic and tissue culture studies. Ophthalmology 1986;93:534-541. Gass JD. Uveal effusion syndrome. A new hypothesis concerning pathogenesis and technique of surgical treatment. Retina 1983;3:159-163. Brockhurst RJ. Vortex vein decompression for nanophthalmic uveal effusion. Arch Ophthalmol 1980;98:1987-1990. Rudnick DE, Noonan JS, Geroski DH, et al. The effect of intraoc- ular pressure on human and rabbit scleral permeability. Invest Ophthalmol Vis Sci 1999;40:3054-3058. Sagara T, Gaton DD, Lindsey JD, et al. Topical prostaglandin F2? treatment reduces collagen types I, III, and IV in the monkey uveoscleral outflow pathway. Arch Ophthalmol 1999;117:794-801. Edwards A, Prausnitz M. Fiber matrix model of sclera and corneal stroma for drug delivery to the eye. Am Inst Chem Eng J 1998;44:214-225. Nagase H, Woessner JF Jr. Matrix metalloproteinases. J Biol Chem 1999;274:21491-21494. Nagase H, Fields GB. Human matrix metalloproteinase specificity studies using collagen sequence-based synthetic peptides. Biopolymers1996;40:399-416. Lindsey J, Weinreb R. Effects of prostaglandins on uveoscleral out- flow. In: Alm A, Weinreb R eds. Uveoscleral Outflow: Biology and Clinical Aspects. London, England: Mosby-Wolfe; 1998;41-56. Gaton DD, Sagara T, Lindsey JD, et al. Matrix metalloproteinase- 1 localization in the normal human uveoscleral outflow pathway. Invest Ophthalmol Vis Sci 1999;40:363-369. Abramovitz M, Boie Y, Nguyen T, et al. Cloning and expression of a cDNA for the human prostanoid FP receptor. J Biol Chem 1994;269:2632-2636. Pierce KL, Fujino H, Srinivasan D, et al. Activation of FP prostanoid receptor isoforms leads to Rho-mediated changes in cell morphology and in the cell cytoskeleton. J Biol Chem 1999;274:35944-35949. Ocklind A, Lake S, Wentzel P, et al. Localization of the prostaglandin F2( receptor messenger RNA and protein in the cynomolgus mon- key eye. Invest Ophthalmol Vis Sci 1996;37:716-726. Gabelt B, Kaufman P. Normal physiology. In: Alm A, Weinreb R, eds. Uveoscleral Outflow. London, England: Mosby-Wolfe; 1998;25-40. Lalani EN, Golding M, Hudson M, et al. Protein extraction and western blotting from methacarn-fixed tissue. J Pathol 1995;177:323-328. Shetye JD, Scheynius A, Mellstedt HT, et al. Retrieval of leukocyte antigens in paraffin-embedded rat tissues. J Histochem Cytochem 1996;44:767-776. Sagara T, Gaton DD, Lindsey JD, et al. Reduction of collagen type I in the ciliary muscle of inflamed monkey eyes. Invest Ophthalmol Vis Sci 1999;40:2568-2576. Raviola G, Freddo TF. A simple staining method for blood vessels in flat preparations of ocular tissues. Invest Ophthalmol Vis Sci 1980;19:1518-1521. Alexander RA, Hiscott PS, Hart RL, et al. Effect of melanin bleaching on immunoperoxidase, with reference to ocular tissues and lesions. Med Lab Sci 1986;43:121-127. Foss AJ, Alexander RA, Jefferies LW, et al. Immunohistochemical techniques: The effect of melanin bleaching. Br J Biomed Sci 1995;52:22-25. Goldberg GI, Strongin A, Collier IE, et al. Interaction of 92-kDa type IV collagenase with the tissue inhibitor of metalloproteinases prevents dimerization, complex formation with interstitial collage- nase, and activation of the proenzyme with stromelysin. J Biol Chem1992;267:4583-4591. Gross DS, Rothfeld JM. Quantitative immunocytochemistry of hypothalamic and pituitary hormones: Validation of an automated, computerized image analysis system. J Histochem Cytochem 1985;33:11-20. 100. van der Loos CM, Marijianowski MM, Becker AE. Quantification in immunohistochemistry: The measurement of the ratios of colla- gen types I and II. Histochem J 1994;26:347-354. 101. Stamer WD, Seftor RE, Snyder RW, et al. Cultured human tra- becular meshwork cells express aquaporin-1 water channels. Curr Eye Res1995;14:1095-1100. 102. Huang Y, Gil DW, Vanscheeuwijck P, et al. Localization of alpha 2- adrenergic receptor subtypes in the anterior segment of the human eye with selective antibodies. Invest Ophthalmol Vis Sci 1995;36:2729-2739. 103. Anthony TL, Pierce KL, Stamer WD, et al. Prostaglandin F2? receptors in the human trabecular meshwork. Invest Ophthalmol Vis Sci 1998;39:315-321. 104. Parsons SL, Watson SA, Brown PD, et al. Matrix metalloproteinas- es. Br J Surg 1997;84:160-166. 105. Stjernschantz J, Selen G, Sjoquist B, et al. Preclinical pharmacolo- gy of latanoprost, a phenyl-substituted PGF2?analogue. Adv Prostaglandin Thromboxane Leukotriene Res1995;23:513-518. 106. Sjöquist B, Almegard B, Khalifeh V, et al. The bioavailability of Xalatan in the human eye. Invest Ophthalmol Vis Sci 1997;38:S248. 107. Lindsey JD, Kashiwagi K, Boyle D, et al. Prostaglandins increase proMMP-1 and proMMP-3 secretion by human ciliary smooth muscle cells. Curr Eye Res1996;15:869-875. 108. Weinreb RN, Kashiwagi K, Kashiwagi F, et al. Prostaglandins increase matrix metalloproteinase release from human ciliary smooth muscle cells. Invest Ophthalmol Vis Sci 1997;38:2772-2780. 109. Nagase H, Okada Y. Proteinases and matrix degradation. In: Kelly J, Ruddy S, Sledge C, eds. Textbook of Rheumatology. Philadelphia, Pa: WB Saunders; 1997;323-341. 89. 66. 90. 67. 68. 91. 69. 92. 70. 93. 71. 94. 72. 95. 73. 96. 74. 97. 75. 98. 76. 77. 99. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 342

  25. Weinreb, Thesis 11/9/01 11:30 AM Page 343 Enhancement of Scleral Macromolecular Permeability with Prostaglandins 110. Woessner JF Jr. The family of matrix metalloproteinases. Ann N Y Acad Sci 1994;732:11-21. 111. Nagase H, Suzuki K, Enghild JJ, et al. Stepwise activation mecha- nisms of the precursors of matrix metalloproteinases 1 (tissue col- lagenase) and 3 (stromelysin). Biomed Biochim Acta 1991;50:749- 754. 112. Vincenti MP, White LA, Schroen DJ, et al. Regulating expression of the gene for matrix metalloproteinase-1 (collagenase): Mechanisms that control enzyme activity, transcription, and mRNA stability. Crit Rev Eukaryot Gene Expr 1996;6:391-411. 113. Denhardt DT, Feng B, Edwards DR, et al. Tissue inhibitor of met- alloproteinases (TIMP, aka EPA): Structure, control of expression and biological functions. Pharmacol Ther 1993;59:329-341. 114. Bode W, Fernandez-Catalan C, Grams F, et al. Insights into MMP- TIMP interactions. Ann N Y Acad Sci 1999;878:73-91. 115. Berman MB. Regulation of corneal fibroblast MMP-1 collagenase secretion by plasmin. Cornea 1993;12:420-432. 116. Cook H, Davies KJ, Harding KG, et al. Defective extracellular matrix reorganization by chronic wound fibroblasts is associated with alterations in TIMP-1, TIMP-2, and MMP-2 activity. J Invest Dermatol 2000;115:225-233. 117. Westermarck J, Holmstrom T, Ahonen M, et al. Enhancement of fibroblast collagenase-1 (MMP-1) gene expression by tumor pro- moter okadaic acid is mediated by stress-activated protein kinases Jun N-terminal kinase and p38. Matrix Biol 1998;17:547-557. 118. Astin M. Effects of prostaglandin E2, F2alpha, and latanoprost acid on isolated ocular blood vessels in vitro. J Ocul Pharmacol Ther 1998;14:119-128. 119. Stjernschantz J, Selen G, Astin M, et al. Microvascular effects of selective prostaglandin analogues in the eye with special reference to latanoprost and glaucoma treatment. Prog Retin Eye Res 2000;19:459-496. 120. Brannon-Peppas L, Peppas NA. Dynamic and equilibrium swelling behaviour of pH-sensitive hydrogels containing 2-hydroxyethyl methacrylate. Biomaterials1990;11:635-644. 121. Bell CL, Peppas NA. Water, solute and protein diffusion in physi- ologicly responsive hydrogels of poly (methacrylic acid-g-ethylene glycol). Biomaterials1996;17:1203-1218. 122. Chapman SA, Ayad S, O'Donoghue E, et al. Glycoproteins of tra- becular meshwork, cornea and sclera. Eye1998;12:440-448. 123. White J, Werkmeister JA, Ramshaw JA, et al. Organization of fib- rillar collagen in the human and bovine cornea: Collagen types V and III. Connect Tissue Res1997;36:165-174. 124. Kimura S, Kobayashi M, Nakamura M, et al. Immunoelectron microscopic localization of decorin in aged human corneal and scleral stroma. J Electron Microsc (Tokyo) 1995;44:445-449. 125. Wessel H, Anderson S, Fite D, et al. Type XII collagen contributes to diversities in human corneal and limbal extracellular matrices. Invest Ophthalmol Vis Sci 1997;38:2408-2422. 126. Shuttleworth CA. Type VIII collagen. Int J Biochem Cell Biol 1997;29:1145-1148. 127. Matrisian LM. The matrix-degrading metalloproteinases. Bioessays 1992;14:455-463. 128. Lindsey JD, Kashiwagi K, Kashiwagi F, et al. Prostaglandin action on ciliary smooth muscle extracellular matrix metabolism: implica- tions for uveoscleral outflow. Surv Ophthalmol 1997;41 Suppl 2:S53-59. 129. Lindsey JD, Kashiwagi K, Kashiwagi F, et al. Prostaglandins alter extracellular matrix adjacent to human ciliary muscle cells in vitro. Invest Ophthalmol Vis Sci 1997;38:2214-2223. 130. Bikfalvi A, Klein S, Pintucci G, et al. Biological roles of fibroblast growth factor-2. Endocrinol Rev 1997;18:26-45. 131. Penc SF, Pomahac B, Winkler T, et al. Dermatan sulfate released after injury is a potent promoter of fibroblast growth factor-2 func- tion. J Biol Chem1998;273:28116-28121. 132. Walicke P, Cowan WM, Ueno N, et al. Fibroblast growth factor promotes survival of dissociated hippocampal neurons and enhances neurite extension. Proc Natl Acad Sci USA 1986;83:3012- 3016. 133. Dreyfus H, Sahel J, Heidinger V, et al. Gangliosides and neu- rotrophic growth factors in the retina. Molecular interactions and applications as neuroprotective agents. Ann N Y Acad Sci 1998;845:240-252. 134. Cummings BJ, Yee GJ, Cotman CW. FGF-2 promotes the survival of entorhinal layer II neurons after perforant path axotomy. Brain Res1992;591:271-276. 135. Nakata N, Kato H, Kogure K. Protective effects of basic fibroblast growth factor against hippocampal neuronal damage following cerebral ischemia in the gerbil. Brain Res1993;605:354-356. 136. Rabchevsky AG, Fugaccia I, Fletcher-Turner A, et al. Basic fibrob- last growth factor (FGF-2) enhances tissue sparing and functional recovery following moderate spinal cord injury. J Neurotrauma 1999;16:817-830. 137. Lipton SA, Wagner JA, Madison RD, et al. Acidic fibroblast growth factor enhances regeneration of processes by postnatal mammalian retinal ganglion cells in culture. Proc Natl Acad Sci USA 1988;85:2388-2392. 138. Pearce RK, Collins P, Jenner P, et al. Intraventricular infusion of basic fibroblast growth factor (FGF-2) in the MPTP-treated com- mon marmoset. Synapse1996;23:192-200. 139. Connor B, Dragunow M. The role of neuronal growth factors in neurodegenerative disorders of the human brain. Brain Res Brain Res Rev 1998;27:1-39. 140. Cunha-Vaz JG, Shakib M, Ashton N. Studies on the permeability of the blood-retinal barrier: I. On the existence, development and site of a blood-retinal barrier. Br J Ophthalmol 1966;50:441-453. 141. Lee VH, Carson LW, Kashi SD, et al. Metabolic and permeation barriers to the ocular absorption of topically applied enkephalins in albino rabbits. J Ocul Pharmacol 1986;2:345-352. 142. Pino RM. Restriction to endogenous plasma proteins by a fenes- trated capillary endothelium: An ultrastructural immunocytochem- ical study of the choriocapillaris endothelium. Am J Anat 1985;172:279-289. 143. Toris CB, Pederson JE. Experimental retinal detachment. VII. Intravenous horseradish peroxidase diffusion across the blood-reti- nal barrier. Arch Ophthalmol 1984;102:752-756. 144. Smith RS, Rudt LA. Ocular vascular and epithelial barriers to microperoxidase. Invest Ophthalmol 1975;14:556-560. 145. Laties AM, Rapoport S. The blood-ocular barriers under osmotic stress. Studies on the freeze-dried eye. Arch Ophthalmol 1976;94:1086-1091. 146. Peyman GA, Bok D. Peroxidase diffusion in the normal and laser- coagulated primate retina. Invest Ophthalmol 1972;11:35-45. 147. Grayson MC, Laties AM. Ocular localization of sodium fluorescein: Effects of administration in rabbit and monkey. Arch Ophthalmol 1971;85:600-609. 148. Tso MOM, Shih C-Y, McLean IW. Is there a blood-brain barrier at the optic nerve head? Arch Ophthalmol 1975;93:815-825. 149. Lindsey JD, To HD, Weinreb RN. Induction of c-fos by prostaglandin F2?in human ciliary smooth muscle cells. Invest Ophthalmol Vis Sci 1994;35:242-250. 343

  26. Weinreb, Thesis 11/9/01 11:30 AM Page 344