Clinical trials found that 95% of patients receiving monthly ranibizumab injections maintained their visual acuity and 34–40% had improved vision (gaining 15 or more letters in 12 months). Bevacizumab (Avastin®, Genentech), a full length humanized anti-VEGF antibody, approved by the FDA to prevent regrowth of ­vessels at tumor sites in patients with colon cancer, breast cancer, and nonsmall cell lung cancer, is currently used as an off-label drug to treat wet AMD. A new VEGF analog that has received increased attention is the VEGF trap (VEGF Trap-Eye™, Regeneron), a modified soluble VEGF receptor analog protein that binds more tightly to VEGF than pegaptanib (Ng et al. 2006) and ranibizumab (Stewart and Rosenfeld 2008). With the development of anti-VEGF therapies, visual acuity of patients suffering from wet AMD and diabetic macular edema (DME) is expected to be significantly increased. In addition, development of therapeutics that target growth factor such as ciliary neurotrophic factor (CNTF) are under development for treating retinal degenerative disorders using the NT-501 intravitreal implant [Neurotech, Inc. developed an encapsulated cell technology (ECT) based implant to deliver macromolecules directly to the site of action after cellular production], which is currently undergoing clinical trials. The results thus far have shown that the implant is safe up to 1 year after injection. Thus, protein and other macromolecule therapeutics are of value in treating disorders of the eye and are currently being explored for their potential long-term effects. The primary target for the current protein therapies of the eye are tissues of the posterior segment of the eye. However, due to the presence of formidable biological barriers, therapeutic macromolecules such as pegaptanib and ranibizumab as well as implants encapsulating cells are ­typically administered in the vitreous humor in order to ensure that therapeutic ­concentrations of the drug reach the target site in the back of the eye.

In this chapter, various routes of administration, delivery strategies, challenges for each delivery system, macromolecule case studies, and standard protocols for formulation development are discussed with a key focus on protein drugs. Several of the approaches discussed might be relevant to nucleic acid therapeutics as well.

17.2  Routes of Protein Administration

Due to the unique anatomy and physiology of the eye, ocular drug delivery is historically­ challenging (Lee and Robinson 1986; Kompella et al. 2010). Protein delivery to the eye has been evaluated for various routes of administration including topical, intracorneal, intracameral, periocular (subconjunctival, sub-Tenon, retrobulbar, and peribulbar), intravitreal, subretinal, suprachoroidal, and intravenous. The route of administration directly influences the extent of drug delivery to various target sites within the eye. Topical, intracorneal, intracameral, and periocular routes typically deliver higher concentration of most therapeutic agents, especially small molecules, to the anterior­ segment as compared to the posterior segment. Whereas, intravitreal, subretinal, and suprachoroidal injections deliver higher concentration of protein and other therapeutic agents to the posterior segment as compared to the

anterior segment. On the other hand, systemic administration or delivery by the intravenous route can potentially deliver proteins and other therapeutic agents, although at low concentrations, to the anterior and/or posterior segments of the eye, provided the drug can overcome the blood-aqueous, blood-retinal, metabolic, and immunologic or other clearance barriers. In the following discussion, some principal routes of administration of therapeutic proteins along with some successful examples are discussed.

17.2.1  Topical

Topical administration of drugs into the inferior fornix of the conjunctiva is ­typically used for treating diseases of the anterior segment of the eye. Due to rapid clearance from eye surface, drugs from an eye drop cannot typically reach the posterior ­segment to a therapeutic level (Lee and Robinson 1986). Topically applied drugs undergo rapid clearance and do not reside for long durations in the precorneal area, due to mixing and dilution of drug with tears, tear turnover or tear drainage [0.7 mL/min in rabbit and 1.0–2.0 mL/min in human (Owen et al. 2007)], and blinking of the eye [once per 18 min for rabbits and 4–16 times per min in human (Congdon et al. 2004; Owen et al. 2007)], leading to poor drug bioavailability. In addition, the tight junctions of the corneal and conjunctival epithelial layers further restrict the drug from entering the eye.

Formulations such as suspensions, ointments, and gels may be used to prolong the precorneal drug residence. In situ forming gels such as Gelrite™ were designed to overcome the precorneal elimination problem to a certain extent (Carlfors et al. 1998). Upon instillation, these drops undergo sol-gel transition in the cul-de-sac of the eye in the presence of monoor di-valent cations of the lacrimal fluid. A formulation of indomethacin using Gelrite sustained drug release for 8 h in vitro and was efficacious in treating uveitis in a rabbit model (Balasubramaniam et al. 2003). Topically applied drugs may be able to reach the posterior segment of the eye to a greater extent if the formulation has enhanced precorneal drug retention.

Interestingly, few protein drugs have been reported to permeate to the back of the eye following topical eye drop instillation. In a recent study, tumor necrosis factor (TNF)-a inhibitory single-chain antibody fragment (scFv; 26 kDa) (ESBA105) when administered as topical drop at high frequency followed by persistent opening of the eyes, showed absorption and distribution to various compartments of the eye as opposed to an intravenous injection of an equivalent dose (Furrer et al. 2009). In this study, rabbits were divided into three groups: two groups received ESBA105 topically as drops and one group received ESBA105 via intravenous administration. In group one, ESBA105 was administered topically as one drop every hour for 10 h, up to 5 mg/day for one single day (after each administration, the eyes were kept still for 30 s). Group two was given one topical drop of ESBA105, 5 times a day, for 6 days up to 15 mg/6 days. Group three received an intravenous bolus injection of 5 mg of ESBA105, one time through the marginal ear vein. Drug concentrations

were recorded in the following tissues: aqueous humor, vitreous humor, neuroretina, retinal pigmented epithelium (RPE)-choroid, and serum. In group one, the tissue levels of ESBA105 were: 12, 295, 214, 263, and 0.5 ng/mL in the aqueous humor, vitreous humor, neuroretina, RPE-choroid, and serum, respectively, after a single day administration. Interestingly, vitreous humor levels were nearly 25 times higher than aqueous humor levels. In group three, the following drug levels were obtained: 175, 63, 66, 2,690, and 89,284 ng/mL in the aqueous humor, vitreous humor, neuroretina, RPE-choroid, and serum, respectively, after intravenous administration. Interestingly, rabbits given multiple topical doses in one day (group one) had 4.7 times higher vitreous humor levels of ESBA105 as compared to rabbits given a single intravenous dose (group three). In addition, the aqueous humor levels of ESBA105 after intravenous administration (group three) was nearly 15 times higher than after topical administration (group one). RPE-choroid drug levels were approximately 10 times higher after intravenous administration (group three) than after multiple topical doses (group one). In summary, the vitreous humor and neuroretina drug levels were nearly 5 times higher after multiple topical doses in one day than after a single bolus intravenous injection of the same dosage, yet aqueous humor levels were 15 times higher after intravenous bolus administration as compared to topical administration. After a single drop of ESBA105, the concentration reached 98 ng/mL in the vitreous humor. The half-life of ESBA105 after multiple topical doses in a single day as well as after a single intravenous administration was significantly longer in vitreous, neuroretina, and RPE-choroid as compared to aqueous humor and serum. Although the half-life of ESBA105 after intravenous administration was 1.5 times higher (24 vs. 15 h) in the vitreous than multiple topical doses in one day, the neuroretinal half-life after multiple topical doses was 1.2 times (27 vs. 23 h) higher than intravenous administration. These results confirm the presence of the ESBA105 in specific locations within the back of the eye up to 27 h after topical administration. Group two (topical administration for multiple days) showed a continuous rise in ESBA105 levels in all tissues and reached a steady state concentration of above 300 ng/mL in retina and above 500 ng/mL in vitreous humor. For both groups that received topical administration of ESBA105 either multiple doses in a single day or multiple doses over multiple days, the systemic exposure of ESBA105 was minimal compared to group three, given a single intravenous administration. The results from this study indicated that daily multiple topical doses of a protein drug may deliver therapeutic quantities to the posterior segment of the eye depending on the protein characteristics and concentration needed for a therapeutic effect.

In another study, eye drops of vasostatin, an endogenous angiogenesis inhibitor containing N-terminal fragments (CGA1-76 and CGA1-113) of chromogranin A, with an apparent molecular weight of 7–22 kDa, have been shown to reduce CNV lesion area for at least up to day 35 following eye drop dosing for 20 days (Sheu et al. 2009). In this study, rats were dosed topically with 1 mg/mL of vasostatin in PBS, 3 times daily for 20 days after induction of CNV lesions by laser photocoagulation. On day 21, CNV lesions decreased to 3.5 ± 1.11 mm2 for vasostatin-treated eyes as compared to 7.01 ± 1.07 and 6.87 ± 2.03 mm2, respectively, for untreated and

vehicle (PBS) treated eyes. On day 28, the lesion sizes were 5.27 ± 1.06, 10.34 ± 1.3, and 8.99 ± 2.03 mm2, in vasostatin, untreated, and vehicle treated groups, respectively. Although the CNV lesion areas did increase in all groups by day 35, the rate of increase was much slower in the vasostatin treated eye. The CNV lesion areas were 6.11 ± 1.33, 11.03 ± 0.72, and 9.75 ± 1.62 mm2, respectively, for vasostatin, untreated, and PBS treated groups on day 35.

Insulin is currently administered only by systemic injection. However, earlier efforts have demonstrated that insulin administered as an eye drop can also reduce blood glucose levels (Yamamoto et al. 1989) and further, the effects of insulin can be enhanced when an absorption enhancers such as glycocholate or fusidic acid are included in the insulin formulation at pH 8.0 (Xuan et al. 2005). When 50 mL drops of 0.5% insulin (either pH 3.5 or pH 8.0) were administered into rabbit eyes topically, the blood glucose level was significantly reduced , indicating a systemic therapeutic effect is possible via eye drops. The blood glucose level was reduced to 65% when insulin was formulated at pH 8.0 (0.5% concentration), whereas the blood glucose level was reduced to 80% when the same concentration of insulin was formulated at pH 3.5. Further, when 1% insulin drops were administered, the blood glucose level decreased to 30 and 70% for pH 8.0 and pH 3.5 formulations, respectively. When either 0.5 or 1% glycocholic­ acid was added to 0.125% insulin at pH 8.0, the blood glucose level decreased to 60% as compared to 80% with 0.125% insulin at pH 8.0 without glycocholic acid. Further, addition of either 0.25 or 0.5% fusidic acid to 0.125% insulin at pH 8.0 reduced the glucose level to 55 and 35%, respectively, as compared to 80% with 0.125% insulin at pH 8.0 without fusidic acid. This study demonstrates that it is possible to reduce blood glucose levels by administering insulin as eye drops in the presence of an absorption enhancer at pH 8.0.

The results from the studies discussed above demonstrate that protein therapeutics can potentially exert therapeutic effects in tissues of the back of the eye or in the system after topical administration. Since all of the earlier studies were conducted in animals, it is still uncertain if therapeutic proteins or other therapeutic agents can reach the posterior segment of the eye after topical administration in humans. Although the drug is capable of reaching the posterior segment of the eye after topical administration, it is expected that drug levels in the posterior segment of the eye will be much less than if the drug was administered by intravitreal injection.

17.2.2  Intracameral

Intracameral injections either into the anterior chamber or aqueous humor are commonly used for delivering anti-infective agents or anti-inflammatory agents during eye surgery (Lee and Robinson 2001; Karalezli et al. 2008). This route is inefficient in delivering therapeutic agents to the posterior segment of eye, and therefore, it might not be suitable for treating diseases such as retinal degeneration (Lee and Robinson 2001). For instance, Lee and Robinson compared the vitreous and aqueous humor drug levels after intracameral and subconjunctival injections of