Danazol binds to androgens and steroid-binding globulins, stimulating the formation of a cortical actin ring that enhances endothelial barrier function. Danazol has been approved by the US Food and Drug Administration (US FDA) for the oral treatment of endometriosis.

The FDA mandated that Ampio Pharmaceuticals perform a confirmatory study on patients with DME who are refractory to approved medications. Since patients will already have failed to respond to anti-VEGF medications, placebo will serve as the control group. The company projects that 80 patients (40 in each of the danazol and control groups) will be needed for a 12-month study. The FDA did not require that safety be a specified endpoint because danazol would be used at doses only 10% of those approved for use in patients with endometriosis [7].

A drug delivery platform based on cyclodextrin microparticles that dissolve in the tear film to form water-soluble drug/cyclodextrin complex microparticles has been developed for ocular pharmacology [44]. Microparticulate 1.5% (wt/vol) dexamethasone-cyclodextrin eye drops can deliver the drug to the retina and vitreous humor in rabbits [63, 80, 81, 122]. Cyclodextrin-based dexamethasone eye drop solutions penetrate well into the anterior segment of the human eye [75, 110]. Since early pharmacology studies in rabbits and humans suggested that dexamethasone-cyclodextrin microparticle eye drops may reach the human retina, clinical trials of dexamethasone eye drops were initiated for the topical treatment of DME.

Nineteen eyes of 19 patients with DME received dexamethasone-cyclodextrin eye drops three or six times a day for 4 weeks and were then observed for 4 weeks without treatment [127]. At weeks 0 (baseline), 4, and 8, logMAR visual acuity (mean ± SD) was 0.52 ± 0.41, 0.37 ± 0.40 (P = 0.0025 vs. baseline), and 0.45 ± 0.41, respectively; central macular thickness (μm) was 512 ± 164, 399 ± 154 (P = 0.0016 vs. baseline), and 488 ± 172 (P = 0.0116 vs. week 4), respectively; and intraocular pressure (mm Hg) was 15.2 ± 3.1, 17.4 ± 4.2 (P = 0.0015 vs. baseline) and 15.8 ± 4.0, respectively. At week 4, central macular thickness had decreased more than 10% in 12 of 19 eyes (63%), and the mean change was −20% (−65% to −10%). In 14 of 19 eyes (74%) visual acuity had improved more than 0.1 logMAR at week 4. No eye drop-related adverse effects were noted.

The efficacy of topical difluprednate ophthalmic emulsion 0.05% (Durezol®, Sirion Therapeutics Inc., USA) on the treatment of refractory DME after vitrectomy was compared to sub-Tenon’s injections of triamcinolone (STTA) [87]. Eleven eyes of ten subjects were treated with STTA (STTA group), and 11 eyes of seven subjects were treated with difluprednate ophthalmic emulsion 0.05% four times daily for the first month and then twice daily for 2 months (eye drop group).

The mean VA (±SD) in the eye drop group was similar at 3 months (0.67 ± 0.29 loMAR) as at baseline (0.67 ± 0.35 logMAR); mean retinal thickness (μm) decreased from baseline (500.6 ± 207.7) to 3 months (341.2 ± 194.8), with a mean minimum retinal thickness during the treatment period (300.6 ± 123.2) that was significantly lower than that at baseline (Mann-Whitney U test: P = 0.003). In the STTA group, mean VA (±SD) was 0.67 ± 0.35 logMAR, and mean retinal thickness was 543.3 ± 132.6 μm at baseline. After 3 months of treatment, mean VA improved to 0.49 ± 0.67 logMAR, and mean retinal thickness had decreased to 378.6 ± 135 μm. The mean minimum retinal thickness during the treatment period (349.9 ± 113.8 μm) was significantly lower than at baseline (Mann-Whitney U test: P = 0.003). The rate of improvement in retinal thickness did not differ between the eye drop group (73%) and STTA group (84%) (Fisher’s exact test: P = 1).

Eyegate Pharmaceuticals is using iontophoresis to deliver the experimental drug EGP-437 (reformulated, topically active dexamethasone phosphate) to the retina of patients with DME. A multicenter, open-label, phase Ib/IIa trial has enrolled 19 patients with DME, retinal vein occlusions, and postsurgical cystoid macular edema. Treatments with an electrical impulse of 4.0 mA-min (3.5 mA) were administered on days 0, 4, and 9 with the primary outcome being the reduction in central subfield thickness (CST) on days 4, 9, and 14. The dexamethasone insert was administered to patients who did not respond favorably by day 14.

The interim results showed that some patients, particularly those with postoperative cystoid macular edema (CME) and those who were pseudophakic, responded positively [36]. Edema re-accumulated when the drug was cleared from the tissues. An extension study will include an additional 15 patients who will be dosed on 3 consecutive days.

Kala Pharmaceuticals is developing nanotechonology-based ophthalmic products to treat DME. They are initiating a phase II clinical trial (KPI-121-C-004) to evaluate KP-121 (LE-MPP), a topically administered loteprednol etabonate mucus-penetrating platform (MPP), for the treatment of macular edema due to DR and retinal vein occlusions [82]. This single-masked, randomized trial will investigate the efficacy and safety of 1% LP-MPP and 0.25% LPP administered QID to 20 patients.

DARPins (designed ankyrin repeat proteins) are small molecular weight (14–20 kDa) molecules with high solubility (>100 mg/L) in saline. They are a flexible design platform that allows for the creation of genetically engineered mimetic proteins that can target any molecule [17]. Several DARPin molecules (at least 15) are being developed for chorioretinal vascular conditions, oncologic indications, and inflammatory diseases. DARPin technology also allows for the design of dual action proteins.

Abicipar pegol (Allergan, Irvine, CA), formerly known as DARPin MP0112, binds all isoforms of VEGF-A. This high-affinity molecule (K _D = 2 pM for VEGF₁₆₅) has a surprisingly long intravitreal half-life in rabbits (6 days) that may be attributed to its pegylation. A phase I/II multicenter, open-label, dose-escalation trial evaluated the safety and bioactivity of abicipar in 18 patients with DME [21]. Patients receiving 1 mg injections experienced excellent reductions in macular thickness and mean improvement in VA (+10 letters) 12 weeks after single injections. Pharmacokinetic analyses based on anterior chamber drug concentrations suggest an extended intraocular half-life of 13.4 days. Multicenter, randomized, double-masked, phase III nAMD trials (CEDAR and SEQUOIA) are comparing q8wk and q12wk abicipar with q4wk ranibizumab. Given abicipar’s high binding affinity, apparently long intraocular half-life, and encouraging results from the early phase trials, the developers are hoping to establish the efficacy of 3-month dosing. Phase III DME trials are being planned but have not yet begun.

Conbercept (KH902, Chengdu Kanghong Biotech Co., Sichuan, China) is a recombinant, fusion protein that, like aflibercept, acts as a decoy receptor. Conbercept (MW of 143 kDa) contains the second immunoglobulin (Ig)-binding domain from VEGF receptor 1 (VEGFR1), the third and the fourth binding domains from VEGFR2, and the Fc region of human IgG. The difference between aflibercept and conbercept is that aflibercept does not contain the fourth domain of VEGFR2 [53, 141]. Conbercept has a high affinity for VEGF because the fourth Ig domain of VEGFR2 is essential for receptor dimerization and it enhances the association rate of VEGF to the receptor [141]. Like aflibercept, conbercept binds all isoforms of VEGF-A, VEGF-B, and placental growth factor. At concentrations from 100 ng/ml to 100 μg/ml, conbercept was not cytotoxic to cultured human retinal endothelial cells (HRECs). A 500 ng/ml solution of conbercept significantly suppresses high glucose-induced migration and sprouting of HRECs by downregulating the expression of PI3K and inhibiting the activation of Src, Akt1, and Erk1/2 [25].

Four weeks after intravitreal injection, conbercept-treated rats had better retinal electrophysiological function, less retinal vessel leakage, and lower levels of PlGF, VEGFR2, PI3K, AKT, p-AKT, p-ERK, and p-SRC than PBS or Avastin-treated rats [58]. The distribution of claudin-5 and occludin in the retinal vessels of diabetic rats treated with conbercept was smoother and more uniform than those of diabetic rats treated by PBS or Avastin. Conbercept has already been approved in China for the treatment of nAMD [79], and a phase III trial evaluating the efficacy of conbercept for the treatment of DME is currently enrolling patients.

Encapsulated cell technology (ECT) was first reported by Bisceglie (1934) as a way to prevent rejection of foreign cell, tissues, or organisms. ECT involves the use of immortalized cells that have been programmed to overproduce a specified biochemical product. The cells are grown in a cylinder lined by semipermeable membranes that allow ingress of nutrients and egress of the synthesized product. The membrane prevents migration of the modified cells and shields them from the body’s immune system. The cylinder is 10 mm long and is surgically implanted through the pars plana and sutured to the sclera.

Clinical studies using ECT to produce ciliary neurotrophic factor (CNF) have been completed in eyes with retinitis pigmentosa and atrophic AMD [70]. The ECT cylinder was well tolerated, but the trials failed to meet their primary therapeutic endpoints. Pharmacokinetic analyses showed that the half-life of CNF production by the cylinder was 54 months. Phase I trials with a cylinder that produced a high-affinity VEGF-binding protein similar to aflibercept have been performed, and a multicenter phase II trial [28] failed to meet its primary efficacy endpoint, thereby calling into question future development and use in patients with DME.

Avalanche Biotechnologies has developed a viral delivery system (AVA-101) to enable the eye to produce long-term anti-VEGF therapy. An adenovirus vector inserts the DNA for a naturally occurring slt-1 (soluble VEGF receptor-1) into retinal pigment epithelial cells. Infected cells synthesize and excrete the soluble VEGF inhibitory protein into the outer retina and choriocapillaris.

In a phase IIa trial, 21 patients with nAMD received AVA-101, with 0.5 mg ranibizumab injected both at baseline and 1 month and again as rescue therapy. Patients underwent core vitrectomy and subretinal injection of AVA-101 adjacent to the macula at day 7. Patients were evaluated monthly and were eligible for rescue ranibizumab therapy based on prespecified criteria. Eleven control patients received only 0.5 mg ranibizumab monthly.

At the 52-week endpoint, mean improvement in BCVA was +2.2 letters in the AVA-101 group compared to +9.3 letters in the ranibizumab group [51]. These differences were statistically significant, but they were largely driven by three subjects in the AVA-101 group who each lost at least four lines of vision. Mean center point thickness improved by −27 μm in the AVA-101 group and −85 μm in the control group. There were no serious ocular adverse events in the AVA-101 group, and no systemic safety signals were noted. All patients in the AVA-101 group who were phakic at baseline developed cataracts, and three (14%) developed moderate vitreous hemorrhages. Gene therapy was well tolerated by patients, but the technology failed to provide a complete or durable anti-VEGF response.

Microelectromechanical system (MEMS) technology is a miniaturized system that is currently used in insulin pumps to deliver drug to the tissues. The Posterior MicroPump Drug Delivery System (PMP, Replenish Inc., Pasadena, CA) uses MEMS technology to deliver drug within the eye. Long-term safety after implantation into animal eyes has been demonstrated [47, 111]. The PMP can reliably deliver 100 programmed doses of an anti-VEGF drug, equivalent to over 8 years of therapy. The PMP was evaluated for 3 months in 11 patients with DME [59]. After episcleral implantation, similar to placement of a glaucoma drainage device, the PMP was well tolerated with no cases of endophthalmitis or strabismus.

PanOptica, Inc. is developing a topical anti-VEGF medication for the treatment of nAMD and PDR. PAN-90806 is a low molecular weight, VEGF receptor blocker administered in eye drop form. Pharmacokinetic studies show excellent drug concentrations in the central retina and choroid as late as 17 h after administration. Animal studies are reported to show control of leakage and bleeding from choroidal neovascular membranes, comparable to that achievable with intravitreal anti-VEGF antibodies, with minimal systemic exposure to the drug. Preliminary results from each of four monotherapy treatment arms in a phase I/II trial were judged by an independent panel of experts to show promise for the treatment of nAMD [94]. Results from a phase II trial of PAN-90806 maintenance therapy after a single anti-VEGF injection are expected to be presented in 2017. The company is moving forward with a phase I trial for the treatment of PDR.

A refillable ranibizumab port delivery system is being codeveloped by Genentech and ForSight Vision 4 to reduce the need for repeated intravitreal anti-VEGF injections. The pre-loaded implant is surgically implanted beneath the conjunctiva through a 3.2 mm scleral incision over the pars plana. The reservoir tip can be accessed easily through the conjunctiva and refilled in the office as needed. The device continuously releases ranibizumab into the vitreous between refills.

A phase I trial for patients with nAMD was performed in Riga, Latvia [105]. At baseline, the reservoir was implanted, and eyes were given 500 μg ranibizumab injections, 250 μg into the vitreous and 250 μg into the reservoir for sustained release. Additional injections were given based on optical coherence tomography (OCT) evaluation of disease activity. The primary endpoint was 12 months with an observation period that extended through 36 months. The primary objective of the study was safety assessment with secondary objectives that included functional measurements.

Four of the patients had significant or serious adverse events (endophthalmitis, vitreous hemorrhage (2), and traumatic cataract), but 3 of these 4 had improved vision by the study’s endpoint. The average visual acuity gains for the cohort were +10 letters; 10 eyes (50%) gained at least 3 lines and 2 (10%) lost at least 3 lines. The mean number of refills was 4.8 per patient.

The planned phase II trial will feature 750 μg injections, with hopes of extending the treatment interval to 4 months.

RTH258 (formerly known as ESBA 1008) is a single-chain, VEGF-binding antibody fragment currently being developed by Alcon (Ft. Worth, TX) for the treatment of nAMD. It has been touted by its developer as having a longer duration of action than currently available anti-VEGF drugs, thereby requiring fewer injections.

A phase II clinical trial compared RTH258 to aflibercept in patients with nAMD [92]. The trial’s primary objective was to compare the efficacy of 6 mg RTH258 against 2 mg aflibercept with the primary endpoint of the study being the mean change in BCVA from baseline to 12 weeks. Secondary endpoints included improvement in central subfield foveal thickness (CSFT) on SD-OCT. Preliminary reports stated that RTH258 produced BCVA gains that were non-inferior to aflibercept with a greater reduction in macular fluid. Patients treated every 3 months experienced a positive effect, suggesting a long duration of action. No new safety signal was seen.

The phase III clinical trial program was initiated in December 2014, with an enrollment goal of 1700 patients in more than 50 countries. These 2-year, double-masked, multicenter trials will randomize patients with untreated nAMD to one of two dosage levels of RTH258 or 2 mg aflibercept bimonthly [29]. The primary endpoint will be changed in BCVA at 48 weeks with several additional secondary functional and morphologic endpoints. No DME trials have yet been announced.

Ziv-aflibercept (Zaltrap®, Regeneron, Tarrytown, NY) is the systemic formulation of Eylea® that is indicated for the intravenous treatment of advanced colorectal carcinoma. Single use vials contain 4 ml (25 mg/ml) of ziv-aflibercept in a buffered solution of polysorbate 20 (0.1%), sodium chloride (100 mM), sodium citrate (5 mM), sodium phosphate (5 mM), and sucrose (20%), with a pH of 6.2. Small series of patients with nAMD that received single injections of ziv-aflibercept had anatomic and visual acuity improvements at 1 month without evidence of toxicity [26]. Two patients with DME had improved VA (20/800 to 20/100; 20/800 to 20/200) and macular thickness (CST: −65 μm and −352 μm) 1 week after intravitreal injections [84]. Additional studies continue to evaluate the use of ziv-aflibercept for nAMD, DME, and retinal vein occlusions.

Intravitreal injections of 2.0 mg adalimumab were ineffective in five eyes with DME that had been refractive to anti-VEGF therapy. No adverse side effects were noted in any of these eyes [136]. In a series of seven eyes of five patients with macular edema from various causes, favorable clinical responses were noted when adalimumab was combined with bevacizumab [9].

Etanercept is a soluble TNF-α receptor that acts as a competitive inhibitor to block TNF-α binding to transmembrane receptors. It reduces leukocyte adherence in retinal blood vessels [67], blood-retinal barrier breakdown, and NF-κB activation in the diabetic retina [66]. Two injections of etanercept (2.5 mg) were performed 2 weeks apart to seven eyes with refractory DME, but no clinical responses were noted at 3 months [131].

Visual acuity changes from baseline to 3 months in 15 patients with DME receiving single injections of 1.0 mg infliximab (1.49 LogMAR to 1.38 LogMAR), and 19 patients receiving 2.0 mg infliximab (0.76 LogMAR to 1.03 LogMAR) were disappointing. Furthermore, 42% of eyes developed severe uveitis with 37% requiring vitrectomy, thereby halting further research with intravitreal infliximab [136].

A double-blind, randomized, placebo-controlled, crossover study in patients with DME that had persisted after two sessions of laser photocoagulation showed that patients receiving intravenous infliximab (5 mg) had significantly improved visual acuity and reduced retinal thickness [118].

Intravitreal injection of the TNF-α-specific inhibitor, pegsunercept, led to a significant reduction in pericyte loss and capillary degeneration in diabetic rats [14, 15].

In clinical studies, low-dose aspirin has shown only little or no benefit in preventing diabetic retinopathy [93]. However, further work is still needed to determine if high-dose aspirin can prevent the development of diabetic retinopathy.

In a randomized trial, 57 eyes with treatment naïve DME received single intravitreal injections of either diclofenac (500 μg/0.1 ml) or bevacizumab. The primary outcome was the change in mean BCVA at 12 weeks, and secondary outcomes included changes in macular thickness, macular leakage, and safety. Eyes receiving diclofenac had better mean improvements in BCVA compared to bevacizumab (Δ −0.08 LogMAR vs. Δ +0.04 LogMAR, P = 0.033), but bevacizumab improved macular edema slightly better [124].