Article Date: 1/1/2010

Exhibition of Inhibition
RPS: From the Podium to the Practice

Exhibition of Inhibition

Inhibiting proteins and enzymes to treat AMD started with VEGF. Where is it heading?


In our final recap of presentations from the 2009 Retinal Physician Symposium in the Bahamas, we focus on inhibition of proteins and enzymes in treating retinal disease. While one of the four experts in the field talked about VEGF, most of them did not. Here is our summary.


A presentation on integrin antagonism given by Jeffrey S. Heier, MD, of Ophthalmic Consultants of Boston, began with consideration of the strides made in the treatment of retinal disease by blocking VEGF. “How can we improve upon these outstanding results?” Dr. Heier asked. “Well, you might be able to do it by increasing efficacy, and to do that, you might look at certain things that limit efficacy right now.” He indicated that fibrosis and atrophy were potential areas to examine for increased efficacy.

Dr. Heier then spent a few moments defining what integrins are. He stated that they are a family of transmembrane heterodimeric proteins that are involved in cell survival, proliferation and differentiation. They mediate adhesion between cells and play direct roles in signaling and modulation of kinase signaling pathways. Dr. Heier turned specific attention to integrin α5β1, which is a receptor of fibronectin that is expressed on granulocytes, macro phages, and growing vessels. “It's been validated in ocular models of neovas-cularization as well as tumor models where suppression has aided tumor angiogenesis,” Dr. Heier said.

After demonstrating the effects of selective expression of α5β1 on pancreatic tumors, as well as in AMD and proliferative vitreoretinopathy, Dr. Heier turned to JSM6247, a molecule developed by Jerini AG of Berlin that potently protects α5β1 binding to fibronectin. JSM6247 has been tested in various animal models, but Dr. Heier stated, “A lot of people say, ‘Well, gosh, that really doesn't have any application to humans.’ But I've never seen a drug not demonstrate efficacy in animal models — assuming appropriate species-specific targets — and demonstrate good efficacy in humans. It's always the other way.”

Dr. Heier briefly reviewed the animal data on JSM6247, concentrating on the molecule's ability to reduce scarring, and then reported phase 1 data. Among patients enrolled in this trial, there was a modest increase in mean ETDRS letters by day 15 (Figure 1). However, OCT measurements for mean changes in retinal thickness were not conclusive. Briefly looking at a few individual cases from the study, Dr. Heier concluded the presentation on JSM6247 by noting that the molecule's very short half-life will dictate the need for a different delivery system.

Figure 1. Mean visual acuity results from the phase 1 trial of JSM6427.

The remainder of Dr. Heier's presentation focused on another drug, volociximab or M200 (Figure 2), developed by Ophthotech of Princeton, NJ. M200 is a chimeric monoclonal antibody that inhibits binding of fibronectin to α5β1. It has been shown in vitro to have the ability to block human umbilical venous endothelial cell (HUVEC) tube formation. Dr. Heier showed slides of the HUVEC experiments as well as avian and rabbit models before informing the audience that M200 is in phase 1 trials and interim data are forthcoming. “So what can we say about the integrins?” Dr. Heier asked by way of closing. “Clearly there's potential here for synergism with the anti-VEGF agents. We've seen early evidence of biologic activity. The JSM agent is well tolerated and, to date, so is Ophthotech's.” Dr. Heier noted that the potential exists with integrin antagonism not only for antiangiogenic activity, but also antifibrotic and anti-inflammatory activities. “The possibility of synergism is certainly very exciting,” Dr. Heier said.

Figure 2. The molecular form of volociximab.


David M. Brown, MD, director of clinical research at Greater Houston Retina Research and a partner at Retina Consultants of Houston, presented one-year data from the CLEAR-IT 2 trial of Regeneron's (Tarrytown, NY) VEGF Trap-Eye. Beginning by echoing Dr. Heier's point that our current animal models for age-related macular degeneration are inadequate, Dr. Brown said, “ We basically have a wound-healing model where you burn a hole in the back of the eye with a laser turned up to the point where you break through Bruch's membrane, and the resulting neovascular response is much more like a myopic membrane. It's not like an AMD membrane. So we test multiple compounds in a bad model, and just because they don't work in a bad model — we abandon that compound's use and move on— The compound might actually work in human AMD — but you have to have demonstrate efficacy before you put the time and the effort in and the FDA regulation to push them into human trials.”

Then Dr. Brown turned to VEGF Trap-Eye itself, remarking, “We're very excited about this compound. Our group has been involved for a long time with VEGF Trap.” He described the drug as a fusion protein that has an amazing affinity for or tight binding to the VEGF molecule. He discussed binding constant numbers for anti-VEGF agents, noting that the lower the constant (in picomoles [pmol]), the higher the binding affinity of an agent. VEGF receptors 1 and 2 have binding constants of 10-30 pmol and 100-300 pmol, respectively. By contrast, the binding constant of VEGF Trap-Eye is 0.5 pmol — 300 times higher than that of ranibizumab and thousands of times higher than that of bevacizumab.

Here, Dr. Brown began presenting the CLEAR-IT 2 data. Noting that the initial 12-week data had already been presented, Dr. Brown stated that the data he would be presenting would consist of the prn dosing phase following the week 12 data. The three main questions, Dr. Brown said, were: “Is this really going to potentially be a drug where we have less injections, or is it a ‘me too’ drug? Are we going to have the gold-standard efficacy we saw with the landmark papers of MARINA and ANCHOR? And is a sequential loading dose important?”

In answering the questions in order, Dr. Brown presented the retreatment criteria (essentially, loss of five letters, plus fluid on the OCT) and then discussed outcomes. The median time to reinjection over all initial dosing regimens was 110 days (Figure 3), and only about 5% of patients required monthly retreatment. Across all cohorts, the mean retreatment was 2.01 injections.

Figure 3. Reinjection data from the CLEAR-IT trial.

Answering the second question, Dr. Brown explained that mean efficacy of VEGF Trap-Eye was demonstrated by changes in central retinal thickness (−130 μm over 52 weeks; Figure 4) and change in area of active CNV lesion size as measured on fluorescein angiography (between −1.42 mm2 and −3.41 mm2 in four statistically significant dosing regimens).

Figure 4. Mean change in CRT and lesion thickness from the CLEAR-IT trial.

Dr. Brown answered the third question by saying, “If you can show that a sequential load matters in a drug with a 0.5-pmol affinity for VEGF, it's really going to matter for ranibizumab at 150 pmol affinity and bevacizumab which has a 1500 pmol affinity for VEGF.” However, Dr. Brown conceded, a loading dose did not seem to have the effect of a longer effect for VEGF Trap-Eye. However, the regimens including a loading dose did have a considerably higher mean change in visual acuity (9.0 and 5.4 letters vs 2.6 – 5.2 letters, depending on dose). Even the highest-dose group, which did not include a loading dose, did not improve at the rate of lower doses with a loading dose. Dr. Brown named the 2 mg sequential load dose the “chicken dinner winner” dose. In closing, Dr. Brown previewed the VIEW1 and VIEW 2 studies, which are still enrolling, and a DME trial being conducted at Johns Hopkins led by Quan Nguyen and Peter Campochiaro.


Dr. Brown also gave a presentation on the topic of tyrosine kinase inhibitors (TKIs). His stated goals in this talk were to compare the tyrosine kinase receptor to the fuel-injection system of a car; to illustrate the mechanics of the tyrosine kinase receptor; and to summarize the results of a six-month study of a TKI.

He began, however, by broadly discussing angiogenesis, describing it as a “very complex process where blood vessels grow where they're not supposed to, under the subretinal space and into the retina and subretinal RPE space.” And, Dr. Brown added, while we know that by inhibiting VEGF, we can shut angiogenesis down, he stated something that may have surprised many audience members: “Until very recently, there's never been any definitive demonstration that there's always increased VEGF in AMD.” Rather, he clarified, the system is like a teeter-totter, where agents help and inhibit growth of blood vessels or leakage. VEGF is probably one of these agents, but it is not the only one.

Dr. Brown likened the process of angiogenesis to NASCAR fuel-injection systems, where the pedals of the car turn on the gas (initiate angiogenesis), turn off the gas (inhibit angiogenesis), and hit the brake (stop it altogether). The current options of anti-VEGF agents (ie, bevacizumab, ranibizumab and VEGF Trap-Eye) block the gas. Pigment epithelium-derived factor (PEDF), endostatins and other agents, Dr. Brown said, may have the effect of hitting the brakes. Targeting the fuel injector itself, Dr. Brown intimated, would involve inhibiting tyrosine kinase (Figure 5).

Figure 5. The role of TKIs in the receptor inhibition mechanism.

After a brief discussion on blockade of all receptors vs selection blockage, Dr. Brown then turned to TKI data in humans. He pointed to a DME study of PKC412 conducted by Peter Campochiaro, MD, at Johns Hopkins. The problem Dr. Campochiaro encountered was that, while PKC412 was effective, it caused elevated liver enzymes.

Dr. Brown then turned to TargeGen's (San Diego, CA) prodrug TG00801, which outperformed bevacizumab in an animal trial. In a laser-induced CNV model, TG00801 reduced CNV. Dr. Brown punctuated this statement with OCT images of a patient with subretinal fluid who improved under treatment with TG00801. “The problem with this drug,” Dr. Brown said, “is that the formulation we tested left red deposits in the cornea in some patients. TargeGen is currently working on changing the formulation”.

He then showed slides from results that had been achieved using a third TKI agent from an undisclosed pharmaceutical company, with month six OCT and fluo-rescein angiography that demonstrated anatomic efficacy. No leakage had returned at six months. “What's the problem?” Dr. Brown asked, answering, “The vitreous looked like a snow-globe, with refractile crystals that decreased the visual acuity.”

The last drug Dr. Brown described was GlaxoSmithKline's (Philadelphia, PA) pazopanib hydrochloride, which effects a broad blockade of three separate VEGF receptors, PDGFR alpha and beta, and C-Kit (Figure 6). This broad inhibition of multiple pathways, Dr. Brown said, could make it more effective but at the same time may make it more likely to have undesired side effects. It's currently being tested in phase 1 trials in eyes that are end-stage. Dr. Brown concluded by saying that, while targeting the fuel injector makes sense, “The formulation and the side effects have yet to get us anything that's ready for prime time.”

Figure 6. The mechanism of action of pazopanib.


Pravin U. Dugel, MD, managing partner of Retinal Consultants of Arizona in Phoenix and founding member of Spectra Eye Institute in Sun City, AZ, spoke about mTOR inhibitors — enzyme inhibitors that target the mammalian target of rapamycin (mTOR). The focus of Dr. Dugel's talk was, of course, rapamycin or sirolimus (Rapamune, Wyeth), which, he explained, has been approved by the FDA for the treatment of organ rejection, coronary restenosis and advanced renal-cell carcinoma. “You'll notice that, in cancer research, there was a tremendous amount of research up until the last few years on VEGF, and now it's really switched to mTOR, and that's because mTOR occupies a key intracellular point of convergence,” Dr. Dugel said (Figure 7).

Figure 7. The position of mTOR in the angiogenesis pathway.

After demonstrating sirolimus's history of safety and efficacy, Dr. Dugel addressed the challenge of formulating rapamycin for ocular use. MacuSight (Union City, CA), the company for which Dr. Dugel acts as a consultant and adviser, received independent confirmation from a Japanese company that the best route of administration for sirolimus is periocular (Figure 8). “So, the challenge for MacuSight then has been to develop a sirolimus formulation that is safe for the eye, can be administered locally with minimal systemic exposure, can be administered easily through a standard 30-gauge needle, and will have a reduced frequency of administration,” Dr. Dugel said.

Figure 8. Local ocular delivery of sirolimus.

First, however, Dr. Dugel turned to trial data from studies looking at sirolimus as a treatment for DME and AMD. In these phase 1 trials, sirolimus's safety profile was confirmed. “What about efficacy? What about a biological signal?” Dr. Dugel asked, answering his own questions with data suggesting that, even in patients with significant fibro-sis because of refractory disease, there was a signal of improved visual acuity and OCT changes detectable out to 90 days following a single intravitreal or subconjunctival injection. Dr Dugel also mentioned that in all trials, DME and AMD, the subconjunctival route was as effective and as safe as the intravitreal route.

Dr. Dugel then turned to the ongoing phase 2 trials of sirolimus, which include AMD and DME trials, as well as trials for uveitis and dry eye. He looked at the EMERALD trial first, which is currently enrolling patients to study sirolimus for wet AMD using subconjunctival injection and redosing every two months. The EMERALD trial is unique, Dr. Dugel noted, because patients who have undergone prior treatment are being included in order to simulate the “real world” situation and to not compete with the CATT trial, which MacuSight and all the study investigators “consider to be of great importance for our profession, our patients, and our healthcare system.” The CATT trial is under way to compare the efficacy and safety of Avastin vs Lucentis for AMD.

Dr. Dugel also mentioned another trial he is carrying out at his own practice — a 20-patient monotherapy trial of sirolimus for AMD — and presented some data from the 19 of 20 patients already enrolled and injected in that study.

Turning finally to the phase 2 trial for sirolimus in DME, Dr. Dugel referred back to the hypoxia-inducible factor-1 alpha pathway involved in the DME disease process before laying out the protocols for the ongoing DME study, called DIAMOND. Dr. Dugel closed his presentation by saying, “I personally believe that this particular formulation has an excellent chance of being a combination agent for DME and AMD, and perhaps a monotherapy agent also. It's rare to find an FDA-approved drug with proven safety and efficacy systemically for over a decade that has such promising characteristics for the treatment of multiple retinal diseases. I sincerely believe all retina specialists will be hearing about sirolimus over the next few years. I look forward to presenting the phase 2 data for both DME and exudative AMD in the near future.”


Peter K. Kaiser, MD, professor of ophthalmology at the Cleveland Clinic's Lerner College of Medicine and vitreo-retinal staff at the Cole Eye Institute, gave the keynote address, discussing platelet-derived growth factor (PDGF). Dr Kaiser noted that while anti-VEGF agents are effective at reducing choroidal neovascularization (CNV) leakage, they rarely lead to regression of the CNV itself. Thus, if the goal is to eliminate CNV, then retinal physicians will have to look at treatments other than anti-VEGF monotherapy despite its efficacy in clearing up CNV leakage.

One reason why anti-VEGF may not successfully eradicate CNV, Dr. Kaiser said, is the age of the neovascularization. He notes that older lesions are more refractory to therapy, and this may be based on the presence of pericytes. Pericytes are recruited to neovascular vessels by PDGF. In particular, PDGF-B is involved in the endothelial cell and pericyte association and recruitment. “So it's a key step in the maturation of neovascular vessels,” Dr. Kaiser said.

He then demonstrated where and how in the angiogenesis cascade PDGF-B enters the picture (Figure 9). PDGF levels coincide with neovascular growth and more PDGF is expressed as CNV occurs. On this basis, Ophthotech has developed an aptamer that blocks PDGF-B, called E10030 (Figure 10). One characteristic of E10030 is that it strips pericytes from new vessels without affecting normal vessel. “This is an interesting idea,” Dr. Kaiser noted. “If you can strip the pericytes, you can make those endothelial cells more susceptible to VEGF blockade. And herein lies the idea behind using combination therapy — an anti-PDGF agent with an anti-VEGF agent: The anti-VEGF agents won't work on these mature vessels that have pericytes surrounding them. By stripping the pericytes, you now make them sensitive once again to anti-VEGF.”

Figure 9. The angiogenesis cascade.

Figure 10. Ophthotech's E10030 acting against PDGF-B.

The association between PDGF and VEGF is made even stronger by the fact that pericytes excrete VEGF. Dr. Kaiser pointed to a paper on Avastin in oncologic tumors published in the FASEB Journal that demonstrated how selected VEGF inhibition could not induce blood vessel regression in tumors, whereas combined blockade of both PDGF and VEGF did induce regression. Dr. Kaiser also pointed to a model of combination blockade in corneal neovascularization animal model, where combination therapy was effective in causing regression of not only for new vessels, but also for established vessels. “The other thing,” Dr. Kaiser said, “is that using this combination treatment, you can actually get apoptosis of neovascular endothelial cells. So you're actually causing these neovascular cells to regress, which would be ideal for CNV.”

Dr. Kaiser then turned to Ophthotech's phase 1b study of E10030, which compared combination therapy with E10030 and Lucentis to a single dose of E10030 followed by Lucentis alone. At four weeks, mean visual acuity in the combination arm had increased by an average of +12.3 letters, increasing to +14.2 letters at week 8 and +16.5 letters at week 12. Three-line gainers ran as high as 64% and mean change in foveal thickness decreased by 255 μm.


The 6th Annual Retinal Physician Symposium will be held this year from April 21 to 24 again at the Atlantis Paradise Island Resort in the Bahamas. More information can be found on 'LASIK' of this issue. RP

Retinal Physician, Issue: January 2010