By Nikolas J.S. London, MD FACS
Director of Clinical Research, Retina Consultants San Diego
Chief of Ophthalmology, Scripps Memorial Hospital La Jolla
Back in the early 2000s as a third-year medical student I recall sitting in clinic as my Ophthalmology explained the newly-developed treatment option for wet age-related macular degeneration (AMD), photodynamic therapy (PDT) with Visudyne (Bausch+Lomb, Inc, Bridgewater, NJ, USA). While at the time I knew very little about the condition, I understood that it was a devastating disease with little-to-no hope to offer affected patients. The attending also explained that PDT did not reverse the disease, but simply slowed its progression. Looking back, this reminds me quite a bit of our current quest to find a treatment for geographic atrophy (GA) associated with AMD, and, in particular, the unrealized promise of lampalizumab. Similar to exudative AMD before the anti-VEGF era, GA is a disease state in desperate need of an effective treatment option. Like PDT for exudative AMD, many of the drugs being investigated for GA, including complement inhibitors like lampalizumab, hope to slow disease progression rather than reverse it. Currently we have nothing to offer our affected patients other than education and preparation for the inevitable. Fortunately, this need is recognized, and many potential treatment options are in various stages of development, from injectable drugs that hope to slow disease progression to surgical transplantation of stem cells to potentially restore lost tissue. This article will focus on the former, using the recent demise of lampalizumab as a platform to discuss the role of the complement cascade in the pathogenesis of GA, points of potential intervention in this pathogenesis, and what the future holds for this space. We are confident that it is only a matter of time before an effective treatment comes to fruition.
Geographic atrophy is a form of advanced AMD that accounts for approximately 10% of severe vision loss associated with AMD, in general. It is characterized by progressive death of the macular retinal pigment epithelium (RPE) in discrete, well-circumscribed, patches that enlarge over the course of years similar to slowly-growing bacterial colonies on an agar plate. As the RPE dies, the overlying neurosensory retina becomes secondarily affected with atrophy and absolute scotomata development. GA lesions grow slowly, on average 1.5 to 2.1 mm per year. Visual acuity may be relatively unaffected until central involvement develops, which occurs at a median time of 2.5 years after GA diagnosis. Bilateral involvement is also common, with the second eye developing GA approximately 7 years after first seen in the first eye. Nearly all GA lesions expand, albeit at a variable rate and, fortunately, faster toward the periphery than toward the central macula. An important clinical goal is the preserve foveal function and to minimize additional scotomata. Given the advanced age of many affected patients along with the slow rate of GA progression, and interventions that can slow progression even 20% could have a profound impact on visual function. This was the promise of lampalizumab.
The complement cascade is implicated in the pathogenesis of GA, making it a prominent target for therapeutic intervention. While the complement cascade has multiple potential points for intervention, it is complicated and still fairly poorly understood, with three separate but intertwined branches and multiple feedback loops. This is very different from exudative AMD where inhibition of a single, highly-upregulated, cytokine is effective in controlling the disease. Investigators have identified several genetic polymorphisms in the alternative complement pathway that are associated with advanced AMD. One of these is complement factor H (CFH), where a deficiency leads to amplification of the alternative pathway. However, it is not known where in the cascade we need to intervene, whether single or multiple points of intervention would be effective, nor do we fully understand the potential undesired implications of altering this delicate system.
Lampalizumab attempted to block the upstream in the alternative complement pathway. As an anti-Factor D monoclonal antibody fragment, this drug inhibited a key feedback loop and rate-limiting step in this pathway. Phase 2 data were impressive. The relatively small MAHALO study demonstrated a 20% reduction in mean change in GA growth at month 18 in patients treated with lampalizumab and a 44% reduction in patients with the complement factor I (CFI) risk allele, which is a negative regulator of the alternative complement pathway. The MAHALO study was followed by two large phase 3 trials, Spectri and Chroma. These studies were identical, 2-year, multicenter, trials with nearly 2000 subjects randomized to one of four study arms: sham injection every four week, a 10mg lampalizumab injection every four weeks (Lq4), a sham injection every six weeks, and a 10mg lampalizumab injection every six weeks (Lq6). The primary outcome was the mean change in GA area measured by fundus autofluorescence (FAF) at 48 weeks with several secondary outcomes evaluating changes in visual function. The study prioritized enrollment of CFI+ subjects as well as subjects with diffuse or banded patterns on FAF, predictive of GA progression. Unfortunately, the drug did not meet its endpoints. In Spectri, all three groups demonstrated GA enlargement by approximately 2mm2 over 48 weeks, with no significant differences between the groups (Lq4 2.089mm2, Lq6 2.019 mm2, sham 1.932 mm2). Moreover, there was no benefit in the CFI+ subgroup (CFI+ Lq4 2.057 mm2, CFI+ Lq6 2.032 mm2, CFI+ sham 2.007 mm2). There were no new safety signals, including a low rate of endophthalmitis. More detailed results will be presented in 2018. Suffice it to say, these results were unexpected and disappointing. However, we should keep in mind that 50% of phase 3 clinical trials fail, despite sound pre-clinical and positive phase 2 data. Moreover, all is certainly not lost. The Chroma and Spectri studies enrolled over 600 patients in the sham arms, providing the largest data set of its kind on the natural history of GA including data on ETDRS visual acuity (both best-corrected and low-luminance), objective assessments of visual function with reading speed and macular microperimetry, subjective assessments of function with the NEI VFQ 25 and Functional Reading Independence Index (FRI Index), high-quality serial imaging with color photography, FAF, spectral-domain optical coherence tomography (SDOCT), near-infrared, and fluorescein angiography. It is exciting to consider what we will learn over the next few years as these data are scrutinized.
While lampalizumab was ineffective, several other complement inhibitors for GA are being developed, including a peptide against C3, an antibody against properdin and C5, gene therapy directed toward the culmination of the cascade, the membrane attack, complex, and an aptamer against C5. These are in various stages of development, including phase 1 and 2 clinical trials.
An anti-C3 peptide may be more effective as it is a central inhibition of complement – C3 is involved in all three complement pathways, lectin, classical, and alternative. In a phase 2 study of APL-2 by Apellis (Crestwood, KY, USA), the drug was given monthly (n=84) or every other month (n=78) with GA lesion growth compared to sham over 12 months. The treatment groups showed a 20% and 29% reduction, respectively, with little to no difference in visual outcomes between groups. One point of concern was that treated eyes appeared to have a higher incidence of new choroidal neovascular membrane (CNVM) formation, up to 30% over 12 months with monthly treatment and a history of CNVM in the fellow eye. Potentially even more concerning, there was also a higher incidence of endophthalmitis in treated groups – 2.3% in the monthly arm and 1.3% in the arm treated every other month, however it is difficult to know the clinical significance of this given the small size of the trial.
Another agent in development by Novartis (LFG316, Basel, Switzerland) is an anti-C5 antibody. C5 is cleaved into 5a and C5b, which activate different parts of the complement cascade. This drug did not meet the primary endpoint as monotherapy in a phase 2 trial, but is now being evaluated in combination with an anti-properdin antibody. Properdin is a molecule that perpetuates the effect of the C3 enzyme and greatly amplifies activation of the alternative pathway. Blocking properdin may blunt an important amplification loop and may be synergistic with blockage of C5.
An anti-C5 aptamer (Zimura, Ophthotech Corp., New York, NY, USA) also hopes to slow the rate of GA progression, and was studied in a 48-week phase 1b/2a trial with two dose levels, 0.3mg (n=24) and 1mg (n=23) and injections per eye. There were no safety signals, including no new CNVM formation seen, and a trend toward a dose-response treatment benefit. A phase 2 randomized study is being planned with a target of 200 subjects in three cohorts (2mg/eye, 4mg/eye, sham).
Finally, another target is the culmination of the complement cascade – the membrane attack complex (MAC). This intriguing approach from Hemera Biosciences Inc. (Boston, MA, USA) involves gene therapy to induce the expression of soluble CD59. CD59 is a natural membrane-bound protein that inhibits MAC formation by blocking incorporation of C9. A phase 1 dose-escalation study has completed enrollment of 17 subjects with no safety issues noted and stable vision. If this approach works, a single intravitreal injection may enable long-term protection through protein expression.
In summary, the complement cascade appears to be a strong candidate for therapeutic intervention in preventing irreversible vision loss due to geographic atrophy associated with AMD. However, it is a very complicated system, and we will likely encounter multiple dead ends before we reach our goal. This is what happened to lampalizumab, with other investigational drugs certain to suffer the same fate. Why is an effective therapy so elusive? Treating GA is not like exudative AMD with a single significant target and will likely require a multifaceted approach. Moreover, treating GA is akin to treating programmed cell death and it may be that the effect of interventions may not become apparent for some time after initiation. There is also not a good pre-clinical model on which to base GA trials, further increasing uncertainty. Finally, there is the issue of stoichiometry. The concentration of complement components can be significantly higher (up to a million times) than levels of VEGF in the eye – making continuous, long-term inhibition a daunting task.
Thomas Henry Huxley (1825-1895) once stated, “The great tragedy of science—the slaying of a beautiful hypothesis by an ugly fact.” As these “ugly facts” pile up in our quest to find a therapy for GA, we hope that we will continue to learn, redirect inevitable frustration, and continue our pursuit to preserve our patients’ vision.
Thanks for reading. Please don’t hesitate to contact me with any questions.
Best wishes, and until next time,
Nikolas London, MD, FACS
Retina Consultants San Diego, Poway, La Jolla, and Coronado
firstname.lastname@example.org (personal email)
email@example.com (RCSD email)