Macular Degeneration

AMD: Shedding New Light on a Dark Problem? - CCR3

News CCR3 - New therapeutic approach in the early stages of wet macular degeneration


CCR3: Shedding New Light on a Dark Problem?

A. Brett Mason* and Josephine Hoh*

Yale School of Public Health, Yale University School of Medicine, 60 College Street, New Haven, CT 06510, USA

*Correspondence to: A. Brett Mason, E-mail: alan.mason{at}; Josephine Hoh, E-mail: josephine.hoh{at}

Received July 13, 2009.
Accepted July 13, 2009.


A recent work by Ambati et al. represents a bold step towards a more effective diagnosis and treatment of age-related macular degeneration, with the new evidence showing that CCR3, a chemokine receptor, is an early marker of and potential therapeutic target for choroidal neovascularization development. In the wake of such a novel and significant finding, additional illumination to confirm and consolidate the promise shown by CCR3 will soon follow.

As life expectancy in the developed world increases, treating disorders such as age-related macular degeneration (AMD) assumes greater importance. The number of people living beyond age 60 is estimated at 600 million in 2000 and 2 billion in 2050 (Bosch, 2002); AMD is the most common cause of blindness in societies with aging populations.

In its advanced clinical manifestations, AMD presents in two distinct forms: geographic atrophy (GA), often called 'dry' AMD, which is characterized by atrophy of the retina, and choroidal neovascularization (CNV), or 'wet' AMD, in which new blood vessels proliferate from the underlying membranes. Both forms of the disease disrupt the tissue near the macula, the area of the retina responsible for high visual acuity. People with AMD may develop one or both forms. Patients with advanced AMD lose vision and may eventually become blind. Of the two forms, GA tends to be slowly destructive, whereas CNV is highly invasive and can rapidly cause legal blindness (de Jong, 2006; Rattner and Nathans, 2006; Jager et al., 2008). Thus, the findings of Ambati et al., recently published in Nature, are welcome news (Takeda et al., 2009). They present evidence that a chemokine receptor, CCR3, holds promise as a novel therapeutic target in treating CNV. Takeda et al. (2009) identify a potential marker of CNV as well as an intriguing new role for CCR3 and its ligands.

To date, CCR3 is known primarily for its expression on eosinophils (Ponath et al., 1996). Eosinophils stimulated by eotaxins, the CCR3 ligands, are typically attracted to the site of inflammation, for example, the airways in allergic asthma. Ambati et al. observe that CCR3 is expressed in choroidal endothelial cells (CECs) isolated from untreated patients with CNV. At the same time, they find no evidence of CCR3 expression in CECs isolated from patients with early stage non-invasive AMD, from age-matched healthy eye tissues or from patients with other proliferating cell conditions. Interestingly, in these CCR3-expressing CNV tissues, the receptor's natural ligands, three eotaxins, are detected, but no eosinophils or mast cells. If the presence of eosinophils or mast cells is the key indication of allergic or inflammatory response, these observations suggest that CCR3 expression bypasses the inflammatory pathway in CNV.

Indeed, the authors point out that CCR3 is involved in CNV angiogenesis rather than inflammation. They show that blocking CCR3 function suppresses new vessel formation, both in vitro, in cultured human CECs, and in vivo, in mice with laser-induced CNV. Anti-CCR3 antibodies prevent tube formation prior to vascularization and reduce the proliferation of CECs following laser-induced injury (Takeda et al., 2009). It is possible that CNV-promoting growth factors present in cell culture medium are responsible for these chemokine-mediated effects, because in human CNV tissues CCR3 expression is only found in the tissue isolated from patients with advanced AMD and not in early AMD or healthy tissues.

There is increasing evidence for the multifaceted roles of eosinophils and mast cells in tissue remodeling and angiogenesis in inflammation. These cells can be both sources and targets for pro-angiogenic mediators. Recent data show a link between inflammation and angiogenesis through which vascular endothelial growth factor (VEGF) and the immune system can interact (Puxeddu et al., 2009). If this is also the case in CNV, is it possible that blocking CCR3 functionally inhibits the inflammatory response in addition to angiogenesis?

Observations by Takeda et al. (2009) suggest that this is not the case in the wild-type lasered mouse model of CNV. First, after laser-induced trauma, inflammatory immune response (IR) cells (i.e. eosinophils and mast cells) are absent from the choroid. More importantly, CNV induced in eosinophil- or mast cell-deficient mouse strains appears no different from that induced in wild-type mice, and anti-CCR3 antibodies are just as effective at reducing CNV in the cellular IR-deficient mice. Collectively, these results imply that the production of laser-induced experimental CNV does not require eosinophil or mast cell participation.

In contrast, neutrophils and macrophages are indispensable for the successful development of experimental CNV (Tsutsumi-Miyahara et al., 2004). Recruitment of these IR cells was not affected by CCR3 targeting. The authors conclude that the anti-angiogenic effect of blocking CCR3 is mediated by direct inhibition of vascularization and not by modification of the cellular inflammatory response.

In the clinical setting, current fluorescein angiographic techniques are only able to detect CNV after the retina has been breached. The ability to detect AMD earlier would be invaluable for diagnosis and timely treatment. Takeda et al. (2009) tested whether Fab fragments of anti-CCR3 antibodies conjugated to Quantum Dots (Qdots), polymer-coated nanocrystals of semiconductor material that produce stable and high-intensity fluorescence, could be used to detect and visualize CNV in strains of mice that develop CNV spontaneously (i.e. mice with knockouts of either monocyte chemoattractant protein-1, Ccl2−/− and/or its receptor Ccr2−/−). Intravenous injection of Qdot-CCR3-Fab facilitated detection of proliferating CCR3+ blood vessels in the choroid that had not yet invaded the retina. Translating these mouse studies to the setting of human disease, it may be feasible to use CCR3 as an early marker for detection of human CNV before disruption of the retina and visual impairment occurs. However, the use of Qdots for in vivo human clinical testing is still under development and so, for the time being, the application of such technology to prevent progression or development of AMD remains a possibility rather than a reality.

Currently, therapeutic interventions for CNV focus on the inhibition of VEGF-A signaling. Ambati et al. compare the capacity of antibodies against CCR3 and VEGF-A to inhibit CNV development in the laser-induced CNV model (Takeda et al., 2009) and conclude that targeting CCR3 is more effective than targeting VEGF-A. They claim that the two molecules follow independent pathways to CNV. On a cautionary note, we do not know how results of CNV inhibition in mice with laser-induced CNV will compare with the results of this treatment in human patients. We do know that the VEGF-inhibiting drugs ranibizumab (Lucentis) and bevacizumab (Avastin), currently used to treat patients with CNV, do not cause significant retinal toxicity (Jeganathan and Verma, 2009). It would be interesting to co-administer both anti-CCR3 and anti-VEGF therapies, either simultaneously or sequentially, to determine if the combination produces a discernable result in their model.

Ambati's group previously generated two mouse models of spontaneous CNV development, one in senescent animals deficient in monocyte chemoattractant protein-1 (MCP1 or CCL2) (Ambati et al., 2003) and the other in younger animals with double knockouts of CCL2 and its receptor CCR2. The authors were obliged to use the Ccl2−/− Ccr2−/− mice to test CCR3 imaging because the laser-induced injury produced the equivalent of 'advanced' CNV but the objective was to detect CCR3 expression well before retinal disruption. As they note in their CCL2 study, 'Current animal models...rely on laser injury to fracture the Bruch membrane, or viral transfection of VEGF into RPE cells. Although these models are useful for experimental study, they are poor facsimiles of the human condition'. Thus, it will be interesting to see the therapeutic effects of CCR3 on their genetic mouse model of CNV.

In the same earlier study, Ambati et al. state that they 'identified CNV with frank evidence of angiographic leakage in 4 of 15 Ccl2−/− and 3 of 13 Ccr2−/− mice older than 18 months…'. In the current CCR3 study, we wonder what percentage of Ccl2−/− Ccr2−/− mice have detectable CCR3 by bio-imaging, what age or clinical stage of AMD they represent and how these variables correlate with the number of animals that subsequently develop CNV.

As for why Takeda et al. did not pursue CCL2 or CCR2 as a therapeutic agent for CNV but chose to study CCR3 instead, we reason that for therapeutic development, it is more challenging to enhance the activity of a molecule that has been lost (as in the case of CCR2 in AMD) than to suppress the activity necessary for disease development (as with CCR3). We are indeed curious how they came across this particular molecule, CCR3, among hundreds of cytokines and chemokines.

Another important goal in developing new therapies is to find a target molecule downstream in the pathway underlying the disease process. For example, in AMD, the genes CFH, ARMD2/HtrA1 and others, discovered in linkage and association studies, play critical roles in AMD development (DeWan et al., 2007). However, they may play this role at the initial stage of their respective processes, with all the processes converging at some point to lead to wet AMD - the CCR3 molecule may be that point of convergence. Thus, although CCR3 is not apparent on genetic screening, it could be a valid and good target for therapeutic development. A potential downside of this type of target is its involvement in fundamental biological processes (in this case, angiogenesis). Blocking its action may have adverse effects. However, the work of Takeda et al. shows that the blocking effect is apparently specific to angiogenesis near the retinal pigment epithelium.

In summary, this study represents a large step in the direction of effectively diagnosing and curing AMD. Despite the convincing results, we remain cautious about the claims until further supporting data appear. With a project of this significance, the wait will likely be a short one, as many AMD investigators will now take an avid interest in CCR3.

© The Author (2009). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved.