FIGURE 23.1. A: Intense angiogenesis leads to a plexus of vessels with shunts and saccular dilations. B: Antiangiogenic therapy can lead to a pruning of vessels, particularly those that are relatively poor in pericyte coverage. The remaining vessels adopt a more ordinary phenotype (B) in a process known as vascular normalization. (Reproduced from Spaide. Am J Ophthalmol. 2006;141:149–156, with permission.)

NORMALIZATION OF TUMOR VASCULATURE

The early vessels in angiogenesis do not organize graceful filigrees with the sequential hierarchy of arterioles, capillaries, and venules; instead, they form a relatively chaotic network of loops, shunts, and saccular dilations. The wall structure of these early vessels is abnormal, with large inter-endothelial junctions, more fenestrations, vesicles, and lack of a normal basement membrane.³⁴ The surrounding perivascular cells are equally abnormal. As a result of this defective vascular organization, leakage occurs and increases tissue hydrostatic pressure, which may alter blood flow as well.³⁵ The resultant local differences in blood flow may lead to inadequate delivery of O² and metabolites in certain areas of the tumor. The resultant imbalance of proangiogenic and antiangiogenic factors is presumably key to theses vascular anomalies and their improper maturation.³⁶ Antiangiogenic treatment can effectively prune the more immature, poorly formed vessels from the vascular complexes, and leave vessels that adopt a more normal phenotype. This process has been called vascular normalization.^37,38 This outcome may occur because more mature vessels with pericyte coverage are protected through pericyte mediated survival mechanisms, which involve Ang-1/Tie2 signaling between the pericytes and the endothelial cells.^7,8 Vascular normalization has been associated with improved outcomes, partly because tumor-associated edema is relieved,³⁹ but possibly also because of improved delivery of chemotherapeutic agents through improved blood flow to and decreased interstitial fluid pressure within the tumor.³⁴ The effect of bevacizumab in colon cancer could have been attributable to the improved delivery of chemotherapeutic agents rather than direct antiangiogenic induced tumor regression in one study.⁴⁰ Interestingly, the effect of radiation also appeared to be enhanced when combined with anti-VEGF treatment in human glioma xenografts,³⁴ and this synergy could only be observed during a time window, when “normalized” vessels were present (Fig. 23.2).

FIGURE 23.2. Choroidal neovascularization secondary to AMD is a tissue invasion of a multitude of cells, many of which are inflammatory. The actual vascular component is a small minority of the lesion and grows into in with and to an extent because of, the extravascular component. (Reproduced from Spaide. Am J Ophthalmol. 2006;141:149–156, with permission; Photograph courtesy of Hans Grossniklaus, MD.)

TWO-COMPONENT MODEL OF CNV

Cancer is constituted of two principle components, tumor cells and vascular cells. Proof of the principle has been provided that treatment of both components may be more efficacious compared with treatment of either component alone.^41,42 CNV is a condition commonly referred to as “neovascularization” because of its angiographic appearance. This neglects the fact that histopathologically, CNV is really a tissue invasion of vascular cells and extravascular cells.^43,44 In analogy to solid cancers, CNV can be thought of as a two-component system as well. Just as we described for cancers, the vascular CNV component consists of endothelial cells, circulating endothelial progenitor cells, and pericytes. The extravascular component is composed of invading inflammatory cells such as macrophages, lymphocytes, granulocytes, foreign body giant cells, as well as fibrocytes, myofibroblasts, retinal pigment epithelial cells, and glial cells.^44–47 In pathologic angiogenesis, resident cells recruit vessel progenitors, and vessels themselves enhance further growth. For example, stromal fibroblasts colocalize with the tumor vasculature even deep inside the tumor,⁴⁸ and this has been shown to be tumorigenic in and in-vivo model of breast cancer.⁴⁹ The same appears to be true for inflammatory cells in cancer, where they may counter-intuitively stimulate angiogenesis and tumor growth, rather than combat it.^50,51 In particular, macrophages can have either proangiogenic or antiangiogenic functions, depending on their phenotypes.⁵² While classical, or M1 macrophages have cytotoxic activity, M2 macrophages suppress the immune response, promote wound healing, angiogenesis (by secretion of factors such as VEGF) and tissue remodeling, and thus are thought to aid in tumor progression. The important role of inflammatory cells in neovascularization that has been demonstrated in tumor angiogenesis appears to apply to CNV as well, where the amount of VEGF expressed is proportional to the number of inflammatory cells present and VEGF in CNV colocalizes with the inflammatory cells.^44,53 This may be further amplified by VEGF induced expression of cellular adhesion molecules in vessels, thus recruiting more inflammatory cells.⁵⁴ Further, supportive evidence for a crucial role that inflammatory cells have in pathogenesis of CNV comes from animal experiments in which depletion of monocytes inhibits the formation of CNV.⁵⁵ Even in health, hypoxia prevails in the outer retina, and this may be a further stimulus for invading cells to produce angiogenic factors.^56,57 Over the course of the disease process CNV produces vascular invasion, proliferation, matrix formation with or without hemorrhage, remodeling of the tissue, and eventual vascular regression in involutional lesions.⁵⁸ This is more consistent with a wound healing response than growth of vessels in response to ischemia alone, where the invading vessels would attempt to recapitulate the organization of the choriocapillaris.^43,58,59

TWO-COMPONENT MODEL AND COMBINATION THERAPY

Therapy for CNV is aimed at the vascular component, the extravascular component, or potentially both simultaneously. Inhibiting the vascular component addresses the neovessels but does not directly affect the extravascular component, although it is possible, that reduced blood supply may lead to reduction of the extravascular component through ischemia. Conversely, ischemia, and resultant tissue damage may lead to increased elaboration of angiogenic and proinflammatory factors.

Limitations of Monotherapies

Several monotherapies directed predominantly against the vascular component have been developed and tested in large clinical trials. They include photodynamic therapy (PDT) with 689 nm wavelength laser-activated verteporforin (Visudyne, Novartis Pharmaceuticals Corporation, Hanover, NJ) and pharmacological approaches such as pegaptanib (Macugen, Eyetech Pharmaceuticals, New York, NY), ranibizumab (Lucentis, Genentech, South San Francisco, CA) and bevacizumab (Avastin, Genentech, South San Francisco, CA). Initial data from studies examining pegaptanib have shown continued growth and leakage, not shrinkage, of the neovascular lesions. This may suggest a process like vascular normalization may be occurring, possibly related to incomplete inhibition of possible mediators of angiogenesis.⁶⁰

Ranibizumab was tested in two large randomized clinical trials: ANCHOR (predominantly classic lesions treated with either PDT or ranibizumab)⁶¹ and MARINA (predominantly occult lesions randomized to ranibizumab or sham injections).⁶² Both trials demonstrated mean visual gains after 24 months, and the 24 month ANCHOR trial data demonstrated that monthly injections of ranibizumab provided patients who had angiographically predominantly classic CNV lesions with better visual outcomes than the previous standard of care, PDT. For the first time did a therapy for wet AMD result in improved vision, not just longterm stabilization at the cost of initial visual loss. Similar results appear to be achieved with bevacizumab,^63,64 which is frequently used in an off-label fashion for the same indication^65–67 and its equivalence is the subject of an ongoing NIH sponsored trial (http://www.nei.nih.gov/CATT).

However, this therapeutic paradigm shift now commits the patient to monthly office visits and invasive treatment, arguably because not all pathogenetic components are addressed, thus maintaining a chronic disease state. While inhibition of VEGF also may decrease the inflammatory effects of induced cellular adhesion molecule expression and inflammatory cell recruitment to the lesion, administration of anti-VEGF agents may induce intraocular inflammation as well, so the net anti-inflammatory effect is not known.

As a case in point, the PIER trial attempted to induce with three monthly injections, followed by quarterly injections, with only about 13% of patients gaining more than 15 letters compared to 33% to 40% of patients receiving monthly injections in the ANCHOR and MARINA trial, demonstrating that the vast majority of patients needed more frequent injections. While several dosing regimens have been designed to tailor a schedule to the individual patient by OCT guidance,^68,69 thus reducing the need for injections, more or less regular injections are still required. No matter which regimen is used, it has become clear that patients will depend on long-term, and possibly lifelong injections, as demonstrated in the HORIZON⁷⁰ and other trials examining the longterm outcomes of AMD treated with anti-angiogenic drugs.^68,69

The strategy of inhibiting the extravascular component would include targeting the action of the extravascular cells or important cytokines (such as with immunomodulators or biologicals),⁷¹ thus hopefully addressing all factors that maintain the pathogenic process and consequently lead to chronicity of disease. In cancer, the large variety of possible cancer cells, their potential to secrete inappropriate cytokines, and their genetic instability make such an indirect approach difficult.⁷² In CNV, however, the makeup of invading cells is fairly consistent from patient to patient, and these cells are genetically stable. There are no published reports of pure inhibitors of the extravascular cellular component of CNV at present, although corticosteroids inhibit many aspects of the extravascular component. Some therapies, such as thermal laser, radiation,⁷³ and the combination of PDT and intravitreal triamcinolone⁷⁴ inhibit both components (Fig. 23.3). Due to its indiscriminate action not only on the vascular and extravascular component, but also innocent neuroretina, thermal laser has been essentially abandoned for the treatment of subfoveal and most other CNV. Therefore, we will focus on current attempts to halt CNV with radiation, PDT, anti-inflammatory agents, or a combination of them.

FIGURE 23.3. A: At presentation, visual acuity was 20/40. The patient had one combined treatment of intravitreal triamcinolone and PDT. B: One year later, there was resolution of the leakage but some persistent staining. Visual acuity was 20/20. (Reproduced from Spaide, et al. Retina. 2005;25:685–690, with permission.)

Radiation Therapy for AMD

Ionizing radiation has widespread cellular effects, affecting neovascularization, inflammation and tissue remodeling, depending on the type, dose and delivery time of radiation, as well as the target tissue. Since the biological effect of ionizing radiation is mediated by breaking double-stranded DNA, it is more effective on rapidly dividing cells, such as lymphomas and carcinomas, less effective on fibroblasts, and least effective on mature cells, such as mature endothelial cells.⁷⁵ Theoretically, this could target all components simultaneously, both active vascular and extravascular while sparing mature neighboring cells. Various treatment strategies have been explored and are evaluated in a major review by the Cochrane Collaboration, with variable results.⁷⁶ Several approaches are currently under investigation. One employs a beta-radiation strontium-90 applicator (EpiRad, NeoVista, Fremont, California), which is inserted into the eye after a limited vitrectomy and held over the macula to deliver 24 Gy of radiation (Fig. 23.4). This epimacular low-dose mono-brachytherapy had promising preliminary results⁷⁷ and its safety and efficacy is currently being evaluated in the CABERNET (CNV Secondary to AMD Treated with Beta RadiatioN Epiretinal Therapy) phase 3 clinical trial in combination with ranibizumab and compared to ranibizumab alone. The MERITAGE (Macular EpiRetinal Brachytherapy in Treated Age Related Macular Degeneration Patients) trial will examine whether the treatment burden can be alleviated in patients who are unable to maintain vision without regular anti-VEGF injections (Box 23.1).

FIGURE 23.4. An illustration of the intraocular, epiretinal beta radiation (strontium-90) delivery device placed in proximity to the CNV complex.

(Reproduced from Avila, et al. Retina. 2009;29:157, with permission.)

It is possible that normalization of the vascular organization by VEGF inhibition prior to radiation, an approach which has been reported in combination with ionizing radiation of solid tumors, would lead to better responses. In the case of PDT, one retrospective study demonstrated that only 48% of patients receiving a single intravitreal injection of 1.25 mg bevacizumab 2 weeks prior to PDT required retreatment, which is much lower than previously reported.⁷⁸

While radiation therapy promises long-term control of wet AMD through addressing both the neovascular and extravascular components of the disease, it is unclear what the longterm effects on the vasculature, as well as “innocent” bystander cells are. Radiation associated choroidal neovasculopathy (RACN) has been described as a late sequel of external beam radiation.⁷⁹ Another concern is that the retinal pigment epithelium (RPE) may suffer as well and that wet AMD might turn into geographic atrophy, for which no effective treatment is available to date.

Combination Therapy with PDT and Anti-VEGF agents

PDT with Verteporforin was the first FDA approved treatment for neovascular AMD. Few patients actually experienced visual improvement, but rather stabilization of vision and lesion regression in the pre-pharmacological era.^80–82 Lipophilic verteporforin is taken up by neovascular endothelial cells through the Low density lipoprotein (LDL)-receptor, which isoverexpressed by cells undergoing rapid division.⁸³ Subsequent activation with a nonthermal laser generates reactive oxygen species, which damage the neovascular endothelial cells, leading to platelet binding, aggregation, and vascular occlusion.⁸⁴ However, this effect is not as specific as desired, and reduction of blood flow in physiologic choriocapillaris has been demonstrated in multiple instances after PDT.^85,86 This leads to varying degrees of outer retinal ischemia as well as inflammation, stimulating regrowth of new vessels, apparently mediated in part by VEGF.⁸⁷ To antagonize this, double and triple combinations of all permutations of PDT, an anti-inflammatory and/or an angiogenic drug have been proposed. Posterior sub-Tenon injection of triamcinolone combined with PDT lead to decreased choriocapillaris ischemia at 3 months compared to PDT alone, but no difference at 12 months.⁸⁸ However, there was a decreased need for retreatment, indicating that the short-term effect on choroidal perfusion may indeed be biologically relevant. However, despite encouraging results from several prior studies, the VERITAS trial, which compared the visual acuity response of PDT with verteporforin as compared with PDT and two different doses of triamcinolone, did not show any significant benefit in the triamcinolone groups.⁸⁹

The multicenter FOCUS trial, examined the efficacy and toxicities of combined PDT and ranibizumab in patients with neovascular AMD over the course of 2 years.⁹⁰ PDT was administered at baseline and then as needed every 3 months to all 162 patients. In addition to obligatory treatment with PDT, 106 patients received monthly injections of 0.5 mg ranibizumab and 56 received sham injections. At the end of 24 months, 88% of those on the combination versus 75% of those on PDT alone had lost <15 letters in visual acuity, while 25% of the combination versus 7% of the PDT-alone patients had gained at least 15 letters (p < 0.05 for both outcomes). Interestingly, serious intraocular inflammation occurred in 12% of those on the combination versus none of those on PDT alone, but it was concluded that this risk was outweighed by the improved efficacy. This is a contentious issue, and the cause for this phenomenon is not certain (Box 23.2).

Unfortunately, it is thus far impossible to titrate a treatment precisely enough to allow for recanalization of physiological choriocapillaris with restoration of physiological choroidal flow on the one hand, but avoid regrowth of the CNV.

Combination Therapy with Anti-inflammatory Agents

Corticosteroids may inhibit vascular proliferation indirectly by affecting the extravascular component through their anti-inflammatory properties. Triamcinolone has been shown to inhibit ocular neovascularization in a variety of models.^91–93 Corticosteroids decrease platelet-derived growth factor (PDGF)-induced VEGF secretion⁹⁴ and in models of oxidative stress,^94,95 and strongly inhibit VEGF production by macrophages.^96,97 Intravitreal injections of triamcinolone showed 100% reduction in CNV in laser induced models of CNV.⁹⁸ In contrast, nonselective cyclo-oxygenase (COX) inhibition, selective COX-1, or selective COX-2 inhibition did suppress laser induced CNV.⁹⁹ Interestingly, corneal cautery-induced neovascularization was inhibited by corticosteroids despite a decrease in VEGF upregulation by only 27%, which suggests that corticosteroids can modulate downstream signaling by VEGF, or antagonize other important proangiogenic pathways.¹⁰⁰ Many cytokine and cellular effects related to inflammation and the breakdown of the blood-ocular barrier are attenuated by corticosteroids.^101–107 For example, triamcinolone inhibits VEGF-induced breakdown of the blood-retina barrier,¹⁰⁸ decreases basic fibroblast growth factor (bFGF) proliferation of retinal vascular endothelial cells in a dose dependent fashion¹⁰⁹ and reduces bFGF induced migration and tube formation in choroidal vascular endothelial cells.¹¹⁰ While there was no benefit in terms of immediate occlusion of lesions in response to the combination of PDT and dexamethasone in a rodent model of laser-induced CNV, 6 days after treatment, only 10% of the lesions treated with monotherapy remained closed versus 31% of the lesions receiving combination therapy.¹¹¹ Interestingly, this effect was accompanied by less photoreceptor apoptosis (Box 23.3).

Stay updated, free articles. Join our Telegram channel

Join