Ординатура / Офтальмология / Английские материалы / Retinal and Choroidal Angiogenesis_Penn_2008

There are many similarities between CNV and a nonspecific wound healing process, including blood clotting, deposition of provisional fibrin matrix, an inflammatory response involving neutrophils and macrophages, angiogenesis, formation of extracellular matrix, scarring, and eventually, reepithelialization.7

Some molecules are present at equivalent stages, e.g., vascular endothelial growth factor (VEGF A), fibroblast growth factor (FGF-2), insulin like growth factor (IGF-1), transforming growth factor (TGF-ß), monocyte chemoattractant protein (MCP-1), and connective tissue growth factor (CTGF). A fibrin front seems to act as a scaffold for sprouting capillaries, but eventually the vascular front is impeded by epithelial involution. All these physiological and pathological relationships in AMD associated with CNV are possible future targets for therapeutic intervention.

1.1Angiogenesis

In order to better understand CNV in AMD, it should be viewed as an example of angiogenesis. New vessel formation has previously been attributed to either vasculogenesis, which is new blood vessel formation from angioblastic precursor cells in the developing embryo, or angiogenesis, in which new vessels form from pre-existing vessels in the adult. Contrary to these definitions, recent experimental CNV papers on mice have reported a substantial contribution of angioblastic precursor cells (up to 25%) in adult angiogenesis.8-14 In this context, the resting pericyte/smooth muscle cell (smc)-coated endothelial cells constitutively express tissue plasminogen activator (t-PA), matrix metalloproteinase-2 (MMP-2), and angiopoietin/Tie 2 balanced by plasminogen activator inhibitor-1 (PAI-1), tissue inhibitor of MMP-2 (TIMP-2) and TGF-β. When the endothelial cells are stimulated at the vascular “front” line, several proand anti-angiogenic factors are upregulated. The stabilizing pericyte/smc coating is shed, transforming the cells to an immature state more sensitive to extrinsic signals (Figure 2).

Figure 24-3. Proteolysis. Three proteolytic systems, the coagulation system, the plasminogen/plasmin system, and the matrix metalloproteinase system, interact in the angiogenic response by proteolytic degradation of fibrin and extracellular matrix (ECM) during transient provisional matrix formation balanced by release of cryptic endogenous angiogenesis inhibitors, e.g., angiostatin and endostatin, as well as promoters, e.g., vascular endothelial growth factor (VEGF).

There are many stages in the process of molecular angiogenesis including an angiogenic shift, stimulation of angiogenic growth factor receptors on endothelial cells, proteolysis of the basal membrane, proliferation and migration of endothelial cells, invasion and proteolysis of ECM with provisional matrix formation, and stabilization and eventual survival of newly formed vessels, including recruitment of pericytes and smooth muscle cells and closing of arteriovenous loops (Figure 4). Since all these factors work in consort during the complicated process of CNV, it seems reasonable to assume that therapeutically, several factors will have to be addressed simultaneously.

Figure 24-4. Molecular angiogenesis. (1) An angiogenic switch is induced by a focal molecular imbalance stimulus, e.g., hypoxia. (2) In response to the molecular switch, angiogenic growth factor receptors e.g., vascular endothelial growth factor (VEGF) receptors on endothelial cells (EC) in the vicinity are stimulated. (3) The supporting basal membrane of the endothelial cells is digested by proteolytic enzymes, e.g., matrix metalloproteinases (MMP). (4) Endothelial cells (EC) proliferate and migrate toward the angiogenic stimulus. (5) The extracellular matrix (ECM) is rearranged by proteolytic enzymes, e.g., MMPs, to facilitate endothelial cell invasion toward the target. (6) For survival, the newly formed EC vessel is stabilized by several molecules. (See also Figure 7.) (7) For further maturation and stabilization, pericytes and smooth muscle cells are recruited to surround the immature vessel. (See also Figure 7.) (8) By a “lock and key” molecular mechanism, arteries and veins form vascular loops. (See also Figure 7.)

1.2VEGF

Recently much focus in AMD therapy has been given to VEGF, the main molecule regulating neovascular growth in the body. VEGF is a member of the platelet-derived growth factor (PDGF) superfamily, which shares a homologous CUB domain that is a target for proteolytic cleavage and activation. The VEGF family has six members that interact with four receptors (Figure 5).16 VEGF is an equally important survival factor for

neuronal and endothelial cells, and the vascular and neuronal systems are often parallel in the body.18,19 The main VEGF receptor is VEGFR2, and the

main VEGF is VEGF-A. VEGF-A is hypoxia-inducible and stimulates proliferation and migration as well as tube formation in endothelial cells. It increases vascular permeability and activates MMP and TF.

VEGF-A has six isoforms with varying numbers of amino acids, ranging from the soluble “mobile” 121-residue isoform through the normal 165residue (164 in mice) isoform to the insoluble “sticky” 206-residue isoform.18,19 At the vascular front, soluble VEGF-A causes diffuse growth, lumen enlargement and reduced branching, whereas insoluble VEGF-A effects local growth, excessive branching and reduced lumen, and normal VEGF-A induces cued growth direction in a gradient fashion resulting in vessel elongation.18 Growing neurons and vessels have “receptor sensors” in digital protrusions. Vascular sprouts can sense VEGF from VEGF receptors in “tip cells.”20 Potential sources of VEGF are hypoxic cells, blood cells, RPE and cryptic VEGF in the ECM.

Figure 24-5. The vascular endothelial growth factor (VEGF) family and its receptors. (Adapted from Matsumoto T., Claesson Welsh L. VEGF receptor signal transduction. SCI STKE 2001; (112)RE21.)

VEGF has angiogenic functions in physiological processes such as reproduction and epiphyseal growth. It has pathological angiogenic properties in trauma, inflammation, tumor growth, and bone marrow recruitment. Examples of VEGF inhibitors are angiostatin, endostatin, glucocorticoids, PEDF, and synthetic antagonists. Several synthetic antiVEGF antibodies (IgG or fragments thereof) have recently proven to be efficacious in clinical trials, targeting some or all VEGF isoforms in AMD associated with CNV.21

1.3MMPs

The ECM surrounding the CNV in AMD and its remodeling during CNV growth is probably very important to the angiogenic process. Matrix metalloproteinases (MMP) are some of the key enzymes regulating ECM structure. MMPs belong to a large family of over 20 members.22 The gelatinases are either secreted (MMP-2 or MMP-9) or membrane bound (membrane-type (MT)-1 MMP). When secreted as inactive proenzymes (zymogens), they are proteolytically activated in the ECM or on the plasma membrane. MMP-2 and MMP-9 preferentially break down collagen IV in basal membranes and are activated in macrophages and RPE. During invasion and proteolytic degradation of ECM in the angiogenic transition, type I collagen in the stroma and fibrin from vascular leakage activate latent MMP-2 and upregulate MT-1 MMP. Fibrin is efficiently degraded by MT-1 MMP.

Inhibitors of secreted MMPs are tissue inhibitor of matrix metalloproteinases (TIMP), angiostatin, and endostatin. TIMP-3 is produced by RPE and accumulates in Bruch’s membrane and drusen over time. It specifically inhibits VEGF chemotaxis and endothelial cell motility. TIMP-3 also inhibits collagen gel invasion but is downregulated with age and blue light exposure. Clinical trials solely targeting MMP in angiogenesis have not proven effective, but given its synergistic role with VEGF, it is reasonable to assume its potential in a future therapeutic drug combination targeting CNV in AMD.23

1.4PEDF

In several but not all experimental and clinical studies on angiogenesis, pigment epithelial derived growth factor (PEDF) has been seen as the counterbalance to VEGF. There are some indications for an altered balance between VEGF and PEDF in the angiogenic switch. PEDF belongs to the serine protease inhibitor (serpin) family in the plasminogen system, but it lacks the protease inhibitor activity normally seen in serpins. It is found in

high concentration in the vitreous, lens, and cornea, and could possibly explain the relative avascularity of these tissues.

PEDF is produced by the non-pigmented ciliary epithelium and RPE at the apical side toward the interphotoreceptor matrix by a factor of 4:1 compared to the basal side.24 The normal distribution of PEDF across RPE is thus opposite that of VEGF (Figure 1).

Similar to VEGF, PEDF is neuroprotective, but it induces apoptosis (cell death) in immature endothelial cells. It does not affect mature endothelial cells. It is also a strong tumor suppressor, but it is downregulated with age, disease and blue light exposure.25-27 PEDF has a dual role of interacting simultaneously with negatively charged binding sites on proteoglycans and heparin on the cell surface and with positively charged binding sites on collagen I in ECM. The binding is pH-dependent. That fact could be important for fortifying the blood-retina barrier.25,28

Given the importance of upholding the integrity of the blood-retina barrier to prevent choroidal neovascular ingrowth in AMD, PEDF holds much future promise in AMD therapy.29

1.5Homing of blood cells in angiogenesis

Similar to other neovascular processes in the body, recruitment of bone marrow-derived cells in CNV angiogenesis is probably signaled via placenta-derived growth factor (PLGF, a VEGF family member) through VEGFR1 stimulation.8-11 Hematopoietic stem cells (HSC) are also major angiogenic contributors.12-14 Adhesion molecules, e.g., intercellular adhesion molecule-1 (ICAM-1), help HSCs stick to the walls of vessels stimulated by VEGF and monocyte chemoattractant protein-1 (MCP-1) but inhibited by angiopoietin 1 (Figure 6). In AMD, antigen-presenting dendritic cells attach to drusen. Vascular progenitor cells (VPCs) are involved in angiogenesis and vascular stabilization. Depending on molecular input, the fate of the VPCs shifts the balance between maturation and immaturity (Figure 7). Targeting several of these factors in experimental CNV in mice has proven effective in reducing the CNV and could lead to future therapeutic modalities in AMD with CNV.

Figure 24-6. Main events leading to the invasion of a target organ by cells from the circulation. Recruitment of bone marrow derived cells in angiogenesis is probably signaled via placenta-derived growth factor (PLGF, a VEGF family member) through VEGFR1 stimulation. Hematopoietic stem cells (HSC) are also major angiogenic contributors. (1) Rolling. The circulating cell is slowed down by rolling in the bloodstream. (2) Cell adhesion. Adhesion molecules, e.g., ICAM-1, help HSCs stick to the walls of vessels stimulated by VEGF and MCP-1, but are inhibited by angiopoietin 1. (3) Transmigration. The cells are stimulated to extravasate through the vessel wall into the extracellular matrix. (4) Chemotaxis. The cells are stimulated to migrate towards the goal through chemotactic stimulation. (5) Activation. The cells produce cytokines in response to local stimulus. (6) Protein release. At the target site the cells release their protein load as programmed. (Reprinted with kind permission from Montan, P. (2000): Immunological and inflammatory mechanisms in ocular allergy with special reference to vernal keratoconjunctivitis. Clinical and experimental studies. Thesis, KI, Stockholm.)

2.MOLECULAR IMPLICATIONS IN CLINICAL AND EXPERIMENTAL CNV

Recent progress in CNV treatment has allowed researchers to slow the process or reduce the extent of CNV. In AMD with CNV,30 VEGF congregates around the vessels,31 MMP-2 is expressed mainly in the membrane core,32 and MMP-9 localizes toward the surface close to the RPE region.32 In experimental CNV, growth is typically induced within 4 to 10 days by a laser-induced break in the Bruch’s membrane and the RPE33 or subretinal injection of growth factors. Alternative methods are subretinal

expression of growth factors in beads34 or delivery by adenovirus vectors targeting the RPE.35 Age seems to be crucial, with 64% less CNV in 2- month-old than 16-month-old mice.36 Macrophage depletion also reduced CNV by 88%.37 Experimental CNV can be reduced by 75% by VEGF inhibition,38 60% by VEGFR1 inhibition,39 and 70% by PLGF inhibition.39

MMPs are also expressed in experimental CNV.40 CNV is impeded 30% by MMP-2 inhibition,41 20% by MMP-9 inhibition,42 and 56% by combined MMP-2/MMP-9 inhibition.43 Overexpression of TIMP-3 reduces experimental CNV by 50%.44 Inhibition of the plasminogen system reduces CNV by 50%.39 Inhibition of the coagulation system by blocking TF through an antibody and a cytolytic immune attack reduced CNV by 90%.45 CNV was reduced 53% by endostatin,46 30% by angiostatin,47 and 50% by PEDF.48 The unique feature of regression of an already established CNV has been noted for PEDF,49 combrestatin, and angiopoietin-2.50 All these findings indicate that future treatment of CNV in AMD will include drugs that are able to impede the progression of CNV.

Figure 24-7. Vascular stability. Depending on molecular input, the fate of vascular progenitor cells (VPCs) shifts the balance between maturation and immaturity. PDGF-BB, angiopoetin-1 (Ang-1)/Tie-2 receptor, and TGF-β favor a pericyte fate, and VEGF and angiopoietin-2 (Ang- 2)/Tie-2 receptor favor an endothelial fate with consequent results. (Yamashita J., Itoh, H. et al. Flk-1 positive cells derived from embryonic stem cells serve as vascular progenitors. Nature 2000; 408(6808):92-96.)

3.CNV TREATMENT

Clinical results of CNV treatment in AMD with photocoagulation,51 photodynamic therapy (PDT),52 transpupillary thermotherapy (TTT),53 radiation,54 or surgery55,56 have previously been rather dismal, with limited visual benefit as well as high recurrence rates. AMD prophylaxis with antioxidants could reduce the risk of CNV by 25% within five years in AMD high-risk groups.57 Several prospective randomized clinical trials have been performed. They normally combined PDT with a molecular intervention such as steroids (Fluocinolone acetonide)58 or steroid analogs (anecortave acetate),59 VEGF inhibition (Rhu-FabV260 or VEGF aptamer61), or PEDF gene delivery (adenovirus vector).62 These treatment modalities interfere with both VEGF and MMP systems. Recently, very promising results have been reported in anti-VEGF therapy of CNV in AMD in clinical trials using IgG or Fab-fragments targeting human VEGF.21 Presently, they require frequent intravitreal re-injections every 1-2 months for a prolonged period. Other treatment modalities, including nanoparticles giving sustained release for a prolonged time period, are now being investigated. If successful, these therapies hold promise for many other types of intraocular neovascularization, e.g., intraocular tumors, inflammation, and diabetes.

4.FUTURE PROSPECTS

The “golden bullet” theory whereby one treatment modality takes care of the problem of CNV in AMD is probably not realistic.63 CNV is a multifactor problem involving a plethora of cross-talking genes expressed differentially in both space and time. A combination treatment approach including prophylaxis would be more beneficial.

Protein blockage through mRNA interference seems to be a promising modality, as well as a lifelong combination pharmacological treatment, e.g., VEGF antagonists, anecortave acetate, and PEDF. Treatment targeting leakage of choroidal neovascular vessels to the subretinal space, causing separation of the photoreceptors from their RPE counterpart, which is the major cause of visual impairment, might be most effective.62 The combination of HSC delivery and a local cytotoxic induction by systemic drug delivery could also hold future promise. Stem cell therapy of experimental brain tumors is already being investigated.64 Another promising treatment involves conjugating an antibody against proliferating endothelial cells to a CNV drug. Initial studies showed that this method allows the targeting of local areas of CNV.65 It is likely that we will see