Chapter 28 Mechanisms of Macular Edema and Therapeutic Approaches

Antonia M. Joussen, Thomas J. Wolfensberger

For additional online content visit http://www.expertconsult.com

Introduction

Macular edema is a common phenomenon in various diseases where fluid accumulates in between the retinal cells. The fluid originates from the intravascular compartment. The focal, diffuse, and cystic forms are all characterized by extracellular accumulation of fluid, specifically in Henle’s layer and the inner nuclear layer of the retina. The compartmentalization of the accumulated fluid is likely to be due in part to the relative barrier properties of the inner and outer plexiform layers.

The breakdown of the blood–retinal barrier (BRB), modulated via different growth factors, results from a disturbance of the integrity of the tight junctions.¹ Starling’s law predicts that macular edema will develop if the hydrostatic pressure gradient between capillary and retinal tissue is increased. That can occur, for example, in the presence of elevated blood pressure, or if the osmotic pressure gradient is decreased by excessive protein accumulation in the extracellular space within the retina.²

Macular edema as A result of various disease mechanisms

Causes of macular edema

The classic pattern of cystoid macular edema (CME) with the petaloid appearance originating from the fluorescein leakage from perifoveal capillaries may be seen in cases of advanced edema of various origins (Fig. 28.1, panel B online). This includes postsurgical CME as well as CME associated with one of the following conditions: diabetes, vascular occlusion, hypertensive retinopathy, epiretinal membranes, intraocular tumors (e.g., melanoma, choroidal hemangioma), intraocular inflammation (e.g., pars planitis), macroaneurysm, retinitis pigmentosa, choroidal neovascularization, and radiation retinopathy.

Fig. 28.1 Cystoid macular edema. (A) Histological section of the fovea area demonstrating cystoid macular edema with large cystic spaces in the outer nuclear and plexiform layer as well as in the inner nuclear layer. Panel B, online.

Fig. 28.1, online (B) Schematic drawing of the central retina with large cysts (top), finally resulting in a pseudohole formation (bottom).

Given the heterogeneous etiology of macular edema, its effective treatment depends upon a better understanding of its pathogenesis. In general, formation of macular edema is related to metabolic changes, ischemia, hydrostatic forces, inflammatory and toxic mechanisms, or mechanical forces that occur to various degrees in different conditions (Table 28.1).

Table 28.1 Causes of macular edema in relation to underlying disorders

Disease group	Disorder	Pathogenesis
Metabolic alterations	Diabetes	Abnormal glucose metabolism Aldose reductase
	Retinitis pigmentosa	CME: leakage at the level of RPE
	Inherited CME (autosomal-dominant)	Müller cell disease: leakage from perifoveolar capillaries
Ischemia	Vein occlusion Diabetic retinopathy	Inner BRB (retinal capillary hypoperfusion)
Ischemia	Severe hypertensive retinopathy HELLP syndrome Vasculitis, collagenosis	Outer BRB (ischemic hypoperfusion of the choroid: serous detachment)
Hydrostatic forces	Retinal vascular occlusions Venous occlusion Arterial hypertension Low IOP	Increased intravascular pressure Failure of the BRB
Mechanical forces	Vitreous traction on the macula	Epiretinal membranes with tangential traction Vitreomacular traction syndrome
Inflammation	Intermediate uveitis	Mediated by prostaglandins CME is considered an indication for treatment
	Postoperative CME	Perivascular leukocytic infiltrates
	DME	Diabetic leukostasis mediates vascular leakage by endothelial cell apoptosis
	Choroidal inflammatory diseases	Vogt–Koyanagi–Harada syndrome Birdshot retinochoroidopathy
Pharmacotoxic effects	e.g., Epinephrine (in aphakia) Betaxolol Latanoprost	Mostly via prostaglandins

CME, cystoid macular edema; RPE, retinal pigment epithelium; BRB, blood–retinal barrier; HELLP syndrome, hemolytic anemia, elevated liver enzymes, and low platelet count; IOP, intraocular pressure; DME, diabetic macular edema.

Metabolic alterations have a causal role in diabetic maculopathy, but also in inherited diseases such as the autosomal dominant form of macular edema or macular edema in retinitis pigmentosa. Furthermore, ischemia of the inner or outer BRB leads to formation of a macular edema. Decreased perfusion of the retinal capillaries is seen, e.g., in vein occlusion and diabetic retinopathy, whereas ischemia plus decreased perfusion of the choroid with associated serous retinal detachment occurs in severe hypertensive retinopathy, in eclampsia, or in rheumatoid disorders. Following retinal vascular occlusion the intravascular pressure increases and leads to dysfunction of the BRB. Similarly, hydrostatic forces are effective in arterial hypertension or in eyes with low intraocular pressure and may cause fluid accumulation in the macula.

Mechanical traction such as in epiretinal membranes or in vitreomacular traction syndrome promotes macular edema by physical forces.

Inflammation is important in the pathogenesis of macular edema in conditions such as intermediate uveitis, postoperative CME (Irvine–Gass syndrome), diabetic macular edema (DME) and various forms of choroidal inflammatory disease, including Vogt–Koyanagi–Harada syndrome and birdshot retinochoroidopathy. All prostaglandin-like pharmacological agents, even if applied topically, can induce macular edema via a cytokine response similar to inflammatory conditions.

The knowledge of the basic mechanisms involved in vascular leakage is essential for the development of an effective clinical treatment. Development of optimal strategies for treating retinal edema may depend on determining the ratio of the contribution of intra- and extracellular mechanisms to edema and measuring how this ratio changes between patients, between different retinopathies, and during disease progression.

With the growing understanding of the pathophysiology of the macular edema, the therapeutic thinking is likely to change from a merely symptomatic treatment (either surgical or medical) to a treatment that targets specifically the causal factors involved in its formation (e.g., cytokine or growth factor inhibition).

Molecular and cellular alterations leading to macular edema

Much of the knowledge on the pathophysiology of macular edema has been determined from extensive experimental studies on diabetic retinopathy and diabetic vascular leakage. A variety of techniques measuring accumulation of material from plasma in the neural retina have been investigated to assess permeability. Such accumulation seems diffuse in nature and focal defects have not been reproducibly described in diabetic mice; as well, interpretations of techniques involving tracer accumulation have not been validated in terms of “gold standard” permeability surface area product.³ Interestingly, edema has not been demonstrated in the retina of diabetic mice based on retinal thickness measurements despite the indication of increased permeability.

The BRB consists of the retinal pigment epithelium (RPE) layer (outer BRB), and the vascular endothelium (inner BRB), that prohibit the passage of macromolecules and circulating cells from the vascular compartment to the extracellular compartment and therefore intraretinal space.⁴ Intracellular edema (or cytotoxic edema) is defined as cellular swelling that occurs without opening of the BRB. Extracellular (or vasogenic) edema is characterized by retinal thickening in association with loss of BRB integrity (Fig. 28.2). While for diabetes and ischemic retinopathies the inner BRB was found to play a dominant role in vascular leakage, the importance of the outer BRB has recently been supported.⁵ The outer BRB separates the neural retina from the choroidal vasculature, which is responsible for approximately 80% of the blood supply in the eye. The outer BRB-specific leakage of fluorescent macromolecules can be visualized in diabetic and ischemic rodents and substantial leakage of macromolecules through the outer BRB can be detected.

Fig. 28.2 Blood–retinal barrier (BRB) and breakdown in vascular disease. (A) Normal retinal tissue: dotted blue line marks the outer BRB at the level of the retinal pigment epithelial cells. The inner BRB is at the level of the endothelial cells of retinal vessels (dashed red line). Further structural guidance is given by Müller cells. (B) Diabetic retinopathy: note the large caveolae of fluid and accumulation of blood cells in the retinal tissue, resulting from a breakdown of the inner BRB.

(Courtesy of Sarah Coupland, MD, Liverpool.)

The breakdown of the inner BRB may occur to a variable extent via dysfunction of intercellular junctions, increased transcellular transport, or increased endothelial cell destruction, and result in an increase in vascular permeability (Fig. 28.3B, online).

Fig. 28.3 Pathogenesis of macular edema and vascular leakage. (B) Vascular leakage is increased by a variety of factors, including growth factors and inflammatory cytokines. Leakage in turn is decreased by steroids, inhibition of tight junction formation, or antibodies to factors inducing leakage. ILM, inner limiting membrane; ELM, external limiting membrane; RPE, retinal pigment epithelium; VEGF, vascular endothelial growth factor; TGF-β, transforming growth factor-beta; TNF-α, tumor necrosis factor-alpha; IGF-1, insulin-like growth factor-1; TJAPs, tight junction proteins; NSAIDs, nonsteroidal anti-inflammatory drugs. Panel A, online.

Fig. 28.3, online (A) In general, water outflow through vessels is possible via three major routes: paracellular via dysfunction of tight junctions, transcellular via increased transport, e.g., mediated via growth factors, and finally directly via endothelial gaps after cell death.

The initial site of damage that results in the increased vascular permeability is controversial to date. Although impairment of the perivascular supporting cells such as pericytes and glial cells might play a role, endothelial cell dysfunction and injury seems more likely to be the first pathogenetic step towards the breakdown of the BRB early in the course of the disease. In order to dissect the molecular and pathophysiologic mechanisms that lead to the accumulation of fluid in the macular area, we have chosen DME as a model.

Cell-to-cell junctions and vascular permeability

Fluid homeostasis and endothelial permeability are mostly regulated by intercellular junctions in the nondiseased retina. Intercellular junctions are complex structures formed by the assembly of a transmembrane and cytoplasmic/cytoskeletal protein components. At least four different types of endothelial junctions have been described: tight junctions, gap junctions, adherence junctions, and syndesmos. Tight junctions are the most apical component of the intercellular cleft (Fig. 28.4, online).

Fig. 28.4, online Intercellular junctions in endothelial cells. Endothelial cells are connected and communicate with each other by tight junctions and adherens junctions. Tight junctions resemble a major part of the inner blood–retinal barrier. They are built by different proteins, including occludin, ZO-1, and the claudin family.

Although the molecular structure of tight junctions generally appears to be similar in all barrier systems, there are some differences between epithelial and endothelial tight junctions, and between tight junctions of peripheral and retinal endothelial cells.⁶ Expression of selected endothelial cell tight junction genes and particularly that of occludin and claudin-5 is reduced in the diabetic retina.⁷ In contrast to tight junctions in epithelial systems, structural and functional characteristics of tight junctions in endothelial cells respond promptly to ambient factors. It is likely that inflammatory agents increase permeability by binding to specific receptors that transduce intercellular signals, which in turn cause cytoskeletal reorganization and widening of the interendothelial clefts. For example, tumor necrosis factor-alpha (TNF-α) signals through protein kinase C (PKC)ζ/nuclear factor-kappa B (NF-κB) to alter the tight junction complex and increase retinal endothelial cell permeability.⁸ Endothelial junctions also regulate leukocyte extravasation. Once leukocytes have adhered to the endothelium, a coordinated opening of interendothelial cell junctions occurs.

Inflammation and vascular permeability

Leukocytic infiltration of the retinal tissue characterizes many inflammatory diseases such as diabetes, pars planitis, or choroidal inflammatory diseases.

In diabetes, activated leukocytes adhere to the retinal vascular endothelium.^9,¹⁰ Increased leukostasis is one of the first histologic changes in diabetic retinopathy and occurs prior to any apparent clinical pathology.

Adherent leukocytes play a crucial role in diabetic retinopathy by directly inducing endothelial cell death in capillaries,¹¹ causing vascular obstruction and vascular leakage. Endothelial cell death precedes the formation of acellular capillaries.¹⁰ With time, however, acellular capillaries prevail and become widespread. Although the mechanism of this destructive process remains elusive, it is clear that the interaction between the altered leukocytes and the endothelial cells and the subsequent endothelial damage represents a crucial pathogenic step^9,^11,¹² (Fig. 28.5).

Fig. 28.5 Inhibition of retinal pathology in long-term hyperhexosemic intercellular adhesion molecule-1 (ICAM-1) and CD18-deficient mice. Trypsin digests demonstrating a large destruction of the capillary network comparable to nonproliferative diabetic retinopathy with acellular capillaries and microaneurysms in 24-month hyperhexosemic mice. The capillary network in mice with a “noninflammatory” phenotype (CD-18- or ICAM-1-deficient) demonstrates almost normal capillaries.

(Reproduced with permission from Joussen AM, Poulaki V, Le ML, et al. A central role for inflammation in the pathogenesis of diabetic retinopathy. FASEB J 2004;18:1450–2.)

Inflammatory cytokines such as TNF-α decrease the protein and mRNA content of the tight junction proteins zonula occludens (ZO)-1 and claudin-5.⁸ TNF-α and interleukin-1 beta (IL-1β) are elevated in the vitreous of diabetic patients and in the retina of diabetic rats associated with increased retinal vascular permeability and leukostasis^13,¹⁴ (Fig. 28.6, online). Furthermore, TNF-α is involved in ischemic vascular changes.¹⁵

Fig. 28.6, online Inflammatory mediators are involved in leukocyte–endothelial interaction that, via reduction of tight junction protein expression and induction of apoptosis, results in vascular leakage. MCP-1, monocyte chemoattractant protein; Il-8, interleukin-8; MIB-1, macrophage-inhibitory factor; TNF-a, tumor necrosis factor-alpha; ZO-1, zona occludens-1.

Growth factors, vasoactive factors, and vascular permeability

The disruption of endothelial integrity leads to retinal ischemia and vascular endothelial growth factor (VEGF)-mediated iris and retinal neovascularization.^9,^16,¹⁷ VEGF is 50 000 times more potent than histamine in causing vascular permeability.¹⁸^–²⁰ Previous work has shown that retinal VEGF levels correlate with diabetic BRB breakdown in rodents²¹ and humans.²² Flt-1(1–3 Ig)Fc, a soluble VEGF receptor, reverses early diabetic BRB breakdown and diabetic leukostasis in a dose-dependent manner.¹⁷ Early BRB breakdown localizes, in part, to retinal venules and capillaries of the superficial inner retinal circulation²³ and can be sufficiently reduced by VEGF inhibition (Fig. 28.7, online). Although VEGF is only one of the cytokines involved in the pathogenesis of the vascular leakage, it is likely to be one of the most effective therapeutic targets.

Fig. 28.7, online Vascular endothelial growth factor (VEGF) is the key mediator of vascular damage and finally proliferation in the diabetic retina. AGE, advanced glycation end-products; IGF, insulin-like growth factor.

On a cellular level, VEGF has been implicated in many different mechanisms, which lead to macular edema. VEGF has, for example, been shown to decrease the proteins responsible for the tightness of the intercellular junctions and induces rapid phosphorylation of the tight junction proteins occludin and ZO-1, resulting in breakdown of the BRB.²⁴ VEGF-induced BRB breakdown appears to be effected via nitric oxide.¹⁷ VEGF also increases paracellular transport without altering the solvent drag reflection coefficient.²⁵ Furthermore, VEGF activation of PKC stimulates occludin phosphorylation and contributes to endothelial permeability.²⁶

There are tight connections between inflammation and VEGF expression.¹⁷ Recently, Müller-cell-derived VEGF was shown to be essential for diabetes-induced retinal inflammation and vascular leakage.²⁷

To determine the significance of Müller-cell-derived VEGF in diabetic retinopathy, VEGF expression in Müller cells was disrupted with an inducible Cre/lox system and diabetes-induced retinal inflammation and vascular leakage was examined in these conditional VEGF knockout (KO) mice. Diabetic conditional VEGF KO mice exhibited significantly reduced leukostasis, reduced expression of inflammatory biomarkers, depletion of tight junction proteins, reduced numbers of acellular capillaries, and reduced vascular leakage compared to diabetic control mice.

Investigations on cell–cell interactions by D’Amore and coworkers demonstrated an inhibitory effect of TGF-β secreted by pericytes on endothelial cell growth. In diabetic retinopathy formation of sorbitol via aldose reductase leads to PKC activation, resulting in a loss of the inhibitory balance ²⁸ (Fig. 28.8).

Fig. 28.8 Transforming growth factor-β (TGF-β) is involved in an inhibitory balance between pericytes and endothelial cells.

There are several other vasoactive factors and biochemical pathways affected by sustained hyperglycemia and known to be involved in DME, which are discussed in more detail in other chapters.

High glucose concentration leads to increased diacylglycerol (DAG) by two pathways: de novo synthesis and through dehydrogenation of phosphatidylcholine. Increased levels of DAG mediate PKC activation. Several studies have shown that a decrease in retinal blood flow occurs with PKC activation. Conversely, inhibition of PKC with LY333531 (Eli Lilly, Indianapolis, IN) normalized decreased retinal blood flow in diabetic rats.^29,³⁰

PKC activation causes vasoconstriction by increasing the expression of endothelins (ET), especially ET-1. The expression of ETs can be induced by a variety of growth factors and cytokines, including thrombin, TNF-α, TGF-β, insulin, and vasoactive substances including: angiotensin II, vasopressin, and bradykinin.

Furthermore, retinal vascular endothelial cells are very sensitive to histamine. Several studies have documented increased vascular histamine synthesis in diabetic rats and humans.³¹^–³³ The administration of histamine reduces ZO-1 protein expression and thus correlates with vascular permeability. The H1 receptor stimulates PKC that has been implicated in increased retinal vascular permeability.³⁴ Interestingly, Aiello and coworkers showed that administration of LY333531, a PKC-β isoform-selective inhibitor, does not significantly decrease histamine-induced permeability but instead VEGF-induced permeability. In contrast, administration of nonisoform-selective PKC inhibitors did significantly suppress histamine-induced permeability.³⁵

Furthermore, in vascular endothelial cells, advanced glycation end-products (AGE) may affect the gene expression of ET-1 and modify VEGF expression. The AGE-stimulated increased VEGF expression is dose- and time-dependent and additive to hypoxia.^36,³⁷

Endothelial cell death and vascular permeability

Where intraluminal pressure falls below a critical closing pressure the tone of the arteriolar wall cannot be maintained and the downstream capillary bed collapses and endothelial cells become “fibrin-locked.” Endothelial cells deprived of their circulation and nutrition die and only acellular basement membranes persists. A reduction in intraocular pressure may cause macular edema with cystoid degenerative changes and secondary atrophic alterations at the outer retina. Similarly, a reduction in retinal perfusion pressure, often linked to carotid/ophthalmic artery insufficiency, can have similar retinal manifestations and in extreme circumstances there may be retrograde filling of arteries from fellow veins. Stasis of the blood flow in capillaries after venous or arterial occlusions results in rapid apoptosis of endothelial cells.³⁸

Similarly, in diabetes, retinal barrier breakdown is at least in part due to endothelial cell damage and apoptosis. The proapoptotic molecule Fas-ligand (FasL) induces apoptosis in cells that carry its receptor Fas (CD 95).³⁹ There is evidence that FasL is expressed on vascular endothelium where it functions to inhibit leukocyte extravasation. The expression of FasL on vascular endothelial cells might thus prevent detrimental inflammation by inducing apoptosis in leukocytes as they attempt to enter the vessel. In fact, during inflammation and ensuing TNF-α release, the retinal endothelium upregulates several adhesion molecules⁴⁰ that mediate the adherence of the leukocytes, but also downregulates FasL thus allowing leukocyte survival and migration to active sites of inflammation. In experimental diabetic retinopathy, inhibition of Fas-mediated apoptotic cell death reduces vascular leakage.⁴¹ The cumulative endothelial cell death during the course of diabetes plays a causal role in the pathogenesis of the diabetic vascular leakage and maculopathy.

Extracellular matrix alterations and vascular permeability

Degradation of the extracellular matrix affects endothelial cell function at many levels causing endothelial cell lability which is required for cellular invasion and proliferation, or influencing the cellular resistance and therefore the vascular permeability. The degradation and modulation of the extracellular matrix are exerted by matrix metalloproteinases (MMPs), a family of zinc-binding, calcium-dependent enzymes.⁴² Elevation of MMP-9 and MMP-2 expression has been shown in diabetic neovascular membranes,^43,⁴⁴ although a direct effect of glucose on MMP-9 expression in vascular endothelial cells could not be shown.⁴⁵ It is probable that MMPs participate at various stages during the course of the BRB dysfunction and breakdown. Their actions include early changes of the endothelial cell resistance with influence on intercellular junction formation and function⁴⁶ to active participation in endothelial and pericyte cell death⁴⁷ that occurs late in the course of the disease.

Transcellular transport and vascular permeability

Disruption of the BRB is an early phenomenon in preclinical diabetic retinopathy. Two vascular permeability pathways may be affected, the paracellular pathway involving endothelial cell tight junctions, and the endothelial transcellular pathway mediated by endocytotic vesicles (caveolae). Despite the fact that pinocytic transport is critically involved in the transepithelial fluid exchange, its role in the pathogenesis of increased vascular leakage in diabetes is just emerging.^48,⁴⁹ The importance of the regulation of fluid homeostasis by active cellular transport of nutrients and fluid via pinocytosis is underlined by recent data suggesting a transient induction of the paracellular pathway and prolonged involvement of transcellular endothelial transport mechanisms in the increased permeability of retinal capillaries in diabetes.⁷

It is currently known that one of the factors involved in the regulation of pinocytic transport is VEGF. VEGF increases vascular permeability not only by disrupting the intercellular tight junctions between the retinal endothelial cells but also by inducing the formation of fenestrations and vesiculovacuolar organelles. The role of VEGF in the disruption of the pinocytic transport that is translated into increased vascular permeability in disease states is still controversial.⁵⁰ Whereas, in higly permeable blood vessels the number of pinocytotic vesicles at the endothelial luminal membrane transporting plasma immunoglobulin G is significantly increased, no fenestrations or vesicles were found in the endothelial cells of the VEGF-affected eyes when examined by electron microscopy.

Neuronal involvement in the formation of macular edema

Recent research on DME emphasizes the role of neuronal cells in the diabetic retinal damage. The retina consists of a network of neurons and glia (astrocytes, Müller cells, and microglial cells) that comprise approximately 95% of the tissue, with blood vessels representing less than 5% of the retinal mass.⁵¹ As the network of retinal neurons and glia is intimately linked, there is no doubt that the neural (photoreceptors, bipolar cells, horizontal cells, amacrine cells, and ganglion cells) and vascular components of the retina are closely associated by metabolic synergy and paracrine communication.^52,⁵³ Neuroglial cells are involved in vision, and blood vessels provide nutrients to facilitate the process.⁵⁴

The functional integration of blood vessels with the neurosensory retina is clinically evident during autoregulation in which retinal arterioles and venules constrict in response to hypertension and hyperoxia and dilate in response to hypercapnia. Likewise, disorders of the neurosensory retina and retinal vasculature are integrally linked, and understanding of the neurodegenerative component of retinal damage is of importance for treatments of macular edema.

In the inner retina, metabolic substrates, such as glucose, flow from vascular endothelium to astrocytes to neurons. In the outer retina, substrates reach Müller cells and photoreceptors from the choroid via the RPE.⁵² Microglia associate intimately with neurons that express molecules, such as CX3CL1 (fractalkine) and CD20, that negatively regulate microglial activation through their respective receptors. As such, perturbation of expression of ligand or receptor during stress would activate microglia to produce proinflammatory cytokines and acquire an activated morphology. Activated microglia produce chemokines such as monocyte chemoattractant protein-1, inducing expression of adhesion molecules, which can promote the leukostasis of neutrophils on endothelium, and potentially inducing the extravasation of inflammatory macrophages.^52,⁵⁴ Induction of glial fibrillary acidic protein (GFAP) is a marker of glial activation and increased expression of this protein occurs in Müller cells from the retinas of diabetic patients, but also after ischemic injury.

Mechanical factors involved in the formation of macular edema

Clinical and anatomic evidence indicates that abnormalities in the structure of the vitreoretinal interface may play an important role in the pathogenesis of DME.⁵⁵^–⁵⁷ It was suggested that vitreoretinal adhesions in diabetic eyes are stronger than the shear forces of traction from vitreous shrinkage and this in turn may lead to the development of vitreomacular traction and subsequently to macular edema.⁵⁸ Nevertheless, the risk of developing diffuse macular edema was 3.4-fold lower in the group of eyes with complete posterior vitreous attachment or complete vitreoretinal separation compared to the eyes with vitreomacular adhesion.⁵⁹

The vitreous humor is a gel-like structure composed mostly of water (99%), hyaluronic acid, and collagen. A structural barrier between the vitreous cavity and the retina is formed by the inner limiting membrane (ILM), which is localized between the innermost layer of the retina and the outer boundary of the vitreous. The ILM shows typical ultrastructural characteristics of a basal lamina, is found in close contact with the foot processes of Müller cells, and contains proteins that are typically found in basal laminae such as collagen type IV and laminin.⁶⁰ Striated collagen fibrils of the vitreous cortex insert into the inner portion of the ILM,⁶¹ which is also known as the hyaloid membrane of the vitreous. Detachment of the posterior hyaloid membrane with aging or pathology results in a condensation of the posterior vitreous surface (membrana hyaloidea posterior). In youth, there is adhesion between the vitreous cortex and the ILM that is stronger than Müller cells themselves and Müller cell foot processes become separated from their main cell body and remain connected to the posterior aspect of the ILM when this is separated from the retinal surface.⁶²

There has been a controversial discussion regarding the embryonic origin of the ILM, which can be demonstrated as early as 4 weeks after gestation in the human eye.^63,⁶⁴ Traditionally, the ILM has been considered to be synthesized by Müller cells. This concept has been challenged by data presented from Sarthy, who investigated the expression of collagen type IV during development of the mouse eye.⁶⁵ ILM proteins appear to originate largely from lens and ciliary body, although a contribution of retinal glial cells in ILM synthesis cannot be excluded. In support of this are data which show that also other ILM proteins such as perlecan, laminin-1, nidogen, and collagen XVIII are expressed predominantly in lens and ciliary body, but are not detected in the retina.⁶⁶

Diffuse DME has been found in association with an attached, thickened, and taut posterior hyaloid.⁶⁷ As immunocytochemical staining for cytokeratin (found in RPE) and GFAP protein (found in astrocytes and Müller cells) demonstrated the existence of cells in the premacular posterior hyaloids, suggesting a possible role for cell infiltration in the development or maintenance of macular edema. It remains to be elucidated whether these cells in the posterior vitreous cause macular edema physiologically rather than mechanically through the production of cytokines.

Treatment of macular edema

The current therapy for macular edema targets conditions where mechanical traction, hydrostatic force, or inflammation, plays a pathogenetic role. Laser coagulation, pharmacological approaches, and surgical measures are the most frequently used therapies.

Laser treatment

Many studies have demonstrated a beneficial effect of photocoagulation therapy for DME.⁶⁸^–⁷³ The exact mechanism of action of laser photocoagulation-induced resolution of DME is unknown. In short, a laser-induced destruction of oxygen-consuming photoreceptors has been discussed as well as cell death and scarring (involving gliosis and RPE hyperplasia) induced by the temporary rise in tissue temperature. Oxygen that normally diffuses from the choriocapillaris into the outer retina can now diffuse through the laser scar to the inner retina, thus relieving inner retinal hypoxia.^74,⁷⁵ There are contrasting data whether an increased preretinal oxygen partial pressure is involved and allows for microvascular repair in the treated areas.^76,⁷⁷

When studying the diameter of retinal arterioles, venules, and their macular branches before and after macular laser photocoagulation in eyes with DME, the macular arteriolar branches were found to be constricted by 20.2% and the venular branches 13.8%. This was attributed to an improved retinal oxygenation caused by the laser treatment leading to autoregulatory vasoconstriction, improving the DME.⁷⁸

According to another theory, the beneficial effect of laser photocoagulation is due to an enhanced proliferation of RPE and endothelial cells leading to repair and restoration of the BRB.⁷⁹ The RPE cells may respond to the injury in several ways: if the lesion is relatively small, the RPE defect can be filled by cell spreading; if the defect is relatively large, the cells can proliferate to resurface the area, and the RPE can produce cytokines (e.g., TGF-β) that antagonize the permeabilizing effects of VEGF.^80,⁸¹

Laser therapy is well established in diabetic retinopathy as well as in diseases with peripheral retinal ischemia. The Early Treatment Diabetic Retinopathy Study (ETDRS) was designed to evaluate the effects of argon laser photocoagulation for DME in a prospective, randomized, multicenter clinical trial. Among the subgroup of eyes with mild to moderate nonproliferative diabetic retinopathy with DME, visual acuity improved in 16%, remained unchanged in 77%, and worsened in 7% of treated eyes, whereas visual acuity improved in 11%, remained unchanged in 73%, and worsened in 16% of untreated eyes after 2 years of follow-up. After 3 years of follow-up, vision worsened in 12% of treated eyes compared to 24% of untreated eyes. There was a statistically significant benefit of laser photocoagulation in this group of eyes. Nevertheless, a current literature review indicates that, at least in selected groups, a beneficial effect of panretinal photocoagulation with respect to DME can be identified. The exact relationship between peripheral ischemia and DME and thus the relevance of peripheral panretinal photocoagulation in these patients as a treatment for DME still remains to be determined.

While focal laser coagulation reduces hypoxic areas and directly occludes leaky microaneurysms, the rationale for grid laser treatment in DME is not yet well established. Potentially, grid laser may have its beneficial effect by thinning the retina, bringing retinal vessels closer to choroidal vessels, permitting the retinal vessels to constrict by autoregulation, thereby decreasing retinal blood flow and consequently decreasing edema formation.⁸²

Potential side-effects are choroidal neovascularization induction (Fig. 28.9) or lack of effect in cases of diffuse edema (Fig. 28.10).