Chapter 57 – Corneal Endothelium
MARK L. MCDERMOTT
HARVINDER K. S. ATLURI
• Hexagonal nonreplicating monolayer of neural crest–derived tissue that regulates the hydration state of corneal stroma.
• A tissue containing large quantities of membrane-bound Na+ , K+ -ATPase with specialized intercellular junctions that establish a pump–leak process in the maintenance of corneal deturgescence.
• A delicate tissue subject to alteration from age, trauma, systemic or ocular disease, contact lens wear, surgery, intraocular solutions, and unique dystrophic conditions. This chapter discusses normal corneal endothelial anatomy and physiology, the effects of extrinsic and intrinsic stressors on edothelial structure and function, and the endothelial dystrophies.
• Endothelial abnormalities: Fuchs’ dystrophy, congenital hereditary endothelial dystrophy, posterior polymorphous dystrophy, and endothelial trauma are discussed.
The effects of external trauma and of ophthalmic and systemic disease on human corneal endothelium are best understood by reviewing the anatomy and physiology of the adult human endothelium.
In early embryogenesis, the posterior cornea is lined with a neural crest–derived monolayer of orderly arranged cuboidal cells. By the 78?mm stage, the cells become flattened and abut one another. At this stage, immediately anterior to the flattened layer is a discontinuous, homogeneous acellular layer that in time becomes Descemet’s membrane. By the 120?mm and 165?mm stages of development, the endothelial monolayer is uniform in thickness, spans the entire posterior corneal surface, and fuses with the cells of the trabecular meshwork. Similarly, Descemet’s membrane becomes continuous and uniform and fuses with the trabecular beams. The fusion site, known as Schwalbe’s line, is a gonioscopic landmark that defines the end of Descemet’s membrane and the start of the trabecular meshwork. At birth, the endothelium is approximately 10?µm thick.
Morphology and Development
The intact human endothelium is a monolayer that appears as a honeycomb-like mosaic when viewed from the aqueous side ( Fig. 57-1 ). The individual cells flatten as the person ages; they stabilize at about 4?µm in height in adulthood ( Fig. 57-2 ). The surface of the endothelium on the aqueous side is devoid of surface
Figure 57-1 Specular photomicrograph of normal endothelium. Note the dark, well-defined cell borders, the regular hexagonal array, and the uniform cell size. (Bar = 50?µm.)
Figure 57-2 Light micrograph of normal endothelium (×100). Note the single-cell endothelial layer with a Descemet’s membrane of uniform thickness (epithelial surface at top of figure). (Courtesy of Dr. David Barsky.)
villi, except in certain pathological conditions, when it may develop epithelial characteristics. Adjacent cells share extensive lateral interdigitations and possess gap and tight junctions spread along their lateral cell borders. The lateral membranes contain a high density of Na+ , K+ -adenosinetriphosphatase (ATPase) pump sites. The basal side of the endothelial membrane contains numerous hemidesmosomes that promote adhesion to Descemet’s membrane. Endothelial cells contain numerous mitochondria and a prominent Golgi apparatus ( Fig. 57-3 ).
Endothelium continuously secretes Descemet’s membrane throughout life, beginning in utero at the 8-week stage. The anterior portion of Descemet’s membrane formed in utero has a distinctive banded appearance when viewed by electron microscopy and is approximately 3?µm thick. Descemet’s membrane thickens with age, reaching up to 10?µm, but any Descemet’s membrane produced after birth is not banded and has an amorphous texture when viewed by electron microscopy.
Figure 57-3 Scanning electron micrograph of normal endothelium (×500). Note the hexagonal cellular array and surface topography.
Throughout life, endothelial cell density and topography change. From the second to eighth decades, endothelial cell density declines from approximately 3000–4000 cells/mm2 to around 2600 cells/mm2 , and the percentage of hexagonal cells declines from about 75% to around 60%. 
The anatomy of the endothelium is ideally suited to its primary physiological role of fluid regulation. Thermodynamically, the hexagonal arrangement of individual cells is the most favorable one for cells to cover a surface without gaps, thereby facilitating barrier function.    This geometrical arrangement also minimizes individual circumferential cell area, thereby allowing a maximal number of cells per unit area and maximizing pump site density.
The primary site of fluid regulation through the endothelium’s activity is the corneal stroma. As a result of this endothelial activity, the stroma is maintained in a relatively deturgesced state (78% water content), which allows an orderly lattice of collagen fibrils to enmesh in glycosaminoglycans and create a transparent tissue. One hypothesis is that this endothelial activity is mediated by a pump–leak process; net fluid egress from the corneal stroma follows movement down an osmotic gradient from a relatively hypo-osmotic stroma toward a relatively hypertonic aqueous humor. This bulk fluid movement requires no energy. The energy-requiring processes are the intracellular and membrane-bound ion transport systems, which generate the osmotic gradient. The two most important ion transport systems are the membrane-bound Na+ , K+ -ATPase sites and the intracellular carbonic anhydrase pathway. Activity in both these pathways produces a net flux from stroma to aqueous humor.
The barrier portion of the endothelium is unique, in that it is permeable to some degree, permitting the ion flux necessary to establish the osmotic gradient.  This permeability can be modulated, depending on ambient calcium ion. Exposure of endothelium to calcium-free media results in large reductions in barrier function.
Assessment of Endothelial Function
In vivo assessment of endothelial function relies on measurements of corneal thickness or observation of the endothelial monolayer using specular microscopy. Measurement of the corneal thickness (pachymetry) indirectly reflects endothelial function, because corneal thickness reflects the state of corneal deturgescence. The average central corneal thickness is around 0.5?mm, which gradually increases toward the periphery to around 0.7?mm. Pachymetry may be carried out ultrasonically or optically. Both methods are reproducible, but they are not always directly comparable. Moreover, pachymetry as an indicator of corneal function is time dependent. A normal patient has slightly thicker corneas immediately upon awaking than he or she does later in the day. This increased thickness is the result of the loss of the open-eye desiccating effect on the cornea, as well as reduced metabolic activity of the endothelium under nocturnal lid closure. Such nocturnal swelling is exaggerated in dysfunctional endothelium and is often noticed by patients as a morning blur in vision that improves during the day.
Direct observation using specular microscopy also is used to evaluate endothelial status. By using a wide-field specular microscope, a photomicrograph can be obtained and subsequently digitalized and analyzed. In cases in which corneal edema may not allow adequate visualization of the endothelium by specular microscopy, confocal microscopy may be of value. Confocal microscopy allows real-time in vivo assessment of different layers of the cornea. Although not precluded by corneal edema, it is technically more difficult to perform than specular microscopy.  Analysis of images from these techniques provides data that reflect endothelial cell density and morphology. At birth, the normal cornea is lined by about 3500–4000 cells/mm2 . Central endothelial cell density decreases at an average rate of 0.6% per year in normal corneas. Little, if any, mitotic potential exists within the endothelium. The exact number of cells/mm2 required to maintain corneal deturgescence is not known, but corneas with cell counts below 1000 cells/mm2 in multiple areas may be at risk for the later development of corneal edema. This prediction is difficult to verify, because nonuniform cell loss results in regional variation in cell density, which leads to sampling errors during specular microscopy. Besides actual cell density, endothelial cell morphology (size and shape) can affect function. Endothelial monolayers with increased variation in cell size (polymegathism) and increased variation in cell shape (pleomorphism) are less effective in achieving deturgescence in a cornea swollen by hypoxic stress than in a cornea with normal morphology. 
Endothelial Function Research Techniques
In vivo assessment of global endothelial function can be carried out by examining the curve obtained by plotting corneal thickness over time after induction of corneal edema by hypoxia. This technique permits the comparison of eyes with morphometric or cell density alterations to the eyes of age-matched controls. Barrier function can be evaluated using fluorophotometry.  In this technique, the change in concentration of the fluorescein molecule in the corneal and aqueous compartments is determined over time and used to calculate transfer coefficients that reflect the relative permeability of the barrier between the two compartments. This technique generally is used in research studies rather than clinical practice because of the duration of the experiment and the requirement for multiple fluorescence measurements.
Endothelial Responses to Stress
The endothelium has a rather restricted response to stress. Mild stress may result in morphometric changes, and greater stresses may result in cell loss as well as morphometric changes. Morphometric changes are believed to result from alterations in the endothelial cytoskeleton. Sources of stress may be metabolic (from hypoxia or hyperglycemia) or toxic (from drugs or preservatives); stress may also be caused by alterations in pH, ionization, or osmolarity or by trauma from surgery.
A common hypoxic stress is that generated by contact lens wear. All contact lenses present a hypoxic stress to the endothelium to varying degrees.  Over time, this results in alteration of the morphometry of the endothelium. With contact lens wear, the coefficient of variation for cell size increases, and the percentage of hexagons decreases. The severity of this change is related
Figure 57-4 Age-matched comparison of diabetic and normal endothelia. Note the reduction in the percentage of hexagons and the increase in the coefficient of variation of cell size in both severe and moderate diabetics compared with the age-matched controls. (Courtesy of Dr. Henry Edelhauser.)
to contact lens composition, duration of use, and type of lens wear. The greatest changes in morphometry are seen in patients who have worn polymethyl methacrylate lenses for multiple decades.  Less severe changes are seen in other forms and types of contact lens use. Although earlier studies found no difference in endothelial cell density among contact lens wearers, recent studies demonstrate a significant reduction in endothelial cell density with increased duration of contact lens wear. This decrease in cell count is preceded by pleomorphism and polymegathism and may be a late indicator of altered corneal endothelium in long-term contact lens users.  The morphometric changes induced by contact lens wear do not regress upon cessation of wear. A physiological correlate to the altered morphometry induced by contact lens wear is the in vivo observation that the corneal deswelling (deturgescence) response after induction of edema by hypoxia is significantly less in long-term contact lens wearers compared with age-matched controls.
Another metabolic stress is hyperglycemia. The corneal endothelium in type I and type II diabetics, when compared with that of age-matched controls, has morphometric and morphological alterations manifested as lower mean cell density and greater pleomorphism and polymegathism. These changes are more pronounced in type I diabetics. There are conflicting results regarding functional abnormalities associated with these morphological alterations, but it is recognized that the diabetic cornea has a higher susceptibility to mechanical or surgical stress or additional metabolic stress ( Fig. 57-4 ).
Besides adverse systemic conditions, the endothelium can be damaged by the introduction of intraocular agents during surgery. The potential for damage to the endothelium and other tissues is related to chemical composition, pH, and osmolality. 
Intraocular surgery commonly results in endothelial damage. During cataract surgery, contact of the endothelium with jets of irrigation fluid, nucleus particles, instruments, or intraocular lenses causes focal endothelial cell death. Repair of these areas takes place via the process of cell slide from adjacent undamaged areas. A permanent reduction in cell density occurs, and individual cell size is increased. Often the mosaic suggests the area of injury by elongation of cells or the production of giant cells with 10 or more sides. Ophthalmic viscoelastic devices (composed of hydroxypropyl methylcellulose, chondroitin sulfate, or sodium hyaluronate) are now widely used in cataract surgery and provide significant protection against trauma to the endothelium. 
Glaucoma has also been associated with endothelial cell loss. A recent study found significantly lower endothelial cell counts in patients who have glaucoma and ocular hypertension compared with age-matched controls. Cell counts were inversely proportional to the mean intraocular pressure in the glaucoma and ocular hypertension groups. Mechanisms of cell loss may include direct damage from intraocular pressure, congenital alterations of endothelium in glaucoma, and medication toxicity.