Ординатура / Офтальмология / Английские материалы / Slatter's Fundemental of Vetrinary Ophthalmology 4th edition_Maggs, Miller, Ofri_2008

immune system to recognize and destroy a variety of infectious and noninfectious “foreign” agents.

The major mechanical and chemical events of inflammation do not differ among the various mammalian species. Considered here are those aspects of inflammation that have particularly important consequences for the eye or that are somehow modified by peculiarities of ocular anatomy or physiology.

When it works as it should, the inflammatory reaction is a beneficial physiologic reaction that is limited only to the immediate area of injury. It should persist only as long as is necessary for its defensive and débridement activities, and it should selectively recruit only those body defenses most effective in combating the specific injurious agent. Inflammation should thus exhibit a remarkable degree of moderation and specificity so that there is no undue injury to bystander tissues or to the overall health of the animal. This remarkable balancing act is achieved by the local release of a wide range of chemical mediators that reside within normal parenchymal tissue, within leukocytes or platelets, and within the plasma itself. It is not practical to attempt to list all of the inflammatory mediators, their origins, and their physiologic activity. Not only does this list grow almost daily, but such lists tend to reinforce the mistaken view that a given mediator has a specific, invariable biologic activity.

Like letters of the alphabet, each inflammatory mediator is a hormonelike member of a complex biologic alphabet. Each “letter” may thus have many different meanings, depending on the company it keeps and in what sequence the letters occur. These inflammatory mediators are part of a larger group of locally acting hormonelike messengers called cytokines, so named because they stimulate some kind of proliferative activity on the part of neighboring cells. These cytokines create the biologic language that carries the instructions for everything from coordinated embryologic development to orderly cell death (apoptosis). One must thus read the supposed activity of any given cytokine with substantial skepticism, because there is no guarantee that the activity that we have determined by in vitro testing of isolated mediators at arbitrary dosages has anything to do with their in vivo activity at physiologic dosages and in the company of many other members of the cytokine alphabet.

Although our understanding of the chemical mediation of inflammation and repair is still quite primitive and is largely limited to making lists of the chemicals involved, we know quite a lot about the mechanical events of inflammation. These events represent a stepwise, highly integrated chronologic sequence that involves changes in microvascular blood flow, endothelial permeability, leukocyte migration, humoral and leukocyte-dependent neutralization of foreign material, and tissue débridement preparatory to parenchymal and stromal repair.

The initial events of inflammation involve microvascular dilation (hyperemia) and endothelial cell contraction to increase the permeability of postcapillary venules to plasma solutes. This creates the redness and serous effusion that typify early inflammation, and that will continue as long as the active phase of inflammation persists. These early events are stereotypic: They are identical regardless of the stimulus, and they have no diagnostic specificity in terms of predicting what type of injury might have occurred. They occur in response to the rapid release of preformed chemical mediators such as histamine from mast cells or platelets within the region of initial injury.

These short-acting vasoactive mediators are then reinforced by a wide variety of mediators that are synthesized de novo from parenchymal cells, leukocytes, and other cells at sites of inflammation.

This serous effusion that characterizes the very early stages of inflammation is beneficial, because it serves to flood the region with a substantial array of such activated humoral defenses as complement and some broadly acting antibodies as well as to engender the leakage of fibrinogen, which will then create the fibrin scaffold that will enhance both the migration and the phagocytic efficacy of the leukocytes that follow.

Within minutes of the initiation of the preceding changes in vascular permeability, leukocytes begin to settle out of circulation, bind to endothelial cells, migrate through the nowpermeable endothelial junctions, and then move through the tissues in search of the cause of tissue injury. We have recognized for many years that the types of leukocytes recruited carry substantial diagnostic value, because certain types of infectious agents or immune responses habitually recruit specific types of leukocytes. As a rule neutrophil-dominated inflammation is equated with bacterial infection, eosinophils predict hypersensitivity reactions, especially to parasites, and macrophagedominated (granulomatous) inflammation is restricted to cell-mediated immune events and to inflammation initiated by a relatively small group of poorly degradable infectious or noninfectious agents. Only recently, however, has the basis for that sometimes remarkable specificity become apparent.

Part of the answer lies in the mixture of mediators/cytokines that are triggered by certain types of infections, but there was always an inexplicable problem: Most described chemotactic mediators have a fairly broad range of activity, so that many of them attract neutrophils, eosinophils, and macrophages with almost equal avidity. How, then, could we explain the empiric observation of almost purely eosinophilic infiltrates in some parasitic diseases, or purely neutrophilic infiltrates in many bacterial infections? The answer seems to lie in a second level of leukocyte “screening” that occurs at the level of the endothelial cells themselves. The locally generated inflammatory mediators rapidly stimulate the expression of leukocyte-specific adhesion molecules on the surface of the venular endothelium, which will bind only to complementary receptors on the surface of stimulated leukocytes that have fallen under the influence of these same local mediators. Thus, although the general chemotactic stimulus may attract many different leukocyte types into the permeable venules, only those leukocytes that can pass this second screening test will actually be allowed to enter the tissue to take up the “search-and-destroy” mission.

The leukocytes that thus enter the tissue become intermingled with the serum and fibrin that may already have accumulated because of the previous increase in vascular permeability, and these mixtures form the inflammatory exudates that for years have formed the basis for the prediction of disease causation based on histologic or cytologic evaluation of such exudates. These exudates are not static but are constantly changing in amount and in cellular and humoral makeup according to the ever-changing nature of the battle between the injurious agent and the tissue defenses. In its simplest form these changes may be nothing more than a gradual reduction in the intensity of the vascular response and leukocytic recruitment as the humoral and cellular defenses accomplish their task of diluting and destroying the offending agent. On the other hand, the nonspecific humoral and cellular defenses may become modified

by the addition of specific immune responses as the battle continues. These immune responses act, in general, to improve both the specificity and the efficacy of what at first are relatively broad and nonselective defensive strategies.

The purpose of these defenses is to confront and destroy the offending agent. In some circumstances this is a messy affair, with spillage of leukocyte contents or excessive diffusion of humoral cytotoxic chemicals (like complement fragments) into the local environment. The resulting bystander injury is a substantial and undesirable side effect of inflammation, and it is sometimes a justification for intervening with drugs to dampen the inflammatory response.

Peculiarities of Ocular Inflammation

The eye seems reluctant to become involved in inflammatory disease, and for good reason: The visual function of the eye is easily disturbed and is sometimes destroyed by minor degrees of inflammation that would be considered inconsequential in most other tissues. The blood-ocular barrier, the protective presence of eyelids, the tear film, the bony orbit, and a carefully regulated system of intraocular immune tolerance all seem designed to spare the globe from the need to participate in inflammatory responses (inflammation is a second level of defense to be triggered in any tissue only when the primary structural barriers have failed). When the intraocular tissues do become involved in an inflammatory reaction, the outcome is strongly influenced by the following three unique factors:

•The globe is a closed sphere that prevents the usual dissipation of inflammatory mediators and the clearing of exudates. Intraocular inflammation is thus virtually always diffuse, because inflammatory mediators and injurious agents easily become dispersed throughout the fluid media. Although clinicians distinguish anterior uveitis from posterior uveitis, on a histologic basis virtually all intraocular inflammation is diffuse (i.e., endophthalmitis). The diffusion of inflammatory mediators also probably explains why, for example, inflammation that predominantly affects the iris will also trigger perivascular inflammation within the distant retina, or why corneal stromal inflammation will inevitably cause increased vascular permeability within the iris and ciliary body, resulting in a clinically detectable increase in protein within the aqueous humor.

•The intraocular environment contains virtually no resident defenses. There are virtually no resident granulocytes, macrophages, or lymphocytes. There are no mechanisms for self-cleansing like coughing or peristalsis.

•The function of the eye is easily disturbed by what ordinarily might seem like minor manifestations of inflammation, such as accumulation of fluid, leukocytes, or tissue debris, or even by transient fluctuations in tightly regulated intraocular physiologic processes like production of aqueous humor. Even mild serous effusion can create vision-threatening serous retinal detachment, corneal edema, and cataract, whereas fibrinous effusions threaten to create traction retinal detachment, pupillary block, or occlusion of other parts of the aqueous drainage pathway, culminating in glaucoma (Figure 4-11).

Inflammation can occur in any of the intraocular and periocular tissues, but the periocular tissues—eyelid, conjunctiva, and orbital soft tissue—react exactly like other soft tissues,

FIGURE 4-11. A layer of granulation tissue (ARROW) has formed on the anterior surface of the iris, representing a source fo hyphema and the potential to cause glaucoma via occlusion of the pupil or the filtration angle.

such as skin and muscle. Our discussion here is limited to those ocular tissues that, for one reason or another, undergo rather specific and unusual patterns of inflammation.

Corneal Inflammation

The causes of corneal injury and subsequent inflammation are thoroughly discussed elsewhere in this text, but it is nonetheless useful to remember that the normal cornea is highly resistant to infectious disease. With the possible exception of viral keratitis, almost all inflammatory corneal disease is triggered by some kind of “environmental” injury to the cornea: abnormalities in the quantity or quality of the tear film, mechanical irritation from eyelids, cilia, or foreign bodies, or outright physical trauma. We nonetheless are often guilty of drawing excessively simplistic (and usually erroneous) conclusions about the significance of various bacterial or fungal isolates in the pathogenesis of corneal disease. Such erroneous conclusions often form the basis for ineffective or even harmful antimicrobial therapy.

Because inflammation is fundamentally a vascular event, the avascular cornea cannot undergo true inflammation until it has acquired blood vessels by ingrowth from the limbus. The very acute manifestations of corneal “inflammation” after injury (neutrophilia and corneal edema) are in fact passive events related to corneal ulceration, so that neutrophils and fluid are imbibed from the adjacent tear film. The first genuine inflammatory reaction to corneal injury occurs in the nearest available vascular bed, ordinarily that of the limbus. Depending on the diffusion of inflammatory mediators, vessels within the conjunctiva or even the iris may participate. A peripheral ring of edema (serous effusion) commonly accompanies increased permeability of limbal perivascular venules, and protein effusion from the iris vessels enters the aqueous humor to create aqueous flare. Leukocytes (usually neutrophils) migrate through these permeable vessels, and those from the bulbar conjunctiva or limbic sclera migrate into the corneal stroma at a level appropriate to the localization of the stimulus. Shallow injuries/ infections thus stimulate only a superficial stromal infiltration.

The corneal stroma may be the victim of bystander injury as the intracellular neutrophilic enzymes diffuse into the surrounding tissue while the neutrophils do battle with various infectious agents. The resulting lesion (Figures 4-12 and 4-13) is known as suppurative keratomalacia (malacia = softening).

FIGURE 4-12. Severe deep mycotic keratitis. The deep corneal stroma has been destroyed by the bystander effects of neutrophil accumulation. Descemet’s membrane, staining magenta, is heavily infiltrated by fungal hyphae.

FIGURE 4-13. Suppurative keratomalacia in an enucleated equine globe. The corneal stroma was destroyed by enzymes released from neutrophils immigrating into the cornea in response to opportunistic bacterial infection.

These leukocytes and the injured cornea itself are the most important sources of the fibroblastic and angioblastic growth factors that, after a delay of 3 or 4 days, initiate the ingrowth of blood vessels and fibroblasts from the limbus. There will also be fibroblastic transformation of the keratocytes themselves, which begin to produce large amounts of extracellular matrix, particularly chondroitin sulphate.

As all clinicians recognize, the transformation of the avascular cornea into a vascularized tissue capable of participating fully in inflammatory responses is a mixed blessing, for the enhanced protective and wound-healing capabilities of this “transformed” cornea are at the same time the events that will lead to some degree of permanent corneal scarring. Incidentally, this 3- to 4-day delay in ingrowth of blood vessels is a “period of grace” during which the cornea (either with or without our assistance) can try to deal with minor injury without running the risk of permanent scarring.

Nomenclature of Intraocular Inflammation

In theory one can have inflammation that is isolated to any of the intraocular compartments. Thus, one can postulate the existence of iritis, cyclitis (inflammation of the ciliary body), choroiditis, retinitis, and even hyalitis (inflammation of the vitreous itself). In reality, however, the interior of the globe almost always acts as a unit when responding to inflammatory stimuli, simply because the fluid ocular media do not allow rigid isolation of inflammatory agents or inflammatory chemical mediators. Thus we habitually use broader terminology, as follows:

•Anterior uveitis (iridocyclitis): Inflammation of the iris and ciliary body, with inflammatory exudate accumulating within the aqueous humor as aqueous flare or (if purulent) hypopyon.

•Posterior uveitis: Inflammation of the choroid, often with accumulation of inflammatory exudate in the subretinal space to cause retinal detachment.

•Chorioretinitis: Inflammation affecting the choroid and adjacent retina, almost always with exudate accumulating within the adjacent vitreous and subretinal space.

•Endophthalmitis: Inflammation of the entire uveal tract along with inflammatory effusion into the aqueous and vitreous humors.

•Panophthalmitis: Endophthalmitis that has extended to involve even the sclera. Although there are traditional clinical criteria to distinguish anterior from posterior uveitis, or choroiditis from chorioretinitis, almost all substantial intraocular inflammatory disease is histologically classified as endophthalmitis. Even when the majority of the inflammation is limited to the iris, there is almost always some inflammation within retina, choroid, and vitreous. Panophthalmitis is actually quite rare and is almost entirely limited to being a late complication of uncontrolled bacterial endophthalmitis after penetrating injury.

Common Sequelae of Intraocular Inflammation

Inflammatory Retinal Detachment

Because the retina is not truly attached to the RPE, there is a potential space between the photoreceptors and the RPE that is the remnant of the original optic vesicle. Any increase in permeability within the choroidal blood vessels, such as with hypertension or with acute inflammation, results in accumulation of fluid between the RPE and the photoreceptors, creating a serous retinal detachment (Figure 4-14). Much less common is the so-called traction detachment that results from organization of vitreal fibrinous exudates. Because the outer half of retina is nourished by diffusion from the choroid, any separation has dire consequences in terms of outer retinal viability. The rapidity with which there is microscopically visible photoreceptor injury, and the time at which that injury becomes irreversible, varies with the nature of the subretinal fluid and with the height of the detachment (i.e., the nutrient diffusion distance). Light-microscopic evidence of photoreceptor damage occurs within about 10 days following experimental infusion of saline into the subretinal space. It seems to occur much more quickly if that fluid is laden with leukocytic or other tissue breakdown products (see Figure 4-1).

Cataract

Because lens nutrition depends totally on the controlled flow of normal aqueous humor to deliver nutrients and remove wastes,

FIGURE 4-14. Diffuse endophthalmitis has caused serous retinal detachment, posterior synechia, and iris bombé. The fixative has caused coagulation of the markedly increased protein within the aqueous and vitreous humors that represents the inevitable outcome of greater vascular permeability during acute inflammation.

inflammation results in cataract because of the reduction in the production of aqueous and the addition of abnormal chemical constituents as by-products of inflammation.

Glaucoma

Most cases of secondary glaucoma are related to consequences of inflammation that, in most other tissues, would have been considered minor inconveniences. Because the iris normally

FIGURE 4-16. Posterior synechia leading to a pupillary block, iris bombé, and secondary peripheral anterior synechia. The pupillary block and the peripheral anterior synechia combined to cause inevitable secondary glaucoma.

contacts the anterior face of the lens, the most common of such complications is posterior synechia, in which the inflamed iris adheres to the surface of lens and results in pupillary obstruction and iris bombé (Figures 4-14 to 4-17). In contrast, iridocorneal adhesion (anterior synechia) is almost always associated with perforating corneal injury, which serves to “suck” the iris up to the corneal defect to facilitate its adhesion at that site (Figure 4-18). In contrast to what is often written, glaucoma caused by accumulation of inflammatory debris within the trabecular meshwork appears to be a very rare event in any of the domestic species. This is hardly surprising, because it is unlikely that exudate will seal a significant proportion of the 360-degree filtration angle. A more important cause of glaucoma is the development of a preiridal fibrovascular membrane on the face of the iris and within the trabecular meshwork itself. Such membranes are nothing more than ordinary granulation tissue, the normal response to the need to rebuild tissue scaffolding preparatory to parenchymal regeneration. Within the unforgiving environment of the eye, however, this granulation tissue can be deadly, because

FIGURE 4-15. The histologic section from the eye shown in Figure 4-14. The endophthalmitis occurred in response to a penetrating corneal injury (cat claw) that could have created traumatic uveitis, lens-induced (phacoclastic) uveitis, and/or bacterial endophthalmitis.

FIGURE 4-17. The iris adheres to the anterior capsule of a liquefied and collapsed lens, resulting in pupillary block and secondary glaucoma.

FIGURE 4-18. Transilluminated bovine globe with anterior synechia. The iris is drawn into a tent as it anchors into the cornea at the site of previous perforating injury.

FIGURE 4-19. Higher magnification of a preiridal fibrovascular membrane. The iris blood vessels are activated in response to angiogenic factors contained within the aqueous humor, and migrate onto the surface of the iris. They may cause pupillary block, peripheral anterior synechia, or hemorrhage.

it may seal the pupillary aperture (pupillary block) or impair outflow through the pectinate ligament (Figure 4-19).

Alterations in the Blood-Ocular Barrier and Implications for Immune-Mediated Disease

One of the fundamental events of inflammation, occurring within minutes of the inflammatory stimulus, is the chemically mediated contraction of endothelial cells that creates gaps through which fluid and, later, leukocytes may leave the vascular compartment. This results in rapid dissolution of the bloodocular barrier at the level of the endothelium of iris vasculature. Disruption of the blood-ocular barrier is not so rapid within ciliary processes, which tend to distend with accumulating inflammatory exudates until, eventually, the tight junctions of the nonpigmented epithelium become disrupted. At this stage these two components of the blood-ocular barrier are temporarily incompetent; restitution of this disrupted barrier takes many months and may never be perfectly restored. As a

result of this disruption there is an abnormally vigorous exchange of molecular substances between the intraocular and extraocular environments. In terms of therapy, this disruption of the blood-ocular barrier allows access of a wide range of therapeutic agents to the eye, a positive side effect of the inflammatory process. It also results in interaction between ocular antigens and the systemic immune system, an interaction that ordinarily is tightly regulated to maintain tolerance by the extraocular immune system to small amounts of leaking intraocular antigens.

The concept that most intraocular antigens, particularly those of lens, are totally “foreign” is incorrect, but the amount that normally reaches systemic (splenic) immunocytes is very small. The concept of anterior chamber–associated immune deviation presumes that intraocular antigens are somehow processed within the eye (probably by dendritic cells within the trabecular meshwork) before draining from the trabecular meshwork into systemic circulation. These antigens, upon reaching the spleen, initiate a typical humeral immune response but an atypical, muted T-cell response. There is no defect in proliferation of antigen-specific cytotoxic T cells, but there is induction of a population of suppressor T cells, which inhibits any cell-mediated hypersensitivity to the sensitizing antigen. The theoretical result is that any lymphocytes that return to the eye are capable of producing only the most localized and specific types of inflammatory events, and the subsets of T cells that produce a lot of nonspecific “bystander” injury are suppressed.

Following virtually any kind of ocular injury, antigens leaving the eye reach the spleen, stimulate lymphoid proliferation, and result in return of splenic lymphocytes to the eye about 1 week later. Typically these lymphocytes are seen as perivascular aggregates in the iris and ciliary stroma. As in inflammatory reactions anywhere else, only a small proportion of these returning lymphocytes are specifically sensitized to the initiating antigen. The rest carry receptors for a wide variety of unrelated antigens, which may be important to our understanding of the recurrent nature of many cases of uveitis in domestic animals. In brief, it is quite plausible that the recurrent bouts of uveitis are triggered by exposure of these polyclonal lymphocytes to a wide variety of antigens, even vaccinal antigens, that have nothing to do with the events that caused the original bout of uveitis.

The very slow restitution of a blood-ocular barrier and the presence of these polyclonal reactive lymphocytes would seem to guarantee that all uveitis be persistent or cyclical. The presence of active transforming growth factor-beta (TGF-B) in normal aqueous humor has a profound inhibitory effect on lymphocyte (particularly T-cell) proliferation and also on the action of such potent effectors of cell-mediated immunity as interleukin-2 and interleukin-4. What we see is the outcome of a very complex interaction of many interacting and overlapping regulatory systems designed to enhance the specificity of the immune response, and to limit its duration and harmful side effects.

Etiologic Implications of Inflammatory Exudates

The admirable specificity of the acute inflammatory response notwithstanding, it is essential to recognize that inflammation is a dynamic and changing system and that the initial events of inflammation blend with those of immune-mediated events

even within a few days. Thus inflammation that is initially suppurative may change into something that looks granulomatous or lymphocytic, and conversely, it is theoretically possible that neutrophilic inflammation may become eosinophilic, or that eosinophilic may become granulomatous. An example is feline infectious peritonitis, which is described as being granulomatous or pyogranulomatous but which may be suppurative, fibrinous, lymphocytic-plasmacytic, or granulomatous, depending on the stage of the disease and on complex immunologic interactions that remain undiscovered.

Another example is phacoclastic uveitis, which varies from suppurative to lymphocytic to granulomatous, depending on the interval between lens injury and histologic examination. This is a delayed, immune-mediated reaction to the release of large amounts of previously normal lens protein into the intraocular environment. Although that lens protein is certainly not completely “foreign,” the sudden release of a large volume of such protein after spontaneous or traumatic lens rupture overwhelms immune tolerance and results in a severe suppurative- to-granulomatous perilenticular uveitis that often destroys the globe. Even if the inflammation can be successfully controlled, its wound-healing sequelae often induce pupillary block and secondary glaucoma. The amount of lens material leaking out of the lens seems to be the most important determinant of whether this fateful reaction is triggered. There are other variables as well, for considerable species differences cannot be explained just on the basis of duration of disease or the size of the lens capsular defect. In rabbits and birds the reaction in phacoclastic uveitis is much more granulomatous than in dogs, cats, or horses (Figures 4-20 and 4-21).

Nonetheless, certain broad principles allow a guess as to the nature of the offending agent until more definitive identification (via culture or other means) is available:

Suppurative exudates imply bacterial infection, particularly when the neutrophils are lytic. Neutrophils also predominate in the early stages of phacoclastic uveitis.

Eosinophilic exudates occur with parasitic migration but are only rarely seen within the globe. They more frequently occur around parasitic migration tracts within the conjunctiva (as with Habronema or Onchocerca lesions in horses). More commonly, eosinophilic infiltrates are seen with allergic reactions and are particularly frequent in the conjunctiva and cornea of cats affected with eosinophilic keratitis.

Granulomatous inflammation is inflammation that is dominated, from its onset, by macrophages. It is a term that is commonly abused, because how many macrophages are required, or what percentage of the total leukocyte population must be macrophages before such a classification can be properly used, has not been established. Because all inflammation will sooner or later become dominated by mononuclear leukocytes, there is a danger that the somewhat specific term “granulomatous” will be used interchangeably with simple chronic inflammation. There are numerous dramatic examples of granulomatous inflammation in ocular pathology. Leakage of sebaceous secretion from inflamed or neoplastic sebaceous glands results in a pure granulomatous periadnexal inflammation known as chalazion. A mixed lymphocytic and granulomatous nodular inflammation of the conjunctiva or adjacent sclera is common in dogs and has variously been called nodular fasciitis, nodular granulomatous episcleritis, or even fibrous histiocytoma. Various systemic mycoses stimulate a granulomatous or mixed neutrophilic-histiocytic (pyogranulomatous) infiltrate, commonly

FIGURE 4-20. Traumatic lens rupture resulting in an intractable, immunemediated pyogranulomatous perilenticular inflammation known as phacoclastic uveitis.

FIGURE 4-21. Phacoclastic uveitis in a rabbit. The still-intact lens fibers are surrounded by a layer of neutrophils, then foamy macrophages, and then lymphocytes. The character of the reaction varies substantially with time and with the species.

more obvious in the subretinal space and choroid than in the anterior uvea (Figure 4-22). Finally, cell-mediated immune responses frequently culminate in granulomatous inflammation, as occurs in phacoclastic uveitis and in the diffuse granulomatous destructive endophthalmitis known as uveodermatologic syndrome (Vogt-Koyanagi-Harada–like syndrome) that appears to be an immune response against intraocular (and occasionally epidermal) melanin pigment.

Lymphocytic-plasmacytic inflammation is the most common pattern of inflammation within the globe. It may reflect specific recruitment of lymphocytes and plasma cells in response to viral infection or to other intraocular antigens incapable of inducing the other humoral and cellular types of inflammation. Often, however, lymphocytic-plasmacytic inflammation reflects

FIGURE 4-22. Choroidal suppurating granuloma in a dog with systemic blastomycosis. With higher magnification, budding yeast was found in similar granulomas throughout the choroid and subretinal space.

nothing more than the normal maturation of the inflammatory response. Virtually all types of inflammation, as they age, recruit growing numbers of long-lived mononuclear leukocytes that eventually predominate within the exudate. Such accumulations may reflect a very specific response to an infectious agent that is now recognized as an antigen, but it is equally possible that these lymphocytes are responding to normal ocular antigens excessively exposed by the inflammatory disruption of the blood-ocular barrier. Lymphocytic anterior uveitis is the usual pattern seen in posttraumatic uveitis even when there has been no penetrating injury. Regardless, lymphocytes and plasma cells may occur as diffuse uveal infiltrates or as discrete perivascular lymphoid aggregates. In all domestic species we see persistent or recurrent lymphocytic-plasmacytic uveitis, and hardly ever do we identify a specific offending agent or antigen (Figures 4-23 and 4-24).

RESTORATION OF HOMEOSTASIS: OCULAR WOUND HEALING

The mechanisms of wound healing within the eye have received a great deal of attention for two reasons. First, healing of ocular injuries is intrinsically important for the well-being of the eye, and even minor defects in this normally well-regulated process may have disastrous consequences for optical clarity, aqueous drainage, or other ocular functions that are exquisitely sensitive to minor changes in anatomic fidelity. Second, the ability to directly visualize the sequelae to injury in such tissues as cornea or even retina has made the eye a much-favored tissue for basic research in the mechanisms and pharmacologic manipulation of wound healing.

Wound healing includes both mitotic parenchymal regeneration and regeneration of the connective tissue scaffolding, sometimes simply referred to as “scarring.” They frequently are portrayed as competing or antagonistic activities, but they are

FIGURE 4-23. Coalescing perivascular lymphoid aggregates create socalled lymphonodular uveitis in a cat. Similar lesions are seen in all species as the common histologic counterpart of persistent or recurrent uveitis. The cause is rarely determined, and such cases are frequently dismissed as “immune-mediated.”

4-24. A higher magnification of the perivascular lymphoid accumulation within the iris of a cat with lymphonodular uveitis seen in Figure 4-23.

interdependent and even complementary. Following any substantial tissue injury, successful parenchymal regeneration requires the removal of any persisting infection and tissue debris, the preservation or reconstruction of a connective tissue scaffold, the survival of an adequate population of mitotically competent germinal cells, and the presence of an adequate blood supply. If the initial injury has so damaged the original tissue stromal scaffold or vasculature that it cannot perform this supportive role, parenchymal regeneration must wait until a new scaffold and/or vasculature has been recruited from the nearest available source. Although this new scaffolding may seem unsightly or even excessive, it is a deliberate and essential prerequisite to parenchymal rebuilding.

At a molecular level wound healing is regulated by a variety of locally produced cytokines, the most important of which are basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), TGF-B, platelet-derived growth factor (PDGF), and vascular endothelial growth factor (VEGF). As with those cytokines that act as inflammatory mediators, the action of these wound-healing cytokines also is greatly influenced by the context in which each occurs (i.e., where, when, how much, and in what company). Although interest in these cytokines has given rise to a huge body of literature, it is wise to remember the four classic “morphologic” prerequisites for regeneration: débridement, stromal scaffold, germinal cells, and blood supply. It is inadequacy in one or more of these traditional prerequisites that most commonly results in a failure of proper healing, and it also is these various phenotypic parameters that we attempt to manipulate by such procedures as mechanical or chemical débridement, suturing, and grafting. Nonetheless, the availability of at least a few of these cytokines as purified therapeutic agents holds great promise in the investigation and treatment of a whole range of conditions typified by apparently defective repair.

Germinal Cells

Tissue regeneration can occur only in those tissues that contain a population of mitotically competent germinal cells. The application of this principle to the eye is obvious when one deals with wound healing in retina. Injuries to the developing retina often result in disorderly reparative proliferation of the retinal neurons, leading to retinal dysplasia. Exactly the same injury to the adult retina results in retinal scarring, because only the glial cells and vasculature are able to proliferate in the mature retina. The age at which the retinal neurons cease mitotic activity varies with the species and results in a different temporal definition of “retinal dysplasia,” depending on the species involved. For cattle and horses lesions of retinal dysplasia (rosettes and disordered intermixing of the various retinal layers) is necessarily an in utero event, but in dogs and cats injury to the retina any time up to about week 6 after birth results in retinal dysplasia. Because, in the latter two species, the retina matures first near the optic disc and ceases mitotic activity last near the ora ciliaris, it is sometimes possible to determine the age at which the insult occurred from the topographic distribution of the dysplastic lesions. In those areas of retina that had ceased mitotic activity at the time of the injury, retinal scars rather than dysplastic foci will develop. In at least some species of fish the retina retains mitotic capability throughout life, and thus lesions of retinal dysplasia could develop at any age.

At the opposite end of the regenerative spectrum is the conjunctival and corneal epithelium. The normal corneal epithelium turns over every 7 to 10 days, fed by a population of germinal cells with a limited number of “preprogrammed” mitotic cycles (the transient amplifying population). The permanent replicative population, with no apparent limit on mitotic capability, is found at the junction of corneal and conjunctival epithelium near the limbus.

After shallow injury to the corneal epithelium, the immediate (after about 1 hour) reaction is one of flattening and sliding by viable adjacent wing cells, and then by basal cells, so that even large (1-cm) defects may be covered in as little as 72 hours if the basal lamina is intact. Within 24 to 36 hours

mitotic figures are seen within the permanent replicative population at the limbus but are inconspicuous in the epithelium adjacent to the ulcer itself. After a single injury mitotic activity becomes maximal at 10 days, dropping rapidly thereafter.

The loss of the hydrophobic epithelial barrier allows for the rapid osmotic absorption of water from the tear film into the superficial stroma. Even in nonseptic shallow ulcers the fluid absorption is usually accompanied by modest numbers of neutrophils (Figures 4-25 and 4-26). In the presence of bacterial or fungal contamination the number of neutrophils is greatly increased. The leakage of neutrophil enzymes into the surrounding stroma may lead to stromal malacia and the risk of corneal perforation and iris prolapse (Figures 4-27 and 4-28).

Persistent or repeated corneal epithelial injury, in the presence of a competent germinal population and an adequate stroma, results in epithelial hyperplasia and even dysplasia. Such stimulated epithelium may acquire rete ridges and even pigmentation, perhaps because the ingrowing cells from the limbus have not entirely abandoned their conjunctival heritage. If the stimulus is chronic desiccation, the epithelium will also undergo adaptive keratinization. In most cases there is accompanying injury to the superficial stroma, resulting in ingrowth of fibroblasts and blood vessels in parallel with the adaptive epithelial changes and leading to so-called corneal cutaneous metaplasia (Figures 4-29 and 4-30).

A variety of ocular tissues, particularly corneal endothelium and lens epithelium, offer interesting adaptations in cellular regeneration. Depending on the species, each of these tissues has relatively little mitotic capability and, under normal circumstances, regenerates poorly. This is true of corneal endothelium, which is claimed not to regenerate at all in adult dogs and cats. Loss of corneal endothelial cells results in sliding of adjacent

FIGURE 4-25. Normal canine cornea. The nonkeratinized, nonpigmented, stratified corneal epithelium adheres to the surface of a cell-poor, avascular, and highly regimented fibrous stroma.

FIGURE 4-26. Acute corneal ulcer. Within hours of injury causing shallow ulceration, viable epithelial cells adjacent to the defect flatten and slide across the surface of the defect. The denuded corneal stroma rapidly absorbs water and neutrophils from the tear film.

FIGURE 4-27. Acute corneal ulcer progressing to suppurative keratitis. If the exposed stroma becomes contaminated with bacteria, the recruitment of neutrophils is greatly increased, and the potential for neutrophilmediated stromal destruction (keratomalacia) is similarly greater.

FIGURE 4-29. Healed corneal ulcer with subepithelial fibrosis and vascularization. The resulting cornea will be permanently opaque.

FIGURE 4-30. Corneal cutaneous metaplasia with subepithelial scarring and pigmentation. This could be a sequel to previous ulceration, or to persistent sublethal corneal irritation to which the cornea must adapt.

FIGURE 4-28. Rapidly progressing corneal ulcer leading to perforation and iris prolapse.

viable endothelium to seal the defect and may not cause any permanent functional impairment of the endothelial monolayer. In contrast, some injuries result in fibroblastic metaplasia of the corneal endothelium. In this fibroblastic disguise the epithelium seems to acquire very dramatic proliferative capabilities,

sometimes resulting in so-called retrocorneal fibrous membrane, which can lead to glaucoma. Similarly the mature lens epithelium appears to have very little replicative capability yet responds with fibroblastic metaplasia to a variety of injuries; in this form, it is capable of dramatic proliferation, which can result in drastic consequences in terms of pupillary obstruction. As one “descends” the phylogenetic scale the proliferative capacity of ocular tissues increases, so that adult corneal endothelial cells and even retinal neurons in reptiles, amphibians, and fish may retain mitotic capability.

Tissue Scaffold

To rebuild a functionally significant anatomic unit, epithelial cells require some kind of tissue scaffold as a “road map” for repair. If the initial injury has not left enough of the original scaffold, a new one must be built with contributions from the epithelial cells themselves (i.e., basement membrane) and from proliferation of adjacent stroma. When properly regulated the

combination of these two phenomena results in perfect, or almost perfect, restitution of normal structure and function.

The best example is healing of corneal ulcers. A shallow defect heals by sliding, replication, and eventual normalization of the corneal epithelium as described earlier. Adhesion to the denuded stroma is at first via fibronectin and laminin produced by reactive stromal fibroblasts, absorbed leukocytes, and the epithelium itself. In as little as 3 days, the epithelium produces at least some fragments of new basement membrane in amounts great enough to be microscopically detected. It tends to do this even if the original membrane is still present, and this thickening or duplication of basement membrane serves as a reliable marker of previous epithelial injury in cornea and lens. Re-formation of the hemidesmosomes and the ultrastructural collagenous anchoring fibrils that represent the normal, firm adhesion between the epithelium and the stroma may require many weeks.

Those defects that involve a loss of more than the superficial 25% of the stroma often require both stromal and epithelial repair (this figure is arbitrary and just an approximation; numerous exceptions exist). In these instances the epithelial regeneration must await the rebuilding of the stromal scaffold that occurs via a process of fibroplasia and angiogenesis identical to wound healing in the skin. The major difference, of course, is the need for the normally avascular corneal stroma to recruit angioblasts from the limbus, and thus the process of corneal stroma repair is considerably delayed compared with that occurring in normally vascularized collagenous tissues such as the dermis. Under experimental circumstances recruitment of the angioblasts begins by budding from venules at the limbus within as little as 72 hours after corneal injury, and these vessels migrate in a laminar fashion into the stroma, moving approximately 1 mm per day.

The combination of stromal fibroplasia and vascularization (usually termed granulation tissue) can be seen migrating all the way from the limbus to the site of stromal injury. Over the bed of granulation tissue, the epithelium slides, proliferates, produces new basement membrane, and adheres. As in wounds elsewhere, the fibroblasts produce a whole sequence of precollagenous and collagenous matrix types that eventually remodel into a stroma that is only slightly less regular than the adjacent normal stroma. The resultant scar never quite matches the normal cornea in terms of histologic architecture or optical clarity, but some cases come very close (see Figure 4-30).

The same phenomena of epithelial proliferation, basement membrane reduplication, and stromal fibroplasia/angiogenesis occur elsewhere in the eye, but in these noncorneal locations they usually seem unwelcome events. For example, injury to lens epithelium via perforation, by adherence of pupillary membranes, or sometimes just through the biochemical events of cataract results in plaquelike proliferation and fibrous metaplasia of capsular epithelium, usually with reduplication of multiple laminae of basement membrane to increase the density and thus the visual significance of the lens opacity. As a more exaggerated example, the devastating consequences of lens rupture (phacoclastic uveitis) seem to be the result of escape of the lens epithelium through a rent in the lens capsule, resulting in perilenticular proliferation of this fibroblastic epithelial membrane to cause pupillary block and secondary glaucoma (see later). The important phenomenon of preiridal fibrovascular membrane (and rarely, preretinal fibrovascular membrane) is no more than a proliferation of granulation tissue

from the stroma of the iris (or retina) that is “accidentally” exposed to growth factors released at some distant site within the eye (see Figures 4-18 and 4-19). Ciliary body tumors and detached retinas are particularly common and potent initiators of preiridal and, occasionally, preretinal membranes. These membranes are initially fragile and may lead to intractable hyphema or, as they mature, may cause pupillary block, retinal detachment, or sealing of the filtration angle. Trabecular endothelial cells within the filtration angle are also susceptible to the same proliferative stimuli and may contribute to the glaucoma that frequently follows the formation of preiridal membranes.

Adequate Nutrition

Lack of adequate nutrition is an important determinant of wound healing in the cornea and lens, two tissues that normally have a precarious nutritional supply. Cornea is nourished by absorption from the tear film, anterior chamber, and vascular network at the limbus. Of these, the presence of a quantitatively and qualitatively adequate tear film appears to be most important. Ingrowth of blood vessels from the limbus is commonly seen during the healing of any severe corneal injury (see Figures 4-29 and 4-30). Although it may have serious implications for later corneal transparency, this ingrowth is an essential ingredient of the wound-healing process. Such vessels are permanent, although they diminish in size and are no longer filled with blood, because the requirement for this augmented nutritional support fades after reconstruction of the corneal wound.

Various types of conjunctival grafts are nothing more than efforts to speed up the arrival of blood vessels into the injured corneal stroma, therefore augmenting whatever nutrition is arriving via the tear film.

Diseases Resulting from Defective Wound Healing

Although many ocular syndromes could be considered the result of inappropriate wound healing—corneal stromal scarring, posterior synechia, traction retinal detachment and postnecrotic retinal dysplasia, which have been mentioned already—the wound healing in those syndromes is perfectly normal and appropriate to the nature of the initial injury. There are, however, a few examples of improperly regulated or inadequate wound healing that cause distinct ocular disease syndromes.

The syndromes canine persistent ulcer, feline corneal sequestration, and equine corneal sequestration are probably all reflections of the same fundamental defect. With every significant corneal epithelial injury, there is a transient degeneration of the most superficial corneal stroma, characterized by cellular apoptosis and stromal matrix disintegration. In normal wound healing that superficial stromal degeneration is only transient and does not interfere with subsequent migration and adhesion of the regenerating epithelium. In some individuals, however, the depth and persistence of that stromal degeneration are excessive and it becomes histologically visible as a corneal sequestrum. Because permanent adhesion of the regenerating epithelium depends on the extension of ultrastructural “anchors” from the epithelial cell membranes into the superficial stroma, the inability of the degenerate stroma to secure those anchors results in ineffective epithelial adhesion and recurrent ulceration. In dogs the most obvious clinical manifestation is the recurring ulceration; in cats the absorption of colored break-