Increased interleukin-1 activity in the injured vitamin A-deficient cornea.

Injury to a vitamin A-deficient cornea leads to severe acute inflammation often culminating in ulceration. We report on possible regulatory mechanisms involved in the pathogenesis of corneal inflammation in vitamin A deficiency. Thymocyte comitogenic assay and interleukin (IL)-6 induction in corneal fibroblasts have shown that thermally injured and mechanically abraded vitamin A-deficient rat corneas produce much higher levels of an IL-1-like factor as compared with uninjured or injured, normal control corneas. This was confirmed by antibody capture enzyme immunoassay, which detected high levels of IL-1 alpha and IL-1 beta in injured vitamin A-deficient corneas. To our knowledge this is the first report describing the induction of IL-1 in the vitamin A-deficient cornea by thermal and mechanical injuries. When mechanically injured corneas were screened for chemotactic activity, they were found to contain significantly higher levels of a chemoattractant as compared with similarly injured, normal control corneas. Chemotactic activity [expressed as a percentage of a known chemotactic tripeptide, formyl-methionyl-leucyl-phenylalanine (fMLP), found in medium harvested from vitamin A-deficient corneas] averaged 58.8 +/- 8.9% (SEM) as compared with 12.6 +/- 5.4% in medium conditioned by normal corneas. Checkerboard analysis confirmed that the activity in vitamin A-deficient cornea conditioned medium was chemotactic and not chemokinetic. These results demonstrate a correlation between IL-1 levels and severity of inflammation in the injured vitamin A-deficient rat cornea.

Increased gelatinolytic and caseinolytic activity in the thermally injured, nutritionally compromised rat cornea: detection of a 27-kDa lymphoreticular cell-associated caseinase.

This study assesses the impact of various forms of injury on matrix degrading enzymes in nutritionally compromised rat corneas. In vitamin A-deficient (nutritionally compromised) and normal control corneas, in vivo or ex vivo mild mechanical abrasion did not appreciably alter the activity of either the 65-kDa or the 92-kDa gelatinases. In contrast, after thermal injury, while no appreciable change was detected in activity associated with the 65-kDa gelatinase in either vitamin A-deficient or normal control corneas, 92-kDa gelatinolytic activity was consistently higher in corneas from both groups, although activity associated with nutritionally compromised corneas was much higher. In these corneas, thermal injury also induced the expression of two high molecular weight (approximately 130-kDa and 225-kDa) gelatinases and a 27-kDa caseinase. While gelatinases were totally inactivated by inhibitors of metalloproteinases such as 1,10-phenanthroline and Galardin MPI, the 27-kDa caseinase showed considerable susceptibility to a mixture of serine protease inhibitors (aprotinin, dichloro-isocoumarin and pA-PMSF [(4-amidino-phenyl)-methane-sulphonyl fluoride]. Furthermore, unactivated-lymphoreticular cells from either nutritionally compromised or normal control animals contained a 24- and 27-kDa caseinase, however most of the activity was due to the 24-kDa caseinase. In contrast, glycogen-activated lymphoreticular cells contained a preponderance of the 27-kDa caseinase. Activated-lymphoreticular cells also expressed 92-kDa, 130-kDa and 225-kDa gelatinases. The presence of low molecular weight caseinases in lymphoreticular cells implicates them as the source of these enzymes.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of vitamin A deficiency on the adhesion of rat corneal epithelium and the basement membrane complex.

PURPOSE: To understand the underlying mechanisms responsible for the easy removal and sloughing of corneal epithelium in vitamin A deficiency. METHODS: An animal model of vitamin A deficiency, the vitamin A-deficient rat (A-rat), transmission electron microscopy, computer-assisted morphometric analysis and indirect immunofluorescence were used to study the adhesion of rat corneal epithelium to its basement membrane with emphasis on structure and molecular composition of the anchoring structures such as the hemidesmosome and bullous pemphigoid antigen. RESULTS: Transmission electron microscopy resolved numerous microseparations of the basal epithelial cell membrane from the basement membrane with intervening segmental basement membrane duplications and electron dense deposits. Morphometric analysis disclosed a statistically significant reduction in the frequency and size of hemidesmosomes. Four weeks after supplementing the diet with retinyl acetate (700 micrograms/week), significant reversal of these same structural abnormalities could be detected. Immunofluorescence staining for bullous pemphigoid antigen, a component of the adhesion complex, showed intense staining of the basal epithelial cytoplasm but weak and discontinuous staining of the basement membrane. Weak staining for laminin was also evident in A- corneas. In contrast, normal corneas displayed no cytoplasmic staining for bullous pemphigoid antigen and intense staining of the basement membrane for bullous pemphigoid antigen and laminin. CONCLUSIONS: The authors propose that structural abnormalities of the epithelial basement membrane complex are responsible for the observed loose epithelial adhesion and sloughing, as well as other known abnormalities of healing in the vitamin A-deficient rat cornea.

Cleavage and activation of corneal matrix metalloproteases by Pseudomonas aeruginosa proteases.

PURPOSE: To examine the effect of Pseudomonas aeruginosa on the expression of corneal matrix metalloproteases and the effect of its proteases on activation of corneal matrix metalloproteases in vitro. METHODS: Rat corneas and human corneal fibroblasts were co-cultivated with two different strains (RPS & 599A) of P. aeruginosa and one strain of Staphylococcus aureus, and the conditioned media were analyzed for proteolytic activity by gelatin and casein zymography. Human corneal fibroblast-conditioned medium was incubated with that from either strain of P. aeruginosa and was analyzed in a similar manner. RESULTS: Normal rat corneas in organ culture produce a 65 kDa gelatinase (inactive matrix metalloprotease-2), whereas thermally injured rat corneas additionally produce gelatinases with molecular masses of 92 kDa (inactive matrix metalloproteases-9) and > 200 kDa. Matrix metalloprotease-2 is also detected in human corneal fibroblast-conditioned medium. Although these matrix metalloproteases are no longer detectable when rat corneas or human corneal fibroblasts are co-cultured with two strains of P. aeruginosa for 48 hr, a 58 kDa gelatinase fragment appears in earlier stages of co-culture. In contrast, S. aureus does not affect matrix metalloprotease-2. The 58 kDa fragment is also evident by incubating human corneal fibroblast-conditioned medium with that from either strain of P. aeruginosa. Conditioned medium from the RPS strain, which produces both elastase and alkaline protease, is more effective in cleaving matrix metalloprotease-2 than that from the 599A strain, which expresses mainly alkaline protease. CONCLUSION: The secreted inactive corneal matrix metalloprotease-2 is activated through limited proteolysis by pseudomonal proteases.

Proteolytic activation of corneal matrix metalloproteinase by Pseudomonas aeruginosa elastase.

Purified Pseudomonas aeruginosa elastase cleaved a 65 kDa gelatinase [inactive proenzyme form of matrix metalloproteinase (MMP-2)] from human corneal fibroblasts into a biologically active fragment with an approximate molecular mass of 58 kDa. However, purified pseudomonal alkaline protease did not cleave MMP-2 appreciably. Since activated MMP-2 is known to degrade native type IV, V and VII collagens, all components of the corneal basement membrane or stroma, our results suggest a new role for pseudomonal elastase in the pathogenesis of corneal infection, inflammation and ulceration.

Clinical and histopathologic studies of two families with lattice corneal dystrophy and familial systemic amyloidosis (Meretoja syndrome).

Lattice corneal dystrophy associated with familial systemic amyloidosis (Meretoja syndrome) has rarely been described other than in patients of Finnish origin. The authors report two North American patients with this disease who manifest blepharochalasis, lattice corneal dystrophy, open-angle glaucoma, and cranial neuropathy. In one patient, a corneal intraepithelial and subepithelial pseudodendrite was managed by superficial keratectomy, and this same patient benefited from surgical brow suspension for facial muscular weakness. In the second patient, penetrating keratoplasty was complicated by a neurotrophic persistent epithelial defect. Corneal tissue from both superficial keratectomy and penetrating keratoplasty exhibited ultrastructurally characteristic amyloid filaments and associated elastoid material. Transmission electron microscopy of conjunctiva and skin biopsies similarly revealed amyloid deposits associated with most basement membranes, the perineurium and endoneurium of most peripheral nerves, and the intima and media of arteries. By immunoperoxidase staining, the corneal amyloid deposits were positive for the amyloid P-component protein but negative for the nonimmunoglobulin amyloid A protein and prealbumin. Serum prealbumin and amyloid A related protein were normal.

Experimental Pseudomonas aeruginosa keratitis from extended wear of soft contact lenses.

We used a rabbit model to investigate the pathogenesis of soft contact lens-induced bacterial keratitis. Rabbit eyes underwent complete tarsorrhaphy for 7 days either with (group A, n = 14) or without (group B, n = 13) new sterile soft contact lenses. On day 7, an increase in mean corneal thickness (20.3% in group A and 17.2% in group B) was detected. New or rabbit-worn soft contact lenses were then inoculated with 10(7) colony-forming units of Pseudomonas aeruginosa or by 0.1 mL of P aeruginosa suspension. On day 9, conjunctival cultures of all eyes yielded P aeruginosa. Corneal infection developed in 11 of 14 eyes wearing new or worn, contaminated soft contact lenses. Bacterial keratitis did not develop in any of the 13 eyes inoculated with P aeruginosa suspension. Light and electron microscopy of infected eyes showed abundant polymorphonuclear neutrophils destroying the epithelium, basement membrane, and stroma. Few bacteria could be detected and only in the deep stroma. Since bacterial suspension alone caused no inflammation, soft contact lens-wear appears crucial to corneal infection in this model.

Histopathology of human keratorefractive lenticules.

Eight human corneal lenticules (three from keratophakia, two from hypermetropic keratomileusis, and three from myopic keratomileusis) were examined by light and electron microscopy. The keratophakic lenticules were removed for optical reasons 3, 3.5, and 6 months after surgery. Microscopically, all displayed hypocellularity, mature collagen fibrils and microfibrils, keratocyte ghosts, and keratocytic debris. The hypermetropic keratomileusis specimens were removed at 13 and 14 months postoperatively, the first because of opacities from enzymatic digestion of the cornea, and the second due to contact lens-induced erosion of Bowman's layer and decreased lacrimal secretion. Ultrastructurally, both lenticules exhibited fractures in Bowman's layer, and the 14-month specimen showed multilayered squamous epithelia. The myopic keratomileusis specimens were removed at 4, 8, and 48 months postoperatively due to opacification of the interface caused by delayed epithelial healing, detergent trauma, and previous epithelization, respectively. Electron microscopy revealed fractures in Bowman's layer, subepithelial fibrocellular growth, sparse keratocyte populations of the anterior stroma, porous collagen bundles, keratocytic debris, and regions of epithelial ingrowth.

Stromal degradation in vitamin A-deficient rat cornea. Comparison of epithelial abrasion and stromal incision.

Epithelial abrasions (3-mm diameter) and linear stromal incisions (50-75% depth) were made in vitamin A-deficient (A-) rat corneas to investigate what are contributory factors to the development of keratomalacia. Sixty-four rats were killed at various times after injury, and the corneas were histologically examined. In pair-fed control corneas, wounds healed without stromal degradation. Although abraded A- corneas were infiltrated by numerous polymorphonuclear neutrophil leukocytes (PMNs), reepithelialization eventually occurred and severe stromal degradation was not evident, suggesting that PMN infiltration alone cannot cause keratomalacia. In incised A- corneas, six (20%) exhibited marked stromal degradation, of which three (10%) were infected. Two other corneas studied shortly after incision were in the initial stages of infection. The results suggest that stromal injury importantly contributes to the development of keratomalacia. Bacterial infection might also be contributory. Cytochrome oxidase staining showed that metabolic functional levels of incised A- corneas were accelerated only in the wound zone, whereas abraded A- corneas were metabolically accelerated throughout.

Time course of human conjunctival mast cell degranulation in response to compound 48/80.

Compound 48/80, a non-immunogenic mast cell degranulatory agent, is known to produce the signs and symptoms of ocular allergy. Maximal mast cell degranulation of human conjunctiva occurred within the first hour after stimulation by a single topical dose of compound 48/80 (20 microliters, 7.5 mg/ml). The average percentage of fully degranulated mast cells in treated specimens (n = 9) was 31% (range 5-60%) versus 6% (range 0-20%) in control specimens (n = 5). Exact correlates of representative granulated, partially degranulated, and fully degranulated mast cells were determined by light and transmission electron microscopy.

Natural history of disease in a rabbit model for keratoconjunctivitis sicca.

We have continued our study of the tear film and ocular surface in our full KCS (keratoconjunctivitis sicca) rabbit model up to 52 weeks post-operatively. Tear film osmolarity remains elevated, conjunctival goblet cell density remains decreased, and the conjunctival epithelium remains abnormal. Corneal epithelial glycogen levels decreased progressively, and at 44 weeks post-operatively rabbits developed abnormal rose Bengal staining of the affected cornea that was shown to be associated with morphologic abnormalities at 52 weeks. Rabbits began rubbing the affected eye after the development of corneal staining. Our full KCS rabbit model demonstrates the features of the human disease.

Gentamicin toxicity in the primate retina.

To study the toxic effect of aminoglycoside antibiotics in the primate retina, gentamicin sulfate was injected into the center of the vitreous cavity of Cebus navrigatus monkeys. At a dose of 1000 to 3000 microgram, a picture consistent with apparent macular infarction appeared on fundus examination and fluorescein angiography by three days and gradually faded by 21 days. While light and electron microscopic examination of the retina showed no primary vascular lesions, striking damage to the inner retinal layers, mainly the nerve fiber layer, ganglion cell layer, and the inner plexiform and nuclear layer, was seen. Less severe effects in the outer retinal layers and the retinal pigment epithelium occurred. These observations suggest that the neurotoxic effect of intravitreal gentamicin was sufficient to cause a complete shutdown of the regional blood flow, perhaps by the mechanism of granulocytic plugging of the capillary bed. Although this toxic effect occurred at doses considerably in excess of what has been recommended for clinical use in humans, the "safe" dose of intravitreal gentamicin, nevertheless, remains to be established unequivocally.

Noninvasive metabolic analysis of preserved rabbit cornea.

With the method of corneal redox fluorometry, the autofluorescence of reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) in rabbit corneal endothelium was measured as a function of storage time in McCarey-Kaufman (MK) medium and K-Sol medium. Measurements were started immediately after preparation of the corneal button, and the corneas were followed up for up to three weeks of storage. In both media, the PN/Fp ratio of the endothelium initially increased slightly in the first or second day and then began to decrease toward a level lower than baseline. This initial increase is possibly a result of an adaptive mechanism. The PN/Fp ratio maintained itself in the range of baseline values up to one week in MK medium but not in K-Sol medium. With scanning electron microscopy, the surface of the membrane and cell border were maintained for a three-week preservation period, and no apparent differences were found between the corneas stored in the two media. With transmission electron microscopy, the intracellular organelles appeared almost normal through one week of preservation in corneas stored in either media. During weeks 2 and 3, however, intracellular edema with increased endothelial thickness became prominent in the corneas stored in both media. Although no visual difference in the morphological features of the endothelium was apparent between corneas stored in either medium, computer-assisted morphometric analysis showed a statistically significant increase in the coefficient of variation of mean cell area for the corneas preserved in K-Sol medium but not for those preserved in MK medium.

Tear film osmolarity and ocular surface disease in two rabbit models for keratoconjunctivitis sicca.

We report the natural history of keratoconjunctivitis sicca (KCS) in two rabbit models. The first one (full KCS model) was created by closing the lacrimal gland excretory duct, and removing the nictitating membrane and harderian gland. We created the second one (lacrimal gland duct only [LGDO]-KCS model) by closing the lacrimal gland excretory duct. Although tear film osmolarity was abnormally high in both models, it was higher in the full KCS model. Decreases in corneal epithelial glycogen and in conjunctival goblet cell density, and morphological abnormalities of the conjunctiva correlated with increases in tear film osmolarity and duration of disease.

Residual bodies in the retinal pigment epithelium induced by intravitreal netilmicin.

Intravitreal injection of aminoglycoside antibiotics is known to induce morphological changes in the retinal pigment epithelium (RPE) resembling a lipidosis. The RPE of netilmicin-treated rabbits displays a dose-related increase in autofluorescence compared to untreated controls. Netilmicin produces an accumulation of membrane-limited osmiophilic lamellated inclusions in the pigment epithelial cell. These inclusions measure from 1 to 3 microns in diameter, and have acid hydrolase activity demonstrated by cytidine monophosphate cytochemistry. These findings suggest that netilmicin-induced inclusions are residual bodies and that the accumulation of these residual bodies is responsible for the observed cellular lipidosis.

Lacrimal gland, cornea, and tear film in the NZB/NZW F1 hybrid mouse.

The NZB/NZW F1 hybrid mouse has been reported to contract a disease similar to Sjögren's syndrome in man. We studied lacrimal gland morphology, corneal morphology, and tear osmolarity in this mouse as a function of age. Lacrimal glands of hybrid mice contained abnormal periductal infiltrates of lymphocytes and plasma cells. Maximum infiltration of the lacrimal gland occurred in the 29-week-old female hybrid mouse and was estimated to involve 12% of the gland, but was insufficient to alter tear osmolarity relative to DBA and Balb/c control mice. Nevertheless, both NZB/NZW F1 hybrid mice and DBA and Balb/c control mice had tear osmolarity and corneal surface morphology similar to that reported for keratoconjunctivitis sicca in man. Although the NZB/NZW F1 hybrid mouse may provide a valuable model for the study of lacrimal gland infiltration, since its tear osmolarity and ocular surface remain normal for a mouse, its usefulness as a model for ocular surface disease in human keratoconjunctivitis sicca may be more limited than previously thought.

Experimental drusen formation induced by intravitreal aminoglycoside injection.

A single intravitreal injection of aminoglycoside antibiotics in the rabbit induces changes in the retinal pigment epithelial layer that consist of disseminated yellow-white dots that are apparent six to ten months after injection. With fluorescein angiography, these dots have the characteristics of window defects or drusen. Histologic examination disclosed subepithelial amorphous material that stained positively with both periodic acid-Schiff and oil red O. Ultrastructural examination disclosed lipidic inclusions in the retinal pigment epithelial cells, basally directed cellular evagination, and basal accumulation of granular material, findings consistent with the reported morphologic features of hard drusen. These findings suggest that aminoglycoside-induced lesions may represent a model for retinal pigment epithelial degeneration and drusen formation.

The influence of aphakia and vitrectomy on experimental retinal toxicity of aminoglycoside antibiotics.

To study the effect of lens and vitreous surgery on the dose threshold of aminoglycoside-induced retinal toxicity, we performed extracapsular lens extraction (Group 1) or lensectomy and vitrectomy (Group 2) on Dutch Belted rabbits. A single dose of amikacin or gentamicin ranging from 100 to 4,000 micrograms was administered intravitreally. Retinal toxicity was examined by light and transmission electron microscopy seven days after injection. Both groups showed retinal toxicity after 400 micrograms of gentamicin or 1,500 micrograms of amikacin, doses identical to those causing toxicity in intact, phakic rabbit eyes. Thus, neither surgical procedure increased the toxic threshold of injected aminoglycoside. Given the accelerated aminoglycoside clearance reported in aphakic eyes, these findings imply that the mechanism of aminoglycoside toxicity may be related to peak drug concentration rather than duration of tissue exposure.

Comparative toxicity of intravitreal aminoglycoside antibiotics.

We compared the toxicity of the aminoglycoside antibiotics (tobramycin, amikacin, netilmicin, and kanamycin) by ophthalmoscopy, light and electron microscopy, and electro-retinography after intravitreal injection in rabbits in doses ranging from 100 to 3,000 micrograms. The earliest manifestations of toxicity were confined to the outer retina with each drug, with lamellar lysosomal inclusions in the retinal pigment epithelium as the earliest finding. However, the aminoglycosides displayed marked differences in the threshold dose required to produce toxic reactions, permitting the following ordering of toxicity: (most toxic) gentamicin greater than netilmicin = tobramycin greater than amikacin = kanamycin (least toxic).

Polymorphonuclear neutrophils delay corneal epithelial wound healing in vitro.

Using an in vitro system for measuring epithelial wound healing, we studied the effect of polymorphonuclear neutrophils (PMNs) and PMN lysate on rat corneal epithelial wound healing. After 22 hr of organ culture, epithelial defects that were originally 3 mm in diameter (7.06 mm2) measured 0.41 mm2 (range, 0.17-0.72 mm2) in control rat corneas, 2.11 mm2 (range, 0.81-3.91 mm2) in corneas incubated in medium containing 5 X 10(6) PMN/ml, and 2.49 mm2 (range, 2.21-2.76 mm2) in corneas incubated with lysate obtained from an equivalent number of PMNs. Scanning and transmission electron microscopy showed similar morphology in the three groups. PMNs selectively adhered to the leading edge of the wound in corneas incubated with PMNs. These data indicate that PMNs and PMN lysate significantly slow corneal epithelial wound healing in vitro.