Comparison of corneal endothelial image analysis by Konan SP8000 noncontact and Bio-Optics Bambi systems.

PURPOSE: Compare corneal endothelial image analysis by Konan SP8000 and Bio-Optics Bambi image-analysis systems. METHODS: Corneal endothelial images from 98 individuals (191 eyes), ranging in age from 4 to 87 years, with a normal slit-lamp examination and no history of ocular trauma, intraocular surgery, or intraocular inflammation were obtained by the Konan SP8000 noncontact specular microscope. One observer analyzed these images by using the Konan system and a second observer by using the Bio-Optics Bambi system. Three methods of analyses were used: a fixed-frame method to obtain cell density (for both Konan and Bio-Optics Bambi) and a "dot" (Konan) or "corners" (Bio-Optics Bambi) method to determine morphometric parameters. RESULTS: The cell density determined by the Konan fixed-frame method was significantly higher (157 cells/mm2) than the Bio-Optics Bambi fixed-frame method determination (p<0.0001). However, the difference in cell density, although still statistically significant, was smaller and reversed comparing the Konan fixed-frame method with both Konan dot and Bio-Optics Bambi comers method (-74 cells/mm2, p<0.0001; -55 cells/mm2, p<0.0001, respectively). Small but statistically significant morphometric analyses differences between Konan and Bio-Optics Bambi were seen: cell density, +19 cells/mm2 (p = 0.03); cell area, -3.0 microm2 (p = 0.008); and coefficient of variation, +1.0 (p = 0.003). There was no statistically significant difference between these two methods in the percentage of six-sided cells detected (p = 0.55). CONCLUSION: Cell densities measured by the Konan fixed-frame method were comparable with Konan and Bio-Optics Bambi's morphometric analysis, but not with the Bio-Optics Bambi fixed-frame method. The two morphometric analyses were comparable with minimal or no differences for the parameters that were studied. The Konan SP8000 endothelial image-analysis system may be useful for large-scale clinical trials determining cell loss; its noncontact system has many clinical benefits (including patient comfort, safety, ease of use, and short procedure time) and provides reliable cell-density calculations.

Non-invasive assessment of the donor corneal endothelium using ocular redox fluorometry.

AIMS: To investigate the usefulness of ocular redox fluorometry for evaluating donor corneal endothelial viability. METHODS: Corneas from 42 recipients of penetrating keratoplasty and four donor corneas were examined by ocular redox fluorometry. Autofluorescence from reduced pyridine nucleotides (PN) and oxidised flavoproteins (Fp) of the human corneal endothelium were measured non-invasively, and the PN/Fp ratio was used as a tissue metabolic indicator. Specular microscopy and electron microscopy were also performed. RESULTS: Both the quality of specular microscopic image and the PN/Fp ratio were significantly correlated with the degree of corneal endothelial damage determined by histological examination. Corneas with poor specular microscopic image showed significantly decreased PN/Fp ratio compared with corneas with good or fair specular images (p = 0.041 and 0.027, respectively). The PN/Fp ratio increased in corneas with mildly damaged endothelium but decreased in corneas with severely damaged endothelium determined by histological examination. Evaluation of corneal endothelium by combination of specular microscopy and ocular redox fluorometry showed excellent association with that of histopathological examination (p < 0.0001). CONCLUSION: Ocular redox fluorometry is useful for assessing donor corneal endothelial viability. Combination of ocular redox fluorometry and specular microscopy may increase the ability of donor cornea selection.

Unification and large-scale structure.

The hypothesis of relativistic flow on parsec scales, coupled with the symmetrical (and therefore subrelativistic) outer structure of extended radio sources, requires that jets decelerate on scales observable with the Very Large Array. The consequences of this idea for the appearances of FRI and FRII radio sources are explored.

Changes of corneal redox state in diabetic animal models.

Metabolic and biochemical changes of the corneal epithelium and endothelium were studied in experimental diabetic animal models. Ocular redox fluorometry was used to noninvasively determine tissue reduction-oxidation (redox) changes in organ cultured rabbit corneas incubated in high glucose concentration media, alloxan-induced diabetic rabbits, and nonobese diabetic mice. The ratio of autofluorescence from reduced pyridine nucleotide to oxidized flavoproteins (PN/Fp) was used as the indicator of the redox state. Chemical assays for NADH and NAD+ were performed on in vitro materials. Analysis of corneal endothelial morphology using specular microscopy was performed to study possible correlations with metabolic changes. Both the PN/Fp and NADH/NAD+ ratios increased in the corneal endothelium under all experimental conditions. Changes in redox state were not observed in the corneal epithelium in any of the models. Morphologic analysis of the corneal endothelium revealed no significant changes. These results indicate that redox changes occur in the diabetic corneal endothelium, but not the corneal epithelium. Ocular redox fluorometry is capable of detecting changes in the corneal endothelial redox state noninvasively.

Epidermal growth factor and insulin use in corneal preservation. Results of a multi-center trial. The Corneal Preservation Study Group.

PURPOSE: The ability of DexSol medium, supplemented with two growth factors, human epidermal growth factor (hEGF) and human insulin, to improve long-term endothelial survival after penetrating keratoplasty was evaluated in a multi-center, randomized, prospective, double-masked clinical trial. METHODS: Donor cornea pairs, one stored in DexSol and the other in DexSol with hEGF (10 ng/ml) and human insulin (10 micrograms/ml) (ProCell), were transplanted into 105 pairs of recipients matched by diagnosis and procedure and followed postoperatively for graft and endothelial survival. RESULTS: No primary donor failures occurred in either group. Graft clarity did not differ between the ProCell and DexSol groups at all postoperative periods: 3 months (98% versus 99%), 6 months (94% versus 98%), and 1 year (95% versus 97%), respectively. Postoperative complications (e.g., glaucoma, rejection) occurred with comparable frequencies in both groups. Mean endothelial cell loss did not significantly differ between the ProCell and DexSol groups at 3 months (5.7% versus 5.1%), 6 months (8.1% versus 10.1%), and 1 year (12.3% versus 15.6%), respectively. Similarly, there were no clinically and statistically significant differences in other endothelial morphometric parameters. CONCLUSIONS: The use of corneas stored in DexSol medium with added hEGF and insulin in corneal transplantation resulted in a safety and efficacy profile comparable with that observed in patients receiving DexSol-stored corneas; however, there were no clinically and statistically significant differences in postoperative endothelial morphometric parameters.

Distribution of autofluorescence in the rabbit corneal epithelium.

Autofluorescence from reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) was measured in order to detect the difference in redox states in rabbit corneal epithelium. The enucleated rabbit eye was mounted in an eye bank eye container with McCarey-Kaufman medium, and the autofluorescence was measured using ocular redox fluorometry as a function of depth. The PN signal distributed evenly whereas the Fp signal was greater in the posterior epithelial region than in the anterior region (p < 0.05). The PN/Fp ratio, a sensitive indicator of tissue redox state, was less in the posterior region. After the application of 1 mM of potassium cyanide in the medium, the ratio increased significantly in each layer (p < 0.001), and the difference between anterior and posterior region diminished. These results indicate that ocular redox fluorometry has the potential to resolve the redox states of the various layers of the corneal epithelium. The posterior region of the epithelium is more active in mitochondrial respiration than the anterior region.

Metabolic and morphologic changes in the corneal endothelium. The effects of potassium cyanide, iodoacetamide, and ouabain.

The metabolic pathways of glycolysis and mitochondrial respiration in the corneal endothelial cell are the primary sources of the adenosine triphosphate (ATP) necessary to maintain endothelial structure and pump fluid to maintain the corneal stroma in its normally dehydrated and transparent state. The correlation between endothelial metabolism and morphology in rabbits was studied for 7 days after the application of three different agents: (1) iodoacetamide, used to inhibit ATP synthesis from both glycolysis and respiration; (2) potassium cyanide (KCN), used to inhibit ATP synthesis from respiration only; and (3) ouabain, used to inhibit fluid pumping but not ATP synthesis. After application of each of these three drugs to the corneal endothelium, changes in endothelial morphology were measured. The greatest change resulted from the use of iodoacetamide. Specular microscopic examination of the endothelium after the application of iodoacetamide showed progressive degradation of the integrity of the cellular structure; after 6 hr, there were no discernible cell borders. In those corneas treated with either KCN or ouabain, only minor changes in the endothelium were seen during the full 7 days of the investigation. Computer-assisted morphometric analysis showed an increase in the coefficient of variation of both cell area and perimeter in all cases. This increase was greater in the corneas treated with ouabain than those treated with either iodoacetamide or KCN. Redox fluorometry showed that the metabolic ratio (autofluorescence of reduced pyridine nucleotides divided by that of oxidized flavoproteins) decreased significantly in the iodoacetamide-treated corneas, increased significantly in the KCN group, and showed no significant change in the corneas in the ouabain group, all compared with a control group. The results showed that (1) when ATP produced by both glycolysis and respiration was inhibited by 0.1 mmol/l iodoacetamide, the endothelial cells could not survive, but (2) when ATP synthesis produced by respiration alone was inhibited by 1.0 mmol/l KCN, the cells could survive for at least 1 wk on the ATP produced by anaerobic glycolysis. Furthermore, the polymegathism seen after application of ouabain, a drug that is not believed to affect ATP synthesis but inhibits the endothelial pump function, is greater than that seen as a result of reduced pump function caused by inhibited respiration produced by 1.0 mmol/l KCN. Combining specular microscopy, computer-assisted morphometric analysis, redox fluorometry, and corneal pachymetry allowed correlations between corneal endothelial metabolism, pump function, and morphology to be studied.

Optisol corneal storage medium.

We compared the safety and efficacy of Optisol (Chiron Ophthalmics, Irvine, California), a new corneal storage medium, with McCarey-Kaufman and Dexsol corneal storage media (Chiron Ophthalmics, Irvine, California) and K-Sol corneal storage medium (Cilco, Huntington, West Virginia). Optisol contains dextran, 2.5% chondroitin sulfate, vitamins, and precursors of adenosine triphosphate (adenosine, inosine, and adenine). In in vitro studies, rabbit and human corneas stored in Optisol were thinner after 12 to 14 days at 4 C than tissue stored in other media. Thymidine uptake showed increased mitotic activity in human corneal endothelial cells cultured in Optisol, compared to Dexsol. Specular microscopic fields showed larger are-as of visibly intact endothelial cells and ultrastructural examination disclosed fewer structural changes in endothelial cells stored in Optisol, compared to tissue stored in Dexsol. In vivo, no clinical signs of epithelial toxicity or histologic evidence of intraocular inflammation were observed in rabbit eyes in which Optisol drops were instilled four times a day for 14 days. Finally, 51 patients undergoing penetrating keratoplasty with tissue stored in Optisol for one to six days (mean, 3.6 days) were enrolled in an uncontrolled, open-label clinical study. The percentage of clear grafts (93%, 41 of 44 patients examined at three months; and 98%, 42 of 43 patients examined at six months) and endothelial cell loss (5.0% and 11.5% at three and six months, respectively) were comparable to data from previous studies that used tissue stored in other short-term and intermediate-term media. The results suggest that Optisol storage preserves corneal endothelial cells for up to two weeks at 4 C, thereby permitting flexibility in the use of donor tissue for corneal transplantation, and that Optisol storage yields thinner tissue, which may allow for more accurate evaluation and more effective surgical manipulation.

[Metabolic analysis of the diseased human corneal endothelium].

Redox states of the corneal endothelium in 42 recipient corneas obtained at the penetrating keratoplasty were measured non-invasively using ocular redox fluorometry. Autofluorescence from reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) were measured, and the PN/Fp ratio was used as an indicator of the redox state. Endothelial damage was graded as normal, mildly damaged, moderately damaged, and severely damaged, based on the histopathological findings. Mildly damaged endothelium showed a significantly higher PN/Fp ratio than the that in normal endothelium whereas the ratio was significantly lower in the severely damaged endothelium. These changes in the redox state may represent compensation and decompensation processes of the endothelial metabolism. Ocular redox fluorometry was shown to be useful for the evaluation of the metabolic state in the human corneal endothelium.

Metabolic changes in the corneal epithelium resulting from hard contact lens wear.

The metabolic state of rabbit corneas was monitored in vivo using the noninvasive method of corneal redox fluorometry. The autofluorescence signals of reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) were measured in the corneal epithelium with and without contact lens wear. The PN/Fp ratio, which is related to the metabolic status of the tissue, was then calculated for each of these conditions. After application of polymethylmethacrylate (PMMA) contact lenses having an oxygen transmissibility (Dk) of less than 0.1, the PN signal increased and the Fp signal decreased. The PN/Fp ratio, generally a more precise indicator of metabolic state than either of these two quantities alone, was 1.93 +/- 0.78 without contact lenses, and increased to 2.78 +/- 0.86 (p less than 0.0001) with contact lenses. When oxygen-permeable silicon contact lenses (Dk = 12.5) were placed on the corneas, the PN/Fp ratio was found to increase slightly, but not as much as with the PMMA lenses. Newly developed highly oxygen-permeable contact lenses (Dk = 58.8) did not increase this ratio. Our findings indicate that redox fluorometry can be valuable in determining the effects of contact lens wear on corneal metabolism.

[Noninvasive metabolic analysis of the diabetic cornea and lens: in vivo measurement].

In vivo measurement of metabolic changes in diabetic cornea and lens were performed using redox fluorometry in nonobese diabetic (NOD) mice. Autofluorescence from reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) were measured, and the PN/Fp ratio was used as a tissue metabolism indicator. The PN/Fp ratios were significantly higher in the diabetic corneal endothelium. Morphometric analysis of the corneal endothelium using specular microscopy revealed no significant differences between the two groups. These results indicate that redox fluorometry is able to detect early metabolic changes in the corneal endothelium and lens epithelium, which are induced by diabetes mellitus. Activation of the polyol pathway may be responsible for the change. Corneal epithelia may be less susceptible to diabetic changes than the corneal endothelium and lens epithelium.

Glycolytic oscillation and effect of metabolic inhibitor on rat lens.

Abnormalities in glucose metabolism are thought to be among the main causes of cataract formation. We have taken noninvasive biochemical measurements of the lens which provides us with information concerning glucose metabolism in the lens epithelium. The autofluorescence of reduced pyridine nucleotides (PN) and oxidized flavoproteins (Fp) in the rat lens epithelium was measured noninvasively as a function of time using redox fluorometry. The oscillations of the metabolic ratio, PN/Fp, were measured in vivo, in situ, and in the organ-cultured lens. The PN/Fp ratio in the organ-cultured lens ranged from 1.05 to 2.57 within a period of 60-90 minutes (mean +/- SD = 1.52 +/- 0.36). This PN/Fp ratio increased by 23% when a respiratory inhibitor (8 mM KCN) was applied to the lens. However, it decreased by 10% in the presence of a complete metabolic inhibitor (8 mM iodoacetamide). The presence of metabolic oscillations in the in vivo, in situ and cultured lens indicates that this oscillation is a local phenomenon. In cell-free extract systems, oscillations of several intermediates in the glycolytic pathway have been previously demonstrated and this PN/Fp oscillation is thought to be a reflection of glycolytic oscillation.

[Observation of cellular communication in the corneal endothelium using a small molecular weight fluorescent dye].

The metabolic and functional connection between corneal endothelial cells is thought to depend on the transfer of small molecules via gap junctions, as has been reported for other tissues. Cell-to-cell communication in the corneal endothelium was studied by monitoring the spread of fluorescence following the direct injection of Lucifer Yellow CH into single endothelial cells of the excised rabbit cornea. The image of the endothelial cells was observed during the injection and post-injection periods using specular microscopy and fluorescence microscopy. In the fresh cornea, the dye transferred readily from the injected endothelial cell to its neighbors. In damaged cells, dye transfer was slower or did not occur. After 5 hours of incubation in tissue culture medium 199, the specular microscopic image degraded, lowered cell membrane potential and decreased dye transfer rate were measured. This study showed that in normal corneal endothelial cells there were efficient cell communication channels between neighboring cells, and as the endothelium becomes less viable, cell communication is inhibited.

Effects of intracameral injection of viscoelastic solutions on corneal endothelium in dogs.

Contact in vivo wide-field specular microscopy was performed on right eyes of 20 healthy dogs after sodium hyaluronate (1%, n = 5), sodium chondroitin sulfate (4%) and sodium hyaluronate (3%, n = 5), hydroxypropyl methylcellulose (2%, n = 5), or balanced salt solution (control, n = 5) was injected into the anterior chamber. Using computerized morphometric analysis and pachymetry, changes in endothelial cell density, cell morphologic features, and corneal thickness from baseline values were evaluated at postinjection hour (PIH) 72 and PIH 168. Changes were not seen in endothelial cell density or cell morphologic features in any treated eye. The mean corneal thickness of all treated eyes at PIH 72 increased 6%, significantly greater than that of the nontreated eyes (P = 0.03). Mean corneal thickness of treated and nontreated eyes was similar at baseline and PIH 168 in all treatment groups.

Correlation of redox fluorometry and analytical measurements of pyridine nucleotide.

Pyridine nucleotide levels in the corneal epithelium were measured using redox fluorometry, a noninvasive method for monitoring the metabolic status of corneal tissue, and a sensitive bioluminescent assay and an improved extraction procedure that allows the simultaneous extraction and measurement of both NADH and NAD. The same corneas were measured using each of the two methods to enable comparison of the results. The NADH/(NADH + NAD) fraction in the normal epithelium measured by the bioluminescent assay was 0.14 +/- 0.06. Incubation of corneas in 1 mM potassium cyanide (KCN) to mimic the anoxic state increased the NADH/(NADH + NAD) fraction significantly to 0.24 +/- 0.03 (P less than 0.001). The autofluorescence from reduced pyridine nucleotide measured by redox fluorometry also increased with KCN from 2840 +/- 605 to 5147 +/- 738 (P less than 0.0001). A plot of the fluorescence and analytical data for each cornea showed a positive correlation between the two methods, with a correlation coefficient (r value) of 0.80. The correlation was improved but was not dependent on the high values of the KCN treated corneas; an r value of 0.73 was obtained for the non-KCN treated corneas alone. Additional measurements of the temperature dependence of the fluorescence intensity of an NADH solution and the cornea gave a decrease in intensity of 17% from 25 degrees C to 35 degrees C for the NADH solution and 11% (P = 0.0004) for the reduced pyridine nucleotide fluorescence in the cornea over the same temperature range.

Lens redox fluorometry: pyridine nucleotide fluorescence and analysis of diabetic lens.

We performed both ex vivo and in vivo fluorometric analyses of pyridine nucleotides (PN) in rabbit and rat lenses. Rabbit lens PN fluorescence (99% NADH) was found to have an excitation maximum at 366 nm and an emission maximum at 462 nm (366:462). The only other fluorescent chromophore in that region of the spectrum has excitation and emission peaks at 328 and 460 nm, respectively. Anaerobic glycolysis in the lens was stimulated by KCN, a known inhibitor of mitochondrial respiration, after which a time-study of fluorescence intensities was performed. Over the course of a 3.5 hr period following treatment with KCN, the PN signal showed a statistically significant increase relative to that in the control lenses (those treated with KCl). while the 328:460 signal (which may be due to some protein involved in energy transfer with the PN) had a significantly greater decrease. We also found that fluorescence intensity of NADH in solution is linearly proportional to physiologic-range concentration. Moreover, there was a close correlation between fluorescence intensity of rat lens PN as measured on a specular microscope-coupled redox fluorometer capable of in vivo use, and the lens PN levels as determined by the analytical cycling assay technique. This fluorometer was then employed to assess the redox state in rats with streptozotocin-induced diabetes. The normalized ratio of PN to flavoproteins (Fp) in the lens epithelium increased from 0.96 +/- 0.12 in the normal state to 1.48 +/- 0.30 2 weeks after diabetes induction. In contrast, the ratio in the diabetic lens treated with an aldose reductase inhibitor, sorbinil, did not increase. The increase in the PN:Fp ratio therefore reflects activation of the polyol pathway and its associated metabolic activities, which results in an increase in the NADH:NAD ratio in the diabetic rat lenses. Our results indicate that the non-invasive, real-time method of redox fluorometry may be useful in the early detection and evaluation of cataracts and other disorders in lens metabolism, long before opacities occur. It can be used to monitor the disease process and evaluate the efficacy of such drugs as aldose reductase inhibitors on a biochemical level.

Effects of EGF and indomethacin on rabbit corneal endothelial wound closure in excised corneas.

Corneal endothelial cells of rabbit corneas stored in M-K medium at 37 degrees C were wounded by touching them lightly with a micropipet under video specular microscope observation. Three groups were studied: control, with EGF, and with EGF + indomethacin. The wound closure process (initial wound area about 8500 microns 2) was observed and recorded with time-lapse videography for 6 hr. The recorded video images were digitized and computer assisted morphometric analysis was performed. (1) Addition of either EGF (10 ng/ml) + indomethacin (1 microM), or EGF (10 ng/ml) alone to the M-K medium statistically significantly shortened the wound closure time as compared with the control group. (2) Both EGF + indomethacin and EGF alone resulted in a greater average percent relative change of the shape factor, more than three times greater with EGF + indomethacin and more than two times greater with EGF alone, than in the control group 150 min after wounding. (3) The maximum cell shape change occurred at about 150 min after wounding in the groups EGF + indomethacin and EGF alone, and at about 200 min in the control group. After this time in all three groups the cells began to approach a normal shape. (4) The cells near the wound boundary moved faster in the EGF + indomethacin and the EGF groups as compared with the control group. These results suggest that EGF and indomethacin may be of therapeutic value in promoting closure of traumatized human corneal endothelium.

Thermal cycling effects on the stored rabbit cornea.

Specular microscopy has proven itself a useful tool for evaluation of donor corneas. In many eye banks, corneas are stored at 4 degree C and warmed to room temperature for specular microscopic evaluation. On occasion it would be desirable to put the cornea back into storage provided there were no detrimental effects of the recooling and subsequent rewarming. The effects of repeated cooling and rewarming (thermal cycling) on the corneal endothelium were determined using rabbit corneas stored in M-K medium at 4 degree C. Some corneas were left in the refrigerator for a 7-day storage interval while others were removed daily, warmed to room temperature, evaluated morphologically and biochemically with the specular microscope and the redox fluorometer, respectively, and then placed back in the refrigerator. At the end of the 7-day storage period there were no statistically significant differences in either the biochemical signals or the specular appearance between thermally cycled corneas and corneas that were continuously stored at 4 degrees C for the same period of time. Repeated warming, specular microscopic observation, and cooling of the cornea appear not to be detrimental to the corneal endothelium.

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.

Corneal endothelial changes following minor trauma.

The healing processes that occur when corneal endothelial cells are subjected to only mild trauma are not known. To study these processes we have developed a system that enables only a few endothelial cells to be traumatized or destroyed under continuous specular microscopic observation in vitro. Experiments in which a small group of cells were traumatized by gentle wounding with a microglass tip produced an immediate and distinct dark area having the same size as the tip. Using this method we have produced and observed two types of wound by controlling the force of wounding. The first type of wound, produced by a gentle touch, recovered within 1 hr. The second type of wound, produced by moderate touch, took about 24 hr to recover completely from the trauma. In the second type of wound, we observed migration, elongation, coalescence and mitosis during the healing process. Histological examination revealed that in the first type of wound, the cells remained intact with no apparent damage seen by vital staining and light microscopy. For the second type of wound, the cells were completely missing although there was no apparent damage to Descemet's membrane.