Neutralizing antibody to TGFbeta modulates stromal fibrosis but not regression of photoablative effect following PRK

PURPOSE

Previous studies have suggested that corneal fibrosis controlled by a TGFbeta-mediated cytocrine pathway underlies the development of clinical corneal haze and associated regression of photoablative effect following excimer laser PRK. Using a unique blocking antibody, we evaluated the role of TGFbeta in post-PRK corneal wound healing as measured by in vivo Confocal Microscopy Through Focusing (CMTF).

METHODS

Twelve rabbits received a monocular, 6-mm diameter, 9.0 D PRK myopic correction. Six animals received 50 microg of anti-TGFbeta blocking antibodies applied topically 3x/day for three days post-PRK, while six animals received vehicle alone. An additional six animals served as unoperated controls. At various times during a four-month-period, animals were evaluated using CMTF, which generates a quantitative image intensity depth profile of the cornea. The location and reflectivity of corneal structures were identified from CMTF-profiles and used to determine epithelial and stromal thickness and corneal light reflectivity as an objective estimate of corneal haze. To correlate in vivo and ex vivo morphology, an additional six rabbits were analyzed at differing temporal intervals post-PRK for the expression and cytoskeletal organization of contractile microfilaments: f-actin (stress fibers) and alpha-smooth muscle actin (a molecular marker for myofibroblast transformation).

RESULTS

Anti-TGFbeta treated corneas showed significantly less CMTF measured light reflectivity (ANOVA, p < 0.02) following PRK compared to vehicle treated corneas with a 34% decrease at two weeks (2513 +/- 758 U compared to 3810 +/- 1262 U) and a 61% reduction in reflectivity at four months (447 +/- 208 U compared to 1154 +/- 585 U). The reduction in early development of light reflecting structures and the more rapid decline appeared related to anti-TGFbeta-mediated inhibition of keratocyte activation and proliferation, myofibroblast transformation, and stromal fibrosis. Between anti-TGFbeta and vehicle treated corneas, no significant differences were detected in either photoablation depth (126 +/- 9 microm versus 126 +/- 7 microm) or regression of photoablative effect (postoperative stromal thickening at four months: 95 +/- 16 microm versus 95 +/- 10 microm). Histologic examination demonstrated that regression of photoablative effect in anti-TGFbeta treated corneas was related entirely to regeneration by corneal growth underlying the photoablated stromal surface. In vehicle treated corneas, fibrosis or deposition of new fibrotic tissue above the photoablated stromal surface was observed but contributed only about 25% of the total postoperative stromal thickening. No epithelial hyperplasia was detected. In unoperated control animals, a physiologic stromal thickening of 5 +/- 2 microm per month (p < 0.001) was observed.

CONCLUSIONS

This study confirms our earlier observations that increased corneal light reflectivity following PRK is predominantly due to: (1) distortion of the photoablated stromal surface leading to prominent reflections; and (2) increased reflections from activated and transformed keratocytes. Anti-TGFbeta reduced keratocyte activation and transformation and inhibited stromal fibrosis, leading to a reduction in early light reflectivity as well as to a more rapid decline. Of greatest interest is the unexpected finding that anti-TGFbeta treatment inhibited stromal fibrosis without reducing or delaying post-PRK stromal re-thickening. Based on these findings we propose that corneal thickness may be tightly and dynamically regulated by an unknown, non-TGFbeta mediated pathway. We propose that anti-TGFbeta treatment may be useful in reducing post-PRK corneal haze development in patients by: (1) inhibiting the recruitment of highly reflective, activated keratocytes, (2) inhibiting myofibroblast transformation, and 3) reducing stromal fibrosis.

Area and depth of surfactant-induced corneal injury correlates with cell death

PURPOSE

In previous studies in which in vivo confocal microscopy (CM) was used, quantifiable differences were identified in the corneal epithelium and stroma for surfactants producing different degrees of ocular irritation. In the present study, in vivo confocal microscopy was used to determine area and depth of the initial corneal changes, and the correlation of the data to cell death was characterized by ex vivo live-dead assay.

METHODS

In four groups of rabbits (12 animals each), 10 microl surfactants known to produce slight, mild, moderate, or severe irritation was applied to the central cornea of one eye; 4 untreated rabbits served as controls. Measurements of group total mean epithelial thickness, epithelial cell area, and depth of keratocyte loss in four corneal regions were made by in vivo CM in 6 rabbits of each group and in 4 control animals at 3 hours and in the remaining rabbits at 3 hours and 1 day. Corneas were then removed and fixed for conventional histologic examination (two eyes/treatment/group), or regions were excised and placed in culture media containing 2 microM calcein-acetoxymethyl ester (calcein-AM) and 4 microM ethidium homodimer. Using laser scanning CM, the number of dead epithelial or stromal cells in a 300 x 300 x 170 microm (in the x, y, and z axes, respectively) volume of the cornea was determined.

RESULTS

Confocal microscopy showed that application of the slight irritant resulted in decreased epithelial thickness at 3 hours (41.2+/-2.6 microm in treated eyes versus 43.6+/-3 microm in control eyes; n=6 and 4, respectively) and a significant decrease (P < 0.001) in epithelial cell size (630+/-203 microm2 versus 1427.2+/-90.7 microm2). On day 1, mild, moderate, and severe irritants caused complete loss of epithelium and disappearance of keratocytes to a depth of 30.8+/-10.7 microm, 47.2+/-10.4 microm, and 764.6+/-159.6 microm (n=6, 5, and 6), respectively. At 3 hours, live-dead assay detected more dead epithelial cells as a percentage of total surface cells (49.2+/-4.5% in slightly irritated eyes versus 20.9+/-3.2% in control eyes), significantly correlating with the measurement by in vivo CM of average epithelial cell size in each eye (r=-0.96; P < 0.005). On day 1, mild and moderate irritants showed increasing stromal cell death from 9.8+/-16.2 cells to 36.4+/-17.7 cells, which significantly correlated with the depth of stromal injury determined by in vivo CM (r=0.79; P < 0.00001). No surviving keratocytes were detected in severely irritated eyes.

CONCLUSIONS

The data support the hypothesis that differences in surfactant-induced ocular irritation are directly related to area and depth of acute corneal injury.

Assessment of stress fiber orientation during healing of radial keratotomy wounds using confocal microscopy

Radial keratotomy (RK) is a refractive surgical procedure in which partial thickness incisions are made in the cornea in order to alter its shape. Previous studies suggest that RK wounds undergo changes in wound gape in response to the ingrowth of myofibroblasts which mediate subsequent wound contraction and may modulate changes in corneal curvature seen after RK. A recent quantitative analysis of f-actin organization in full-thickness incisional wounds in the rabbit demonstrated that microfilament bundles (stress fibers) present in myofibroblasts align parallel to the long axis of the wound during wound contraction. To investigate whether the same pattern of alignment occurs after RK, a similar analysis of f-actin organization was undertaken using the cat RK model. Radial keratotomy wounds were studied from 10 to 28 days after surgery using en block staining with fluorescein isothiocyanate (FITC) phalloidin, and three-dimensional (3-D) datasets (z-series of en face optical sections) were collected using laser confocal microscopy at various regions within the wound. In addition, conventional en face sections were double-labeled using combinations of phalloidin and antibodies to fibronectin and alpha 5 beta 1 integrin. Myofibroblast ingrowth started in the bottom of the wound and progressed anteriorly. At 10 to 14 days, f-actin was predominantly distributed in long, thick bundles (stress fibers) within the wound. These fibers appeared to be randomly oriented anteriorly, but became progressively more aligned with the long axis of the wound posteriorly. At 21 days, the stress fibers were predominantly oriented parallel to the long axis of the wound at all levels. F-actin, fibronectin and integrin were coaligned at both the 14 and 21 day time points. Since the majority of wound closure occurs between 14 and 28 days after surgery, we conclude that parallel alignment of the actin filament-fibronectin-integrin assembly in the cat RK model is associated with wound contraction.

Confocal microscopic characterization of wound repair after photorefractive keratectomy

PURPOSE

Development of postoperative corneal haze and regression of refractive effect are unfavorable clinical complications of excimer laser photorefractive keratectomy (PRK). Although exact mechanisms remain to be elucidated, these outcomes have been attributed to post-PRK corneal wound healing. The purpose of this study was to evaluate corneal wound repair quantitatively after PRK in a rabbit model using a newly developed in vivo technique, termed confocal microscopy through focusing (CMTF).

METHODS

Twelve rabbit corneas received a monocular, 6-mm diameter, 9.0-diopter PRK myopic correction. Animals were evaluated sequentially up to 6 months after surgery by in vivo CMTF, which uses an image-intensity depth profile to measure epithelial and stromal thickness and uses corneal light reflectivity as an objective estimate of corneal haze. At differing temporal intervals, in vivo morphology was correlated with ex vivo histology using fluorescence microscopy.

RESULTS

One week after PRK, an acellular layer of 86 +/- 24 microns was found anteriorly in the remaining stroma, which demonstrated surgically induced keratocyte death. Underlying keratocytes became activated and migrated toward the wound bed; repopulation was completed within 3 weeks. One week after PRK, there was a significant increase (P < 0.001) in light reflections detected from the photoablated stromal surface (1745 +/- 262 U) and from the underlying activated fibroblasts (713 +/- 607 U). Corneal reflectivity peaked at 3 weeks (4648 +/- 1263 U) and decreased linearly to 889 +/- 700 U by 6 months after the PRK; this corresponded to a reflectivity six times greater than the level seen in unoperated corneas. Two weeks after PRK, initial corneal edema had resolved, revealing an actual ablation depth (maximal stromal thinning) of 118 +/- 8 microns. Starting at 2 weeks after surgery, the stroma underwent gradual rethickening that reached 98% of the preoperative thickness at 6 months after PRK; at that time, only 6% of the initial photoablation depth persisted. By contrast, the central corneal epithelium showed no significant postoperative hyperplasia.

CONCLUSIONS

Rabbit corneas treated by PRK showed a remarkable stromal wound-healing response that ultimately led to the restoration of the original stromal thickness by 6 months after surgery, demonstrating complete regression of the initial photoablative effect. Additionally, corneal wound healing was associated with increased light reflections from both the photoablated stromal surface and the activated wound-healing keratocytes underlying this area. Based on these findings, the authors hypothesize that the development of clinically observed corneal haze in PRK patients may be related, in part, to activation of corneal keratocytes and to putative changes in the extracellular matrix.

An in vitro force measurement assay to study the early mechanical interaction between corneal fibroblasts and collagen matrix

An in vitro force measurement assay has been developed to quantify the forces exerted by single corneal fibroblasts during the early interaction with a collagen matrix. Corneal fibroblasts were sparsely seeded on top of collagen matrices whose stiffness was predetermined by micromanipulation with calibrated fine glass microneedles. The forces exerted by individual cells were calculated from time-lapse videomicroscopic recordings of the 2-D elastic distortion of the matrix. In additional experiments, the degree of permanent reorganization of the collagen matrices was assessed by lysing the cells with 1% Triton X-100 solution at the end of a 2-hour incubation and recording the subsequent relaxation. The data suggest that a cell can exert comparable centripetal force during either extension of a cell process or partial retraction of an extended pseudopodia. The rates of force associated with pseudopodial extension and partial retraction were 0.180 +/- 0.091 (x 10(-8)) N/min (n = 8 experiments) and 0.213 +/- 0.063 (x 10(-8)) N/min (n = 8 experiments), respectively. Rupture of pseudopodial adhesion associated with cell locomotion causes a release of force on the matrix and a complete recoil of the pseudopodia concerned; a simultaneous release of force on the matrix was also observed at the opposite end of the cell. Lysis of cells resulted in 84 +/- 18% relaxation of the matrix, suggesting that little permanent remodeling of matrix is produced by the actions of isolated migrating cells.

Confocal microscopic characterization of initial corneal changes of surfactant-induced eye irritation in the rabbit

We have previously demonstrated with slightly and severely irritating surfactants that the new technology of noninvasive, in vivo confocal microscopy (CM) can be a useful approach to a better understanding of the pathobiology of ocular irritation in situ. In this study, in vivo CM was used to qualitatively and quantitatively characterize the initial microscopic corneal changes occurring with surfactants of slight, mild, moderate, and severe irritation. Surfactants were directly applied to the corneas of rabbits (6/group) at a dose of 10 microl. Eyes and eyelids were examined macroscopically and scored for irritation beginning at 3 hr after dosing and periodically through Day 35. Concurrently, the corneas were evaluated by in vivo CM; 3D data sets extending from the surface epithelium to the endothelium were assessed for surface epithelial cell size, epithelial layer thickness, total corneal thickness, and depth of keratocyte necrosis. The average macroscopic scores at 3 hr for the slight, mild, moderate, and severe irritants were 6.0, 39.3, 48.5, and 68.7, respectively, of a possible 110. At 3 hr, in vivo CM revealed corneal injury with the slight irritant limited to the epithelium, resulting in reductions in epithelial cell size and thickness to 59.0 and 82.4% of controls (p < 0.001 and p < 0.01, respectively). These parameters returned to normal by Day 3. For the mild irritant, at 3 hr the epithelium was absent, corneal thickness was increased to 157.6% of controls (p < 0.001), and necrosis of keratocytes extended to an average depth of 4.3 microm (0.8% of the corneal thickness); these parameters were essentially normal by Day 14. For the moderate irritant, at 3 hr the epithelium was markedly attenuated, corneal thickness was increased to 155.8% of controls (p < 0.001), and keratocyte necrosis extended to an average depth of 19.0 microm (3.6% of corneal thickness; statistically greater than with the mild irritant, p < 0.001); these parameters were essentially normal by Day 14. For the severe irritant, at 3 hr the epithelium was significantly thinned, corneal thickness was increased to 165.9% of controls (p < 0.001), and keratocyte necrosis occurred to an average depth of 391.1 microm (70.1% of corneal thickness). These findings demonstrate that significant differences in area and depth of injury occur with surfactants of differing irritancy. The data suggest that differences at 3 hr can be used to distinguish different levels of ocular irritation. Data such as these will be important in the development and evaluation of future mechanistically based in vitro alternatives for ocular irritancy testing.

Inhibition of corneal fibrosis by topical application of blocking antibodies to TGF beta in the rabbit

Previous studies have shown that TGF beta 1 induces activation and myofibroblast transformation of cultured rabbit corneal keratocytes. To determine whether TGF beta has a similar function in vivo, we evaluated the effect of TGF beta-blocking antibodies on corneal fibrosis after lamellar keratectomy (LK) in the rabbit. A total of 51 rabbits received standard LK wounds, and eyes were treated with 50 microliters of Celluvisc/PBS, containing 10, 50, or 100 micrograms of 1D11, a mouse monoclonal anti-TGF beta-blocking antibody. Control wounds received either 100 micrograms of an irrelevant mouse monoclonal antibody or vehicle alone. At days 14, 28, 42, and 56, eyes were evaluated by in vivo confocal microscopy (CM) and the mice were killed for light microscopy (LM) and immunostaining with antibodies to human fibronectin. In vivo CM of LK wounds clearly identified a disorganized layer that contained irregularly arranged fibroblasts and reflective extracellular matrix overlying normal corneal stroma. In a subset of 11 eyes stained with 5-(4,6-dichlorotriazinyl) aminofluorescein (DTAF) immediately after injury, the thickness of the disorganized layer identified by in vivo CM significantly correlated with both anterior corneal fibrosis (r = 0.627; p < 0.025) and depth of keratocyte activation (r = 0.8980; p < 0.0005), indicating that in vivo CM can be used quantitatively to assess anterior stromal fibrosis. In eyes treated with an irrelevant monoclonal antibody, in vivo corneal fibrosis averaged 100 +/- 26 microns thick at day 14, whereas treatment with 10, 50, and 100 micrograms anti-TGF beta significantly reduced (p < 0.0005) the anterior disorganization in a dose-dependent fashion to 101 +/- 32, 45 +/- 11, and 56 +/- 18 microns, respectively. Semiquantitative measurement of anti-fibronectin staining within the wound revealed that anti-TGF beta significantly reduced the intensity of anti-fibronectin staining in the anterior 50 microns of the corneal stroma (p < 0.003). These findings indicate that TGF beta plays an important in vivo role in keratocyte activation and myofibroblast transformation. Furthermore, the in vivo use of TGF beta-blocking antibody effects may allow modulation of corneal fibrosis after refractive surgery.

Quantification of stromal thinning, epithelial thickness, and corneal haze after photorefractive keratectomy using in vivo confocal microscopy

PURPOSE

The authors establish, for the first time, observer-independent quantification of stromal thinning, epithelial thickness, and corneal haze after excimer laser photorefractive keratectomy (PRK) using a unique, new form of in vivo confocal microscopy.

METHODS

Rapid, continuous z-scans of high-resolution confocal images, termed confocal microscopy through focusing (CMTF), were performed in the central corneal area of 17 patients before and 1 month after PRK for low- to moderate-grade myopia (-2.88-9.13 diopters [D]). Corneal, epithelial, and stromal thickness measurements and an objective haze estimate were obtained from each CMTF scan by digital image analysis.

RESULTS

Epithelial thickness averaged 51 +/- 4 microns before and 45 +/- 10 microns 1 month post-PRK (P < 0.005), whereas stromal thinning ranged from 20 to 154 microns, representing a direct estimate of the actual photoablation depth. Corneal thickness averaged 560 +/- 36 microns before PRK and 462 +/- 52 microns at 1 month. The change in corneal thickness correlated closely with the change in spherical equivalent refraction (r = 0.94, P < 0.0001); linear regression analysis revealed a value of 14.3 microns corneal thinning per diopter of correction. A significant correlation was found between the objective CMTF haze estimate and a clinical haze grading obtained by slit-lamp examination (r = 0.73, P < 0.001).

CONCLUSIONS

Confocal microscopy through focusing is a new, powerful in vivo tool that enables quantitative, unbiased evaluation of PRK procedures over time by providing epithelial and stromal thickness analysis, photoablation depth assessment, and unbiased haze measurement. The method is uniquely valuable in the pre- and postoperative assessment of PRK patients and for determining the optimal treatment strategy, especially in assessing refractive and visual outcomes in individual cases.

Epithelial and corneal thickness measurements by in vivo confocal microscopy through focusing (CMTF)

PURPOSE

To study the feasibility of measuring total corneal thickness, as well as the thickness of the epithelium and Bowman's layer, using a novel in vivo confocal microscopy through-focusing (CMTF) methodology.

METHODS

The central cornea was scanned from the epithelium to endothelium at an average focal plane speed of 32 microns/sec for rabbits, and 64 microns/sec for humans. Scans were initially video-recorded and later digitized. From digital images, CMTF intensity curves were generated by calculating the average pixel intensity in the central 180 x 180 pixel region (285 microns x 285 microns) of each image in the scan, and plotting as a function of z-depth. Peaks in this intensity profile were then empirically correlated to unique corneal layers using a program which interactively displayed images corresponding to the mouse cursor position along the intensity profile curve. Sublayer thickness values were then calculated from the z-axis positions of the relevant peaks in the intensity curve. Ten normal rabbits and seven human volunteers were evaluated in the study. Both CMTF and ultrasonic pachymetry (UP) measurements were performed on rabbit eyes to determine the agreement between CMTF and UP.

RESULTS

Distinct epithelial, basal lamina, and endothelial peaks were identified for all 10 rabbit eyes. The mean central corneal thickness in the rabbit was 381.6 +/- 27.3 microns by CMTF and 384.4 +/- 28.7 microns by UP. The mean difference in central corneal thickness between CMTF and UP was -2.8 +/- 7.1 microns which was not statistically significant (p > 0.2 by paired t-test). Central epithelial thickness in the rabbit measured by CMTF was 47.7 +/- 2.2 microns. The average coefficients of variation for repeated scans were 2.5% and 0.7% for epithelial and corneal thickness, respectively. The standard errors for both epithelial and corneal thickness were less than 1.5 microns for all rabbits. The reproducibilities for epithelial and corneal thickness measurements were 2.2 microns and 2.6 microns, respectively, calculated as the square root of the within group variances of One-Way ANOVA. Intensity profiles for human corneas showed strong epithelial and endothelial peaks, as well as smaller peaks corresponding to the basal-epithelial nerve plexus and the denser anterior layer of stromal keratocyte nuclei. The mean central corneal thickness in the human was 532.1 +/- 18.8 microns; central epithelial thickness was 50.6 +/- 3.9 microns; central Bowman's layer thickness was 16.6 +/- 1.1 microns. The average coefficients of variation for repeated scans were 5.9%, 13.2%, and 1.6% for epithelial, Bowman's layer, and corneal thickness, respectively. The standard errors for all measurements were less than 2.4 microns. The reproducibilities for epithelial, Bowman's layer, and corneal thickness measurements were 3.2 microns, 2.3 microns, and 10.0 microns, respectively.

CONCLUSIONS

CMTF is a novel, reproducible technique for obtaining epithelial and corneal thickness measurements during clinical in vivo confocal microscopy of the cornea. More importantly, this methodology provides the first objective, quantitative approach for measurement and analysis of depth and thickness of corneal sub-layers which may prove uniquely valuable in temporally assessing corneal function.

ZO-1 reorganization and myofibroblast transformation of corneal endothelial cells after freeze injury in the cat

Corneal endothelial wound healing following scrape injury in the rabbit and cat is characterized by cell spreading and maintenance of a normal endothelial phenotype consisting of apically-localized, circumferential microfilament bands and cell border-associated ZO-1, a tight junction protein and marker for endothelial differentiation. In contrast, after freeze injury in the rat and rabbit endothelial cells develop basally organized microfilament bundles (stress fibers), and appear to proliferate and form a multilayered zone at the wound margin. The purpose of the present study was to determine if similar phenotypic changes are observed after freeze injury in the cat corneal endothelium, which like human, normally has limited growth potential. In addition, changes in ZO-1 and alpha-smooth muscle actin (a marker for myofibroblast transformation) distribution were evaluated for the first time following freeze injury. In vivo endothelial healing of standard 3 mm diameter freeze injury was evaluated at 4 hr, 12 hr, 24 hr, 48 hr, 3 days and 5 days after injury in 22 cat eyes. Corneas were stained with phalloidin, propidium iodide, and anti-ZO-1, anti-alpha-smooth muscle-specific actin or anti-fibronectin antibodies. Protein organization was then evaluated using immunofluorescence and laser scanning confocal microscopy. Beginning at 12 hr after injury, endothelial cells appeared to extend and elongate over the wound area. By 48 hr after injury, migrating endothelial cells formed a multilayered activated zone (AZ) at the wound margin. Endothelial cells immediately adjacent to the AZ maintained a normal circumferential organization of f-actin colocalized with cell border-associated anti-ZO-1 staining at all time points observed. However, within the AZ there was an abrupt increase in phalloidin staining and development of prominent microfilament bundles (stress fibers), as well as a loss of normal anti-ZO-1 staining. The AZ also stained positively for anti-alpha-smooth muscle actin and anti-fibronectin antibodies. Changes in the distribution of ZO-1 were observed as early as 4 hr after injury, and appeared to precede f-actin reorganization. These data indicate that endothelial healing after freeze injury in the cat involves a loss of normal endothelial differentiation and cell connectivity, and transformation to a myofibroblastic phenotype.

Collagen fiber architecture of a cultured dermal tissue

Advances in tissue engineering have led to the development of artificially grown dermal tissues for use in burn and ulcer treatments. An example of such an engineered tissue is Dermagraft, which is grown using human neonatal fibroblasts on rectangular sheets of biodegradable mesh. Using small angle light scattering (SALS), we quantified the collagen fiber architecture of Dermagraft with the mesh scaffold contributions removed through the use of a structurally based optical model. Dermagraft collagen fibers were found to have a preferred direction nearly parallel to the long dimension of the kite-shaped mesh opening with small spatial variations over the mesh. This study demonstrated the utility of SALS as a rapid and inexpensive technique for the evaluation of gross collagen fiber architecture in engineered tissues.

Characterization of SV40-transfected cell strains from rabbit keratocytes

The process of corneal wound healing involves the transformation of adjacent corneal keratocytes to myofibroblast-like cells characterized by the development of prominent microfilament bundles containing alpha-smooth muscle-specific actin (alpha-SM), a contractile protein thought to be important in mediating wound contraction. Recent studies have shown that the expression of alpha-SM in cultured corneal keratocytes can be induced by serum and TGF beta 1. To study the cellular and molecular mechanisms underlying this transformation process and to begin to identify the role of alpha-SM in wound contractile events, we generated immortalized rabbit corneal cell strains with extended life by using SV40 transfection. Two unique strains were isolated (TRK-36 and TRK-43). TRK-36, which appears similar to normal corneal keratocytes, maintains a stellate, keratocyte morphology when grown in the absence of serum and transforms to a myofibroblast-like cell when treated with TGF beta 1 (1 ng/ml), as indicated by the induced expression of alpha-SM actin. TRK-43 exhibits features characteristic of myofibroblasts in that it constitutively expresses alpha-SM actin under serum-free conditions. Both strains show in vitro contraction of collagen gels < or = 80% in 24 h in serum-containing medium. Interestingly, under serum-free conditions, TRK-43 cells showed significantly greater contraction of collagen gels compared with those of TRK-36. Overall, the establishment and further study of these cell strains may provide important insights into the molecular mechanisms underlying myofibroblast transformation.

Adherence of Pseudomonas aeruginosa to shed rabbit corneal epithelial cells after overnight wear of contact lenses

PURPOSE

Previous studies have shown that contact lens oxygen transmissibility correlates with binding of Pseudomonas aeruginosa to the rabbit cornea after overnight lens wear. Studies of human lens wear stratified by oxygen transmissibility will be required to validate these animal results. In humans, bacterial binding to shed cells obtained through corneal irrigation cytology may provide an indirect measure of in vitro binding. The purpose of this study was to establish the relationship between binding to shed cells and to the residual corneal surface in an animal model of lens wear prior to initiation of human studies.

METHODS

The test contact lenses used were: rigid lens A (Dk/L = 10 x 10(-9) [cm/ sec][mL O2/mL mmHg]); rigid lens B (Dk/L = 97); soft lens A (Dk/L = 9); soft lens B (Dk/L = 20); and, soft lens C (Dk/L = 39). There were six rabbits in each group, except for the soft lens C group, which had seven rabbits. After overnight lens wear, the corneal surface was irrigated with a corneal irrigation chamber to collect surface cells before exposure to a bacterial suspension (1 x 10(7) CFU/mL) for 30 minutes. The number of bacteria adherent to the residual corneal surface was then assessed by CFU determination. Cells collected from the corneal surface (9 mL) were incubated with 1 mL bacterial suspension containing 10(8) (CFU/mL) for 30 minutes. The number of bacteria adherent to shed cells was assessed by staining with acridine orange and direct counting by epifluorescence microscopy.

RESULTS

The differences in the number of bacteria adhering to shed epithelial cells between the treated and the control eyes were 2.90 +/- 1.20 and 0.23 +/- 0.41 for rigid lenses A and B, respectively, and 5.97 +/- 1.54, 3.67 +/- 2.32, and 0.90 +/- 1.45 (bacterial/cell) for soft lenses A, B, and C, respectively. Overnight contact lens wear induced a significant increase in bacterial binding to shed corneal epithelial cells for rigid lens A and for soft lenses A and B. There were significant differences among lens groups (P = 0.00017, ANOVA), with significant differences between rigid lenses A and B, soft lenses A and C, and soft lenses B and C. The binding of bacteria to shed cells was significantly correlated with the binding of bacteria to the residual corneal surface, both confirming and extending previous results (R = 0.78, P < 0.001).

CONCLUSION

These results demonstrate a positive correlation between P. aeruginosa adherence to shed corneal cells and to the residual corneal surface in the rabbit eye following contact lens wear. In light of the results from prior animal studies, examination of the behavior of P. aeruginosa binding to exfoliated cells appears to be a promising and valid method for future assessment of similar lens-induced increases in bacterial binding in prospective human clinical studies.

Induction of alpha-smooth muscle actin expression and myofibroblast transformation in cultured corneal keratocytes

The effects of serum, transforming growth factor (TGF) beta 1, bFGF, and heparin on in vitro myofibroblast transformation was studied. Primary rabbit corneal keratocytes were grown under serum-free conditions or in media supplemented with serum (10% fetal calf serum), TGF beta 1 (0.1-10 ng/ml), basic fibroblast growth factor (bFGF) (0.1-10 ng/ml), or heparin (10 U/ml). Cells were analyzed for expression of alpha-smooth muscle actin (alpha-SM actin), alpha 5 beta 1 integrin (the high-affinity fibronectin receptor) and fibronectin by immunoprecipitation, Western blotting, and immunofluorescence. Corneal keratocytes grown in the presence of serum showed a typical fibroblast morphology with induction of alpha-SM actin expression in 1 to 10% of cells. Addition of bFGF blocked serum-induced alpha-SM actin expression, whereas addition of TGF beta 1 enhanced alpha-SM actin expression (100%), which in combination with heparin (10 U/ml), led to a pulling apart of the fibroblastic sheet, simulating contraction. Under serum-free conditions, with or without bFGF and heparin, primary corneal fibroblasts appeared morphologically similar to in situ corneal keratocytes, demonstrating a broad, stellate morphology with interconnected processes and no alpha-SM actin expression. Addition of TGF beta 1 to serum-free cultures resulted in a dramatic transformation of corneal keratocytes to spindle-shaped, fibroblast-like cells that expressed alpha-SM actin in 100% of cells and exhibited a 20-fold increase in fibronectin synthesis and a 13-fold increase in alpha 5 beta 1-integrin synthesis. These effects were blocked by the addition of neutralizing antibodies (16 micrograms/ml). Overall these data suggest that TGF beta 1 is a potent modulator of myofibroblast transformation under serum-free conditions. In addition, the growth of keratocytes in serum appears to mimic, in part, in vivo activation and myofibroblast transformation. We conclude that detailed study of TGF beta 1-induced myofibroblast transformation under defined serum-free conditions will provide important insights into the myofibroblast transformation process.

Effects of basic FGF and TGF beta 1 on F-actin and ZO-1 organization during cat endothelial wound healing

Previous studies suggest the existence of two separate and distinct mechanisms of endothelial wound healing (i.e., cell migration and cell spreading), which may be controlled by unique, injury-dependent, wound-related factors. The purpose of our study was to evaluate potential biologic mediators regulating healing of the growth arrested cat endothelium by using an ex vivo, organ-culture model. Three buttons were punched from each cornea of 11 cats with a 6-mm trephine. A 1- to 2-mm diameter endothelial scrape injury (SI) was made, and buttons were cultured in (a) serum-free media (SFM), (b) serum plus media (20% fetal calf serum), (c) SFM plus basic fibroblast growth factor (bFGF), (d) SFM plus bFGF and heparin, (e) SFM plus transforming growth factor-beta 1 (TGF beta 1), or (f) SFM plus TGF beta 1 and anti-TGF beta 1. At various times from 8-48 h after injury, buttons were stained with phalloidin and anti-ZO-1, and imaged by using laser scanning confocal microscopy. Evaluation of SI in cat corneal buttons under serum-free conditions showed maintenance of normal endothelial differentiation, indicating that the organ-culture SI model mimics in vivo SI. Addition of TGF beta 1 produced a dramatic reorganization of apical F-actin and development of stress fibers, as well as the loss of normal cell border-associated ZO-1 distribution. The effects of TGF beta 1 were blocked by the neutralizing antibodies to TGF beta 1. Addition of serum or bFGF produced much less pronounced changes in F-actin and ZO-1 distribution. These results suggest that TGF beta 1 may play a critical role in modulating the wound-healing response of the corneal endothelium.

Application of in vivo confocal microscopy to the understanding of surfactant-induced ocular irritation

The purpose of this study was to assess the ability of in vivo confocal microscopy (CM) to provide noninvasively derived histopathologic correlates of surfactant-induced eye irritation from which specific pathologic mechanisms can be identified and later evaluated in alternative in vitro models. Rats and rabbits, divided into groups of 5, received 10 microliters of an anionic or cationic surfactant in one eye with the other eye used as a control. At specified times, eyes were examined and scored for ocular irritancy using a penlight and slit-lamp. Subsequently, corneas were evaluated by in vivo CM to evaluate epithelial layer thickness and surface epithelial cell area, corneal thickness, depth of necrosis, inflammation, fibrosis, and endothelial injury. At 3 hr, the anionic surfactant produced slight irritation with peak scores of 12.4 and 8.0 out of a possible 110 in the rats and rabbits, respectively. In vivo CM revealed changes limited to the corneal epithelium that decreased in thickness to 78% in rats and 81% in rabbits at 3 hr. This decrease in the thickness correlated with a significant decrease in surface epithelial cell area from 2,061 +/- 395 microns2 to 567 +/- 330 microns2 in the rats and 1,523 +/- 185 microns2 to 934 +/- 71 microns2 in the rabbits (p < 0.005 and 0.005, respectively). The cationic surfactant produced severe irritation in both the rats and rabbits with peak scores of 85.4 and 80.2 occurring at day 2, respectively. In vivo CM in the rats showed complete loss of corneal epithelium, lysis of keratocytes, and loss of corneal endothelium. In the rabbits, injury appeared limited to the anterior cornea with complete loss of epithelium and loss of keratocytes extending to 52% of the corneal thickness. These findings establish the application of noninvasive, in vivo CM to qualitatively and quantitatively characterize the pathobiology of ocular irritation in situ. This information will be important in the development and evaluation of mechanistically based in vitro alternatives for ocular irritancy testing.

Measurement of surgically induced corneal deformations using three-dimensional confocal microscopy

The goal of this study was to develop and apply a new set of experimental techniques for measuring the local deformations induced by partial-thickness corneal incisions in situ. Eight adult cat eyes were enucleated and cannulated, with corneal viability maintained as close to in vivo conditions as possible and intraocular pressure (IOP) carefully controlled. Experimental measurements were made pre/post radial keratotomy (RK) surgery in situ at IOPs of 15, 30, and 45 mm Hg. Incision depth and cross-sectional profiles were measured at the midpoint of selected incisions using three-dimensional (3-D) tandem scanning confocal microscopy (TSCM); central corneal curvature was estimated using a commercial corneal topographical analysis system, and corneal thickness was assessed by both 3-D TSCM and ultrasonic pachymetry. Corneas were then processed for light microscopy and incision depth was measured histologically. Finite element models were developed for comparison with the experimental measurements. There was no significant change in central corneal thickness (-5.3 +/- 3.9%, n = 8) over the course of the experiments, demonstrating that normal endothelial cell function and normal stromal hydration was maintained. The in situ TSCM incision depth measurements were significantly correlated with the histological measurements (slope = 0.95, R = 0.854, p < 0.01, n = 13 incisions). Measured incision gape at the top (anterior) of the stroma was 64.9 +/- 13.4, 87.3 +/- 12.6, and 108.7 +/- 14 microns at IOPs of 15, 30, and 45 mm Hg, respectively. The 3-D incision profiles were nonlinear in shape; comparison with the finite element models suggests that the shape of the wound profile may provide unique information regarding the shear stiffness of the cornea. Overall, the data suggest that TSCM measurements of the cross-sectional profile of the incisions immediately after RK under controlled in situ conditions provide important data regarding the mechanical behavior of the cornea after refractive surgery. These data should provide the foundation for future studies into the relationships between local tissue mechanics and corneal wound healing.

Quantitative three-dimensional confocal imaging of the cornea in situ and in vivo: system design and calibration

A new depth encoding system (DES) is presented, which makes it possible to calculate, display, and record the z-axis position continuously during in vivo imaging using tandem scanning confocal microscopy (TSCM). In order to verify the accuracy of the DES for calculating the position of the focal plane in the cornea both in vitro and in vivo, we compared TSCM measurements of corneal thickness to measurements made using an ultrasonic pachymeter (UP, a standard clinical instrument) in both enucleated rabbit, cat, and human eyes (n = 15), and in both human patients (n = 7). Very close agreement was found between the UP and TSCM measurements in enucleated eyes; the mean percent difference was 0.50 +/- 2.58% (mean +/- SD, not significant). A significant correlation (R = 0.995, n = 15, p < 0.01) was found between UP and TSCM measurements. These results verify that the theoretical equation for calculating focal depth provided by the TSCM manufacturer is accurate for corneal imaging. Similarly, close agreement was found between the in vivo UP and TSCM measurements; the mean percent differences was 1.67 +/- 1.38% (not significant), confirming that z-axis drift can be minimized with proper applanation of the objective. These results confirm the accuracy of the DES for imaging of the cornea both ex vivo and in vivo. This system should be of great utility for applications where quantitation of the three-dimensional location of cellular structures is needed.

Assessment of f-actin organization and apical-basal polarity during in vivo cat endothelial wound healing

PURPOSE

To assess the relationships between cytoskeletal changes and apical-basal polarity during healing of mechanical scrape injuries in the cat corneal endothelium.

METHODS

Ten cats (20 eyes) were used in this study. One mechanical scrape injury was created in the corneal endothelium of each eye using a blunt olive tip cannula. Tandem scanning confocal microscopy (TSCM) was performed at sequential time points after injury for in vivo assessment of cell morphology and wound healing rates. In two eyes, scanning electron microscopy was performed to allow verification of TSCM observations. Ten eyes were collected between 6 and 48 hours after wounding for in situ labeling of f-actin, ZO-1, or both.

RESULTS

Cat endothelial cell morphology observed using in vivo microscopy was identical to that shown using scanning electron microscopy. During healing, endothelial cells always remained attached to the endothelial sheet, although some showed extensions of lamellipodia into the open wound area. The in situ localization of f-actin also correlated with the TSCM in vivo wound morphology. Quantitative analysis showed that there was a decrease in the intensity of phalloidin-fluorescein isothiocyanate staining at the leading edge of the wound, suggesting a decrease in f-actin; a significant correlation was found between the relative intensity of f-actin staining and the distance from the wound margin (R = 0.98, P < 0.01). At 24 and 48 hours after injury, both ZO-1 and f-actin maintained an apical localization within cells immediately adjacent to the leading edge, despite the considerable distance of movement and dramatic decrease in the intensity of f-actin staining.

CONCLUSIONS

Overall, these data demonstrate that after scrape injury in the cat, endothelial cells exhibit a pattern of healing in which total intracellular f-actin is reduced, but normal cell connectivity and apical-basal polarity are maintained throughout.

Basic science and applications of in vivo microscopy

Confocal microscopy creates a scanned image from a point light source and point detection or a scanning slit to remove scattered light and improve optical resolution. This also results in optical sectioning of tissues. These capabilities can be employed to image structures in the human cornea, in vivo, both for research and for the diagnosis and treatment of human disease. Optical sections, when recombined, can lead to three-dimensional reconstructions from which very useful information is obtained. Investigators have found keratocyte density decreases from anterior to posterior in the stroma of the rabbit cornea. Surface epithelial desquamation can also be studied and the effects of contact lens use can be demonstrated. The instrument is also useful for diagnosing and guiding therapy for some human diseases such as Acanthamoeba keratitis. Colonies of bacteria may also be observed and treatment evaluated in patients with infectious crystalline keratopathy. Confocal microscopy can also image the retina.

The spatial organization of corneal endothelial cytoskeletal proteins and their relationship to the apical junctional complex

PURPOSE

To determine the spatial organization of the major cytoskeletal proteins and their relationship to the apical junctional complex (AJC) in the normal rabbit corneal endothelium.

METHODS

Normal endothelial cytoskeletal structure in three dimensions was studied in rabbit eyes by laser scanning confocal microscopy after en bloc immunocytochemical staining of whole corneal tissue with various antibodies and fluorescent probes; specificity of antibodies to rabbit corneal endothelial cell proteins was established by Western blot analysis.

RESULTS

Normal actin microfilament network organization was seen predominantly as a complex apical array forming a circumferential bundle. The tight junction-associated protein ZO-1 was positive at the apical junctions, forming a hexagonal pattern that was localized between and just proximal to the circumferential actin microfilament bundles. The distribution of ZO-1 was discontinuous around the cell, with the largest gaps (1 micron in diameter) occurring at the Y-junction between adjacent endothelial cells; transmission electron microscopy of the apical face of the endothelium confirmed the existence of 1-micron diameter gaps in the adherens junctions located at the Y-junction. Antivimentin antibodies showed a ring of intermediate filaments located just below the circumferential actin microfilament band. This ring appeared to be continuous with a basal mat of filaments, which together formed a basketlike structure within endothelial cells. An intricate cytoplasmic, perinuclear network of microtubules was observed by antitubulin antibodies that appeared unrelated either to the apical circumferential actin microfilament bundle or to intermediate vimentin filament ring. Staining of endothelial cells with NBD-ceramide identified a prominent, perinuclear Golgi complex suggesting an association between microtubules and Golgi.

CONCLUSIONS

The organization of cytoskeletal elements and the tight junction-associated protein ZO-1 is similar to the classical AJC of transporting epithelia, comprised of a zonulae occludens (ZO) located apical to a zonulae adherens (ZA) and desmosomes. The organizational pattern seen in corneal endothelial cells, however, is distinct from transporting epithelia in that the ZO and ZA are discontinuous, with large gaps in the ZO-1 distribution at the Y-junction between adjacent endothelial cells. The authors propose that the structural differences in the AJC underlie the functional differences between classical transporting epithelia, which actively pump fluid from the lumen to the mucosa, and the corneal endothelium, which has a "pump-leak" fluid transport mechanism.

Temporal, 3-dimensional, cellular anatomy of corneal wound tissue

We have evaluated temporally the 3-dimensional cellular anatomy of corneal wound tissue in the rabbit eye using in vivo tandem scanning confocal microscopy. In vivo microscopic studies showed that corneal fibroblast migrated into the wound as an interconnected cellular meshwork with long, thin, randomly oriented cell processes. Interconnection of fibroblasts was further confirmed by localisation of monoclonal antibodies to connexin 43 which demonstrated prominent staining of putative gap junctions between fibroblasts. Temporal observations indicated that the interconnected cells and cellular processes undergo sequential positional changes leading to orientation of cells and interconnected cell processes parallel to the wound margin. Laser scanning confocal microscopy of en bloc, phalloidin-stained corneal wounds showed prominent intracellular f-actin bundles (i.e. stress fibres) within cell processes which formed an extensive interwoven pattern within the wound.

Expression of alpha-smooth muscle (alpha-SM) actin during corneal stromal wound healing

PURPOSE

The purpose of this study was to correlate the temporal expression of alpha-smooth muscle specific actin (alpha-SM actin), a molecular marker for myofibroblast transformation, with corneal wound contraction.

METHODS

After full-thickness, central corneal injury in rabbit eyes, the anterior width of the wound (wound gape) was measured in the same animals using in vivo confocal microscopy. In addition, animals were sacrificed at various times after injury for the determination of alpha-SM actin expression by immunofluorescent microscopy using a mouse monoclonal antibody specific for human alpha-actin. Antibody specificity was confirmed by Western blot analysis of normal and wound fibroblasts. Expression of alpha-SM actin also was related spatially to f-actin and the wound margin by co-localization with phalloidin and DTAF (5([4,6-dichlorotriazin-2yl]amino)fluorescein), a fluorescent marker bound to the wound margin.

RESULTS

Wound contraction was most evident from days 7 to 42, when wound gape progressively decreased from 574 +/- 120 microns to 250 +/- 61 microns. Thereafter, the wound remained stable to day 84 (304 +/- 58 microns). Expression of alpha-SM actin directly correlated with wound contraction--appearing across the wound at day 7, the full thickness of the wound at day 14, and the posterior wound at day 28. alpha-SM actin was localized exclusively to phalloidin-stained, f-actin microfilament bundles or stress fibers within wound healing fibroblasts, and the disappearance of alpha-SM actin correlated with the concomitant disappearance of stress fibers at days 28 to 42. Staining of the wound margin with DTAF confirmed that the expression of alpha-SM actin was limited to fibroblasts within the wound.

CONCLUSIONS

The expression of alpha-SM actin was directly correlated to corneal wound contraction, appearing at the initiation of and disappearing at the completion of the contraction process. Furthermore, the exclusive expression of alpha-SM actin by fibroblasts present only within the wound suggests that local environmental factors unique to the wound may play an important role in myofibroblast transformation.

Quantitative assessment of anteroposterior keratocyte density in the normal rabbit cornea

The anteroposterior keratocyte density distribution in the rabbit cornea was measured. Unsectioned tissue blocks from the central cornea of five rabbits were stained with propidium iodide and imaged using a Leica laser scanning confocal microscope. A z-series of images was acquired confocal microscope. A z-series of images was acquired in each sample, from anterior to posterior stroma in either 3- or 8-microns steps. Software was developed to allow interactive marking of the keratocyte nuclei within each section of the z-series and for calculating cell density. For convenience, cell density was expressed as the number of cells per corneal volume element (CVE), where CVE is a newly defined volume unit with x, y, and z dimensions of 250, 250, and 10 microns, respectively. The calculated keratocyte density was 20.2 +/- 1.0 cells/CVE (n = 5), which is equivalent to 32,360 +/- 1,660 cells/mm3. The greatest density was underneath the epithelium (26.3 +/- 2.5 cells/CVE), the density then decreased linearly with depth to 15.2 +/- 1.4 cells/CVE; there was a slight increase in density pre-Descemets membrane to 18.5 +/- 3.5 cells/CVE. A 30% decrease in cell density over the entire anteroposterior stromal thickness was observed. To facilitate statistical analysis, the cell density was averaged over 5% thickness intervals from anterior to posterior cornea. A significant difference in mean cell density of these intervals was found (ANOVA, n = 20, p < 0.01). To further assess the density distribution, linear regression analysis was performed. A significant correlation was found between keratocyte density and stromal depth (R = -0.94, n = 20, p < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Laser and tandem scanning confocal microscopic studies of rabbit corneal wound healing

The process of corneal endothelial wound healing was studied using laser and tandem scanning confocal microscopy (LSCM and TSCM). Following transcorneal freeze (TCF) injury, rabbit corneas were observed using ex vivo LSCM and in vivo TSCM. LSCM revealed the intracellular actin filament organization which, stained with phalloidin-FITC, in migrating endothelial cells, transformed fibroblast-like cells, stroma keratocytes, and epithelial cells during wound healing in corneal tissue. The TSCM provided sequential spatial observation of morphologic changes from endothelium to epithelium of the cornea during in vivo cellular repair of wound healing noninvasively on the same cornea without animal sacrifice. Ex vivo LSCM supported the morphologic analysis of the in vivo TSCM observations.

The effects of daily wear of rigid gas permeable contact lenses treated with contact lens care solutions containing preservatives on the rabbit cornea

We evaluated the effects on the rabbit cornea of daily wear of rigid gas permeable (RGP) contact lenses treated with preserved care solutions by measuring concomitant tear lactate dehydrogenase (LDH) activity followed by in vivo tandem scanning confocal microscopy (TSCM). In vivo morphologic changes were confirmed by in vitro scanning electron microscopy (SEM). Two standard commercial RGP lens wetting and soaking solutions from the same manufacturer were tested: solution A with 0.004% benzalkonium chloride (BAK) and solution B with 0.003% chlorhexidine digluconate (CHX) and 0.002% thimerosal. Two experimental PBS-based wetting and soaking solutions were also tested: solution C with 0.005% BAK and 2% hydroxypropylmethylcellulose (HPMC) and solution D with 0.005% BAK without HPMC. Instillation of solution A without contact lens wear caused significant (P < 0.01) increases in desquamation of the superficial corneal epithelium and tear LDH activity compared with control eyes. After 3 weeks of RGP contact lens daily wear (8 hours/day), modified Draize scores of ocular surface lesions on the eyes wearing RGP lenses treated with solution A increased according to the duration of lens wear. Solution B did not produce significant change. With daily wear for 4 days (8 hours/day), RGP lenses treated with solution C and solution D produced increased corneal epithelium desquamation and an increase of LDH activity in tears. These effects were greater with HPMC (solution C) than without HPMC (solution D).(ABSTRACT TRUNCATED AT 250 WORDS)

In vivo confocal imaging: general principles and applications

It is well established that confocal microscopy provides higher resolution images with better rejection of out-of-focus information than conventional light microscopy. The optical sectioning ability of confocal microscopy allows images to be obtained from different depths within a thick tissue specimen, thereby eliminating the need for processing and sectioning procedures. Thus, confocal microscopy has made it possible to view biological tissues under more physiologic conditions than previously possible. The most widespread biological application of confocal microscopy has been in the localization of immunofluorescently labeled proteins in cell culture or within excised blocks of tissue. Because of its noninvasive optical sectioning capability, confocal microscopy is also ideally suited to the study of tissue in intact living animals, although the potential in vivo applications of this paradigm have received less attention. In this paper we trace the development of in vivo confocal microscopy and present examples of current capabilities for both research and clinical use.

In vivo osmotic pertubation of intercellular fluid channels in the rabbit corneal endothelium

An in vivo rabbit corneal model was used to evaluate morphological changes in the corneal endothelium associated with osmotically increasing fluid movement from the anterior chamber into the stroma. When the corneal stroma is rendered more hypertonic than normal by immersing the scraped epithelial side of the cornea in a hypertonic sucrose solution, intercellular channels and apical pores at the Y-junctions between endothelial cells become greatly enlarged. The foregoing changes are reversible and do not appear to result in damage to the corneal endothelium. These observations suggest that specific intercellular channels in the corneal endothelium may provide pathways for the movement of fluid from the aqueous humor into the stroma.

The relation between contact lens oxygen transmissibility and binding of Pseudomonas aeruginosa to the cornea after overnight wear

PURPOSE

To assess adverse effects of contact lens-induced hypoxia on the rabbit cornea in vivo and determine the relation between binding of Pseudomonas aeruginosa and oxygen transmissibility for rigid and hydrogel lenses.

METHODS

Six rigid lenses with Dk/Ltotal values between 0 and 97 x 10(-9) (cm/second) (ml O2/ml mmHg) and four hydrogel lenses (Dk/Ltotal 9, 20, 39, 51) were tested. All lenses had 14.0-mm diameters and a thickness (parallel) of 0.12 or 0.15 mm. Tear lactate dehydrogenase activity and tandem scanning confocal microscopy determinations were performed after the lens was worn for 24 hours. Binding of P. aeruginosa then was separately determined by the colony-forming unit method. Scanning electron microscopy was used to confirm in vivo tandem scanning confocal microscopy findings.

RESULTS

Lens oxygen transmissibility determines binding of P. aeruginosa to the cornea after the lens is worn for 24 hours; epithelial damage produced by lenses of lower Dk/Ltotal appears to be the dominant biologic factor for P. aeruginosa binding and not lens rigidity.

CONCLUSIONS

These results suggest that the risk of P. aeruginosa keratitis developing with overnight wear will be enhanced significantly for contact lenses with Dk/Ltotal values less than 50 x 10(-9) (cm/second) (ml O2/ml mmHg) (human equivalent oxygen percentage < or = 15%), and this risk will increase with further decreases in oxygen transmissibility. Because no hydrogel lenses approved by the Food and Drug Administration are available with oxygen transmission at this level, patients should be made aware of the increased risk of infectious keratitis associated with the overnight wear of current extended wear hydrogel lenses. Results of this study also demonstrate that quantitative clinical tandem scanning confocal microscopy imaging and tear lactate dehydrogenase activity measurements can provide prospective, noninvasive methods for assessing the ongoing interaction between contact lens and cornea in vivo.

Corneal keratocytes: in situ and in vitro organization of cytoskeletal contractile proteins

PURPOSE

Recent studies of corneal wound healing suggest that activated corneal keratocytes develop myofibroblast-like characteristics including a putative contractile apparatus comprised, in part, of intracellular microfilament bundles (i.e., stress fibers) containing f-actin, myosin, and alpha-actinin; extracellular fibronectin fibrils; and fibronectin surface membrane receptors (alpha 5 beta 1 integrin). The purpose of this study was to determine the expression and organization of specific components of the contractile apparatus in normal, quiescent (in situ) corneal keratocytes, and to compare the in situ organization with that of activated, tissue culture (in vitro) corneal keratocytes that potentially mimic wound healing fibroblasts.

METHODS

Cat corneal tissue was obtained immediately after sacrifice and was either fixed for in situ studies or cultured with MEM supplemented with 10% fetal calf serum for in vitro studies. Keratocytes (in situ and in vitro) were stained with the following probes: phalloidin, a mushroom toxin that specifically binds to f-actin; rabbit anti-bovine aortic myosin; monoclonal anti-human alpha-actinin; monoclonal anti-human vimentin; rabbit anti-human alpha 5 beta 1 integrin; monoclonal anti-human alpha 5 integrin; monoclonal anti-human connexin 43; and goat anti-human fibronectin. The cytoskeletal organization and co-localization were evaluated using epifluorescent and confocal microscopy.

RESULTS

Normal, quiescent corneal keratocytes were distributed within the cornea as a lattice network, interconnected by broad, cellular processes extending from a flattened cell body. The f-actin distribution of in situ keratocytes was predominantly cortical and appeared to be closely associated with the plasma membrane. In addition, punctate areas that appeared to correlate with the localization of adhesion sites were identified. These punctate regions appeared to stain with antibodies to alpha 5 beta 1 but to not alpha 5. These data suggest that the fibronectin receptor, alpha 5 beta 1 integrin, is not present on normal corneal keratocytes. Based on co-localization studies, rabbit anti-bovine aortic myosin and monoclonal anti-alpha-actinin staining had similar distributions to FITC-phalloidin. Interconnections between keratocytes also showed staining for connexin 43, indicating the presence of gap junctions. By contrast, activated, cultured (in vitro) keratocytes showed an FITC-phalloidin staining pattern localized predominantly along intracellular stress fibers not detected in normal, quiescent keratocytes. Myosin and alpha-actinin staining had a similar stress fiber distribution, arranged in alternating bands and suggesting a sarcomeric distribution. Associated with stress fibers there was both anti-alpha 5 beta 1 and anti-alpha 5 staining, indicating the presence of focal adhesions.

CONCLUSIONS

This study demonstrates that there are major structural differences in the organization of contractile cytoskeletal proteins between normal, quiescent (in situ), and activated (in vitro) keratocytes. In situ, contractile proteins appear to be associated with the cortical f-actin network, probably related to maintenance of cell shape and interconnectivity. Alternatively, activated keratocytes were characterized by the presence of a putative contractile apparatus comprised of f-actin, myosin, and alpha-actinin organized into sarcomeric, muscle-like bundles (stress fibers) associated with focal contacts containing alpha 5 beta 1 integrin. These data suggest that activation of keratocytes, i.e. myofibroblast transformation, must involve the reorganization of cytoplasmic contractile proteins as well as the expression of alpha 5 beta 1 integrin and the formation of focal contacts.

The application of confocal microscopy to the study of living systems

A unique tandem confocal microscope (TSCM) has been developed that permits noninvasive imaging in vivo of the eye and many other organ systems in real time in situ. The application to the study of microphysiological processes in vivo is described and illustrated for the cornea, kidney, liver, epididymis, muscle, and adipose tissue. Novel applications are shown for studying the healing of wounds in four dimensions (x, y, z, t) in single animals over time at the cellular level. Application to clinical diagnostic use in humans is also demonstrated. When combined with Laser Scanning Confocal fluorescence microscopy, the TSCM offers a unique new imaging paradigm for experimental biology and medicine with great potential for use in neuroscience and many other disciplines.

Clinical and diagnostic use of in vivo confocal microscopy in patients with corneal disease

BACKGROUND

The purpose of this article is to introduce the practicing ophthalmologist to the optical principles and images produced by a tandem scanning confocal microscope (recently approved by the Food and Drug Administration for general clinical use). The tandem scanning confocal microscope allows real-time viewing of structures in the living cornea at the cellular level in four dimensions (x, y, z, and time).

METHODS

Nine patients (2 males, 7 females), ranging in age from 7 to 52 years, were examined. Images were recorded on super VHS videotape, digitized and processed on a computer workstation, and photographed for presentation.

RESULTS

Two-dimensional (x, y) 400 x 400-microns images (9-microns z-axis thickness) are presented for normal corneal structures and for the clinical conditions of herpetic keratitis, wound healing after myopic excimer ablation, Acanthamoeba infection, corneal dystrophies (granular, Reis-Buckler), contact lens abrasion, and the irido-corneal endothelial syndrome.

CONCLUSION

Clinical confocal microscopy has the unique potential of providing noninvasive assessment of corneal injury and disease at the cellular level that is not available currently from other technologies.

In vivo confocal microscopic studies of endothelial wound healing in rabbit cornea

Corneal endothelial wound healing in living rabbit eyes after mechanical scrape (MS) and transcorneal freeze (TCF) injury was studied using tandem scanning confocal microscopy (TSCM). MS injury was created on the central corneal endothelium with an olive tip cannula; TCF injury was created using a 3-mm-diameter stainless steel probe cooled with liquid nitrogen. In vivo observation of wound healing using TSCM was correlated with scanning electron microscopy (SEM) for fixed tissues. At 6 h after MS, migrating endothelial cells at the leading edge showed lamellipodial processes on in vivo TSCM and SEM. After 24 h, the denuded area was almost fully resurfaced by migrating endothelial cells showing wide spaces between nuclei by TSCM. After 28 days, resurfaced endothelial cells showed normal hexagonal mosaic appearance with enlarged cells by TSCM and SEM. TCF injury produced fibroblastic changes in the endothelial cells with elongation and spreading by 24 h after injury. After 3 days, the wounded area was resurfaced with two cell types: (a) migrating endothelial cells at the peripheral area, which appeared polygonal in shape with wide intracellular spaces and (b) fibroblast-like cells at the center of the wound, which formed a retrocorneal fibrous membrane (RCFM). The RCFM was posteriorly covered with normal endothelium after 28-60 days. TSCM of the stroma demonstrated spindle-shaped, activated keratocytes migrating into the wounded stroma at 3-14 days. In conclusion, TSCM allows viewing of dynamic four-dimensional morphologic changes (x, y, z, and time) during in vivo cellular repair of corneal wound healing after either MS or TCF injury.

Actin filament organization during endothelial wound healing in the rabbit cornea: comparison between transcorneal freeze and mechanical scrape injuries

PURPOSE

To compare and contrast the in vivo mechanism of wound healing after mechanical scrape and transcorneal freeze (TCF) injury in a rabbit eye model by examining changes in the cytoskeletal organization of contractile, filamentous actin (f-actin) microfilaments as relates to differences in cell migration or translocation during endothelial repair.

METHODS

Endothelial wound healing after mechanical scrape and transcorneal freeze injury was studied in rabbit eyes using laser scanning confocal microscopy (LSCM). Central corneal mechanical scrape injury was made using an olive tip cannula, and TCF injury was made using a 3-mm diameter stainless steel probe cooled with liquid nitrogen. Cytoskeletal changes in f-actin stained with phalloidin-FITC were observed during wound healing using LSCM.

RESULTS

At 6 hours after mechanical scrape, the leading edge of the migrating sheet showed a decrease in the intensity of phalloidin-FITC staining, suggesting a decrease in cortical f-actin. Migrating endothelial cells in vivo did not appear to develop stress fibers after mechanical scrape, which is consistent with an in vitro cell spreading mechanism of endothelial wound healing. By 24 hours, the denuded area was almost fully resurfaced by migrating endothelial cells. On the other hand, TCF injury produced fibroblastic changes in the endothelial cells with extension and elongation of spindle-shaped endothelial cells at the leading edge by 24 hours after injury. Fibroblastic endothelial cells developed prominent actin stress-fibers, which is consistent with an in vitro cell migration mechanism of endothelial wound healing. Three days after TCF, the wounded area was resurfaced with two cell types: rough, fibroblast-like cells forming a retrocorneal fibrous membrane having prominent f-actin bundles or stress fibers with few cell-cell junctions, and smooth, polygonal-shaped endothelial cells having tight cell junctions with a cortical distribution of f-actin. After 28 days the retrocorneal fibrous membrane was posteriorly covered with normal endothelium.

CONCLUSIONS

These data support the hypothesis that endothelial wound healing involves two separate, injury-dependent, mechanisms--cell spreading and cell migration.

Three-dimensional imaging of corneal cells using in vivo confocal microscopy

Confocal microscopy is a unique and powerful imaging paradigm which allows optical sectioning through intact tissue. Real-time tandem scanning confocal microscopy has previously been used to generate high-magnification two-dimensional (2-D) images of cells in living organ systems. Inherent problems with movement, however, have prevented the in vivo acquisition of complete 3-D datasets. The development of a new objective lens, used in combination with specialized real-time image acquisition procedures, has allowed sequential serial sections to be obtained in vivo from the rabbit cornea for the first time. These sections can be digitally registered and stacked on the computer to provide a 3-D reconstruction of the corneal cells. This technique should serve as a useful method for studying 3-D structures and analysing 4-D phenomena at the cellular level in living animals. Three-dimensional images of a stromal nerve in normal rabbit cornea and of fibroblasts within a rabbit corneal wound are presented as examples of current capabilities.

Noninvasive microscopic evaluation of the intact living nephrotic kidney

BACKGROUND

Tandem Scanning Confocal Microscopy (TSCM) is a new form of microscopy that allows one to noninvasively "optically section" into intact structures and record microscopic images in real-time. In this study, we have evaluated the ability of this new technology to distinguish histopathologic changes in unstained living kidneys that occur coincident with the onset of puromycin aminonucleoside-induced nephrosis.

EXPERIMENTAL DESIGN

The rats were anesthetized and a laparotomy was performed to expose the kidneys. Using a TSCM equipped with a 24x water-immersion objective, we viewed and recorded real-time images of subcapsular living uriniferous tubules and glomeruli of puromycin aminonucleoside-induced nephrosis rats exhibiting different levels of proteinuria.

RESULTS

TSCM revealed a variety of histopathologic changes in the puromycin aminonucleoside-induced nephrosis uriniferous tubules including dilation of proximal tubular lumens, loss of microvillous brush border, cellular debris in the tubule lumens, and focal regions of extensive tubular necrosis. Using the fluorescent probe carboxyfluorescein, we were able to demonstrate significant heterogeneity in the movement of this fluorescent probe through the uriniferous tubules (e.g., some tubules were completely blocked), and irregular patterns in flow of carboxyfluorescein through nephrotic glomeruli.

CONCLUSIONS

These observations indicate that TSCM is capable of revealing a variety of histopathologic changes in unstained, intact, living kidneys. In addition, many of the histopathologic features of the uriniferous tubules revealed by TSCM are difficult or impossible to distinguish in biopsied samples of renal tissue.

Morphological and biochemical evaluation for rigid gas permeable contact lens extended wear on rabbit corneal epithelium

We studied the effects of 24-hour wear of rigid gas permeable (RGP) contact lenses of varying oxygen transmissibilities on the rabbit cornea by measuring concomitant lactate dehydrogenase (LDH) activity in tears and by in vivo tandem scanning confocal microscopy (TSCM). We used a PMMA lens and rigid gas permeable (RGP) lenses that had Dk/L values ranging from 7 to 64 x 10(-9) (cm/sec) (mL O2/mL mmHg) and a uniform 0.15 mm thickness. After 6- and 24-hour contact lens wear, rabbit tear LDH activity increased according to the decrease in the Dk of RGP lenses. Tear LDH activity after 24 hours of lens wear was higher than after 6 hours. The observed increase in tear LDH activity was correlated with in vivo corneal epithelial morphology by tandem scanning confocal microscopy. The observed severity of desquamation and swelling of corneal epithelial cells was dependent upon the Dk/Ltotal of contact lenses worn, which directly related to the contact lens induced corneal hypoxia. Based on the results of this study, we conclude that: 1) a nap or accidental overnight wear of contact lenses with less than 20 x 10(-9) Dk/Ltotal could cause severe corneal epithelial damage; 2) the ultra high Dk lens appeared to alter the ocular surface least; and 3) TSCM accompanied with tear LDH assay is an objective, non-invasive in vivo method to assess the effect of contact lens wear on the ocular surface over time at the cellular level.

Quantitative analysis of stress fiber orientation during corneal wound contraction

Previous studies of actin and actin-binding proteins in corneal myofibroblasts suggest the development of a contractile apparatus composed, in part, of F-actin micro-filament bundles, i.e. stress fibers. To better understand the mechanics of wound contraction and the relationship between microfilament bundles and wound closure, we have analyzed the spatial and temporal organization of stress fibers during the process of corneal wound healing. Rabbit corneas (26 eyes) received 6 mm full-thickness, central incisions and were studied at various times for F-actin organization using en bloc (whole cornea) staining with FITC-phalloidin, as well as conventional histological techniques. 3-D datasets (z-series of 40 en face optical sections, 1 micron steps) were collected using the Biorad MRC-600 laser scanning confocal microscope at various regions within the wound. At 7 days, 3-D analysis showed randomly oriented, interconnected F-actin filament bundles (stress fibers). Between 7 and 28 days, stress fibers appeared to organize gradually into planes parallel to the wound surface, with a large population achieving a final orientation nearly parallel to the long axis of the wound. Using Fourier Transform analysis techniques, an orientation index (OI) was calculated to quantitate global fiber orientation at each time point. Analysis of variance demonstrated a significant change (P &lt; 0.001) in overall stress fiber orientation from a random distribution at day 7 to an alignment more parallel to the lateral wound borders at day 28. Overall, these data suggest that stress fibers undergo temporal changes in spatial organization that correlate with wound closure, and that wound closure does not involve the development of previously described contractile or tractional forces aligned directly across the wound.

In vivo confocal microscopy in clinical dental research: an initial appraisal

Until recently, the in vivo microscopic investigation of intraoral tissues at high resolution has been virtually impossible. Confocal microscopy enables high-resolution imaging to be achieved below semitransparent surfaces in intact living specimens, but this may still be impractical for intraoral applications because of the need to stabilize the sample. The development of a steadying objective (x 240 overall mag.) which is held against the sample surface and is focused by moving internal elements, avoids the need for fine adjustment of the living sample under the microscope to achieve a change of focus. It is therefore more comfortable and also reduces the problems of movement due to the pulse. The objective was used with a tandem scanning microscope, with images recorded via a SIT video camera. Using this system internal tooth structure (e.g. enamel prisms/adhesive restoration interfaces) and the lining cells of the gingival crevice through to the junctional epithelium may be examined. It is also possible to image the oral mucous membrane, focusing to the capillary loops in the basal layers, where streaming red blood cells can be seen. Access is limited to the anterior regions as far back as the premolar teeth. Applications could include caries research, soft and hard tissue responses to biomaterials (e.g. implants), wound healing and monitoring the effect of periodontal treatment regimens. This new technique offers numerous exciting opportunities for the microscopic investigation of many clinical operative procedures in vivo, allowing the response of the tissues to be non-destructively monitored, over time, at high resolution.

Radial keratotomy. 1. The wound healing process and measurement of incisional gape in two animal models using in vivo confocal microscopy

Using in vivo confocal microscopy, corneal wound healing was evaluated in both rabbit and cat eyes after radial keratotomy. A total of six rabbit and six cat eyes were evaluated sequentially over time for 1 mo after surgery by in vivo confocal microscopy, and quantitative measurements of changes in incisional wound gape were determined. In vivo histopathologic changes were correlated with conventional histopathologic findings in 18 rabbit and 4 cat eyes; the animals were killed at various intervals from 0-30 days after surgery. In the rabbit, in vivo corneal wound healing was characterized by the initial ingrowth of corneal epithelium followed by persistence within the wound without a marked fibrotic response. Measurement of incisional wound gape showed increasing gape from 144 +/- 32 microns on day 0 to 976 +/- 155 microns on day 26 at a distance of 2.4 mm from the optical zone. These in vivo measurements were not significantly different (P = 0.996) from those obtained using conventional histopathologic techniques which showed an incisional wound gape of 252 +/- 112 microns on day 0 and 917 +/- 216 microns on day 26 at 2.5 mm from the optical zone. In the cat eyes, healing of radial keratotomy wounds showed an initial increase in incisional wound gape from 135 +/- 56 microns on day 0 to 245 +/- 88 microns on day 7 at a distance of 2.4 mm from the optical zone. Starting at day 14 and continuing to day 30, there was a progressive decrease in incisional wound gape from 198 +/- 41 microns to 92 +/- 35 microns. Sequential, in vivo histopathologic analyses indicated that increasing incisional wound gape correlated with the retention of corneal epithelium in the wound. Initiation of decreasing incisional wound gape was associated with replacement of the incisional epithelial plug with fibroblastic tissue. These changes in the incisional wound gape observed in the cat suggest that healing of radial keratotomy wounds involves contraction of the wound in response to the ingrowth of fibroblastic cells. Furthermore, the contractile response appears to be biphasic involving a precontractile and contractile phase. Overall these data indicated that in vivo confocal microscopy provides quantitative histopathologic data on living tissue comparable with that obtained with conventional techniques on dead, fixed, and sectioned tissue. Additionally, the absence of wound fibrosis in the rabbit radial keratotomy model raises important questions as to the appropriateness of this experimental model for human radial keratotomy.

Radial keratotomy. II. Role of the myofibroblast in corneal wound contraction

The cellular mechanism of corneal wound contraction after radial keratotomy (RK) was studied in a feline eye model. A total of 10 cat eyes were evaluated at various times from 0-30 days after surgery. Changes in the distribution of intracellular filamentous actin, nonmuscle myosin, alpha-actinin, surface membrane alpha 5 beta 1 integrin, and extracellular fibronectin were studied using immunofluorescence and laser confocal and electron microscopy. From day 3-7, staining for fibronectin increased along the wound margin. By day 7, keratocytes adjacent to the wound margin showed increased f-actin staining with intense staining for fibronectin compared with normal keratocytes. Myosin and alpha 5 beta 1 integrin expression was very weak at this time; alpha-actinin was not found. By day 14, fibroblasts within the wound formed f-actin microfilament bundles (stress fibers) which colocalized with fibronectin. Wound-healing fibroblasts also stained positively for alpha 5 beta 1 integrin, myosin, and alpha-actinin (the latter two were colocalized). The presence of myosin and alpha-actinin in the wound fibroblasts and the re-organization of f-actin into stress fibers by day 14 correlated with the development of wound contraction. A comparison of the cellular distribution of actin, myosin, and alpha-actinin with alpha 5 beta 1 integrin 14 days after injury suggested that integrin was localized along stress fiber bundles during wound contraction. The data from this study suggest that modulation of wound gape during healing of RK wounds may involve transformation of the corneal keratocyte to a myofibroblast-like cell and the subsequent formation of intracellular stress fibers composed of f-actin, nonmuscle myosin, and alpha-actinin. Based on the colocalization of fibronectin filaments and f-actin filaments and the unique distribution of alpha 5 beta 1 integrin, these findings support the hypothesis that the tension within the wound is generated by the formation of intracellular stress fibers and the interactions between stress fibers and the extracellular matrix, mediated by specific membrane receptor molecules.

Radial keratotomy. III. Relationship between wound gape and corneal curvature in primate eyes

The relationship between changes in wound gape and corneal curvature after radial keratotomy (RK) was evaluated in five primates. Four-incision RK was performed using a diamond knife set to 100% of central corneal thickness with a 3-mm optical zone. In vivo measurements of wound gape were obtained using tandem scanning confocal microscopy at 3, 7, 14, and 45 days after surgery. The changes in corneal contour were measured at the same time points using a corneal modeling system with a specially designed primate cone. Wounds progressively increased in width to a maximum of 38 +/- 1 microns (n = 5) at day 7. After day 7, wounds showed increasing fibrosis which correlated with decreasing wound gape to 20 +/- 1 microns at day 45. A similar temporal change was detected in central corneal curvature (K), with maximum flattening occurring at day 7 (delta K = -3.17 +/- 0.90 diopters, n = 5), and progressive regression of effect to -1.32 +/- 0.61 diopters (n = 5) at day 45. Although there was interanimal variation, the mean temporal changes in corneal curvature significantly paralleled the changes in wound gape (r = -0.96, n = 4, P < 0.05). Based upon these findings, a simple geometric model was proposed which provides a hypothetic foundation for the relationship between corneal curvature and wound gape after RK. Calculations of wound gape made from this analytic model (using the measured topographic data) showed significant correlation with the actual wound gape measurements (r = 0.96, n = 4, P < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of increasing Dk with rigid contact lens extended wear on rabbit corneal epithelium using confocal microscopy

The effects of 24-h wear of various Dk-rigid gas-permeable (RGP) contact lenses on the rabbit corneal epithelium were studied by in vivo tandem scanning confocal microscopy (TSCM), and confirmed by scanning electron microscopy (SEM). Lenses used were polymethylmethacrylate (PMMA) (Dk/L = 0), RGP experimental A lens (siloxanylmethacrylate-fluoromethacrylate-methylmethacrylate , 33), experimental B (siloxanylmethacrylate-fluoromethacrylate, 56), and experimental C (siloxanylstyrene-fluoromethacrylate copolymer, 64 x 10(-9)) (cm/s) (ml O2/ml mm Hg) with 0.15-mm thickness (Dk/L measured by polarograph including boundary layer effect). After 24-h PMMA lens wear, TSCM showed no superficial epithelial cells but only exposed, underlying wing cells. The cornea with experimental A showed partial superficial epithelial desquamation. With experimental B wear, slight superficial epithelial cell swelling and desquamation were observed on the surface of the cornea. No changes were observed for the eye with experimental C and control. The observed severity of desquamation of superficial epithelial cells was dependent on the oxygen transmissibility (Dk/L) of RGP lenses worn. All in vivo findings were confirmed by SEM observations. Based on the results of this study, we conclude that (a) although Dk/L = 56 lens B shows no residual overnight corneal swelling, surface damage is still produced; (b) Dk/L = 64 lens C is best for epithelium showing the same corneal images as control; and (c) TSCM is a good way to evaluate the contact lens safety and efficacy in vivo at the cellular level noninvasively.

Confocal microscopic studies of living rabbit cornea treated with benzalkonium chloride

The effects of benzalkonium chloride (BAK) on the living rabbit cornea were studied by in vivo Tandem scanning confocal microscopy (TSCM) and confirmed by conventional scanning electron microscopy (SEM). Two drops of saline or phosphate-buffered saline (PBS) containing BAK in concentrations of 0.02, 0.01, and 0.005% was applied to rabbit eyes 15 times at 5-min intervals. The solutions were pH 5.5-5.9 (saline) and pH 7.5 (PBS), and osmolarity was 275-280 (saline) and 300-307 mOsm (PBS). Immediately after application of 0.02 and 0.01% BAK, no normal corneal superficial epithelial cells could be imaged by in vivo TSCM. No swelling of the superficial epithelial cells was observed for the control solution without BAK; however, there was a small amount of desquamation. Application of as little as 0.005% BAK caused the superficial epithelial cells to swell and desquamate. The observed desquamation of corneal superficial epithelial cells increased with higher BAK concentrations applied to the eye. One hour after final drug application, inflammatory cells appeared on the surface of the cornea treated with 0.02% BAK. These findings were correlated with SEM observations. Based on the results of this study, we believe that BAK used frequently can produce clinical corneal toxicity and that the cytotoxicity of any topical ophthalmic solutions can be evaluated by in vivo TSCM.

Comparison of in vivo and ex vivo cellular structure in rabbit eyes detected by tandem scanning microscopy

Using the tandem scanning microscope, in vivo confocal microscopic images of living eyes were compared to images obtained from ex vivo, freshly enucleated or fixed tissue in the rabbit. In the normal cornea, microscopic details of the superficial epithelium, basal lamina, stromal fibrocyte nuclei, nerves and endothelial cell borders were easily discernible. Removal of the eye from the intact animal resulted in loss of detail with distortion of the normal structural interrelationships within the corneal stroma whilst enhancing details of the corneal epithelium. Formalin fixation further enhanced details of the basal and suprabasal corneal epithelial cell nuclei and the stromal fibrocyte cell borders whilst inducing prominent brightly reflecting folds in the thickened stroma with concomitant enhancement of the edge contrast of the collagen lamellae. These changes appeared to be related, in part, to hydration of the cornea and artefactual pooling of water between structures that may enhance reflectivity by increasing the difference between the refractive index of the cellular and extracellular elements. We conclude that microscopic examination of ex vivo preparations of corneal tissue, although providing increased resolution similar to conventional light microscopic techniques, significantly altered the normal structural relationships and could lead to erroneous measurements of the physiological properties of the tissue as compared to in vivo microscopy of undisturbed, intact tissue.

Digital image acquisition in in vivo confocal microscopy

A flexible system for the real-time acquisition of in vivo images has been developed. Images are generated using a tandem scanning confocal microscope interfaced to a low-light-level camera. The video signal from the camera is digitized and stored using a Gould image processing system with a real-time digital disk (RTDD). The RTDD can store up to 3200 512 x 512 pixel images at video rates (30 images s-1). Images can be input directly from the camera during the study, or off-line from a Super VHS video recorder. Once a segment of experimental interest is digitized onto the RTDD, the user can interactively step through the images, average stable sequences, and identify candidates for further processing and analysis. Examples of how this system can be used to study the physiology of various organ systems in vivo are presented.