The Effect of Phacoemulsification Energy on the Redox State of Cultured Human Corneal Endothelial Cells

Objective  To evaluate the effects of phacoemulsification energy on the redox state and mitochondrial distribution of cultured human endothelial cells.

Methods  Human corneal endothelial cells from fresh banked human donor tissue not suitable for transplantation were harvested and cultured. Cellular autofluorescence images were obtained using an inverted microscope. The redox fluorometric ratio, which can be related to oxidative stress, was calculated as the net value of fluorescence from the 4,6-diamidino-2-phenylindole channel divided by the net value of fluorescence from the fluorescein isothiocyanate–conjugated channel after subtraction of background. For determining the mitochondrial distribution patterns, the cell area was divided by drawing a line halfway between the nuclear and cell membranes. The average fluorescence in the central area was divided by the average fluorescence in the peripheral area. This ratio was compared.

Results  Human corneal endothelial cells exposed to increasing phacoemulsification times and increasing ultrasonic energy levels displayed dose-dependent decreases in measured redox ratios. Lower redox ratios in response to phacoemulsification did not associate with decreases in cell size or altered patterns of mitochondrial localization.

Conclusion  Redox fluorometry may serve as a useful indicator for the in vitro study of human corneal endothelial cell physiological response to ultrasonic stressors and potentially other nonoxidative stressors.

Clinical Relevance  Redox fluorometry in combination with human corneal endothelial cell morphometric measurements has potential to serve as an indicator of human corneal endothelial cell injury resulting secondary to ultrasound phacoemulsification.

Ultrasound (US) phacoemulsification is currently the most widely used means by which cataract surgery in the developed world is carried out. Phacoemulsification is associated with a number of potentially injurious effects on human corneal endothelial cells (HCECs), including mechanical trauma,¹ bubbles,² and oxidative tissue damage.³ Endothelial cell damage during US phacoemulsification can interfere with early visual rehabilitation as well as lead to permanent corneal injury, including persistent edema requiring keratoplasty.

Currently, measurement of cell density,⁴ hexagonality,⁵ and size in vivo are all used to assess corneal damage following US phacoemulsification. Methods to evaluate the effect of phacoemulsification on cultured corneal endothelial cells have been described but are few. To our knowledge, the effect of phacoemulsification on the redox state in cultured human endothelial cells in vitro has not been previously reported. Such a method would provide a simple and quantitative means to experimentally assess the effects of current and evolving phacoemulsification techniques on HCECs in vitro as well as to screen for the effects of potential HCEC protective agents. The redox ratio is reported to be correlated with oxidative stress⁶ because it reflects changes in the relative amounts of reduced nicotinamide nucleotides and oxidized flavins.^7,8 The same ratio has not been evaluated in response to mechanical trauma. Observed differential effects on HCECs may provide insights into the differences between oxidative and mechanical stressors on HCECs. In this study, we evaluate the effects of phacoemulsification, as a function of time and power, on the redox ratio, intracellular mitochondrial distribution, and average HCEC size (area) in cultured HCECs.

Endothelial cells were cultured according to previously published methods.^9-12 Corneal endothelial cells from fresh banked human donor tissue not suitable for transplantation (Central Florida Lions Eye and Tissue Bank, Tampa, Florida, and Tissue Banks International, Baltimore, Maryland) were harvested attached to the Descemet membrane on or before the seventh day after death. The age of the donor tissue was 56 years. After overnight incubation of the Descemet membrane/endothelial cell complex in OptiMem-I (Gibco, Invitrogen Corp, Carlsbad, California) containing 8% fetal bovine serum, the complex was centrifuged and washed in Hank balanced salt solution (Mediatech, Inc, Herndon, Virginia). Next, the endothelial cells and Descemet membrane complex were incubated for 1 hour in 0.02% of EDTA solution, stirred vigorously with a flame-polished pipette to disrupt cell junctions, centrifuged for 5 minutes at 3000g (g-force), and seeded onto culture plates coated with FNC coating mix (Athena Enzyme Systems, Baltimore) containing bovine fibronectin (10 mg/mL) and bovine type I collagen (35 mg/mL). The cells were cultured in OptiMem-I media supplemented with 8% fetal bovine serum, calcium chloride (200 mg/L), chondroitin sulfate (0.08%), ascorbic acid (20 μg/mL), pituitary extract (100 μg/mL), epidermal growth factor (5 ng/mL), nerve growth factor (20 ng/mL), gentamicin (1:200), penicillin (1:100), streptomycin (1:100), and amphotericin (1:100) under 10% carbon dioxide. Medium was changed every 2 days. At confluence, the cells were split 1 to 3, and passage 4 cells were used for experiments. The experiments were performed at 70% confluence.

The HCECs, cultured in 8-well chamber slides (Laboratory-Tek II chamber slide; PGC Scientific Corp, Gaithersburg, Maryland), were washed with phosphate-buffered saline (PBS) and fixed for 20 minutes with 4% paraformaldehyde solution. Cells were permeabilized with 0.1% Triton X-100 for 10 minutes and blocked with 5% donkey serum for 1 hour at room temperature. After washing, cells were incubated with rabbit polyclonal antibody to the α 2 chain of bovine collagen VIII (1:1500; kindly provided by Paul F. Davis, Wellington, New Zealand) in 1% donkey serum overnight at 4°C and then washed with PBS. Cells were incubated with Cy3-conjugated donkey antirabbit IgG antibody (1:100) for 1 hour at 37°C in the dark. After extensive washing with PBS, the slide was mounted in a drop of Vectashield mounting medium with 4,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, California) to reduce photobleaching. Negative control staining was performed in parallel with the omission of primary antibodies.

For zonula occludens-1 (ZO-1) studies, the cells were plated on 8-well chamber slides precoated with cell attachment reagent (FNC coating mix) and were allowed to grow to confluence in corneal endothelial cell culture medium. The cells were then fixed in 2% paraformaldehyde for 20 minutes and treated with 0.1% Triton X-100 for 15 minutes at room temperature. Nonspecific antibody binding was blocked by incubating cells in 5% donkey serum for 1 hour before incubation with the primary rabbit antibody against ZO-1 (Zymed Laboratories Inc, San Francisco, California) overnight at 4°C. The cells were further incubated in fluorescein isothiocyanate–conjugated (FITC) donkey antirabbit secondary antibody (1:100; Jackson ImmunoResearch, West Grove, Pennsylvania) for 1 hour at room temperature and were counterstained with Hoechst nuclear staining dye (1:2000; Molecular Probes, Carlsbad) according to the manufacturer's recommendations. All images were obtained using an inverted microscope (Axiovert 200M; Zeiss, Thornwood, NY).

The Alcon Legacy unit was used (Alcon Surgical, Fort Worth, Texas). A phacoemulsification probe with a 30° round, 1.1-mm TurboSonics ABS Tip (Alcon Surgical) was introduced into culture dishes, taking care to avoid touching the cells at a predetermined and constant distance from the dish culture surface of 2 mm, which was carefully maintained by sustaining the position of the handpiece under the operating microscope after marking the probe. The cells in 2-mL media without serum were treated with 30% US phacoemulsification for 0, 3, 5, and 5 seconds and treated with 0%, 30%, 50%, and 70% US power for 3 seconds. The cells selected for analysis were taken from the same area of the culture plate such that the distance between the probe and the cells examined was constant.

Cellular autofluorescence images were obtained using an inverted microscope (Axiovert 200M). The microscope was equipped with a mercury lamp (HB 103) and a cooled charge-coupled device camera (Axiocam MR5; Zeiss) for taking images. A DAPI filter set (excitation, G365; emission bandpass, 445/50) was used to detect intrinsic reduced pyridine nucleotides and an FITC filter set (excitation bandpass, 450-490; emission bandpass, 515-565) was used to detect oxidized flavoproteins. To minimize photobleaching and light stimulation, the illumination source was turned off during fluorescence imaging. All the images were processed and analyzed using AxioVision 4.5 software (Zeiss). Prior to autofluorescence microscopy imaging, all cells were equilibrated in balanced salt solution (Alcon) and then imaged at room temperature under room air.

After outlining the cell border, the average intensities of cellular and background fluorescences were automatically calculated by the AxioVision 4.5 software. Twenty cells were randomly selected in each group and analyzed. The net value of cellular fluorescence was obtained by subtracting background intensity from cellular intensity. The redox ratio, which is inversely proportional to the cellular metabolic rate, was determined as the net value of fluorescence from the DAPI channel divided by the net value of fluorescence from the FITC channel (Figure 1).

The cell area was divided by drawing a line halfway between the nuclear and cell membranes. The net value of cellular fluorescence in each of the 2 zones was obtained by subtracting background intensity from cellular intensity. The average fluorescence in the central area (CF) was divided by the average fluorescence in the peripheral area (PF). The CF:PF ratios in the range of more than 1.70 were defined as having a perinuclear mitochondrial arrangement (Figure 2A), while ratios of less than 1.70 defined the homogenous mitochondrial pattern (Figure 2B). The CF:PF ratio was compared.

Cell viability was assessed by cell counting with trypan blue (Gibco, Invitrogen Corp) staining. Four dishes per each group were used. After the cells were trypsinized (0.025% Trypsin-EDTA), the cells were suspended in media. To determine the number of live cells, cells were stained with 0.4% trypan blue, and the unstained live cells and stained dead cells were counted with a hemocytometer. The average cell counts of 4 culture plates were recorded.

Data are expressed as mean (SD). Comparisons between controls and treated groups were performed using the Mann-Whitney test and χ² test. A P value <.05 was considered statistically significant.

Corneal endothelial phenotype was verified by the intense positive staining with type VIII collagen α 2 monoclonal antibodies (Figure 3A). The bright red signal within cytoplasm indicates collagen VIII α 2 synthesis. Immunocytochemical study of ZO-1, a protein associated with tight junctions, demonstrated its presence in cultured cells (Figure 3B).

Redox ratios decreased with longer phacoemulsification time and increased power. Redox ratios of cells at 5 seconds and 7 seconds with 30% US power were lower compared with normal control (Figure 4) (P = .002 and <.001, respectively, Mann-Whitney test). Cells with 50% and 70% US power at 3 seconds showed lower redox ratios compared with normal control (Figure 5) (Table) (P < .001 and <.001, respectively, Mann-Whitney test).

The percentage of cells with a perinuclear mitochondrial distribution was not different between groups.

The percentage of nonviable, stained cells increased with longer phacoemulsification time (Figure 2A) (P = .01, .01, and .01, Mann-Whitney test) and stronger power (Figure 2B) (P = .01, .01, and .01, Mann-Whitney test).

There was no difference in size between normal control and cells with US phacoemulsification (Figure 6).

There have been various methods to evaluate the effects or damage of US phacoemulsification on endothelial cells in vivo including cell density,⁴ morphology,⁵ corneal volume,¹³ and corneal thickness.⁴ To our knowledge, the method of redox autofluorescence microscopy to evaluate the effect of US phacoemulsification on endothelial cells in vitro has not been described. This method is an important next step as, unlike prior morphometric measures, the data derived from this method are potentially reflective of both a physiological state of the injured HCEC as well as a potential indicator of the cell's ability to respond to the stressor being studied.

One mechanism by which phacoemulsification energy can reduce the redox ratio in cells is by generation of extracellular reactive oxygen species (ROS), such as hydroxyl radical and other free radical species. These are known to be formed during phacoemulsification and to come into contact with the HCECs.^14-17 Cameron et al¹⁶ reported that such radicals generated near cellular membranes may also result in protein oxidation as well as lipid oxidation. These free radicals can act as external oxidative stressors in a manner similar to exposure to a chemical oxidant such as tert-butylhydroperoxide (tBHP). Tert-butylhydroperoxide is now understood to cause oxidation of mitochondrial pyridine nucleotides followed by mitochondrial production of ROS.^18,19 Detection of redox fluorometry shifts induced by tBHP in HCECs and other ocular cells has now been well described.^6,20 Moreover, the technique has sufficient sensitivity to detect the intracellular redox shifts induced by small changes in extracellular tBHP concentration.⁶

A second mechanism by which phacoemulsification energy can reduce the redox ratio in cells is through phacoemulsification-induced intracellular apoptosis-mediated ROS. Apoptotic processes can decrease intracellular redox ratios due to oxygen consumption and oxidative stress in the presence of complex I or complex II substrates that are consumed during the apoptotic process.^21,22 In the present study, redox ratios were significantly reduced in a dose-dependent manner in response to increasing phacoemulsification times as well as increasing phacoemulsification power, each reflecting an indicator of increasing total energy transfer to the HCEC.

Predictably, the percentage of nonviable, stained cells increased with both longer phacoemulsification times and higher phacoemulsification powers. Redox ratios were well correlated with the percentage of nonviable, stained cells. However, in the present study, cell size (area) was not significantly altered by either greater phacoemulsification time or increased ultrasonic power. Moreover, phacoemulsification did not have a significant effect on mitochondrial distribution patterns. While US phacoemulsification of HCECs results in shifts in the redox ratio that implicate ROS, the injury pattern is distinct from that seen following extracellular exposure of HCECs to the chemical oxidant tBHP. Concentrations of tBHP sufficient to cause both injury and apoptosis do result in similar redox ratio shifts but are also associated with characteristic changes in cell morphology such as decreased cell size (area) and a perinuclear pattern of mitochondrial distribution. Therefore, the redox fluorometry data presented herein support a role for oxidative injury as a component of phacoemulsification injury as well as implicate additional modes of injury in this complex HCEC stressor.

Finally, we used balanced salt solution as the medium when obtaining autofluorescence images. BSS Plus (Alcon) is supplemented with glutathione, which is thought to prevent oxidative stress placed on the cells.^23,24 Further study is necessary to investigate the effect of BSS Plus on the redox state in HCECs during phacoemulsification.

In this study, cultured cells, and not whole organs, were used. Monolayer cultured cells were necessary to obtain live cell images for redox ratio calculation and mitochondrial distribution determination. It is more difficult to perform such imaging in endothelial cells through hazier banked corneas. The purpose of this study was to compare different phacoemulsification parameters in vitro and to develop in vitro model conditions that allow quantitative assessment of phacoemulsification-induced endothelial cell damage and response to protective measures. Further experimental study as well as clinical correlation will be necessary to establish the clinical relevance of this novel laboratory test system. The ultrasound tip was directed at the cells at 2 mm, because some surgeons prolapse the cataractous lens into or above the iris plane rather than operate in the bag. Two millimeters, the distance of the tip from endothelial cells, was chosen to provide a “worst-case” scenario. These were pilot experiments, and geographic parameters were chosen arbitrarily.

As shown in Figure 3A, unlike in vivo, cultured HCECs under these environmental conditions were not hexagonal. Staining for ZO-1 also revealed a different distribution compared with the in vivo state. Many different methods to evaluate mitochondrial distribution, including transmission electron microscopy^24-26 and Mitotracker staining,^24,27,28 have been used. Although transmission electron microscopy can be helpful in mitochondrial evaluation, it was not necessary in this study and can be technically difficult and prone to artifact. In this study, standard trypan blue exclusion staining was used as a cell viability test. Other cell viability tests, which may or may not be more accurate, including fluorometric assays and the MTT (3-[4,5-dimethylthizol-2-yl] 2,5-diphenyltetrazolium bromide) assay, were not tested.^29-34

In conclusion, redox fluorometry in combination with HCEC morphometric measurements has potential to serve as an indicator of HCEC injury resulting secondary to US phacoemulsification. Differences in injury response between chemical oxidant injury and US phacoemulsification injury may provide fundamental insights into the mechanisms of injury involved. Redox fluorometry is at present an in vitro, inexpensive, rapid, and reproducible means to quantitatively assess HCEC injury following US phacoemulsification and may prove a useful means by which to examine, modify, and optimize current and evolving phacoemulsification tools and techniques. The method may prove of particular value in the preclinical identification and screening of potential HCEC protective agents applicable to the prevention and treatment of corneal edema following cataract surgery.

Correspondence: Roy S. Chuck, MD, PhD, Wilmer Ophthalmological Institute, Johns Hopkins University, 255 Woods Bldg, 600 N Wolfe St, Baltimore, MD 21287 (rchuck1@jhmi.edu).

Submitted for Publication: September 4, 2008; final revision received January 11, 2009; accepted January 21, 2009.

Financial Disclosure: None reported.