Abstract
Purpose:
The long-term success of endothelial keratoplasty is related to the number of viable cells transplanted. Traditional methods of viability assessment combine global areas of cell death with sample cell densities. Inherent problems with this method occur as a result of variations in cell density from the center to the periphery and handling damage when attempting to divide and mount living tissue. We describe a method of endothelial assessment at an individual cell level that can be applied to the whole graft.
Methods:
Twelve Descemets membrane endothelial keratoplasty grafts were prepare using a standard peeling technique. Grafts where incubated with Hoechst 33342 (10 micro molar), ethidium homodimer-1 (4 micro molar), and calcein-AM (2 micro molar) in DMEM with 2% fetal calf serum and 36 degrees for 30 minutes. Grafts were imaged macroscopically using an upright fluorescent stereomicroscope and then transferred to a custom built curved viewing chamber using a ‘no touch’ technique. The graft was unrolled and covered with cohesive viscoelastic. The whole graft are was then imaged using an inverted Nikon fluorescent microscope with a x4 objective lens. Image stacks at 45-micron intervals were acquired and in focus areas were combined using the integrated software. 30 tiled images were combined to produce a complete map of the graft. Images were exported to the NIH ImageJ software package. Calcein AM staining areas in the FITC channel were thesholded and used to create an image mask. Ethidium and Hoechst staining cells were counted in the masked areas using the ITCN plugin and the total number of viable cells calculated. This was compared to the number of viable cells as calculated using the Calcein AM stained area and average cell density calculated using 5 representative fields. A Bland-Altman plot was used to compare the two methods.
Results:
No significant damage to the endothelial monolayer occurred using the no touch technique to mount and image the DMEK grafts The method described allowed every cell in the living graft to be imaged and used to determine total graft viability. A difference of up to 19% in viable cell number was observed with no clear linear relationship between the 2 methods.
Conclusions:
Traditional methods of assessing viability may be less accurate than the method described here which is able to determine viability on an individual cell level.