July 2011
Volume 52, Issue 8
Free
Cornea  |   July 2011
Pan-Corneal Endothelial Viability Assessment: Application to Endothelial Grafts Predissected by Eye Banks
Author Affiliations & Notes
  • Aurélien Pipparelli
    From the Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France;
    Department of Ophthalmology, University Hospital, Saint-Etienne, France;
  • Gilles Thuret
    From the Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France;
    Department of Ophthalmology, University Hospital, Saint-Etienne, France;
  • David Toubeau
    Normandy Regional Eye Bank and
  • Zhiguo He
    From the Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France;
  • Simone Piselli
    From the Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France;
  • Sabine Lefèvre
    Normandy Regional Eye Bank and
  • Philippe Gain
    From the Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France;
  • Marc Muraine
    Normandy Regional Eye Bank and
    Department of Ophthalmology, Charles Nicolles Hospital, Rouen, France.
  • Corresponding author: Gilles Thuret, CHU, service d'Ophtalmologie, avenue Albert Raimond, F-42055 Saint-Etienne Cédex 02, France; gilles.thuret@univ-st-etienne.fr
Investigative Ophthalmology & Visual Science July 2011, Vol.52, 6018-6025. doi:10.1167/iovs.10-6641
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Aurélien Pipparelli, Gilles Thuret, David Toubeau, Zhiguo He, Simone Piselli, Sabine Lefèvre, Philippe Gain, Marc Muraine; Pan-Corneal Endothelial Viability Assessment: Application to Endothelial Grafts Predissected by Eye Banks. Invest. Ophthalmol. Vis. Sci. 2011;52(8):6018-6025. doi: 10.1167/iovs.10-6641.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To present an experimental method for determining the viable cell pool of corneal endothelia and its application to assessing predissected endothelial grafts.

Methods.: The endothelial cell density (ECD) of five pairs of human organ cultured corneas was determined using a standard counting method with a calibrated image analysis system. A thin posterior graft (30–50 μm) was manually predissected from a cornea chosen at random. Predissected and control corneas were shipped to the remote center, where standard ECD determination was repeated and was immediately followed by a triple Hoechst/ethidium/calcein labeling coupled with image analysis of the whole graft surface. Numeration of nuclei (H+), dead cells (E+), and total area covered by viable cells (C+) allowed the calculation of viable ECD corresponding to the cell density that the cornea may have after redistribution of viable cells over the whole Descemet surface.

Results.: The median (range) viable ECD was lower than the standard ECD determined immediately earlier in predissected and control corneas: 1628 (1138–2379) and 2065 (1492–2876) cells/mm2 (P = 0.043), corresponding to −20% (−1%-38%) and −12% (−3%-26%), respectively (P = 0.08).

Conclusions.: Standard counting by eye banks overestimates the actual pool of viable endothelial cells. This may be the main explanation for the initially rapid decrease in ECD universally described in patients after all types of keratoplasty. Early low postoperative ECD may indicate that surgeons graft fewer living cells than the eye banks' ECD let suppose, rather than a massive pre- and postoperative cell death. The novel concept of viable ECD can be useful for assessing all types of corneal processing.

Eye bank assessments of endothelial graft quality during storage is based on three criteria: cell density, cell morphometry, and the detection of any endothelial abnormalities such as guttae. 1 A fourth criterion, applied by a small proportion of the banks using organ culture (OC), primarily in Europe, is assessment of endothelial mortality after trypan blue staining. 2,3  
Endothelial cell density (ECD) is the main recognized criterion for corneal transplantation. Its threshold value, between 2000 and 2400 cells/mm2, 2 4 was determined empirically in the 1980s for the delivery of grafts for penetrating keratoplasty (PKP). It is the best compromise between acceptable graft survival time in the recipient and a discard rate compatible with eye banks' medical economics. Morphometric parameters (polymegethism and polymorphism) provide additional information, 5 although their influence on postoperative survival has yet to be clearly established. Trypan blue vital staining is used primarily to detect striae, or lines of dead endothelial cells (ECs), that may indicate trauma during removal or handling at the bank or, exceptionally, herpetic endothelial necrosis. 6 Incidentally, some banks use it to measure instant endothelial mortality (percentage of ECs with blue-stained nuclei), but this information has low relevance. It has been clearly demonstrated that trypan blue does not highlight the onset of apoptosis in ECs. 7  
In eye banks, ECD is most often determined by computer analysis of specular microscopy (SM) images for cold storage at +4°C 8,9 (primarily in the United States) and of transmitted light microscopy (TLM) images for OC storage. 5,10 14 Although counting software programs have become widespread in recent years, this routine ECD imperfectly reflects the pan-corneal pool of viable cells, for four reasons: (1) Only visible ECs can be counted, whether spontaneously visible in SM or made visible by osmotic preparation with 0.9% sodium chloride or 1.8% sucrose for TLM. 2 However, areas in which ECs are invisible often correspond either to desquamated ECs (particularly in corneal folds) 7 or to areas in which cell contours are invisible or poorly visible either spontaneously or after osmotic preparation. (2) EC contour recognition is often difficult and may distort counting, despite computer extrapolation (pointing of EC centers in SM) or drawing (completing cell contours in TLM). 15 (3) Counting is done on a very small EC sample (50–300), i.e., less than 1/1000th of the pan-corneal endothelial cell pool in adults (200,000 to 300,000 cells). (4) Routine cell density determination takes account of EC viability either very imperfectly (OC) or not at all (cold storage). These four reasons together suggest it is likely that routine ECD substantially overestimates the pan-corneal pool of viable ECs. 
Several experimental methods have been proposed since alizarin red by Sperling in the 1970s for assessing endothelial quality during corneal storage or in other applications (staining described in 1946 by Vonwiller 16 ). These are summarized in Table 1. None measures pan-corneal endothelial viability, unlike the method presented here. 
Table 1.
 
Experimental Methods for In Situ Assessment of Corneal Endothelial Quality
Table 1.
 
Experimental Methods for In Situ Assessment of Corneal Endothelial Quality
Reference Year Method Species Application
Aquavella 17 1975 Trypan blue, SEM H Stored corneas
Sperling 18 1977 Alizarin red + trypan blue H Stored corneas
Basu 19 1978 Trypan blue, para nitroblue tetrazolium, SEM H, R, GP Stored corneas
Binder 20 1978 Nitroblue tetrazolium, SEM, TEM R Stored corneas (postmortem cell damage)
Schrapel 21 1982 Toluidine blue P Stored corneas
Singhl 22,23 1985 Trypan blue H, P Stored corneas (toxicity tests)
Madden 24 1987 Trypan blue, nitroblue tetrazolium, acridine orange, fluorescein diacetate, ethidium bromide, SEM H Stored corneas
Hartmann 25 1989 Janus green P Stored corneas (toxicity tests)
Means 26 1995 Alizarin red + trypan blue, calcein-AM + ethidium homodimer H Stored corneas
Salla 27 1995 Succinate dehydrogenase staining H Stored corneas
Wusteman 28 1997 Acridine orange + propidium iodide, confocal microscopy R Stored corneas
Kent 29 1997 Calcein-AM + ethidium homodimer H High-diopter myopic PRK or LASIK
Wusteman 30 1999 Acridine orange + propidium iodide, nuclear magnetic resonance spectroscopy P Cryoconservation
Albon 31 2000 Hoechst, TUNEL assay, immunostaining (active caspase 3) H Stored corneas
Koh 32,33 2000 Calcein-AM + ethidium homodimer H, B Stored corneas
Gain 7 2002 Trypan blue + TUNEL assay H Stored corneas
Engelman 34 2004 Immunostaining (integrins and tight junction), perfusion studies H Bioengineered endothelium
Joyce 35 2004 Calcein-AM + ethidium homodimer, TUNEL R Viability after endothelial transfection
Sikder 36 2006 Calcein-AM + ethidium homodimer H Precut lamellar graft (femtosecond laser)
Steinhardt 37 2006 Calcein-AM B Stored corneas
Suwan-Apichon 38 40 2006 Alizarin red + trypan blue, ultrasonic pachymetry H Precut lamellar graft (manual, femtosecond laser, DSAEK)
Ide 41,42 2007/8 Alizarin red + trypan blue H Whole lamellar graft process (DSAEK)
Mehta 43 2008 Trypan blue, SEM H Lamellar graft introduction techniques (DSAEK)
Slettedal 44 2008 Immunostaining (n-cadherin), SEM H Stored corneas
Amato 45 2009 Trypan blue H Precut lamellar graft (DSAEK)
Kim 46 2009 Alizarin red, pachymetry, SEM P Femtosecond laser trephination for PKP
Proulx 47 2009 Alizarin red, immunostaining (Na/CO3, ZO-1, Na/K-ATPase), SEM, TEM F Bioengineered endothelium
Yoeruek 48 2009 Calcein-AM + ethidium homodimer, immunostaining (ZO-1, connexin-43, Na/K-ATPase, and cytokeratin-3), phase-contrast microscopy H Bioengineered endothelium
Wolf 49 2009 Calcein-AM + ethidium homodimer + Hoechst, * pachymetry, optical coherent tomography H Stored corneas (deswelling)
He 50 2010 Calcein-AM + ethidium homodimer + Hoechst,* immunostaining (ZO-1) H Endothelial transfection (electroporation of stored corneas)
Endothelial keratoplasty (EK) is fast being adopted in the United States and Europe. 3 In the United States, operations rose by 135% between 2006 and 2007 (14,159 in 2007) 51 and reached 18,221 in 2009 (i.e., 43% of grafts delivered by banks). 52 This procedure, offered mainly in cases of Fuchs' dystrophy and pseudophakic corneal dystrophies, significantly reduces complications reported after PKP (astigmatism, scar fragility, rejection) and allows faster visual recovery. 53,54 The diseased endothelium is removed through a small incision and is replaced by an endothelial graft that may include a stromal lamella. The graft can be dissected manually using a microkeratome or a femtosecond laser. Descemet stripping automated endothelial keratoplasty (DSAEK) using a microkeratome is the most common technique because it is the easiest to perform, but it obtains thicker grafts than do manual techniques and is far more costly. Manual dissection obtains extremely thin grafts, and even endothelium-Descemet membrane complexes; the latter allow greater visual recovery 55 by limiting interface and hypermetropization phenomena. However, the difficulty with the manual technique (which involves a steep learning curve) is restricting its adoption by surgeons. Consequently, even more than for DSAEK, 56 manual preparation of grafts by cornea bank specialists highly skilled in this technique and shipment to remote operating theaters would limit these drawbacks. 
This article presents an experimental method for pan-corneal measurement of the viable EC pool supplied by the graft and its application to assessing endothelial grafts predissected manually by an eye bank and sent to a remote location. 
Materials and Methods
Human Corneas
Five pairs of healthy human corneas assigned to scientific use and procured by the Normandy Eye Bank (Rouen, France [center 1]) were used after informed consent of the relatives, as authorized by French bioethics laws. All procedures conformed to the tenets of the Declaration of Helsinki for biomedical research involving human subjects. Median donor age was 83 years (range, 49–89 years), which is typical in France. 3 The male/female ratio was 1.5. Median time from death to procurement was 11 hours (2 hours 50 minutes to 13 hours 45 minutes). Corneas were procured by in situ excision with 16- to 18-mm-diameter trephination, according to the procedure recommended in France for corneas intended for transplantation. They were then immediately placed in 100 mL of the OC medium (Corneamax; Eurobio, Les Ulis, France) at +31°C in a dry incubator, in accordance with the OC preservation procedure of most European eye banks. Median storage time before dissection was 18 days (range, 13–25 days), close to the mean storage time observed in most European eye banks. 3 For each pair, one cornea chosen at random was prepared as described here and was immediately returned to the OC medium. The mate cornea (control) remained in its storage medium. Both corneas were sent at +31°C (controlled and registered temperature) by car transporter to the remote center 398 miles (640 kilometers) away (Laboratory for Corneal Graft Biology, Engineering and Imaging, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France [center 2]), replicating the typical conditions of shipping to a remote operating theater. 
Manual Endothelial Dissection
At center 1, dissection was performed by a sole experienced operator (MM) using a technique published elsewhere. 57 Briefly, the graft was placed, endothelial side up, on an artificial anterior chamber (Moria, Antony, France). A surface mark 8.25 mm in diameter was then made with a disposable trephine blade. For easier viewing of the endothelium-Descemet membrane complex, the endothelium was covered with 0.4% trypan blue (Eurobio) for 1 minute, then washed with balanced salt solution (BSS; Alcon, Rueil-Malmaison, France). The anterior chamber was then filled with air to invert the cornea, which became convex in shape, endothelium side up. During dissection, the endothelium was constantly bathed by BSS. Lamellar dissection was then performed using Melles-type spatulas (Dorc, Zuidland, Netherlands), passing just below the Descemet membrane. Dissection was begun with a pointed spatula and continued with a foam-tipped spatula, remaining in the same cleavage plane in the posterior stroma, just below the endothelium. This technique allowed dissection of a very thin posterior graft of 30 to 50 μm (histologic and postoperative confocal microscopy data). 57 The graft perimeter was then cut using Vannas micro scissors, after the superficial trephination mark was initially made. Because the objective was to ship the graft to a distant location, the cut was left incomplete so as to prevent detachment of a graft in the OC medium. The 10° hinge was marked by a nick next to it on the sclera. After predissection, the cornea was immersed in a new bottle of medium and shipped to center 2. All endothelial lamellae remained adherent to the underlying stroma during shipping. However, on receipt, detachment was easy and atraumatic after sectioning of the hinge. Both corneas of each pair were assessed in parallel with two complementary methods: a standard cell density determination was performed at center 1 before dissection; at center 2, after a mean of 3.8 ± 0.8 days (4; range, 3–5 days), the same standard cell density determination was repeated, and immediately afterward a specific pan-corneal endothelial live/dead assay was performed. 
Standard Cell Density Determination (Centers 1 and 2)
The determination was performed by a single observer (SP), as described elsewhere. 10,12 Briefly, after washing in BSS, corneas were placed endothelial side up in a sterile petri dish. Dead cells were identified using 0.4% trypan blue only, to eliminate corneas with extensive EC necrosis. The endothelial surface was incubated with 0.9% sodium chloride (Aguettant, Lyon, France) for 4 minutes to dilate the intercellular spaces. Once the cell contours were optimally discernible, the endothelium was viewed through a long working distance ×10 objective using a light direct microscope (Eclipse 50i, Nikon Corp., Japan, in center 1; DMLB, Leica Microsystems GmbH, Wetzlar, Germany, in center 2). Endothelial photographs were acquired using a monochrome CCD video camera (XC-ST50CE; Sony Corp., Tokyo, Japan) installed in both centers and digitized using a video frame grabber (DT-3155; Data Translation, Marlboro, MA). Three wide-field (1000 × 750 μm) images of three randomly chosen nonadjacent zones of the endothelium contained within the central 8-mm-diameter were taken at 768 × 576 pixel resolution in 8-bit gray level and saved in bitmap format. Image analysis was centralized at center 2 after software calibration of the two microscopes. ECDs were determined on the three wide-field images with the intensively validated computer-assisted eye bank endothelial analyzer (Sambacornea; Tribvn, Chatillon, France), using a semiautomated method, on >300 EC with manual correction of cell contours when necessary. 5,10,12 14 This method has low intraobserver and interobserver variability. 13  
Pan-Corneal Endothelial Live/Dead Assay (Center 2)
Triple Labeling.
The Hoechst/ethidium homodimer/calcein-AM combination was used (Fig. 1). Hoechst 33342 (H), a vital nucleic acid dye, is a common cell-permeant nuclear counterstain that emits blue fluorescence when bound preferentially to A-T base-pairs of double-stranded DNA. It is excited by ultraviolet light at around 350 nm and emits blue/cyan fluorescent light with a peak of 461 nm. Ethidium homodimer-1 (E) enters cells with compromised membranes and is fluorescent when bound to nucleic acids. Ethidium produces bright red fluorescence in the nuclei of dead cells. The dye is excited at 528 nm and emits red fluorescent light with a peak of 617 nm. Calcein acetoxymethyl (calcein-AM) (C) is a nonfluorescent, hydrophobic compound that easily permeates intact live cells. Hydrolysis of Calcein-AM by ubiquitous intracellular esterases in live cells produces a hydrophilic green fluorescent product. Hydrolyzed calcein is retained within cells. It is excited at 495 nm and emits green fluorescent light with a peak of 520 nm. The corneas, placed endothelial side up in a sterile petri dish, were incubated for 45 minutes at +31°C with 100 μL Hoechst 33342 (10 μM), ethidium homodimer-1 (4 μM), and calcein-AM (2 μM) in phosphate-buffered saline (PBS) and then were gently rinsed in PBS. The endothelium was coated with a viscoelastic substance (Viscoat; Alcon) to avoid artifactual lesions during subsequent manipulation. For predissected grafts, the hinge was cut with Vannas micro scissors. The endothelial lamella was gently transferred onto a glass slide under an operating microscope. Controlled corneas were trephined at 8.25 mm (Hessberg-Baron trephine). In all grafts, a radial cut was made to allow a flat mount without folding under a coverslip using an antifading agent (Fluorescent Mounting Medium; DakoCytomation, Glostrup, Denmark). For the full-thickness buttons, the coverslip was held by adhesive tape. 
Figure 1.
 
Triple endothelial labeling with Hoechst 33342 (H), ethidium homodimere (E), and calcein-AM (C). ×10 objective. Green arrowhead: living cell C positive. Red arrowhead: dead cell E positive. Blue arrowhead: dying cell without metabolic activity (C negative) but still adherent, either isolated or at fold edges. White arrowheads: fold deprived of cell. Scale bar, 200 μm.
Figure 1.
 
Triple endothelial labeling with Hoechst 33342 (H), ethidium homodimere (E), and calcein-AM (C). ×10 objective. Green arrowhead: living cell C positive. Red arrowhead: dead cell E positive. Blue arrowhead: dying cell without metabolic activity (C negative) but still adherent, either isolated or at fold edges. White arrowheads: fold deprived of cell. Scale bar, 200 μm.
Pan-Corneal Image Acquisition for Viable Cell Surface Area Determination.
Tagged image format file images were acquired using a microscope (IX81; Olympus, Tokyo, Japan), with motorized XYZ, run by imaging software (CellP; Cell Imaging, Hamburg, Germany). Acquisition of a mosaic of 40 overlapping images using a ×4 objective made it possible, after reconstruction, to view the whole surface of the grafts (Fig. 2). Viability was measured on the whole graft surface, after manual tracing of the circumference but excluding the artifact area bordering the radial incision required for flat mounting. Total median analyzed surface was 57 mm2 (range, 53–64 mm2) for both groups. After background noise removal, signal thresholding, and binarization, the calcein-stained viable endothelial surface area was automatically calculated, as was the ratio of viable surface area to total analyzed area (percentage of graft surface area covered by live ECs). 
Figure 2.
 
Pan-corneal image acquisition for viable cell surface determination. (1) Acquisition of the entire graft surface with 40 overlapping images after calcein staining (×4 objective). (2) Reconstruction. (3) Extraction of green channel. (4) Delineation of graft contour by elimination of the radial cutting area necessary for the flat mount. (5) Standardized thresholding. (6) Binarization. (7) Automatic calculation of surface area in mm2. (8, 9) Representative paired grafts. (8) Full-thickness graft (control). Folds with no living cells were clearly visible. The radial cut necessary for the flat mount caused artifactual cell destruction, visible as a darker area around the cut (white arrowheads). This was more pronounced than with the lamellar graft because of stromal compression during cutting and was, therefore, excluded during image analysis. The blurred surrounding rim (blue arrow) is attributed to stromal autofluorescence (absent in 9). Yellow arrow: cell damaged by trephination. (9) Endothelial graft. Folds with no living cells were also present. Red arrow: hinge that kept the graft attached to the stroma for shipping and was then cut after staining. Green: percentages of areas covered by living cells. Median viable endothelial surface was higher in the control group than in the predissected group, respectively 90% (86%–92%) versus 82% (80%–87%) (P = 0.043).
Figure 2.
 
Pan-corneal image acquisition for viable cell surface determination. (1) Acquisition of the entire graft surface with 40 overlapping images after calcein staining (×4 objective). (2) Reconstruction. (3) Extraction of green channel. (4) Delineation of graft contour by elimination of the radial cutting area necessary for the flat mount. (5) Standardized thresholding. (6) Binarization. (7) Automatic calculation of surface area in mm2. (8, 9) Representative paired grafts. (8) Full-thickness graft (control). Folds with no living cells were clearly visible. The radial cut necessary for the flat mount caused artifactual cell destruction, visible as a darker area around the cut (white arrowheads). This was more pronounced than with the lamellar graft because of stromal compression during cutting and was, therefore, excluded during image analysis. The blurred surrounding rim (blue arrow) is attributed to stromal autofluorescence (absent in 9). Yellow arrow: cell damaged by trephination. (9) Endothelial graft. Folds with no living cells were also present. Red arrow: hinge that kept the graft attached to the stroma for shipping and was then cut after staining. Green: percentages of areas covered by living cells. Median viable endothelial surface was higher in the control group than in the predissected group, respectively 90% (86%–92%) versus 82% (80%–87%) (P = 0.043).
Viable ECD Determination.
Five additional images (a central image and one in each quarter) were acquired with a ×10 objective for the Hoechst and ethidium stainings, which required higher magnification than calcein. In homogeneous areas of these five manually delineated images, all the positive Hoechst nuclei were counted by automatic detection with manual correction, if necessary. Positive ethidium nuclei were counted using the same principle in five other fields randomly selected.(Fig. 3). Image analysis was performed by a single observer (AP) masked to the group. These measurements were made on 3359 (range, 2457–3964) versus 2252 (range, 2114–3631) (P = 0.225) nuclei in the control and predissected groups, respectively. The mortality rate was the number of positive ethidium nuclei out of the total number of nuclei (Hoechst + ethidium). Density of nuclei (nuclei/mm2) in the homogeneous areas corresponded to the ratio between the number of nuclei and the delineated measurement area. The average of the five images was considered. Viable ECD (number of viable ECs/mm2) was calculated by multiplying this density of nuclei by the percentage of viable endothelial surface area. Lastly, the total number of viable ECs on each graft was calculated by multiplying this viable ECD by the total surface area. 
Figure 3.
 
Procedure for calculation of endothelial cell density. (A) Example of 1 of 5 images taken with a ×10 objective, one central and one per quarter. (B) Selection of Hoechst acquisition channel. (C) Delineation of largest possible region of interest by eliminating areas without nuclei to determine homogeneous density of nuclei. These areas without nuclei corresponded to areas of naked Descemet, mostly folds, without ECs. They were by definition eliminated during the processing of calcein images (see Fig 2); thus, it was logical not to select them to determine the number of nuclei per surface unit. (D) Counting of nuclei (here, 1099). (E) Selection of ethidium acquisition channel. (F, G) Similar to the previous Hoechst process. Median mortality rate (ethidium+) did not differ between the control group and the predissected group, respectively 0.05% (0%–0.28%) versus 0.12% (0.06%–0.30%) (P = 0.068).
Figure 3.
 
Procedure for calculation of endothelial cell density. (A) Example of 1 of 5 images taken with a ×10 objective, one central and one per quarter. (B) Selection of Hoechst acquisition channel. (C) Delineation of largest possible region of interest by eliminating areas without nuclei to determine homogeneous density of nuclei. These areas without nuclei corresponded to areas of naked Descemet, mostly folds, without ECs. They were by definition eliminated during the processing of calcein images (see Fig 2); thus, it was logical not to select them to determine the number of nuclei per surface unit. (D) Counting of nuclei (here, 1099). (E) Selection of ethidium acquisition channel. (F, G) Similar to the previous Hoechst process. Median mortality rate (ethidium+) did not differ between the control group and the predissected group, respectively 0.05% (0%–0.28%) versus 0.12% (0.06%–0.30%) (P = 0.068).
Statistical Analysis
Considering the sample size, data are expressed as median (range). Quantitative data were compared using nonparametric tests (Wilcoxon for paired data) using statistical software (SPSS 11.5; SPSS, Chicago, IL). P < 0.05 was deemed significant. 
Results
Standard ECD
In center 1, endothelial baseline ECDs before dissection were comparable, with 2610 (range, 1898–3611) versus 2619 (range, 1675–2877) cells/mm2 in the control and predissected corneas respectively (P = 0.686). Cell morphometry was also comparable (data not shown). 
After shipping to center 2, the EC image quality of predissected corneas was poor, with cells scarcely visible because of poor intercellular space dilation and high background noise caused by trypan blue staining of stromal lamellae. Nevertheless, standard cell density determination remained possible and was performed with complete manual drawing of cell contours. ECDs were 2328 (range, 1935–3252) versus 2043 (range, 1235–2793) cells/mm2 (P = 0.08), respectively, for control and predissected corneas, corresponding to cell losses of 7% (range, −6%–14%) and 12% (−7%–29%), respectively (P = 0.345). These cell losses were attributed to endothelial cell damage during the dissection and shipping procedures. Cell morphometry remained comparable (data not shown). 
Viable ECD and Comparison with Standard ECD
Viable ECD was 2065 (range, 1492–2876) versus 1628 (range, 1138–2379) cells/mm2, significantly lower in the predissected group (P = 0.043). Compared with standard ECD determined immediately before HEC labeling, viable ECD was 12% (range, 3%–26%) lower for control corneas and 20% (range, 1%–38%) for predissected corneas (P = 0.08). Considering the areas of the grafts, the median number of viable ECs on the graft was 116,101 (85,576,164,792) in controls versus 100,300 (65,885,129,940) on predissected corneas (P = 0.043). The main results are shown in Figure 4
Figure 4.
 
Box plots of the ECD measurement before and after predissection and shipping to a remote center and comparison with viable ECD determined immediately thereafter using triple labeling by Hoechst/ethidium/calcein staining (HEC) combined with pan-corneal analysis. Horizontal line: the threshold of 2000 cells/mm2 conventionally used to deliver grafts for penetrating keratoplasty. Circle: case at >1.5 box length from the upper or lower edge of the box.
Figure 4.
 
Box plots of the ECD measurement before and after predissection and shipping to a remote center and comparison with viable ECD determined immediately thereafter using triple labeling by Hoechst/ethidium/calcein staining (HEC) combined with pan-corneal analysis. Horizontal line: the threshold of 2000 cells/mm2 conventionally used to deliver grafts for penetrating keratoplasty. Circle: case at >1.5 box length from the upper or lower edge of the box.
Discussion
The technique of HEC triple staining combined with image analysis of the whole endothelium allows us to introduce the original concept of viable ECD. Applied for the first time to assessing the pan-corneal viable cell pool of grafts delivered by eye banks, it confirms that standard ECD overestimates the viable EC pool of grafts, corresponding to the actual endothelial pool, regardless of whether these are sent unhandled for PKP (+12% of ECD) or are predissected by the eye bank for EK (+20%). This study suggests that standard ECD overestimation is the main explanation for the initial rapid decrease in ECD universally described in patients, whatever the type of keratoplasty. 
Until now, with one exception in which calcein + ethidium was used, 36 endothelial dissection methods have been assessed simply by double staining with alizarin red and trypan blue, 18 which does not allow reliable ascertainment of EC viability. Double staining with the calcein-AM/ethidium homodimer 58 60 has been widely validated for assessing the viability/toxicity of numerous cell types. Image analysis is simple thanks to the high intensity of label fluorescence. In our study we combined it with Hoechst 33342, which highlighted all cell nuclei irrespective of cell status (living, dying, or dead). This experimental destructive method (not usable by eye banks) for pan-corneal image analysis allows consideration of a cell sample (approximately 3000 ECs analyzed, or roughly 1% of all ECs) >10 times greater than is used in standard ECD, as well as the instant mortality rate and the endothelial surface actually covered by viable ECs. By combining these data, we can propose the concept of viable ECD (i.e., the actual pool of viable ECs on a graft). This also allows expression of the absolute number of viable ECs donated to the recipient. This novel concept corresponds to the ECD after all living ECs have recolonized, without dividing, the acellular spaces or those containing ECs that are committed to dying. Such acellular areas of varying extent, already described after endothelial dissection, 36,61 frequently exist even on non-predissected grafts, as shown by the acquired images of the whole control-graft surface (Fig. 3). It should be noted that the percentage of viable surface never reaches 100% (90% for the control group vs. 83% for the predissected group). Most of these acellular areas are located in hypotonia-induced folds. These folds became more pronounced with the edema that occurs during OC but were initially described in storage at +4°C in McCarey-Kaufmann medium. 62 The status of the present but unviable ECs, revealed by our study as ECs whose nuclei are well displayed by Hoechst but remain calcein-negative, remains to be defined, but they may be apoptotic. 7,31  
This new concept, applied to assessing grafts predissected by hand and sent to a remote location, shows that cell loss is greater than when a full-thickness graft is stored. Even when predissection is performed by an expert, as in our study, graft manipulations (removal from storage medium for approximately 10 minutes, exposure to air, inversion of corneal curve, insertion of spatula, lesions to underlying stroma, final hinge cut) and shipping time explain this additional mortality. Storage aside, full-thickness grafts are subject only to trephination-related trauma. 63  
The most important point, in our view, is the coherent difference observed between the standard ECD supplied by the eye bank to the surgeon and the viable ECD as we have defined it, with the rapid decrease in cells reported in the clinical series for both PKP and EK. All these data seem to indicate that the “cell loss” calculated postoperatively using ECD provided by the bank and specular microscopy performed on recipients primarily reflects the difference between the bank-estimated ECD and the actual viable EC pool on the graft (rather than massive EC death in the first postoperative days or weeks). For PKP, the difference in ECD of 12% ± 9% (mean ± SD) measured in our study is comparable with that reported in the early postoperative phase of series using OC: 17.5% (95% confidence interval, 13.1%–21.9%) in a series of 25 perforating grafts for which, in 75% of cases, ECD could be measured reliably at 5 days after graft, 64 and 9% ± 11% at 1 month postoperatively in another series of 24 perforating grafts. 65  
Interestingly, although we did not demonstrate it, the same observation can be made for series using 4°C stored corneas: 11% ± 20% at 6 months postoperatively in the Cornea Donor Study. 66,67 Because no difference in either clinical results or ECD has been shown between 4°C and OC stored corneas, 65,68 we think our results could be extrapolated to 4°C storage, where SM performed in eye banks may have the same gap with viable ECD as does TLM in eye banks using OC. This remains to be investigated. Similarly, for EK, the difference in ECD of 20% measured in our study is also comparable with the clinical series using either OC or 4°C stored corneas: a loss of 25% ± 15% at 6 months in a prospective series of 100 DLEKs reported by Terry 69 ; 30% at 6 months, then stable in the series of 100 DSAEKs using grafts predissected in-bank reported by Terry 56 ; 38% ± 22% at 6 months in the series of 173 DSAEKs reported by Price 67 ; 29% at 6 months in the series of 58 DMEKs reported by Melles 70 ; and 40% at 1 month in two series of 37 DMEKs with 4°C stored corneas and 45 OC corneas reported recently by Laaser. 71 Additional losses caused by surgical trauma during insertion of the endothelial graft 41,43 into the anterior chamber and unfolding maneuvers, not considered in our study, are responsible for the remainder of measured cell loss. Further, the hitherto unknown difference between standard ECD and viable ECD probably explains why the most recent studies find no correlation between the ECD of bank-supplied grafts and either post-PKP graft survival 72 or postoperative ECD after DSAEK. 73  
Predissection of endothelial grafts for DSAEK and shipping to a remote center is now common practice in the United States, using grafts stored at +4°C in medium (Optisol GS; Bausch & Lomb, Rochester, NY). In 2009, of the 18,221 grafts delivered for EK, 12,071 were predissected at an eye bank (12,037 by microkeratome, 9 manually, 25 by femtosecond laser). Postoperative cell loss with grafts predissected in-bank and sent to a remote center was comparable with that in graft series involving tissue cut intraoperatively. 56 There is no comparable data for grafts that underwent in-bank manual predissection of thin grafts or pure Descemet with shipping to a remote center. A sole European study of 10 pure Descemet grafts stored in OC for 4 weeks reported cell loss 1 week after dissection, measured at 3.7% ± 6.3% by routine cell density determination. Our series population did not allow the study of predissected ECD influence on final ECD. It remains to be determined whether it is necessary to select grafts with high initial ECD liable to better resist cutting trauma and the stress caused by storage and shipping. 
In conclusion, the technique of triple HEC staining coupled with pan-endothelial analysis allows us to introduce the novel concept of viable ECD as the best indicator of the functional EC pool given to the recipient. This technique can be applied to the analysis of any surgical or nonsurgical procedure that is liable to directly or indirectly alter the corneal endothelium. 
Footnotes
 Disclosure: A. Pipparelli, None; G. Thuret, None; D. Toubeau, None; Z. He, None; S. Piselli, None; S. Lefèvre, None; P. Gain, None; M. Muraine, None
References
Borderie VM Sabolic V Touzeau O Scheer S Carvajal-Gonzalez S Laroche L . Screening human donor corneas during organ culture for the presence of guttae. Br J Ophthalmol. 2001;85:272–276. [CrossRef] [PubMed]
Pels E Rijneveld WJ . Organ culture preservation for corneal tissue: technical and quality aspects. Dev Ophthalmol. 2009;43:31–46. [PubMed]
European Eye Bank Association. European Eye Bank Association Directory. 18th ed. Amsterdam; European Eye Bank Association; 2010.
Sugar J Montoya M Dontchev M . Donor risk factors for graft failure in the cornea donor study. Cornea. 2009;28:981–985. [CrossRef] [PubMed]
Acquart S Gain P Zhao M . Endothelial morphometry by image analysis of corneas organ cultured at 31°C. Invest Ophthalmol Vis Sci. 2010;51:1356–1364. [CrossRef] [PubMed]
Sengler U Reinhard T Adams O Krempe C Sundmacher R . Herpes simplex virus infection in the media of donor corneas during organ culture: frequency and consequences. Eye. 2001;15:644–647. [CrossRef] [PubMed]
Gain P Thuret G Chiquet C . Value of two mortality assessment techniques for organ cultured corneal endothelium: trypan blue versus TUNEL technique. Br J Ophthalmol. 2002;86:306–310. [CrossRef] [PubMed]
Lass JH Gal RL Ruedy KJ . An evaluation of image quality and accuracy of eye bank measurement of donor cornea endothelial cell density in the Specular Microscopy Ancillary Study. Ophthalmology. 2005;112:431–440. [CrossRef] [PubMed]
Benetz BA Gal RL Ruedy KJ . Specular microscopy ancillary study methods for donor endothelial cell density determination of Cornea Donor Study images. Curr Eye Res. 2006;31:319–327. [CrossRef] [PubMed]
Gain P Thuret G Kodjikian L . Automated tri-image analysis of stored corneal endothelium. Br J Ophthalmol. 2002;86:801–808. [CrossRef] [PubMed]
Thuret G Manissolle C Acquart S . Urgent need for normalization of corneal graft quality controls in French eye banks. Transplantation. 2004;78:1299–1302. [CrossRef] [PubMed]
Deb-Joardar N Thuret G Racine GA . Standard microlithographic mosaics to assess endothelial cell counting methods by light microscopy in eye banks using organ culture. Invest Ophthalmol Vis Sci. 2006;47:4373–4377. [CrossRef] [PubMed]
Deb-Joardar N Thuret G Gavet Y . Reproducibility of endothelial assessment during corneal organ culture: comparison of a computer-assisted analyzer with manual methods. Invest Ophthalmol Vis Sci. 2007;48:2062–2067. [CrossRef] [PubMed]
Thuret G Deb-Joardar N Manissolle C . Assessment of the human corneal endothelium: in vivo Topcon SP2000P specular microscope versus ex vivo samba cornea eye bank analyser. Br J Ophthalmol. 2007;91:265–266. [CrossRef] [PubMed]
Deb-Joardar N Thuret G Zhao M . Comparison of two semiautomated methods for evaluating endothelial cells of eye bank corneas. Invest Ophthalmol Vis Sci. 2007;48:3077–3082. [CrossRef] [PubMed]
Vonwiller P . Pigmentepithel und hornhautendothel. Verhandl Schweitz Naturforsch Gesellsch. 1946;194.
Aquavella JV Van Horn DL Haggerty CJ . Corneal preservation using M-K medium. Am J Ophthalmol. 1975;80:791–799. [CrossRef] [PubMed]
Sperling S . Combined staining of corneal endothelium by alizarine red and trypan blue. Acta Ophthalmol (Copenh). 1977;55:573–580. [CrossRef] [PubMed]
Basu PK Hasany SM Doane FW Schultes K . Can steroid reduce endothelial damage in stored corneas? Effect on cell viability and ultrastructure. Can J Ophthalmol. 1978;13:31–38. [PubMed]
Binder PS Wickham MG . M-K medium and postmortem cytologic damage. Invest Ophthalmol Vis Sci. 1978;17:159–170. [PubMed]
Schrapel A Letko G Giessmann HG . An objective rapid testing method for viability of corneal endothelium after conservations. Klin Monatsbl Augenheilkd. 1982;180:53–57. [CrossRef] [PubMed]
Singh G Bohnke M von Domarus D Draeger J . Toxicity of methods of implant material sterilization on corneal endothelium. Ann Ophthalmol. 1985;17:727–730. [PubMed]
Singh G Bohnke M von-Domarus D Draeger J Lindstrom RL Doughman DJ . Vital staining of corneal endothelium. Cornea. 1985;4:80–91. [CrossRef] [PubMed]
Madden PW . The evaluation of endothelial damage following corneal storage: a comparison of staining methods and the value of scanning electron microscopy. Curr Eye Res. 1987;6:1441–1451. [CrossRef] [PubMed]
Hartmann C Rieck P . A new test for endothelial viability: the Janus green photometry technique. Arch Ophthalmol. 1989;107:1511–1515. [CrossRef] [PubMed]
Means TL Geroski DH Hadley A Lynn MJ Edelhauser HF . Viability of human corneal endothelium following Optisol-GS storage. Arch Ophthalmol. 1995;113:805–809. [CrossRef] [PubMed]
Salla S Redbrake C Becker J Reim M . Remarks on the vitality of the human cornea after organ culture. Cornea. 1995;14:502–508. [CrossRef] [PubMed]
Wusteman MC Boylan S Pegg DE . Cryopreservation of rabbit corneas in dimethyl sulfoxide. Invest Ophthalmol Vis Sci. 1997;38:1934–1943. [PubMed]
Kent DG Solomon KD Peng Q Whiteside SB Brown SJ Apple DJ . Effect of surface photorefractive keratectomy and laser in situ keratomileusis on the corneal endothelium. J Cataract Refract Surg. 1997;23:386–397. [CrossRef] [PubMed]
Wusteman MC Armitage JW Wang LH Busza AL Pegg DE . Cryopreservation studies with porcine corneas. Curr Eye Res. 1999;19:228–233. [CrossRef] [PubMed]
Albon J Tullo AB Aktar S Boulton ME . Apoptosis in the endothelium of human corneas for transplantation. Invest Ophthalmol Vis Sci. 2000;41:2887–2893. [PubMed]
Koh SW Waschek JA . Corneal endothelial cell survival in organ cultures under acute oxidative stress: effect of VIP. Invest Ophthalmol Vis Sci. 2000;41:4085–4092. [PubMed]
Koh SW . Ciliary neurotrophic factor released by corneal endothelium surviving oxidative stress ex vivo. Invest Ophthalmol Vis Sci. 2002;43:2887–2896. [PubMed]
Engelmann K Bednarz J Valtink M . Prospects for endothelial transplantation. Exp Eye Res. 2004;78:573–578. [CrossRef] [PubMed]
Joyce NC Harris DL Mc Alister JC Ali RR Larkin DF . Effect of overexpressing the transcription factor E2F2 on cell cycle progression in rabbit corneal endothelial cells. Invest Ophthalmol Vis Sci. 2004;45:1340–1348. [CrossRef] [PubMed]
Sikder S Snyder RW . Femtosecond laser preparation of donor tissue from the endothelial side. Cornea. 2006;25:416–422. [CrossRef] [PubMed]
Steinhardt RA Alderton JM . Poloxamer 188 enhances endothelial cell survival in bovine corneas in cold storage. Cornea. 2006;25:839–844. [CrossRef] [PubMed]
Suwan-Apichon O Reyes JM Griffin NB Barker J Gore P Chuck RS . Microkeratome versus femtosecond laser predissection of corneal grafts for anterior and posterior lamellar keratoplasty. Cornea. 2006;25:966–968. [CrossRef] [PubMed]
Suwan-Apichon O Reyes JM Griffin NB Barker J Gore P Chuck RS . Microkeratome preparation of lamellar corneal grafts. Eye Contact Lens. 2006;32:248–249. [CrossRef] [PubMed]
Suwan-apichon O Rizen M Reyes JM . A new donor cornea harvesting technique for posterior lamellar keratoplasty. Br J Ophthalmol. 2005;89:1100–1101. [CrossRef] [PubMed]
Ide T Yoo SH Goldman JM Perez V O'Brien TP . Descemet-stripping automated endothelial keratoplasty: effect of inserting forceps on DSAEK donor tissue viability by using an in vitro delivery model and vital dye assay. Cornea. 2007;26:1079–1081. [CrossRef] [PubMed]
Ide T Yoo SH Kymionis GD Perez VL Goldman JM O'Brien TP . Descemet-stripping automated endothelial keratoplasty (DSAEK): effect of nontoxic gentian violet marking pen on DSAEK donor tissue viability by using vital dye assay. Cornea. 2008;27:562–564. [CrossRef] [PubMed]
Mehta JS Por YM Poh R Beuerman RW Tan D . Comparison of donor insertion techniques for Descemet stripping automated endothelial keratoplasty. Arch Ophthalmol. 2008;126:1383–1388. [CrossRef] [PubMed]
Slettedal JK Lyberg T Roger M Beraki K Ramstad H Nicolaissen B . Regeneration with proliferation of the endothelium of cultured human donor corneas with extended postmortem time. Cornea. 2008;27:212–219. [CrossRef] [PubMed]
Amato D Oddone F Nubile M Colabelli Gisoldi RA Villani CM Pocobelli A . Precut donor tissue for Descemet stripping automated keratoplasty: anterior hinged lamella On versus Off. Br J Ophthalmol. 2010;94:519–522. [CrossRef] [PubMed]
Kim JH Choi SK Lee D . The comparison of femtosecond laser-assisted penetrating keratoplasty with conventional surgery in terms of endothelial safety: ex vivo study using porcine eyes. Cornea. 2009;28:812–816. [CrossRef] [PubMed]
Proulx S Audet C Uwamaliya J . Tissue engineering of feline corneal endothelium using a devitalized human cornea as carrier. Tissue Eng Part A. 2009;15:1709–1718. [CrossRef] [PubMed]
Yoeruek E Saygili O Spitzer MS Tatar O Bartz-Schmidt KU Szurman P . Human anterior lens capsule as carrier matrix for cultivated human corneal endothelial cells. Cornea. 2009;28:416–420. [CrossRef] [PubMed]
Wolf AH Welge-Lussen UC Priglinger S . Optimizing the deswelling process of organ-cultured corneas. Cornea. 2009;28:524–529. [CrossRef] [PubMed]
He Z Pipparelli A Manissolle C . Ex vivo gene electrotransfer to the endothelium of organ cultured human corneas. Ophthalmic Res. 2010;43:43–55. [CrossRef] [PubMed]
2007 Eye Banking Statistical Report. Washington, DC: Eye Bank Association of America; 2008.
2009 Eye Banking Statistical Report. Washington, DC: Eye Bank Association of America; 2010.
Terry MA . Endothelial keratoplasty: history, current state, and future directions. Cornea. 2006;25:873–878. [CrossRef] [PubMed]
Price MO Price FWJr . Endothelial keratoplasty: a review. Clin Exp Ophthalmol. 2010;38:128–140. [CrossRef]
Dapena I Ham L Melles GR . Endothelial keratoplasty: DSEK/DSAEK or DMEK—the thinner the better? Curr Opin Ophthalmol. 2009;20:299–307. [CrossRef] [PubMed]
Terry MA Shamie N Chen ES Phillips PM Hoar KL Friend DJ . Precut tissue for Descemet's stripping automated endothelial keratoplasty: vision, astigmatism, and endothelial survival. Ophthalmology. 2009;116:248–256. [CrossRef] [PubMed]
Muraine M . Endothelial keratoplasty. J Fr Ophtalmol. 2008;31:907–920. [CrossRef] [PubMed]
Wang XM Terasaki PI Rankin GWJr Chia D Zhong HP Hardy S . A new microcellular cytotoxicity test based on calcein AM release. Hum Immunol. 1993;37:264–270. [CrossRef] [PubMed]
Poole CA Brookes NH Clover GM . Keratocyte networks visualised in the living cornea using vital dyes. J Cell Sci. 1993;106(pt 2):685–691. [PubMed]
De Clerck LS Bridts CH Mertens AM Moens MM Stevens WJ . Use of fluorescent dyes in the determination of adherence of human leucocytes to endothelial cells and the effect of fluorochromes on cellular function. J Immunol Methods. 1994;172:115–124. [CrossRef] [PubMed]
Lie JT Birbal R Ham L van der Wees J Melles GR . Donor tissue preparation for Descemet membrane endothelial keratoplasty. J Cataract Refract Surg. 2008;34:1578–1583. [CrossRef] [PubMed]
Nartey IN Ng W Sherrard ES Steele AD . Posterior corneal folds and endothelial cell damage in human donor eyes. Br J Ophthalmol. 1989;73:121–125. [CrossRef] [PubMed]
Legeais JM Parel JM Simon G Ren Q Denham D . Endothelial damage by the corneal Hessburg-Barron vacuum trephine. Refract Corneal Surg. 1993;9:255–258. [PubMed]
Thuret G Chiquet C Bernal F . Prospective, randomized clinical and endothelial evaluation of 2 storage times for cornea donor tissue in organ culture at 31 degrees C. Arch Ophthalmol. 2003;121:442–450. [CrossRef] [PubMed]
Frueh BE Bohnke M . Prospective, randomized clinical evaluation of Optisol vs organ culture corneal storage media. Arch Ophthalmol. 2000;118:757–760. [CrossRef] [PubMed]
Lass JH Gal RL Dontchev M . Donor age and corneal endothelial cell loss 5 years after successful corneal transplantation: specular microscopy ancillary study results. Ophthalmology. 2008;115:627–632. [CrossRef] [PubMed]
Price MO Gorovoy M Benetz BA . Descemet's stripping automated endothelial keratoplasty outcomes compared with penetrating keratoplasty from the Cornea Donor Study. Ophthalmology. 2010;117:438–444. [CrossRef] [PubMed]
Rijneveld WJ Beekhuis WH van Rij G Rinkel-van Driel B Pels E . Clinical comparison of grafts stored in McCarey-Kaufman medium at 4 degrees C and in corneal organ culture at 31 degrees C. Arch Ophthalmol. 1992;110:203–205. [CrossRef] [PubMed]
Terry MA Wall JM Hoar KL Ousley PJ . A prospective study of endothelial cell loss during the 2 years after deep lamellar endothelial keratoplasty. Ophthalmology. 2007;114:631–639. [CrossRef] [PubMed]
Ham L van Luijk C Dapena I . Endothelial cell density after Descemet membrane endothelial keratoplasty: 1- to 2-year follow-up. Am J Ophthalmol. 2009;148:521–527. [CrossRef] [PubMed]
Laaser K Bachmann BO Horn FK Schlotzer-Schrehardt U Cursiefen C Kruse FE . Donor tissue culture conditions and outcome after Descemet membrane endothelial keratoplasty. Am J Ophthalmol. 2011;151:1007–1018. [CrossRef] [PubMed]
Lass JH Sugar A Benetz BA . Endothelial cell density to predict endothelial graft failure after penetrating keratoplasty. Arch Ophthalmol. 2010;128:63–69. [CrossRef] [PubMed]
Terry MA Shamie N Chen ES Hoar KL Phillips PM Friend DJ . Endothelial keratoplasty: the influence of preoperative donor endothelial cell densities on dislocation, primary graft failure, and 1-year cell counts. Cornea. 2008;27:1131–1137. [CrossRef] [PubMed]
Figure 1.
 
Triple endothelial labeling with Hoechst 33342 (H), ethidium homodimere (E), and calcein-AM (C). ×10 objective. Green arrowhead: living cell C positive. Red arrowhead: dead cell E positive. Blue arrowhead: dying cell without metabolic activity (C negative) but still adherent, either isolated or at fold edges. White arrowheads: fold deprived of cell. Scale bar, 200 μm.
Figure 1.
 
Triple endothelial labeling with Hoechst 33342 (H), ethidium homodimere (E), and calcein-AM (C). ×10 objective. Green arrowhead: living cell C positive. Red arrowhead: dead cell E positive. Blue arrowhead: dying cell without metabolic activity (C negative) but still adherent, either isolated or at fold edges. White arrowheads: fold deprived of cell. Scale bar, 200 μm.
Figure 2.
 
Pan-corneal image acquisition for viable cell surface determination. (1) Acquisition of the entire graft surface with 40 overlapping images after calcein staining (×4 objective). (2) Reconstruction. (3) Extraction of green channel. (4) Delineation of graft contour by elimination of the radial cutting area necessary for the flat mount. (5) Standardized thresholding. (6) Binarization. (7) Automatic calculation of surface area in mm2. (8, 9) Representative paired grafts. (8) Full-thickness graft (control). Folds with no living cells were clearly visible. The radial cut necessary for the flat mount caused artifactual cell destruction, visible as a darker area around the cut (white arrowheads). This was more pronounced than with the lamellar graft because of stromal compression during cutting and was, therefore, excluded during image analysis. The blurred surrounding rim (blue arrow) is attributed to stromal autofluorescence (absent in 9). Yellow arrow: cell damaged by trephination. (9) Endothelial graft. Folds with no living cells were also present. Red arrow: hinge that kept the graft attached to the stroma for shipping and was then cut after staining. Green: percentages of areas covered by living cells. Median viable endothelial surface was higher in the control group than in the predissected group, respectively 90% (86%–92%) versus 82% (80%–87%) (P = 0.043).
Figure 2.
 
Pan-corneal image acquisition for viable cell surface determination. (1) Acquisition of the entire graft surface with 40 overlapping images after calcein staining (×4 objective). (2) Reconstruction. (3) Extraction of green channel. (4) Delineation of graft contour by elimination of the radial cutting area necessary for the flat mount. (5) Standardized thresholding. (6) Binarization. (7) Automatic calculation of surface area in mm2. (8, 9) Representative paired grafts. (8) Full-thickness graft (control). Folds with no living cells were clearly visible. The radial cut necessary for the flat mount caused artifactual cell destruction, visible as a darker area around the cut (white arrowheads). This was more pronounced than with the lamellar graft because of stromal compression during cutting and was, therefore, excluded during image analysis. The blurred surrounding rim (blue arrow) is attributed to stromal autofluorescence (absent in 9). Yellow arrow: cell damaged by trephination. (9) Endothelial graft. Folds with no living cells were also present. Red arrow: hinge that kept the graft attached to the stroma for shipping and was then cut after staining. Green: percentages of areas covered by living cells. Median viable endothelial surface was higher in the control group than in the predissected group, respectively 90% (86%–92%) versus 82% (80%–87%) (P = 0.043).
Figure 3.
 
Procedure for calculation of endothelial cell density. (A) Example of 1 of 5 images taken with a ×10 objective, one central and one per quarter. (B) Selection of Hoechst acquisition channel. (C) Delineation of largest possible region of interest by eliminating areas without nuclei to determine homogeneous density of nuclei. These areas without nuclei corresponded to areas of naked Descemet, mostly folds, without ECs. They were by definition eliminated during the processing of calcein images (see Fig 2); thus, it was logical not to select them to determine the number of nuclei per surface unit. (D) Counting of nuclei (here, 1099). (E) Selection of ethidium acquisition channel. (F, G) Similar to the previous Hoechst process. Median mortality rate (ethidium+) did not differ between the control group and the predissected group, respectively 0.05% (0%–0.28%) versus 0.12% (0.06%–0.30%) (P = 0.068).
Figure 3.
 
Procedure for calculation of endothelial cell density. (A) Example of 1 of 5 images taken with a ×10 objective, one central and one per quarter. (B) Selection of Hoechst acquisition channel. (C) Delineation of largest possible region of interest by eliminating areas without nuclei to determine homogeneous density of nuclei. These areas without nuclei corresponded to areas of naked Descemet, mostly folds, without ECs. They were by definition eliminated during the processing of calcein images (see Fig 2); thus, it was logical not to select them to determine the number of nuclei per surface unit. (D) Counting of nuclei (here, 1099). (E) Selection of ethidium acquisition channel. (F, G) Similar to the previous Hoechst process. Median mortality rate (ethidium+) did not differ between the control group and the predissected group, respectively 0.05% (0%–0.28%) versus 0.12% (0.06%–0.30%) (P = 0.068).
Figure 4.
 
Box plots of the ECD measurement before and after predissection and shipping to a remote center and comparison with viable ECD determined immediately thereafter using triple labeling by Hoechst/ethidium/calcein staining (HEC) combined with pan-corneal analysis. Horizontal line: the threshold of 2000 cells/mm2 conventionally used to deliver grafts for penetrating keratoplasty. Circle: case at >1.5 box length from the upper or lower edge of the box.
Figure 4.
 
Box plots of the ECD measurement before and after predissection and shipping to a remote center and comparison with viable ECD determined immediately thereafter using triple labeling by Hoechst/ethidium/calcein staining (HEC) combined with pan-corneal analysis. Horizontal line: the threshold of 2000 cells/mm2 conventionally used to deliver grafts for penetrating keratoplasty. Circle: case at >1.5 box length from the upper or lower edge of the box.
Table 1.
 
Experimental Methods for In Situ Assessment of Corneal Endothelial Quality
Table 1.
 
Experimental Methods for In Situ Assessment of Corneal Endothelial Quality
Reference Year Method Species Application
Aquavella 17 1975 Trypan blue, SEM H Stored corneas
Sperling 18 1977 Alizarin red + trypan blue H Stored corneas
Basu 19 1978 Trypan blue, para nitroblue tetrazolium, SEM H, R, GP Stored corneas
Binder 20 1978 Nitroblue tetrazolium, SEM, TEM R Stored corneas (postmortem cell damage)
Schrapel 21 1982 Toluidine blue P Stored corneas
Singhl 22,23 1985 Trypan blue H, P Stored corneas (toxicity tests)
Madden 24 1987 Trypan blue, nitroblue tetrazolium, acridine orange, fluorescein diacetate, ethidium bromide, SEM H Stored corneas
Hartmann 25 1989 Janus green P Stored corneas (toxicity tests)
Means 26 1995 Alizarin red + trypan blue, calcein-AM + ethidium homodimer H Stored corneas
Salla 27 1995 Succinate dehydrogenase staining H Stored corneas
Wusteman 28 1997 Acridine orange + propidium iodide, confocal microscopy R Stored corneas
Kent 29 1997 Calcein-AM + ethidium homodimer H High-diopter myopic PRK or LASIK
Wusteman 30 1999 Acridine orange + propidium iodide, nuclear magnetic resonance spectroscopy P Cryoconservation
Albon 31 2000 Hoechst, TUNEL assay, immunostaining (active caspase 3) H Stored corneas
Koh 32,33 2000 Calcein-AM + ethidium homodimer H, B Stored corneas
Gain 7 2002 Trypan blue + TUNEL assay H Stored corneas
Engelman 34 2004 Immunostaining (integrins and tight junction), perfusion studies H Bioengineered endothelium
Joyce 35 2004 Calcein-AM + ethidium homodimer, TUNEL R Viability after endothelial transfection
Sikder 36 2006 Calcein-AM + ethidium homodimer H Precut lamellar graft (femtosecond laser)
Steinhardt 37 2006 Calcein-AM B Stored corneas
Suwan-Apichon 38 40 2006 Alizarin red + trypan blue, ultrasonic pachymetry H Precut lamellar graft (manual, femtosecond laser, DSAEK)
Ide 41,42 2007/8 Alizarin red + trypan blue H Whole lamellar graft process (DSAEK)
Mehta 43 2008 Trypan blue, SEM H Lamellar graft introduction techniques (DSAEK)
Slettedal 44 2008 Immunostaining (n-cadherin), SEM H Stored corneas
Amato 45 2009 Trypan blue H Precut lamellar graft (DSAEK)
Kim 46 2009 Alizarin red, pachymetry, SEM P Femtosecond laser trephination for PKP
Proulx 47 2009 Alizarin red, immunostaining (Na/CO3, ZO-1, Na/K-ATPase), SEM, TEM F Bioengineered endothelium
Yoeruek 48 2009 Calcein-AM + ethidium homodimer, immunostaining (ZO-1, connexin-43, Na/K-ATPase, and cytokeratin-3), phase-contrast microscopy H Bioengineered endothelium
Wolf 49 2009 Calcein-AM + ethidium homodimer + Hoechst, * pachymetry, optical coherent tomography H Stored corneas (deswelling)
He 50 2010 Calcein-AM + ethidium homodimer + Hoechst,* immunostaining (ZO-1) H Endothelial transfection (electroporation of stored corneas)
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×