To evaluate the early effect of standard and transepithelial collagen cross-linking on human corneal nerves in donor eyes by ex vivo confocal microscopy and acetylcholinesterase staining.
Experimental laboratory investigation.
Eight human eye bank corneal buttons (mean age, 73.6 years) were included. Ultraviolet A collagen cross-linking was performed postmortem on 3 corneas with the standard protocol involving epithelial debridement and 4 corneas by the transepithelial approach. One cornea served as a control. Corneal nerves were evaluated using confocal microscopy and acetylcholinesterase histology.
Confocal microscopy demonstrated the absence of subbasal nerves in corneas treated by the standard technique. These nerves were preserved in corneas treated by the transepithelial approach. Stromal nerves were visible in both groups. Histology of corneas treated by the standard technique revealed localized swellings of the stromal nerves with disruption of axonal membrane and loss of axonal continuity within the treatment zone. These changes were absent in corneas treated by the transepithelial approach.
This study highlights the immediate effects of collagen cross-linking on the corneal nerves in an ex vivo model. The absence of subbasal nerves in the early phase of treatment appears to be attributable mainly to mechanical removal of epithelium, rather than ultraviolet light–induced damage. Localized swelling of the stromal nerves was the main difference between the 2 treatment protocols. Further research on laboratory animals would be necessary to verify these changes over a specified time course without the super-addition of postmortem changes.
Riboflavin (vitamin B2)/ultraviolet A (UVA) (370 nm) collagen cross-linking (CXL) is a new therapeutic modality in the management of keratoconus and the first treatment available to mechanically strengthen the cornea and thus slow the progression of keratoconus.
Contemporary in vivo, ex vivo, and in vitro investigations have revealed significant alterations in the normal architecture and histology of corneal tissues, mainly at the anterior 300 micrometers of the cornea.
Despite several clinical and laboratory investigations of the effect of CXL on the corneal microstructure, very limited information is available on the status of corneal sensitivity and effect of CXL on corneal innervation in the immediate and long term posttreatment. All reported studies have used in vivo confocal microscopy to examine corneal nerve changes after CXL.
The aim of this study was to evaluate the early effect of CXL on human corneal nerves in donor eyes by relating ex vivo confocal microscopy with histology. The effects of standard (following epithelial debridement) and transepithelial approaches were examined.
Material and Methods
Collagen Cross-linking of Corneas
Eight human eye bank corneal buttons with scleral rims were included in the study. The mean donor age was 73.6 years (range 56 to 83 years). The average death to enucleation time was 7 hours (range 4 hours and 45 minutes to 10 hours and 20 minutes). The mean storage time (between eye bank procedures and cross-linking procedure) was 28 hours (range 20 to 48 hours).
For the cross-linking procedure the corneoscleral buttons were mounted on an anterior chamber maintainer system (Barron artificial anterior chamber; Katena Products Inc, Denville, New Jersey, USA), filled with Eusol-C storage media (Alchimia s.r.l., Ponte S. Nicolò, Italy), to obtain adequate pressure and stability of the corneal tissue.
Three corneas were treated according to a standard protocol, which consisted of removal of the central 10 mm of corneal epithelium and instillation of riboflavin 0.1% (Ricrolin; SOOFT Italia S.p.A., Montegiorgio, Italy) every 2 minutes and 30 seconds for 12 minutes and 30 seconds. This was followed by 5-minute phases of UVA irradiation (3.0 mW/cm 2 , at 370 nm) with instillation of riboflavin 0.1% every 2 minutes and 30 seconds (6 phases, 30 minutes in total). Four corneas were cross-linked via a transepithelial approach. One drop of Ricrolin TE (riboflavin 0.1% plus enhancers [including destrane T500, edetate sodium, tromethamine, bihydrate sodium phosphate monobasic, and bihydrate sodium phosphate bibasic]; SOOFT Italia S.p.A.) was instilled every 10 minutes for 2 hours. Corneas were then presoaked with riboflavin for 15 minutes before UVA light was applied. This was followed by 6 phases (5 minutes each) of instillation of 1 drop of Ricrolin TE and then exposure to UVA light (3.0 mW/cm 2 , at 370 nm). The enhancers in Ricrolin TE facilitate riboflavin penetration into the corneal stroma in the presence of an intact epithelium. In addition, this was augmented by a longer duration of pretreatment with Ricrolin TE as per reported protocols.
One cornea was used as a control, in which the epithelium was left intact and not treated with UVA light. Riboflavin was instilled in the same dose and frequency as described above in the standard approach but the epithelium was not removed. Before every UVA light treatment, the irradiance of the CBM (Caporossi-Baiocchi-Mazzotta)-VEGA X-Linker (SOOFT Italia S.p.A.) was checked with a LaserMate-Q UVA meter (Laser 2000 GmbH, Wessling, Germany). After treatment, all corneas were stored in Eusol-C storage media at room temperature for a varying duration before formalin fixation ( Table 1 ). The aim was to check whether there was any change in the effect of CXL at 3 time points (1/2, 1, and 2 hours) after the procedure.
|Age of Donor||Death to Enucleation Time (Hours)||Treatment Protocol||Time Between Treatment and Formalin Fixation|
The corneoscleral buttons were scanned by laser scanning confocal microscope (Heidelberg Retina Tomograph II Rostock Corneal Module [RCM]; Heidelberg Engineering GmbH, Heidelberg, Germany).
To facilitate the examination, the corneoscleral buttons were mounted on an anterior chamber maintainer system (Barron artificial anterior chamber; Katena Products Inc). The corneas were then pressurized with balanced salt solution to create an artificial anterior chamber and maintain rigidity of the corneal tissue. The system was held in front of the microscope using a clamp holder designed for such purpose, as described before. This holder had 3 joints (including 1 ball-and-socket joint), which allowed accurate positioning and orientation of the eye for examination. Pressure-free contact with the cornea was monitored during the examination using a colored digital camera system that gives a side view of the microscope objective and eye. The average examination time for each eye was about 20 minutes.
Acetylcholinesterase Staining Technique
A whole-mount staining of corneas for nerve demonstration was performed using the Karnovsky and Roots method of acetylcholinesterase enzyme as described previously. The flat mounts were photographed using a Leica microscope system (Leica DM4000B; Leica Microsystems, Nussloch, Germany). Images were processed by Adobe Photoshop CS4 (Adobe Systems Inc, San Jose, California, USA).
Epithelial cells were visible and showed abnormal hyperreflectivity. The cell borders were unclear and epithelial cell layer appeared thin (around 15 μm) ( Figure 1 , Top left). Few subbasal nerves were detected ( Figure 1 , Top right). Stromal nerves could be seen ( Figure 1 , Bottom left). The stroma appeared edematous, demonstrating a honeycombed hyperreflective pattern of large keratocytes with long cytoplasmic extensions mainly at the anterior and mid stroma ( Figure 1 , Bottom left and right).
Corneas treated by standard collagen cross-linking
Epithelial cells were absent and the subbasal nerve plexus could not be visualized. The most superficial frames showed an abnormally hyperreflective Bowman zone ( Figure 2 , Left). Stromal nerves and keratocytes could be detected ( Figure 2 , Middle). The stromal architecture was altered with radially oriented dark striae that were scattered within a highly reflective extracellular matrix ( Figure 2 , Right). Small hyperreflective dots resembling granular deposits were observed mainly in the anterior stroma up to a 200-μm depth ( Figure 2 , Right).
Corneas treated by transepithelial collagen cross-linking
Corneal epithelial cells were visible and nerves of the subbasal plexus could be easily detected running beneath and within the basal epithelial cells ( Figure 3 , Top left). Some of these nerves were found to terminate in the characteristic bulb-like hyperreflective structures, as described before ( Figure 3 , Top right). Focal areas of abnormal hyperreflectivity were observed at the level of the Bowman zone ( Figure 3 , Middle left). The stromal changes appeared similar to the changes that were observed in corneas treated by the standard CXL protocol. The stromal nerves were normal in appearance ( Figure 3 , Middle right). The honeycombed pattern of stromal edema was also seen and appeared similar to that of the control corneas ( Figure 3 , Bottom left). A reduced number of keratocyte nuclei were visible in the anterior stroma. The granular deposits observed in these corneas were more abundant and extended deeper within the corneal stroma (up to 300 μm in depth) compared to those that were noted in the previous group ( Figure 3 , Bottom right).
Epithelial cells and their nuclei were visible. No subbasal nerves could be detected. Stromal nerves appeared normal and showed a characteristic dichotomous branching pattern ( Figure 4 , Top left).