BACKGROUND: Corneal crosslinking (CXL) has revolutionized the treatment of keratoconus during the past decade. In the present study, the morphological changes in the corneal collagen fibrils (CFs) following crosslinking treatment are described.
METHODS: Ten pairs of porcine and rabbit corneas were retrieved. In each pair, one cornea was the control and the other underwent CXL treatment. The central corneal thickness (CCT) was measured before and after CXL treatment. Each treated and control cornea was examined with light microscopy and by transmission electron microscopy.
RESULTS: (a) The mean CCT was significantly reduced following treatment. (b) CFs were more closely packed in the anterior region and loosely packed in the posterior region. (c) CF diameter increased significantly in the anterior and intermediate regions but declined gradually towards the deeper regions. (d) There was a statistically significant decrease in the interfibrillar distance over the different regions of the cornea, except for the posterior region in porcine corneas, where there was no change. (e) The distance between adjacent collagen lamellae was significantly decreased in all regions of treated rabbit corneas. There was no change in porcine corneas.
CONCLUSIONS: CXL treatment resulted in increased the CF diameter and decreased interfibrillar distance in the anterior and intermediate regions, while its effects on the posterior region differed among species. The effect on interlamellar distance was more prominent in the rabbit model than the porcine model. CXL treatment stiffened the corneas by increasing CF diameter and decreasing interfibrillar distance in both rabbit and pig corneas.

Nidogen-2 is a basement membrane (BM) glycoprotein that could be a key to understanding why defects in BM regeneration occur after severe trauma to the cornea. We monitored the location and expression of nidogen-2 during corneal repair after alkali burn in rabbits. In rabbits that received both general and ocular topical anaesthesia, the central cornea of the left eye was burned by placing an 8-mm diameter filter paper soaked in 0.5 N NaOH for 60 s. Right corneas were used as controls. The eyes were evaluated at 2, 7, 15, and 30 days after burning and analysed by immunohistochemistry for nidogen-2 and α-smooth muscle actin, a myofibroblast marker. Nidogen-2 mRNA expression levels were determined by quantitative real-time polymerase chain reaction. In control corneas, nidogen-2-positive cells were in all epithelial layers, the endothelium, and the anterior and posterior stromal regions. At Day 2 after the alkali burn, the wound area epithelium and the peripheral epithelium were made up of only 1 to 2 cell layers, all of them nidogen-2 positive. At Day 7 in the wound area, the epithelium consisted of two cell layers, and the basally located cells were mostly nidogen-2 positive. The greatest change was observed at Day 30. At this time, the ulcer prevalence in the alkali-burned corneas was approximately 50% and the central epithelial defects remained. In unepithelialized corneas, frequent epithelial detachments were present, in which almost of the epithelial cells were nidogen-2 negative. The injured stroma was repopulated by activated stromal cells that synthesized nidogen-2. The nidogen-2 was retained in the newly secreted, but disordered, matrix produced mainly by the myofibroblasts localized in the stroma at 7, 15, and 30 days after burning. Thus, even though nidogen-2 was present, it was unable to contribute to the effective regeneration of the BM.

In vitro drug release testing and ex vivo transcorneal drug permeation can provide valuable information on the performance of the Q1/Q2 equivalent ointments prior to any animal studies. Good correlation between in vitro and ex vivo drug release may be indicative of good in vitro and in vivo correlation. Accordingly, it is important to investigate in vitro as well as ex vivo drug release from Q1/Q2 equivalent ophthalmic ointments and evaluate whether a correlation between these release profiles can be established. Four Q1/Q2 equivalent loteprednol etabonate ointments were prepared using different processing methods and excipient sources. The rheological parameters (crossover modulus and K value) of the four formulations were determined. The in vitro drug release testing of the four ointment formulations were performed using three different apparati (Franz diffusion cells, USP apparatus 2 with enhancer cells and USP apparatus 4 with semisolid adapters). Three models (zero order, logarithmic and the Higuchi model) were used to study the release kinetics of the ointment formulations. The transcorneal (rabbit corneas) permeation studies were performed using spherical joint Franz diffusion cells. The USP apparatus 4 method demonstrated better discriminatory ability compared to the USP apparatus 2 and the Franz diffusion cell methods. The in vitro release profiles of the four Q1/Q2 equivalent ointments with manufacturing differences showed a better fit using the Higuchi model (R2 > 0.98) for all three release testing methods, compared to the other two models. Ex vivo drug release through the rabbit corneas displayed zero order release kinetics. A logarithmic correlation between rheological parameters (crossover and K value) and transcorneal flux were established. In addition, a plot of the in vitro release rate against the ex vivo release flux of the four ointment formulations, yielded a straight line (R2 > 0.98) for all three release methods. Accordingly, the rheological parameters may be useful in predicting in vitro as well as ex vivo release properties.