An early contributor since central corneal neovascularization does not become apparent. Although technical limitations prevented us from evaluating structural evidence of endothelial decompensation, the fact that radiolabeled SM adducts develop on the lens after 5 min of vapor exposure raises the possibility that corneal endothelium may be directly exposed at sufficiently high doses. When comprehensive, full-thickness corneal repair was observed, stromal histopathology scores at eight weeks were dramatically improved versus MGK corneas, and slightly improved compared to two weeks. Alternatively, in MGK corneal repair appeared frustrated, with edematous stromal regions refractory to fibrocyte infiltration and deteriorated histopathology scores. Several mechanisms may underlay the inability of fibrocytes to penetrate the entire wound region, including altered chemokine signaling due to significant increases in stromal volume; disruption of the stromal matrix due to edema or enzymatic activity; pathological assembly of collagen matrices; or chemical derivatization of collagen by SM. A fully stratified epithelium that excludes fluorescein is present in all corneas by two weeks after exposure. MGK corneas subsequently begin to retain fluorescein again starting at three weeks. Although it was initially believed that these lesions represented the loss of extensive regions of epithelium, gross epithelial lesions were not observed by histology. TEM confirmed that fluorescein retention occurs through focal disruptions in the integrity of the corneal epithelium with the demonstration of necrotic basal epithelial cells, destabilized CE and localized epithelial edema, but no instances of gross epithelial detachment except in cases of epithelial bullae. Ultrastructural characterization revealed scattered evidence of basal cell GDC-0199 necrosis at two weeks after exposure and beyond in MGK corneas, but not in resolved corneas. The origin of this persistent BCN is unclear, but appears specific to SM exposure since other means of producing large scale epithelial loss, such as corneal scraping or excimer laser debridement, have not been associated with BCN. The temporal delay between the exposure and the appearance of BCN between one and two weeks suggests that necrosis is either due to delayed SM toxicity or a second-order effect indirectly induced by SM exposure, and elucidating the mechanism underlying this delay is crucial to therapeutic development. The development of widespread necrosis in the stratum basale is likely to have serious implications to corneal integrity. For example, necrosing epithelial cells release keratoactive proteins, such as pro-inflammatory mediators, matrix metalloproteinases and other signaling effectors that can activate receptor-mediated pathways in proximal cells. Basal epithelial cell necrosis structurally disrupts overlying and adjacent cells, and unless those cells are replaced will result in epithelial attenuation and a loss in differentiative capacity to regenerate the CE. Fluorescein uptake in MGK corneas appears to be restricted to the area of the acute lesion, suggesting the necrotic mechanism may be bounded by some aspect of the exposure or recovery. It has been hypothesized that injury to the limbal stem cell niche results in chronic CE cell loss, similar to LSC deficiency. Other possibilities include delayed toxicity of basal CE cells that were originally located at the periphery of the exposure; destabilizing chemical modification of the basement membrane; or cell death in response to aberrant interactions with other corneal or adnexal tissues.