In the quest for the development of a functional keratoprosthesis, the biocompatibility of the porous skirt material in the Chirila keratoprosthesis (KPro) was investigated. The population of live and dead cells within, and the inflammatory response to, a tissue-integrating poly(2-hydroxyethyl methacrylate) (PHEMA) sponge were studied. Samples of the hydrogel sponge were implanted in rabbit corneas and explanted at predetermined time points up to 12 weeks. The explanted sponges were subjected to cell viability assay using two types of fluoroprobes, 5-chloromethylfluorescein diacetate and ethidium homodimer-l. A semiquantitative analysis was performed to assess the number of dead cells within the sponge and in the area of corneal stroma proximal to the sponge. Five rabbits were used for each end point (2, 4 and 12 weeks). To investigate the inflammatory response to the sponge, immunocytochemistry, using specific antibodies to rabbit macrophages, enzyme histochemistry of chloroacetate esterase (to detect neutrophils) and transmission electron microscopy (TEM) were also employed at 24 h, 2, 4 and 12 weeks after implantation. Four weeks after implantation, fewer viable cells were observed in the sponge when compared to the 2-week implant. However, the proportion of viable cells increased dramatically by 12 weeks. The proportion of nonviable cells decreased gradually with time; central sponge contained 34 +/- 11% dead cells after 2 weeks, and 15 +/- 4.3 % after 12 weeks. The staining of inflammatory cells demonstrated the presence of macrophages and neutrophils up to 12 weeks after implantation. TEM confirmed the presence of these cell types and others, including eosinophils and myofibroblasts, as well as blood capillaries. The presence of a significant number of viable cells at each time point and the uniform reduction of the nonviable cell proportion with time suggests that the sponge is a conducive environment supporting a prolific, viable cellular colonization. Dead cells observed in the first instance indicate a normal injury pattern. However, the presence of a small but significant proportion of invading inflammatory cells 12 weeks after implantation confirms a characteristic pattern of wound healing within the sponges. (C) 1998 Published by Elsevier Science Ltd. All rights reserved.