Concentration-Dependent Effects of Silicone Oil Components on Corneal Endothelial Permeability

Abstract

To understand further the toxicity induced by several contaminants often found in silicone oils, we examined the effects of specific individual compounds on rabbit corneal endothelial permeability. The chemicals were selected based on toxic responses exhibited when they were used as mixtures of closely related chemical structures. These chemicals have been shown to remain in oils after incomplete polymerization. Rabbit eyes were perfused in vivo with a nontoxic oil to which various concentrations of different individual low-molecular-weight contaminants had been added. One week later the in vitro permeability to inulin and dextran was determined. Dose-response relationships were generated for certain linear or cyclic chemicals known to show widespread occurrence in nonpurified oils, including eicosamethylnonasiloxane (MD₇M), hexacosamethyldodecasiloxane (MD₁₀M), hexamethylcyclotrisiloxane (D₃), octamethylcyclotetrasiloxane (D₄), and decamethylcyclopentasiloxane (D₅). These low-molecular-weight compounds are of interest because they are postulated to be those most likely to be required to be removed before successful clinical use. These compounds are both more reactive than polymeric oil constituents and also can migrate from the oils into tissues because of their inherent volatility. A cyclic series (D₃, D₄, D₅), when used at 10 mg/ml each in a mixture, has been previously shown to induce different responses when each series was obtained from different sources. In one case corneal permeability increased, while the same materials from another source did not induce a permeability change. MD₇M is of particular interest since it had been found previously to lower permeability at a concentration of 10 mg/ml. It was important, therefore, not only to examine concentrations other than 10 mg/ml for potential toxicity but also to examine individual chemicals, not mixtures. The present data show that, where a response is evoked, there is little effect of concentration, from 1 mg/ml to 25 mg/ml, on induced permeability changes. In several cases, the effects on permeability are anomalously greater at lower concentrations. MD₇M had no influence on permeability except marginally at 1 mg/ml; at other concentrations the permeability was lower than that caused by oil alone. MD₁₀M showed a greater effect at 1 mg/ml, no effect on inulin at 2.5 mg/ml, and a reproducible effect at 10 and 25 mg/ml. Cyclic compounds from both commercial sources tended to influence inulin and dextran permeability, albeit in a different manner. D₃ from source A slightly increased permeability, especially at 1 mg/ml, while that from another source (B) caused essentially no change. D₄ from both sources increased dextran permeability in a consistent manner at all concentrations employed and enhanced inulin permeability at 1 mg/ml. D₅ from source A increased inulin/dextran permeability at 10 and 25 mg/ml, increasing with higher concentrations. D₅ from source B, however, increased dextran permeability at all concentrations, albeit with a greater increase at 1 mg/ml, but only increased inulin permeability at 1 mg/ml. In general, no concentration effect was noted although lower concentrations tended to cause greater inulin and dextran permeability increases. The exception was D₅ from source A, where a concentration-dependent permeability increase was noted. Dextran permeability, which is a measure of a smaller pathway across the endothelium than is inulin, tended to be influenced to a greater degree than inulin. In summary, individual contaminants tend to allow us to come to similar conclusions concerning their behavior on corneal endothelial permeability as do mixtures.