Regulation of corneal endothelial barrier function by adenosine, cyclic AMP, and protein kinases

To determine which processes or factors that regulate corneal hydration are responsible for the hydration-modulating effects of adenosine. Influx of fluid to the stroma and efflux to the aqueous humor are governed, respectively, by the imbibition pressure of the stromal matrix and the transendothelial ionic gradients determined by the permeability and active transport characteristics of this monolayer. The focus of this study was to assess the effects of adenosine on these endothelial parameters.

Isolated corneas freshly dissected from rabbit eyes were used throughout. Active ion transport was assessed by measurement of 86Rb+ uptake by the endothelial cells of intact corneas incubated for 30 minutes in 25 mM HCO3(-)-Ringer with agents promoting corneal deturgescence or corneal swelling. Intracellular and extracellular fluid in the scraped endothelial cell mass was estimated from simultaneous counts of 3H-mannitol and 14C-urea, allowing calculation of tissue-to-medium (T-M) ratios of 86Rb+ in cell water. Permeability of the endothelium was determined by measuring the efflux into the superfusate of 5-carboxyfluorescein (CF) applied to the stroma of deepithelialized corneas superfused at the endothelial surface with the same media described for 86Rb+ uptake. Thickness of these corneas and of others fixed for scanning electron microscopy was monitored with a specular microscope.

In the control medium, 25 mM HCO3(-)-Ringer, 86Rb+ was accumulated to yield a T-M ratio of 6.21. Neither adenosine nor other agents that increase cyclic adenosine monophosphate (cAMP)--that is, forskolin and dibutyryl cAMP--changed this value to a significant extent. Bumetanide had no effect, but ouabain caused a decrease in T-M to 1.30, a 79% inhibition. Elimination of Na+ or HCO3- also caused marked decreases in uptake. Permeability to CF in control medium was 3.40 x 10(-4) cm/min. A decrease of more than 20% (P < 0.05) was seen in the presence of adenosine and cAMP promoters and also with the protein kinase inhibitor H-8, whereas phorbol myristate acetate caused an increase to 4.50 x 10(-4) cm/min (P < 0.01). Ouabain caused no change, but blocked the effects of adenosine. Reducing the Ca2+ concentration of the superfusing medium caused time-dependent increases in permeability to 4.57 at 15 to 45 minutes and 12.5 at 80 to 110 minutes. At the earlier time, this increase in permeability could be prevented by the addition of adenosine or H-8. Elimination of Na+ or HCO3- ions from the medium caused a small decrease in permeability and, like ouabain, blocked the effect of adenosine. Changes in thickness of corneas were consistent, in most cases, with the observed alterations in 86Rb+ uptake or permeability to CF. Scanning electron microscopy showed contraction and rounding of endothelial cells in low Ca2+ medium, with stretching of intercellular borders, features that were largely eliminated when adenosine was also present.

Adenosine, through increasing cAMP, decreases permeability of the corneal endothelium. This effect, rather than a change in the active transport (fluid pump) mechanism, is responsible for the promotion of deturgescence and maintenance of lower steady state thickness of corneas exposed to adenosine. The mechanism may involve the phosphorylation state of cytoskeletal proteins and seems to be dependent on an undisturbed environment of monovalent ions.