No PAR1 immunostaining was observed in umbrella cells. Similarly, basal and intermediate cells also showed MIF immunostaining with umbrella cells showing either slight or no MIF immunostaining. Therefore, the co-existence of PAR1 and MIF is restricted to deeper layers of the rat urothelium. We also document that thrombin stimulation of urothelial cells, whether in vitro or in vivo results in MIF release, and this effect occurs quickly after thrombin application. Since thrombin contained a small amount of endotoxin and because endotoxin can elicit MIF release, it was possible that our thrombin results were due to endotoxin contamination. Heatinactivated thrombin was ineffective in stimulating MIF release from UROtsa cells which argues against this possibility. Much higher temperatures and longer heating times are needed to abolish the activity of endotoxin, therefore it is highly unlikely that our heating conditions affected the activity of the small amount of endotoxin present. Finally, while endotoxin can elicit MIF release it also results in MIF downregulation which is opposite to the effect of up-regulation seen in our studies. Consequently, our findings indicate that the effects observed from thrombin stimulation are not due to endotoxin. Our results thus confirm earlier findings of thrombininduced MIF release from human endothelial cells and we extend those results by showing that the same phenomenon occurs in vivo. Given that both human and rat urothelial cells were shown to express both MIF and PAR1 we consider it likely that thrombin stimulated PAR1 receptors on rat urothelial cells to elicit MIF release. Since PAR1 and MIF TWS119 containing cells in rat urothelium are not located superficially, our findings that intravesical thrombin can induce MIF release from the rat bladder suggest that intravesical thrombin was able to reach those cells to activate PAR1 receptors. Although local thrombin formation during inflammation is likely to occur in the suburothelial compartment and thus stimulate basal and intermediate cells to elicit MIF release, our findings suggest that proteases present in the urine may also be able to activate PAR1 receptors in the urothelium, elicit MIF release and thus contribute to the initiation or maintenance of cystitis. In fact, both mast cell tryptase and neutrophil elastase were documented to be increased in the urine of patients with interstitial cystitis, thus raising the possibility that PAR1 receptors could be activated in interstitial cystitis. Activation of PAR1 receptors by neutrophil elastase was reported to induce apoptosis in lung epithelial cells while activation of PAR1 and PAR2 receptors were shown to increase epithelial permeability in intestinal epithelia. Whether these effects can also be seen in urothelial cells, particularly in clinical conditions such as interstitial cystitis, remains to be determined. Treatment with intravesical PAR1, PAR2 and PAR4 agonists induced inflammation in the mouse bladder.