Targeting naproxen to non-parenchymal liver cells protects against endotoxin induced liver damage.

Non-steroidal anti-inflammatory drugs (NSAID's) could be of value in the treatment of liver disease; however, their use in this situation is limited by renal side effects. Therefore, we explored whether naproxen covalently bound to human serum albumin NAP-HSA) was able to reduce toxicity in an acute model of liver disease induced by endotoxin in rats pretreated with Corynebacterium parvum. In the isolated perfused liver of such animals endotoxin induced cholestasis (0.62 +/- 0.05 vs. 0.24 +/- 0.09 microliter.min-1.g liver-1; p < 0.05), increased vascular resistance (11300 +/- 400 vs. 311000 +/- 2000 dyn.s.cm-5; p < 0.05) and alanine aminotransferase release (22 +/- 9 vs. 149 +/- IU/l; p < 0.05). At the highest dose tested (22 mg/kg, corresponding to 6.0 mumoles naproxen), NAP-HSA normalized ALT release (21 +/- 10 IU/l: p < 0.05) while an equimolar amount of non-targeted naproxen was only partially effective (56 +/- 19 IU/l). A conventional dose of naproxen similarly prevented transaminase release. Cholestasis and increased vascular resistance were also prevented by NAP-HSA. Drug targeting by linking drugs to proteins is a potentially useful approach to maximizing drug effect while minimizing adverse events; this could be particularly useful for compounds with potentially serious adverse effects in patients with chronic liver disease such as the nonsteroidal anti-inflammatory agents used in the present study.