Abstract
Clioquinol (CQ) was widely used as oral antibiotic before being taken off the market in many countries in 1970, after it was linked to subacute myelo-optic neuropathy (SMON) in Japan, leading to vision loss with many patients left wheelchair-bound. The common pathology of CQ-associated SMON was reproduced in animals but none of the proposed modes of toxicity explained the restriction of CQ-induced SMON to Japan. Given a re-emergence of CQ and related analogues as neuroprotectants, it is crucial to understand the underlying mechanism of CQ-induced toxicity to prevent any potential CQ-associated risks to future patients. A small molecule screen to find drugs that induce mitochondrial dysfunction in vitro identified CQ and the structurally related 8-hydroxyquinoline (8-OHQ). Their mitochondrial liability, pro-oxidative and cytotoxic activity was subsequently confirmed in some cell lines but surprisingly not in others. Subsequent studies in isogenic cell lines demonstrated that the antioxidant protein NQO1 is differentially expressed in the cell lines tested and potently protects against CQ toxicity. CQ-induced reduction of cellular ATP levels, increased lipid peroxidation and elevated cell death was also attenuated by antioxidants, implicating oxidative stress as the core mechanism of CQ-induced toxicity. These in-vitro findings were replicated in zebrafish. Visual acuity in zebrafish larvae that do not express NQO1, was reduced by CQ in a dose-dependent manner, while CQ did not affect visual function in the adult zebrafish that express NQO1. Similarly, pharmacological inhibition of NQO1 activity resulted in CQ-induced oxidative stress in the retina and severe acute systemic toxicity in the adult fish. Given the much higher prevalence of the inactivating C609T NQO1 polymorphism in the Japanese population compared to the European population, the results of this study could for the first time indicate how the geographic restriction of SMON cases to Japan could be explained. Importantly, if CQ or its derivatives are to be used safely for the treatment of neurodegenerative diseases, it seems imperative that NQO1 levels and activity of prospective patients should be ascertained.