Diabetes doubles the risk of cardiovascular disease (CVD) independently of other risk factors (1). A 50-year-old with diabetes is likely to die, on average, 6 years earlier than a counterpart without diabetes, with vascular deaths being the major contributor to reduced survival (2). In keeping with the predicted rise in diabetes prevalence, the proportion of CVD deaths attributable to diabetes (currently 10% in developed countries ) is likely to increase substantially. Although intensive research efforts have identified the molecular mechanisms contributing to diabetes-related CVD, these discoveries have not been mirrored by major pharmaceutical advances. As a blockbuster drug to reduce CVD in diabetes has failed to emerge, other approaches need to be considered as a matter of urgency.
Although robust evidence supports the benefits of blood pressure reduction and lipid lowering in diabetes, the appropriateness of intensive glucose lowering as a tool to reduce cardiovascular risk is now questionable. In individuals newly diagnosed with diabetes, the UK Prospective Diabetes Study trial showed that intensive glycemic control with insulin or sulphonylurea resulted in a nonsignificant 16% risk reduction in myocardial infarction (3). Further, it was only after 10 years of follow-up that a 15% relative risk reduction emerged, suggesting a possible legacy effect of intensive control early in the disease process (4). In contrast, a series of large randomized trials investigating intensive glucose control in patients with diabetes of longer duration and/or established CVD has failed to demonstrate benefit. The Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation and Veterans Affairs Diabetes Trial studies reported no benefit from intensive glucose lowering on cardiovascular events or mortality (5,6). The Action to Control Cardiovascular Risk in Diabetes study, which randomized 10,251 diabetic patients at high risk for cardiovascular events, was terminated early because of increased mortality in the intensive intervention group (7). A high incidence of hypoglycemia associated with intensive glucose lowering is a likely explanation for the increased mortality (8). More recently, the Outcome Reduction with Initial Glargine Intervention study, which tested use of insulin glargine to normalize fasting plasma glucose, also failed to demonstrate a reduction in cardiovascular events (9).
If intensive lowering of blood glucose is ineffective in reducing CVD events, what about targeting the cellular consequences of hyperglycemia rather than glucose per se? Might this approach deliver CVD prevention without the potentially unfavorable effects of hypoglycemia? Endothelial dysfunction (characterized by reduced bioavailability of nitric oxide and increased production of reactive oxygen species [ROS]) plays a critical role in the pathogenesis of diabetic vascular dysfunction. Although multiple cellular sources have been implicated in endothelial ROS generation (10), mitochondrial ROS is the principal contributor to hyperglycemic endothelial dysfunction (Fig. 1) (11). Cross-talk between mitochondria and NADPH oxidase facilitates a vicious feed-forward cycle of endothelial ROS generation (12), highlighting mitochondrial ROS as a suitable target for pharmacological inhibition (13).
In this issue of Diabetes, Gerö et al. (14) used a cell-based screening approach to identify potential inhibitors of hyperglycemia-induced endothelial ROS generation. They coupled this strategy with a drug repositioning approach, screening a library of existing clinical drugs and drug-like molecules to identify compounds that reduced mitochondrial ROS generation without jeopardizing cell viability. Of the handful of compounds so identified, the antidepressant paroxetine was selected for further study. Paroxetine reduced hyperglycemia-induced endothelial ROS generation, mitochondrial protein oxidation, and DNA damage without interfering with mitochondrial electron transport or cellular bioenergetics. To confirm a favorable effect on vascular phenotype, the investigators then showed that acute and chronic paroxetine treatment improved (though did not completely reverse) endothelial dysfunction in rat aortic rings exposed to hyperglycemia. Although these findings are persuasive, Gerö et al. acknowledged that certain questions remain unanswered. For example, although it is likely that the principal site of antioxidant action is within the sesamol moiety of paroxetine, the molecular mechanisms by which it inhibits mitochondrial ROS require further evaluation. Furthermore, the observation that paroxetine reduces xanthine oxidase–derived ROS in a cell-free system indicates that its antioxidant properties are not specific to mitochondria, thus arguing for detailed characterization of paroxetine’s action on all cellular sources of ROS.
Drug repositioning offers an alternative to conventional drug discovery by finding new uses for existing medicines or compounds outside the scope of their original indication (15). The concept is not new: sildenafil is a well-known example of a drug identified serendipitously for erectile dysfunction following its original development as an antiangina medication. In the cardiovascular arena, systematic drug repositioning approaches have been used to identify drugs to prevent ischemia reperfusion injury or promote angiogenesis, but they have not previously been reported for diabetes-specific vascular dysfunction. As the safety profiles of repositioned drugs are often known, time-to-market is potentially reduced and less risky than de novo drug development. This is pertinent to the diabetes field, where requirements for preauthorization of cardiovascular risk assessment introduced by the U.S. Federal Drug Administration after safety concerns emerged over rosiglitazone may be restricting drug development (16).
Identification of paroxetine as a novel mitochondrial ROS inhibitor warrants its further evaluation in other experimental models, in particular its ability to reduce atherosclerosis. However, identifying new roles for other drugs by this approach comes with important caveats. First, the phenotypic screen carried out by Gerö et al. discounted statins because they reduced cell viability—contrasting with the unequivocal evidence for cardioprotective effects in diabetes. Second, insulin resistance, which usually precedes the development of hyperglycemia, is associated with ROS generation that is mediated by exposure to circulating cytokines and free fatty acids. Focusing exclusively on hyperglycemia-induced ROS will miss opportunities for ROS inhibition in this critical early phase of atherogenesis. Finally, it is now apparent that low-level mitochondrial ROS generation is critical to endothelial physiology by modulating cell differentiation, immunity, autophagy, and metabolic adaptation (Fig. 1) (17). The divergent functions of mitochondrial ROS to promote cell damage and promote cellular adaptation render it a potentially challenging therapeutic target and may explain why nonselective antioxidant strategies failed to prevent CVD and increased mortality (18).
With paroxetine, we have the reassurance of many years of clinical experience with no signal for harm. An association between selective serotonin uptake inhibitors and reduced cardiovascular risk in depression (19,20) provides a springboard to pursue the drug repositioning strategy initiated by Gerö et al. Paroxetine should now continue its journey from identification as a mitochondrial ROS inhibitor through further preclinical studies to clinical trials in individuals with diabetes. Ultimately, the new trick of this old drug might ease the burden of CVD in diabetes.
See accompanying original article, p. 953.
No potential conflicts of interest relevant to this article were reported.