Reactive Oxygen Species Can Provide Atheroprotection via NOX4-Dependent Inhibition of Inflammation and Vascular Remodeling

Objective Oxidative stress is considered a hallmark of atherosclerosis. In particular, the superoxide-generating type 1 NADPH oxidase (NOX1) has been shown to be induced and play a pivotal role in early phases of mouse models of atherosclerosis and in the context of diabetes mellitus. Here, we investigated the role of the most abundant type 4 isoform (NOX4) in human and mouse advanced atherosclerosis. Approach and Results Plaques of patients with cardiovascular events or established diabetes mellitus showed a surprising reduction in expression of the most abundant but hydrogen peroxide (H2O2)-generating type 4 isoform (Nox4), whereas Nox1 mRNA was elevated, when compared with respective controls. As these data suggested that NOX4-derived reactive oxygen species may convey a surprisingly protective effect during plaque progression, we examined a mouse model of accelerated and advanced diabetic atherosclerosis, the streptozotocin-treated ApoE(-/-) mouse, with (NOX4(-/-)) and without genetic deletion of Nox4. Similar to the human data, advanced versus early plaques of wild-type mice showed reduced Nox4 mRNA expression. Consistent with a rather protective role of NOX4-derived reactive oxygen species, NOX4(-/-) mice showed increased atherosclerosis when compared with wild-type mice. Deleting NOX4 was associated with reduced H2O2 forming activity and attenuation of the proinflammatory markers, monocyte chemotratic protein-1, interleukin-1, and tumor necrosis factor-, as well as vascular macrophage accumulation. Furthermore, there was a greater accumulation of fibrillar collagen fibres within the vascular wall and plaque in diabetic Nox4(-/-)ApoE(-/-) mice, indicative of plaque remodeling. These data could be replicated in human aortic endothelial cells in vitro, where Nox4 overexpression increased H2O2 and reduced the expression of pro-oxidants and profibrotic markers. Interestingly, Nox4 levels inversely correlated with Nox2 gene and protein levels. Although NOX2 is not constitutively active unlike NOX4 and forms rather superoxide, this opens up the possibility that at least some effects of NOX4 deletion are mediated by NOX2 activation. Conclusions Thus, the appearance of reactive oxygen species in atherosclerosis is apparently not always a nondesirable oxidative stress, but can also have protective effects. Both in humans and in mouse, the H2O2-forming NOX4, unlike the superoxide-forming NOX1, can act as a negative modulator of inflammation and remodeling and convey atheroprotection. These results have implications on how to judge reactive oxygen species formation in cardiovascular disease and need to be considered in the development of NOX inhibitory drugs.