Mitochondria in hypoxic pulmonary vasoconstriction: potential importance of compartmentalized reactive oxygen species signaling.

About 15 years ago, Paul Schumacker and his then student, Nav Chandel, reported that in cultured Hep3b cells hypoxia caused mitochondrial production of reactive oxygen species (ROS) that were necessary for accumulation of the master transcriptional regulator in hypoxia, Hif1, and attendant gene expression (1). Since then—and including an important contribution in this issue of the Journal (pp. 424–432) (2)—Greg Waypa and the Schumacker research team have shown that specific disruption of mitochondrial complex III function by genetic deletion of the nuclear gene encoding the Riske iron-sulfur protein (RISP) inhibits hypoxia-induced ROS generation, calcium mobilization, and acute pulmonary vasoconstriction. The current work is significant for at least two reasons. First, it translates previous observations in cultured pulmonary arterial smooth muscle cells (PASMCs) to intact RISP-deficient mice, thus confirming in an integrated system that complex III is indeed a source of hypoxia-induced ROS generation. And second, the current article continues an evolving theme that ROS-mediated signaling is compartmentalized. Regarding the first issue, the murine reagent described by Waypa and colleagues should enable more detailed studies of the links between complex III–derived ROS and pathophysiologic processes that can be examined only in an intact model. Among many possible uses of the model, several seem especially pertinent. For example, it may now be possible to determine whether the long-appreciated suppression of localized hypoxic vasoconstriction and dysregulation of ventilation–perfusion matching in sepsis (3) are associated with defective signaling at the level of mitochondrial complex III. In addition, and as the authors point out, although the available data support the view that complex III–derived ROS are important for the calcium mobilization underlying acute hypoxic vasoconstriction, the question of whether a similar ROS-dependent pathway drives sustained pulmonary vasoconstriction and vascular remodeling in chronic hypoxic pulmonary hypertension should now be amenable to resolution. Finally, the article by Waypa and colleagues raises intriguing questions about the cellular basis of hypoxic pulmonary vasoconstriction. Whereas the current findings support the concept that the PASMC is both a sensor and an effector of hypoxic pulmonary vasoconstriction (3), Wang and coworkers, using multiple strategies to inhibit connexin 40 (Cx40)-mediated gap junctional signaling in intact mice, recently presented evidence that the vasoconstrictor response to alveolar hypoxia is initiated by depolarization of pulmonary capillary endothelial cells (4). In their paradigm, identified as “out of the box” in an accompanying editorial (5), the pulmonary microvascular endothelium, not the pulmonary arterial smooth muscle, functions as the cellular oxygen sensor that initiates a signal conducted retrogradely via endothelial Cx40-containing gap junctions to activate smooth muscle contraction in upstream muscular arteries. Some of the observations reported by Waypa and colleagues may bear on this divergence of evidence. They noted in cultured PASMCs that hypoxia caused only a transient increase in cytosolic calcium that was inhibited by RISP depletion, whereas in small arteries observed in precision-cut lung slices, the RISP-sensitive calcium response was sustained over the course of hypoxic exposure. Perhaps these temporal differences in cytosolic calcium accumulation reflect the contribution of endothelial cells to regulation of hypoxic vasoconstriction in the intact lung tissue that does not occur in cultured PASMCs. Taking this notion one step further, it is reasonable to consider whether the molecular mechanism of the hypoxia-induced depolarization of the pulmonary capillary endothelial cell is similar to that identified in the PASMC; that is, is the endothelial cell depolarization triggered by increased mitochondrial ROS production? Hopefully, the sophisticated strategies and reagents used by the Schumacker (2) and Kuebler (4) labs can be combined to address these questions. The article by Waypa and colleagues also contributes to the evolving concept that ROS-dependent signaling is highly compartmentalized. Using reduction-oxidation–sensitive green fluorescent protein (roGFP) redox probes targeted to the mitochondrial matrix, intermembrane space, and cytoplasm, they found that hypoxia exerted divergent effects on the redox status of these cellular compartments; whereas the mitochondrial matrix was progressively reduced, the intermembrane space and cytosol became more oxidized. This compartmentalized pattern of oxidant stress makes sense; the hypoxia-induced oxidant stress originating at complex III is “vectored” away from the mitochondrial matrix where oxidative damage to the sensitive mitochondrial genome could be expected to disrupt mitochondrial transcription, possibly triggering a bioenergetic crisis and/or cell death (6). And as shown by Waypa and coworkers and discussed below, the oxidant stress in the cytosol triggers calcium accumulation necessary for hypoxic pulmonary vasoconstriction. Because the roGFP used in the study by Waypa and colleagues was diffusely distributed throughout the cytoplasm, it was not possible to determine if the hypoxia-induced, mitochondria-dependent oxidant stress was more or less prominent in specific cytoplasmic domains. This becomes an interesting issue because it is widely appreciated that mitochondria are motile organelles whose distribution has the potential to determine their functional activities (7). For example, kinesin-dependent movement of mitochondria to a submembrane region in close proximity to the “immunologic synapse” in activated immune cells increases local calcium buffering capacity to sustain transmembrane calcium influx through calcium release–activated membrane calcium channels (8). More germane to the focus of the study by Waypa and coworkers, Al-Mehdi and colleagues recently showed in hypoxic pulmonary artery endothelial cells that dynein-driven perinuclear clustering of ROS-producing mitochondria creates a nuclear oxidant stress (9). This nuclear oxidant stress leads to oxidative base modifications in promoter sequences of hypoxia-inducible genes that seem to be important for transcriptional activation. Noted but not pursued in this latter report was the observation that perinuclear mitochondrial clustering was accompanied by a diminution in the peripheral cytosolic density of mitochondria. Putting these findings in the context of Waypa and colleagues’ article, it is tempting to speculate that acute hypoxia-induced mitochondrial ROS production interacts with time-dependent changes in mitochondrial distribution to create signaling microdomains (Figure 1). One critical domain could be at the mitochondrial–sarcoplasmic reticulum (SR) interface, where local ROS might function to regulate SR calcium release and cytosolic calcium accumulation. By contrast, the relative diminution of mitochondria in the vicinity of the cell membrane could serve to create a local environment relatively deficient in mitochondria-derived signaling molecules or functions, thereby impacting the operation of plasma membrane ion channels and other processes. Figure 1. Signal compartmentalization by source translocation. Mitochondria (purple), being motile organelles, can create localized signaling environments by clustering to specific cellular compartments. In normoxia, mitochondria in pulmonary arterial smooth muscle ... The mechanism underlying the reported link between mitochondrial-derived oxidant stress and the hypoxia-induced cytosolic calcium accumulation required for pulmonary vasoconstriction is not understood. It is possible, however, that a compartmentalization of the oxidant stress may contribute to these regulatory processes. In this regard, experiments in cultured PASMCs suggest that hypoxia may induce SR calcium release by causing ROS-mediated dissociation of FK506 binding protein 12.6 from ryanodine receptor 2 (10, 11). Another contributory mechanism might be that perimitochondrial ROS impair mitochondrial calcium sequestration, as reported for carotid glomus cells (12), thereby exaggerating cytosolic calcium accumulation triggered by SR calcium release. In this scenario, SR calcium release and mitochondrial calcium uptake would be reciprocally regulated by complex III–generated ROS. The concept of compartmentalized regulation of ROS signaling in hypoxia as advanced above and inferred from the data of Waypa and colleagues is far from proven. However, we think that it makes intuitive sense. After all, ROS are intrinsically dangerous and can react with macromolecules necessary for cell survival as well as adaptation. Mechanisms restricting access of otherwise cytotoxic ROS to targets important for their signaling function have been described and include their generally short half-lives, the various ROS-scavenging enzymes, the proximity of reactive target molecules, etc. But the Schumacker group’s findings that ROS produced from complex III in the mitochondria—a motile organelle—are directed away from the mitochondrial matrix and into the cytosol point to new concepts by which ROS compartmentalization could be engendered.