Letter by Hüttemann et al., Regarding Article, “Ndufs2, a Core Subunit of Mitochondrial Complex I, Is Essential for Acute Oxygen-Sensing and Hypoxic Pulmonary Vasoconstriction”

To the Editor:

Oxygen sensing in the lung is a complex phenomenon and essential to induce hypoxic pulmonary vasoconstriction (HPV) during local alveolar hypoxia to escape potentially life-threatening hypoxemia. In contrast to the systemic vasculature, which expands when oxygen is scarce to increase oxygen delivery to target organs, acute HPV leads to vasoconstriction in certain areas of the lung to shunt blood from poorly- to well-ventilated alveoli, thereby optimizing oxygen extraction. Thus, HPV requires an oxygen sensing and/or signaling mechanism that is distinct from the systemic vasculature. There is still much debate about the molecular oxygen sensor that initiates a cascade eventually leading to HPV, and evidence suggests that altered superoxide generation by the mitochondrial electron transport chain (ETC) is an early event^{1, 2}. This leads to activation/inhibition of sarcoplasmic and/or plasmalemmal ion channels (including closure of voltage gated potassium channels), which causes an increase in intracellular calcium in pulmonary artery smooth muscle cells (PASMCs), triggering vasoconstriction^3–5. However, an essential question remains: what is the molecular mechanism that senses the oxygen concentration in the lung and how does this lead to increased superoxide?

We saw with interest the recent article “Ndufs2, a Core Subunit of Mitochondrial Complex I, Is Essential for Acute Oxygen-Sensing and Hypoxic Pulmonary Vasoconstriction,”⁶ adding another potential player to the oxygen sensing unit. There is one important misinterpretation that we would like to bring to the readers’ attention, which is not supported by data provided in the manuscript.

As already stated in the abstract, the authors claim that the lung-specific subunit of cytochrome c oxidase, COX4I2, previously reported by us to be essential for HPV⁷, is not required for HPV. The authors draw this false conclusion from data presented in Supplemental Figure 5I, where they show a partial siRNA-based knockdown of COX4I2 in cultured cells. Despite the lack of replicates for the control cells presented in the Western blot with very weak signals that are only slightly above background, the authors report that COX4I2 expression is about 50% reduced in the knockdown cells. Using this system, they show that there is still a measurable increased calcium release in COX4I2 knockdown cells when exposed to hypoxia (Supplemental Figure 5J), and then conclude that the gene is not essential for HPV. Importantly, to draw such a conclusion, the authors must presume that partial knockdown leads to full haploinsufficiency, which is incorrect as discussed below. First, we have previously unambiguously shown in vivo, using Cox4i2 knockout mice, that Cox4i2 is essential for acute HPV, since HPV is abolished in the knockout mice⁷. Second, we also analyzed HPV of Cox4i2^+/− heterozygous mice at the same time as wild-type and knockout mice in Figure 1B in Sommer et al.⁷ in a randomized fashion but the data were not included in the original manuscript: isolated mouse lungs were ventilated with normoxic (21% O₂, 5% CO₂, 74% N₂) or hypoxic (1% O₂, 5% CO₂, 94% N₂) gas for 10 min. Hypoxic ventilation resulted in HPV, detected as an increase in pulmonary arterial pressure compared to pressure during normoxic ventilation (ΔPAP). Wild-type and Cox4i2^−/− knockout mice were identical with, and have been taken from, the groups of Figure 1B in Sommer et al.⁷ for comparison with Cox4i2^+/− mice. ΔPAP (mm Hg±SE) was 1.0±0.11 for wild-type (n=5; P<0.001 versus Cox4i2^−/− mice), 0.75±0.07 for Cox4i2^+/− mice (n=6; P<0.001 versus Cox4i2^−/− mice), and 0.1±0.05 for Cox4i2^−/− mice (n=6). Therefore, heterozygous mice have an HPV response that is present and only about 25% reduced compared to the wild-type animals whereas the knockout mice show no response. In conclusion, Cox4i2 is essential for HPV and its complete absence is required to show this effect.