Metabolic reprogramming, oxidative stress, and pulmonary hypertension

Abstract

Mitochondria are highly dynamic organelles essential for cell metabolism, growth, and function. It is becoming increasingly clear that endothelial cell dysfunction significantly contributes to the pathogenesis and vascular remodeling of various lung diseases, including pulmonary arterial hypertension (PAH), and that mitochondria are at the center of this dysfunction. The more we uncover the role mitochondria play in pulmonary vascular disease, the more apparent it becomes that multiple pathways are involved. To achieve effective treatments, we must understand how these pathways are dysregulated to be able to intervene therapeutically. We know that nitric oxide signaling, glucose metabolism, fatty acid oxidation, and the TCA cycle are abnormal in PAH, along with alterations in the mitochondrial membrane potential, proliferation, and apoptosis. However, these pathways are incompletely characterized in PAH, especially in endothelial cells, highlighting the urgent need for further research. This review summarizes what is currently known about how mitochondrial metabolism facilitates a metabolic shift in endothelial cells that induces vascular remodeling during PAH.

1. Introduction

Pulmonary hypertension (PH) is a general term used to describe high blood pressure in the lung. Importantly, PH does not indicate a cause but describes the resulting high blood pressure state, which may stem from various factors. A subtype and severe form of PH is pulmonary arterial hypertension (PAH). PAH is an incurable disease where the pulmonary arteries constrict and undergo remodeling. The resulting high blood pressure increases the workload on the right side of the heart, as the heart exerts more force to overcome the increased pulmonary pressure and deliver blood to the lungs. This extra strain on the right side of the heart leads to right heart dysfunction, failure, and eventually death [1]. PAH has a poor prognosis, with a three-year survival rate of only 58%. There is no cure for PAH; vasodilator drugs only temporarily relieve symptoms. As the disease progresses, lung transplantation becomes the only option [2]. The lack of treatments for PAH stems from the fact that we still do not understand the complex molecular mechanisms that lead to the development of the disease; however, mitochondrial dysfunction, inflammation, fibrosis, and endothelial dysfunction are all implicated in developing the disease [3] and warrant further exploration.

Patients with idiopathic PAH either have no known genetic cause or have mutations in genes that are associated with heritable forms of PAH. The most prevalent genetic mutation in PAH occurs in the bone-morphogenetic protein receptor-2 (BMPR2), a member of the transforming growth factor-β (TGF-β) superfamily [4,5]. Around 75% of familial PAH cases and 25% of idiopathic PAH patients have a BMPR2 mutation. However, BMPR2 has a low penetrance of only 20–30% [6,7]. This low penetrance suggests BMPR2 mutations may predispose an individual to PAH, but a second stimulus is needed to develop the disease, e.g., other genetic, physiological, or environmental factors. This is supported by animal models, in which BMPR2 deletion is insufficient to induce PAH in most cases [[8], [9], [10]]. However, BMPR2 plays a significant role in the development of PAH and is still considered causally linked to PAH [4,5,8,[11], [12], [13], [14], [15], [16]].

Despite the different genetic components in familial versus idiopathic PAH, vascular remodeling is likely influenced by the same or similar molecular pathways. Thus, both forms of PAH can give valuable insight into understanding the disease. Additionally, women are more likely to develop PAH than men; therefore, there may be a sex component that affects the disease initiation [17]. To date, the etiology behind PAH remains largely unknown and highly complex. For example, early in PAH development, endothelial cells are more prone to apoptosis. In contrast, in the later stages of PAH, endothelial cells are more hyper-proliferative and anti-apoptotic [18]. What is clear is that PAH involves multiple cell types and pathways. Dysfunctional mitochondria and metabolic reprogramming favoring aerobic glycolysis, i.e., a Warburg phenotype, appear to influence the pulmonary vascular remodeling seen in PAH [[19], [20], [21], [22]]. Targeting cellular metabolism in endothelial cells offers a novel approach to finding a treatment for PAH.

Most research on lung diseases has revolved around alveolar epithelial cells, lung macrophages, smooth muscle cells, and fibroblasts. However, mounting evidence suggests that there are important contributions from vascular endothelial cells to disease development. Furthermore, increasing evidence suggests that pulmonary endothelial cells exhibit aberrant mitochondrial morphology and function compared to other cell types with well-defined mitochondrial functions; therefore, information obtained from different cell types cannot be assumed directly applicable to endothelial cells. In this review, we will highlight the involvement of endothelial cell mitochondria and metabolic changes in the development of PAH to identify possible novel therapeutic strategies to treat this disease with limited therapeutic options.

2. Pulmonary endothelial cells require mitochondrial function

The lungs are frequently mistaken as passive, non-metabolically active conduits for gas exchange. However, emerging evidence highlights that dysfunction of lung metabolism is a vital component of lung disease. Mitochondria are crucial to supporting cellular metabolism and central to maintaining normal lung homeostasis. However, mitochondria are also involved in cell death, heat generation, intracellular calcium regulation, thermoregulation, and production of reactive oxygen species (ROS), among other functions [23]. Therefore, alterations in mitochondrial function can significantly alter cellular homeostasis, affecting a multitude of pathways. Understanding how metabolic pathways are dysregulated in pulmonary diseases could lead to the discovery of novel and specific therapeutic targets.

In the alveoli of the lungs, there is a delicate membrane formed by vascular endothelial and alveolar epithelial cells. These cells have fused basal lamina, which allows for rapid gas exchange while preventing air bubbles from forming in the blood and blood from entering the alveoli. The endothelial cells are a continuous single-cell layer that lines the pulmonary vasculature and provides a semi-permeable barrier between the blood and interstitium [24]. Maintaining this endothelial barrier is one of the most important metabolic functions of the lungs. The pulmonary endothelium is a highly metabolic and interconnected cell network [[25], [26], [27], [28]], with energy required to maintain and adapt to environmental changes [29,30]. Endothelial cells require mitochondrial metabolism to produce macromolecules, cell signaling mediators and maintain overall endothelial cell health and function. Therefore, mitochondria are essential for endothelial cell function and the overall health of the lungs. In the following sections, we will discuss how mitochondria are at the center of glucose and lipid metabolism and how these affect overall endothelial cell homeostasis and function.

3. Metabolic pathways involved in ATP generation in the lung endothelium

Glucose is the primary source of cellular energy for most tissues, including the lungs. During glycolysis, glucose is converted into two pyruvate molecules, two net ATP molecules, and two NADH (nicotinamide adenine dinucleotide) molecules. Pyruvate can then be oxidatively metabolized to carbon dioxide (CO₂) in the mitochondria using the tricarboxylic acid cycle (TCA). The TCA cycle generates byproducts, including NADH, used as electron donors to fuel the electron transport chain (ETC), producing 36 ATP molecules through oxidative phosphorylation [31]. Under normal aerobic conditions, this is the preferred pathway to metabolize glucose, as it generates the most ATP per glucose molecule (Fig. 1A). Alternatively, pyruvate can be reduced to lactate, a process known as fermentation. Glucose fermentation generates significantly less ATP than when pyruvate enters the TCA cycle, but it does not require oxygen; therefore, this pathway is termed anaerobic respiration, although it can occur in the presence of oxygen.

Fig. 1.

Glucose and Fatty Acids are the Major Carbon Sources in Endothelial Cells. (A) Under normal, aerobic conditions, pyruvate is converted into Acetyl-CoA, which can enter the TCA cycle, generating byproducts that can participate in oxidative phosphorylation, generating a net of 32 ATP molecules. (B) Fatty acid metabolism uses the carnitine palmitoyl transferase (CTP) system. The CPT system allows FA to be transported into the mitochondrial matrix and undergo FAO. First, a FA is conjugated to a molecule of CoA by the enzyme fatty acyl-CoA synthetase. Subsequently, the CoA is replaced with a molecule of carnitine and brought into the intermembrane space via the enzyme CPT1A. The fatty-acyl-carnitine is then transported to the mitochondrial matrix via carnitine acylcarnitine translocase. Once the fatty-acyl-carnitine is in the mitochondrial matrix, CPT2 converts the fatty-acyl-carnitine back into fatty acyl-CoA, where it can then undergo fatty acid oxidation. Both carnitine and acetyl-carnitine can be transported back into the cytosol by carnitine acylcarnitine translocase. CrAT is a mitochondrial matrix enzyme that catalyzes acetyl-CoA conversion to acetyl-carnitine.

Abbreviations: ATP: adenosine triphosphate, NADH: nicotinamide adenine dinucleotide, FADH₂: flavin adenine dinucleotide, TCA: Tricarboxylic Acid Cycle, CPT: carnitine palmitoyl transferase, FA: Fatty Acids, FAO: fatty acid oxidation, CoA: coenzyme A, CrAT: Carnitine acetyltransferase.

The TCA cycle occurs in the mitochondrial matrix, and the two significant metabolites generated are nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FADH₂). NADH and FADH₂ can enter the ETC to participate in oxidative phosphorylation [32]. The ETC is a series of protein complexes that undergo electron-transfer reactions that ultimately generate ATP through oxidative phosphorylation. Notably, the ETC contributes to the generation of a mitochondrial membrane potential. Under normal conditions, the mitochondrial membrane potential remains stable and is essential for ATP generation [33]. However, loss of mitochondrial membrane potential leads to energy collapse and cytochrome c release from the mitochondrial intermembrane space to the cytosol, initiating the intrinsic apoptotic pathway. Although Acetyl-CoA is considered the beginning of the cycle, other intermediates can also serve as alternative entry points. For example, the metabolism of amino acids can generate α-ketoglutarate (α-KG), oxaloacetate, fumarate, and succinyl-CoA. These metabolites can then enter the TCA cycle directly [34]. Importantly, TCA cycle intermediates are not purely fuel for the cells, as recent studies have highlighted the significance of TCA intermediates in controlling cell function. For example, acetyl-CoA regulates chromatin dynamics [[35], [36], [37]], α-KG is a crucial modulator of the hypoxic response [38], and fumarate is involved in anti-inflammatory signaling [39]. These are just a few examples of the many functions played by TCA intermediates, and their roles in disease are only now starting to emerge. Unfortunately, alterations in TCA metabolism are largely unexplored in the context of endothelial dysfunction during PAH.

The other primary carbon source for endothelial cells is fatty acids (FA). FA are metabolized by fatty acid oxidation (FAO) in the mitochondria through a cyclic series of reactions that converts FA into acetyl-CoA, resulting in the shortening of the FA, 2 carbons at a time, while generating NADH, FADH₂, and acetyl-CoA. Ultimately, FAO is a critical energy source, as this process produces 2.5 times more ATP per mole when compared to glucose oxidation [40]. Furthermore, energy homeostasis depends on mitochondrial FAO during prolonged fasting, exercise, and metabolic stress [41]. Central to the mitochondrial FAO pathway is the carnitine palmitoyl transferase (CPT) system (Fig. 1B) [42].

The mitochondrial membrane is impermeable to FA; therefore, the CPT system is required to facilitate their transport into the mitochondrial matrix for FAO [43]. There are 3 isoforms of CPT1 (CPT1A, CPT1B, and CPT1C), with CPT1A being the predominant isoform in lung tissues [44]. CPT1 is considered one of the critical rate-limiting enzymes of FAO, and its activity is often used to measure FAO. Once in the mitochondria, FA can undergo a series of reactions to produce metabolites, such as acetyl-CoA, that can subsequently enter the TCA cycle. Key enzymes of this process include acyl-CoA dehydrogenases (CAD) which mediate dehydrogenation, hydration, oxidation, and thiolysis of fatty acids [41]. One of the major mitochondrial regulators of FAO is the peroxisome proliferator–activated receptor (PPAR) γ coactivator (PGC)-1α, which induces CPT1 expression [43]. It is essential to recognize that PGC-1α has extensive roles in cellular function, including mitochondrial biogenesis, FAO, triglyceride metabolism, and angiogenesis [45]. Both deficiency and over-activation of the FAO cause disruptions in normal energy metabolism and result in inflammatory oxidative damage, thus disrupting immune homeostasis and leading to acute or chronic inflammatory diseases.

4. The Warburg effect in the lung endothelium

Mounting evidence highlights the complexity of energy metabolism regulation. For example, there are examples of fast-growing unicellular organisms that rely on glucose fermentation for proliferation, regardless of oxygen availability [46]. One of the most famous examples of this “metabolic shift” is the Warburg effect. Initially described in cancer cells, the Warburg effect is an energy shift from mitochondrial oxidative phosphorylation to aerobic glycolysis [47], where pyruvate is preferentially fermented to lactate, even in aerobic conditions. To ferment pyruvate into lactate, NADH donates its electron to pyruvate. This is a critical reaction, as it regenerates NAD+, allowing it to participate again in glycolysis (Fig. 2A). Initially, it was thought that the Warburg effect was an adaptation response to dysfunctional mitochondria; however, this is no longer the case, as in some cases, the mitochondrial function in cancer cells is enhanced. The current theory is that aerobic glycolysis is a survival mechanism providing metabolic advantages in certain conditions. For example, even though aerobic glycolysis produces less ATP than mitochondrial respiration, ATP is produced at a faster rate when compared to oxidative phosphorylation.

Fig. 2.

Endothelial cells exhibit a “Warburg-like” shift in metabolism. (A) The Warburg effect is characterized by a metabolic shift towards aerobic glycolysis. This results in more lactate production, regeneration of NAD⁺, and less ATP production. (B) The pentose phosphate pathway generates intermediates needed for proliferation.

Abbreviations: ATP: adenosine triphosphate, TCA: Tricarboxylic Acid Cycle, NADH: nicotinamide adenine dinucleotide, PPP: pentose phosphate pathway, NADPH: nicotinamide adenine dinucleotide phosphate.

Additionally, glycolysis supplies the intermediates for the pentose phosphate pathway (PPP) necessary to support proliferation. For example, glycolysis produces glucose-6-phosphate, which can participate in the PPP to ultimately produce intermediates needed for cholesterol, FA, and nucleotide synthesis (Fig. 2B) [46]. This highlights a possible advantage of using aerobic glycolysis instead of oxidative phosphorylation for ATP generation. It is now widely accepted that the Warburg effect is observed not only in cancer but in pulmonary endothelial cells. Endothelial cells are highly glycolytic and produce large amounts of lactate, even in high O₂ tension [48]. Furthermore, when angiogenesis is stimulated, endothelial cells further increase the levels of aerobic glycolysis, highlighting that aerobic glycolysis is dynamically regulated and is influenced by the tissue environment and external stressors. Interestingly, the rate of glycolysis also differs between the endothelial cell subtypes: arterial, microvascular, and venous. Arterial endothelial cells are more oxidative, while microvascular endothelial cells are more glycolytic and proliferative [25,49].

It is still unresolved why endothelial cells preferentially use glycolysis to generate energy. This may allow endothelial cells to use glucose, the most abundant extracellular nutrient, to produce more ATP faster if glucose uptake and flux are increased [48,[50], [51], [52]]. Furthermore, it may confer a survival or functional advantage over oxidative phosphorylation. For example, endothelial cells are more resistant to hypoxia than other cell types because they use glycolysis anaerobically for as long as glucose is available. When glucose levels decrease, endothelial cells become sensitive to oxygen, suggesting glycolysis is directly responsible for their resistance to hypoxia [50]. Fascinatingly, endothelial cells can store glucose intracellularly as glycogen, highlighting that glucose metabolism is so critical that they create a reserve of glucose, like skeletal muscles and the liver [53,54]. However, how endothelial cells utilize glycogen is still unknown. Nor is it fully understood why endothelial cells have this “Warburg-like” metabolic phenotype. Although this has implications for how endothelial cells function, it is critical to fully define how these pathways differ in endothelial cells compared to other cell types.

The enzyme hexokinase catalyzes the first and rate-limiting step in glycolysis. Hexokinase converts glucose to glucose-6-phosphate (G6P). This is a critical step, as converting glucose to G6P commits glucose to the cell, as once glucose is phosphorylated, it can no longer leave the cell. Aggressive cancers exhibit higher levels of hexokinase II, bound to the mitochondrial membrane, which confers a survival advantage [55]. Although well-established in cancer, the role of changes in hexokinase expression during pulmonary disease is less understood. However, available evidence suggests that some of the principles in cancer can also be applied to pulmonary disease. For example, elevated levels of hexokinase II are protective against oxidative stress [56], a significant component of many pulmonary diseases, including PAH.

The PPP depends on G6P, which is generated during the beginning of the glycolytic pathway, demonstrating the tight interconnectedness of glycolysis and the PPP (Fig. 2). A unique feature of the PPP is that it neither generates nor consumes ATP. Instead, the PPP has a variety of roles, one of the most important being its ability to reduce NADP⁺ to NADPH (nicotinamide adenine dinucleotide phosphate). NADPH is essential for fatty acid and cholesterol synthesis and for regenerating glutathione for redox homeostasis. Furthermore, the PPP generates ribose-5-phosphate, a component of ribonucleotides, a precursor of nucleic acid synthesis. Therefore, the PPP is an essential metabolic pathway for proliferating and transcriptionally active cells. In the lung, the PPP is responsible for most NADPH production [57]. In rats, an estimated 10% of glucose metabolized by the lung is used in the PPP [[57], [58], [59]]. Glucose-6-phosphate dehydrogenase (G6PD) is the rate-limiting PPP enzyme that converts G6P to 6-phosphogluconolactone and reduces NADP⁺ to NADPH. When G6PD is inhibited, endothelial cell viability decreases, suggesting that the PPP is essential for endothelial survival [54,60].

5. Lactate metabolism

Lactate is produced in the lung at a higher rate than in other tissues and at levels similar to cancer cells exhibiting the Warburg effect. In the rat lung, carbon labeling methods have determined that 40% of all consumed glucose is converted to lactate [61,62]. Furthermore, lactate production in the lung is not primarily dependent on oxygen conditions [63,64], while lactate metabolism is oxygen dependent in most tissues. The role of lactate in lung function is unresolved and highlights our fundamental lack of understanding of energy metabolism in lung physiology. However, these observations suggest that the lung uses aerobic glycolysis for a specific purpose, even in aerobic conditions. No consensus has formed in the field, with work indicating that aerobic glycolysis acts as a buffer to maintain the correct redox homeostasis by regenerating NAD⁺ or increasing fatty acid synthesis [65]. While others suggest aerobic glycolysis as a strategy that the lung use to reduce oxygen consumption to ensure optimal oxygen delivery efficiency [66,67]. Another theory is that lactate is an energy source such that when pulmonary cells receive an inadequate oxygen supply, they can use lactate to obtain energy [68]. This resembles a strategy some cancers use: lactate is secreted in the tumor microenvironment and used by cells not receiving sufficient oxygen [66]. The cells that receive lactate reverse the lactate dehydrogenase reaction to form pyruvate. Pyruvate then enters the TCA cycle and produces energy. This process ultimately decreases the dependence of the cell on blood glucose, as it can utilize lactate as an energy source. However, this excessive lactate production's role in the lung remains unresolved.

The conversion of pyruvate to lactate (and vice versa) is the enzyme lactate dehydrogenase (LDH). Elevated levels of LDH are associated with worse prognosis in lung cancers, and LDH levels have been used as a cancer clinical marker [69,70]. LDH has multiple isoforms, and the highest expressed form in the lung is LDH3 [71]. LDH3 is composed of two subunits, LDHA and LDHB. Although each isoform of LDH contains varying ratios of LDHA and LDHB, LDH3 is a heterotetramer composed of 2 LDHA subunits and 2 LDHB subunits (Fig. 3A) [72,73]. LDHA and LDHB differ in their affinity for lactate and pyruvate. LDHA has a higher affinity towards pyruvate and preferentially converts pyruvate to lactate. LDHB has a higher affinity for lactate and preferentially converts lactate into pyruvate [74,75] (Fig. 3B). It is important to note that although LDHA and LDHB preferentially bind to different substrates, they can facilitate the reactions in either direction. This is highlighted by the fact that lactic acid production is unchanged if LDHA expression is silenced [76].

Fig. 3.

LDH3 comprises LDHA and LDHB subunits with different affinities for the substrates. (A) LDH3 consists of 2 LDHA subunits and 2 LDHB subunits (B) The conversion of lactate to pyruvate is a reversible reaction. LDHA preferentially binds pyruvate and therefore catalyzes the conversion of pyruvate to lactate at a higher rate than converting lactate to pyruvate. LDHB preferentially binds lactate and catalyzes lactate to pyruvate at a greater rate than the reverse reaction.

Abbreviations: LDH: lactate dehydrogenase, NADH: nicotinamide adenine dinucleotide.

The fact that LDH3 is composed of equal units of LDHA and LDHB suggests that the conversion of lactate to pyruvate is physiologically essential to the lung. This is potentially significant when compared to the LDH5 isoform in skeletal muscle and the liver, which is composed solely of LDHA [77], suggesting that these tissues do not need to convert lactate into pyruvate. Although lactate's role in the lung's normal functioning remains unclear, mounting evidence suggests that lactate plays an essential role in lung disease. For example, idiopathic pulmonary fibrosis (IPF) patients have LDH tetramers with higher ratios of LDHA:LDHB and elevated lactate levels, suggesting that converting pyruvate to lactate is favored [78]. This shift indicates that converting lactate to pyruvate is essential in normal lung physiology and needs to be better defined to fully understand the development of pulmonary disease.

6. Dysfunctional mitochondria influence the anti-apoptotic, pro-proliferative phenotype in the lung endothelium during PAH development

PAH is characterized by the dysregulation of proliferative and apoptotic signaling pathways, cellular metabolism, metabolic flux, and mitochondrial function in various cell types, including endothelial cells [79,80]. It is still unclear how these pathways intersect and where they could be therapeutically targeted. However, targeting mitochondrial function is already a promising strategy for treating PAH. Here we provide evidence highlighting endothelial cell mitochondria’s role in promoting the “neoplastic-like” phenotype observed in PAH. Although most of the PAH mitochondrial research has been performed on pulmonary smooth muscle cells [81,82], we propose that endothelial cell mitochondria play a pivotal role in PAH development and could be leveraged for identifying novel druggable targets.

When pulmonary endothelial cells reach confluency, the pathways responsible for contact inhibition of proliferation are activated, and the cells become quiescent to maintain the endothelial barrier. However, in PAH, endothelial cells within plexiform lesions exhibit a pro-proliferative and anti-apoptotic phenotype and do not maintain a monolayer [83]. This has led to the hypothesis that initial, widespread endothelial apoptosis selects for apoptotic-resistant endothelial cells. The remaining cells proliferate and eventually form the plexiform lesions observed in PAH [84]. Very little is known regarding the initial apoptosis phase, nor is it well understood what leads to the rise of the anti-apoptotic endothelial cells.

During PAH, endothelial cells exhibit a Warburg-like metabolic shift, preferentially performing aerobic glycolysis over oxidative phosphorylation [80,85,86]. Aerobic glycolysis benefits proliferating cells, allowing cells to achieve increased metabolic requirements while rapidly proliferating. This is further supported by PAH being associated with increased PPP activity, replenishing NADPH levels, and supplying other anabolic intermediates. NADPH is a critical cofactor for NOX (NADPH oxidase) enzymes, which generate ROS [87]. Therefore, in addition to helping cells meet their increased metabolic demand, increasing the PPP activity also increases ROS levels. This metabolic shift is influenced by hypoxia, as demonstrated by chronic hypoxic rats having increased G6PD expression and activity, the rate-limiting enzyme of PPP [88]. PPP inhibition decreases NAPDH levels and stimulates vasodilation [89], suggesting that at least part of the PAH phenotype is influenced by hypoxia-influenced metabolism that can be pharmacologically modulated to alleviate symptoms.

RNA sequencing on pulmonary arterial smooth muscle cells in patients with PAH identified the most highly upregulated metabolic gene as aldehyde dehydrogenase family 1 member 3 (ALDH1A3). Although this study focused on smooth muscle cells, increased levels of ALDH1A3 were also observed in some of the endothelial cells from PAH patients [37]. ALDH1A3 has been linked to glycolysis and cancer cell proliferation [90]. Nuclear ALDH1A3 converts acetaldehyde to acetate to ultimately produce acetyl-CoA. Acetyl-CoA can be used to acetylate histones, leading to increased expression of metabolic and proliferation genes [37]. Reducing ALDH1A3 levels in both control and PAH PAEC reduces glycolysis and proliferation rates and enhances apoptosis in lung endothelial cells isolated from PAH patients, suggesting that ALDH1A3 and, by extension, acetyl-CoA levels may be involved in the anti-apoptotic phenotype observed in PAH.

Integral to cellular metabolism is the mitochondrial uncoupling protein 2 (UCP2). UCP2 is an anion carrier protein that uncouples oxidative phosphorylation from ATP production. UCP2 dissipates the proton gradient across the mitochondrial inner membrane by directly transporting anions from the matrix to the intermembrane space, facilitating the reverse transfer of protons to lower the membrane potential, making UCP2 a negative regulator of mitochondrial ATP production and an inhibitor of ROS generation [91,92]. UCP2 is widely expressed in the lungs, where it regulates FA metabolism [93] and protects vascular cells from oxidative stress [94]. PTEN-induced kinase 1 (PINK1)-mediated mitophagy also affects UCP2 and the mitochondrial membrane potential. Under normal conditions, healthy mitochondria repress PINK1 by importing it to the inner mitochondrial membrane and degrading it. However, when the mitochondrial membrane potential decreases, PINK1 is no longer transported and accumulates on the outer mitochondrial membrane. Through its kinase activity, PINK1 recruits the E3 ligase, Parkin, from the cytosol to that impaired mitochondrion. Parkin then ubiquitinates outer mitochondrial membrane proteins to induce autophagic elimination of the ubiquitinated mitochondrion [95].

Endothelial-specific UCP2 knockout mice exposed to intermittent hypoxia exhibit excessive PINK1-mediated mitophagy, decreased mitochondrial biosynthesis, increased endothelial apoptosis, and a more severe PH phenotype than control mice exposed to intermittent hypoxia [96]. Interestingly, extensive endothelial apoptosis is observed after week 1 of intermittent hypoxia in UCP2 knockout mice, but not after 5 weeks, supporting the theory of an early mass-apoptosis event that selects for a subpopulation of endothelial cells that are anti-apoptotic and hyperproliferative [96]. Ablating PINK1 alleviates PH development in the endothelial-specific UCP2 knockout mouse, connecting mitophagy to UCP2 activity [84,96]. These animal data match endothelial cells from PAH patients, which have increased expression of PINK1 and decreased expression of UCP2 [96]. Taken together, it is plausible to speculate that UCP2 is essential in maintaining mitochondrial health, and when UCP2 expression is downregulated, mitophagy levels increase, leading to a mass-endothelial-apoptotic event that contributes to the phenotypic change seen in PAH endothelial cells. Although more evidence is needed, this may provide a novel therapeutic target strategy to prevent PAH development.

7. BMPR2 signaling induces vascular remodeling associated with changes in fatty acid metabolism

The endothelial-to-mesenchymal transition (EndoMT) is a critical pathway involved in PAH pulmonary vascular remodeling [79,97] and has been observed in lung tissue samples from patients with PAH [98]. During EndoMT, the endothelial phenotype is lost and replaced with cells with unspecialized mesenchymal phenotypes [[99], [100], [101]] (Fig. 4). EndoMT is detrimental to the integrity of the endothelial barrier, as cells that undergo EndoMT lose tight gap junctions, dissociate from the basement membrane, and migrate to the medial layer. Additionally, the endothelial cells lose their typical endothelial markers and start to express α-smooth muscle actin and vimentin [102]. Mechanical stress, inflammation, oxidative stress, and hypoxia promote EndoMT through inflammatory molecules such as transforming growth factor β (TGF-β) (Fig. 5A) [79,103,104]. ROS can also stimulate EndoMT by inducing endogenous TGF-β expression and activating latent TGF-β. TGF-β can also stimulate ROS production, generating a positive feedback loop [105]. TGF-β activates TGFβR2, a member of the TGF-β superfamily. Different TGF-β isoforms are involved in PH vascular remodeling [106], including TGF-β1. Specifically, there are high levels of TGF-β1 in the plexiform lesions and endothelial cells of advanced PAH patients [107,108]. We have shown that high levels of TGF-β1 increases vascular remodeling in lamb models of PH. This is associated with increased pulmonary blood flow and pressure, resulting in vascular remodeling and disrupting FAO [109].

Fig. 4.

The Endothelial-to-Mesenchymal Transition. When exposed to noxious stimuli, the endothelial phenotype can be lost and replaced with an unspecialized mesenchymal phenotype. During this transition, referred to as the endothelial-to-mesenchymal transition (EndoMT), endothelial cells lose their endothelial markers and express the smooth muscle proteins: smooth muscle actin and vimentin. This transition disturbs normal mitochondrial bioenergetics and causes the endothelial cells to detach from the basement membrane, perturb the endothelium's integrity, and allow the cells to migrate.

Abbreviations: EndoMT: endothelial-to-mesenchymal transition, ROS: reactive oxygen species.

Fig. 5.

Altered TGF-β and BMPR2 signaling promote EndoMT in PAH. (A) TGF-β signaling increases during PAH while BMPR2 signaling decreases, ultimately activating pathways that induce EndoMT. (B) Shown are the canonical signaling pathways mediated by TGF and BMP signaling.

Abbreviations: EndoMT: Endothelial-Mesenchymal Transition, TGF: transforming growth factor, BMP: Bone-morphogenetic protein, ROS: reactive oxygen species.

Another member of the TGF-β superfamily is BMPR2, which is primarily expressed on the pulmonary endothelium [16] and is constitutively active. BMPR2 is essential to canonical bone morphogenetic protein (BMP) signaling, also termed the Smad-dependent pathway, because it activates the Smad transcription factors. BMP canonical signaling activates Smad1, 5, and 8, while TGF-β/TGFβ2 signaling activates Smad2 and 3 (Fig. 5B). Once activated, the Smad transcription factors associate with a co-Smad (Smad4) and translocate to the nucleus to modulate gene expression. BMPR2 and TGF-β also activate non-canonical signaling pathways such as extracellular signal-regulated kinase (ERK), p38 mitogen-activated protein kinase (MAPK), the Wingless-Int (Wnt) pathway, RhoA, and NOTCH [110,111]. Usually, BMPR2 activates pathways that ultimately inhibit EndoMT [[112], [113], [114]]; therefore, endothelial cells with impaired BMPR2 signaling are more prone to undergo EndoMT [115]. This is supported by BMPR2 deficient rats that spontaneously undergo pulmonary vascular remodeling and have increased expression of inducers of EndoMT [98]. Additionally, BMPR2 administered to the pulmonary vascular endothelium through adenoviral gene delivery to PAH rat models significantly attenuates the severity of the disease. PAH rats receiving the BMPR2 therapy show significantly less right ventricular hypertrophy, less pulmonary vascular resistance, improved cardiac function, and reduced vascular remodeling, highlighting the therapeutic potential for upregulating BMPR2 [16].

Unfortunately, the molecular mechanisms behind the EndoMT are not yet fully understood, but emerging evidence suggests that mitochondrial processes, like FAO, play an extensive role. For example, the deletion of endothelial CPT2 in mice results in higher levels of EndoMT, suggesting that endothelial FAO is a critical regulator of EndoMT [116]. Additional supporting metabolomic analysis using ¹³C-labeled metabolites shows decreased levels of FAO in human pulmonary artery endothelial cells (PAEC) from patients with PAH [117]. EndoMT is also characterized by other metabolic disturbances. For example, when EndoMT is induced in human lung endothelial cells, there is an increase in short-chain acylcarnitines, a decrease in glycolytic and TCA cycle metabolites, decreased CPT1A expression, decreased fatty acid metabolism, and reduced levels of acetyl-CoA. Acetate supplementation directly increases acetyl-CoA levels and inhibits EndoMT, and evidence suggests that FAO directly generates acetyl-CoA that is then used for mitochondrial protein acetylation [118]. Therefore, cells that undergo EndoMT in pathological states exhibit decreased FAO and levels of acetyl-CoA, which can directly alter protein acetylation levels/functions and gene expression.

Modulation of FAO in PAH has been explored as a therapeutic strategy. For instance, administering pioglitazone, an agonist of PPARγ (peroxisome proliferator-activated receptor type gamma), reverses severe PAH and the associated vascular remodeling in rat models of PAH. PPARγ is a master regulator of FAO and overall mitochondrial function [119]. In this model, activation of PPARγ normalizes epigenetic and transcriptional regulation of lipid metabolism and mitochondrial function in the pulmonary vasculature [120], suggesting that regulation of FAO is critical in the etiology of PAH and can be therapeutically targeted. Nonetheless, because PPARγ has extensive roles in other cellular functions unrelated to FAO, translating these results into an effective treatment for PAH patients may be challenging. Multiple targets will need to be modulated to successfully treat PAH, as PAH is a multifactorial and complex disease.

8. BMPR2 signaling affects mitochondrial bioenergetics and contributes to an endothelial phenotypic shift in PAH

As discussed above, decreased BMPR2 signaling is causally implicated in PAH and vascular remodeling. In vitro silencing of BMPR2 in endothelial cells reduces the mitochondrial membrane potential, levels of PGC-1α expression, ATP generation, and induces apoptosis. These findings are recapitulated in endothelial cells from PAH patients containing a mutant BMPR2, suggesting that BMPR2 function preserves mitochondrial function and promotes endothelial survival [121]. Thus, when BMPR2 is mutated, the cells undergo mitochondrial dysfunction and subsequent apoptosis. This may be a potential mechanism for that first apoptotic event that leads to the selection of apoptotic-resistant endothelial cells.

In endothelial BMPR2 knockout mice, levels of PGC-1α are decreased during hypoxia-reoxygenation, compared to their non-transgenic littermates, and the mice develop PH accompanied with mitochondrial dysfunction [121]. This suggests that reduced levels of endothelial BMPR2 contribute to mitochondrial dysfunction and promotes PH development. Importantly, PGC-1α activates PPARγ [122,123], and BMPR2 regulates PPARγ transcription in endothelial cells [124]. This has been confirmed in the plexiform lesions of PAH patients, which exhibit significantly reduced levels of PPARγ [125]. Decreasing PPARγ expression reduces the production of nitric oxide (NO) and induces abnormal endothelial cell growth [126,127]. We have shown that inhibiting PPARγ alters multiple endothelial genes implicated in the development of PAH and that PPARγ inhibition leads to cell proliferation and exacerbates endothelial barrier disruption [128]. Transgenic and heterozygous BMPR2-mutant mice also have significantly higher levels of mitochondrial DNA damage [129], and patients with BMPR2 haploinsufficiency PAH exhibit endothelial metabolic dysfunction [130]. Further, pulmonary endothelial cells from mice with mutated BMPR2 exhibit dysfunctional mitochondria and undergo a metabolic reprogramming that is shifted towards glutamine utilization [131]. This increase in glutamine metabolism allows endothelial cells to maintain their hyperproliferative phenotype [131].

Transcriptomic and metabolomic analyses have revealed that human pulmonary endothelial cells with BMPR2 mutations exhibit increased aerobic glycolysis, upregulation of the PPP, decreased carnitine shuttle activity, and decreased FAO, with significant impairment of the TCA cycle [130], suggesting that endothelial cell metabolism is heavily influenced by mitochondrial activity, which is regulated, in part, by BMPR2. Further evidence for abnormal bioenergetics in PAH comes from proteomic profiles of PAEC isolated from PAH patients. PAH PAEC have increased levels of both isocitrate and cis-aconitate and decreased alanine and glycine levels [132]. Overall, PAH PAEC exhibit abnormal TCA cycle and glutamate metabolism, dysfunctional arginine and nitric oxide (•NO) pathways, and increased oxidative stress [132]. Together, these studies prove that mitochondria are dysfunctional during PAH and that BMPR2 signaling is involved. Moreover, these findings highlight changes in BMPR2 expression in the widespread endothelial apoptotic event. Although many are observational findings, these studies highlight a myriad of abnormal metabolic pathways in PAH and suggest that targeting one pathway may be insufficient. Therefore, more research is needed to provide greater insight into the molecular mechanisms of dysregulated bioenergetics in PAH.

9. Endothelial fatty acid metabolism differs in early versus late stages of PAH

Mice lacking the gene for malonyl–coenzyme A decarboxylase (MCD) are resistant to developing PAH during chronic hypoxia [133]. Mice lacking MCD cannot perform FAO, as MCD is the substrate for fatty acid synthase (FAS) and a regulator for fatty acid transport into the mitochondria for FAO. Since MCD-deficient mice are resistant to the development of PAH, fatty acid metabolism is implicated in PAH development. Therefore, inhibiting FA metabolism in endothelial cells may be a strategy to promote a protective metabolic shift toward glucose oxidation (GO) that increases endothelial cell resistance to hypoxia [133]. However, we have reported that FAO is reduced in the early stages of PAH [[134], [135], [136], [137]] and that stimulating FAO prevents endothelial dysfunction [138,139]. A potential explanation for this seemingly paradoxical finding is that FA metabolism may be needed for normal lung endothelial homeostasis, but in those prone to developing PAH, FA metabolism may contribute to PAH development and progression. Therefore, the cells downregulate FAO to protect themselves during the early stages of the disease. As the disease progresses, the cells rely more on FAO and can no longer downregulate it. Furthermore, the phenotypic shift observed in PAH endothelial cells adds a layer of complexity when interpreting these results, as the endothelial cells from early-stage versus late-stage PAH may be completely different cells with different metabolic requirements and functions. Therefore, other therapeutic targets may exist at varying stages of disease development.

Despite the complexity of FAO in PAH, growing evidence supports FAO inhibition as therapeutically beneficial in late-stage PAH. For example, in pulmonary vessels isolated from MCT‐treated rats, acetylcholine‐induced relaxation is impaired, and inhibition of FAS improves endothelial function. Additionally, FAS inhibition increases the levels of phosphorylated endothelial nitric oxide synthases (eNOS) in the lungs, suggesting improved endothelial function [140] as phosphorylation at Ser1177 activates eNOS to generate nitric oxide (•NO) [141]. Both •NO supplementation and overexpression of eNOS have been shown to attenuate PAH symptoms [142]. This suggests that fatty acid metabolism (both synthesis and degradation) affects •NO production, highlighting that abnormal mitochondrial activity influences multiple cellular pathways during PAH. However, the role of fatty acid metabolism in PAH development is still unresolved, and conclusions should be cautiously made.

10. Mitochondrial ROS and metabolic reprogramming in PAH

Reactive Oxygen and Nitrogen Species (ROS and RNS, respectively) play relevant roles in inflammatory conditions such as those found in PAH. ROS is a broad term that encapsulates oxygen species with an enormous range of reactivity and general chemical nature. ROS play an essential role in physiological and pathological processes [143]. ROS includes radicals with unpaired electrons, such as superoxide (O₂^•-) and hydroxyl radical (•OH), as well as hydrogen peroxide (H₂O₂), a chemically stable molecule with no unpaired electrons. Because H₂O₂ reactivity with macromolecules is relatively low, it can quickly leave the mitochondria, independent of the mitochondrial energy state [102]. On the other hand, RNS can introduce nitrosative and nitrative stress to the cell [144]. RNS include free radicals such as nitric oxide (•NO) and nitrogen dioxide (•NO₂), the potent oxidant, peroxynitrite (ONOO⁻), and nitrite/nitrate (NO₂⁻/NO₃⁻). Enzymes are the primary antioxidant defense, and the activity of these enzymes are regulated by the levels of ambient O₂ and various signaling pathways. Major antioxidant enzymes include superoxide dismutases (SOD), glutathione peroxidases (GPx), and catalase (Fig. 6A). In addition to enzymes, cells utilize non-enzymatic means for antioxidant defense, such as Vitamin C (ascorbate) and vitamin E (tocopherol), both found in the lung endothelium (Fig. 6A). Both Vitamin C and E function by donating electrons to stabilize the electron-acceptor [145]. When there is an imbalance in the production of ROS/RNS and the antioxidant defenses, leading to increased ROS/RNS levels, oxidative stress ensues.

Fig. 6.

The major sources of Reactive Oxygen Species. (A) There are many antioxidant defense mechanisms, including enzymatic and nonenzymatic. (B) A byproduct of the mitochondrial ETC is superoxide (O₂•^-). Under normal conditions, MnSOD converts O₂•^- to hydrogen peroxide (H₂O₂) (C) There are many endogenous sources of ROS, including xanthine oxidase, nitric oxidase, ETC, peroxisomes, and NADPH oxidase.

Abbreviations: ETC: electron transport chain, ROS: Reactive oxygen species, SOD: superoxide dismutase, GPx: glutathione peroxidase, ATP: adenosine triphosphate, NADH: nicotinamide adenine dinucleotide, ER: endoplasmic reticulum.

Mitochondrial redox signaling is pervasive in responses to hypoxia (low oxygen conditions) and is critical for reducing oxidative stress, including extramitochondrial oxidants [146]. Mitochondria not only produce O₂•^- and H₂O₂ in vascular cells, but they also are targets of ROS [147]. Mitochondrial membranes, proteins, and mitochondrial DNA are susceptible to oxidative damage [148,149]; therefore, increased ROS production is linked to mitochondrial dysfunction [150]. Oxidative stress has been shown to damage mitochondrial DNA more extensively and persist longer than the damage caused to the nuclear DNA [148]. Redox signaling, hypoxia, and oxidative/nitrosative stress are extensively involved in PAH. Superoxide is also a byproduct of the ETC. Mitochondria produce O₂•^- and subsequent H₂O₂ mainly through the mitochondrial ETC complexes I and III, and to a less extent complex II, and through the mitochondrial matrix and/or inner membrane–bound dehydrogenases [[151], [152], [153], [154]] (Fig. 6B). Although ROS have important cellular roles, such as signaling, maintenance of vascular tone, and oxygen sensing, excess ROS generation can be detrimental to cellular and organ health [23,155]. In pulmonary endothelial cells, other sources of ROS besides the ETC include NADPH oxidase, uncoupled nitric oxide synthase, xanthine oxidase, and monoamine oxidase (Fig. 6C) [156].

When exposed to hypoxic conditions, cells activate signaling pathways that promote the transcription of adaptive factors, reduce cellular oxygen usage, and decrease overall energy consumption. Paradoxically, despite reduced oxygen availability, mitochondrial ROS production is increased during hypoxia. One hypothesis is that mitochondrial ROS produced during the early stages of hypoxia may serve as an upstream regulator of many cellular responses to hypoxia [143,[157], [158], [159], [160], [161], [162], [163]]. However, despite considerable research on the role of mitochondrial ROS in hypoxic signaling, it is still unclear if mitochondria are solely responsible for this initiation of signaling or if other cellular factors are needed. Initial reports suggested that hypoxia decreased mitochondrial-ROS during. However, these reports relied on dyes that are sensitive to oxidation and lack specificity. Moreover, these dyes have difficulty reaching organelles, which is especially problematic when studying mitochondrial ROS [164]. Better methods to measure mitochondrial-ROS are now commonplace and demonstrate that hypoxia increases mitochondrial ROS, including ratiometric fluorescent protein ROS probes [159,163]. Additionally, during hypoxic conditions, DNA and lipid oxidation products accumulate, suggesting that overall oxidant production is increased during hypoxia [143,165,166].

The primary cellular response to hypoxia is the induction of hypoxia-inducible transcription factors (HIFs), which subsequently promote a series of cellular responses that combat the effects of hypoxia, often referred to as HIF-signaling (Fig. 7). HIFs consist of a stable β-subunit and one of three labile α-subunits (HIF-1α, HIF-2α, and HIF-3α). During hypoxic conditions, the labile HIF-α subunits are stabilized, allowing for gene expression involved in erythropoiesis, glycolysis, angiogenesis, cell cycle, and survival [167]. Because HIF regulates multiple metabolic enzymes, it is considered a critical effector of metabolic shifts observed in PAH [168]. Hypoxia likely induces ROS production from mitochondrial complex III, creating a cytosolic signal stabilizing HIF. Interestingly, the ability of complex III to generate ROS does not depend on its ability to pump protons or perform oxidative phosphorylation [160,169,170]. Global Hif2a deletion induces embryonic lethality and is associated with mitochondrial dysfunction [171] and defective lung and vascular development [172], signifying that HIF2 is critical for normal mitochondrial and pulmonary function. Additionally, we have shown that hypoxic conditions increase levels of HIF-2α expression in endothelial cells, differing from smooth muscle cells that have increased levels of HIF-1α in response to hypoxia [173].

Fig. 7.

Regulation of HIF-2 under normoxic and hypoxic conditions. Under normoxic conditions, HIF-2α is hydroxylated by PHD2. Hydroxylated HIF-2α is then recognized by VHL, which then subsequently promotes ubiquitination of HIF-2α, ultimately leading to its degradation. In hypoxic conditions, the labile HIF-2α is stabilized by ROS and is allowed to interact with HIF-2β. HIF-2α/HIF-2β can then translocate to the nucleus and promote transcription of various hypoxia-signaling genes, promoting angiogenesis and erythropoiesis.

Abbreviations: HIF: hypoxia-inducible factor, PHD2: HIF prolyl hydroxylase domain-2, VHL: von Hippel-Lindau tumor-suppressor protein, ROS; reactive oxygen species, EndoMT: Endothelial-Mesenchymal Transition, TGF: transforming growth factor.

Hypoxia signaling has also been implicated in inducing EndoMT. We specifically have connected increased levels of HIF-2α in PAH endothelial cells to the EndoMT inducer, SNAI. siRNA silencing of HIF-2α decreased SNAI expression in PAH endothelial cells, but silencing HIF-1α does not. Moreover, prolyl hydroxylase domain protein 2 (PHD2, promotes HIF-2α degradation) is downregulated in PAH endothelial cells, which increases levels of HIF-2α. Genetic silencing of the endothelial PHD2 gene in mice induced SNAI expression, EndoMT, and PAH development under normoxic conditions, suggesting that HIF-2α regulates EndoMT and PAH development. In support of this, mice with genetic deletion of endothelial HIF-2α are protected from developing hypoxia-induced PAH. This effect appears to be endothelial and HIF-2α specific, as deletion of the endothelial HIF-1α gene or the smooth muscle HIF-2α gene negligibly affects PH development. Our studies indicate that increased HIF-2α in pulmonary endothelial cells is involved in the development of PAH by upregulating SNAI, inducing EndoMT, and causing vascular remodeling [173]. Others have proposed that hypoxia induces mitochondrial ROS production, stabilizing the labile HIF subunits, such as HIF-2α [143,174,175]. As mentioned earlier, ROS can also stimulate EndoMT by inducing endogenous TGF-β expression and activating latent TGF-β [105]. Furthermore, mitochondrial ROS generation is increased during hypoxic conditions. This suggests that mitochondria and mitochondrial ROS may also be involved in HIF-2α/EndoMT signaling and should be investigated further.

Another critical function of the cellular hypoxic response is to induce the endocytosis of the ATP-generating sodium-potassium pump (Na/K-ATPase) [176]. The Na/K-ATPase is ubiquitously found in cellular membranes and is a significant consumer of oxygen [177]. The endocytosis of the Na/K-ATPase requires both mitochondrial ROS and AMP-activated protein kinase (AMPK). AMPK is a ubiquitously expressed enzyme that facilitates ATP production and suppresses ATP use during low-energy states, shifting the cell metabolism from anabolism to catabolism. During hypoxia, the mitochondria increase ROS production, which subsequently triggers the activation of AMPK, linking ROS with Na/K-ATPase inhibition [143,176].

11. Mitochondrial redistribution of uncoupled eNOS and effects on cellular metabolism

•NO is a highly diffusible gas with many physiological functions relevant to the lung. For example, •NO induces vasodilation, prevents neutrophil adhesion to endothelial cells, regulates apoptosis, and maintains the endothelial barrier. Additionally, •NO directly affects mitochondrial respiration and biogenesis and increases mitochondrial-generated ROS and RNS [178]. •NO is synthesized from l-arginine by nitric oxide synthases (NOS), with endothelial NOS (eNOS) being the endothelial isoform (Fig. 8A) [179]. Because eNOS is responsible for generating •NO in endothelial cells, here we discuss in depth how eNOS is regulated and how dysfunctional eNOS affects mitochondrial function and health.

Fig. 8.

Mechanisms of eNOS uncoupling during PAH. (A) Each monomer of eNOS has a reductase and an oxygenase domain, linked together by a calmodulin-binding region. Electrons flow through the reductase domain to the oxygenase domain of an adjacent eNOS monomer. BH₄ stabilizes the eNOS dimers and the interaction of eNOS with l-arginine. The electrons then flow to oxygen, which reacts with l-arginine to form citrulline and •NO. When BH₄ is oxidized into BH₂, l-arginine no longer associates with eNOS. Therefore, oxygen receives the electron from the heme/iron group and forms O₂•- instead of •NO, uncoupling the electron transfer from l-arginine hydroxylation. (B) eNOS predominantly localizes to the caveolae, interacting with caveolin-1, inhibiting its activity. Recruitment of Hsp90 and calmodulin displaces caveolin-1 and activates eNOS that produces nitric oxide from arginine. Various conditions, such as low arginine levels, can uncouple eNOS activity, where eNOS generates superoxide instead of nitric oxide. Superoxide and nitric oxide can then react and form peroxynitrite, leading to protein tyrosine nitration and altered protein function, further promoting eNOS uncoupling and effectively decreasing nitric oxide levels.

Abbreviations: eNOS; endothelial nitric oxide synthase, Cav-1; caveolin-1, HSP90; heat shock protein 90, BH4; tetrahydrobiopterin; BH₂; dihydrobiopterin; O₂•⁻; superoxide; CaM: calmodulin, FMN: flavin mononucleotide, FAD: flavin adenine dinucleotide, NADPH: nicotinamide adenine dinucleotide phosphate, ADMA: asymmetric dimethyl-l-arginine.

Functional eNOS exists as a dimer and synthesizes •NO in a pulsatile manner. All NOS bind calmodulin, contain heme, and require cofactors such as tetrahydrobiopterin (BH₄) for full enzymatic activity. eNOS activity is significantly increased when intracellular calcium levels rise and can be activated by other stimuli, such as shear stress from blood flow [180]. Other proteins also interact with eNOS to regulate its activity. For example, Heat shock protein 90 (Hsp90) associates with eNOS and acts as an allosteric activator. eNOS predominantly localizes to the caveolae, where it can interact with caveolin-1, inhibiting eNOS activity [181]. It is currently accepted that the recruitment of calmodulin and Hsp90 to eNOS displaces caveolin-1, leading to enzyme activation [182] (Fig. 8B).

Oxidative stress can induce eNOS uncoupling. When eNOS is uncoupled, •NO bioavailability decreases, directly affecting the vasculature. Uncoupling occurs when the flow of electrons from NADPH becomes uncoupled from •NO production such that the ferrous-dioxygen complex dissociates, and O₂•^- is generated from the oxygenase domain instead of •NO [183] (Fig. 8). Of relevance, •NO can interact with O₂•^- to form peroxynitrite (ONOO⁻). Therefore, when eNOS is uncoupled, •NO bioavailability is decreased even more, further affecting the vasculature. This is because uncoupled eNOS produces O₂•^- that can react with •NO produced by functional, coupled eNOS enzymes. Furthermore, peroxynitrite is a powerful oxidizing radical that causes oxidative damage to proteins, lipids, and DNA [184]. It can induce posttranslational oxidative modification in various proteins, ultimately affecting protein function and overall cellular health (Fig. 8) [[185], [186], [187]]. Therefore, eNOS uncoupling increases peroxynitrite formation and decreases •NO generation and bioavailability, impacting multiple pathways and cellular functions, including mitochondrial functions.

Each monomer of an eNOS dimer has a C-terminal reductase domain and an N-terminal oxygenase domain, linked together by a calmodulin-binding peptide. The reductase domain contains binding sites for flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), and nicotinamide adenine dinucleotide phosphate (NADPH). These cofactors allow for the flow of electrons from NADPH to FAD to FMN and then to the oxygenase domain of an adjacent eNOS monomer. The oxygenase domain consists of binding sites for heme (which contains iron), tetrahydrobiopterin (BH₄), and arginine. BH₄ stabilizes the eNOS dimer and increases the interaction of eNOS with l-arginine. Therefore, when eNOS is coupled, the electrons flow from FMN to the heme/iron site and then to oxygen, with l-arginine bound. This then allows for an oxidation reaction between l-arginine and oxygen, hydroxylating arginine into citrulline by displacing arginine’s terminal nitrogen with an oxygen atom. The remaining oxygen and nitrogen atom are then able to form •NO. However, when BH₄ is oxidized into dihydrobiopterin (BH₂), l-arginine no longer associates with eNOS, oxygen receives the electron from the heme/iron group and forms O₂•^- instead of •NO, uncoupling the electron transfer from l-arginine hydroxylation [[188], [189], [190]]. As discussed earlier, mitochondria are highly susceptible to oxidative damage [[148], [149], [150]]. Therefore, uncoupling of eNOS results in less production of •NO and subsequent pathways, and directly harms the mitochondria. Therefore, understanding how eNOS becomes uncoupled and how this affects mitochondrial health is essential to understand vascular diseases where •NO dysregulation is implicated, including PAH.

Uncoupling of eNOS can occur through various mechanisms, including suboptimal levels of the substrate l-arginine or the cofactor tetrahydrobiopterin (BH₄), decreased dimerization, or increased levels of the endogenous NOS inhibitor asymmetric dimethylarginine [191]. Phosphorylation at Threonine 495 increases eNOS uncoupling [191,192] and has been linked to metabolic reprogramming in the lung endothelium [192,193], suggesting that eNOS uncoupling is associated with altered mitochondrial metabolism. The mechanism of eNOS uncoupling is still being explored. However, it is apparent that multiple factors affect the enzymes’ function. For example, it has been shown that the eNOS cofactor, BH₄, is very susceptible to oxidative reactions. In high oxidative environments, BH₄ is oxidized to BH₂, depleting the critical eNOS cofactor and subsequently affecting eNOS activity [194]. Evidence for BH₄ involvement in eNOS uncoupling includes that •NO production by eNOS correlates with the intracellular concentration of BH₄ [195,196]. In addition, BH₄ levels are decreased in many animal models of cardiovascular disease [[197], [198], [199], [200]] and in patients who exhibit endothelial dysfunction [[201], [202], [203]]. However, BH₄ supplementation only partially restores eNOS activity and eNOS-dependent vasodilation in pulmonary models of oxidative stress, suggesting other redox-related mechanisms are involved in eNOS activity. One is the S-glutathionylation of eNOS, which has been shown to uncouple the enzyme [204].

Additionally, l-arginine supplementation has been shown to improve endothelial dysfunction [[205], [206], [207]], suggesting that eNOS activity does affect endothelial function. However, human endothelial cells do not depend on l-arginine uptake from the extracellular space as they can effectively recycle l-citrulline to l-arginine and receive l-arginine from breaking down proteins [208,209]. Importantly, endothelial cells express arginase 2 (Arg2), a mitochondrial enzyme that catalyzes the hydrolysis of arginine. Arg2 can compete for arginine, preventing eNOS from generating •NO [[210], [211], [212], [213]]. Arg2 is translated in the cytosol and then trafficked to the mitochondria; however, Arg2 is active in both the cytosol and the mitochondrion. The physiologic role of Arg2 is poorly understood, but it may play a role in •NO metabolism by inducing a local arginine deficiency in the eNOS vicinity, affecting its ability to produce •NO.

In the lungs of PAH patients, regular eNOS expression is observed. However, arginine levels are inversely related to pulmonary artery pressures, and Arg2 activity is higher in the serum of PAH patients. Furthermore, immunostaining of PAH pulmonary endothelial cells showed high levels of Arg2 expression, and these cells produced lower NO [210]. This suggests that even though eNOS expression remains unchanged, its activity can be affected by a local environment deficiency of arginine. Without arginine, eNOS cannot produce NO and therefore becomes uncoupled, producing superoxide, ultimately increasing both oxidative and nitrosative/nitrative stress. Additionally, whole exome sequencing has revealed a frameshift caveolin-1 mutation in PAH patients with no identifiable mutation in TGF-β signaling [214]. Caveolin-1 is an inhibitor of eNOS; thus, those with this mutation are susceptible to eNOS overactivation and uncoupling. This corroborates the finding that caveolin-1 null mice develop PH [215]. These results highlight that eNOS activity is highly influenced by its local environment and that changes in eNOS regulators or substrates can drastically alter its function.

The administration of monocrotaline (MCT, an inducer of PAH) to rats increases eNOS activity and lung tissue nitrotyrosine levels. In addition, administration of endothelin-1, a vasoconstrictor and pro-proliferative agent, to cultured rat pulmonary endothelial cells results in eNOS activation and induction of EndoMT. When eNOS is inhibited prior to endothelin-1 administration, EndoMT is blocked, suggesting eNOS activity is directly linked to EndoMT [216]. However, eNOS activity is critical for endothelial health, and complete eNOS inhibition would not be a viable therapeutic strategy. For example, eNOS-deficient mice have mild PH under normoxic conditions, and hypoxia exaggerates their pulmonary vasoconstrictive response [217], suggesting that underactive and overactive eNOS can be harmful. However, we have shown that endothelin-1 induces eNOS uncoupling and translocation of eNOS to the mitochondria, altering carnitine metabolism, mitochondrial bioenergetics, and ROS-mediated activation of HIF, cumulatively leading to a glycolytic metabolic shift [192]. Because uncoupled eNOS is solely a pathological phenomenon, targeting the uncoupling mechanism can be a novel therapeutic strategy for PAH.

Further evidence for eNOS involvement in PAH is that elevated levels of asymmetric dimethyl-l-arginine (ADMA) are considered a risk factor for cardiovascular death, including in those with PAH [218]. ADMA is an inhibitor of eNOS, and increased levels of ADMA have been associated with eNOS uncoupling [219]. It has been shown that the activity of one of the key enzymes that produce ADMA, protein arginine N-methyltransferase (PRMT), is redox-sensitive [220]. In addition, DDAH (dimethylarginine dimethylaminohydrolase), an enzyme that degrades ADMA, is also redox-sensitive [221]. Specifically, oxidative stress increases the activity of PRMT and decreases the activity of DDAH, ultimately increasing the amount of ADMA [[220], [221], [222]] and eNOS uncoupling. This is highly suggestive that ROS can increase ADMA levels and subsequently affect eNOS activity.

PAH is a common complication of congenital heart disease (CHD) [223]. We have generated a lamb model of congenital heart disease (CHD) with increased pulmonary blood flow to model PAH induced by CHD. In this model (termed shunt lambs), mitochondrial dysfunction is associated with disrupted carnitine metabolism and increased levels of ADMA. Moreover, we found that ADMA increases carnitine acetyltransferase (CrAT) nitration, decreasing its activity by redistributing eNOS from the plasma membrane to the mitochondria, disrupting overall mitochondrial bioenergetics. l-arginine supplementation prevents the translocation of eNOS to the mitochondria, prevents the nitration-mediated inhibition of CrAT, decreases eNOS uncoupling, and enhances •NO generation [109,224].

HIF is a heterodimer of HIF-1α or -2α with HIF-1β that mediates the responses to low oxygen levels. Both HIF-1α and HIF-2α have been implicated in NO bioavailability and PAH development. For example, HIF-1α has been implicated in mediating the metabolic shift of endothelial cells towards aerobic glycolysis. Specifically, endothelial cells from PAH patients have greater HIF-1α expression and transcriptional activity, regardless of oxygen availability, and is associated with low levels of manganese superoxide dismutase (MnSOD) activity and •NO [225]. HIF-1α is an essential transcriptional regulator that affects the expression of many genes, including genes related to proliferation and carbohydrate metabolism, such as glycolytic enzymes [226]. Significantly, HIF-1α increases the transcription of pyruvate dehydrogenase kinase (PDK), which is an inhibitor of pyruvate dehydrogenase. By inhibiting pyruvate kinase, HIF-1α activation can prevent the conversion of pyruvate to acetyl-CoA, subsequently blocking pyruvate entry into the TCA cycle. This increases ATP generation by increasing anaerobic glycolysis and inhibiting oxidative phosphorylation [226]. In rat models of PAH, the pharmacological administration of dichloroacetate, an inhibitor of PDK, has been shown to improve pulmonary hemodynamics and survival by decreasing aerobic respiration [227].

Furthermore, HIF-1α^+/− [228] and HIF-2α^+/− [229] heterozygote mice are protected against hypoxia-induced PH. Additionally, the deletion of endothelial HIF-2α protects mice from hypoxia remodeling, suggesting that vascular remodeling is directly influenced by endothelial cell activity. Deleting endothelial arginase-1 (a downstream target of HIF-2α) protects against hypoxic remodeling. Chronic hypoxia enhances HIF-2α stability, which increases arginase expression and subsequently dysregulates normal vascular •NO homeostasis by depleting local arginine levels [230]. Although this work did not address the role of mitochondria, it has been shown that hypoxia induces mitochondrial ROS production. This mechanism may stabilize the labile HIF subunits, such as HIF-2α [143]. These results highlight the need to investigate mitochondrial control of HIF signaling further.

Because PAH is highly associated with decreased bioavailability of •NO, substantial work has been done targeting the •NO pathway to restore •NO levels [231,232]. The cGMP pathway is currently targeted for treating PAH. For example, activators of sGC and PKG, as well as inhibitors of PDE-V, have been explored and implemented to treat PH and PAH. However, no cure for PH/PAH remains, and additional therapies are still needed. The lack of effective therapies stems from the fact that the etiology of PAH is exceptionally complex. For example, it is known that endothelium-derived •NO prevents apoptosis of endothelial cells induced by proinflammatory cytokines and pro-atherosclerotic factors, such as ROS [233]. However, a disconnect exists between the decreased bioavailability of •NO observed in PAH patients and the anti-apoptotic phenotype we observe in PAH endothelial cells. Logically, if the bioavailability of •NO is lower, the endothelial cells should be more prone to apoptosis. However, this is not what we observe after the initial stages of PAH development. Therefore, this indicates that we still do not fully understand the mechanism of •NO, eNOS, apoptosis, and endothelial cell function in developing PAH.

12. Conclusion: are mitochondria the missing puzzle piece?

Mitochondria are intimately involved in regulating cellular metabolic processes and cell death; thus, mitochondrial dysfunction can result in disease. However, our understanding of the role of mitochondria in endothelial cells during PAH is still severely lacking. The complexity of the dysregulated pathways in PAH has significantly hindered the search for potential therapeutic targets and strategies. Instead of focusing on individual signaling pathways or cell types, entire biological systems must be investigated to understand better how the lung responds holistically to PAH progression. Additionally, pathways and functions in one cell type or disease may differ in another, further complicating efforts to understand a complex disease like PAH. For example, in PAH, there is evidence that UCP2 is under-expressed, resulting in higher levels of mitophagy that damage endothelial cells [96]. Conversely, during acute lung injury UCP2 is overexpressed, resulting in a decrease in mitophagy, which also damages endothelial cells [[234], [235], [236]]. This apparent dichotomy highlights the complexity of cellular signaling, as both over and under-expression can be detrimental. Therefore, these targets must be modulated to have protective effects when designing therapeutics. This review highlights that mitochondria are intimately related to cell homeostasis, actively responding and contributing to phenotypes seen in cells, including inducing vascular remodeling. Therefore, mitochondria may be the missing puzzle piece we have overlooked as we search for therapeutics that can reverse, rather than slow, PAH progression.

Acknowledgements for images

All images were created using BioRender.com. Fig. 1, Fig. 2A were adapted from “Cellular Respiration”, Fig. 5B was adapted from “TGF-Beta and BMP Signaling Pathway”, Fig. 6B was adapted from “Electron Transport Chain”, Fig. 6C was adapted from “Sources of Reactive Oxygen Species (ROS) with Cell Background”, Fig. 7 was adapted from “HIF Signaling”. All templates are from BioRender.com (2023) and retrieved from https://app.biorender.com/biorender-templates.

Declaration of competing interest

None.

Data availability

No data was used for the research described in the article.