Mitochondrial and Metabolic Adaptations to Exercise-Induced Fluid Shear Stress in Endothelial Cells

Abstract

Recent studies have greatly advanced our understanding of the central role of mitochondria on endothelial function. Here, we propose a hypothesis that unidirectional laminar (pulsatile) flow and disturbed laminar (oscillatory) flow may differentially modulate mitochondrial phenotypes in the context of their bioenergetic, signaling, and biosynthetic functions, providing novel insights into subcellular mechanisms underlying how exercise benefits the improvement of vascular health.

Summary for the table of contents:

Exercise-mediated hemodynamic flow confers mitochondrial remodeling and metabolic reprogramming in the vascular endothelial cell.

INTRODUCTION

The endothelium is one of the largest organs in the human body (1). In adults, the intricate network of arteries, veins, and capillaries, when aligned end-to-end, add up to a truly remarkable length of 90,000 km (or 56,000 miles), the equivalent of more than two times the Earth’s circumference. The endothelium coats the entire luminal surface of the vascular circulatory system, covering the total surface area of 7,000 m² (the size of 1½ soccer fields), and weighing approximately 1 kg. That is comparable to the weight of both the right and left lungs combined (2). The endothelium does not merely provide such a seemingly inert interface between circulating blood and vessel walls, but it also actively participates in several vital processes, including the control of vasomotor tone, the trafficking of bloodborne cells and other circulating substances (i.e. nutrients and hormones), the growth of new blood vessels, and the maintenance of blood fluidity. To achieve these goals, endothelial cells demand a continuous energy supply and require a large quantity of metabolic intermediates derived from mitochondria-centric metabolism.

Until recently, however, the functions of mitochondria within endothelial cells have been neglected, at least to some degree, because of several preconceived ideas: (1) energy supply for endothelial cells is highly dependent on glycolysis even in the presence of sufficient oxygen, which is similar to those seen in cancer cells (so-called “Warburg effect”: the phenomenon of prevalent aerobic glycolysis in cancer cells (3)); (2) endothelial cells survive well under hypoxic conditions; and (3) aerobic respiration rates are relatively low in isolated endothelial cells. However, these observations were primarily made under a cell culture system without a proper hemodynamic flow condition, which was non-physiological.

In fact, endothelial cells contain a relatively high mitochondrial content composing approximately 5–11% of the cell volume. Comparatively, in type 2a and type 2x muscle fibers, mitochondria occupy only 2.3% and 4.5% of the cell volume, respectively. In addition, endothelial cells are highly capable of managing coupled respiration when bioenergetic demand increases, suggesting that endothelial mitochondria have a considerable bioenergetic reserve function (4). Furthermore, recent studies demonstrated that the endothelial mitochondria participate in various non-bioenergetic functions, such as biosynthesis (5) and intracellular cell signaling (6), especially under environmentally challenging conditions. Thus, the concept of maintaining endothelial mitochondrial integrity has emerged as a novel protective mechanism against the development of endothelial dysfunction and vascular incompetence.

This review focuses on consolidating available data to analyze how changes in the magnitude (i.e. low to high) and the pattern (i.e. disturbed laminar (oscillatory) flow to unidirectional laminar (pulsatile) flow) of hemodynamic flow, often occurring during aerobic exercise, serve as important stimuli to confer mitochondrial remodeling and metabolic reprogramming within the endothelial cell. From an anatomical standpoint, the apical surface of an endothelial cell is in direct contact with the bloodstream, whereas its basal surface is tethered by the extracellular matrix, providing structural support through parenchymal cells. At a given “bridgehead” position, this review further focuses on a concept that endothelial cells sense the flow changes and initiate signal transduction pathways, resulting in mitochondrial remodeling along with the activation of mitochondrial quality control machinery.

A large body of evidence indicates that exercise elevates endothelial function and prevents the progression to cardiovascular disease (CVD). Unidirectional laminar shear stress (LSS) is well-known to cause positive effects on endothelium, such as (1) increase vasodilation by nitric oxide (NO) production, (2) inhibition of leukocyte adhesion, (3) less intracellular reactive oxygen species (ROS) production, (4) less thrombogenic, more fibrinolysis, (5) less permeability, (6) less inflammatory signals (such as monocyte chemoattractant protein 1 (MCP1) and interleukins), while exact opposite effects of oscillatory shear stress (OSS) on the endothelial cell phenotype have been reported (7). Herein, we assess the body of evidence on the effect of fluid shear stress (or exercise) on mitochondria in the endothelial cells and relate it to the effect of hemodynamic flow on vascular adaptations in response to exercise. For detailed information on the vascular effect of exercise, the reader is referred to an excellent review paper by Green et al. (2017) (8). Therefore, this review focuses on mitochondrial biogenesis, mitochondrial dynamics, mitochondrial autophagy (mitophagy) and mitochondrial metabolism in the context of exercise endothelial cell biology, thereby readers will gain a better insight into subcellular adaptations to exercise in the vascular system.

ENDOTHELIAL CELL BIOLOGY OF EXERCISE: EVOLUTIONARY PERSPECTIVES

All animals have evolved to survive and reproduce, and to do so, a larger body size is a major trend in animal evolution. These animals share common tasks that include the absorption and distribution of nutrients, the distribution of acquired oxygen for cellular respiration, and the disposal of metabolic waste. While simple diffusion is an energetically inexpensive means to passively transport these substances, it is a very slow process and works only over small distances (diffusion path of less than 1 mm). Therefore, the evolutionary enlargement of an organism’s size has mandated the formation of a circulatory system to reduce functional diffusion distance, thereby, being capable of delivering enough nutrients, gases, and metabolic waste products rapidly across the larger body size.

Likewise, from evolutionary perspectives, the emergence of the vascular endothelium is a critical event that allows mammals to achieve rapid, enduring movement. Existing evidence suggests that the circulatory system likely first appeared over 600 million years ago in a precursor lifeform, a protostome-deuterostome ancestor (a tiny flatworm). Initially, the circulatory system was simply a matrix without an endothelial lining, and therefore carried only slow disturbed blood flow, but was sufficient enough to meet their low metabolic demands. Approximately, 100 million years later, ancestral vertebrates evolved into more rapidly moving predators, requiring increased body size and higher metabolic rates. In addition, the transition from aquatic to terrestrial life provided additional selection pressure for more efficient circulatory systems to overcome gravitational force. To survive under these conditions, vertebrates have evolved a second smooth cell lining in the vessel walls, thereby creating the endothelium, now capable of rapid laminar blood flow.

Only vertebrate vessels possess a true endothelial lining, defined as a layer of epithelial cells with basoapical polarity (the apical side facing the lumen), an intercellular junction (expressing tight junction proteins in endothelial cells within the brain and lungs), and are anchored to a basement membrane constituted of connective and elastic tissues. In humans, the endothelium forms a 0.2 – 0.4 μm thick monolayer composed of one trillion (10¹²) endothelial cells, lining the entire surface of the vascular tree. As such, the human vascular circulatory system has evolved into a closed and intricate network of arteries, veins, and capillaries, which can deliver nutrients and oxygen rapidly enough to supply these energy substrates to distant tissues via the resulting laminar blood flow. Thus, the existence of vascular endothelium is a key feature with which the human circulatory system can supply continuous oxygen, nutrients, and hormones for enabling the body to move (exercise), and the system has evolved to maximize the circulatory function via differential adaptations under different hemodynamic flow patterns it encounters.

MITOCHONDRIA IN ENDOTHELIAL CELLS

Endothelial mitochondria play essential roles in maintaining endothelial cell homeostasis by serving as: (1) a bioenergetic reservoir, (2) an intracellular signaling regulator, and (3) a metabolic hub of biosynthesis. In this section, we briefly discuss these functions played by mitochondria within endothelial cells. Readers can gain further information on this topic by reviewing an elegant review paper by Kluge et al. (2013) (9).

Metabolically, as mentioned above, endothelial cells are highly glycolytic, and oxidative phosphorylation (OXPHOS) is not their primary source of ATP. However, endothelial mitochondria are tightly coupled (basal respiratory control ratio (RCR) = 8.9) and possess a substantial bioenergetic reserve (maximal RCR = 23) compared to other tissues. Note that RCR is determined by the ratio of state 3 and state 4 oxygen consumption. For instance, endothelial cells can increase oxygen consumption by three folds when switching from quiescence to vascular sprouting during angiogenesis, and, likewise, mitochondrial uncouplers (reagents that dissociate mitochondrial respiration from mitochondria-dependent ATP synthesis) have been shown to impair neoangiogenic tumor growth and wound healing in vivo. In addition, the mitochondria reserve capacity plays an important role in responding to a surge of oxygen- or nitrogen-derived reactive species (ROS or RNS, respectively) under various conditions such as vascular inflammation (4).

There are multiple cellular sources of ROS in endothelial cells, including mitochondrial electron transport chain (ETC), nicotinamide adenine dinucleotide phosphate oxidase (NADPH oxidase), xanthine oxidase, uncoupled nitric oxide synthase, and cytochrome P450 (10). Mitochondria present a major source of ROS (so called mitochondrial ROS or mtROS) that are generated from the partial oxygen reduction to form superoxide (O₂^•) at complex I and complex III in the ETC. In isolated bovine aortic endothelial cell (BAEC) mitochondria, O₂^• formation was approximately 4-fold greater during state 4 (absence of ADP or no ATP production) compared to state 3 (active ATP synthesis) respiration. The use of ETC complex-specific chemical inhibitors have shown to increase O₂^• formation by inhibiting electron flow through ETC. Enhanced O₂^• formation induced by complex III inhibitor, antimycin, was significantly reduced by mitochondria uncoupler, p-(trifluoromethoxyl)-phenyl-hydrazone. This result suggest that increased proton conductance combined with uncoupled respiration and the resulting low mitochondrial membrane potential (ΔΨm) protects against ROS generation.

The percentage of mitochondrial respiratory complexes attributable to ROS production have been estimated using isolated mitochondria from primarily muscle tissues, and these estimates in endothelial cells in vivo remains uncertain. Nevertheless, it is widely accepted that the major sites of superoxide (O₂^•) generation under physiological conditions are complexes I, II, and III in endothelial cells. At the ETC, unpaired electrons (e^–) react with O₂ to produce superoxide (O₂^•) (O₂ + e^– = O₂^•). Reverse electron transport has been shown to contribute to mtROS formation, which is increased, in part, by defects in coupled mitochondrial respiration. Koziel and Jarmuszkiewicz (2017) (11) reported that human endothelial cell line EA.hy926 cultured under hypoxic condition (1% O₂ tension for 6 days) resulted in an increase in succinate-sustained mtROS formation, primarily through increased reverse electron transport from complex II to complex I. Increased mitochondrial membrane potential (ΔΨm) is another major source of mtROS shown in uncoupling proteins (UCPs) deficient models (12). To this end, mitochondrial biogenesis has been shown to decrease ΔΨm. The inner mitochondrial membrane is inherently leaky to protons, therefore, accounting for approximately 5–50% of basal oxygen consumption rate is due to respiration required to support the proton leak across the mitochondrial inner membrane depending on the cell/tissue types one is studying (for reviews see (13)). Endothelial cells present abundant expression of uncoupling proteins 2 and 3, which control mitochondrial proton leak (14). Therefore, mitochondrial content and net mitochondrial respiration rates within a cell may reflect the intracellular concentration of the free oxygen ([O₂]_IC). Although the precise mechanism is yet to be clear, essentially, the production of O₂^• primarily depends on [O₂]_IC (a substrate for O₂^•) and the unpaired leaky e^–.Also, mitochondrial reserve capacity may be an integral part of O₂^• generation, thereby cellular oxidant (both ROS and RNS) production.

Beyond the obvious function in bioenergetics, mitochondria in endothelial cells have emerged for their role in signaling events. For instance, under a high oxidant condition, nitric oxide (NO) facilitated the production of mtROS that activates the AMP-activated protein kinase (AMPK) independent of changes in nucleotide levels (6). Moreover, endothelial mitochondria are involved in regulating intracellular calcium signaling and apoptosis. Endothelial mitochondria synthesize cellular building blocks such as amino acids (i.e., aspartate), and nucleotides, which are critical for cell proliferation (5). Defective bioenergetic/oxidant regulatory, biosynthetic, and signaling functions of endothelial mitochondria have shown to present significant implications for the development of a wide range of cardiovascular diseases and cancer. Readers can further gain information on this topic from a couple of review papers (15). Furthermore, recent efforts in targeting mitochondrial dysfunction for CVD prevention and treatment highlights the importance of future research in developing novel interventions that can improve endothelial mitochondria function leading to enhancing cardiovascular health (16).

WALL SHEAR STRESS, FLOW PATTERNS, AND EXERCISE

The vascular endothelium is the innermost layer of blood vessel walls, supported by connective tissue and elastic tissue, together named tunica intima (New Latin “inner coat”); so that the luminal surface of the endothelium is constantly exposed to circulating blood flow thereby shear stress (or friction force) is applied to the vascular wall as a result of the flow of viscous blood. (Note that the viscosity (μ) of whole blood is 3–4 times higher than pure water.) Briefly, the magnitude of wall shear stress (τ_s) can be determined by the force-displacement per unit of cross-sectional area (F (force) / A (area)), and thus the unit of shear stress is presented as dynes/cm², where 1 dyne = 10⁻⁵ N. Shear stress is positively proportional to blood flow rate (Q) and viscosity (μ), and inversely proportional to the third power of internal radius (R³). The relationship can be explained by the equation of Poiseuille’s law (7):

$${\tau }_s= \frac{4\mu Q}{\pi R^3}$$

Under resting conditions, both venous and arterial sides of the vascular system are exposed to unidirectional laminar flow with wall shear stress at levels of 1–6 dynes/cm² in veins or >15 dynes/cm² in arteries. Some regions (i.e., inner curvature, bifurcation, or vessel openings) are exposed to a different type of flow, swirling disturbed flow, that generates oscillatory wall shear stress (± 4 dynes/cm²). These regions present higher expression of proinflammatory cytokines, cell adhesion molecules, and ultimately atherosclerotic plaque formation.

Studies have shown that exercise significantly elevates the magnitude of wall shear stress and alters flow patterns as summarized in Table 1. During exercise, the magnitude of wall shear stress is increased to higher levels ranging from 15 to 30 dynes/cm² in human arteries (17). Using a functional magnetic resonance imaging (fMRI) technique with a customized stationary bike set inside, several studies by Taylor’s group directly measured real-time exercise hemodynamics. Specifically, even at moderate exercise conditions (i.e. HR = 133 bpm and $\dot{Q}$ = 7.5 L/min), wall shear stress increases up to ~25 dynes/cm² and significantly cut the flow oscillation by near 50% and eliminates adverse hemodynamic conditions including flow recirculation and low wall shear stress in atheroprone regions. In this review, we cited a number of in vitro studies that used a cell culture model with a shear stress apparatus, as human studies are limited due to technical constraints. In those papers, static or OSS represented 0 dynes/cm² or ± 3–5 dynes/cm², respectively, where LSS was commonly set at 10–20 dynes/cm².

Table 1.

Alterations in wall shear stress and flow rate during exercise

Refs	subjects	n	Interventions	Methods	Findings
Taylor et al. 1999 (45)	N/A	N/A	Resting, light & moderate exercise	3D computation modeling	Increased flow rate and wall shear stress; reduced oscillatory shear index (OSI) at AA region
Tang et al. 2006 (17)	Young healthy adults (20–30 y.o.)	5	Resting and exercise condition (50% HRmax	PC-MRI	Elevated blood flow and shear stress magnitude during exercise condition at AA
Cheng et al. 2004 (46)	Children (10–14 y.o.)	10	Resting and exercise condition (150% HRrest)	PC-MRI	Increased blood flow at proximal pulmonary arteries with reduced reverse flow index
Cheng et al. 2003 (47)	Healthy subjects	11	Resting and moderate exercise (150% HRrest)	0.5T PC-MRI	Elevated wall shear stress intensity with reduced blood flow oscillation at AA
Suh et al. 2011 (48)	Patients with abdominal aortic aneurysm	10	Resting and light intensity low-limb exercise (60–70% HRrest)	0.5T PC-MRI	Increased blood flow and shear stress with reduced OSI index during exercise at AA
Schlager et al. 2011 (49)	Young healthy subjects (25 y.o.)	46	Resting and exercise (30 toe raises)	Duplex ultrasound	Elevated peak wall shear stress at femoro-popliteal area

MRI, magnetic resonance imaging; PC-MRI, phase contrast magnetic resonance imaging; AA, abdominal aorta; HR, heart rate; y.o., year old

EFFECT OF FLUID SHEAR STRESS ON MITOCHONDRIAL BIOGENESIS

Mitochondrial biogenesis is a complex process by which a cell increases its intracellular mitochondrial volume. In this process, activation of the genetic machinery in both nuclear and mitochondrial genomes is essential, increasing the expression levels of almost 1,000 mitochondrial proteins as building blocks in a coordinated fashion. Peroxisome-proliferator-activated receptor γ co-activator-1α (PGC-1α) has been identified as a master regulator for this process. Particularly, PGC-1α promotes transactivation of nuclear respiratory factor 1 (Nrf1) and 2 (Nrf2), which subsequently transactivates the genes encoding ETC subunits and mitochondrial transcription factor (transcription factor of mitochondria A or TFAM) which is primarily responsible for mitochondrial DNA (mtDNA) transcription and replication (18).

A summary of literatures regarding the effect of flow pattern on mitochondrial biogenesis is shown in Table 2. We have previously demonstrated that prolonged exposure of LSS to primary human endothelial cells increased mitochondrial biogenesis in a shear dose-dependent manner. Kim et al. (2014) showed that physiological levels of LSS (5–20 dynes/cm², up to 48 hours) induces PGC-1α, Nrf1 and TFAM expression levels and increased mitochondrial volume determined by mitochondria-specific fluorescent molecular probes (i.e., MitoTracker^™ Green FM) and mtDNA copy number, a surrogate marker for mitochondrial content, in human aortic endothelial cells (HAECs) (19). Conversely, Chen et al. (2010) showed that OSS caused by disturbed flow does not elicit the mitochondrial biogenesis seen under LSS, suggesting the mitochondrial biogenesis is occurring in a flow pattern-specific manner (20). Mechanistically, our group and others showed that a NAD⁺-dependent deacetylase sirtuin 1 (SIRT1) is the upstream mechanism for the activation of PGC-1α for the flow-induced mitochondrial biogenesis (20, 21). Kim et al. (2015) further showed that LSS-induced mitochondrial biogenesis was associated with reduced release of proinflammatory (CD62E⁺) / proapoptotic (CD31⁺/CD42a⁻) endothelial microparticles, whereas SIRT1 knockdown resulted in increasing proinflammatory and proapoptotic microparticle production in endothelial cells (21). These data suggest that enhanced LSS, which is what systemically occurs during aerobic exercise in the vessel wall, mitigates endothelial activation by promoting endothelial mitochondrial biogenesis.

Table 2.

Mitochondrial remodeling under different fluid shear stress in endothelial cells

Refs		Shear stress		Cell types	Effects
	Type	dynes/cm2	Duration
Mitochondrial biogenesis
Kim et al. 2015 (21)	LSS	5–20	36h	HUVEC	Increase mitochondrial content; increase PGC-1α, TFAM, and Sirt1
Kim et al. 2014 (19)	LSS	20	48h	HAEC	Increase mitochondrial content; increase mitochondrial proteins
Chen et al.2010 (20)	PS OSS	12 1±4 (1Hz)	4–16h 4–16h	HUVEC	Increase mitochondrial biogenesis mediated by Sirt1 expression No change in mitochondrial mass with unchanged Sirt1 expression
Mitochondrial dynamics
Wu et al. 2018 (31)	LSS	12	6–12h	HUVEC	Increase mitochondrial fusion; Increase Mfn2, OPA1, protease YME1L; decrease phospho-ser 616-Drp1; increase phospho-ser 637-Drp1
Breton-Romero et al. 2014 (32)	LSS	12	5–30 min (short)	BAEC and HUVEC	Increase mitochondrial fission; Increase Drp1 accumulation on outer mitochondrial membrane
Giedt et al. 2011 (30)	LSS	10	1h	HUVEC	Increase Drp1 oligomer formation; increase phosphor-ser616-Drp1
Mitophagy capacity (Autophagic flux)
Vion et al. 2017 (37)	LSS	20	24h	HUVEC	Enhance intact autophagy process
Li et al. 2015 (35)	PS OSS	23±8 0.02±3 (1Hz)	4h 4h	HAEC	Intact Sirt1-mediated autophagy process Impair autophagic flux; increase mtDNA damage
Liu et al. 2015 (36)	LSS	12, 20	8–24h	HUVEC	promote autophagy; inhibition of Sirt1 blunted LSS-induced autophagy
Mitochondrial functions
Wu et al. 2018 (31)	LSS	12	6–12h	HUVEC	Increase SHDA and COX-IV expression; Increase membrane potential and ATP generation
Kim et al. 2014 (19)	LSS	20	48h	HAEC	Increase mitochondrial oxygen consumption
Kalucka et al. 2018 (50)	LSS	10	24h	HUVEC	Increase CPT1A expression
Kudo et al. 2000 (51)	LSS	10	48h	PAEC	Increase membrane potential by 20%
Wu et al. 2017 (40)	LSS OSS	N.A. N.A.	24–72h 24–72h	HAEC	Enhance intact mitochondrial-dependent ATP production and OCR Reduced mitohondrial-dependent ATP production and OCR
Yamamoto et al. 2018 (25)	LSS	1, 3, 8	250 sec	HPAEC	Increase mitochondrial ATP level
Breton-Romero et al. 2014 (32)	LSS	12	5–30 min (short)	BAEC and HUVEC	Decrease oxygen consumption and enhanced ECAR

HUVEC, human umbilical vein endothelial cell; HAEC, human aortic endothelial cell; HPAEC, human pulmonary aortic endothelial cell; BAEC, bovine aortic endothelial cell; PAEC, porcine aortic endothelial cells; LSS, unidirectional laminar shear stress; OSS, oscillatory disturbed shear stress; PS, pulsatile shear stress; SIRT1, sirtuin 1; OCR, Oxygen consumption rate; ECAR, extracellular acidification rate

Note that the studies cited in Table 2 have used different types of endothelial cells derived from different vessel beds, mainly from large vessels within arteries (i.e., HAECs) and veins (i.e., human umbilical vein endothelial cells (HUVECs)). HAECs and HUVECs are the most studied primarily cultured human endothelial cells. Because these primary cells can change their properties if they grow in vitro for long time, early passaged cells (3–7 passages) are often used for cell experiments. There are a few studies comparing the gene expression profile between these endothelial cell types. For example, at baseline under static flow, expression patterns of both messenger RNA and microRNA are generally consistent across different vascular endothelial cell types. Furthermore, Butcher et al. (2006) compared the gene expression profiles of aortic endothelial cells and aortic valve endothelial cells in response to shear stress and reported a similar expression patterns (22). Recently, Maurya et al. (2021) reported that the temporal transcriptomic responses to LSS (pulsatile) and OSS (oscillatory) are similar between HAECs and HUVECs for endothelial function-related genes and several pathways such as inflammation, redox regulation, and angiogenesis (23). Note that although we consistently observed flow-mediated mitochondrial biogenesis in both HUVECs and HAECs, HAECs tend to require higher and longer LSS to elicit the similar degree of mitochondrial biogenesis compared to HUVEC. This might be attributed to the nature of the flow these cell types originated from.

Conversely, readers should interpret data with caution when endothelial cells from distinct organs and tissues are considered. Chi et al. (2003) evaluated fifty-three cultured human endothelial cell lines from arteries, veins, and microvessels from 14 different tissue sites that include intestine, lung, myocardium, bladder, and skins from various body parts (24). Not only did they find distinct gene expression profiles between those of veins and arteries, but they also linked the expression patterns to functional roles including lipid transport, cell migration, neurogenesis, and tracheal branching. Furthermore, endothelial cells play an integral role in the regional specialization of vascular structure and function, and thus they may require different metabolic rates and present different mitochondrial features, undertaking distinct adaptive processes in response to flow for mitochondrial remodeling. Understanding of the organ-specific metabolic pathways that maintain endothelial normalcy and quiescence is in its early stage. Further studies are warranted to determine different mitochondrial and metabolic features of endothelial cells from different organs and tissues.

In order to assess the effect of LSS on endothelial mitochondrial biogenesis in vivo, we hypothesized that the exercise-mediated increase in wall shear stress would enhance mitochondrial biogenesis in mouse endothelium. Mice were subjected to 5-weeks of voluntary running exercise training. The average running distance was 8.4 ± 0.3 km per day, and the average running duration was 5.5 ± 0.4 hours per day. As shown in Figure 1, two-photon microscopic images of the en face immunostaining of a constitutive mitochondrial protein (porin, green color) as well as western blot analysis with antibody against porin indicated a significant increase in endothelial mitochondrial content in the descending thoracic aorta from exercised mice compared to the sedentary control. Also, mtDNA copy number, a surrogate marker for mitochondrial content, was significantly higher in the exercised endothelium compared to the sedentary control.

Figure 1. Voluntary exercise training increases mitochondrial content in the arterial endothelium.

For the exercise training group, mice were subjected to 5-weeks of voluntary wheel running exercise. During the period of training, the average running distance was 8.4 ± 0.3 km/day, and the average running duration was 5.5 ± 0.4 hour/day. (mean ± SEM). During the first 2 weeks, mean speed and total distance tend to increase, which can be taken as an indicator of cardiovascular as well as muscular adaptations. VW increased the left ventricular weight by 17% (EXTR: 4.1 ± 0.1 mg/g body weight; SED: 3.5 ± 0.1 mg/g body weight, P = 0.009). (A) Immunofluorescence staining was performed on the mouse thoracic aorta for visualizing endothelial mitochondria using anti-VDAC/Porin (mitochondria) and anti-CD31 (endothelial cells) antibodies. The secondary antibodies used for the staining were Alexa Fluor 488-conjugated goat anti-mouse (green) and Cy3-conjugated goat anti-hamster (red). Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI). For the en face imaging, stained tissues were splay opened and mounted, and images of the endothelial monolayer were acquired using Leica TCS SP5 II two-photon laser scanning microscope. (B) Quantification of fluorescence intensity for porin, a mitochondria-specific protein marker (green). (C) Representative Western blot images for mitochondrial protein marker (porin) in the vessel walls. The bar chart is the result of the densitometric analysis performed using Image J. n=5 (D) Mitochondrial DNA (mtDNA) copy number assay in the mouse thoracic aorta. mtDNA content is expressed as a function of total genomic DNA (nDNA) copy number. mtDNA and nDNA were concurrently isolated from aorta tissues. Specific primer sequences used for the qPCR are as follows: NDII (as a mtDNA encoded gene), F: 5’-CCTATCACCCTTGCCATCAT-3’; R: 5’- GAGGCTGTTGCTTGTGTGAC-3’ and 18S rRNA (as a nDNA encoded gene), F: 5’-CTTAGAGGGACAAGTGGCGTTC-3’; R: 5’-CGCTGAGCCAGTCAGTGTAG-3’. The ratio between mtDNA and nDNA copy numbers were calculated. Data present results from a total of four mice in each group. mean ± SD *P < 0.05, **P < 0.01. (Park JY and Kim B, Unpublished data, 2014.)

Then, why are more mitochondria needed with LSS? As mentioned above, endothelial cells rely on glycolysis rather than on oxidative phosphorylation (OXPHOS) for ATP production. Moreover, previous studies demonstrate that exposure of endothelial cells to LSS is likely to promote a more quiescent state, resulting in reduced cell proliferation and suppression of the glycolytic activator that controls the formation of filopodia/lamellipodia and directional migration. These findings are seemingly contradictory to the current observations. It is possible that there is a metabolic shift from glycolysis to more mitochondria-driven ATP production under a high-magnitude LSS condition. To this end, we have shown that LSS-preconditioned endothelial cells decrease lactate production and increase oxygen consumption rate (19). Moreover, Yamamoto et al. (2018) recently reported that LSS (pulsatile) increased mitochondrial ATP production through reducing plasma membrane cholesterol, which triggers the purinergic receptor-mediated Ca²⁺ -dependent mechanotransduction pathway, suggesting enhanced mitochondrial signaling function (25, 26). We also hypothesize that LSS may facilitate the biosynthetic function of mitochondria for amino acids and nucleotides to facilitate the control of cell biomass and mtDNA replication and repair. Future research is needed to determine the specific mechanisms and the pathophysiological implications of these adaptive responses.

EFFECT OF FLUID SHEAR STRESS ON MITOCHONDRIAL DYNAMICS

Mitochondria constantly change their size, shape, and network, depending on the tissue and the local energy state in which they reside. Mitochondrial fission and fusion events, referred to as mitochondrial dynamics, are essential adaptive processes responsible for this phenomenon. Considering double membranous organelles, mitochondrial fusion is not a simple merging event of two daughter mitochondria. Instead, there are specific and distinct mechanisms accounting for each outer mitochondrial membrane (OMM) and inner mitochondria membrane (IMM) fusion. Large dynamin-related GTPase proteins mitofusin 1 (MFN1) and 2 (MFN2) mediate outer mitochondria membrane fusion whereas optic atrophy 1 (OPA1) plays a critical role in the merging of inner mitochondria membranes. Chen et al. (2005) (27) demonstrated that MFNs or OPA1 knockdown cells display fragmented mitochondria with defective mitochondrial respiratory function and cell senescence. On the other hand, mitochondrial fission is primarily driven by a mitochondrial fission protein, dynamin-related protein 1 (Drp1). Upon stimulation, cytosolic dynamin-related protein 1 (Drp1) is attracted to the OMM and bound to the Drp1 receptor protein complex consisting of mitochondrial fission 1 protein (FIS1), mitochondrial fission factor (Mff), and mitochondrial dynamics proteins of 49 (MID49) and 51 kDa (MID51). These fission processes ultimately result in the formation of a Drp1 oligomeric ring around the constricting site, which brings about mitochondrial fission.

Recent advancements in the understanding of mitochondrial fusion/fission dynamics shed light on mitochondrial quality control mechanisms. Successful mitochondrial dynamic events implicate a mix of contents (i.e., intact mtDNA and damaged mtDNA), segregation and removal of the impaired mitochondrial contents, typically followed by mitochondrial autophagy (mitophagy). Indeed, impairment of the mitochondrial fusion/fission events have been associated with various diseases such as diabetes mellitus and hypertension. For example, upregulation of Drp1 expression increased mitochondrial ROS production, apoptosis and endothelial nitric oxide synthase (eNOS) uncoupling in endothelial cells. Lugus et al. (2011) reported that knockdown of MFN decreased VEGF-stimulated endothelial cell migration and differentiation into angiogenic tubes, diminished endothelial cell viability, and resulted in blunting of the gene expression of components of ETC subunits (28). We also previously reported that excess mitochondrial elongation induced by genetic and chemical inhibition of Drp-1 prohibited normal myogenic differentiation (29). Together, previous studies have shown that imbalance between fusion/fission events are associated with a defective mitochondrial quality control mechanism such as mitophagy, leading to pathological conditions in various cell types.

Recently, several studies documented the effect of LSS on mitochondrial fusion/fission activities. Giedt et al. (2012) reported that the exposure of LSS (10 dynes/cm²) for one hour showed a similar mitochondrial network that was observed in a static flow condition, although the Drp1 activation state determined by phosphorylation at Ser616 and Drp-1 oligomer formation were elevated (30). It was found that mitochondrial depolarization (for example, due to earlier cell exposure to hypoxia) was necessary for steady LSS to cause mitochondrial fragmentation. These results suggested that there would be a concurrent increase in mitochondrial fusion activity that may result in enhanced balance between fusion/fission activities, and thus no morphological changes. Wu et al. (2018) (31) confirmed this concept that prolonged exposure of LSS (12 dynes/cm², 6–12 hours) elicited an increase in mitochondrial network (fusion) by enhancing Mfn2 expression and decreasing Drp1 expression in HUVECs. We also reported that LSS increased both mitochondrial fusion (Mfn1, Mfn2, and OPA1) and mitochondrial fission (Drp1 and Fis1) proteins (19). Interestingly, Breton-Romero et al. (2014) (32) reported that short-term exposure of LSS (12 dynes/cm², 5–15 min) increased the influx of extracellular Ca²⁺, which subsequently caused mitochondrial fragmentation (fission) via calcium-dependent activation of Drp-1. This study also found that transient LSS exposure decreased mitochondrial respiratory function, increased ΔΨm, and promoted the generation of mtROS with the subsequent oxidation and activation of the antioxidant enzyme peroxiredoxins (PRXs) activation (32). These studies suggest that there is a temporal response pattern in the morphological changes under LSS, in which LSS elicits an acute stress-response (i.e., Ca²⁺ influx, mitochondrial fragmentation and mtROS production) that is followed by the enhancement in both mitochondrial fusion/fission activities, which may enhance mitochondrial quality control capacity.

Recently, endothelial mitochondrial morphology in vivo has been reported using a endothelial-specific photo-activatable mitochondria mouse model (EC-PhAM) (33). These studies found that mitochondrial morphology in the endothelium in vivo was associated with a shorter, less interconnected shape than those in primary cultured aortic endothelial cells in vitro. We also recently created EC-PhAM mice and further analyzed mitochondrial morphology in different vessel regions including descending thoracic aorta and lesser curvature of the aortic arch that chronically encounter unidirectional laminar flow and disturbed (oscillatory) flow, respectively. As shown in Figure 2, mitochondria in the thoracic aorta display a more interconnected, elongated shape compared to those in the lesser curvature.

Figure 2. Flow pattern alters the morphometric characteristics of endothelial mitochondria.

Mitochondrial morphology in the endothelium of the descending thoracic aorta (left) and aortic lesser curvature (right) were imaged using an endothelial cell-specific photo-activatable mitochondria mouse model (EC-PhAM mouse). Descending thoracic aorta is a straight region of the vessel where unidirectional flow (pulsatile) is generated. In contrast, the lesser curvature of the aortic arch generates disturbed flow (oscillatory). In order to examine endothelial cell-specific mitochondrial morphology at a tissue level, an EC-PhAM mouse model was generated by breeding mito-Dendra2 (PhAM floxed) and VE-cadherin-Cre mice. Counter staining was performed using anti-VE-cadherin, an endothelial cell-specific membrane protein, antibody to confirm the level of mitochondria signal (mito-dendra 2) at the endothelial level of the en face tissue preparation. Note that endothelium from lesser curvature displays a short-tubule or fragmented shape where those in the thoracic aorta shows more elongated, interconnected mitochondrial shape. (Hong SG, Unpublished data, 2021.)

EFFECT OF FLUID SHEAR STRESS ON MITOPHAGY (AUTOPHAGIC FLUX)

Mitophagy is a selective form of autophagy that functions to remove and recycle amino acids, proteins, and mtDNA within defective mitochondria and is often viewed as the major process in the mitochondrial quality control mechanism. Impairment of mitophagy perturbs mitochondrial function and causes progressive accumulation of defective mitochondria, leading to cell and tissue damage. Indeed, mitophagy impairment has been associated with a number of pathological conditions, such as neurodegenerative diseases, myopathies, metabolic disorders, and cardiovascular disease. The mitophagy processes can be initiated by several different pathways based on cellular conditions via different signaling cascades. One of the best studied mitophagy pathways is the phosphatase and tensin homologue (PTEN)-induced putative kinase 1 (PINK1)/Parkin pathway. Readers can find the details for these regulatory mechanisms in an article by Palikaras et al. (2018) (34). In both pathways, when immobile and damaged mitochondria are sequestered by autophagosomes, the formation of autophagolysosomes facilitates the degradation of the encapsulated mitochondrial contents.

Although mitophagy processes under different flow conditions have, to our best knowledge, yet to be directly investigated, recent studies have implied that LSS enhances autophagic flux, whereas disturbed flow impairs autophagic flux that is caused by a deficiency in autophagy (Table 2). Li et al. (2015) (35) showed that OSS (±3 dynes/cm², 1Hz, 4 hours) decreased autophagic flux determined by a p62 increase that led to the accumulation of damaged mtDNA in HAECs. Liu et al. (2015) (36) also showed that LSS (12 and 20 dynes/cm², 8–24 hours) increases Sirt1-mediated autophagy in endothelial cells. In addition, Vion et al. (2017) (37) showed that LSS (20 dynes/cm², 24 hours) stimulated endothelial autophagic flux in HUVECs and murine arteries. Conversely, low shear stress (2 dynes/cm², 24 hours) was associated with inefficient autophagy as a result of the mammalian target of rapamycin activation, AMP-activated protein kinase α inhibition, and the blockade of the autophagic flux. Furthermore, inhibition of autophagy caused endothelial apoptosis, senescence, and inflammation, suggesting that efficient autophagic flux under high shear stress promotes endothelial cell homeostasis.

METABOLIC ADAPTATIONS TO FLUID SHEAR STRESS

Recent studies have shown that the metabolic shift from mitochondrial-dependent OXPHOS to glycolysis is a critical step for endothelial cell activation, such that inhibition of the metabolic transition is thought to be a method to prevent the progression of vascular CVDs. Despite highly glycolytic cells, emerging evidence has implicated mitochondrial oxidative metabolism as necessary to maintain endothelial cell homeostasis. For instance, the silencing of carnitine palmitoyltransferase 1A (CPT1A), a rate-limiting enzyme of β-oxidation, leads to vascular sprouting defects, and dysfunctional barrier function (5). Thus, mitochondria-dependent metabolism may be crucial for endothelial cell homeostasis under a quiescent state, but also for endothelial cells survive under oxygen-poor conditions by shifting their metabolism to mitochondria-independent glycolysis.

LSS induces a quiescent endothelial phenotype by suppressing glycolytic flux. Kim et al. (2014) (19) showed that LSS (20 dynes/cm², 48 hours) reduced glycolysis-related gene expression including hexokinase 2 (HK2) and 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFKFB), decreased lactate accumulation, and enhanced oxygen utilization compared to the static control. Furthermore, Doddaballapur et al. (2015) (38) reported that athero-protective LSS increased Krϋppel-like factor 2 (KLF-2) expression, which reduced glucose uptake and gene expression of key glycolysis enzymes, such as PFKFB3, phosphofructokinase-1 (PFK1), and HK2, resulting in the quiescence phenotype. Recently, Venturini et al. (2019) (39) showed, using integrated proteomics and metabolomics analysis, that LSS (20 dynes/cm², 24 hours) increased KLF-2 expression and lead to intact lipid metabolism, whereas OSS (5 dynes/cm², 24 hours) downregulated KLF-2 expression and lipid metabolism, especially cholesterol metabolism.

Recent studies demonstrated that disturbed flow promoted endothelial glycolytic flux through hypoxia-inducible factor 1α (HIF-1α) stabilization. These studies demonstrated that OSS upregulated expression of glycolytic enzymes (HK2, glucose transporter-1, and pyruvate dehydrogenase kinase-1), and knockdown of HIF-1α attenuated OSS-induced endothelial glycolytic flux and restored KLF-2 expression under disturbed flow, indicating that OSS exacerbates glycolytic metabolism by HIF-1α-mediated reduction of KLF-2 activation. Wu et al. (2017) (40) showed that knockdown of key glycolytic regulators (siHIF-1α, siPDK1, siNOX4 or siSLC2A1) significantly reduced OSS-induced endothelial cell activation by decreasing inflammatory gene expression levels such as vascular cell adhesion molecule 1 (VCAM1), and chemokine MCP-1 (CCL2), indicating disturbed flow induced endothelial cell activation via enhancing glycolysis. Yang et al. (2018) (41) showed that OSS leads to glycolysis via protein kinase AMP-activated (PRKA)/ AMPKs pathway, and further demonstrated that loss of endothelial PRKAA1, coding for protein kinase AMP-activated catalytic subunit alpha 1, reduced glycolysis, compromised endothelial cell proliferation, and accelerated the formation of atherosclerotic lesions in hyperlipidemic mice. This result indicates that enhanced glycolysis under OSS provides endothelial cells sufficient energy at a rapid rate, which may be necessary for the cells to escape from the stressful condition. A comprehensive summary of literatures regarding the effect of flow pattern on endothelial metabolism is presented in Table 3.

Table 3.

Metabolic adaptation to fluid shear stress in endothelial cells

Refs	Shear stress			Cell types	Effects
	Type	dynes/cm2	Duration
Wu et al. 2017 (40)	LSS OSS	N.A. N.A.	24–72h 24–72h	HAEC	Reduce glycolysis; maintain mitochondrial-dependent respiration Elevate glycolysis via NOX4-mediated HIF-1α stabilization pathway; HIF-1α reduces KLF2.
Feng et al. 2017 (52)	LSS OSS	11–13 4–5	72h 72h	HUVEC	Reduce HIF-1α-mediated glycolysis Elevate glycolysis via cezanne-mediated HIF-1α stabilization.
Doddaballapur et al. 2014 (38)	LSS	20	72h	HUVEC	Reduce glycolysis via KLF2-mediated PFKFB3 suppression
Kim et al. 2014 (19)	LSS	20	48h	HAEC	Reduce glycolysis-related gene expression; reduced lactate accumulation; but enhanced O2 consumption
Venturini et al. 2019 (39)	LSS OSS	15 5	24h 24h	HUVEC	Increase KLF2; upregulate lipid metabolism; high LDLR glycosylation Reduce KLF2; downregulated lipid metabolism; low LDLR glycosylation
Kalucka et al. 2018 (50)	LSS	10	24h	HUVEC	Increase KLF2 and CPT1A; increased fatty acid oxidation flux.
Yang et al. 2018 (41)	LSS OSS	15 ±5 (1 Hz)	24h 24h	HUVEC	Downregulate glycolysis; reduce HIF-1α, slc2a1, PFKFB3 expressions Elevate PRKAA1/AMPKα1-driven glycolysis; enhance HIF-1α, slc2a1, PFKFB3
Breton-Romero et al. 2014 (32)	LSS	12	5–30min	BAEC and HUVEC	Reduce oxygen consumption rate; Increase glycolysis.

HUVEC, human umbilical vein endothelial cell; HAEC, human aortic endothelial cell; BAEC, bovine aortic endothelial cell; LSS, unidirectional laminar shear stress; OSS, oscillatory shear stress; KLF2, krüppel-like Factor 2; KLF4, krüppel-like Factor 4. LDLR, low density lipoprotein receptor; PFKFB3, 6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3; NOX4, NADPH oxidase 4; Slc2a1, glucose transporter 1 (GLUT1)

Collectively, LSS seems to induce an endothelial cell quiescent phenotype suppressing glycolytic flux via KLF-2-dependent pathway with intact mitochondria-dependent respiration. Conversely, OSS is likely to enhance glycolysis via HIF-1α-mediated KLF-2 suppression, which seems to act as a protective mechanism under OSS.

FUNCTIONAL IMPLICATIONS FOR THE FLOW-MEDIATED MITOCHONDRIAL AND METABOLIC ADAPTATIONS

The endosymbiotic theory proposes that oxygen consuming mitochondria evolved from aerobic proteobacteria engulfed by primordial anaerobic eukaryotes about 1.5 billion years ago, when the Earth’s ambient oxygen was thought to be merely 1% (42). After existing in hypoxic conditions for a long time, the present ambient oxygen (21% O_{2 atm}) was established rapidly over the past 200 million years, which accounts for only 5% of the estimated Earth’s age (43). This dramatic increase of ambient oxygen likely required organisms to evolutionarily adapt to life in a high-oxygen environment. The detoxifying role of mitochondria against the oxygen toxicity has recently just begun to be understood. Indeed, recent studies suggest that improving mitochondrial function reduces the deleterious effects of oxidative stress associated with aging, atherosclerosis, metabolic and cardiovascular diseases.

It has been demonstrated that LSS not only increases the mass of the mitochondria, but also improves its function. For example, LSS has been shown to improve coupled mitochondrial respiration indicated by increasing both oxygen consumption rate and ATP production. Also, long-term LSS facilitates the maintenance of ΔΨm at a normal physiological range (slightly decreased compared to static flow control) (19). This result is consistent with other reports showing that shear stress increases ΔΨm in the beginning of shear exposure within 5–10 minutes, but gradually decreases at the later phase, suggesting a biphasic temporal response. Studies often showed that hyperpolarization of the mitochondria (>−140 mV) triggers the release of excess superoxide. Also, there is a U-shaped relation between ΔΨm and ROS formation. Therefore, it is plausible that maintaining the normal ΔΨm is an important feature for healthy mitochondria, and LSS (or exercise) promotes mitochondrial homeostasis.

LSS likely enhances the activity of the ETC, establishing a greater electrochemical gradient across the inner mitochondrial membrane that is well coupled with OXPHOS (or ATP synthesis). This orchestrated electron handling mechanism is also beneficial because it blocks electron leakage through the ETC, especially at complex I and III, thereby reducing O₂^• production. Also, we recently observed that LSS dramatically elevates UCP2 levels, which is likely to inhibit the formation of ROS by regulating reentry of H⁺ back from inner-membrane space to the matrix without substrate (ADP) metabolism (Unpublished data). LSS also enhances the expression of mitochondrial antioxidant enzymes including manganese superoxide dismutase 2 (MnSOD), thioredoxin 2 (Trx2), and Peroxiredoxin 3 and 5 (Prx3 and Prx5). Therefore, LSS is likely to improve the cellular redox state, in part, by enhancing coupled respiration, maintaining ΔΨm, and increasing mitochondrial antioxidant enzyme contents all in favor of reducing mitochondrial ROS production.

Distinct from other intracellular organelles presented in eukaryotic cells, mitochondria contain their own genome (mtDNA) in the mitochondrial matrix. mtDNA is a circular, intron-free genome that is 16,569 bp in length. Each mitochondrion contains multiple copies of mtDNA (5–10 copies/cell), and, thus, each mammalian somatic cell contains 10³ to 10⁴ mtDNA copies. mtDNA contains 37 genes that encode 13 ETC proteins (subunits of complexes I, III, IV and V), 22 tRNAs, and 2 rRNAs(12S and 16S). Compared to the nuclear genome, mtDNA is more vulnerable to oxidative damages and mutations mainly due to their proximity to the ETC, a lack of histone-like proteins as well as less efficient DNA repair systems unlike those in the nucleus. We previously showed that prolonged LSS increased mtDNA copy number along with an increase in TFAM expression level in human endothelial cells and murine endothelium (19, 21). More recently, our lab has conducted experiments to investigate whether unidirectional laminar flow and exercise preconditioning protected mtDNA from angiotensin II-induced ROS-mediated mtDNA damage. Interestingly, we found that LSS preconditioning and endurance exercise training significantly alleviated mtDNA damage in a p53-dependent fashion in human endothelial cells and mouse abdominal aorta, respectively (Unpublished data).

CONCLUSIONS AND FUTURE PERSPECTIVES

Traditionally, the role of mitochondria in endothelial cell function has been widely neglected because of the metabolic characteristics of these cell types. Recent studies have made a significant advancement in understanding of not only bioenergetics of the endothelial mitochondria but also their biosynthetic and signaling roles under pathological conditions. In this review, we outlined mitochondrial and metabolic adaptations in endothelial cells in response to different physiological flow patterns that can be altered by exercise as outlined in Figure 3.

Figure 3.

Summary of the mitochondrial and metabolic responses to exercise-induced flow in endothelial cells.

Future studies are warranted: (1) to elucidate a specific mechanotransduction pathway, responsible for the metabolic adaptations to a different flow patterns, (2) to determine the role of mitochondria as mechanosensing organelles and its downstream retrograding signaling mechanisms, (3) to identify the relationship between mitochondrial fission/fusion events with the endothelial cell metabolic shift under different flows, and (4) to understand the effect of flow patterns on mitophagy as well as lysosomal degradation pathways in the context of mitochondrial quality control.

Furthermore, a highly diverse nature of mitochondrial population isolated from metabolically distinct organs and tissues have been recently reported (44). Therefore, endothelial cell mitochondria-specific parameters including mitochondrial content marker, metabolic profiles, and relative ETC supercomplex composition should be further examined to allow more accurate assessment of mitochondrial biogenic processes, metabolic shift, and mitochondrial quality control. Accumulating knowledge in these areas would allow better understanding about the molecular basis of exercise supporting the concept of “Exercise is the mitochondrial medicine” within the vascular system.

Key Points.

• Endurance aerobic exercise alters the magnitude and pattern of hemodynamic flow.

• Exercise-induced high magnitude unidirectional flow promotes mitochondrial biogenesis in vascular endothelium.

• Unidirectional laminar shear stress enhances mitochondrial network, mitochondrial dynamic events, and mitophagy in endothelial cells.

• Unidirectional laminar shear stress decreases glycolysis through the Krϋppel-like factor 2-mediated suppression of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 and may promote mitochondrial lipid metabolism in endothelial cells.

• Oscillatory shear stress increases glycolysis via a hypoxia-inducible factor-1 dependent mechanism, which is coupled with endothelial cell activation.