Vascular Metabolism as Driver of Atherosclerosis: Linking Endothelial Metabolism to Inflammation

The endothelium is a crucial regulator of vascular homeostasis by controlling barrier integrity as well acting as an important signal transducer, thereby illustrating that endothelial cells are not inert cells. In the context of atherosclerosis, this barrier function is impaired and endothelial cells become activated, resulting in the upregulation of adhesion molecules, secretion of cytokines and chemokines and internalization of integrins. Finally, this leads to increased vessel permeability, thereby facilitating leukocyte extravasation as well as fostering a pro-inflammatory environment. Additionally, activated endothelial cells can form migrating tip cells and proliferative stalk cells, resulting in the formation of new blood vessels. Emerging evidence has accumulated indicating that cellular metabolism is crucial in fueling these pro-atherosclerotic processes, including neovascularization and inflammation, thereby contributing to plaque progression and altering plaque stability. Therefore, further research is necessary to unravel the complex mechanisms underlying endothelial cell metabolic changes, and exploit this knowledge for finding and developing potential future therapeutic strategies. In this review we discuss the metabolic alterations endothelial cells undergo in the context of inflammation and atherosclerosis and how this relates to changes in endothelial functioning. Finally, we will describe several metabolic targets that are currently being used for therapeutic interventions.


INTRODUCTION
The endothelium is a crucial barrier between blood and tissue and is essential for maintaining vascular homeostasis [1]. The monolayer of endothelial cells (ECs) covering the vascular wall are exposed to several mechanical (stretch, shear stress, pressure) and circulating factors (cytokines, chemokines, humoral agents, chemical factors, lipoproteins) that can all affect their phenotype. Throughout vascular homeostasis ECs are in a quiescent state, characterized by the formation of nitric oxide (NO) by endothelial nitric oxide synthase (eNOS). NO has been considered to be atheroprotective due to its anti-inflammatory role by regulating vasodilatation, inhibiting thrombosis and the adhesion of leukocytes and platelets [2]. However, when ECs are exposed to disturbed or low flow conditions, they exhibit a loss in eNOS activity and an enhanced activated phenotype [3]. This EC activation can result in (1)  immune cells to the site of infection [3][4][5][6]. Furthermore, during the advanced stages in atherosclerosis, the hypoxic regions in atherosclerotic plaques can lead to local production of the pro-angiogenic factor VEGF, resulting in plaque neovascularization [7][8][9]. Overall, these processes can contribute to further progression of atherosclerosis, thereby aggravating clinical outcome.
To sustain these pro-atherogenic processes, a certain amount of energy and biomass is necessary. In the field of cancer biology, rewiring of cellular metabolism has been extensively explored as a way for cancer cells to gain a substantial amount of energy and biomass required for proliferation, invasion and metastasis [10,11]. Cancer cells switch from oxidative phosphorylation to aerobic glycolysis for the production of ATP [12,13]. This metabolic switch is referred to as the Warburg effect.
Interestingly, this rewiring of cellular metabolism has also been observed during atherosclerosis, which has predominantly been described in macrophages [14,15]. Recently, it has been established that vulnerable human atherosclerotic lesions exhibit an enhanced expression of glycolytic markers compared to stable plaques [16,17]. Several landmark studies of the group of Carmeliet have shown that ECs are highly glycolytic, a phenomenon which is also the case under quiescent conditions [18].
Immunometabolism. 2021;3(3):e210020. https://doi.org/10.20900/immunometab20210020 However, to date the role of EC metabolism in atherosclerosis has been studied to a lesser extent. In this review we will discuss the intricate role of EC metabolism in fueling vascular inflammation and atherogenesis.
Lastly, we will address various signaling routes targeting microRNA-124, the mitochondria, the glycolytic enzyme 6-phosphofructo-2kinase/fructose-2,6-biphosphatase 3 (PFKFB3), lipoprotein (a) [Lp(a)] and oxidized phospholipids as potential interventions that target endothelial metabolism. It is important to note that all the metabolic states described in this manuscript are a reflection of the 'activation state' of the cells (i.e., proliferating cells, inflammatory cells) and should therefore be extrapolated into different contexts and disease pathologies with caution.

ALTERED METABOLISM AS A MARKER FOR PLAQUE VULNERABILITY
Although the molecular mechanisms that underlie cellular metabolic changes or rewiring are still being unraveled, the concept of an altered vascular metabolism has already been exploited for years. The tracer 18 [24][25][26][27]. In contrast to other cell types, ECs carry a low mitochondrial content accounting for approximately 2-5% of their cytoplasm, which suggests that in this context, mitochondrial respiration appears not the preferred route for ATP generation in ECs [28].
In the field of cancer biology, it has been established that ECs undergo a metabolic switch towards glycolysis to promote neovascularization, which facilitates tumor growth and metastasis [10,11]. Equivalently, neovascularization is pivotal in atherosclerotic lesions. These newly formed unstable and leaky vessels provide novel routes for the influx of pro-atherogenic lipoproteins, red blood cells, inflammatory cells andmediators, and thereby contribute to plaque instability by forming thincap fibroatheromas that are more prone to rupture [8,9,29]. The formation of these new blood vessels rests on ECs specializing into leading tip cells that extend their filopodia and trailing stalk cells, which support extension of the sprouts by proliferation [30][31][32]. To date, a collection of studies describes the changes in EC metabolism that are essential in driving angiogenesis (neovascularization) and are extensively reviewed elsewhere [26,33,34]. In this review we aim to provide an overview of the candidates that are of interest in the context of atherosclerosis (a schematic overview can be found in Figure 1).

Glycolysis
De Bock et al. demonstrated that in human umbilical venous endothelial cells (HUVECs), as well as arterial, lymphatic, and microvascular ECs glycolysis is the predominant bioenergetic pathway [35]. To investigate the role of glycolysis in ECs, they focused on the glycolytic enzyme 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3). Upon knock-down of PFKFB3, an in-vitro sprouting assay showed a marked decrease in the number and length of sprouts (p < 0.05) [35].

Mitochondrial Respiration
Yetkin-Arik and colleagues demonstrated that silencing the mitochondrial respiration enzyme pyruvate dehydrogenase E1 subunit alpha 1 (PDHA1) in HUVECs, resulted in an increased number of apoptotic tip cells and a decrease in proliferating cells [38]. This data underpins that besides glycolysis, mitochondrial respiration is also of importance in driving angiogenesis. Similarly, blocking pyruvate transport into mitochondria using 2-cyano-3-(1-phenyl-1H-indol-3-yl)-2-propenoic acid (UK5099), targeting the mitochondrial pyruvate carrier, resulted in a 30% reduction in the number of tip cells, indicating that mitochondrial respiration is essential for tip cell survival and EC proliferation. Similar effects were observed by the group of Diebold in HUVECs upon inhibition of the mitochondrial complex III using antimycin A and was attributed to decreases NAD + /NADH ratios [39]. The importance of mitochondrial respiration in angiogenesis is further highlighted by the observation that silencing Pdha1 expression resulted in a 2.3-fold reduction in sprout length (p < 0.05) in in-vitro spheroid assays, followed by a decrease in branching points (p < 0.01) and total sprout length (p < 0.05) in the in-vivo chicken chorioallantoic-membrane photodynamic therapy (CAM-PDT) assay [38].
Furthermore, the group of Lapel reported diminished tubular formation of vasa vasorum ECs (VVECs) upon exposure to the OXPHOS inhibitors rotenone, oligomycin, and FCCP [40]. Collectively, these studies indicate a significant role for mitochondrial respiration in neovascularization. In conclusion, although the expression of glycolysis related markers has been established to be associated with increased plaque vulnerability, the importance of mitochondrial respiration in driving angiogenesis is nowadays becoming increasingly clear. It would be of interest to extrapolate these findings and assess the role of mitochondrial respiration in driving atherosclerosis in order to combat its progression.

Fatty Acid Oxidation
Lastly, the function of mitochondrial fatty acid oxidation (FAO) in angiogenesis was studied by Schoors and colleagues [41]. Carnitine palmitoyltransferase 1 (CPT1) is a rate-limiting enzyme in FAO and is essential for beta oxidation of long chain fatty acids in the mitochondria.
Silencing of the CPT1 isoform CPT1A in HUVECs resulted in a decrease in vessel sprout numbers and length (p < 0.0001), which was due to reduced

KLF2 AND FOXO1; GATEKEEPERS OF EC QUIESCENCE
Since endothelial cells form the inner lining of blood vessels, they are exposed to a force of laminar blood flow and shear stress. Disturbance of this blood flow can lead to disturbed shear stress, resulting in NF-κBinduced hypoxia-inducible factor 1α (HIF1α) transcription, promoting EC proliferation, activation and inflammation [45]. Besides the NF-κB-HIF1α signaling pathway, the AMPK/mTOR/ULK1-axis also have been demonstrated to be induced by shear stress [46]. This axis induces autophagy and thereby modulates vascular smooth muscle cells (VSMCs) phenotype. Similarly to VSMCs, autophagy is also essential in ECs for maintaining alignment [47].

KLF2
In ECs the transcription factor Krüppel-like factor 2 (KLF2) promotes endothelial quiescence by upregulating anti-inflammatory and antithrombotic proteins and by downregulating pro-inflammatory and prothrombotic proteins. Upon exposure to laminar shear stress for 72 h, HUVECs induced KLF2 expression, which was accompanied by decreased glucose uptake and mitochondria per EC compared to static conditions [48]. This reduction in glycolysis was mediated by KLF2-induced downregulation of PFKFB3. This furthermore resulted in increased intracellular hyaluronan (HA) substrate availability and HA synthesis [49].
These results suggest that the KLF2-PFKFB3 axis has an important role in regulating EC metabolism and thereby altering the quiescent or activation state of the endothelium. It is therefore tempting to speculate that at sites

FOXO1
In addition to KLF2 as a gatekeeper for endothelial quiescence, the transcription factor forkhead box O1 (FOXO1) has been described as a metabolic checkpoint. Similarly to KLF2, FOXO1 is essential in regulating neovascularization [50]. Upon endothelial selective Foxo1 deletion,

YAP-TAZ Signaling
The YAP/TAZ signaling pathway is also of importance in EC quiescence. homeostasis. Emerging evidence is accumulating that the YAP/TAZ pathway is intertwined with cellular metabolism [52][53][54]. The group of Enzo presented that YAP/TAZ, similarly to KLF2 and FOXO1, is regulated by glycolysis in MDA-MB-231 breast cancer cells [55]. In turn activation of the YAP/TAZ pathway in pulmonary arterial ECs has also been shown to modulate the metabolic enzyme glutaminase (GLS1), involved in glutaminolysis and glycolysis [56]. Collectively, these studies imply the existence of a YAP/TAZ metabolism positive feedback loop that could lead to the progression of atherosclerosis.

THE RISE OF MIRNAS
Over the years microRNAs (miRNAs) have been emerging as significant regulators in atherosclerosis with novel functions being discovered regularly. The current status of miRNAs and their therapeutic potential in atherosclerosis have been extensively discussed by Feinburg and Moore [58]. This review aims to focus on miRNAs that are specifically involved in cellular metabolism in the context of atherosclerosis.

Inflammation
Oxidized phospholipids (OxPLs) are also known as Danger Associated Molecular Patterns (DAMPs) that can be carried by lipoprotein(a) and oxLDL, resulting in accumulation in atherosclerotic lesions [59]. Here oxPLs can induce an inflammatory response and thereby aggravate disease progression [60]. Upon exposure of HUVECs to 30 μg/mL oxidized However, further research is necessary to further investigate the intricate connection between miR-93, inflammation and glycolysis.

Pulmonary Arterial Hypertension
Lastly, the effect of anomalous miRNA expression in cardiovascular disease was highlighted in-vivo by Caruso

CELLULAR METABOLISM
Atherosclerosis is a multifactorial process that drives cardiovascular disease, and has been associated with several risk factors, including age.
Especially, in the Western world we are being confronted with a growing aging population, which increases the risk of major adverse cardiovascular events (MACE) [63,64]. To be able to treat this expanding patient population, it is necessary to understand how aging affects the atherosclerotic process.
One of the key hallmarks of ageing is the growing number of cells that turn senescent, thus an increase in cells that are in proliferation arrest. Recently, Sabbatinelli and colleagues extensively reviewed the metabolic rewiring that senescent ECs undergo in order to sustain their activities. This rewiring is characterized by an even higher dependency on glycolysis, the production of ROS, a decrease in nitric oxide (NO) production and induction of proinflammatory processes [65], demonstrating a metabolism-senescenceinflammation axis in aging individuals.
During aging, mitochondrial function declines, thereby contributing to the acceleration of atherosclerosis [66]. Using wild-type mice in a low cholesterol environment, Tyrrell and colleagues demonstrated that, along with an increase in mitochondrial dysfunction, aged mice also exhibit elevated levels of IL-6 within the aorta [66]. Mitochondrial damage-associated molecular patterns expressed by dysfunctional mitochondria activate the TLR-9-MyD88 axis, resulting in the production of pro-inflammatory cytokines including IL-6. In turn IL-6, can further aggravate mitochondrial dysfunction, suggesting a positive feedback loop within the aorta of aging mice. Furthermore, this enhanced mitochondrial dysfunction was also characterized by an increase of mitophagy, which is the degradation of mitochondria by autophagy.
Recently, encouraging evidence is accumulating about the potential of targeting mitochondrial function as strategy for maintaining vascular function [67]. By supplementing the mitochondrial-targeted antioxidant MitoQ in aged (approximately 27 months) mice, it has been demonstrated that targeting mitochondrial fitness reduced the production of mitochondrial derived ROS and restored endothelium-dependent dilation [68]. These promising pre-clinical data were confirmed in a randomized, placebocontrolled, double-blind, crossover design study with healthy older adults Besides changes on a cellular level, aging individuals are also confronted with increased stiffness of the large arteries. There are several mechanisms underlying this arterial stiffness, which have been extensively reviewed by several groups [70][71][72]. Vessel stiffening has been shown to increase endothelial permeability [73]. In tumor vasculature, PFKFB3 has been implicated in vessel destabilization due to VE-cadherin internalization [11]. Inhibiting PFKFB3 led to increased barrier integrity, due to increased expression of VE-cadherin at the membrane. This suggests that increased PFKFB3 activity can lead to vessel destabilization and increased endothelial permeability. Additionally, vessel stiffness and shear stress-mediated EC alignment are also linked. Bovine aortic ECs cultured on hydrogels mimicking older, stiffer vessels form less tight junctions after 24-hour exposure to fluid shear stress compared to ECs cultured on hydrogels mimicking younger vessels [74]. To this end, loss of eNOS activity due to disturbed shear stress, results in lower NO production, an increase in blood pressure and thereby an increased likelihood of vessel wall injury [2], a process that is accelerated in aging individuals and hallmarked by increased blood vessel stiffening.

Trained Immunity
In the context of trained immunity, the influence of inflammation on rewiring metabolism in monocytes and the subsequent sustained inflammatory effects have been well documented [59,75,76].
Accumulating evidence suggests that circulating lipoproteins elicit trained immunity in monocytes [59,77]. The pro-atherogenic lipoprotein oxLDL has been recognized to induce trained immunity in primary monocytes demonstrated by an enhanced secretion of IL-6, TNFα, IL-8 and MCP-1, thereby contributing to the persistent low-grade inflammation observed in atherosclerosis [77]. Recently it has been demonstrated that metabolic reprogramming is required for oxLDL induced trained immunity [78].
Using extracellular flux analysis Keating and colleagues observed an increased ECAR (p < 0.05) in oxLDL trained macrophages, that was accompanied by an upregulation of the glycolytic enzyme PFKFB3 (p < 0.05). Along with this increase in ECAR, there was also an enhanced OCR (p < 0.05). Overall, these results indicate a metabolic switch upon oxLDL induced training. This was further validated by demonstrating that the susceptibility for trained immunity in individuals was associated with genetic variations in glycolytic genes, including PFKFB3, PFKP and HK1.

Endothelial Cells
Increasing evidence suggests that a similar metabolism-inflammationaxis may also exist in endothelial cells [4,11,57]. As previously described,

A Novel Approach; Directly Targeting Vascular Metabolism
Altered endothelial metabolism is inextricably linked to atherosclerosis, especially PFKFB3 has been illustrated as a key regulator of glycolysis in ECs, and could therefore be a potential drug target (Table   1). In cancer research, PFKFB3 has already been extensively studied as a Along with restoring vascular homeostasis, 3PO has also been shown to be effective in reducing pathological angiogenesis in ocular and inflammatory models [87]. Previous studies have also shown that neovascularization in atherosclerotic plaques contributes to increased plaque instability [8,9]. Therefore, the observations that 3PO can reduce Plaques from the PFK158 treated group had less incidence of fibrous cap atheroma (p < 0.05), accompanied by a significant reduction in necrotic core area (p < 0.05) and apoptotic cell (TUNEL) staining area (p < 0.005) [16]. Moreover, there was increase in vascular smooth muscle content (p < 0.005). And thickening of the fibrous cap area (p < 0.05). Altogether, these aspects contribute to plaque stability, as indicated by the significant increase in stability index area (p < 0.05) of the PFK158 treated group.
Taken together, pharmacological therapeutic interventions directly or in-directly targeting vascular metabolism appear to be beneficial by increasing plaque stability, diminishing inflammation and reducing neovascularization in in-vitro and/or in-vivo models.as summarized in Table 1.

Metabolism
In parallel to directly targeting vascular metabolism, it would be advantageous to reduce the atherogenic stimuli that induce metabolic reprogramming in ECs in the first place. As described previously, Lp(a) induces vascular glycolysis, thereby initiating a pro-inflammatory endothelial phenotype that facilitates leukocyte extraversion [4]. This

Conclusions
In this review, we discussed the impact of vascular metabolism in atherosclerosis and its progression along with shedding some light on the potential of targeting these altered metabolic pathways. Although there are several treatment options on the market for slowing the progression of atherosclerosis, CVD remains the number one cause of death worldwide and is still increasing, in part due to our rapidly growing aging population [63,64]. Aging as well as the exposure to atherosclerotic stimuli are able to rewire cellular metabolism in the vasculature [2, 4,11,65,66,78]. This metabolic rewiring in ECs results in endothelial activation, consequently inducing neovascularization and creating a pro-inflammatory environment that facilitates leukocyte extravasation [3][4][5][6][7][8][9]. Both processes drive the progression of atherosclerosis and contribute to plaque instability, illustrating the importance of these pathways [8,9,29].

Limitations and Future Perspectives
As stated previously multiple studies showed the beneficial therapeutic effect of targeting altered EC metabolism in several atherosclerosis models (Table 1). However, currently most of these interventions have not entered clinical trials. In order to be able to translate these experimental findings into the clinical arena, new scientific advances in the field of vascular and immunometabolism are warranted. For instance, most in-vitro studies discussed in this review were performed with HUVECs. HUVECs are a preferred endothelial model, since they are easily to retrieve and have a high proliferation rate [90,91]. Additionally, HUVECs can migrate and invade, making them suitable for several angiogenesis and transmigration assays [90,92]. However, HUVECs do not fully recapitulate the vascular bed affected in atherosclerosis [91]. It is therefore important to take this into account when extrapolating the data into the context of their respective disease etiology. Therefore, future studies could take the different disease pathologies into account as well as the tissue of interest and adapt their cell lines accordingly. To illustrate, HAECs could be one of the preferred cell-types when studying atherogenesis [93,94]. Alongside utilizing the appropriate cell lines, the field of vascular metabolism could also benefit from the use of advanced in-vitro models, such as organ-on-a-chip technologies, co-culture systems and human induced pluripotent stem cells. These in-vitro models provide a platform to mimic the complex multifactorial aspects of the vasculature, making the results accessible to translate towards the clinic [95].
Besides exposure to atherogenic stimuli, the role of aging is significant for the elevated dependency on glycolysis, increased mitochondrial dysfunction, ROS production and inflammation as well as the decrease in NO production [65,66]. However, the plasticity of EC metabolism in aging individuals has been discussed to a lesser extent. This generates the Immunometabolism. 2021;3(3):e210020. https://doi.org/10.20900/immunometab20210020 question whether novel therapeutic interventions targeting metabolism can switch EC metabolism to their original state, thereby restoring EC phenotype and consequently vascular homeostasis. Implementing these outstanding questions in future research in the field of vascular metabolism will help move the field forward.
In this review we described various metabolic pathways that can be altered in ECs, where the glycolytic pathway being the one being that has been extensively investigated and therefore mostly discussed. PFKFB3 inhibition have been described in the context of cancer by several landmark studies by the group of Carmeliet [11,26,78]. In the context of atherogenesis, inhibition of PFKFB3 showed promising results in the first in-vitro studies as well as in-vivo studies, demonstrating the therapeutic potential of targeting of vascular metabolism as a therapeutic strategy to combat atherosclerosis [4,16].
However, it is important to realize that-just like any other cellular processes-the adaptation of metabolism is, amongst others, dependent on time, spatial localization, their 'cellular state' (i.e., quiescent, proliferative, activated/inflamed) but also on the available energy supply and demand [96]. This makes extrapolation of the metabolic state of ECs from one disease to another extremely challenging. While some data suggests that EC activation and inflammatory responses precede the observed increase in glycolysis [18], the opposite could also be true for example in diabetic patients. Here, the sustained glucose supply and increased glycolytic flux by itself may also cause EC activation and inflammation [97,98]. Therefore, further unraveling of the metabolic-inflammatory axis in ECs in the proper (patho)physiological context is necessary to provide this and other exciting fields with detailed insight in which metabolic regulators could be targeted to reduce the atherosclerotic burden.

CONFLICT OF INTERESTS
JK has received a research grant from Oxitope Pharma BV. The other authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.

FUNDING
This work was supported by the Netherlands Organization for Scientific Research. JK received a VENI grant from ZonMW (91619098).