Abstract
In the microcirculation, the exchange of nutrients, water, gas, hormones, and waste takes place, and it is divided into the three main sections arterioles, capillaries, and venules. Disturbances in the microcirculation can be measured using surrogate parameters or be visualized either indirectly or directly.
Arginase is a manganese metalloenzyme hydrolyzing L-arginine to urea and L-ornithine. It is located in different cell types, including vascular cells, but also in circulating cells such as red blood cells. A variety of pro-inflammatory factors, as well as interleukins, stimulate increased arginase expression. An increase in arginase activity consequently leads to a consumption of L-arginine needed for nitric oxide (NO) production by endothelial NO synthase. A vast body of evidence convincingly showed that increased arginase activity is associated with endothelial dysfunction in larger vessels of the vascular tree. Of note, arginase also influences the microcirculation. Arginase inhibition leads to an increase in the bioavailability of NO and reduces superoxide levels, resulting in improved endothelial function. Arginase inhibition might, therefore, be a potent treatment strategy in cardiovascular medicine. Recently, red blood cells emerged as an influential player in the development from increased arginase activity to endothelial dysfunction. As red blood cells directly interact with the microcirculation in gas exchange, this could constitute a potential link between arginase activity, endothelial dysfunction and microcirculatory disturbances.
The aim of this review is to summarize recent findings revealing the role of arginase in regulating vascular function with particular emphasis on the microcirculation.
Introduction
The microcirculation
The microcirculation constitutes the perfusion of the smallest vessels. In this part of the circulatory system, the exchange of nutrients, water, gas, hormones, and waste takes place. Arterioles, capillaries, and venules are the three main sections of the microcirculation. A vessel diameter of 100μm defines the transition from arteries to arterioles. In the intima-media, vascular smooth muscle cell-layers get thinner ultimately ending in a monolayer distally [1]. A vital function of the arterioles, rich in vascular smooth muscle cells, is the regulation of blood flow. Arterioles can change from complete lumen closure to a 50% diameter dilatation. Different stimuli regulate this vasomotion. Increased flow dilates arterioles and augmentation of the intra-vasal pressure constricts arterioles. Further, oxygen tension influences the vascular tone [1].
The capillaries are characterized by thin walls which enable the exchange between blood and tissues. Active change of the diameter of around 10μm is not possible. However, capillaries can change the properties of the surface and the number of perfused capillaries influencing vascular resistance to a minor extent [2]. The capillary network consists of an endothelial tube with a basal membrane and pericytes. Many branches of vessels form the capillary network.
When several capillaries merge, they are called postcapillary venules. These vessels usually have a diameter of 30–50μm. Venules have a structure similar to that of capillaries and regulate post-capillary resistance [3]. Erythrocytes change capillary perfusion flow properties dependent on volume status and flow velocity [4].
In addition to regulating systemic pressure, the microcirculation regulates volume status, ensuring gas and nutrients exchange. The hydrostatic, osmotic pressure and macromolecules regulate the fluid transfer. In inflammation, cytokines open fluid pores in postcapillary venules. Furthermore, specific cells, especially leukocytes, actively pass the venules. The leukocyte transmigration is of high pathophysiological relevance in inflammation [5].
Both for clinical and research applications, several methods to evaluate the microcirculation are available. Direct measures visualizing the smallest vessels versus indirect methods without visualization can be differentiated. Clinicians most often use methods indirectly estimating surrogate parameters of malperfusion, including reduced body surface temperature, the capillary refill-time, impaired renal output, and serum lactate concentration [6–8]. Dedicated electrodes measure the balance between oxygen transport and oxygen consumption within a tissue of interest — however, parameters independent of the microcirculation influence all these indirect methods [9]. Indirect microcirculatory measurements include Laser-Doppler flow measurements, which quantify the flow using a signal reflected by erythrocytes. These methods enable only indirect skin measurements without distinguishing between arteriole, capillary and venules [9]. Direct techniques using micro-videoscope techniques emit light on the surface of distinct tissues visualize remissions of the microcirculation and filter reflecting light. Video capillaroscopy on the nail fold illuminates the border between the nail and the skin using light microscopy and transparent oil [10]. Other direct methods include Orthogonal Polarization Spectral Imaging and Sidestream Darkfield Imaging (SDF) (Fig. 1) [11]. Polarization filters eliminate surface reflection enabling to detect only reflecting light by underlying tissue. In SDF light emission, the wavelength of the emitted pulsed light is optimized for red blood cell light depolarization to focus on vessels [9, 12]. The contrast agent are red blood cells (RBC), and the wavelength of the emitted light (530 nm) enables an optimum absorption of both oxygenated and deoxygenated hemoglobin [9, 14]. Using online or offline computer-aided analysis, characterization of the microcirculation is possible. The perfused vessel density (PVD, 10–100μm) and the perfused capillary density (PCD, 10–20μm) can be determined [12]. Microvascular flow in highly prognostic small vessels can be measured [15, 16]. The sublingual microvascular network was evaluated using the SDF microcirculation camera [3, 17–19]. Recently, a new technique evolved called Incident dark-field imaging (IDF). SDF and IDF in conjunction with accelerated image analysis, improved image resolution, and even an investigation of the glycocalyx is possible [20, 21]. It is unclear which microcirculatory parameter has the highest outcome predictability [16, 23]. The characteristics of an ideal microcirculatory monitoring system are discussed elsewhere [24]. The association between local and systemic microcirculation is of importance using surface sites such as sublingual region, bladder, muscle, or subcutaneous tissue.

Representative picture of Sidestream Darkfield Imaging (SDF) of sublingual microcirculation in a healthy individual.
Especially in intensive care patients, the correlation between organ microcirculation alterations and prognosis constitutes a challenge [25, 26]. Still, the advancements in the imaging technology supported the central pathophysiological importance of the microcirculation in the development of (multi-)organ failure [3]. Several studies showed the existence of microcirculatory impairment and confirmed its prognostic relevance in cardiogenic shock and sepsis [16, 17]. Observations include decreased vessel density, changes in the proportions of perfused capillaries, and the microvascular flow [19]. Treatment strategies targeting at improving the microcirculation in critically ill patients could improve outcomes in the critically ill [27].
Impaired microcirculation is associated with an increase of endothelin-1 and catecholamine concentration. These biomarkers increase vascular tone [13]. Therefore, therapeutic strategies tailored to increase the bioavailability of nitric oxide (NO) and hence dilate the vessels could be beneficial [28]. The enzyme arginase is involved in the regulation of NO metabolism and could, therefore, be a target in the treatment of microcirculatory disturbances.
Arginase is a manganese metalloenzyme hydrolyzing L-arginine to urea and L-ornithine. Arginase is present in two isoforms, arginase I and II, that share approximately 60% sequence homology [29]. Both isoforms are expressed throughout the body and arginase I is a cytosolic enzyme mainly localized in the liver. Hepatic arginase I constitutes the majority of the body’s total arginase activity. In the liver, arginase I plays a central role in the urea cycle in the elimination of nitrogen formed during amino acid and nucleotide metabolism. Further, arginase I expression is expressed in extra-hepatic tissues, including endothelial cells, vascular smooth muscle cells, and red blood cells [30].
Arginase II is understood as a mitochondrial enzyme and widely distributed. Arginase II is expressed in the kidney, prostate, gastrointestinal tract, and the vasculature. The enzyme was postulated to be involved in the L-arginine homeostasis, the production of L-ornithine for polyamine, and the proline synthesis necessary for cell proliferation [31]. In the vasculature, both isoforms of arginase are expressed, but vasculature arginase expression might be species-dependent [32].
A variety of pro-inflammatory factors, as well as interleukins, stimulate increased arginase expression [31, 33–36]. Other stimuli for arginase expression are oxidized low-density lipoprotein (oxLDL) [37], glucose [38], thrombin [39], hypoxia [40, 41] and angiotensin II [42]. Reactive oxygen and nitrogen species, including H2O2 [43] and peroxynitrite [44, 45] derived from endothelial NO synthase (eNOS) [38] and nicotinamide-adenine-dinucleotide phosphate (NADPH) oxidase [46] increase arginase expression. Intracellular signalling cascades activated by these factors include protein kinase C/RhoA/Rho kinase (ROCK) pathway [45, 47], mitogen-activated protein kinase [42], tyrosine kinases and cyclic adenosine monophosphate/protein kinase A [48]. Further, several transcription factors regulate arginase expression [36].
Arginase and eNOS utilize L-arginine as their common substrate, which leads to reciprocal interactions. Increased arginase activity leads to the consumption of L-arginine necessary for NO production by eNOS, which results in reduced NO production and ultimately endothelial dysfunction. Experimental models of hypertension [49], atherosclerosis [37], diabetes [38, 50], pulmonary hypertension [51] and aging [52] established a link between arginase activity and endothelial dysfunction.
Further, uncoupling of eNOS, a situation when eNOS produces superoxide instead of NO as a result of substrate and/or co-factor deficiency, is another factor aggravating endothelial dysfunction [53]. Thus, increased arginase activity leads to increased superoxide production due to uncoupling of eNOS besides reduced production of NO. Of note, superoxide production further aggravates NO inactivation. On the contrary, inhibition of arginase activity increases the bioavailability of NO and reduces superoxide levels [38, 52], which results in improved endothelial function.
Further, increased cytosolic arginase II is co-localised with eNOS during hypoxia [40]. Both enzymes share L-arginine as a substrate, which might hint at a mechanism for NO synthesis control. Substrate availability for eNOS might be further reduced by arginase inhibiting L-arginine transport in endothelial cells [40]. Arginase might, therefore, play an essential role in cardiovascular disease.
Implications of arginase and microcirculation in distinct pathologies
Arginase and microcirculation in acute myocardial infarction and reperfusion injury
Usually, ischemic heart disease and myocardial infarction are thought to be diseases of larger epicardial vessels. In acute myocardial infarction (MI), percutaneous coronary intervention and restoring adequate coronary flow – macrocirculation – remains gold standard [54]. Still, after reperfusion, a phenomenon called microvascular obstruction occurs in up to 50% of patients presenting with ST-segment elevation myocardial infarction and is associated with adverse outcomes [55]. There are other comprehensive reviews covering microvascular obstruction in the setting of acute myocardial infarction and cardiovascular disease in general [56–59].
Serum arginase activity is increased in patients with MI and correlates with myocardial necrosis [60, 61]. Compared to normal myocardium, arginase activity in infarcted human myocardial tissue is increased [62]. Various experimental models investigated the relevance of arginase expression following myocardial ischemia/reperfusion (I/R). The expression of arginase in coronary arterial endothelial cells was increased following I/R in a study by Hein et al. [63]. In a mouse model of I/R, endothelial cell arginase I expression was increased [64]. Inflammatory cytokines might regulate myocardial expression of arginase during I/R [64]. After I/R in vitro arginase inhibition prevented impairment of ex vivo endothelium-dependent vasodilatation in coronary arteries [63]. These effects may contribute to the reduction in infarct size obtained with arginase inhibition following myocardial I/R in vivo [65, 66].
Arginase and microcirculation in atherosclerosis
Arginase is of importance also in atherosclerosis as endothelial dysfunction occurs early as a result of reduced bioavailability of NO [67, 68]. Compared to age-matched wild-type mice apolipoprotein E knockout (apoE-/-) mice, fed a cholesterol-rich diet, evidenced increased aortic arginase activity [39, 52]. Endothelial cell contribution is suggested as removal of the endothelium decreased arginase activity [52]. Further, atheromatous lesions of hyperlipidemic rabbits evidenced increased arginase activity [69]. Arginase II is considered to be the predominant isoform of arginase in apoE–/– atherosclerotic mice [39, 70]. Distinct pathways, including the RhoA/ROCK are involved in atherosclerotic arginase activity regulation [39, 71]. Further oxLDL, potent proatherogenic, increases arginase activity [37, 73] via the endothelial lectin-like oxidized low-density lipoprotein scavenger receptor and RhoA/ROCK [72]. The role of arginase in atherosclerosis and possible treatment targets are reviewed elsewhere [32].
The role of microcirculatory disturbances in the vasa vasorum and therefore in the pathogenesis of atherosclerosis has been extensively reviewed elsewhere [74].
Arginase and microcirculation in heart failure
Low NO bioavailability might contribute to worsening heart failure due to increasing levels of plasma asymmetric dimethylarinine (ADMA) [75]. In HF, increased plasma arginase activity might stem from an arginase spill-over of injured tissues, especially the congested liver and damaged myocytes [76]. In an HF rabbit model, inducing decreased left ventricular function by left ventricular pacing the serum arginine concentration was decreased, whereas increased cardiac arginase II expression could be demonstrated [77]. Further, arginase-mediated reduction of NO bioavailability might regulate contractility cardiomyocytes [78]. Therefore, increased cardiomyocyte arginase activity could aggravate HF [79]. In rats, the combined administration of arginase and ADMA decreased stroke volume and cardiac output [80]. On the contrary, distinct arginase inhibitors increased contractility in rat myocytes in a dose-dependent manner [81].
Plasma arginase I levels are elevated in patients with HF, especially among those with severe HF [82]. Toya et al. evaluated the pathophysiological role of arginase in heart failure [83]. In a mouse model of doxorubicin-induced HF, lungs, aortic and liver arginase, expression, and activity were increased. Further, the administration of an arginase inhibitor reversed the ejection fraction decrease induced by doxorubicin. Also, arginase inhibition lowered the systolic blood pressure and recovered the decline of NO concentration in serum, lungs, and aorta induced by doxorubicin [83]. Increased arginase activity might therefore play a role in distinct etiologies of HF [82]. However, the exact mechanisms remain to be elucidated: General tissue hypoxia in HF might increase arginase activity. Further, liver congestion could contribute to increased arginase I.
Endothelial dysfunction is present in patients with HF [84, 85]. Increased ADMA levels, decreasing the bioavailability of NO, hence leading to macro- and microvascular dysfunction could contribute to endothelial dysfunction in HF [86]. Increasing NO bioavailability could, therefore, improve the microvascular perfusion in HF patients. In patients with cardiogenic shock and acute decompensated heart failure NO in a dose-dependent manner improved the microcirculation [16, 28]. Nitroglycerin increased sublingual perfused capillary density at low doses even before changes in systemic hemodynamics became evident. Via a mechanism related to NO production from NOS the administration of an arginase inhibitor improved microvascular function [82]. Therefore, increased arginase could be of importance for the impaired microvascular function in HF via reducing NO bioavailability.
Arginase and microcirculation in hypoxia
Only limited data about the hypoxia/arginase-relationship is available. Hypoxia induces the upregulation of arginase activity, mRNA, and protein levels are supported by in vitro data in human microvascular endothelial cells [87]. Of note, in this cell type, arginase II but not arginase I is expressed. In a chronic intermittent hypoxia rat model, both arginase expression and activity were increased in lung and heart tissues [88]. In a similar model of chronic hypoxia, arginase I protein concentration in the carotid artery was elevated; of note eNOS levels were reduced [89]. The impaired endothelial function could be restored by arginase inhibition. This effect was entirely blocked in presence NOS inhibitor [89]. Therefore, the pharmacological inhibition of arginase activity might constitute a potent new therapeutic option.
Hypoxia was shown to increase arginase activity [90, 91]. Both in healthy volunteers and patients following cardiopulmonary resuscitation (CPR) hypoxia led to an increase in systemic arginase I levels, which could contribute to microvascular dysfunction [92, 93]. Sublingual incubation of nor-NOHA significantly increased capillary perfusion and vessel density in patients after cardiac arrest. The effect was inhibited by the addition of L-NMMA, a NOS inhibitor, which could indicate the involvement of NO formation. In patients after cardiac arrest the microcirculatory dysfunction is present; in turn increased microcirculatory function could be associated with favorable neurological outcome [94, 95]. On the other hand, in therapeutic hypothermia, microcirculatory flow was decreased, possibly due to vasoconstriction and a reduced metabolism [96]. Further, tissue I/R injury, inflammation and reduced NO bioavailability, leukocyte, platelet activation, and coagulation could aggravate microcirculatory disturbances [97, 98].
One of the most important predictors of prognosis after initially successful CPR is the neurologic outcome. A correlation of neuronal-specific enolase (as a measure of neurologic damage) and arginase I concentrations on day one in cardiac arrest patients could implicate involvement of arginase in neuronal microcirculatory flow and hence neuro-prognostication [93].
Arginase and microcirculation in shock
An impaired macrocirculation with low mean arterial pressure in critically ill patients subsequently leads to microcirculatory disturbances [9]. Adequate macrocirculatory perfusion pressure seems essential to ensure sufficient microvascular perfusion. Still, the interaction between the macrocirculation and the microcirculation is complex and normal systemic hemodynamics do not guarantee unimpaired microcirculation. Abnormal microcirculatory parameters in a patient with stable hemodynamics were reported [92, 93]. Other studies in patients following resuscitation, septic shock, cardiogenic shock, and HF with optimized systemic hemodynamics suggested a similar independent role of the microcirculation [17, 99].
The microcirculation has been evaluated in distinct shock etiologies [17, 99]. Septic shock and cardiogenic shock are characterized by high death rates and impaired microvascular perfusion [18, 22]. However, in a randomized study investigating the microcirculation in cardiogenic shock no acute (five minutes stop of IABP support) or chronic (randomized into the IABP versus control group), microcirculatory alterations could be detected [22]. Given real-time reliable bedside assessment of the microcirculation in shock, clinical decision-making could be influenced by this information. Cases in patients on extracorporeal membrane oxygenation have already been published [18, 19]. The German clinical guidelines suggest the investigation of the microcirculation to be useful [23]. Clinical studies are warranted to confirm the clinical use of microcirculatory devices.
In patients with septic shock, plasma arginase activity was shown to be increased [100]. In a study of 44 sepsis patients and 25 controls, arginase activity was associated with neutrophil counts. The authors speculated that an increase in arginase activity in sepsis decreases L-arginine bioavailability [100]. In an ovine model for acute lung injury, both ADMA concentration and arginase activity in the lung were reported [101]. Both ornithine aminotransferase and ornithine decarboxylase, associated with collagen and cell proliferation respectively, were also increased and a potential mediation of lung collagen excess and hence reduced lung function after burn injury through the arginase pathway was proposed [101]. In a rat model of hemorrhagic shock, vascular arginase levels were increased [102]. Further, endothelial function of arterioles isolated from rat skeletal muscles was investigated: In arterioles from hemorrhage rats endothelial dysfunction could be restored by treatment with N-hydroxy-nor-l-arginine, an arginase inhibitor [102].
Further evaluation of arginase in distinct shock etiologies could deepen our understanding of the pathophysiological disturbances occurring during multi-organ failure. Potentially, inhibiting arginase could constitute a potential treatment target in shock.
Arginase and microcirculation and red blood cells
Red blood cells (RBC) transport oxygen and carbon dioxide to tissue and lungs, respectively. Still, recent evidence suggests a role of RBC beyond gas transportation emerged. RBC play a role in vasodilatation and vasoconstriction, altering adenosine triphosphate export and NO bioactivity [103, 104]. RBC take up NO, and form nitrate and methemoglobine through oxyhemoglobin. Through NO/nitrate RBC regulate vasodilatation, and NO was even postulated to be the “third gas” in the respiratory cycle [105]. Although it is a subject of debate how exactly RBC are capable of exporting NO bioactivity, arginase, likely plays an essential role by tightly regulating eNOS-generated NO by hydrolysing the substrate L-arginine [30, 106]. Inhibition of RBC arginase was cardioprotective, an effect which could not be observed after blockade of RBC eNOS or in RBC lacking the eNOS protein [30]. Recently, RBC alterations were shown to contribute to endothelial dysfunction in type 2 diabetes [50, 107–109]. RBC from diabetic patients induced endothelial dysfunction both in rats and humans [50, 107]. Further, RBC from patients or mice with type 2 diabetes reduced cardiac recovery from I/R injury [107]. Besides reducing NO bioactivity, increased RBC arginase activity was proposed to increase reactive oxygen species due to eNOS uncoupling [104, 107]. Recently, reduced eNOS RBC expression could be demonstrated in patients with stable coronary artery disease and endothelial dysfunction [110, 111]. Of note, red blood cells are a significant source of exosomes, vesicles known to exhibit cardioprotective but also pathophysiological effects [112, 113]. These and other potentially clinically relevant pathophysiological alterations of RBC attenuating endothelial dysfunction have been reviewed elsewhere extensively [103, 114]. The term “erythropathy” to describe the contribution of RBC to cardiovascular diseases has been proposed.
Besides vascular actions, diseased RBC might evidence altered deformability. In RBC, the plasma membrane has to adapt to distinct vessel diameter, especially in the capillaries, where the gas exchange takes place [115, 116]. Reduced RBC deformability is described in different diseases including metabolic syndrome and might contribute to impaired microcirculation [117–120]. Further, an impaired RBC deformability has been linked to unfavorable rheology and arterial stiffness in obese patients [121, 122]. Therefore, RBC could play a central role in the development of microcirculatory endothelial dysfunction due to increased arginase activity (Fig. 2). We speculate that these changes might have a particular impact on the microcirculation; however, this needs to be confirmed in translational studies. We speculate about an interplay between endothelial cells, RBCs and arginase and NOS (Fig. 3). Therefore, further research investigating the relationship between “erythropathy” aggravated by increased arginase activity and impaired microcirculation is warranted.

RBCs could play a central role in the development of microcirculatory endothelial dysfunction: Increased arginase activity reduces NO bioactivity and increases superoxides, impaired RBC deformability and RBC dysfunction further aggravate microcirculation. NOS – nitric oxide synthase; RBC - Red blood cells.

We speculate about an interplay between endothelial cells, RBCs and arginase and NOS: Increased arginase activity reduces NO bioactivity through L-arginine availability and increases reactive oxygen species due to eNOS uncoupling. eNOS – endothelial nitric oxide synthase; RBC - Red blood cells.
Increased arginase activity is associated with atherosclerosis, hypoxia, and inflammation. Arginase, however, is not just an innocent bystander, but it directly contributes to endothelial dysfunction. Specifically inhibiting arginase activity could constitute a potent target in cardiovascular medicine. Recently, the RBC emerged as a new potential linkage between diabetic endothelial dysfunction, atherosclerosis, and adverse vascular outcomes. RBC might be a relevant factor aggravating vascular diseases through multiple pathways but including an increase in arginase activity and alterations in shapeshifting. As RBC need to pass the capillaries, they might constitute a link between increased arginase activity and impaired microcirculatory states.
Developments in imaging will enable us to visualize and understand the microcirculation better. Especially in the critically ill, disturbances in the microcirculation are well-established predictors of adverse outcome. A specific treatment to improve the microcirculation guided by direct, online, and reliable measurement could help to improve patient outcomes.
Conflict of interest
None declared.
Footnotes
Acknowledgments
Parts of this review have been published as part of the PhD thesis “Role of arginase in vascular function” by Christian Jung. All rights of the thesis are with the author. This work was funded by the German Research Foundation (DFG SFB1116-B06), the Forschungskommission of the Faculty of Medicine of the Heinrich-Heine-University Düsseldorf, Germany, the Swedish Research Council (2016-01284) and the Swedish Heart and Lung Foundation.
