Oxidative loss of tetrahydrobiopterin (BH₄) as a trigger for eNOS uncoupling is the best characterized “redox switch” in eNOS and meanwhile well documented in hypertension, diabetes, and atherosclerosis ^{[8][9][10][11][12][13]}[33,114,115,116,117,118]. Reviews on the role of BH₄ deficiency in almost all cardiovascular diseases provide detailed insights on the mechanisms ^{[14][15][16][17][18]}[23,119,120,121,122]. The enzymatic source for BH₄ synthesis, GTP-cyclohydrolase-1 (GCH-1), was identified as an important regulator of eNOS and endothelial function ^[19][123]. The stoichiometry between eNOS and GCH-1 expression controls endothelial function and eNOS overexpression without matched increase in BH₄ levels will ultimately result in eNOS uncoupling ^[12][117]. In accordance with the concept of direct oxidative depletion of BH₄ by peroxynitrite, an oxidant-driven proteasomal degradation of the GCH-1 has been demonstrated via peroxynitrite- or superoxide-mediated activation of the proteasome 26S ^[20][21][22][124,125,126]. Recently, Chuaiphichai et al. have demonstrated that genetic endothelial-specific GCH-1 deficiency in Gch1^fl/flTie2cre mice causes eNOS uncoupling (evident by increased endothelial superoxide formation) and an impaired endothelium-dependent relaxation in arterial resistance arteries ^[23][24][127,128]. These data indicate that endothelial GCH-1 is the essential regulator of eNOS functionality and it can prevent uncoupling of eNOS. In a just published work, the same group show that BH₄ deficiency in endothelial and macrophages is required to trigger endothelial dysfunction and the enhancement of atherosclerosis by using endothelial/myeloid-specific GCH-1 and ApoE global knockout (Gch1^fl/flTie2CreApoE^−/−) mice and bone marrow chimeras thereof ^[25][26][129,130]. These data indicate that NOS isoforms in the vasculature (eNOS) and immune cells (iNOS) need tight regulation by BH₄ in order to prevent the progression of atherosclerosis, as already suggested ^[27][131], thereby providing another strong link between vascular function, cardiovascular health, and inflammation. Besides the GCH-1 dependent de novo synthesis of BH₄, the so called “salvage pathway”, is of high physiological importance, consisting of the recycling of oxidized BH₂ back to BH₄ by dihydrofolate reductase (DHFR) ^[28][29][132,133]. Also, DHFR is subject to 26S proteasome-dependent degradation, a process that was prevented by S-nitros(yl)ation of DHFR by eNOS-derived NO ^[30][134]. Therefore, the BH₄ regulatory system itself provides multiple pharmacological targets for therapeutic prevention of endothelial dysfunction and the progression of cardiovascular disease^[31][135]. As suggested previously, a combination of antioxidant therapy and BH₄ supplementation may be required in order to successfully treat cardiovascular diseases ^[32][40].

Another direct redox-regulatory pathway for eNOS function is the oxidative disruption of the zinc-sulfur-complex (ZnCys₄) in the binding region of the eNOS dimer, resulting in a loss of sodium dodecyl sulfate (SDS)-resistant eNOS dimers, which has been first described by Zou and coworkers for peroxynitrite-mediated oxidation of eNOS ^[33][136]. The reports on this “redox switch” reflected by a decreased eNOS dimer/monomer ratio were previously summarized until the year 2010 ^[1][14][20,23]. More recent studies reported impaired endothelial function and a decrease in eNOS dimer/monomer ratio by CRP treatment of rats ^[34][137], in old rats with improvement by arginase inhibition ^[35][138], in mild hyperhomocysteinemia in mice with heterozygous gene deletion of methylenetetrahydrofolate reductase (Mthfr^+/−) ^[36][139], and in 8-nitro-cGMP treated mice ^[37][140]. Similar observations were made in caveolin-1 depleted and angiotensin-II stimulated endothelial cells ^[38][141], in diabetic rats with improvement by nicorandil ^[39][142], in diabetic rats with improvement by green tea extract ^[40][143], in hypoxic pigs with improvement by L-citrulline ^[41][144], and in tachycardia/sympathetic over-activation in mice by isoproterenol ^[42][145]. eNOS dimer/monomer ratio was also decreased in diabetic db/db mice and normalized by saxagliptin ^[43][146], in erectile tissue of type 2 diabetes mellitus (T2DM) men ^[44][147], in mice with doxorubicin-induced cardiomyopathy and improvement by folic acid ^[45][148], in spontaneously hypertensive rats ^[46][149], and the improvement of eNOS coupling by increased eNOS dimer/monomer ratio by exercise training ^[47][48][150,151]. The critical role of this zinc-sulfur-complex for proper eNOS dimer formation was demonstrated by significant monomerization in a knock-in mouse expressing a C101A-eNOS mutant with impaired zinc-sulfur-complex forming ability ^[49][152]. Of note, these C101A-eNOS-transgenic mice displayed normal systolic blood pressure despite higher levels of eNOS, whereas mice overexpressing wild type eNOS showed significant hypotension ^[50][153]. According to another hypothesis, the dimer formation is mainly regulated by BH₄ binding in a pocket of the dimer interface and prevention of the BH₄ binding by mutation of the critical tryptophan 447 caused monomerization of the eNOS enzyme [154]. We ^[51]think that both monomerization mechanisms could play a synergistic role and oxidative disruption of zinc-sulfur-complex is a very prominent biological redox reaction, as seen for zinc-finger transcription factors, phorbol ester domain in PKC [111], or the active site of alcohol dehydrogenase [155]. More recent reports also suggest that the dimer/monomer ratio can be also used as a surrogate parameter of neuronal NOS (nNOS) uncoupling [156]. Although eNOS dimer/monomer ratio might not represent a drug-targetable pathomechanism, it might represent a surrogate marker to monitor the successful recoupling of eNOS by pharmacological or non-pharmacological therapy.

S-glutathionylation represents another important “redox switch” in eNOS. Zweier and coworkers showed that eNOS is adversely regulated and uncoupled (leading to superoxide formation) by S-glutathionylation at cysteine residues Cys689 and Cys908 in the reductase domain ^[54][55][157,158]. The reports on this “redox switch” that were reflected by eNOS S-glutathionylation were previously summarized until the year 2011 ^[1][14][56][20,23,43]. More recent in vivo studies reported on an association of eNOS S-glutathionylation with eNOS uncoupling and/or endothelial dysfunction in isosorbide-5-mononitrate treated mice ^[57][159], in hypertensive mice with improvement by inhibition of mitochondrial permeability transition pore opening ^[58][160], in aged mice with potentiation by glutathione peroxidase-1 deficiency ^[59][161], in atherosclerotic mice treated with carbamylated low-density lipoprotein ^[60][162], in mice with lung injury after lipopolysaccharide challenges ^[61][163], in mice with cardiac pressure-overload and improvement by N-acetylcysteine ^[62][164], in rats with T2DM and erectile dysfunction ^[63][165], and in mice with doxorubicin-induced cardiomyopathy and improvement by folic acid ^[45][148]. Also, environmental risk factors, such as traffic noise exposure, were associated with eNOS S-glutathionylation in mice exposed to aircraft noise for several days ^[64][65][166,167]. eNOS S-glutathionylation in mice with pressure overload produced by transverse aortic constriction was normalized by physical exercise ^[66][168]. eNOS S-glutathionylation was also observed in a knock-in mouse expressing a C101A-eNOS mutant with impaired zinc-sulfur-complex forming ability, which was also associated with increased superoxide production and protein tyrosine nitration ^[50][153]. Also, ex vivo experiments yielded similar results and eNOS S-glutathionylation with eNOS uncoupling was observed in human aortic endothelial cells that were treated with ultrafine particles ^[67][169] and in endothelial cells upon hypoxia/reoxygenation (associated with BH₄ depletion) ^[68][170]. This redox-regulatory mechanism gained even more biological relevance when Zweier and colleagues showed that eNOS S-glutathionylation can be reversed by glutaredoxin-1 ^[69][171]. An essential role of glutaredoxin-1 for the control of eNOS S-glutathionylation was also demonstrated in mice with necrotizing enterocolitis ^[70][71][172,173]. eNOS S-glutathionylation in experimental fibrosis is directly connected to glutathione biosynthesis ^[72][174]. eNOS S-glutathionylation is obviously connected to BH₄ deficiency ^[73][175] and also a direct connection between nNOS function and eNOS S-glutathionylation was shown in the heart ^[74][176]. It remains to be established whether eNOS S-glutathionylation represents a drug-targetable pathomechanism. However, it at least represents a surrogate marker to monitor the successful recoupling of eNOS by pharmacological or non-pharmacological therapy.

Regulation of eNOS is also mediated by phosphorylation via redox-sensitive kinases. Whereas, eNOS phosphorylation at Ser1177 via Akt pathway is of activating nature ^[75][177], phosphorylation at Tyr657 mediated by the protein tyrosine kinase-2 (PYK-2) is of inactivating nature ^[76][178]. Also, phosphorylation at Thr495 that is mediated by protein kinase C (PKC) leads to the inactivation and potential uncoupling of eNOS ^[77][78][79][179,180,181], and PKC is also a potent activator of NADPH oxidases NOX-1 and NOX-2. PKC has redox-sensitive thiols in its phorbol ester/diacylglycerol binding site and it can be activated by hydrogen peroxide ^[80][81][182,183]. PYK-2 was reported to be stimulated by authentic hydrogen peroxide as well as by angiotensin-II stimulated ROS formation in cell culture and in vivo ^[76][178]. The reports on this “redox switch” reflected by adverse phosphorylation by redox-sensitive kinases were previously reviewed in detail ^[1][14][56][20,23,43], but some old reports and new studies were not considered and they are highlighted here. Homocysteine obviously causes dysregulation of eNOS via PKC activation and adverse (de)phosphorylation of eNOS in human platelets ^[82][184]. Treatment of human aortic endothelial cells with CRP caused the dephosphorylation of eNOS at Ser1177 and phosphorylation at Thr495 ^[83][185]. In isolated, pre-contracted, endothelium-intact porcine coronary arteries, amlodipine increased eNOS phosphorylation at Ser1177 and it decreased the one at Thr495, which was associated with an improvement of bradykinin-dependent relaxation ^[84][186]. Likewise, the AT1-receptor blocker telmisartan improved the ratio of phosphoThr495/phosphoSer1177 of eNOS in cultured endothelial cells ^[85][187] and nitroglycerin-treated rats ^[86][188]. Simvastatin was reported to increase NO formation, eNOS phosphorylation at Ser1177, to decrease phosphorylation at Thr495, and improve cardiac function in patients undergoing cardiac surgery ^[87][189]. Other therapeutic strategies for improvement of the ratio of phosphoThr495/phosphoSer1177 in eNOS were reported for exercise of mice ^[88][89][190,191], inhibition of the ROCK and ERK kinase pathways by MEKK1/2 inhibitor U0126 and ROCK inhibitor Y27632, thereby preventing the oxidative activation of PKC in human umbilical vein endothelial cells (HUVECs) ^[90][192]. Conversely, there seems to be a feedback mechanism to prevent over-activation of the “NO pathway”, as demonstrated by a decrease in Ser1177 and an increase in Thr495 phosphorylation of eNOS by chronic administration of high dose inorganic nitrate, which also resulted in impaired endothelium-dependent relaxation ^[91][193], providing a possible explanation as to why chronic NO donor therapy is not suitable for targeting the NO/cGMP pathway.

Asymmetric dimethylarginine (ADMA) is a potent endogenous inhibitor of eNOS ^[92][194], which may trigger the uncoupling of eNOS ^[93][195]. ADMA serum/plasma levels have prognostic value for future cardiovascular events in patients at increased risk ^[94][95][196,197]. ADMA as an eNOS “redox switch” was previously reviewed in more detail and is based on increased ADMA generation (e.g., by the activation of S-adenosylmethionine-dependent protein arginine methyltransferase (PRMT, type I)) and on decreased ADMA degradation (e.g., by inhibition of hydrolyzing dimethylarginine dimethylaminohydrolase (DDAH) enzymes) under oxidative stress conditions ^[1][14][56][20,23,43]. The relationship between ADMA, eNOS activity, and oxidative stress was also highlighted by Sydow and Münzel ^[93][195]. According to an overview on circulating biomarkers in diabetic patients, ADMA may be used as an indicator of endothelial dysfunction in diabetes ^[96][198]. Rosuvastatin attenuated isoproterenol-induced hypertrophy, remodeling/dysfunction of the ventricle, and increased myocardial NO in rats, which was associated with normalized ADMA and PRMT1/DDAH2 expression levels ^[97][199]. Simvastatin increased DDAH1 expression levels and decreased ADMA concentrations in cultured endothelial cells ^[98][200]. Another study reported that coronary artery disease patients with too high ADMA serum concentrations are not protected anymore by simvastatin therapy against cardiovascular events, which however could be prevented by DDAH2 overexpression in a related experimental animal model ^[98][200]. Serum ADMA levels were significantly increased and endothelial function and DDAH activity were inhibited in hyperlipidemic rabbits, all of which were improved by therapy with the angiotensin converting enzyme (ACE) inhibitor captopril ^[99][201].

L-Arginine is the endogenous substrate of eNOS and “L-arginine deficiency” may contribute to the uncoupling of eNOS ^[100][202]. However, the K_m (the concentration of L-arginine that is necessary for half maximal saturation) of eNOS for L-arginine is approximately 2.9 µM, whereas intracellular L-arginine concentrations are usually in the mM range ^[1][14][56][20,23,43]. Therefore, L-arginine depletion as a regulator of eNOS was considered to be unlikely. Nevertheless, a number of pre-clinical and clinical studies reported on highly beneficial effects by oral L-arginine supplementation and also increased superoxide formation by isolated NOS enzymes in the absence of L-arginine was observed. The direct antioxidant effects of the guanidino-group L-arginine may be one explanation for these observations ^[101][203]. Improved export of ADMA from endothelial cells by high dose L-arginine represents another explanation, which would become even more attractive when the y⁺L amino acid transporters (y⁺LAT-1 and -2) are impaired (e.g., by genetic deficiency) and the cationic amino acid transporter (CAT-1) takes over using cationic amino acids, such as L-arginine for the export of ADMA [204]. We had described this concept in a case report [204]. Moreover, the uptake ^[102]of ADMA into brain may make use of these transport mechanisms and L-arginine may beneficially modulate this uptake [205]. There is clinical evidence that pharmacological treatment of patients with acute congestive heart failure can change the expression levels of the cationic amino acid transporter (CAT-1) and thereby modulate ADMA levels in serum and tissues [206].

L-arginine deficiency can also occur as a result of arginase induction. Arginases are major L-arginine-consuming enzymes that metabolize L-arginine to urea and L-ornithine. The up-regulation of arginase limits L-arginine bioavailability for eNOS, which represents a mechanism of eNOS uncoupling ^[105][207]. In the PVAT of diet-induced obese mice, arginase expression was enhanced and L-arginine content was reduced. These changes were associated with an uncoupling of PVAT eNOS ^[106][98]. Importantly, the incubation of PVAT-containing aorta with a combination of L-arginine and an arginase inhibitor for 30 min ex vivo in organ chambers restored the vasodilator effects of PVAT-containing aortas from obese mice ^[106][98]. These results demonstrate that L-arginine deficiency is indeed a reason for PVAT eNOS dysfunction and that eNOS uncoupling is causally involved in obesity-induced vascular dysfunction ^[107][96].