Sirtuin 5, vascular endothelium and fibrinolysis: a deadly embrace?

This editorial refers to ‘Sirtuin 5 promotes arterial thrombosis by blunting the fibrinolytic system’ by L. Liberale et al., pp. 2275–2288.

Vascular endothelial cells (EC) synthesize and/or secrete several molecules in the bloodstream and extracellular matrix that ensure vascular homeostasis. Nitric oxide (NO), bradykinin, prostacyclin, tissue-type plasminogen activator (tPA) from EC are critical for vasodilation, laminar blood flow maintenance, platelet inhibition, control of clot formation and lysis, respectively, thus warranting physiological haemostasis and protecting from thrombosis.^1,2 Traditional cardiovascular risk factors (ageing, hypertension, hypercholesterolaemia, diabetes, smoking) induce EC dysfunction and accelerate EC senescence, disrupt vascular homeostasis thus leading to a pro-thrombotic phenotype characterized by platelet adhesion/activation to EC, clot formation and reduced fibrinolysis.^1–3 In particular, tPA, expressed on the EC surface and released in soluble form,³ in the presence of fibrin powerfully converts plasminogen to plasmin.³ Plasmin degrades not only fibrin but also fibrinogen, coagulation factors, extracellular matrix, thus impairing the haemostatic balance.^3,4 At the EC site, plasmin is tightly regulated by inhibitors, including the PA inhibitor (PAI)-1.^3,4 PAI-1 is a serpin released by different cell types (EC, hepatocytes, adipocytes, fibroblasts, and platelets) and upregulated in dysfunctional and senescent EC.⁴ PAI-1 inactivates both tissue-type and urokinase-type PA with high affinity through a suicide substrate mechanism, generating soluble tPA/PAI-1 complexes and preventing plasmin formation. The haemostatic relevance of PAI-1 in humans is suggested by the provoked bleeding diathesis (post-surgery or trauma) in subjects with inborn PAI-1 defects.³ On the other hand, increased circulating PAI-1 antigen levels (but not activity) are associated with major atherothrombotic events or high cardiovascular risk conditions such as diabetes or obesity.³ Moreover, PAI-1 regulates not only fibrinolysis but also other pivotal pathophysiological processes including EC senescence: PAI-1 overexpressing mice develop age-dependent coronary artery thrombosis, alopecia, amyloidosis, while aged wild type (WT) mice show increased PAI-1 expression in different tissues and in stress-induced thrombosis (Figure 1).⁴ PAI-1 regulation involves diverse signalling pathways including p53, reactive oxygen species (ROS), cAMP, protein kinase (PK)A activity, adenosine monophosphate-activated PK (AMPK) phosphorylation via nuclear factor kappa-B (NF-kB).⁴

Figure 1

Simplified scheme of major SIRTs activities and regulation on the endothelial cell (EC) under physiological (left panel) and dysfunctional/senescent (right panel) conditions. Left panel: the deacetylation enzymatic activity of SIRT1 in physiologic conditions and healthy EC activates eNOS promoting NO production, inhibits p16, p21, p53, and NF-kB, reduces the transcription of PAI-1 in the nucleus and activates MMP14. Altogether, these mechanisms contribute to maintain laminar blood flow, reduce platelet reactivity and EC adhesion properties and promote fibrinolysis. SIRT3 deacetylation of FOXO3 decreases ROS generation, SIRT6-mediated deacetylation inactivates the FOXM1 transcription factor, which promotes EC senescence, thus exerting a protective effect on EC as well. ROS generation is likely inhibited by SIRT3 as well, through deacetylation of FOXO3 and by SIRT5 through desuccinylation of SOD1. Right panel: In dysfunctional EC cells, different microRNAs can inactivate SIRT1 (MiR 19b, 138, 217) and SIRT6 (miR-92a-3p). The imbalance between increased ROS and decreased NO generation downregulates vascular-protective SIRT1 and SIRT6 as well. Furthermore, SIRT5 enzymatic activity may promote the transcription and secretion of PAI-1, which in turn binds t-PA blocking fibrinolysis. SIRT5 may also increase pAMPK and promote ROS formation. Moreover, SIRT5 may also activate p53, activate glutaminase through desuccinylation, and inhibit STAT3 in the mitochondria. Altogether those effects on and from different SIRTs and related metabolic paths favour ROS-mediated EC damage, lack of NO protection, platelet adhesion/activation, thrombosis, clot growth, EC apoptosis and vascular fibrosis. Turbulent blood flow generated under these conditions can further trigger VWF unfolding, amplifying platelet adhesion and activation. Data are from references 5–11. Red lines and text indicate inhibitory effects and inhibited/reduced molecules. Green lines and text indicate promoting effects and increased/activated molecules. AMPK, adenosine monophosphate-activated protein kinase; eNOS, endothelial-NO-synthase; ERK, extracellular signal-regulated kinase; FDP, fibrinogen degradation products; FOXO3, forkhead-box-protein-O3; FOXM1, forkhead box M1; miR, microRNA; MMP, matrix metalloprotein; NF-kB, nuclear factor kappa-B; NO, nitric oxide; P, platelets; PA, plasminogen activator; PAI-1, plasminogen activator inhibitor 1; ROS, reactive oxygen species; STAT3, signal transduction and activator of transcription 3; SOD, superoxide dismutase; VWF, von Willebrand factor.

Silent information regulator 2 proteins or sirtuins are enzymes highly conserved along the phylogenesis. Mammalian sirtuins include seven (SIRT1-SIRT7) nicotinamide adenine dinucleotide (NAD)⁺-dependent enzymes with deacylase activity that remove lipid acyl (i.e. acetyl, succinyl, malonyl, glutaryl, and long-chain acyl) groups from histones, transcription factors, and/or proteins and regulate pivotal mechanisms including cell stress, glucose metabolism, senescence, and apoptosis.⁵ Sirtuins differ for the type of deacylase activity, substrate specificity, intracellular location, and tissue distribution. SIRT1 (nuclear and cytosolic), SIRT2 (cytosolic), SIRT3 (mitochondrial), and SIRT7 (nuclear) catalyse deacetylation reactions only.⁵ SIRT4 (mitochondrial) and SIRT6 (nuclear) can also catalyse adenosine di-phosphate (ADP)-ribosylation of proteins and remove longer-chain acyl groups, particularly myristyl groups.⁵ SIRT5 (nuclear, cytosolic, and mitochondrial) can remove long-chain succinyl or malonyl groups.⁶ In particular, SIRT5 appears to regulate glycolysis, ATP production, ketogenesis, and fatty acid oxidation,⁷ and exerts anti-oxidant activity through desuccinylation of superoxide dismutase-1.⁸ SIRT5 regulation remains poorly known, microRNA-19b can induce SIRT5 in fasting animals.⁷

Sirtuins regulate EC function as well (Figure 1).^6,7 Data from EC lineages in vitro, mice deleted (^−/−) for endothelial NO synthase (eNOS) or for vascular SIRT1, or overexpressing SIRT1, altogether indicate that SIRT1 predominantly protects EC from dysfunction and senescence by inducing eNOS, and reducing inflammatory and pro-atherosclerotic signals.^6,7 Notably, SIRT1 can de-acetylate histones and inhibit transcription of the PAI-1 promoter.⁴ SIRT1-mediated PAI-1 inhibition appears a relevant protective and anti-aging mechanism for EC, beyond the anti-fibrinolytic effect.⁴ SIRT6 deletion in ApoE^−/− mice is associated with impaired EC-dependent vasorelaxation, increased adhesion molecules on EC, inflammation, and atherosclerotic lesions,⁷ suggesting an overall vascular-protective role for SIRT6 as well. The effects of the other SIRTs on vascular EC have been less investigated. The CXCR4/JAK2/SIRT5 signalling improves mitochondrial function and angiogenic capacity of EC progenitors.⁷ Moreover, SIRT5^−/− mice display reduced adenosine tri-phosphate (ATP) levels under fasting conditions and develop hypertrophic cardiomyopathy.⁹ Hearts from SIRT5^−/− mice show enhanced ischaemia/reperfusion injury and larger infarct size, likely due to hypersuccinylation and activation of succinylate-dehydrogenase (SDH). In fact, ischaemia is associated with fumarate and succinate increase due to augmented formation of pyruvate, the end product of glycolysis.¹⁰ Upon reperfusion, the accumulated succinate is re-oxidized by SDH and leads to reverse electron flow through complex I generating ${\mathrm{O}}_2^{-}$⁠, responsible for the injury.¹⁰

Liberale et al.¹¹ add new evidence to the complex sirtuin system in arterial vessels and EC, providing data on a pro-thrombotic role of SIRT5, likely through PAI-1 induction (Figure 1). As compared to WT animals, SIRT5-overexpressing mice showed increased EC-dependent arterial thrombosis. Conversely, SIRT5^−/− animals showed reduced thrombus formation associated with an increased embolization, D-dimer levels, reduced soluble and EC-associated PAI-1 and reduced phosphorylation of both AMPK and ERK1/2. Consistently with data from genetically-modified animals, SIRT5-silencing in vitro on a human aortic EC lineage reduced PAI-1 transcripts and ERK1/2 phosphorylation. Peripheral mononuclear cells (PMN) isolated from acute coronary syndrome (ACS) patients showed SIRT5 transcript levels higher than non-ACS subjects. Moreover, PAI-1 mRNA increased across SIRT5 quartiles in PMNs from all (ACS and non-ACS) subjects. Thus, SIRT5 seems associated with a downregulation of plasmin generation in animal models, in EC in vitro and in PMNs from human subjects in vivo.

These interesting novel data call for further research. The trigger(s) of SIRT5 expression in damaged EC, the type of enzymatic activity and molecules accounting for phosphorylation of ERK1/2 and AMPK and for the regulation of PAI-1 transcription/expression remain to be refined. The relevance of the SIRT5/PAI-1 path in promoting human vascular dysfunction, senescence, and atherothrombosis as compared to other PAI-1 regulators, such as SIRT1, would require further insights as well. Although PAI-1 antigen increases in association with cardiovascular diseases and complications in humans, it remains unclear whether PAI-1 is simply a biomarker of cardiovascular complications and/or EC senescence or it is rather a mediator, thus representing a relevant therapeutic target. Moreover, SIRT5 appears not only to promote EC dysfunction and arterial thrombosis in animals but also to protect from ROS-mediated damage in ischaemia–reperfusion injury and related infarct.⁹ Finally, the interplay between SIRT5 and the other sirtuins, such as SIRT1, SIRT6 and SIRT3, known to be relevant in regulating EC ageing and dysfunction, including PAI-1 transcription, will deserve further understanding.

In conclusion, the role of SIRT5 activity in EC regulation is getting increasingly complex and puzzling, calling for further studies on the interplay between sirtuins and PAI-1 in human vessel protection, dysfunction and ageing.

Conflict of interest: none declared.

The opinions expressed in this article are not necessarily those of the Editors of Cardiovascular Research or of the European Society of Cardiology.