COVID‐19‐induced endotheliitis: emerging evidence and possible therapeutic strategies
Eleonora Calabretta, José M. Moraleda, Massimo Iacobelli, Rubén Jara, Israël Vlodavsky, Peter O’Gorman, Antonio Pagliuca, Clifton C. Mo, Rebecca M. Baron, Alessio Aghemo, Robert J. Soiffer, Jawed Fareed, Carmelo Carlo‐Stella, Paul G. Richardson
Abstract
The coronavirus disease 2019 (COVID-19) pandemic, a viral illness caused by the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2),1 has produced at the time of this writing nearly 33 million cases of infection, with over a million deaths in 235 countries,2 causing an unprecedented burden on healthcare systems and a severe global socioeconomic crisis. As the pandemic spreads, knowledge on the disease course, as well as potential risk factors and predictors of severity is increasing daily, and initial data from randomised controlled studies have allowed care providers to refine therapeutic strategies. Nonetheless, mortality is markedly elevated among those presenting with severe disease, long-term sequelae among survivors are unknown, and vaccine-based therapies currently remain at early stages of development. Most reported cases are asymptomatic or present with mild symptoms; however, 7–26% of hospitalised patients experience severe disease, often requiring admission to intensive care units (ICUs), with progressive multiple organ dysfunction and high mortality.3-5 Such differences in clinical outcomes have led physicians to initiate diverse pharmacological therapies at various stages of the disease, generating challenges as to the most appropriate therapeutic choice for COVID-19. In this context, the use of dexamethasone has significantly reduced mortality rates in critically ill patients requiring supplemental oxygen or mechanical ventilation,6 and remdesivir has demonstrated clinical benefit in hospitalised patients, but with unknown survival benefit to date7; additional effective treatment options are therefore urgently needed. In an initial attempt to provide a uniform and widely reproducible methodology to guide systematic treatment strategies, a three-stage classification of COVID-19 has been proposed.8 The Stage I or 'early infection' occurs at the initial establishment of disease with high viral replication, and commonly presents with a range of complaints that can include mild and often non-specific influenza-like signs and symptoms. Stage II is the 'pulmonary phase', with preferential viral-mediated injury of the lung parenchyma and this is characterised by shortness of breath, hypoxia and pulmonary infiltrates with some degree of lung inflammation. Stage III is characterised by an exaggerated host immune-inflammatory response to the virus, leading to acute respiratory distress syndrome (ARDS) and multi-organ failure (MOF). Emerging evidence suggests that endothelial damage and subsequent morphological and functional changes in the endothelium play important roles in COVID-19-induced hyperinflammation. The virus, which binds to the angiotensin-converting enzyme 2 (ACE2) receptor,9 displays a profound tropism for human lung and small intestine epithelium, as well as the vascular endothelium.10 In an important case series from Varga et al.,11 postmortem histology from three patients affected by late-stage COVID-19, revealed viral inclusions in endothelial apoptotic cells and microvascular lymphocytic endotheliitis, with infiltration of inflammatory cells around the vessels and endothelial cells (ECs), as well as evidence of endothelial apoptotic cell death in the lung, kidney, small bowel and heart. Additionally, autopsy findings of 27 patients in another series confirmed the detection of the SARS-CoV-2 in multiple organs, including the respiratory tract, pharynx, heart, liver, brain, and kidneys.12 Immunofluorescence of kidney specimens from six of the 27 patients showed the presence of SARS-CoV-2 protein in all renal compartments, and in three of the patients preferentially in the endothelium of the glomerulus. Similar microscopic findings were also noted in lung specimens from seven patients with COVID-19, which displayed small vessel endotheliitis, microvascular thrombosis and angiogenesis, along with the presence of SARS-CoV-2 in pulmonary ECs, an observation strongly supporting the vascular tropism of the virus.13 Lastly, Stahl et al.14 have identified in the plasma and serum of 19 critically ill patients with COVID-19 evidence of disruption of the endothelial glycocalyx, reflected by increased levels of the Tie-2 receptor and syndecan-1 (SDC-1), a heparan sulphate (HS) proteoglycan. This particular observation is of interest as the endothelial glycocalyx covers the luminal surface of ECs, and its integrity is vital for the maintenance of vascular homeostasis. Such findings suggest that virus-mediated apoptosis may promote endothelial barrier disruption with interstitial oedema and increased recruitment of circulating activated immune cells, thus causing widespread endothelial dysfunction, as well as activation of platelets and the coagulation cascade leading to venous and arterial thrombosis.15 Altered pro-inflammatory and pro-thrombotic status is confirmed by the presence of elevated inflammation-related indices (e.g. C-reactive protein and serum ferritin), humoral biomarkers [interleukin (IL)-2, IL-6, IL-7, granulocyte-colony stimulating factor (G-CSF), tumour necrosis factor-alpha (TNF-α)], and indicators of an increased pro-coagulant-fibrinolytic state [e.g. von Willebrand factor (VWF), D-dimer, fibrinogen]. Further, factor VIII (FVIII), a potent and key factor in the coagulation process, is greatly increased in ICU patients with COVID-19.16 In this setting, it has been proposed that the systemic hyperinflammation observed in severe COVID-19 is comparable to a cytokine release syndrome (CRS), or cytokine storm. IL-6 has a central role in the generation of the cytokine storm, and is commonly elevated in the serum of severely ill patients with COVID-19.17 High levels of IL-6 activate ECs, thus resulting in vascular leakage, further cytokine secretion and activation of the complement and coagulation cascades.18 Interestingly, a population of IL-6 producing monocytes was found to be expanded in the peripheral blood of ICU patients19 and an aberrant macrophage response exhibiting increased levels of pro-inflammatory cytokines has been detected in bronchoalveolar fluid, especially in severely ill patients.20 Although the exact driver of monocyte activation remains unclear, such cells are attracted to the endothelium, where the release of highly noxious molecules, such as reactive oxygen species (ROS), contributes to endothelial dysfunction and promotes hyperinflammation.21 Further, activated monocytes enhance tissue factor expression and form aggregates with platelets through P-selectin interaction and hence augmenting the pro-coagulant response.22 Indeed, significantly increased levels of VWF and FVIII and thrombomodulin,16 aberrant coagulation, thrombosis and microangiopathy are very common in critically ill patients with COVID-19, resulting in a disseminated intravascular coagulation (DIC)-like syndrome characterised by massive fibrin formation and organ dysfunction.23 Likewise, adaptive immunity actively participates in the establishment of the inflammatory response; specifically, activated and proliferating CD8+ T cells are prevalent in mild COVID-19, whereas critically ill patients display higher levels of hyperactive IL-6-producing CD4+ T cells, which may contribute to disease severity, even after viral clearance.17 Interestingly, T cells show phenotypical signs of an exhausted, functionally unresponsive state, thus allowing viral escape from immune surveillance.22 Once initiated, the endotheliitis and resultant cytokine storm become self-sustaining, leading to widespread organ damage. Some patients may also display features of haemophagocytic lymphohistiocytosis, such as cytopenias, hyperferritinaemia and rapid onset of MOF. Overall, once hyperinflammation and CRS develop, rates of mortality significantly increase.24-26 As direct viral activation of the vascular endothelium has an important role in initiating and maintaining the hyperinflammatory response, attempting to blunt such a response with endothelial-protective agents is a very rational strategy. Controlled clinical trials focussing on the use of anti-cytokine antibodies, including tocilizumab (IL-6 inhibitor), have failed to show significant activity in this stage of the disease.27 However, increasing evidence suggests that the altered homeostasis of the endothelium may be a key initiating event in the pathogenesis of the disease, therefore representing a potentially more promising target.28 Clinically and histopathologically, COVID-19-associated endotheliitis resembles a spectrum of post-bone marrow and stem cell transplantation (BMT) syndromes characterised by disruption of endothelial homeostasis and consequently dysregulation of coagulation, vascular tone, endothelial permeability and vascular inflammation.29 These disorders include hepatic veno-occlusive disease (VOD)/sinusoidal obstruction syndrome (SOS), idiopathic pneumonia syndrome (IPS), transplant-associated thrombotic microangiopathy and graft-versus-host disease (GvHD). Hepatic VOD/SOS develops as a result of endothelial damage to hepatic sinusoids and subsequent hepatocyte necrosis30. Damage to the ECs leads to a hypercoagulable state, production of inflammatory mediators, and the upregulation as well as release of heparanase.31 Heparanase degrades the heparan sulphate scaffold of the subendothelial basement membrane, consequently allowing the extravasation of blood-borne cells, including activated T lymphocytes, neutrophils and macrophages.32 This cascade of events leads to postsinusoidal hypertension, hyperinflammation and ultimately MOF. Severe VOD/SOS associated with MOF without effective therapy is fatal in >80% of cases.33 Interestingly, the histopathological examination of lung lesions in VOD/SOS shows early alveolar epithelial and lung endothelial injury, resulting in accumulation of protein- and fibrin-rich inflammatory oedematous fluid in the alveolar space and progression to interstitial fibrosis,34, 35 as is also seen in fatal COVID-19 cases. Similarly, IPS, a widespread alveolar injury in the absence of identifiable infectious or non-infectious causes, is characterised by histological evidence of EC injury with fibrin accumulation, luminal thrombosis and fibrotic processes. Adhesion molecules, such as intercellular adhesion molecule 1 (ICAM-1) and/or vascular cell adhesion molecule 1 (VCAM-1), are commonly upregulated, thus reflecting profound endothelial activation.36 It has been suggested that TNF-α directly causes endothelial injury, and increased levels of angiopoietin-2 (Ang-2), have been recently reported in cases of acute exacerbations of IPS,37 similar to that seen in severely ill patients with COVID-19.38 Likewise, multifactorial endothelial damage has been also implicated in the development of transplant-associated thrombotic microangiopathy, where micro-vessel intimal swelling and necrosis lead to the formation of luminal microthrombi and subsequent microangiopathic haemolytic anaemia. Plasma levels of markers of EC injury and inflammation, such as thrombomodulin, plasminogen activator inhibitor-1 (PAI-1), ICAM-1, VCAM-1, IL-1, TNF-α, interferon gamma and IL-8 are commonly elevated.39, 40 Endothelial dysfunction predominantly affects the kidneys and the brain, but may become widespread and progress to MOF, which in turn is associated with high mortality. Lastly, acute GvHD (aGvHD) develops as a consequence of the activation of the immune system. Antigen-presenting cells become activated by endothelial and tissue damage derived from direct toxicity of the conditioning regimen, thus initiating an alloreactive T-cell response directed against recipient tissues.41 As a result, SDC-1 is commonly elevated in the serum of patients with GvHD and correlates with disease severity.42 In addition to cell-mediated cytotoxic damage, the cytokine storm generated in response to T-cell activation and proliferation causes targeted organ damage involving mainly the skin, liver and gut.43 It has recently been suggested that endothelial vulnerability and pro-thrombotic shift precedes clinically evident aGvHD and that angiogenesis driven by early endothelial activation is an initiating event.44 Indeed, increased plasma levels of VWF,45 Ang-246 and TNF receptor 147 have been detected in patients prior to development of aGvHD, correlating with response to therapy. Similarly, markers of endothelial dysfunction and inflammatory activation have been detected also in the serum of patients with sickle cell disease (SCD), especially during vaso-occlusive episodes. SCD is characterised by a chronic course of relapsing-remitting episodes of ischaemia and then reperfusion. The polymerisation of defective haemoglobin S upon deoxygenation initiates many pathological processes, such as complement activation, generation of ROS and pro-thrombotic molecules, secretion of numerous pro-inflammatory cytokines and chemokines and ultimately leucocyte recruitment.48 Oxidative stress and endothelial dysregulation plays a key role in vaso-occlusion; ECs activated by substances released by the haemolytic process and by red blood cell adhesion initiate production and release of soluble mediators such as IL-1β, IL-8, IL-6, IL-1α and PAI-1,49, 50 and increase the expression of adhesion molecules such as VCAM-1, ICAM-1, E-selectin and P-selectin,51, 52 reflecting a pro-inflammatory and pro-thrombotic shift. Vaso-occlusive phenomena commonly affect the lung vasculature, provoking acute chest syndrome (ACS), a spectrum of diseases ranging from mild pneumonia to ARDS and MOF, which is the leading cause of morbidity and mortality in SCD.53 Lung specimens from ACS cases showed micro-thrombotic occlusion, endothelial VWF deposition and arterial vessel re-modelling with initial fibrotic processes,54 fascinatingly all comparable to the histopathological findings in COVID-19.55 Importantly, heme-mediated endothelial damage to alveolar cells is regulated by the p38 mitogen-activated protein kinase (MAPK) pathway, which plays a crucial role in the biosynthesis of pro-inflammatory cytokines and collagen production.56 This key pathway is also upregulated in COVID-19 as a result of decreased ACE2 tissue functionality consequent to viral binding, and may consequently promote endotheliitis, hypercoagulation and end-stage fibrosis.57, 58 In summary, post-BMT syndromes, vaso-occlusive organ dysfunction in SCD and COVID-19-associated endotheliitis share common pathological mechanisms including: i) dysregulation of the homeostasis of the endothelial milieu toward a pro-inflammatory and pro-thrombotic phenotype with thrombotic microangiopathy; ii) hyperproduction of inflammatory cytokines such as IL-6, IL-8 and TNF-α;59-62 and iii) small vessel endotheliitis and endothelial barrier dysfunction, leading to oedema of the microvascular bed, protein and fibrin accumulation and subsequent fibrotic shift.34, 36 All these conditions if untreated irremediably lead to MOF and display similar microscopic and macroscopic features in target organs upon pathological examination. At the molecular level, the p38 MAPK pathway may also be critical in promoting vasoconstrictive and inflammatory phenomena; its activation is described in SCD, COVID-19 and also as a result of conditioning regimen-induced endothelial damage in BMT.63 Together, these findings support the notion that the pleiotropic character of the endothelium as a key regulator of the internal homeostasis, vascular tone, blood coagulation and the inflammatory process and therefore of so called 'immune-thrombosis' events, make it an intriguing therapeutic target for post-BMT disorders, SCD, and COVID-19. Classically, heparins have been the most widely used drugs for the treatment and prevention of endothelial cell disorders. Several animal studies and clinical trials have suggested that, in addition to its well-known anticoagulant effects, heparin also possesses anti-inflammatory properties, mainly mediated by inhibition of IL-6 release and its activity,64 a phenomena also demonstrated in patients with COVID-19 treated with low-molecular-weight heparin.65 Further, heparin is structurally related to HS,66 a negatively charged glycosaminoglycan as described earlier, which serves as binding sites for growth factors, cytokines, selectins, extracellular-matrix molecules, and a large number of human viruses,67 including the SARS-CoV-2 virus.68, 69 Indeed, Clausen et al.70 have recently demonstrated that the SARS-CoV-2 spike protein must bind both the ACE2 receptor and HS to enter human cells. The structural analogies between heparin and HS may result in competitive inhibition, where heparin and related compounds compete with the cell surface HS for viral binding to target cells,68, 69 thus potentially blocking or at least attenuating viral entry. The beneficial effects of heparin-based therapies are also linked to their inhibition of circulating heparanase enzymatic activity.32 Heparanase, an endo-β-glucuronidase, physiologically cleaves HS chains located in extracellular matrices and on cell surfaces.32 It is often overexpressed during viral infections and act as a regulator of virus release after replication has occurred, promoting its dissemination.71-74 Additionally, it may be upregulated by pro-inflammatory molecules such as IL-1 and TNF-α. Once activated, heparanase stimulates the expression and release of pro-inflammatory cytokines, including TNF-α, IL-1 and IL-6.75 The enzyme has been implicated in cancer progression, inflammation,76 VOD/SOS development77 and other vascular pathologies.71 Currently, numerous clinical trials are underway to investigate the therapeutic potential of intravenous and subcutaneous heparin, as well as the appropriate dose regimen in COVID-19. Further, nebulised heparin delivered directly to the airways may be effective in preventing infection and mitigating lung disease (clinicaltrials.gov; NCT04545541, NCT04511923). Notwithstanding their anticoagulant, anti-inflammatory and anti-viral properties, the use of heparins is associated with a substantially increased risk of systemic bleeding, and other challenging 'off-target' effects, making its use potentially part of the standard of care, but not without qualification, as well as highlighting the need for combination approaches. The use of defibrotide (DF), which has both comparable but distinct properties from heparins and negligible haemorrhagic risk78 may therefore be warranted, especially given the established propensity for the development of DIC later in the COVID-19 clinical course. DF is a naturally derived, complex mixture of poly-deoxyribonuleotides extracted originally from bovine lung and now exclusively from porcine gut mucosa.79, 80 Since its original isolation >30 years ago, DF has demonstrated locally acting pro-fibrinolytic,81-84 anti-thrombotic,85, 86 anti-ischaemic and anti-inflammatory activities, which exert protective effects on small vessel endothelia. It is currently approved for the treatment of paediatric and adult hepatic VOD/SOS with MOF.87-89 In this setting, DF has demonstrated efficacy and safety in critically ill patients with MOF, as well as a significant reduction in PAI-1 and other markers of endothelial stress in patients with VOD/SOS and MOF successfully treated with DF.89-91 Furthermore, in a pivotal Phase III trial, DF prophylaxis reduced the incidence and severity of VOD/SOS in high-risk children undergoing BMT.92 In a more recent study, Palomo et al.93 demonstrated that DF directly interacts with the cell membrane and becomes internalised by ECs, thus providing physical evidence of its endothelial-protective properties. In particular, DF appears to decrease levels of pro-inflammatory proteins, such as TNF-α,94 IL-6, vascular endothelial growth factor (VEGF)95 and to downregulate major histocompatibility complex (MHC) Class I and Class II molecules,96, 97 therefore attenuating both the inflammatory and immune responses. Furthermore, it appears to decrease interaction between leucocytes and ECs by downregulating P-selectin,98 ICAM-195 and VCAM-1.99 Lastly, DF displays potent adenosine agonism.100 Such activity may be clinically relevant, not least based on substantial improvement observed in an animal model of acute lung injury upon treatment with adenosine receptor agonists.101 Based on such properties, the use of DF can be reasonably extended to other post-BMT syndromes and other microangiopathies involving CRS complicating a variety of disease states and treatment modalities, such as chimeric antigen receptor (CAR) T-cell therapy.102 Indeed, paediatric and adult patients receiving DF as VOD/SOS prophylaxis also exhibited a reduced incidence of aGvHD,92, 103 a finding that is strongly supported by a preclinical model of aGvHD.99 Additionally, a retrospective survey from paediatric patients treated with DF for transplant-associated thrombotic microangiopathy showed of clinical disease in of Currently, Phase II studies the use of DF for prevention of transplant-associated thrombotic microangiopathy and and in the the treatment of ACS are (clinicaltrials.gov; DF the expression of heparanase cell surface expression and enzymatic that DF may have anti-viral properties, this remains to be Heparanase is upregulated by the cytokine storm of COVID-19 and may contribute to further inflammation, oedema of the microvascular and DF is a potent of heparanase in of both cell surface and and therefore is especially Furthermore, the therapeutic use of DF in a model of significantly survival to untreated by among other the levels of which is to with ARDS and is markedly elevated in critically ill patients with COVID-19.38 In patients with COVID-19 may display increased activation and subsequent formation of aggregates upon interaction with thus stimulating and P-selectin and other adhesion molecules DF may inflammatory and pro-coagulant and most DF has also been to decrease the activity of p38 MAPK and its the of which is in the pathogenesis of the COVID-19 hyperinflammation syndrome and this may be a key therapeutic target in this In summary, the therapeutic properties of DF and its as well as its make it an potential therapeutic for the treatment of COVID-19 vascular In to heparin, DF also anti-inflammatory and properties. Importantly, by acting on the DF may viral given its to i) compete with HS and adhesion and ii) heparanase enzymatic activity and virus and and iii) activation of immune cells and upregulation of pro-inflammatory cytokines and the associated systemic inflammatory host response Phase II clinical trials are now underway and critical on therapeutic potential in patients with COVID-19 include critically ill paediatric patients treated with DF for a inflammatory syndrome and with studies supporting the effects described as well as effects seen on complement Similarly, from the studies as part of the as and of and of support both safety and promising potential efficacy to the activity of heparin, has been directed of its to molecules with activity anticoagulant and heparins are promising agents presenting such Indeed, of in of lung and injury and subsequent oedema by tissue infiltration and IL-6, and TNF-α Furthermore, the most pathological renal damage caused by by release of pro-inflammatory cytokines and established fibrotic processes, thus tissue histology in preclinical This is especially relevant, the formation of and damage in post-BMT syndromes, SCD and COVID-19. Additionally, interest has been directed the a with heparan properties. is a potent of 1 cells IL-6 macrophage angiogenesis and tumour progression in Furthermore, it mild anticoagulant activity and is well et have recently demonstrated that interacts with the SARS-CoV-2 spike protein binding and this is with its heparan was found to markedly SARS-CoV-2 supporting its clinical as a therapeutic for prophylaxis and treatment of COVID-19. heparanase as a factor that once activated viral and inflammatory These data that heparanase currently development are for inflammatory such as COVID-19, thrombotic microangiopathies and but as of now studies remain preclinical with clinical In increasing evidence suggests that the SARS-CoV-2 directly ECs, promoting the release of pro-inflammatory and pro-thrombotic Endothelial dysfunction appears to be a crucial initiating in the pathogenesis of the disease and its morbidity and mortality. with the hyperproduction of cytokines leading to and thrombotic microangiopathy are by COVID-19, VOD/SOS and other endothelial injury syndromes, by and including the vaso-occlusive of SCD, so providing a common these Most endothelial-protective such as a promising and rational therapeutic in COVID-19, with DF currently in a variety of and As a heparanase inhibition, with the of related and other effects on endothelial stress may thus be crucial in anti-viral and anti-inflammatory In particular, as this to endotheliitis, it may directly CRS and its which in turn may lead to This was supported in part by a from the for to and the all in the critical and of the from has and from is on a and a for serves on the of for and The for and and for is an for is a for and The to