The function of the lungs in gas exchange means the endothelium here is unique from endothelium in other parts of the body. The lungs are a low-pressure system where the entire blood volume passes through. Hypoxic pulmonary vasoconstriction (HPV) is a physiological response of the lungs to low oxygen levels in the alveoli (the tiny air sacs where gas exchange occurs). This response involves the constriction of blood vessels in the lungs, which helps to redirect blood flow to areas of the lungs with better oxygenation and improve the efficiency of gas exchange. While the systemic circulation vasodilates in response to low oxygen signals, the pulmonary vessels constrict to redirect blood flow to better aerated alveoli. This response is more pronounced as the diameter of the vessels decreases (23). The pulmonary endothelium is highly metabolically active and involved in maintaining vascular tone through nitric oxide, prostacyclin, serotonin and endothelin production (24).

One of the main functions of the pulmonary vascular endothelium is to preserve the airspaces against vascular fluid. The permeability of this membrane is concerned in several disease states, indicative of the health of different organs. Permeability leads to trans-endothelial movement of fluid and cells into airspaces (Figure 1).

The lungs are also an important site of immunological defence, so the endothelium is specialised at vascular branch points, which act as filters to trap and transport potential antigens to local lymph nodes (25). The immune response recruits inflammatory cells to the lung where adjacent cells respond by releasing inflammatory substances such as arachidonate, ATP and reactive oxygen species which damage the endothelium (26–28). ARDS precipitants such as gastric acid and mechanical stress increase mitochondrial transcription of leukocyte adhesion molecules (28). ARDS generating conditions increase the concentration of mitochondrial cytosolic calcium which leads to nuclear factor kappa B (NFκB) production through a hydrogen peroxide pathway. This activation leads to production of E-selectin and P-selectin, important inflammatory and adhesion molecules in the lung.

The alveolar glycocalyx is a layer of glycoproteins and proteoglycans that covers the surface of the alveolar epithelium, which is the tissue that lines the alveoli in the lungs (29). The glycocalyx is a part of the extracellular matrix of the alveoli and plays an important role in maintaining the integrity and function of the alveolar barrier (30). The alveolar glycocalyx is involved in regulating the exchange of fluids, electrolytes, and other molecules across the alveolar epithelium, which is essential for normal lung function (Figure 1). It also helps to protect the alveolar epithelium from damage caused by mechanical stress and inflammation. Recent studies have suggested that the alveolar glycocalyx may play a role in the pathogenesis of ARDS (15, 30, 31). Damage to the glycocalyx has been shown to increase the permeability of the alveolar barrier and contribute to the development of pulmonary oedema (Figure 2). Animal models of inhaled pulmonary insults lead to alveolar epithelial glycocalyx breakdown, loss of surfactant and reduced lung function (11). Studies have shown that direct pulmonary insults lead to epithelial glycocalyx breakdown and are correlated to PaO2/FiO2 and outcomes (duration of mechanical ventilation) in ARDS patients (15). This distinction between the pathogenesis of direct and indirect lung injury in ARDS could lead to new diagnostic and treatment targets (32).

The destruction of the endothelial glycocalyx has been associated with increased vascular permeability, leukocyte adhesion, and inflammation in the lungs (33, 34). Moreover, clinical studies have demonstrated that patients with ARDS have increased levels of glycocalyx components in their circulation, suggesting that glycocalyx damage may be a feature of the disease (35–37). Although the exact mechanisms underlying glycocalyx destruction in ARDS are not fully understood, it is thought to be related to a combination of factors, including oxidative stress, inflammation, and mechanical stress.

Animal models have been used to demonstrate the relationship between endotoxemia and loss of endothelial thickness (38–41). LPS injection causes increased pulmonary capillary permeability with endothelial leak that recovers over 96 h (40).

Tight junctions between endothelial cells maintain barrier integrity, which is dependent on vascular endothelial cadherin (VE-cadherin) (42). Endothelial permeability is regulated by adherens junctions and tight junctions, particularly on the venular side where there are molecular targets for inflammatory mediators such as tumour necrosis factor (TNF), interleukin-1 (IL-1) and vascular endothelial growth factor (VEGF) (42, 43). Endothelial leak is promoted by intracellular sequestration of VE-cadherin which is initiated by mediators such as thrombin and VEGF (44). Inflammatory mediators thrombin and VEGF stimulate dismantling of adherens and tight junctions in a calcium dependent manner by phosphorylating VE-cadherin, leading to actin chain contraction and increased intercellular permeability (45).

Angiopoeitin 2 (Ang-2) is a marker of endothelial dysfunction and has therefore been studied extensively in patients with or at risk for ARDS. Ang-2 works with Tie-2 and in opposition to Ang-1 to control permeability throughout the vascular network. A retrospective analysis of 931 patients from the NHLBI ARDS Network randomised controlled trial (RCT) fluid liberal vs. fluid conservative management strategy in ALI (FACTT) trial found that Ang-2 was an independent predictor of mortality in ARDS as well as being statistically significantly associated with infectious aetiology of acute lung injury (ALI) (37). A meta-analysis of prospective cohort studies focussing on Ang-2 and risk of mortality in ARDS found increased odds of death in patients with a raised baseline Ang-2 level (OR 1.56, 95% CI 1.3–1.89) (46). In patients with infection related ARDS, Ang-2 levels higher on day 3 than on day 0 was independently associated with death at an OR 2.29 (95% CI 1.54–3.43, p < 0.001). Ang-2 levels on day 3 were associated with mortality in infection related ARDS (OR 1.64, 95% Ci 1.32–2.03, p < 0.001) as well as in non-infection related ARDS (OR 2.03, 95% CI 1.31–3.16, p = 0.002). Patients in the fluid conservative group had a 13% greater decline in Ang-2 over the 3 days (p = 0.005) and this effect was more in those patients who were not in shock and experienced a 19.9% fall in Ang-2 levels (37).

Using Mendelian randomisation a causal link between Ang-2 and ARDS was investigated (47). In this study the authors found that patients with European ancestry who had sepsis were more likely to develop ARDS if they expressed the ANGPT2 genetic variant. Five ANGPT2 gene variants had associations with ARDS in sepsis patients with European ancestry. This not only describes a potential subset of patients but also highlights how genetic data could be used to classify and phenotype cohorts in ARDS.

A study of single nucleotide polymorphisms (SNPs) in 225 ARDS patients found an association with FLT1, which encodes a VEGF receptor (48). The authors proposed this as a potential mechanistic link between endothelial damage and ARDS. Forty-nine percent of patients in the cohort had positive blood cultures with Gram-negative bacteria identified and 49% had direct pulmonary injury resulting in ARDS (48). Utilizing omics data represents a promising approach for elucidating the underlying mechanisms of intricate pathologies, such as ARDS.

Intercellular adhesion molecule (ICAM) is a cell surface glycoprotein that plays an important role in leukocyte adhesion and trans-endothelial migration during inflammation (49). ICAM expression on endothelial cells can induce actin stress fibre formation, which is a cytoskeletal structure composed of contractile actin filaments that are important for maintaining cell shape and motility (50). ICAM expression can also increase endothelial permeability through a positive feedback loop involving the activation of intracellular signalling pathways, such as the RhoA/ROCK pathway. The formation of actin stress fibres induced by ICAM expression can activate RhoA/ROCK signalling, which in turn leads to the assembly of actin-myosin contractile structures and increased endothelial permeability (51). This positive feedback loop between ICAM expression, actin stress fibre formation, and increased permeability is thought to play an important role in the pathogenesis of various inflammatory disorders, including acute lung injury and sepsis. In these conditions, increased ICAM expression and activation of RhoA/ROCK signalling can lead to a breakdown of the endothelial barrier, resulting in the leakage of fluid and cells into the interstitial space and impaired organ function (49).

Alveolar red blood cells and pulmonary haemorrhage are also features of ARDS. RBC rheology in sepsis is affected by several factors, including increased nitric oxide release affecting membrane deformability as well as increased aggregation and increased viscosity (52). Altered membrane deformability can lead to haemolysis and the release of cell free haemoglobin, which travels into alveoli and causes damage (53). This was proposed as a mechanism causing transfusion related acute lung injury (TRALI) however, two RCTs have disputed this mechanism, showing no benefit of fresh RBCs in overall survival of critically ill patients (54, 55). Cell-free haemoglobin precipitates release of reactive oxygen species (ROS) which impairs the endothelial integrity, leading to vasculopathy (56, 57). Large cohort studies have been employed to investigate the influence of cell free haemoglobin and pulmonary hypertension in sickle cell anaemia (56, 58, 59) RBC haemolysis products can be considered danger-associated molecular patterns (DAMPs) that could cause lung injury by damaging the endothelium (60).

The pulmonary endothelium plays an important role in COVID-19 related lung injury by blood flow regulation, maintaining vascular integrity, and preventing the extravasation of fluid and cells into the lung tissue (65). In COVID-19, the virus can infect pulmonary endothelial cells through the ACE2 receptor, which is expressed on the surface of these cells (66). The viral infection can lead to endothelial dysfunction, inflammation, and increased vascular permeability, which may contribute to the development of severe respiratory symptoms and organ dysfunction in COVID-19 patients (67–69). The endothelial dysfunction can also lead to the pro-inflammatory and pro-thrombotic state in the pulmonary endothelium with formation of microthrombi further exacerbating tissue damage (70).

Fluid therapy for patients with sepsis is the standard of care worldwide (72, 73). Excessive fluid resuscitation and hyper-oncotic solutions have been shown to increase damage to the glycocalyx (74). The FINNAKI trial showed that fluid administration is correlated to amount of fluid transfused and non-survivors had more SDC-1 and thrombomodulin circulating than survivors (75). Glycocalyx damage assessed by peripheral heparan sulfate levels was also shown to correlate with each litre of intravenous fluids administered (74). Excess fluid administration is associated with poorer outcomes in sepsis (76–78). Albumin also preserves the glycocalyx and reduced leukocyte adhesion and SDC-1 concentrations (79). A recent review found that plasma and albumin were superior to crystalloids and colloids in preserving the glycocalyx (Table 3) (84). Conservative fluid resuscitation strategies in ARDS patients aim to prevent pulmonary oedema and may act by preventing glycocalyx breakdown and leakage.

Recombinant thrombomodulin has also been investigated as a potential therapy to induce endothelial glycocalyx repair in ARDS. Thrombomodulin is a key component of the endothelial glycocalyx with anticoagulant effect, binding thrombin to generate activated protein C. By neutralizing high mobility group B1 (HMG-B1) released by necrotic cells it attenuates inflammation (85). Recombinant thrombomodulin has been investigated as a treatment for sepsis induced coagulopathy due to its effect on protein C (86). However, the SCARLET trial found no benefit at 28-days for patients with sepsis and DIC who received rhTM. A study investigating the possible effects of recombinant thrombomodulin on LPS-induced ARDS in mice administered recombinant thrombomodulin, which was associated with less capillary endothelial disruption and oedema than control (39). Genetic analysis suggested that recombinant thrombomodulin treatment may affect anti-inflammatory, cell proliferative and glycocalyx synthesis pathways and recombinant thrombomodulin might enhance glycocalyx synthesis. Survival of recombinant thrombomodulin treated mice was significantly higher than control mice after 48 h. Another study found that recombinant thrombomodulin might be a regulator of inflammation in the lung endothelium and protect against endothelial damage due to Streptococcus pneumoniae (87). The effect of recombinant thrombomodulin to reduce inflammation and possibly influence glycocalyx repair in ARDS warrants further investigation.

Neutrophil elastase, released from activated neutrophils in the immune response, damages pulmonary endothelium and is a factor contributing to ARDS (41). Sivelestat is a competitive inhibitor of neutrophil elastase, preventing pulmonary endothelial permeability but not affecting the immune response elsewhere (88). Though it has shown some efficacy in treating ALI, ARDS and SARS-CoV2,poorly designed, negative and small trials have hampered its adoption. Although some trials have shown a connection between PaO2/FiO2 and sivelestat therapy, there was no improvement in ventilation days or mortality (89–92). Retrospective studies have also had their validity questioned based on how they selected patients to show benefit in sepsis patients with disseminated intravascular coagulation and ARDS (93). Despite this controversy, the potential for a neutrophil elastase inhibitor to improve outcomes in ARDS is intriguing. A well designed randomised trial, enriched with subgroups of at risk patients could answer these questions in the future (94).

The negative association of high Ang-2 with mortality in ARDS has also gained attention as a possible therapeutic target. Ang-2 is a product of the renin-angiotensin-aldosterone system (RAAS) that is responsible for blood pressure, vascular permeability, vasodilation and sodium absorption. GSK2586881 is a recombinant human angiotensin converting enzyme type 2 that was studied in ARDS and aimed to reduce levels of Ang-2, increasing Ang (1–7) thereby promoting anti-inflammatory and vasodilatory effects (95). The study was terminated early for failing to meet primary endpoints, though it has gained some attention for its potential treating COVID-19, there have been no further trials in ARDS, that we are aware of (80, 96). The involvement in of the RAAS in development of ARDS or patient’s failure to progress from mechanical ventilation could represent an important biologic marker for future prognostic enrichment (97, 98).

Imatinib mesylate is a small molecule therapy that was initially developed for cancer treatment with several targets including Abl, Arg (Abl-related gene), the stem-cell factor receptor (c-KIT) and platelet-derived growth factor receptor (PDGF-R) (99). Through its effect on Rac1 transmission and Arg, it improves endothelial cell barrier function and prevents vascular leak (100). Imatinib protects the endothelial barrier through inhibition of the ABL-2 tyrosine kinase, which is a key regulator of barrier function (100, 101). In vitro and mouse models have highlighted its possible differential effect in treating LPS-induced and ventilator induced lung injury models of ARDS (102). This raises an interesting question about the potential uses of the drug in homogenous populations of ARDS patients, depending on the inciting cause. A multicentre randomised placebo controlled trial of imatinib in COVID-19 disease found imatinib reduced mortality (14% placebo group vs. 8% imatinib group; adjusted hazard ratio 0.52, CI 0.26–1.05, p = 0.068) (103). Although the primary outcome of this trial, time to discontinuation of supplemental oxygen, did not reach statistical significance. Secondary analysis found that the reduction in mortality attributed to imatinib was completely mediated through immunomodulation and reversal of endothelial barrier dysfunction (81). Imatinib has shown some efficacy in treatment of pulmonary vaso-occlusive disease and post-chemotherapy fibrotic lung disease (104, 105). Case reports and series have emphasised imatinib’s usefulness treating various syndromes of endothelial leak, meaning it could have an important role in the future treatment of a subset of ARDS patients (82, 105, 106).

Other treatments for ARDS targeting the endothelium have been sought out. Including old Chinese remedies such as Crocin, that may inhibit inflammation signalling as well as heparanase and matrix metalloproteinase 9 to preserve heparan sulfate and syndecan-4 in LPS induced ARDS (107). Therapies such as mesenchymal stromal stem cells have also been investigated for their anti-inflammatory and endothelial restorative abilities (83). Statins have also been investigated for their ability to reduce vascular leak in ARDS. HARP-2 and SAILS trials did not find any benefit of simvastatin or rosuvastatin in ARDS (108, 109). Analyses of the results of these trials did subsequently identify sub-phenotypes that may benefit from statin therapy. ARDS patients with hyperinflammatory sub-phenotypes had a higher survival rate with simvastatin compared to placebo (110).