Biomedicines, Vol. 10, Pages 3150: Endothelial Damage and the Microcirculation in Critical Illness

1. IntroductionPublication of research into the endothelial glycocalyx and microcirculation has increased exponentially in the last decade. Our understanding of the glycocalyx has changed from that of it being a ubiquitous, jelly-like layer to taking an active role in the interaction between the intravascular and interstitial space. The glycocalyx plays a role in chronic inflammation, diabetes, trauma, sepsis and ischaemia-reperfusion injury. The microcirculation is deranged in approximately 17–20% of a heterogenous population of patients in ICU [1,2]. Microcirculatory derangements can persist despite adequate macrocirculation corrected by vasopressors or transfusion [3]. In those that the microcirculation cannot be recruited or restored, morbidity and mortality is increased.The glycocalyx is a carbohydrate rich layer responsible for maintaining not only the oncotic pressure and barrier function within the circulatory system, but also antithrombotic and inflammatory signalling functions. It covers the luminal surface of endothelial cells throughout the vascular network, but changes between tissues depending on its primary function [4].Made up of proteoglycans and glycosaminoglycans, the glycocalyx creates a strong negative charge within the blood vessels to repel large molecules from escaping into the tissues. In vitro experiments have shown that the damaged glycocalyx becomes sensitised to atherogenic and inflammatory mediators. This sensitisation causes release of chemotactic molecules and by increasing the production of inflammatory mediators, it precipitates a cytokine storm. Circulating components of the inflammasome bind to glycocalyx receptors in a cycle of increasing inflammation. This leads to widening of gap junctions and relaxing of the barrier between the lumen and the interstitium, marked by shedding of glycocalyx molecules. The functionality of the endothelium is dependent on remaining intact with all relevant sidechains and molecular structures in place. Glycocalyx shedding and deterioration leads to loss of endothelial integrity and function [5,6]. Multiorgan dysfunction syndrome including encephalopathy, acute kidney injury, acute liver injury, coagulopathy and acute respiratory distress syndrome can all be associated with endothelial damage. Mediators released in sepsis act on the glycocalyx to produce a global response. By assessing the microcirculation in patients with sepsis, we can better understand the changes that occur in response to insults such as endotoxin or oxidative stress and link these back to markers of endothelial damage. The microcirculation has been described as the largest organ in the body, comprising the capillaries and venules 7]. Sidestream dark field (SDF) imaging has been assessed as a potential prognostic tool to guide therapy. New technologies are emerging that would allow clinicians to directly examine the EG, a potential huge step forward in personalised medicine and point-of care diagnostics.

This narrative review explores the relationship between endothelial damage, how the glycocalyx relates to the clinically observable microcirculation and how we can use this connection to improve patient outcomes. Personalised medicine revolves around our ability to treat each patient and their unique pathology or phenotype individually.

Endothelial Glycocalyx in Clinical PracticeThe loss of glycocalyx function, defence and configuration impairs vessel mechano-transduction, platelet and leucocyte adhesion to the endothelial surface and causes invasion of the vascular compartment with fluid and plasma proteins [8]. At the level of myocardium, it was found that endothelial leak was responsible for swelling in the subendothelial space, resulting in the compression of the capillary lumen and leading to oedema and myocardial dysfunction [9]. Continuation of inflammation increases availability of leucocytes to adhesion molecules by attacking the surrounding EG. Inflammatory mediators directly influence the glycocalyx and its constituents and adjust the structure. Degranulation of activated inflammatory mediators such as mast cells and macrophages release reactive oxygen species (ROS), reactive nitrogen species (RNS) that also participate to the degradation of the EG [10]. Neutrophils are the most abundant circulating cells in the human body and release proteases that damage the glycocalyx also.Glycocalyx dysfunction can occur in any organ and so can be recognised in several clinical conditions. One of the first syndromes that had recognised glycocalyx damage was diabetes. The first studies that quantified the glycocalyx found that patients with diabetes type 1 had a reduced volume of glycocalyx by 500 mL, compared to healthy subjects [11]. Glycocalyx involvement has been found in cardiovascular disease, including hypertension, stroke and left ventricular remodelling after myocardial injury, as well as cancer, renal failure, diabetes, obesity, cognitive impairment, pre-eclampsia, advanced age and COVID-19 [12,13,14,15,16,17,18,19,20,21,22]. An analysis of the ProCESS trial patients showed that elevated markers of glycocalyx damage in blood, angiopoietin-2 (Ang-2), vascular endothelial growth factor-1 and -2 (VEGF) and soluble fms-like tyrosine kinase (sFLT-1) were associated with increased 60-day in hospital mortality at baseline and at timepoints 6 and 24 h [23]. The widespread pathological effects of SARS-CoV-2 infection across various organ systems made a strong case for a glycocalyx driven disease. Before widespread effective vaccination campaigns many infected patients required hospitalisation and up to 43% who required invasive mechanical ventilation, after failure of non-invasive ventilation, and ICU support would die [24,25]. Involvement of the angiotensin converting enzyme-2 (ACE-2) receptor, the prevalence of systemic microthrombi, and large vessel thromboembolic phenomena suggested a vascular pathology. The presence of this receptor throughout not only the pulmonary epithelia but the renal, vascular endothelium and arterial smooth muscle cells can explain these features [26]. Patients with COVID-19 had higher serum concentrations of the glycocalyx marker syndecan-1 but had improved microcirculation at Day-2 of admission than non-COVID sepsis patients in one observational study of 28 ICU patients [27]. The authors concluded that the worse glycocalyx damage with conserved microcirculation could represent a new sub-phenotype of septic shock with endothelial remodelling. There was also evidence of persistent endothelial damage months after infection that was attributed to oxidative stress, endothelial and vascular dysfunction [22]. The MYSTIC study demonstrated not only higher circulating plasma markers of endothelial damage but also reduced small capillary density and an association of increased perfused boundary region (PBR), glycocalyx damage and outcome [28]. The authors showed that investigational biomarkers of glycocalyx damage ADAMST-13 and VEGF were better correlated with outcomes than CRP and IL-6. Although this study included only a small number of patients and should be considered as hypothesis generating, the results are compelling. A large trial comparing moderate, severe COVID-19 and sepsis and septic shock ICU patients would be intriguing, though possibly no longer feasible post-vaccine. The multisystem inflammatory syndrome in children (MIS-C) frequently associated with shock emphasises the multisystem nature of the disease [29]. Endothelial involvement leads to disease sequelae in almost every organ [22]. The cardinal features of MIS-C are hyperinflammation and cytokine storm, features also recognised in the adult illness [30]. Distributive shock results from endothelitis and systemic capillary leak while there is also potential cardiogenic shock through myocardial oedema [22,31]. In a study including COVID-19 paediatric patients [29], the authors found a significant negative correlation between left ventricular ejection fraction (LVEF) and Ang-2 (p = 0.01). Varga et al. also found extensive endothelial cell involvement with macrophage activation, capillary leak and micro-thrombosis [32]. 2. Measuring the GlycocalyxDespite our increasing knowledge about the EG, it remains remarkably difficult to assess. The glycocalyx is composed of sugar and proteins that are reactive with many common laboratory fixation methods [33]. It was first visualised by staining with ruthenium red, a substance with high affinity for the acidic mucopolysaccharides, generating detectable electron density visible with an electron microscope [33,34]. Ruthenium red however is a relatively large molecule and there were concerns that its charge induces conformational change in the EG, leading to inaccurate characterisation of the glycocalyx structure [33]. Efforts were made with smaller molecular dyes (alcian blue) but other techniques were developed as classic perfusion fixation was possibly removing side chains and structural components of the system being examined.The components of the glycocalyx are constantly being generated and shed, so damage to the glycocalyx can be assessed by measuring the concentrations of circulating endothelial components in plasma. The most reliable and widely used is syndecan-1, however levels of heparan sulfate, chondroitin sulfate, endocan and hyaluronan have also been used. Syndecan-1 is a member of the family core glycocalyx proteoglycans varying from 25–40 kDa that is measured by ELISA. They have a single-span transmembrane domain connecting to the cell membrane. Syndecans have 4 subtypes that each binds a different sidechains, either 3–5 chains of heparan sulfate or chondroitin sulfate [35]. As the syndecan sidechains are shed from the endothelium they can be measured in circulation.Soluble shed portions of syndecans can be used as biomarkers as the process of shedding is specifically regulated under disease conditions [36]. Leukocyte-derived proteases and growth factors, associated with cellular injury or wound healing, can initiate shedding [36]. Thus, shed syndecans are found in inflammatory fluid and associated with tissue damage in a variety of disease and critical illness. During inflammation the total expression of syndecans is increased [37,38]. SDC1 plays an important role in leukocyte adhesion, vascular permeability and mechanosensation [39]. It has been studied as a biomarker in a wide range of diseases including kidney disease, heart failure and as an indicator after major surgery [40]. Soluble SDC1 is found in the peripheral blood of patients with sepsis, ischemia-reperfusion injury and graft-versus-host disease [41,42,43,44]. SDC2 plays a role in endothelial damage and vascular dysfunction when endothelial cells are damaged [45]. Inflammatory signals such as hypoxia and TNF-α increase expression of SDC2 in fibroblasts, endothelial cells and intestinal epithelia [46,47]. SDC3 is the largest of the syndecans but is the least studied and understood. It has been implicated in alzheimer’s disease, human immunodeficiency virus-1 (HIV) disease, angiogenesis and arthritis [48]. Cleaved portions of SDC3 disassemble endothelial cell junctions in the lung which has implications for sepsis and diseases where thrombin is activated [49]. Knockout experiments show that lack of SDC1 or SDC4 increases the inflammatory response, possibly indicating an anti-inflammatory role as well [50,51]. SDC4 is involved the development of fibrosis in the lung during inflammation [52,53]. Levels of SDC4 increase in acute pneumonia and correlate with pneumonia severity, indicating it could be a useful biomarker in these patients [54]. Glycosaminoglycans are disaccharide polymers of L-iduronic acid, D-glucuronic acid or D-galactose linked to either D-N-acetyl galactosamine or D-N-acetylglucosamine [35]. Proteoglycans, mainly heparan sulfate, provide abundant binding sites for circulating mediators courtesy of their various sulfation combinations [55]. Heparan sulfate also performs vital antioxidant function binding superoxide dismutase to protect the glycocalyx from oxidative stress. This mechanism is challenged in sepsis and septic shock, leading to glycocalyx damage and extravasation of plasma proteins and fluid into the subendothelial layer (Figure 1). Reduced concentrations of heparan sulfate in serum subsequent to exposure to damaging enzymes increase coagulation and micro-thrombosis, increase adhesion molecule expression and increase leucocyte tracking along the glycocalyx [35]. In an observational study of 38 patients, blocking heparanase, an enzyme that targets heparan sulfate, with heparin eliminated glycocalyx damage in vitro [56]. Hyaluronan is attached to the cell surface via CD-44 receptor; it is not a core protein but contributes to the glycocalyx volume by its length and by binding water ~10,000 times its mass [5].A prominent drawback of using plasma measurements is their dependence on renal clearance, which can be altered in critical illness, impacting on reliability of these tests [57,58]. Another drawback is that chronic inflammatory state also leads to increase in circulating endothelial components [6,10,35,59,60]. Metabolic, vascular and surgical diseases such as diabetes, atherosclerosis, hypertension, ischaemia reperfusion injury and trauma result in increased numbers of plasma glycosaminoglycans that correlate to inflammatory marker serum concentrations. Other biomarkers for endothelial damage include hyaluronic acid, angiopoietin-2, VEGF and vonWillebrand Factor cleaving protease ADAMTS-13, soluble thrombomodulin and soluble angiopoietin receptor (TIE-2) [28,61,62]. Ang-1 and Ang-2 are in opposition to one another, their action on the glycocalyx being mediated by the TIE-2 receptor. Ang-2 is the leakage inducing form and is raised in systemic inflammatory syndromes, indicating glycocalyx damage [63]. In vitro studies on human sepsis sera showed that the TIE-2 pathway regulates the glycocalyx in sepsis in a non-redundant fashion. When endothelial cells were incubated with sepsis serum and TIE-2 pathway inhibitors, the damage to glycocalyx was prevented [63,64]. 3. Visualising the Microcirculation and the Endothelial GlycocalyxFollowing observations that 40 kDa dextrans equilibrate with the EG, efforts to visualise and quantify the glycocalyx began by comparing dilution of fluorescently labelled RBCs to dilution of 40 kDa dextran at the time of injection [65]. Studies of the glycocalyx in cremaster muscle of mice found that the glycocalyx repelled RBCs and slowed plasma while being compressed by passing leucocytes it serves as both a barrier and a gateway to the tissues [65,66]. Visualising the glycocalyx in vivo and how it behaves in clinical practice has become more important as we come to understand its importance. The development of intra-vital microscopic techniques has transformed this area of practice. Developed to examine the movement of RBCs within the circulation, Orthogonal Polarisation Spectroscopy (OPS) allowed clinicians to have a view of the microcirculation in clinical practice. The most recent iterations of this technology—SDF and Incident Dark Field (IDF) imaging, have improved the clinical applicability of the microcirculation. The implications of damage to the microcirculation in a variety of diseases in ICU has been studied since these devices have been available [67,68,69].As techniques have improved, our field of view has grown. The most recent descendant of the OPS devices, the IDF microscope has an increased field of view and improved contrast to better identify cells and perfused capillaries [70]. In the past, the image would be manually divided into sections and the boundaries of vessels individually marked out and perfused vessels counted individually. This is presently done objectively by a piece of software, AVA (Microvision Medical, Amsterdam, The Netherlands). Similarly, because the device uses 540 nm light in a dark field created by circumferential light-emitting diodes (LEDs), it highlights the RBCs themselves [71]. While this gives excellent information about availability of haemoglobin and functional capillary perfusion, the glycocalyx that controls the flow remains invisible.

The PBR is the area at the limit of a blood vessel where RBCs can permeate, representing the luminal aspect of the glycocalyx accessed by the RBCs in circulation. It is quantified by observing the microscopic lateral motion of the cells under SDF microscopy in combination with proprietary Glycocheck™ 5.2 software (Capiscope handheld, KK Research technology Ltd., Honiton, UK). If the glycocalyx is shed or disturbed, the lateral motion of RBCs increases so PBR has an inverse relationship to glycocalyx thickness.

The GlycoCheck™ system makes it possible to calculate the degree of lateral motion of RBCs within small capillaries [72]. The reliability of this system has been established both due to its interobserver consistency and accessibility to all clinical staff as a potential standard monitoring tool [14,73]. The success of the GlycoNurse study established the system’s potential to bring the microcirculation from the research realm into daily clinical practice in a busy Emergency department environment [74]. The GlycoCheck system is a great leap forward from other in vivo glycocalyx measurement methods such as atomic force microscopy and microparticle image velocimetry (μ-PIV), used on animal models in laboratory conditions. The PBR may be elevated in microvascular thrombosis, inflammation or sepsis, and it has been used to visualise the glycocalyx in vivo [75]. The NOSTRADAMUS study used RBC velocity measurements together with PBR thickness to improve discrimination between patients with sepsis and healthy controls [76]. This study showed that the PBR tends to increase when the velocity of RBCs decreases, indicating increased permeability and porosity of the glycocalyx in an environment of reduced shear stress. 3.1. The Sublingual Target Region

The sublingual region is most commonly used area to study the microcirculation because of the proximity to the lingual artery as a branch of the external carotid artery, giving the clinician insight into the reactivity of the central circulation. However, other vascular beds such as the intestinal bed, renal bed, conjunctival and peripheral muscular microvasculature have also been used to study the microcirculation.

The sublingual region is the most clinically accessible however, its reliability relies on how representative it is of all vascular beds. In a pig model, where septic cholangitis was induced by Escherichia coli into the common bile duct, OPS imaging of the intestine and the sublingual region correlated well in timing and specific observable microcirculation changes [77]. A prospective study of patients with sepsis after formation of an intestinal stoma correlated OPS images from within the stoma with sublingual images [78]. This study found no relationship between the two regions on postoperative day 1, but the relationship normalised by day 3. MFI in the stoma of the sepsis group was significantly lower than healthy controls and the non-septic new stoma group. Sublingual region MFI at day 1 correlated well with macrohaemodynamic measures such as sequential organ failure assessment (SOFA) and length of stay, this relationship was not significantly related on day 3. However, another clinical observational study of postoperative ostomy patients before and after fluid challenge found dissociation of the intestinal and sublingual microvascular beds [79]. In response to a fluid challenge on the first postoperative day, the sublingual but not the intestinal microcirculation showed increased RBC velocity. This study did not perform follow up imaging to see if this dissociation resolved or persisted. A study of patients undergoing gastrointestinal surgery were assessed by SDF imaging of their bowel and sublingual region intraoperatively. Studying the sublingual region allowed for more stable image acquisition, less pixel loss and faster image acquisition [80]. There was good correlation of MFI, PVD and TVD between sublingual and gastrointestinal microcirculation. 3.2. Near Infrared Spectroscopy and the MicrocirculationNear infrared spectroscopy (NIRS) is a non-invasive tool that measures microvascular reactivity by oxygenation in muscle, commonly the deltoid or thenar eminence. Studies on patients with sepsis have associated low thenar eminence saturations with poor outcome in sepsis. Using a vascular occlusion test in the forearm, microcirculation reactivity can be assessed by analysis of tissue saturations changes during an ischaemic challenge. This illustrates oxygen extraction by tissues and reactivity of the microvascular bed. A meta-analysis of static and dynamic NIRS and mortality in sepsis found that septic patients had lower tissue saturations, decreased reperfusion slope and lower reperfusion hyperaemic maximum tissue saturation. These results were also associated with higher mortality in septic patients [81]. A prospective study of patients in septic shock found an association between septic shock and lower initial tissue saturations, impaired occlusion slope and recovery slope, implying microcirculation dysfunction. SOFA score was also associated with recovery slope in the septic shock cohort with AUC 0.81, meaning NIRS could be an interesting non-invasive prognostic monitor in the future [82]. Moreover, other studies ICU found microcirculation failure measured by NIRS predicted mortality [83]. 5. Haemodynamic Coherence and Personalised Treatment in ICURestoration of macro-haemodynamic stability does not reliably re-establish the microcirculation [111]. This has been dubbed haemodynamic coherence and ICU research and resuscitation should aim to understand and improve it [87,112]. A physiological state where despite the gross improvement of macrohaemodynamic markers such as blood pressure and heart rate, the microcirculation remains impaired [87]. ICU resuscitation relies on appropriate restoration of cellular respiration. Haemodynamic coherence represents the potential downfall of many large trials of heterogeneous groups of ICU patients [113].Reclassification of acute respiratory distress syndrome (ARDS) biological and clinical phenotypes has increased prognostic and predictive enrichment by defining homogenous groups within this particular disease [114]. By recognising separate cohorts within large heterogeneous groups, treatments can be targeted at those that will benefit most. Those at increased risk of a particular adverse outcome may be more likely to benefit from a certain intervention, increasing a study’s power or a biologically homogenous group may be more likely to benefit from an intervention targeting a specific biological mechanism. For example, the PaO2:FiO2 ratio 115,116]. Similarly, by recognising fluid responders and non-responders, treatments for sepsis can be studied more effectively.Recognising the changes in the microcirculation in different pathological states could help to identify homogenous patient cohorts [67,117,118]. Studies of the microcirculation response to RBC transfusion have shown a heterogeneous response of groups of patients clinically diagnosed as sepsis or septic shock [119]. These results indicate the existence of subsets of microcirculation changes that may respond differently to therapies. Previous studies have shown that despite individual haemodynamic incoherence, sepsis induced dysfunction of the microcirculation can recover following resuscitation of arterial pressure. 6. Prognostic Value of Glycocalyx Damage in Critical IllnessThe connection between glycocalyx degradation, microvascular parameters and systemic clinical markers has been difficult to identify. In non-septic ICU patients only a weak correlation could be found between syndecan-1 and the glycocalyx thickness measured in the sublingual region [120]. Rovas et al. found that PBR, MFI and PPV correlated with measures of critical illness including mean arterial pressure, CRP, IL-6 and procalcitonin (PCT) [121]. They also found an association with systemic inflammatory response (SIRS) and SOFA score. The interest of this study was to attempt to draw together disparate prognostic indicators and to associate bedside microcirculation assessment with glycocalyx function. However, another study showed that PBR and syndecan-1 serum concentrations did not correlate with microcirculation variables. The NOSTRADAMUS study attempted to link the macro and microcirculation by suggesting the Microvascular Health Score (MVHS). The MVHS depends on the correlation Rovas et al. found between flow dependent capillary density and SOFA [76]. This study used RBC velocity measurements together with PBR thickness to improve discrimination between patients with sepsis and healthy controls [76]. One of the largest biomarker trials conducted was the Protocolized Care for Early Septic Shock (ProCESS) randomised controlled multicentre trial [122]. An analysis of 1341 of these patients showed that elevated markers of endothelial permeability in blood, angiopoietin-2 (Ang-2), vascular endothelial growth factor-1 and -2 (VEGF) and soluble fms-like tyrosine kinase (sFLT-1) were associated with increased 60-day in hospital mortality at baseline and at timepoints 6 and 24 h [23]. Though no difference was found between the treatment groups of the trial, there was a significant difference in mortality according to baseline serum concentrations of endothelial markers. A systematic review of 17 studies investigating the relationship between markers of glycocalyx degradation and outcomes in sepsis showed that concentrations of syndecan-1 and endocan were higher in patients who died, developed MODS or experienced renal failure [123]. In a prospective study of 21 sepsis patients PBR correlated positively with plasma concentrations of Ang-2 (R = 0.52, p = 0.03) but not with APACHE, SOFA, lactate or syndecan-1 [124]. Increased endothelial permeability can be clinically detected as microalbuminuria-urinary creatinine ratio (MACR), as a result of glomerular inflammatory injury. MACR is an early marker of sepsis and a marker of severity that correlates with Acute Physiological Score II (APACHE), SOFA, Simplified Acute Physiology Score II (SAPS) [125,126,127]. 8. Future DirectionsThe concept of personalised medicine arises from clinical enrichment, referring to patient subgroup selection of those who are more likely to respond to particular therapy as opposed to an unselected population, was mainly developed in the field of oncology. Its success in that field led to publication of Food and Drug Administration (FDA) guidelines and a statement of intent from the Obama Whitehouse to emphasize, prioritise and pursue enrichment strategies to develop novel therapies for diseases [166,167]. Prognostic enrichment is important in the design of clinical trials, identifying those patients more likely to encounter an outcome or complication, thus increasing the power of a study and reducing the required sample size [168]. Predictive enrichment requires exact knowledge of a biological mechanism to select patients that will respond to an intervention.

Prognostic enrichment has been used in ICU research to define acute respiratory distress syndrome phenotypes, which enhanced research uncovering therapeutic strategies. However, ICU syndromes such as sepsis lack a specific biological target, precluding predictive enrichment.

Microvascular imaging of glycocalyx behaviour and response to treatment could be the biological target needed to stratify patients into clinically relevant phenotypic groups. Using bedside diagnostics and imaging techniques together with machine learning and latent class analysis, better trials could be developed to identify effective therapies for patient subclasses.

Further tests on drugs like Sulodexide, a combination of heparin-sulfate like compound shown to regenerate the glycocalyx in a mouse model of sepsis, that also restores glycocalyx volume in diabetics, should be studied further in critical care [169]. FFP has shown benefit in animal models as well as models of haemorrhagic shock but there are no high quality studies of its effects restoring the glycocalyx in critical illness [170,171].The glycocalyx spans all organs and therefore is exposed to variable rates of flow, as well having different thresholds for onset of glycocalyx damage leading to spatiotemporal uncoupling of insult and reaction. This could explain the differences in microcirculation measurements between organs seen in studies relating intestinal SDF microcirculation measurement to sublingual imaging [79]. Several studies noted the potential uncoupling of glycocalyx damage and the microcirculation variables. It is possible that the effect of glycocalyx damage undergoes a certain lag or that the recovery of the glycocalyx while bathed in septic plasma takes longer than in vitro. There could be immediate precipitation of glycocalyx change with delayed changes in perfusion, followed by prolonged repair of the glycocalyx. In studies of patients on CPB the microcirculation is affected almost immediately at the point of initiation with glycocalyx degradation markers not returning to baseline levels for 72 h [98,99]. We do not know at what point in sepsis the microcirculation becomes impaired, or similarly, how long it takes to improve. Further delineation of the relationship between the functional microcirculation and the detectable markers of glycocalyx damage could elucidate a novel therapeutic target in this syndrome.

留言 (0)

沒有登入
gif