Severe forms of COVID-19 associated with acute pulmonary failure (acute respiratory distress syndrome, COVID-19 ARDS) are a multisystemic, thrombotic and inflammatory disorder with high mortality (1, 2). Histopathological studies in COVID-19 ARDS showed wide spread endotheliitis involving multiple organs (3, 4) beside the affected lungs (5) which is in marked contrast to “classic” ARDS (6). This pathophysiologic difference is probably due to a combination of direct viral infection of endothelial cells (7) and the systemic inflammatory response of the organism (cytokine storm) (8) both resulting in vascular pathology (9, 10).

Under physiological conditions, vascular integrity is preserved by a network of glycocalyx proteins located at the interface between the endothelial cell surface and the vascular environment. Major components of the glycocalyx are hyaluronic acid and heparan sulfate. The endothelial transmembrane protein syndecan serves as connecting structure for both. Severe infection and inflammation lead to degradation of this protective layer by hyaluronidase and heparanase resulting in immunologically active hyaluronic acid and heparan sulfate fragments, both markers of glycocalyx shedding. In patients with COVID-19, elevated levels of these glycocalyx fragments were strongly associated with organ failures and increased inflammatory cytokine levels (11).

There is little information, however, regarding the mediators of intercellular communication between immune and endothelial cells and the molecular signaling cascades of glycocalyx fragmentation in COVID-19 pulmonary failure and microvascular injury.

An important part of intercellular communication are extracellular vesicles (EVs) (12). They play a major part in the regulation of immune response (13) in multiple human disorders such as pneumonia (14) or vascular pathology in general (15–17). In COVID-19, EVs are also key regulators of the immune response and disease progression (18–20).

EVs are categorized as either exosomes or ectosomes. Exosomes are of endosomal origin with a size between 40 to 160 nm whereas ectosomes are larger EVs generated by the direct outward budding of the plasma membrane with a diameter in the range between 50 nm to one µm in diameter.

EVs contain many components of their cells of origin including RNA, as well as DNA, lipids and cytosolic and cell-surface proteins (21). Mechanistic studies show that these molecules serve as biological signals transported by EVs from their cells of origin to specific target cells, underlining their important function in intercellular communication (12). Non-coding RNA - in particular - microRNAs (miRNAs) - as EV components are taken up by target cells where they regulate gene transcription.

In this study, we measured the circulating glycocalyx elements hyaluronic acid and heparan sulfate and its connecting transmembrane protein syndecan-1 along with the von Willebrand factor-cleaving protease ADAMTS13 in hospitalized patients with COVID-19 disease and identified molecular signaling networks targeted by these glycocalyx components and ADAMTS13. Network analysis was based on data from high-throughput RNA sequencing (RNAseq) of mRNA isolated from blood cells and miRNAs derived from EVs as mediators of intercellular communication between blood, endothelial cells and immune cells. Our aim was to characterize the role of circulating glycocalyx components in immune and endothelial cell activation in COVID-19.

Figure 1 gives a summarized overview of the methodological approach used in this study.

Severely affected COVID-19 patients are in a state of endothelial injury with widespread microthrombosis coupled with hyperactivation of the immune system (10). Immune cell activation is reflected by high levels of IL-6, IL-1β, IL-18 known as the COVID-19 cytokine storm (42). High concentrations of pro-inflammatory cytokines in-turn lead to endothelial cell activation. The reduced levels of ADAMTS13 as seen in our study result in high concentrations of the endothelial adhesion protein von Willebrand Factor on the “naked” endothelial cell surface along with loss of its protective glycocalyx layer. This leads to platelet wall adhesion with microangiopathy and microthrombosis (43).

Less investigated, however, was the direct effect of circulating glycocalyx components and the decrease in ADAMTS13 on molecular signaling in immune and endothelial cells. We addressed this research question in our study and showed that circulating glycocalyx components during SARS-CoV-2 infection are not biologically inactive markers of severe microvascular injury but have distinct feedback effects on signaling networks in immune and endothelial cells resulting in immune and endothelial cell activation with an increase in inflammation and further enhancement of microthrombosis formation.

The construction of glycocalyx component signaling networks in our study was based on RNA expression levels determined by high-throughput sequencing of miRNAs from EVs and mRNA transcripts from blood cells. Extracellular vesicles as a source for miRNAs were selected because our group and others have demonstrated their important role in regulating the immune response and the development of immunothrombosis (18, 20). In addition, platelets release extracellular vesicles with procoagulant activity in COVID-19 (44). miRNAs associated with extracellular vesicles can regulate mRNAs transcription in target cells during systemic inflammation and in vascular disorders such as atherosclerosis (17, 45, 46).

When comparing plasma concentrations of syndecan-1 and hyaluronic acid in COVID-19 pneumonia to healthy controls, both glycocalyx components were significantly elevated ( Figure 2 ). The corresponding molecular signaling networks demonstrated that several important mRNA transcripts are regulated by these glycocalyx components in immune and endothelial cells. Syndecan-1 interacts directly with ITGB3 (CD61 or integrin beta 3) ( Figure 3 ). Additionally, ITGB3 is activated by downregulated miR-32-5p. In COVID-19, ITGB3 is indicative of a large number of megakaryocytes in the pulmonary circulation (47). Megakaryocytes as the source of platelets (48) lead to an increase risk for platelet activated thrombus formation. Furthermore, ACE2 expression (angiotensin−converting enzyme 2) as the key host protein of COVID-19, correlates with expression levels of ITGB3 in post-mortem lung tissues obtained from patients who died from COVID-19 ARDS (49). In our study, expression in extracellular vesicles is downregulated in COVID-19 patients in comparison to healthy controls. Downregulation of miR-32-5p showed potential activating effects on ACE2 expression in a recent in-silico investigation (50). In combination with signaling effects of syndecan-1, a lower expression of miR-32-5p could therefore be associated with immune cell activation, ACE2 mediated virus entry and pulmonary microthrombi formation.

A further important finding in the COVID-19 pneumonia network was the direct interaction between hyaluronic acid with TLR4 (toll-like receptor 4) (51) ( Figure 3 ). TLR4 activation results in increased ACE2 expression on the cellular surface of circulating monocytes leading to increased susceptibility to infection with SARS-CoV-2. Extracellular vesicles (exosomes) were found as a source of ACE2 in this context (52). Blocking of TLR4 signaling has been suggested as a means of limiting pulmonary injury in viral infections (53).

The significant increase in syndecan-1 and hyaluronic acid, the decrease in ADAMTS13 and the associated activation of procoagulant and inflammatory pathways seen in COVID-19 pneumonia indicates that patients with this clinically milder variety of COVID-19 disease are also at risk of microvascular complications.

As a next step, we constructed glycocalyx component and ADAMTS13 signaling networks which compared patients with COVID-19 ARDS - the most severe form of the disease - to healthy controls. In this analysis, all glycocalyx components are significantly higher and ADAMTS13 levels are even lower than in COVID-19 pneumonia, a reflection of the progress in disease severity towards pulmonary failure. In the resulting molecular network, the activating effect of heparan sulfate on TLR4 is maintained and, in addition, this proteoglycan is associated with an increase in VEGFA signaling as heparan sulfate is an endogenous agonist on VEGFA and TLR4 (54–56). VEGFA acts as a potent inductor of vascular leakage (55), a hallmark of pulmonary injury leading to ARDS. The signaling network involving TLR4 and VEGFA in COVID-19 ARDS is more complex than in pneumonia and shows additional activating effects of downregulated miR-150-5p and miR-126-3p on VEGFA. Downregulation of miR-150-5p is associated with severe COVID-19 disease (57). miR-126-3p amongst other miRNAs differentiated between survivors and non-survivors in patients with COVID-19 ARDS in an earlier study (58). Another interesting connection in this network is the binding of heparan sulfate to downregulated NRP1 (neurotropin 1 or neuropilin-1), an antagonist of VEGF (59). Lower expression of NRP1 could result in even higher activity of VEGF, further enhanced by additional downregulation of NRP1 through upregulated miR-1-3p derived from extracellular vesicles in our data ( Figure 4 ). Upregulated miR-1-3p in bronchial aspirates also distinguished between survivors and non-survivors with COVID-19 ARDS (58).

This analysis was followed by comparing patients with pneumonia to those who had developed severe ARDS. Syndecan-1 and heparan sulfate plasma concentrations were significantly higher and ADAMT13 levels significantly lower than seen in COVID-19 pneumonia ( Figure 2 ) but no direct regulatory effects of these molecules in the network could be identified. In addition, this network ( Figure 5 ) was smaller and more fragmented than the signaling cascades comparing the healthy state to pneumonia and ARDS. This indicates that many pathophysiologic changes occur during the early phase of the disease when COVID-19 has not yet progressed to pulmonary failure. In the network, ADAMTS13 was targeted by downregulated miR-1228-5p and miR-4433b-5p ( Figure 5 ). As ADAMTS13 levels were already low, counter regulatory effects of these miRNAs could explain this finding; the inhibition of ADAMTS13 most likely occurs by other mechanisms not reflected by the network. It is of interest to note, however, that miR-1228-5p was upregulated in patients with COVID-19 pneumonia ( s. Table E 2 in the online supplement) in contrast to the comparison between COVID-19 pneumonia to ARDS where this EV-derived miRNA was downregulated ( Figure 5 ). This points to the possibility that miR-1228-5p might be involved in an activation of ADAMTS-13 in COVID-19 ARDS. miR-1228-5p also upregulated CARD6 (caspase recruitment domain family, member 6) which plays a role in immune defense, but the exact pathophysiologic function of CARD6 is poorly defined (60). Nevertheless, CARD6 is highly expressed in neutrophils, endothelial progenitor and venous endothelial cells ( s. Figure E 2 in the supplement ). miR-4433b-5p upregulated PKM (pyruvate kinase isozyme). Pyruvate kinase is responsible for net ATP production within the glycolytic sequence. In contrast to mitochondrial respiration, energy regeneration by pyruvate kinase is independent from oxygen and allows survival of organs under hypoxic conditions, which are present in ARDS. The highest expression of PKM in COVID-19 ARDS patients is found in circulating mono-CD14⁺ cells (61).

Interestingly, the subgroup of patients in our study originally admitted with pneumonia who later progressed to ARDS showed significantly higher levels of hyaluronic acid, while hyaluronic acid plasma levels of all pneumonia patients were not statistically significantly different to ARDS patients.

A limitation of our findings results from the fact that our study is missing a proteomics analysis of blood and additional cellular studies of blood and endothelial cells but our findings can form the basis for further proteomic and cell-based analyses of glycocalyx signaling networks and suggests a novel mechanism for immune modulation by glycocalyx components involving EV-derived miRNAs. Furthermore, the regulatory consequences of glycocalyx components and ADAMTS13 on immune and endothelial cells identified by the networks were not directly experimentally confirmed in our data set. The canonical pathways identified in our study were based on the Knowledge Base item underlying the IPA^® software. This Knowledge Base is created by millions of manually curated data obtained from scientific journals, publicly available molecular content databases, textbooks and more and allows the query, visualization and computation across the Knowledge Base in relationship to the researchers own dataset of miRNA, mRNA and protein findings uploaded into IPA^® (62). This holistic approach has been applied to COVID-19 in an earlier study (63) and results in networks which allow the identification of signaling cascades not previously studied. As a disadvantage, the regulatory effects of glycocalyx components in our networks could not explicitly be proven as causal and some of the interactions could still be correlational. However, most regulatory effects underlying the network connections were experimentally confirmed in earlier experimental studies or showed at least a high biologic plausibility.

Circulating glycocalyx components in COVID-19 are not harmless byproducts of microvascular injury but result in further activation of immune and endothelial cells and may fuel a positive feedback loop resulting in immune activation and more severe systemic inflammation leading to microcirculatory failure. These effects already appear during the early phase of COVID-19 when patients present with pneumonia prior to progression to pulmonary failure and ARDS. The timely identification of patients at risk for ARDS is therefore critical and glycocalyx component levels may serve as biomarkers for this negative outcome. A therapeutic consequence for COVID-19 patients with early evidence of microvascular injury would then be the targeted use of antiviral agents (e.g., Ritonavir-boosted nirmatrelvir or Remdesivir) or the timely administration of pharmacologic compounds that protect the glycocalyx layer (10).

The studies involving human participants were reviewed and approved by Ethics Committee of the Medical Faculty of the Ludwig‐Maximilians‐University of Munich. The patients/participants provided their written informed consent to participate in this study.

All authors contributed to the study conception and design. Material preparation and data collection was performed by MB, FB, AM and MK. ELISA-was performed by MB. Molecular analysis was executed by MR and BK. Data analysis was done by AM, MB, FB, BK and J-NB. AM, MR, MP and GS worked on the manuscript. All authors commented on previous versions of the manuscript. All authors contributed to the article and approved the submitted version.