The blood-gas barrier in COVID-19: an overview of the effects of SARS-CoV-2 infection on the alveolar epithelial and endothelial cells of the lung

ABSTRACT

Blood-gas barrier (BGB) or alveolar-capillary barrier is the primary tissue barrier affected by coronavirus disease 2019 (COVID-19). Comprising alveolar epithelial cells (AECs), endothelial cells (ECs) and the extracellular matrix (ECM) in between, the BGB is damaged following the action of multiple pro-inflammatory cytokines during acute inflammation. The infection of AECs and ECs with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the pathogen behind COVID-19, triggers an inflammatory response at the BGB, inducing the release of interleukin 1 (IL-1), IL-6, tumor necrosis factor alpha (TNF-α), transforming growth factor beta (TGF-β), high mobility group box 1 (HMGB1), matrix metalloproteinases (MMPs), intercellular adhesion molecule-1 (ICAM-1) and platelet activating factor (PAF). The end result is the disassembly of adherens junctions (AJs) and tight junctions (TJs) in both AECs and ECs, AEC hyperplasia, EC pyroptosis, ECM remodeling and deposition of fibrin clots in the alveolar capillaries, leading to disintegration and thickening of the BGB, and ultimately, hypoxia. This commentary seeks to provide a brief account of how the BGB might become affected in COVID-19.

COVID-19 and the blood-gas barrier

Consisting of tightly connected ‘alveolar epithelial cells’ (AECs), capillary ‘endothelial cells’ (ECs) and the ‘extracellular matrix’ (ECM) in between, the ‘blood-gas barrier’ (BGB) or ‘alveolar-capillary barrier’ (ACB) is a highly specialized wall spanning throughout the pulmonary alveoli, separating the inhaled air from the internal environment of the body. Damage to this extremely delicate wall¹, which is only 0.2 to 0.3 μm thick², might occur as a result of the inflammation caused by lower respiratory tract infections such as ‘coronavirus disease 2019ʹ (COVID-19)³. An acute inflammatory process, mediated chiefly by macrophage-derived cytokines like ‘interleukin-1ʹ (IL-1), IL-6 and ‘tumor necrosis factor alpha’ (TNF-α), can adversely affect the components of the BGB (Figure 1), and result in ‘acute respiratory distress syndrome’ (ARDS); a life-threatening condition with a relatively high mortality due to impaired gas exchange across the BGB¹.

Figure 1.

A flowchart illustrating the cause-and-effect relationship between several well-known mediators and markers of inflammation in COVID-19. All arrows represent cause-and-effect relationship and the difference in color is to distinguish pathways spanning multiple columns. Plain connectors, without an arrowhead, are used to connect fields with the same effector. Abbreviations used in the image: SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2), AEC (alveolar epithelial cell), ECM (extracellular matrix), EC (endothelial cell), IL-1β (interleukin 1 beta), IL-6 (interleukin 6), TNF-α (tumor necrosis factor alpha), TGF-β (transforming growth factor beta), STAT3 (signal transducer and activator of transcription 3), CLDN (claudin), ZO-1 (zonula occludens-1), EMT (epithelial–mesenchymal transition), MMP (matrix metalloproteinase), HGMB1 (high mobility group box 1), PAF (platelet activating factor), AJs (adherens junctions), TJs (tight junctions), ICAM-1 (intercellular adhesion molecule-1), EndMT (endothelial–mesenchymal transition)

Hypoxemia, defined as an oxygen saturation below 90%, is a common early clinical finding in patients with COVID-19. Low oxygen saturation in the absence of gross pulmonary pathology or respiratory dysfunction, as observed with COVID-19, is strongly indicative of an ongoing destructive process at the BGB; which mediates the exchange of O₂ and CO₂ in an otherwise healthy individual. It is suggested that the COVID-19-associated hypoxemia might be due to a dysfunctional BGB⁴ characterized by endothelial injury⁵ and extensive distortions in the pulmonary vascular bed⁶. These distortions could be attributed to the disintegration of ECM proteins and disruption of intercellular junctions that might result in the loosening of cell–matrix and cell–cell contact, respectively⁷. Endothelial injury, on the other hand, was suggested to be strongly associated with increased plasma levels of ‘intercellular adhesion molecule-1ʹ (ICAM-1) in patients with COVID-19 (Figure 1). Upregulated in response to inflammation, ICAM-1 is a marker of endothelial injury known to induce the release of pro-inflammatory cytokines in the alveoli⁵.

SARS-CoV-2 and its target cells in the BGB

The damage inflicted to the BGB throughout the course of COVID-19, however, is not solely driven through the pro-inflammatory cytokines produced by the migrating neutrophils, since the cells comprising the BGB might also be pathologically affected by ‘severe acute respiratory syndrome coronavirus 2ʹ (SARS-CoV-2); the pathogen behind COVID-19. The present hypothesis is supported by several reports^8,9,10 indicating the expression of ‘angiotensin-converting enzyme, in both AECs⁸, and ECs of the BGB^9,10. In addition to the ECs forming the capillaries responsible for gas exchange, ACE is also expressed in the arterioles and venules immediately preceding and succeeding these capillaries, and the smooth muscle cells residing in the vascular wall⁹.

Aside from ACE2, several other molecules are thought to be expressed by AECs, most notably ‘type II AECs’ (AECIIs), that might facilitate the entry of SARS-CoV-2 into these cells. Of particular interest are ‘transmembrane serine protease 2ʹ (TMPRSS2) and ‘tripartite motif-containing 28ʹ (TRIM28). While the former is a serine protease suggested to prompt the direct cytopathic effects of SARS-CoV-2 by facilitating the viral entry through the activation of Spike protein⁸, TRIM28 is a transcriptional factor¹¹ presumed to mediate both direct and indirect cellular damage associated with COVID-198⁸ through the direct regulation of type I interferon response¹¹ and subsequent indirect upregulation of ACE2. Both molecules are reported to be co-expressed with ACE2 by a certain, albeit low, proportion of the AECIIs⁸.

SARS-CoV-2 targets the EC-neighboring pericytes

Located at the basement membrane underlying the ECs of the arterioles, capillaries and post-capillary venules, pericytes (PCs) are fibroblast-like cells belonging to the connective tissue that cover variable portions (22–99%) of the basolateral surface of ECs at certain intervals¹². The significance of PCs lies within their secretory and contractile function, enabling them to maintain a crosstalk with the ECs, while influencing the luminal caliber of the vessels¹³. The crosstalk between PCs and ECs is mostly mediated through the interaction of ‘angiopoietin ligands’ (ANGPT1/2) with TIE receptors, of which the TIE-2 variant is more predominantly involved. This particular type of cell–cell signaling is thought to improve the survival of ECs, and counteract inflammation-induced vascular leakage¹⁴ by regulating the expression of tight junction proteins in ECs¹⁵. Accordingly, the loss of PCs could contribute to the insidious distortion of capillaries observed in COVID-19 patients¹⁶. SARS-CoV-2 may directly target PCs as well, since these cells express ACE2 on their surface, in a way similar to ECs.^14,16 The resulting infection might prompt the apoptosis of PCs, which in turn, can result in decreased number of PCs, and become manifest in the form of the micro-vasculopathy reported in COVID-19 patients¹⁶.

COVID-19 and the disassembly of intercellular junctions

The stability of a barrier is ensured by the appropriate connection of its components to one another. In the case of BGB, this integral cell–cell connection is established by the consistent distribution of junctional proteins, among AECs¹⁷, that together form the ‘apical junctional complexes’ (AJCs). Adherens junctions (AJs) and tight junctions (TJs) are the two key constituents of the AJCs¹⁸.

Near 80% of the alveolar epithelium, in the adult lung, is composed of specialized cells termed ‘type I AECs’ (AECIs) that mediate the exchange of gases. The remaining portion consists mostly of AECIIs. Given the significantly larger surface area of AECIs in the alveoli, it can easily be assumed that these cells serve as the primary epithelial component of the BGB. Thus, AJCs between the immediately adjacent AECIs, particularly TJs, could be considered the chief regulators of the alveolar epithelial barrier permeability¹⁹.

TJs comprise a variety of transmembrane proteins, the most well recognized of which are occludin, ‘claudins’ (CLDNs) and ‘junction adhesion molecules’ (JAMs). The intercellular connections established by these transmembrane proteins are further strengthened through numerous associations with peripheral membrane proteins such as ‘zonula occludens’ (ZO). Beside the TJs nearing the apical surface of the cell, AJs contribute majorly to the structure of the AJCs as well, albeit, on a more basal level¹⁸.

Since TJs are shown to be negatively affected by IL-1 and TNF-α (Figure 1), two pro-inflammatory cytokines associated with the acute phase of COVID-19, it can be argued that SARS-CoV-2 infection might result in damage to TJs²⁰. This hypothesis is supported by a study suggesting potential binding of the SARS-CoV-2 ‘Envelope protein’ (E protein) to ‘Proteins Associated with Lin Seven 1ʹ (PALS1), as a factor enhancing the pathogenicity of the virus²¹. Binding with the viral E protein might lead to rearrangement of PALS1, a class of proteins involved in the maintenance of cell polarity. As PALS1 is indirectly linked with the TJs, an unexpected rearrangement in its structure could result in the disruption of TJ proteins, most notably ZO-1¹⁸.

ZO-1 is essential to the integrity of TJs, since beside serving as a component of the AJCs, ZO-1 (along with ZO-2) is thought to promote the assembly of CLDNs into TJs¹⁹. Moreover, ZO-1 is believed to enhance the stability of the BGB through the upregulation of ‘trans-epithelial resistance’ (TER) of the AECs. This is mediated in association with occludin and CLDN4²². The significance of the latter has further been emphasized by an investigation reporting downregulation of certain CLDNs in rats with vitamin D receptor deficiency; including CLDN4 and CLDN18²³, both of which contribute to the normal barrier function of the AECs¹⁹. SARS-CoV-2 infection particularly results in downregulation of CLDN18 in AECIIs, interfering with the normal cell–cell tethering in pulmonary alveoli²⁴.

CLDNs and ZO-1 might become subject to several complications such as hypoxia²² and overproduction of IL-1β and TNF-α²⁰, all of which have been reported in patients with COVID-19.^4,20 Monocyte-derived IL-1β could disrupt TJs, and exert a destructive impact on the AJs of ECs through phosphorylating β-catenin. TNF-α, on the other hand, could interfere with the proper localization of ZO-1 in the lungs, and negatively regulate the expression of this protein²⁰, repressing the assembly of CLDNs into the TJs¹⁹ (Figure 1).

COVID-19 and the disintegration of the ECM

The thickest component of the BGB in mammals², ECM is the middle layer of intertwined proteins and proteoglycans upon which the AECs and ECs reside. Aside from providing a scaffold for the cells of the BGB, the significance of ECM is also due to its role in the development and differentiation of AECs²⁵. Throughout the course of COVID-19, ‘hypoxia-inducible factor alpha’ (HIF-α) is overexpressed in the cells receiving low oxygen as a consequence of the COVID-19-induced hypoxia. The resulting upregulation of HIF-α, while aggravating the secretion of pro-inflammatory cytokines, may also enhance the release of VEGF-A into the blood²⁶; a finding reported in COVID-19 patients²⁷.

The subsequent binding of VEGF-A to its receptor, VEGFR, on ECs prompts disintegration of the endothelial TJs and cleavage of the ECM proteins²⁰. The latter is thought to be mediated through the binding of VEGF-A to the ECM and several ‘matrix metalloproteinases’ (MMPs), since VEGF-A is not stored intracellularly²⁸. MMPs, a family of monocyte-derived ECM-degrading enzymes secreted in response to inflammation (Figure 1), might positively regulate the expression of VEGFR in ECs through the NF-κB signaling pathway²⁰. This could potentially be the case with COVID-19, particularly in patients with high serum levels of glucose, since hypoxia, under hyperglycemic conditions, was shown to upregulate MMP-2 in fibroblasts²⁹, a type of ECM-synthesizing cell residing in the pulmonary interstitium³⁰. It is implied that the inhibition of MMP-2 might be associated with downregulation of VEGF-A²⁹.

The COVID-19-associated disintegration of ECM could also occur following the induction of certain MMPs, particularly MMP-2 and MMP-9, in response to the increased activity of CD147 or ‘extracellular matrix metalloproteinase inducer’ (EMMPRIN)³¹. CD147 is a transmembrane glycoprotein expressed in the ECs of cerebral vessels, which contributes to the formation of an important tissue barrier termed ‘blood-brain barrier’ (BBB). SARS-CoV-2 might target CD147 during its entry to the cerebral ECs and trigger an MMP response, leading to the disruption of ECM in the BBB³². The BGB could be affected in a similar way, as CD147 is also expressed in the AECIIs residing at the margins of the fibrotic alveolar area in ‘idiopathic pulmonary fibrosis’ (IPF)³¹, which is positively correlated with the severity of COVID-19³³.

COVID-19 and the STAT3-induced alveolar epithelial hyperplasia

With a larger population compared to that of the AECIs, AECIIs only cover 10% of the alveolar surface area at most, owing to their small size and cuboidal morphology. Known to regulate the alveolar surface tension through the production of surfactant, AECIIs constitute the stem cell niche of the BGB, as they are precursors of AECIs. Any injury exceeding the capacity of AECIIs for proliferation of AECIs can result in the death of AECIIs and formation of scar tissue or fibrosis in the alveoli³⁴.

In the case of COVID-19-induced lung injury, however, the ongoing inflammation at the BGB might act in favor of AECIIs, albeit, in an unfavorable way³⁵. The transition from AECIIs to AECIs is mediated, in part, through the activation of a transcription factor known as ‘signal transducer and activator of transcription 3ʹ (STAT3), which is part of the STAT3–BDNF–TrkB axis³⁶. STAT3 was shown to be stimulated by IL-6, a pro-inflammatory cytokine associated with SARS-CoV-2 infection. The resulting upregulation of STAT3 in COVID-19, induced either indirectly by the action of IL-6 or directly through the cytopathic effects of SARS-CoV-2, might lead to excessive proliferation of AECIIs, which can progress into AECII hyperplasia³⁵. A common postmortem histopathological finding in patients with COVID-19, AECII hyperplasia could contribute to alveolar fibrosis, even in the early stage of the disease³⁷ (Figure 1).

In addition to its hyperplastic effects on AECIIs, STAT3 may also trigger the activation of ‘indoleamine 2,3-dioxygenase 1ʹ (IDO), an immune-modulating intracellular enzyme with a key regulatory role in the metabolism of tryptophan. The STAT3-induced upregulation of IDO in the alveolar ECs might be partly responsible for the vascular dilatation and distortions reported in COVID-19 patients³⁸, indicating the dual disruptive effect of STAT3 on the BGB.

COVID-19, TGF-β and the transition of epithelial and endothelial cells

COVID-19-related alveolar fibrosis might be accelerated through the local effects of ‘transforming growth factor beta’ (TGF-β); a pro-inflammatory cytokine released by neutrophils recruited into the alveoli. This was confirmed by detection of TGF-β in the ‘broncho-alveolar lavage’ (BAL) samples obtained from COVID-19 patients³⁹. A known inducer of ‘reactive oxygen species’ (ROS), TGF-β1 is suggested to upregulate a number of proteins associated with the ‘endoplasmic reticulum’ (ER) stress response by positively regulating the production of ROS in the fibroblasts of the respiratory tract⁴⁰.

TGF-β1 can induce the expression of ‘high mobility group box 1ʹ (HMGB1), an extracellular mediator of inflammation, in AECs⁴¹ (Figure 1). High levels of HMGB1 might be associated with increased vulnerability to SARS-CoV-2⁴², as it stimulates the ‘epithelial–mesenchymal transition’ (EMT) of AECs in response to TGF-β1⁴¹. STAT3 activation can also trigger EMT⁴³. A factor contributing to fibrosis, EMT is not limited to the AECs. In fact, it has been suggested that SARS-CoV-2 infection might influence the ECs in a similar way, resulting in ‘endothelial–mesenchymal transition’ (EndMT), which is characterized by the loss of vascular E-cadherin and TIE1/2, along with subsequent MMP-induced degradation of the ECM (Figure 1). Together, EMT and EndMT could lead to the formation of fibrotic tissue in the BGB⁴⁴. Additionally, HMGB1 might also affect the TJs of AECs by downregulating CLDNs⁴¹.

COVID-19 and the pyroptosis of pulmonary epithelial and endothelial cells

Pyroptosis is an inflammatory type of caspase-mediated programmed cell death associated with the disruption of plasma membrane and release of cellular contents that can prompt an acute inflammatory response through the activation and recruitment of neutrophils⁴⁵ that secrete the pro-inflammatory IL-1β and HMGB1. Increased levels of HMGB1 were indicated to trigger pyroptosis in bronchial epithelial cells⁴²; hence, it can be argued that AECs could also be affected by HMGB1 and undergo pyroptosis⁴⁶. Unlike the AECs, there is solid evidence regarding the caspase-1-mediated pyroptotic effect of HMGB1 on ECs in Kawasaki disease⁴⁷. Importantly, Kawasaki-like disease was reported in a number of pediatric cases of COVID-19⁴⁸. The pyroptosis associated with COVID-19 is mostly regulated by the expression of SARS-CoV-2 S-RBD, Env and ORF3a genes, which code for the receptor-binding domain of S protein, E protein and ORF3a protein, respectively. The latter is involved in the replication of SARS-CoV-2. Transfection of human bronchial cells with S-RBD and ORF3a primers was demonstrated to incite pyroptosis accompanied by elevated extracellular levels of HMGB1, hence, the positive feedback loop between SARS-CoV-2-induced pyroptosis and HMGB1⁴² (Figure 1).

COVID-19, PAF and the formation of microthrombi in the alveolar capillaries

Increased platelet activation is a complication of SARS-CoV-2 infection in critically ill patients⁴⁹. In addition to increased activity, platelet count might also rise in patients with certain comorbidities such as ‘immune thrombocytopenia’ (ITP), resulting in thrombocytosis. The facilitated thrombotic activity reported in COVID-19 is thought to be mediated by ‘platelet activating factor’ (PAF)⁵⁰; a type of lipid inflammatory molecule produced by neutrophils, monocytes, ECs and activated platelets. The upregulation of PAF could lead to the disruption of TJs and AJs in ECs, increasing vascular permeability¹⁵ (Figure 1). PAF and ACE2 influence each other through ‘angiotensin II’, which is the primary substrate of ACE2. Angiotensin II enhances the formation of PAF, while PAF positively upregulates the activity of ACE2, the principal receptor of SARS-CoV-2. PAF is also secreted by ECs in response to the increased levels of IL-1. Taken together, overproduction of PAF during an acute inflammation⁵⁰ could lead to the formation of microthrombi in the BGB capillaries, as reported in COVID-19 patients⁶.

Another culprit for the microthrombotic events affecting alveolar capillaries is suggested to be ‘plasminogen activator inhibitor-1ʹ (PAI-1 or SERPINE1), whose plasma levels are increased in COVID-19. A factor associated with higher mortality rates in patients with ARDS, PAI-1 is often released from platelets and ECs. Similar to PAI-1, ‘tissue plasminogen activator’ (tPA or PLAT) is elevated in the plasma of COVID-19 patients. Nonetheless, its anti-fibrinolytic effect is dominated by the fibrinolytic effect of PAI-1.⁴⁹ The ultimate result is increased deposition of fibrin in the alveolar capillaries, facilitating the migration of neutrophils and their recruitment into the BGB²⁰.

Candidate therapeutic agents for restoring the BGB in COVID-19

Considering the pathogenesis of SARS-CoV-2 (Figure 1) and its effect on each of the three primary components of the BGB, several therapeutic agents could be considered for the treatment of COVID-19-associated alveolar-capillary injury. The candidate therapeutic agents, listed in Table 1, target different cellular pathways and pro-inflammatory cytokines to reverse the inflammation-induced damage to the BGB. While some of them, such as Quercetin²², are considered natural medicinal agents, others, like Ebselen⁵², are synthesized in the lab. ‘Leukemia inhibitory factor’ (LIF), on the other hand, is a naturally occurring cytokine in the lung, with protective effects, which is gradually diminished as a result of aging³. Regardless of their origins, natural, synthetic and endogenous therapeutic agents can be administrated to COVID-19 patients in an exogenous manner to prevent deterioration of the BGB.

Table 1.

List of candidate therapeutic agents for the treatment of COVID-19-associated alveolar-capillary injury

Candidate therapeutic agent	Target component(s) in the BGB	Target mediator(s)	Potential therapeutic effect(s)	Citation(s)
LIF	AEC EC	IL-6 STAT3	Protecting the capillaries of the BGB against cytokine-induced inflammation through downregulation of IL-6 Maintaining the population of the AECs and preventing EMT by balancing the STAT3-associated regeneration of AECIIs, particularly in conditions such as hyperoxia	3
BMS-3455	AEC	NF-κB	Reversing the disruption of TJs in AECs by restoring the assembly of CLDN18 into TJs	19
Quercetin	AEC EC ECM	VEGFA/VEGFR PAI-1/tPA MMPs STAT3	Counteracting vascular leakage by maintaining the integrity of endothelial TJs through upregulation of ZO-1, JAM-C, CLDN4 and occludin Inhibiting the VEGFA/VEGFR-induced disruption of the ECM Offsetting deposition of fibrin in the BGB caused by balancing the counter-regulatory effects of PAI-1 and tPA Reversing MMP-associated disintegration of the ECM by downregulating MMP2 and MMP7 Hindering EMT and AECII hyperplasia through inhibiting STAT3 activation	20,22,43
Nonspecific anti-MMP agents; tetracycline and doxycycline analogues, e.g. COL-3 and CMT	EC ECM	MMPs	Opposing the MMP-driven pathologic remodeling of ECM and increased permeability of the BGB Inhibiting the hypoxia-induced angiogenesis and thickening of the BGB through downregulation of VEGFA (under hyperglycemic conditions)	29,51
EW-7197	AEC	TGF-β1 receptor kinase	Preventing the EMT of AECs caused by TGF-β1/HMGB1 signaling Protecting the integrity of TJs in AECs by enhancing the expression of CLDN1, 2 and 7	41
Glycyrrhizin	EC AEC (?)	HMGB1 IL-1β IL-6	Preventing pyroptosis of ECs and bronchial epithelial cells by inhibiting HMGB1 Counteracting neutrophil recruitment into the BGB by attenuating the release of IL-1β and IL-6 Reversing bronchial epithelial cell death by terminating the replication of SARS-CoV-2 through unspecified mechanism(s)	42
Ebselen	AEC EC ECM	Mpro ICAM-1 TGF-β	Terminating the replication of SARS-CoV-2 in AECs by targeting Mpro; the main protease of SARS-CoV-2 Reversing the ARDS-related endothelial injury through downregulation of ICAM-1 Downregulating TGF-β in the lung, possibly by targeting the neutrophils recruited into the BGB Counteracting fibrosis by mitigating the TGF-β-induced activation of fibroblasts through downregulation of ER stress-related proteins	52,53
Rupatadine	EC	PAF IL-1β IL-6	Repressing inflammation at the BGB through the inhibition of PAF-induced perivascular mast cell activation and the concomitant release of PAF, IL-1β and IL-6 from mast cells	54

Conclusions

The BGB is perhaps the first tissue barrier in the body to be selectively targeted by SARS-CoV-2 following exposure to this virus. As the principal barrier directly involved in the regulation of gas exchange, any insult to this extremely thin part of the alveoli could result in impaired ventilation, and eventually, hypoxia. The triple layer structure of the BGB as a whole, while serving its biologic purpose, could further perplex the pathogenesis of COVID-19, since both AECs and ECs could be damaged following SARS-CoV-2 infection, and the resulting inflammation might very well be accompanied by thickening of the ECM, rendering the BGB an obstructive barricade instead of a protective barrier. While it is necessary to understand the principles of COVID-19-induced alveolar-capillary injury, it is also important to investigate and identify therapeutic agents capable of selectively targeting the components of the BGB in hopes of reversing the damage inflicted to the BGB.

Disclosure of potential conflicts of interest

No conflicts of interest, financial or otherwise, are declared by the author. The author has no financial relationship with a biotechnology and/or pharmaceutical manufacturer that has an interest in the subject matter or materials discussed in the submitted manuscript.