SARS-CoV-2 as a Factor to Disbalance the Renin–Angiotensin System: A Suspect in the Case of Exacerbated IL-6 Production

A mammalian naive organism exposed to a new pathogen can activate two different branches of the immune system: the innate and the adaptive immunity. The innate immunity is involved in the activation of nonspecific processes such as inflammation, whereas the adaptive immunity is related with Ag-specific processes such as, among other, Ab production or immunological memory. The most serious consequences of COVID-19 infection come from an acute respiratory distress syndrome that is aggravated by exacerbated inflammation (1). Immune cells of the white lineage infiltrate the affected tissue to become macrophages that are subsequently activated to produce and release a variety of cytokines. The plasma levels of one of them, IL-6, directly correlates with COVID-19 severity. Often, the macrophage response is limited and does not aggravate viral infections. However, for unknown reasons, some COVID-19 patients develop the so-called cytokine storm, which correlates with symptom aggravation, and the outcome can be fatal (2–4). Patients (n = 30) in the critical care unit of one of the main hospitals in Barcelona were assessed for plasma/serum levels of four cytokines (IL-6, IL-1β, IL-8, and TNF-α). IL-6 levels led to high interindividual variability, whereas IL-1β levels were within the reference value range. There was not any common trend in IL-8 or TNF-α values, which were either normal or increased. In summary, the most noticeable finding in the cytokine storm was the level of IL-6, which in some patients, may be high enough to lose the dynamic range of measurement (i.e., two orders of magnitude increase or even higher) (Refs. 5, 6, and T. Herold, V. Jurinovic, C. Arnreich, J.C. Hellmuth, M. von Bergwelt-Baildon, M. Klein, and T. Weinberger, manuscript posted on medRxiv, M.J. Cummings, M.R. Baldwin, D. Abrams, S.D. Jacobson, B.J. Meyer, E.M. Balough, J.G. Aaron, J. Claassen, L.E. Rabbani, J. Hastie, B.R. Hochman, J. Salazar-Schicchi, N.H. Yip, D. Brodie, and M.R. O’Donnell, manuscript posted on medRxiv, and J. Gong, H. Dong, S.Q. Xia, Y.Z. Huang, D. Wang, Y. Zhao, W. Liu, S. Tu, M. Zhang, Q. Wang, and F. Lu, manuscript posted on medRxiv). The marked increase in IL-6 levels has prompted the design of a clinical assay using tocilizumab, a monoclonal anti–IL-6 Ab (7, 8). More promising, and already used in COVID-19 patients treated in Spanish hospitals, corticoids may attenuate IL-6 production. It is known that the production of the cytokine is inhibited by glucocorticoids in a wide range of cell types (9). In critical illnesses, the α isoform of the glucocorticoid receptor (or nuclear receptor subfamily 3, group C, member 1 [NR3C1]) arises as a key regulator of homeostatic processes (see Ref. 10 for review).

The link between the renin–angiotensin system (RAS) and coronaviruses was serendipitously discovered. The angiotensin-converting enzyme (ACE) 2 was identified as the receptor for viruses of the SARS family, SARS-CoV-2 being no exception (11–15). Importantly (see below), ACE2 has peptidase activity; it is one of the newest members of the RAS, which has been widely studied in relationship with the control of blood pressure. Among its components, there are ACE1, which produces angiotensin II (Ang II), and ACE2, which converts Ang II to angiotensin 1–7 (Ang1–7), Ang II receptors (Ang II receptor type 1 [AT₁R] and Ang II receptor type 2 [AT₂R]), and the Ang1–7 receptor. The latter was first identified as a product of an oncogene and because of resemblance to the mitochondrial assembly gene from Saccharomyces cerevisiae was named as Mas-related proto-oncogene (16). The receptor whose endogenous agonist is Ang1–7 is now known as Mas receptor (MasR). All these angiotensin-related receptors belong to the superfamily of G-protein–coupled receptors (GPRCRs). It should be finally noted that a novel family of receptors has been named as Mas-related GPCRs (Mrgprs) that, interestingly, also respond to Ang1–7 (17–19) but whose endogenous agonist is alamandine, another new RAS member. The relationships between RAS members is shown in Fig. 1.

The intense multidisciplinary research on HIV-1 infection led to the discovery of one of the main anchoring molecules, the chemokine CXCR4 GPCR, and of other cell surface proteins acting as virus coreceptors. A relevant HIV-1 coreceptor was identified as dipeptidyl peptidase IV (DPPIV), also known as CD26. In human lymphocytes, the link between CXCR4 and CD26 was proven using different approaches (20). Identified in bats but also occurring in primates, CD26 acts as a receptor for Middle East respiratory syndrome (MERS) coronavirus (21, 22). Using the online Search Tool for the Retrieval of Interacting Genes/Proteins (STRING), ACE2 is directly connected with CD26/DPPIV (Fig. 2A). By default, ACE2 is connected to Ang II, which, in turn, is connected to its cognate receptors, AT₁R and AT₂R (Fig. 2A). Interestingly, the appearance of a GPCR related to DPPIV is required to force the system to include chemokine receptors in the search for connections (Fig. 2B). In contrast, with these easy-to-find connections, there are almost no papers linking SARS viruses to GPCRs. In sharp contrast, there are several articles devoted to the link between HIV-1 and GPCRs. This fact probably derives from the exhaustive research on HIV-1 and the AIDS it produces, which was more deadly than coronavirus infections. A physiological substrate of CD26 is the stromal-derived factor 1 (SDF-1), alternatively known as CXCL12, which is the endogenous agonist of the CXCR4 chemokine receptor (Fig. 3). On the one hand, SDF-1 is protective against viral entry into cells, and CD26/DPPIV catalytic activity reduces the concentration of this protective factor (20, 23). On the other hand, CXCR4 is a factor in HIV-1 viral entry (24–30). Unfortunately, SARS-CoV-2 arrived, and no background on GPCRs and the viral proteins and/or viral entry into the cells exists. Therefore, there is a need to put the focus on the blind point of COVID-19 related research: involved GPCRs, how GPCR functionality is affected by coronaviruses, and whether GPCRs may or not favor SARS entry into cells. Obviously, it is also needed to know how the viral cycle affects the functionality of GPCRs that are directly involved in viral targeting, membrane fusion, and/or viral entrance into cells.

Extensive research is available related to the receptor of SDF-1, the CXCR4 chemokine receptor, and HIV-1. A search in PubMed on SDF-1 and HIV-1 leads to ∼500 articles and 40 reviews, the most recent (review) in May 2020 (31). The number of articles for HIV-1 and CXCR4 are ∼8 times higher, and the last review appeared in May 2020 (25). Interestingly, the formyl peptide GPCR binds with high-affinity peptides derived from proteins of HIV-1, SARS, and Ebola viruses (32, 33). However, the physiological and/or pharmacological relevance of such findings reported in 2006 is unclear. With the exception of the results related with the formyl peptide receptor, there are few or no articles devoted to studying links between GPCRs and the SARS family of coronaviruses.

The RAS is widely distributed in the mammalian body, and data in the CNS have provided evidence of a relevant role in immune function modulation. All receptors related to both Ang and Ang1–7 are expressed in microglia and contribute to the regulation of cell activation. There is interest in knowing whether the neurologic alterations observed in some COVID-19 patients are somehow mediated by the RAS system expressed in neurons and/or microglia (34–48). However, our focus in this review is RAS in macrophages, which release IL-6 when activated and also express RAS components. RAS has been for years under scrutiny from assuming that Ang II was proinflammatory and a potential profibrotic agent (49). It becomes evident that proinflammation and profibrotic actions depend on an exquisite balance within RAS, whose activity changes depending on the differential cell surface expression of its protein components (enzymes and receptors).

Activated macrophages appear in two main phenotypic variants, the proinflammatory or M1 and the anti-inflammatory or M2. From a pharmacological point of view, it is of interest to know how to skew macrophages to the M2 phenotype and whether it is possible that, at some stage of activation, M1 macrophages could be converted into M2 macrophages. Such polarization depends on inhibiting the anti-inflammatory pathway to reinforce the proinflammatory or vice versa. As pointed out in an excellent review in this topic (50), this cross-feedback regulation is due to activation/deactivation of transcription factors: PPARγ, IRF4, and STAT to promote M2 skewing and AP1, NF-κB, STAT1, and IRF5 to promote M1 skewing.

Unlike in renal cells, the RAS Ang1–7-MasR branch has been better characterized in macrophages than the Ang II–AT₁R–AT₂R branch. Although the expression pattern and the role of Ang II receptors in macrophages are not fully elucidated yet, the general idea is that Ang II is proinflammatory and that inhibitors of ACE1 or antagonists of Ang II receptors could be beneficial for patient with inflammatory diseases coursing with inflammation (51). Unfortunately, the system is more complex, as the two Ang II receptors usually have opposite functional activities. The present review complements another review (1) that appeared recently and tackles the issue of whether ACE2 expression is beneficial or not in SARS-CoV-2 infection (52).

In principle, Ang II has opposite effects depending on the receptor that is activated (Fig. 4). Activation of AT₁R by agonists exacerbates inflammation via enhancement of expression of proinflammatory cytokines. In contrast, activation of the AT₂R in macrophages regulates the activity of inducible NO synthase and negatively modulates the production of TNF-α, NF-κB, IL-6, and IL-1β (53, 54). The AT₂R also mediates the production of IL-10 by the activated macrophage (55), thus attenuating the inflammatory response. In physiological scenarios, it is assumed that inflammation occurs with preponderance of AT₁R-mediated signaling but that time passing leads to the downregulation of this receptor and to the upregulation of those RAS receptors whose function is to be anti-inflammatory and/or to mitigate inflammation (Fig. 4, left). There is a variety of ways by which RAS disbalances.

The burden of SARS-CoV-2 infections is much higher in men than in women. In April, it was reported that 2.4 times more men than women died of COVID-19 (56). As the ACE2 gene is in the X chromosome, it is tempting to speculate that gender differences in symptoms and mortality are due to decreased (overall) expression of ACE2 in men. Currently, there are no data to support the benefit of having more or less ACE2 expression. Also, one of the two X chromosomes in women is inactive. Surely 20–30% of genes escape inactivation, and the ACE2 gene seems to be one of them (57), but it is also true that ACE2 expression is greatly affected by sex hormones (58). ACE2 polymorphisms have been described (59), and therefore, males express one protein, whereas females may express two isoforms. If a given isoform represents a higher vulnerability to infection, men would be more exposed because the second isoform in women could exert a compensatory/protective effect. In summary, ACE2 expression may be different in men and women, but it cannot yet be confirmed that differential expression of equal or polymorphic ACE2 correlates with symptom’s severity and with mortality.

A hypothesis based on evidence from SARS research is virus-induced local reduction of ACE2 activity by either reducing its catalytic activity, downregulation, or shedding. In any of those circumstances, Ang II accumulates, Ang1–7 is reduced, and inflammation is prolonged and/or exacerbated (Fig. 4, right).

It is not known whether the two angiotensin receptors interact in macrophages as they do in microglia, where they are upregulated upon activation and where the AT₂R counteracts the proinflammatory effects mediated by the AT₁R (37, 60). Activation of the AT₂R, which is usually upregulated in diseases with inflammatory component, reduces the action of M1 macrophages and, accordingly, the synthesis of TNF-α and IL-6 (61). In summary, more effort is needed to assess the expression of AT₁R and AT₂R in resting and activated macrophages and to define the role of each receptor when macrophages activate in response to different pathogens.

The first evidence linking Ang1–7 to the regulation of cytokine release by LPS-activated macrophages came from studies using ACE2 knockout mice whose macrophages, upon activation, showed altered expression of adhesion molecules and of cytokines (IL-6 included) (62). When ACE2 is overexpressed, there is a marked reduction of Ang II–induced MCP-1 production, which is seemingly mediated by Ang1–7 (63). Those results led us to propose that the Ang1–7-MasR branch of RAS was relevant for regulating macrophage activation (64). The anti-inflammatory action of Ang1–7 was later confirmed in LPS-treated peritoneal macrophages (65) and in a polymicrobial sepsis rat model (66). Very recent results in an endotoxemia rat model show systemic anti-inflammatory action of Ang1–7 that is mediated by the MasR (67).

MasR activation poses a brake in macrophage activation. In fact, genetic ablation of the receptor leads to increased infiltration of macrophages in a variety of tissues and to increased expression of proinflammatory genes (68). Already, in 2012, upregulation of MasR in LPS-treated cells and regulation by Ang1–7 of TNF-α and of IL-6 production by activated macrophages were reported (65). The effect of the most recently identified member of the angiotensin family, alamandine (Fig. 1), and of Ang1–7 depends on the activated macrophage phenotype. In an in vitro model using macrophages activated using different protocols and in subsets of infiltrating lung macrophages isolated after inducing pleurisy in mice, activation of either MasR or Mrgprs is ineffective on resting cells but reduces inflammation because there were fewer cells producing IL-1β and TNF-α (i.e., fewer cells with the M1 proinflammatory phenotype). In contrast, receptor activation by Ang1–7 or alamandine leads to an anti-inflammatory reprograming of activated macrophages (69).

Based on the involvement of RAS in macrophage biology, alterations induced by SARS-CoV-2 may lead to aberrant macrophage activation. The details of the likely mechanisms operating in macrophages responding to the viral infection are schematized in Fig. 5. First of all, SARS-CoV-2 influences the homeostatic function of its receptor. This issue has not been fully addressed, but even if the SARS-CoV-2 is not affecting peptidase activity, it has been shown that 1) the virus needs ACE2 for entering into cells (70), and 2) there is a coronavirus-induced shedding of ACE2 mediated by transmembrane serine protease 2 (TMPRSS2) and/or disintegrin and metalloproteinase domain-containing protein 17 (ADAM17) (71, 72). About 30 y ago, a selective reduction in cell surface expression of CD26 in cells targeted by HIV-1 was discovered (73). Analogously, ∼15 y ago, it was shown that coronavirus infection markedly reduces cell surface ACE2 expression (74, 75).

A decrease in ACE2 in the cell surface leads to RAS disbalance because of an increase in Ang II and a decrease in Ang1–7/alamandine extracellular concentrations. This unbalance of RAS leading to increasing the concentration of the proinflammatory peptide while decreasing the concentration of the anti-inflammatory peptide would exacerbate inflammation and proinflammatory IL production (Fig. 4).

Other scenarios remain speculative, as there are a number of unknowns to fully understand macrophage-associated pathology and propose effective therapies. For instance, it is not known whether ACE2 may interact with the AT₂R, or even with MasR or with Mrgprs. SARS viruses are surrounded by membrane proteins of the host; are ACE2 and angiotensin receptors among them? Downregulation of ACE2 would result in a similar outcome as enzyme inhibition. If the AT₁R is identified as coreceptor, is it disappearing from the cell surface together with ACE2 and/or SARS-CoV-2? This possibility would lead to the preponderance of the action of AT₂R, which, in general terms, is anti-inflammatory; therefore, it is not occurring or only occurs in those patients who do not develop the cytokine storm. Overall, it would be informative to investigate whether ACE2 interacts with other RAS components, thus leading to the possibility that receptors of the RAS are coreceptors for the virus. Also needed is to investigate how the virus itself and/or the spike protein of the virus affect the expression and function of all components of RAS in macrophages and targeted lung cells. As in the case of CD26, ACE2 may have catalytically independent function in those cells; indeed, CD26 in lymphocytes acts as a costimulatory molecule of relevance in immune and endocrine responses (76). In addition, the noncatalytic function of CD26 is altered by HIV-1 viral particles and by the envelope HIV-1 gp120 glycoprotein (77). This action is mechanistically dependent on the expression of the main HIV-1 receptor, CD4, and of CXCR4 (78). Already, in 2008, a role of the C-terminal domain of ACE2 in infections by SARS-CoV was identified (79) but was not investigated further. In summary, it would be important to know how SARS-CoV-2 affects the expression and function of protein RAS components, including the possible heteromers that, hopefully because of COVID-19–related research pressure, may now be discovered (labeled with * in Fig. 5).

Inhibiting ACE2 and altering the RAS balance in macrophages surely lead to the increase in proinflammatory cytokines but probably not to the extent of triggering a cytokine storm. Then, other mechanism may operate in cases with serious symptoms, very abnormal IL-6 levels, and increased death risk. It should be noted the title of a recent preprint: “Level of IL-6 predicts respiratory failure in hospitalized symptomatic COVID-19 patients” (T. Herold, V. Jurinovic, C. Arnreich, J.C. Hellmuth, M. von Bergwelt-Baildon, M. Klein, and T. Weinberger, manuscript posted on medRxiv). A recent metanalysis confirms that elevated IL-6 is one of the common findings in fatal outcomes (5). It is then reasonable to hypothesize that in the most serious cases, cells of the patient’s lungs may be able to release IL-6, thus engaging a harmful, vicious cycle.

The most serious COVID-19 cases course with pneumonia, and ACE2 is heavily expressed in the lung (80). The air-exposed internal surface of the lungs is composed of a variety of cells: ciliated, secretory, pneumocyte, basal, stromal, endothelial, and epithelial. The most abundant (i.e., epithelial and endothelial) are capable of producing IL-6 and express almost every RAS component. ACE2 is expressed in type II pneumocytes and in most respiratory-related epithelial cells, with the notable exception of those in the upper respiratory tract, including nasal and oral mucosa (80). In addition, ACE2 is located in the apical side of polarized cells (81) (i.e., readily available to SARS-CoV-2). How expression of RAS components is affected by SARS-CoV-2 in pneumocytes and lung epithelial or endothelial cells has not been tested yet.

Epithelial cells may produce IL-6. The first article on this issue reported production and secretion of IL-6 by stimulated epithelial cells of the human retinal pigment (82). As demonstrated in a variety of experimental models, endothelial cells may produce IL-6 (83–85). Furthermore, IL-6 and TNF-α produced by macrophages may lead to endothelial dysfunction (86).

From a multicenter validation study in critically ill patients because of sepsis, which also occurs in some COVID-19 patients (87), it is known that markers of endothelium activation are predictive of final outcome and that soluble FLT-1 (sFLT-1) may be a sepsis biomarker (88). The sFLT-1 is a marker of preeclamptic hypertension in which RAS is dysbalanced, and there is an increase in the expression in platelets of the heteromer formed by the AT₁R and the bradykinin B2 receptor (89). Interestingly, the B2 receptor may interact directly with the AT₂R (90). In addition, it has been recently hypothesized that dysregulated bradykinin signaling is behind respiratory complications because of COVID-19 (91), whereas it is known the direct link between bradykinin, platelets, and coagulopathy (92), which is another complication in COVID-19 patients. In fact, some patients not only present a cytokine storm, but signs of disseminated intravascular coagulation, which increased serum levels of fibrin degradation products, such as the d-dimer (93, 94). IL-6 may negatively impact on coagulation control mechanisms (8).

Also known is that human coronavirus spike proteins downregulate ACE2 (95) and that ACE2 expression is protective against lung failure (in a murine model) (74). After COVID-19 infection, sequalae include lung fibrosis that can be mediated by altered RAS. Ang II reportedly upregulates the AT₁R, downregulates the AT₂R (AT₁R/AT₂R ratio going from 0.4 to 1.4), and increases the activity of a profibrotic enzyme, hyaluronidase (96).

The interindividual response to COVID-19 depends on the viral load, the rate of viral replication, which varies from individual to individual, and the differential expression of the RAS component in the attacked cells (lung cells exposed to the inspired air) and in infiltrating macrophages. However, it cannot be ruled out that opportunistic infections take over and influence disease outcome. Bacterial infection can lead to the production of ILs by cells of the nonimmune system. In fact, bacterial LPS endotoxin can induce IL-6 synthesis by a variety of cells as diverse as osteoblasts (97) and endothelial cells (98). Existing data cannot test the hypothesis that pneumonia is caused by more than one pathogen, but it maintains that possibility. If this were the case, RAS would also have a relevant role.

It should be noted that acute respiratory distress syndrome animal models can be achieved by sepsis induction; therefore, superinfection may be also impacting in the respiratory problems in patients, requiring mechanical ventilation (99). Potential pathogens include, but are not limited to, Mycoplasma and Chlamydia, which, in parallel to pneumonia induction, lead to increases in serum of IL-6 levels (100–104), with features such as IL-17–mediated effects that are common in patients infected with MERS-CoV or SARS-CoV viruses (105, 106). Even pneumonia induced by Escherichia coli courses with elevated IL-6 concentration in serum, Gram-positive bacteria, leading to lower IL-6 and TNF-α levels than Gram-negative bacteria (107, 108). The very informative report on the findings in interstitial pneumonia versus nonspecific interstitial pneumonia/fibrosis is recommended, which includes the sources of IL-6 upon analysis of lung biopsy specimens (109). The imbalance of RAS in idiopathic pneumonia, the involvement in fibrosis postpneumonia, and the therapeutic possibilities by targeting some of the components are reviewed elsewhere (110).

MERS viruses attach to DPPIV/CD26 peptidase in targeted cells. SARS viruses attach to and alter ACE2 peptidase function with local downregulation and reduction of the catalytic activity, thus altering the Ang II/Ang1–7 and Ang II/alamandine ratios. Such RAS disbalance may lead to serious consequences in cases of inflammation. SARS-CoV-2 not only alters RAS in targeted cells but in macrophages infiltrating the affected tissue(s). The evidence points to a reduction of ACE2 function, by inhibition, downregulation, and/or shedding caused by viral proteins. Further alterations in the function of RAS proteins are likely to occur after coronavirus attack. The link between an altered RAS and cytokine storm and pulmonary fibrosis also needs to be addressed in detail. Such knowledge would likely open new perspectives to combat infection and guide the appropriate management of disease aggravation and of comorbidities.

The authors have no financial conflicts of interest.