The RAAS regulates blood composition and circulating volume, having a role in blood pressure homeostasis in various animal species (2). The classic RAAS pathway involves the production of Renin by juxtaglomerular cells whose only substrate is angiotensinogen, forming Angiotensin I (ANG I) (3). In turn, ANG I is then cleaved into ANG II by Angiotensin-Converting Enzyme (ACE), which is primarily transcribed as a membrane-bound enzyme and later detached from the cell membrane by proteolytic cleavage (4, 5). Specifically, the C-terminal portion of ACE (also known as ACE/Kininase II) is responsible for ANG I metabolization to ANG II, while the N-terminal portion cleaves Bradykinin into other peptides (6, 7). However, ANG II might also be generated by other enzymes in tissue-specific RAAS (cardiovascular, pulmonary and renal) (8). The main enzyme involved in ANG II generation in tissue-specific RAAS are serine proteases (specially kallikrein-like enzymes also called tonins), and also cathepsin G and chymase (9). In the kidneys, 40% of ANG II is generated by non-ACE pathways, while in human heart, coronary arteries and atherosclerotic aorta, chymase is the dominant generator of ANG II, suggesting that these non-classical pathways might have a role in the development of several diseases (9, 10). The best described physiological effects of ANG II involves its binding to AT1 receptors (AT1R), leading to vasoconstriction and reducing renal sodium excretion, via aldosterone stimulation (11).

A recent description of the RAAS has revealed several ANG II derived peptides, highlighting the importance of this system as well as providing a better understanding of its role (3). ACE2, which was initially described as an ACE homologous, sharing 40% sequence similarity with ACE (12), counteracts the effects of ANG II by converting ANG I into Angiotensin 1-9 (ANG 1-9), and ANG II into Angiotensin 1-7 (ANG 1-7) (13). While binding to both ANG I and II, ACE2 preferentially binds to ANG II, with a catalytic efficiency 400 times greater than for ANG I. Consequently, plasma levels of ANG 1-7 are higher than those of the other isoforms, such as ANG 1-9 (12, 14). ANG 1-7, by the actions of ACE and aminopeptidases, generate two other peptides, Angiotensin 1-5 and Angiotensin 2-7, respectively (3). Angiotensin 2-7 will then generate Angiotensin 3-7 via aminopeptidases action, while Angiotensin 1-5 is converted into Angiotensin 1-4 by neprilysin or carboxipeptidases. Finally, both Angiotensins 3-7 and 1-4 are converted into Angiotensin 3-4 by aminopeptidases and endopeptidases (3). Besides ANG 1-7 generation, ANG II, when cleaved by a carboxylase, will generate Angiotensin A, whereas when cleaved by aminopeptidase A, ANG II generates Angiotensin III (3). Angiotensin III is then cleaved by aminopeptidase N, forming Angiotensin IV, which is then cleaved by another aminopeptidase, forming angiotensin 5-8. The pathway ends with angiotensin 5-8 being processed by a carboxypeptidase, generating angiotensin 5-7 (3). Most of these new RAAS members are ANG II-derived peptides, and also exert their actions via AT1R and AT2R, having local physiological effects that can be similar to those produced by ANG II, according to the tissues and systems were they occur.

ACE2 is expressed in multiple human epithelial cell types, including vascular endothelial cells, nasal and intestinal epithelial cells, and within the airways, in alveolar epithelial type II cells of the lung parenchyma (15). It is a metalloproteinase type I transmembrane enzyme, characterized by a small cytoplasmic domain, a transmembrane region and an ectodomain (16). Due to conformational differences in the catalytic site, ACE2 is not blocked by classic ACE inhibitors (12) nor angiotensin-receptor blockers. The ACE2 ectodomain is recognized as a cleavage site for ADAM-17 and TMPRSS2 sheddases (17). ADAM-17 is a disintegrin and metalloproteinase 17 domain (17) that cleaves ectodomains of diverse cytokines, receptors and enzymes anchored in cell membranes (18). TMPRSS2 is a type II serine protease, acting in conjunction with ADAM-17. Both enzymes share the same specificity for transmembrane ACE2 (tACE2), however, the final product, soluble ACE2 (sACE2), is different (19). ADAM-17 cleavage leads to the shedding of a form of sACE2 that is still biologically active, being able to cleave ANG II and counteracting its effects, but also blocking circulating viral particles (14). Studies using recombinant sACE2 have demonstrated that it also cleaves ANG I into ANG 1-9, which also counteracts ANG II effects. However, little is known about physiologically produced sACE2 and ANG I cleavage (20). tACE2 cleavage by TMPRSS2, on the other hand, does not release the biologically active form of the enzyme (21) and has been shown to be involved in SARS-CoV and SARS-CoV-2 infections, being the main pathway by which the virus enters the cell (13, 22–24). Therefore, the balance between the two sACE2 forms generated by both enzymes might have an important role in COVID-19 outcomes. This is summarized in Figure 1A and is relevant to COVID-19 symptoms.

The best scenario is probably related to a limited entry of the virus into the cell and consequently a better prognosis. The first step implicates TMPRSS2 and ADAM-17 sheddases, which are involved in tACE2 cleavage. The mechanisms leading to tACE2 shedding by TMPRSS2 are different from the ADAM-17 pathway (16). Studies with SARS-CoV suggest tACE2 shedding by the virus involves TMPRSS2 activation in the short cytoplasmic domain of tACE2. By contrast, ADAM-17 constitutively sheds tACE2 to produce its soluble form (sACE2), which retains its biological activity (14, 16, 21).

Regarding SARS-CoV-2 infection, ADAM-17 and TMPRSS2 provoke opposite effects. SARS-CoV-2 has been shown to bind to tACE2, which then is partially or fully internalized to the cell cytoplasm along with the virus (24). In this process, TMPRSS2 is activated, leading to tACE2 shedding and thus reduced tACE2 levels (14, 24, 28, 36, 37). As a consequence, TMPRSS2 activation leads to increased SARS-CoV and SARS-CoV-2 cell penetration in the host cell as well as to increased ANG II levels, as shown in Figure 1B (13, 16, 22, 24). It has been shown that African Americans and the elderly present increased levels of TMPRSS2, which might to some extent explain the increased severity of COVID-19 seen in these two populations (19).

Interestingly, ANG II is known to upregulate ADAM-17 activity in different tissues, such as heart muscle, kidney, pancreas, etc (38). In this case, if ANG II and ADAM-17 levels increase, some of the tACE2 may be shedded, which might be an important biological response to reduce SARS-CoV-2 infection in some tissues. Moreover, studies have demonstrated sACE2 generated from ADAM-17 cleavage is capable of binding SARS-CoV-2, sequestrating the virus and preventing it from attaching to the tACE2 (28). Curiously, people from blood group A are at higher risk of developing severe forms of COVID-19 and exhibit lower plasmatic ACE levels (39). Hence, patients with blood group A are more likely to have lower levels of ANG II, negatively regulating AT1R activation and ADAM-17, which seems to support our theory.

In the same context, AT1R activation by ANG II downregulates tACE2 expression, reducing ANG 1-7 and favoring ANG II vasoconstrictor activities (40). Chronic AT1R blockade has been shown to lead to a fivefold increase in tACE2 levels in rat aorta (40). These data indicate tACE2 reduction following AT1R activation by ANG II may trigger a positive feedback mechanism in ANG II regulation (40). Positive feedbacks are self-limiting and are rarely the best biological choice to control essential physiological mechanisms, such as blood pressure control. Nevertheless, an ANG II peak may be an interesting biological strategy to reduce viral entry into the cell via tACE2, reducing its expression and attenuating the possible viral damage ( Figure 1B ).

The worst-case scenario would occur if ANG II production and/or AT1R activity are reduced for any reason, e.g. using ACE inhibitors (30) ( Figure 1C ). In this case, ADAM-17 would not be consistently activated, leading to two concomitant situations: a) there would be more tACE2 available to facilitate the entry of the virus into host cells, and b) less biologically active sACE2 might result in more circulating free virus, with the potential to enter host cells via tACE2. Therefore, the TMPRSS2/ADAM-17 balance would be shifted towards TMPRSS2 activation, thus favoring infection by the virus. Corroborating this idea, recent studies have suggested ACE inhibitors are associated to poor prognosis in COVID-19 patients (41, 42). Paradoxically, a literature review from Mackey et al. (43) concludes direct ANG II suppression with ACE inhibitors would stabilize cell membrane complexes between tACE2 and AT1R, thus reducing the ability of the virus to enter host cells, which would be beneficial in COVID-19. This observation requires additional research and the assessment of intervening variables in infected patients (43).

While increased ANG II levels in COVID-19 has been reported (44), as yet there are no studies addressing sACE2 levels in the disease. A similar increase in ANG II levels is also seen in H5N1 avian influenza and H7N9 patients, indicating this biological response could be common in some viral infections (45, 46). Indeed, tACE2 levels are significantly reduced after H1N1 infection, which is probably related to tACE2 degradation via the proteasome pathway (directly cleaved by the influenza neuraminidase protein) (47). Moreover, mice infected with H5N1 show tACE2 expression downregulation in the lung and increased ANG II plasmatic levels, as proposed in our model (45, 46, 48). The literature suggests different viruses, not only SARS-CoV and SARS-CoV-2, might use tACE2 as a carrier. One accepted theory is that bat coronavirus can infect small mammals, such as the palm civet (49). This would suggest this biological defense against viral entry is a well conserved mechanism in mammals. These animals act as intermediate hosts, providing a suitable environment for the coronavirus Receptor Binding Domain (RBD) to develop point mutations compatible with tACE2 binding (50, 51). Studies based on sequences from coronavirus found in bats suggest their RBDs contain deletions spanning key residues needed to mediate viral and tACE2 interaction (49, 52).

Several studies corroborate the occurrence of the tACE2 and sACE2 changes proposed in Figure 1B (14, 24, 36, 37). The increased sACE2 resultant from TMPRSS2 cleavage, and the decrease in its biologically active form from ADAM-17 cleavage, will lead to a mild increase in sACE2, which might not be sufficient to generate ANG 1-7 from the already low levels of ANG II, which is proposed in Figure 1C . Moreover, sACE2 from TMPRSS2 will favor increased SARS-CoV-2 infection in the host cell (21).

These physiological changes produce an imbalance favoring the ACE/Ang II/AT1R axis, which is a pro-inflammatory, vasoconstrictor, fibrotic pathway, while reducing the ACE2/ANG 1-7/MasR signaling pathway that promotes vasodilatation and anti-inflammatory effects (53). Indeed, a recent study suggests the activation of the ACE2/ANG 1-7/MasR axis in order to prevent excessive inflammatory response, especially in lungs (54). Currently, there are three ongoing clinical trials to evaluate the effects of ANG 1-7 administration on hyper-inflammatory response in critically ill COVID-19 patients (55).

sACE2 is a marker of severity and clinical prognosis in several inflammatory diseases. In patients with chronic obstructive coronary artery disease, sACE2 levels were higher than in healthy controls, which was considered an independent predictor of cardiovascular mortality (53). sACE2 levels have been shown to increase in type II diabetes mellitus and coronary arterial disease, being a potential marker for atherosclerosis (56). Also, increased sACE2 levels are considered a marker of poor cardiac outcomes in several cardiovascular diseases (57). Pro-inflammatory cytokines IL1-B and TNF- α also induce ADAM-17 activation, thus increasing sACE2 levels (16). It is worth noting that patients with obesity, diabetes and cardiovascular diseases are risk groups for COVID-19, presenting the worst outcomes (58, 59). Some interesting hypotheses may be related to poor prognosis. First, in patients with diabetes and cardiovascular diseases, ANG II plasmatic levels are typically high, indicating the RAAS is hyperactive (60). Consequently, it would be unlikely to produce a new peak of ANG II to combat the virus infection. Second, the increased ANG II plasmatic levels may also lead to AT1R downregulation. These two hypotheses support the patterns of ANG II and AT1R levels proposed in Figure 1C .

An observation should be made regarding the circadian variations in the levels of the RAAS components, which might also affect COVID-19 outcomes (61). It has been shown, in both animals and humans, that several components of the RAAS present a circadian pattern throughout the day (62–64); in mice, for example, both ACE and tACE2 expression follows a circadian pattern, and the ACE/tACE2 ratio is reduced during the night (65). The pathogenicity of a virus has also been shown to be affected by the circadian rhythm through direct regulation of viral replication inside the host cell (61). It is not only the time of day in which the infection occurs that influences the disease outcome, but also disruption of the circadian rhythm increases the viral replication, as observed in herpes and influenza A infections (66). This suggests that SARS-CoV-2 pathogenicity might be influenced by time of infection due to circadian alterations in the ACE/tACE2 ratio. A lower ACE/tACE2 ratio means reduced ANGII levels and increased tACE2, lending SARS-CoV-2 increased pathogenicity; while with a high ratio, the virus would be less pathogenic. This is also supports our hypothesis that ANG II peaks in response to SARS-CoV-2 infection. Since ACE2 counteracts the ANG II effects by cleaving it to ANG 1-7, a high ACE/tACE2 ratio would favor a peak in ANG II levels. In fact, a study by Liu and colleagues (2020), observed markedly high ANG II levels in COVID-19 patients when compared to healthy controls. Moreover, these levels were negatively correlated to both viral load and lung injury, suggesting an important role for ANG II during early SARS-CoV-2 infection (44).

Regarding T lymphocytes, a marked decrease in CD3+, CD4+ and CD8+ T cells have been shown in critically ill patients (28, 85, 86). The CD8+ decrease may be significant in patients with death outcome in comparison with those who recovered. Recent studies (85) demonstrated that cytotoxic CD8+ T cells from COVID-19 patients presented severe inflammatory and antiviral activity compared to healthy controls (85, 87, 88). Therefore, reduction in CD8+ T cells observed in COVID-19 might actually represents an exhaustion-related effect in this cell subpopulation due to clonal expansion of activated cells. As a result, there is an imbalance between the adaptive response and the disease progression.

The RAAS has a critical role in the early response to SARS-CoV-2 infection and, consequently, COVID-19 better or worse prognoses. In addition, ANG II plays an important role in the onset and progression of inflammatory responses. Taken together, it is possible that an ANG II peak is our first biological response against virus entry (89, 90). To the best of our knowledge, this is the first viewpoint to explore the logical interaction of these pathophysiological mechanisms. Thus, monitoring ACE, tACE2, sACE2, ANG I, ANG II, ANG 1-7, and ANG 1-9 during COVID-19 may provide clues to develop effective treatments.