There is growing evidence of cytolysis (mostly pyroptosis but not necroptosis) in patients infected with SARS-CoV-2 (7, 20). “Pyroptosis” refers to the biological process that relies on caspase-1-dependent activation of gasdermin D to form membrane pores. “Necroptosis” is dependent upon the intracellular signal transduction of receptor-interacting protein kinases, and membrane pores are formed through phosphorylation of mixed lineage kinase domain-like pseudokinase, which results in cytolysis (21). In vitro studies have shown that caspase-1 is activated in patients with severe COVID-19, downstream secretion of IL-1β increased, and decomposition of gasdermin D accelerated (22). Those data show that cytolysis during SARS-CoV-2 infection is mainly pyroptosis rather than necroptosis. Nucleotide-binding oligomerization domain-like receptors (NLRs) recognize the danger signals (which are derived from homologous hosts or microorganisms) and form large supramolecular chemical inflammasomes (9). Caspase-1 activation may be related to SARS-CoV-2 recognition by NLRs in cells. After caspase-1 has been activated, it initiates the pyroptosis pathway, activates gasdermin D to form plasma-membrane pores, and allows water influx, cell swelling, and osmotic lysis. Activated caspase-1 can also activate IL-1β and IL-18, which are released mainly through plasma-membrane pores (9). Many PAMPs, DAMPs, and virus products, which are released by SARS-CoV-2-induced pyroptosis, recruit various types of immune cells (including DCs) to infiltrate lung tissue and promote secretion of cytokines, particularly IL-6, IL-1β, IL-10, TNF, granulocyte macrophage-colony stimulating factor, IFN-induced protein-10, IL-17, monocyte chemoattractant-3, and IL-1ra (23).

ACE2 is the main target of SARS-CoV-2 (11). We postulate that ACE2 is also the inducer of pyroptosis caused by SARS-CoV-2. ACE2 is a transmembrane protease and is the main receptor for virus invasion (24). The N-terminal extracellular domain of ACE2 comprises a “claw-like” protease domain (PD). The receptor-binding domain (RBD) of SARS-CoV-2 combines with the PD of ACE2 to form an RBD–PD complex (24). The C-terminus is a transmembrane domain, also known as the collectrin-like domain (24). Studies have suggested that ACE2 is the major receptor of SARS-CoV and SARS-CoV-2 and binds to transmembrane serine protease (TMPRSS)2-activated spike proteins to induce virus entry into endosomes (25). ACE2 and TMPRSS2 are expressed primarily in cells of the upper respiratory tract and lungs (11, 25–27). These respiratory-tract cells with special cell-membrane receptors are ideal targets for SARS-CoV-2 invasion. This phenomenon may explain why SARS-CoV-2 is transmitted through aerosols and is very infectious (7).

The increased levels of lactate dehydrogenase (LDH) (28) and D-dimer in the plasma of patients infected with SARS-CoV-2 suggest that there is a high level of tissue injury in patients with severe COVID-19 (22, 23, 29, 30). Patients with severe COVID-19 with high LDH levels and leukopenia have impaired integrity of cell membranes (23, 31–33). In addition, the full blood count and biochemical findings of patients with severe COVID-19 reveal that the leukopenia observed in such patients appears to precede the “cytokine storm” (31, 34). SARS-CoV-2 infects human primary monocytes in vitro, resulting in their lysis and death. Flow cytometry and fluorescence microscopy have demonstrated the membrane disruption triggered by SARS-CoV-2 (22). In conclusion, we believe that SARS-CoV-2 invades cells through the ACE2, induces pyroptosis, and finally, causes the subsequent cytokine storm, leading to disease progression.

Products which are induced by SARS-CoV-2 attract DCs selectively into inflammatory sites (35–38). DC aggregation has been observed in the bronchoalveolar lavage fluid of COVID-19 patients, which suggests that DCs infiltrate into lung tissues by the products of SARS-CoV-2-induced pyroptosis (35–38) (Figure 1). Flow cytometry of the bronchoalveolar lavage fluid of COVID-19 patients has revealed that only type-2 conventional dendritic cells (cDC2) accumulate in the lungs (35, 36). cDC1 and plasmacytoid dendritic cells (pDCs), which are involved in IFN-I secretion, are absent in infected lungs (35, 36). cDC2 support the cluster of differentiation (CD)4+ T-cell response, stimulate follicular T-helper (Th) cells (which activate a humoral antiviral adaptive immune response), and produce inefficient antiviral proinflammatory cytokines (39–41).

SARS-CoV-2 attacks nasal cells and lung cells through ACE2, activates pyroptosis, and leads to the production of many DAMPs, PAMPs, and progeny viruses. These products activate innate immunity (Figure 1).

SARS-CoV-2 enters cells through the ACE2. ACE2 is a key factor in the renin–angiotensin–aldosterone system (RAAS) (42, 43). SARS-CoV-2 may affect the RAAS and lead to disease development by affecting the shedding of ACE2. The RAAS mediates blood-pressure control, inflammation, sodium reabsorption, and fibrosis (44). RAAS disorders can lead to heart failure, low blood pressure, atherosclerosis, and diabetes mellitus (45). The factor related most closely with ACE2 in the RAAS is angiotensin (Ang)II. The latter stimulates vascular contraction, secondary inflammation, and atherosclerosis through the type-1 angiotensin II receptor (AT1R) (46). Another receptor, type-2 angiotensin II (AT2R), in contrast to AT1R, is activated by AngII to promote vascular dilation, platelet aggregation, and promotion of insulin action. However, AT2R is rarely expressed in healthy adults (47). Therefore, the regulation and balance of AngII are dependent mainly on ACE2. The latter can convert AngII to Ang-(1–7), which is similar to AT2R stimulation (44, 47). An excess of AngII promotes pulmonary vascular contraction, inflammation, and cytokine-induced organ damage, increases the permeability of cell membranes and apoptosis, and induces acute kidney injury and acute respiratory distress syndrome (ARDS). AngII overactivation is associated with an increase in ACE2 shedding in patients with COVID-19. In a cohort of 12 COVID-19 patients, the circulating level of AngII was significantly higher than that in healthy controls (linearly correlated with viral load), which suggested a direct link between RAAS imbalance and multiorgan damage caused by SARS-CoV-2 infection (1, 48, 49).

Among hospitalized COVID-19 patients, the increase in AngII level is accompanied by an increased level of IL-6, with the highest mortality rate (29). This phenomenon has also been found in the patients with avian influenza A (H7N9) infection. Within 4 weeks of H7N9 infection, AngII levels increase gradually, which is associated with a worse clinical prognosis (50). In COVID-19 patients, AngII levels have been found to be closely related to viral titer and partial pressure of oxygen in arterial blood/fraction of inspired oxygen (49). COVID-19-induced AngII aggregation has been shown to promote acute lung injury (ALI) by activating cytokine-induced inflammation, activating the nicotinamide adenine dinucleotide hydrogen (NADH)/nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) oxidation system and vasoconstriction (49, 51, 52). Researchers found that COVID-19 was more severe and carried a worse prognosis in older people and males, if ACE2 expression declined (49, 53, 54). Coincidentally, patients with COVID-19 with diseases associated with RAAS overactivation (e.g., hypertension, diabetes mellitus) have been shown to carry a higher risk of transfer to the intensive care unit and mortality (55). RAAS overactivation appears to be closely related to a poor prognosis in COVID-19 patients, and the loss of ACE2 (which can regulate the effects of an overactive RAAS during the disease) will lead to worse consequences.

The ACE expressed on the DCs is still able to participate in the RAAS disorders (56–58). It has been proved that the ACE expression increased was correlated with the differentiation and stimulation of DCs (59). Elevated ACE is involved in the production of AngII and the peptide repertoire trimming as part of the MHC-II complex (59, 60). Accumulation of AngII due to the shedding of ACE2 leads to the phosphorylation of the ERK and promotes the secretion of IL-6 and TNFα (61). In the meantime, AngII has the ability to improve the migration, maturation, and antigen presentation of DCs to improve Th1 cells (62).

The increase in the AngII level is accompanied by an increase in the aldosterone level because AngII can stimulate the adrenal cortex to secrete aldosterone. DCs are key immune cells involved in severe COVID-19-induced lung damage mediated by aldosterone, which can stimulate DCs to produce IL-6 and transforming growth factor-β1 via the mineralocorticoid receptor (Figure 1) (63, 64). AngII and aldosterone can alter the proliferation and maturation of DCs, leading to DC dysfunction (36). Supplementation of ACE2 may remedy this problem. Recombinant human (rh)ACE2 has a protective effect in SARS-CoV-induced ALI, and injection of rnACE2 can reduce inflammation and improve lung function (65–67). Therefore, rhACE2 may be a potential treatment for SARS-CoV-2-induced ALI.

Yang and colleagues (68) showed that SARS-CoV-2 enters DCs and macrophages through dendritic cell-specific intercellular adhesion molecule 3-grabbing non-integrin (DC-SIGN) and furin rather than through ACE2 and TMPRSS2. DC-SIGN (CD209) is a C-type calcium-dependent lectin and a type-II membrane protein. DC-SIGN consists of three domains: extracellular, transmembrane, and intracellular (69–72). The intracellular domain is an N-terminal domain responsible for the binding, phagocytosis, and intracellular transport of molecules associated with signal transduction. The transmembrane domain anchors DC-SIGN to the cell membrane. The extracellular domain consists of two portions: a neck domain (which forms a tetramer that stabilizes the extracellular part of the molecule) and a c-type carbohydrate-recognition domain (a calcium-dependent receptor with a highly conserved sequence) (43, 69, 73). The extracellular domain is essential for binding and recognizing high-mannose oligosaccharides (43, 74, 75). DC-SIGN is expressed exclusively by mature and immature DCs in the skin, mucosa, and lymphoid organs (76). It is a PRR/adhesion receptor in DCs that promotes migration and adhesion and mediates the inflammatory response by activating innate immune and adaptive immune systems. Remarkably, DC-SIGN has a pivotal role in the “immune escape” of pathogens and tumor cells (70, 71). As an antigen-capture receptor, DC-SIGN recognizes, internalizes, and decomposes antigens and, eventually, presents them to CD4+ T cells to trigger an immune response (77). However, virus-infected DCs, through DC-SIGN, also spread SARS-CoV-2 to other tissues and organs (Figure 1). This phenomenon is typical in human immunodeficiency virus (HIV) infection, where the DC-SIGN signal is activated by the HIV and induces its migration to lymph nodes. In lymph nodes, DC-SIGN promotes the migration of HIV-1 from DCs to CD4+ T cells, thereby promoting cell infection and virus transmission (78–80). Hence, whether this “Trojan horse” phenomenon exists in SARS-CoV-2 infection is worthy of further study.

The role of DC-SIGN in SARS-CoV-2 infection is not known. Recent studies have shown that SARS-CoV binds DC-SIGN through its activated spike protein and takes part in the inflammatory response of DCs (Figure 1) (43, 81, 82). DC-SIGN also plays an important part in Middle East respiratory syndrome-related coronavirus (MERS-CoV) infection and is too crucial to neglect in SARS-CoV-2 infection (83). DC-SIGN could become an alternative receptor for SARS-CoV-2 to invade DCs. It has been suggested that the interaction of SARS-CoV-2 or its envelope proteins (including the products of pyroptosis) with DC-SIGN result in the loss of a specific DC subpopulation (36). Moreover, activated DC-SIGN receptors downregulate the expression of major histocompatibility class (MHC)-II molecules (84), which is a key factor of impaired presentation of antigens. In addition, the combination between DC-SIGN and HIV can promote a Th2 cell-based immune response, thereby reducing levels of IL-12 and IFNs (42), which helps to explain the innate immunosuppression in patients with severe COVID-19.

The DC-SIGNR (L-SIGN, CD209L), which is the homology receptor of the DC-SIGN, is broadly expressed in both endothelial cells and epithelial cells (85, 86). The DC-SIGN/L-SIGN may be an alternative receptor, due to the fact that ACE2-deficient endothelial cells but not DC-SIGNR-deficient endothelial cells allow SARS-CoV-2 entry (85, 86). Michel et al. confirm the important role of DC-SIGN in the Trojan Horse model of DCs during SARS-CoV-2 infection (87). They think that SARS-CoV-2 adheres to the surface of DCs through DC-SIGN and secondly present SARS-CoV-2 to susceptible cells in the process recognized as trans-infection which relies on the characteristics of DCs’ migration (87). Intriguingly, the DC-SIGN gene expression is interestingly decreased in lung DCs but increased in circulation DCs which would undoubtedly increase the amount of SARS-CoV-2 carried by DCs and enhance its global transmission ability (88).

As in HIV infection, DC-SIGN-mediated virus internalization is a critical mechanism for immune escape in SARS-CoV-2 infection (75). DC-SIGN has become another potential invasion portal for SARS-CoV-2, and its function deserves further exploration in COVID-19 patients. However, after SARS-CoV-2 infection, the depletion and dysfunction of DCs result in persistent virus infection. The exact mechanism by which SARS-CoV-2 infection reduces the number and function of DCs is discussed below.

cDC1, cDC2, and pDCs are recruited into lung tissue during infection (39). cDC1s (also known as CD141+ DC) are found in peripheral blood and among resident DCs of the lymph nodes, bone marrow, and spleen (39, 89–93). They participate in cross-presentation of antigens via MHC-I molecules to activate CD8+ T cells and promote Th1 cells and natural-killer-cell responses though IL-12 (39). cDC1s also secrete IFN-III (including IFN-λ) (39, 94). cDC2s (also known as CD1c+ DC) are present mainly in peripheral blood, lymphoid organs, and tissues. They are activated to become robust producers of IL-12 and are excellent cross-presentation cells. cDC2s are also major producers of IL-23, IL-1, TNF-α, IL-8, and IL-10 (39, 95–97). pDCs (which were identified first in peripheral blood and tonsils) “sense” and respond to viral infection through rapid production of IFN-I (39, 98, 99). These three types of DCs are activated by PAMPs and DAMPs to promote the innate immune response and to activate the adaptive immune response.

The number of DCs in the blood of COVID-19 patients is reduced. Among the three types of DCs, only cDC2s accumulate in the lungs of COVID-19 patients (35, 100–103). The mechanisms that result in a decline in the number of DCs may be caused by an alteration in distribution of DCs, increased apoptosis of DCs, and the inhibitory effects of myeloid-derived suppressor cells (MDSCs) (Figure 2).

Not only does the number of DCs decline, but also the ability of DCs to present antigens is impaired during SARS-CoV-2 infection (101, 105, 115, 116). Zhou et al. studied the antigen-presentation capacity of circulating DCs in the acute phase of COVID-19. They found that expression of the co-stimulatory molecule (CD80, CD86) and antigen-presentation molecule (human leukocyte antigen-DR) decreased after stimulation with a “cytokine cocktail” (IL-1β, IL-6, TNF-α, prostaglandin E2) (101). DCs in the acute phase of COVID-19 failed to stimulate the proliferation of CD4+ and CD8+ T cells according to a mixed lymphocyte reaction assay with allogeneic CD4+ and CD8+ T cells (101). The antigen-presentation capability of DCs could act as a bridge between innate immunity and adaptive immunity. Determining the reason for the decline in the antigen-presentation capacity of DCs is crucial, and some reasons are discussed in the following (Figure 2).

IFN-I is a powerful weapon for DCs in their fight against viral infection. However, SARS-CoV-2 often leads to the effects of IFN-I being hampered. Zhou et al. found that the ability of infected cDC1 and cDC2 to produce proinflammatory cytokines (IL-1β, IL-6, TNF-α) was not influenced, but the ability of infected pDCs to secrete IFN-I was suppressed significantly, in COVID-19 patients (101). Achille et al. found that the ability of cDC1 to produce IFN-λ was not impaired, which may contribute to disruption of the epithelial barrier in the lungs upon SARS-CoV-2 recognition (121). Various studies have shown that SARS-CoV-2 inhibits only pDCs from secreting IFN-I but does not affect cDC production of other proinflammatory cytokines (68, 101, 115, 116). An IFN-I deficiency promotes SARS-CoV-2 to escape recognition by the immune system. Such an escape by SARS-CoV-2 leads to the body producing many inefficient antiviral proinflammatory cytokines that will initiate the cytokine storm that damages lung tissues and results in ARDS or death. Three main reasons have been postulated as to how SARS-CoV-2 may specifically impair the production and transformation of IFN-I (Figure 2).

First, the IFN-I level is reduced significantly in coronavirus infections (e.g., SARS-CoV, MERS-CoV) and such inhibition is a hallmark of coronavirus infections (122–124). Little is known about how SARS-CoV-2 suppresses IFN signaling, but clues can be obtained from SARS-CoV. IFN-I production in SARS-CoV-2 infection is even less than that observed in SARS-CoV infection (124). SARS-CoV-2 encodes various proteins that have been shown to inhibit IFN-I signaling in SARS-CoV infection (122, 124) (Figure 2).

Second, some researchers have found congenital deletion of TLR7 in critically ill patients and that the pDCs of patients with a TLR7 deletion cannot produce IFN-I (125). In addition, anti-IFN antibodies have been found in a small number of critically ill patients; anti-IFN antibodies bind to IFN and prevent IFN transformation (126). The failure of the immunometabolism reprogramming such as the impairment of the glycogenolysis-mediated glycolysis which is stored as a source of energy of DCs will impair the IFN-I generation in DCs (108, 127, 128). SARS-CoV-2 inhibits the early induction of glycolysis via suppression of IFN-I generation, and in turn, the impaired induction of glycolysis inhibits IFN-I generation in DCs (129). This positive feedback loop further reduces IFN-I generation.

Third, the high viremia of COVID-19 is associated with impaired generation of IFN-I in peripheral blood mononuclear cells (130). Supplementation of IFN in the early stage of SARS-CoV-2 infection is beneficial to patients. IFN-β combined with lopinavir and ribavirin can improve physical status (131) and alter the production of proinflammatory cytokines in the latter stages of COVID-19 (97, 132). There is growing recognition that inappropriate, excessive, and mistimed IFN treatments are deleterious in viral infections (131, 133). In the early stage of COVID-19, IFN treatment could improve patient outcomes (122, 131, 134–136), but immunomodulation therapy is recommended at the latter stage of COVID-19 (122, 137, 138). The delayed response to IFN-I leads to accumulation of pathogenic mononuclear macrophages, which cause vascular leakage, lung disorders, and inappropriate T-cell responses (139, 140). Understanding the exact immunopathogenesis of impaired IFN production and restoring the IFN system at the early stage of COVID-19 is a top priority.