Introduction

Coronavirus Disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) rapidly spread worldwide and was declared a pandemic in early 2020. COVID-19 destroyed people’s mental and physical health and staggered global economic growth. As of May 18, 2021, 163 million infections, including 3.38 million deaths, have been recorded (source: World Health Organization). SARS-CoV-2 invades the host by virtue of angiotensin-converting enzyme 2 (ACE2) receptors broadly distributed on various tissues and immune cells.1,2,3,4,5 The virus can cause a wide range of clinical manifestations from mild forms such as fever, cough, and myalgia to moderate forms requiring hospitalization (pneumonia and localized inflammation) to severe/critical forms with fatal outcomes.6,7 Severe or critical infection often manifests as pneumonia,8,9 disseminated intravascular coagulation (DIC), acute respiratory distress syndrome (ARDS), low blood pressure, and multiorgan failure (Fig. 1).9,10,11

Fig. 1
figure 1

A systemic clinical manifestations of COVID-19. SARS-CoV-2 infects airway epithelial cells or immune cells via binding to ACE2 receptors, causing tissue damage and release of DAMPs, as well as production of inflammatory cytokines by epithelial cells and immune cells. Then, the crosstalk between epithelial cells and immune cells leads to a wide range of clinical manifestations, from mild forms (e.g., fever, cough, and myalgia); to moderate forms requiring hospitalization (pneumonia and localized inflammation); to severe/critical forms with a fatal outcome that are manifested as pneumonia, ARDS, DIC, CS, and multiorgan failure. DAMP danger-associated molecular pattern, ARDS acute respiratory distress syndrome, DIC disseminated intravascular coagulation

Several lines of evidence have shown that immunopathological damage may be responsible for the deterioration of COVID-19. Particularly, multiple studies have reported that highly elevated levels of pro-inflammatory cytokines are produced during the crosstalk between epithelial cells and immune cells in COVID-19, which has linked the cytokine storm (CS) with the severe complications and poor outcomes in this infection.12,13,14

CS is a fast-develo**, life-threatening, clinical condition in which the overproduction of inflammatory cytokines and excessive activation of immune cells lead to complicated medical syndromes from a persistent fever, nonspecific muscle pain, and hypotension, to capillary leak syndrome, DIC, ARDS, hemophagocytic lymphohistiocytosis (HLH), multiorgan failure, and death if treatment is not adequate.15 Therefore, the timing of diagnosis and treatment of CS could be life-saving. The term CS was first used in 1993 in graft-versus-host disease,16 and later, in many inflammatory diseases such as autoimmune conditions, organ transplantation, cancer chimeric antigen receptor (CAR) T cell therapy, and, most recently, in COVID-19.17,18,19,20,21,22,23 However, the profile and causative effect of CS in different conditions can greatly vary. Thus far, precise diagnosis and treatment guidelines for CS in most of the conditions are lacking. Understanding the definite alterations and pathogenic roles of individual cytokines involved in the COVID-19-related CS (COVID-CS) is hence extremely important for the development of precise diagnosis and effective treatment.

Although some aspects of this topic have been partly reviewed previously, a comprehensive view of COVID-CS to facilitate its diagnosis and treatment is still lacking with unmet clinical needs. Herein, we provide an updated and full scenario of COVID-CS from basic research to clinical diagnosis, treatment, and trials. Initially, we discuss the currently identified immunopathological features of COVID-19, especially the CS; its mechanism of action and differences with respect to CS in other disease conditions; and individual cytokines involved in the COVID-CS including their pathological role, downstream signaling, and existing interventions. In addition, the challenges and prospects in the diagnosis and treatment of COVID-CS are also discussed.

The immunopathology of COVID-19

In general, patients with COVID-19 present with an abnormal immune landscape, characterized by overactivated inflammatory, innate immune response, and impaired protective, adaptive immune response. This is primarily responsible for the immunopathology of severe COVID-19. Thus far, evidence from both clinical trials and basic research has revealed several key features of immunopathology in severe COVID-19, including lymphopenia, antibody-dependent enhancement (ADE), neutrophilia, dysregulation of monocytes and macrophages, reduced or delayed type I interferon (IFN-I) response, and CS (Fig. 2).

Fig. 2
figure 2

The key immunopathology of severe COVID-19. The immunopathological manifestations of COVID-19 include lymphopenia, dysregulation of monocytes and macrophages, neutrophilia, ADE, reduced or delayed IFN-I response, and CS. Lymphopenia is commonly observed in severe COVID-19. In addition to decreased counts, lymphocytes often exhibit exhaustion phenotypes with the expression of higher levels of exhaustion markers PD-1, Tim-3, or NKG2A. Peripheral monocytes present a phenotype shift from CD16+ to CD14+, and BALF macrophages are increased with a blood-to-BALF transition course. Neutrophil counts are increased with the presence of neutrophil precursors in peripheral blood, especially in patients with severe COVID-19. The possible existence of ADE enhances the entry of SARS-CoV-2 into cells through interaction between Fc regions and Fc receptors, leading to the aggravation of COVID-19. A CS is characterized by highly elevated levels of pro-inflammatory mediators and is a particularly central feature for poor outcomes in patients with severe or critical infection. Reduced or delayed IFN-I response impedes viral clearance and induces paradoxical hyperinflammation, thus leading to the deterioration of prognosis in COVID-19 patients. BALF bronchoalveolar lavage fluid, ADE antibody-dependent enhancement

Lymphopenia

Lymphopenia was commonly found in COVID-19 patients10,11 and is closely correlated with the illness severity.24,25 Laboratory results showed that the counts and percentages of lymphocytes including CD4+ T, CD8+ cytotoxic T,26,27 natural killer (NK),24 and B cells25 were all reduced in COVID-19 patients.28,29,30,31 Evidence from single-cell sequencing,32 flow cytometry,33 and nonhuman primate models of COVID-1934 confirmed the involvement of lymphopenia to the maximum extent. In addition, T cells exhibited exhaustion phenotypes with the expression of higher levels of exhaustion markers including programmed cell death protein-1 (PD-1) and T cell immunoglobulin and mucin domain-3, suggesting that these T cells may have regulatory activities.26,27 Other investigations showed that the PD-1- or NK group 2 member A-positive NK cells were also significantly increased in the peripheral blood of COVID-19 patients compared to healthy controls.35,36

Several mechanisms may contribute to SARS-CoV-2-induced depletion and exhaustion of lymphocytes: (1) SARS-CoV-2 may directly infect T cells via ACE2 receptor expressed on T cells,37 which results in T cell death;38,39 (2) several pro-inflammatory or anti-inflammatory cytokines can accelerate the depletion and exhaustion of T cells with their respective functions. In addition, the virus may destroy secondary lymphoid tissues spleen and lymph nodes, leading to lymphopenia, which is supported by the observations of splenic atrophy, lymph node necrosis, and decreased lymphocyte numbers;24,40,41,42 (3) a nonhuman primate COVID-19 model showed that the impaired anti-viral T cell response may be attributed to the increased numbers of regulatory T cells (Tregs).34 However, we cannot exclude the possibility that depletion and exhaustion of lymphocytes resulted from anti-inflammatory therapies such as the administration of glucocorticoids.

Collectively, lymphopenia may represent a poor outcome of the illness. A retrospective, multicenter, emergency clinical trial in a Han Chinese population including 14,117 normal controls and 69 hospitalized COVID-19 patients (25 severe or critical and 44 mild) showed that lymphopenia occurred in almost 100% of the severe/critical cases, and the number of lymphocytes progressively decreased with the progression of the disease and deterioration of clinical status. The CD4+ and CD8+ T lymphocyte subsets showed a marked difference between mild and severe cases. Therefore, the authors suggested that analyzing the counts and percentages of lymphocytes at admission immediately contributes to improved clinical prognosis.43

Antibody-dependent enhancement

B cells are considered protective in SARS-CoV-2 infection by producing neutralizing antibodies against the ACE2 receptor to prevent viral entry.44 However, B cell-produced neutralizing antibodies are not necessarily protective, depending on the virus element it targets and also the Fc region of the antibody. ADE is a phenomenon wherein pre-existing neutralizing antibodies targeting one serotype of a given virus enhance the entry of not only that virus but also another serotype of the virus into leukocytes through the interaction between the Fc regions of the antibody and Fc receptors or complement receptors on such cells.45,46,47,48,49 The ADE phenomenon has been found in various viral infections such as dengue, Flavivirus, SARS-CoV, MERS-CoV, and Ebola.45,48,50,51,52

Recent studies have shown that relatively high levels of B cells or antibodies are positively associated with COVID-19 severity,32,53,54 implying the potential involvement of ADE in SARS-CoV-2 infection. In addition, a study has reported that the monoclonal antibody MW05 targeting SARS-CoV-2 may also induce ADE activity by binding to FcγRIIB receptors on the target cells in vitro. However, administration of an engineered antibody with Fc region mutation in vivo effectively protected animals from SARS-CoV-2 infection.55 This highlights the importance of develo** protective neutralizing antibodies against SARS-CoV-2. Nevertheless, further investigations about ADE in SARS-CoV-2 infection are required to facilitate the development of vaccine- or antibody-based therapy for COVID-19.

Neutrophilia

An increase in neutrophil count in COVID-19 patients is widely recognized. It is well known that under normal conditions, neutrophils play a protective role against infections by producing neutrophil extracellular traps (NETs) to kill extracellular pathogens;56 however, excessive neutrophil activation can also damage the surrounding cells and dissolve connective tissues.57

An earlier clinical trial including 138 patients from Wuhan, China, showed that neutrophil counts were increased in non-survivors compared to survivors and continued to increase until death in the non-survivors.11 Another clinical study that integrated transcriptomic, proteomic, and metabolomic platforms showed that neutrophil counts were increased in patients with severe, but not mild, COVID-19 as compared to healthy controls, and molecules associated with NETs were significantly upregulated in severe COVID-19 cases.58

The increased neutrophils manifested as both increased numbers of mature and immature cells. In a clinical trial that integrated single-cell RNA-sequencing with single-cell proteomics of blood and peripheral blood mononuclear cells (PBMCs), immature neutrophil precursors, and dysfunctional mature neutrophils expressing programmed death-ligand 1 appeared in severe COVID-19 cases.59 In addition, a single-cell sequencing analysis by Wilk et al.32 and a flow cytometry analysis by Ronit et al.33 also identified the appearance of neutrophil progenitors at various developmental stages in PBMCs or bronchoalveolar lavage fluid (BALF) of COVID-19 patients with ARDS.

Although the mechanism by which the virus promotes neutrophil development in COVID-19 is still poorly understood, McElvaney et al.60 found that the levels of pyruvate kinase M2 (PKM2), a regulator of glycolysis61 and coactivator of hypoxia-inducible factor-1α,62 as well as phosphorylated PKM2 were higher in the neutrophils of COVID-19 patients in the ICU than in those of non-ICU COVID-19 patients. This indicates that neutrophils undergo immunometabolic reprogramming in severe COVID-19 cases, which represents a potential intervention target for excessive neutrophil generation and activation in severe or critical COVID-19.

Dysregulation of monocytes and macrophages

Monocytes and macrophages are the major innate immune cells in infection and inflammation not just by virtue of their higher numbers but also by their functions. A single-cell RNA-sequencing analysis showed that classic CD14+ monocytes were significantly increased, whereas nonclassical CD16+ monocytes and intermediate CD14+CD16+ monocytes were remarkably reduced in the blood of COVID-19 patients with severe symptoms. Classical monocytes can differentiate into macrophages in tissue to initiate an inflammatory response, whereas nonclassical monocytes were viewed as anti-inflammatory as they can maintain vascular homeostasis,63 which may explain the phenotypic shift of circulating monocytes from CD16+ to CD14+. Analysis of the differentiation profiles of BALF and circulating monocyte–macrophages from the same patient revealed a transition course of blood-toward-BALF. More importantly, multiple pro-inflammatory cytokines and chemokines were highly expressed by the BALF monocyte–macrophages, suggesting that the cells are inflamed.64 Another single-cell sequencing analysis of peripheral blood samples also showed that CD16+ monocytes were remarkably depleted in COVID-19 patients with ARDS, with a phenotypic shift from CD16+ to CD14+. However, significant upregulation of genes encoding pro-inflammatory cytokines or chemokines were not found in peripheral monocytes, indicating that peripheral monocytes may not be responsible for the progression of CS in COVID-19.32 Moreover, phenoty** leukocyte subpopulations in BALF and blood of COVID-19 patients with ARDS showed that the expression of activation markers such as CD16, CD64, CD69, and HLA-DR was higher in BALF macrophages than in peripheral macrophages.33 Collectively, these existing studies were generally consistent and revealed the course of blood-toward-BALF transition and the contribution of pulmonary monocyte–macrophages to CS via the release of multiple pro-inflammatory cytokines and chemokines during severe COVID-19.

Interestingly, a two-cohort study showed that activated HLA-DRhighCD11chighCD14+ monocytes were increased in the PBMCs of patients with mild COVID-19, whereas dysfunctional HLA-DRlowCD163high (indicative of anti-inflammatory function) CD14+ monocytes were observed in severe COVID-19 cases.59 This merits further investigation to understand the underlying mechanism and clinical significance.

Reduced or delayed IFN-I response

The IFN-I response is the first line of protective response and critical to combat viral infections by promoting viral clearance and regulating innate and adaptive immune responses.65 Although the detailed mechanism is still unknown when the infection occurs, the RNA of SARS-CoV-2 virus may be recognized by innate immune cells via pattern recognition receptors (PRRs) including toll-like receptor (TLR); retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs)/melanoma differentiation-associated gene 5 (MDA5); and NOD-like receptors (NLRs).66 Subsequently, downstream transcription factors including nuclear factor-κB (NF-κB), activator protein-1 (AP-1), and IFN regulatory factor 3/7 (IRF3/7) are activated to promote the transcription of pro-inflammatory cytokines and IFN-I. The IFN-I can activate the Janus kinase 1 (JAK1)/tyrosine kinase 2–signal transducer and activator of transcription 1/2 (STAT1/2) pathway, promoting the formation of the STAT1/2/IRF9 complex and initiating transcription of IFN-stimulated genes (ISGs) (Fig. 3).66,67

Fig. 3
figure 3

The signaling pathways for the production and function of IFN-I after SARS-CoV-2 infection. After infection, the genomic ssRNAs and replicative dsRNA intermediates of SARS-CoV-2 are recognized by endosomal toll-like receptors TLR3, 7, 8, and cytosolic RNA sensors, RIG-1/MDA5; next, downstream transcription factors including NF-κB and IRF3/7 are activated to induce the production of pro-inflammatory cytokines and IFN-I. IFN-I can activate the JAK1/TYK2–STAT1/2 pathway, promoting the formation of the STAT1/2/IRF9 complex and initiating the transcription of ISGs to produce anti-virus mediators, and it can also nonconventionally activate inflammatory pathways such as NF-κB and MAPK pathways to induce the expression of pro-inflammatory cytokines and paradoxical hyperinflammation in COVID-19

However, accumulating evidence has suggested that the protective IFN-I response was remarkably reduced in severe COVID-19 patients.26,68,69 At least two mechanisms have been proposed to explain the deficient IFN-I response: (1) previous studies have suggested that SARS-CoV employs various mechanisms to inhibit IFN response, especially through the components of its structural proteins such as M protein,70 N protein,71 open-reading frame 3a (ORF3a) protein,72 and ORF6 protein.73 Considering that the structure of SARS-CoV-2 is similar to that of SARS-CoV, it can be speculated that SARS-CoV-2 may exert similar effects on IFN response. For example, Yang et al.74 found that the NSP1 protein of SARS-CoV-2 can inhibit STAT1 phosphorylation and ISG transcription. (2) Decreased plasmacytoid dendritic cells (pDCs) may be partly responsible for the deficient IFN-I response. Sufficient evidence has suggested that pDC is a prominent producer of IFN-I upon viral infection.75,76,77 However, existing studies showed that counts of pDCs were decreased in the blood of COVID-19 patients, especially in severe cases.26,32

Of note, although the levels of systemic IFN-I were low, local IFN-I and ISGs were noticeable in the BALF of some critically ill patients,78 which are related to the phenomenon of delayed IFN-I response.69 Evidence has confirmed that a delayed IFN-I response not only impedes viral clearance but also induces paradoxical hyperinflammation, thereby aggravating the immunopathological response.79,80 Collectively, these studies suggest that IFN-I-based therapy for COVID-19 should be applied as early as possible after the infection is confirmed.

Cytokine storm

In this section, we systemically review and discuss the characteristics, possible induction mechanism, pathogenesis, and diagnosis of CS in COVID-19.

Sufficient evidence has revealed the components and characteristics of CS in the patients with severe COVID-19, which are composed of an array of cytokines including interleukin-1 (IL-1), 2, 6, 7, 8, 10, 12, 17, 18; tumor necrosis factor-α (TNF-α); IFN-γ; granulocyte colony-stimulating factor (G-CSF); granulocyte–macrophage colony-stimulating factor (GM-CSF); and monocyte chemoattractant protein-1 (MCP-1).26,33,60,81,82,83,84,85,86 Reports of hemophagocytosis and clinical benefits from cytokine-targeted therapies in severe COVID-19 patients further confirmed the existence and pathogenesis of COVID-CS.87,88 Existing evidence has characterized and distinguished COVID-CS from CS in variable conditions such as HLH induced by specific viral infections,89 macrophage activation syndrome (MAS) occurring after autoimmune disorders,90 and cytokine release syndrome (CRS) caused by CAR T cell therapy23 in several aspects. First, COVID-CS involves more inflammatory cytokines than other CS conditions, thereby providing an explanation for the aggressive nature of COVID-19. Second, lymphopenia, although relatively less frequent in other CS, was often observed in patients with COVID-CS,88 suggesting that COVID-CS may be mainly attributed to innate—rather than adaptive immune cells. Finally, compared with bacterial infection-induced CS (e.g., sepsis), the treatment of COVID-CS is more challenging, because blocking inflammatory cytokine function without effective anti-viral drug support may exacerbate the infection.

The initiation of COVID-CS induction during infection and the predominant causative cytokine in COVID-19 immunopathology remain largely unknown. Despite the lack of definite pathogen-associated molecule pattern (PAMP) of SARS-CoV-2, in analogy with SARS-CoV and MERS-CoV, it can be speculated that upon cellular entry of SARS-CoV-2 via its ACE2 receptor, viral genomic single-stranded RNA or other RNA compositions (double-stranded RNA) as PAMPs can be sensed by the related PRRs, including TLRs and RLRs in host cells. The downstream transcription factors IRF3/7 and NF-κB are activated following PAMP recognition to induce the production of IFN-I and pro-inflammatory cytokines, respectively.91,92,93 However, as mentioned above, the protective IFN-I response is quickly and selectively abrogated by SARS-CoV-2 via different mechanisms. This is accompanied by an overwhelming production of pro-inflammatory cytokines in the context of COVID-19, which not only impairs viral clearance but also promotes paradoxical hyperinflammation including CS. Therefore, from the immunology perspective, COVID-CS may be an unfortunate event whereby the intended host immune response combating the SARS-CoV-2 has lost control and transformed into an inflammatory type.15

In SARS-CoV-2 infection, the virus infects the respiratory epithelial tissue and activates local innate immune cells to release inflammatory cytokines such as IL-1, IL-6, IL-8, IL-12, TNF-α, and other chemokines. These inflammatory cytokines and chemokines then recruit more innate immune cells (monocytes, macrophages, neutrophils, DCs, and NK cells) and activate adaptive immune cells (CD4+ and CD8+ T cells) from the peripheral tissues to produce sustained inflammatory cytokines like IL-2, IFN-γ, and TNF-α, which induce myelopoiesis and emergency granulopoiesis that further aggravate lung and epithelial damage (Fig. 4). In addition, overproduction of systemic cytokines, particularly IL-2, IFN-γ, GM-CSF, and TNF-α, triggers macrophage activation (i.e., MAS) and erythro-phagocytosis (i.e., HLH), resulting in anemia,94,95 as well as causes perturbation of coagulation and vascular hemostasis, resulting in capillary leak syndrome, thrombosis,96 and DIC. These events together lead to ARDS, multiorgan failure, and death (Fig. 4).15 Of note, the host immunoregulatory system is usually capable of retaining and fine-tuning the protective inflammation to an appropriate level. Regulatory cells such as Tregs97,98 can produce regulatory cytokines like IL-10 and tumor growth factor-β to antagonize overactivated immune responses.99,100 However, aggressive inflammatory conditions such as CS cannot be calmed by the regulatory system’s natural ability.

Fig. 4
figure 4

The immunopathological mechanisms of COVID-CS. SARS-CoV-2 infects the epithelial cells or immune cells, causing tissue damage and release of inflammatory cytokines (e.g., IL-1, IL-6, IL-12, and TNFα) by epithelial cells and immune cells. These inflammatory cytokines then recruit innate immune cells (monocytes, macrophages, neutrophils, DCs, and NK cells) and activate adaptive immune cells (CD4+ T cells and CD8+ T cells) to induce the occurrence of myelopoiesis and emergency granulopoiesis, as well as the production of sustained and excessive circulating cytokines that can further aggravate epithelial damage. In addition, overproduction of systemic cytokines triggers macrophage activation (i.e., MAS) and erythro-phagocytosis (i.e., HLH), resulting in anemia and gives rise to perturbation of vascular hemostasis, resulting in capillary leak syndrome, thrombosis, and DIC. These events together lead to ARDS, multiorgan failure, and death. HLH hemophagocytic lymphohistiocytosis, MAS macrophage activation syndrome, ARDS acute respiratory distress syndrome, DIC disseminated intravascular coagulation

Although the immunological and pathological understanding of COVID-CS has provided valuable information regarding the development of diagnosis and treatment strategies, detailed guidelines are still lacking. Develo** scoring systems such as HScore, MS score, HLH-2004, Penn grading scale, and the Common Terminology Criteria for Adverse Events may be beneficial to predict COVID-CS or related outcomes. Caricchio et al.101 proposed predictive criteria for COVID-CS diagnosis. These criteria comprise three clusters: (1) albumin <2.87 mg/mL, lymphocytes <10.2%, neutrophil absolute count >11.4 × 103/mL; (2) alanine aminotransferase >60 IU/L, aspartate aminotransferase >87 IU/L, d-dimer >4930 ng/mL, lactate dehydrogenase >416 U/L, troponin I >1.09 ng/mL; and (3) anion gap <6.8 mmol/L, chloride >106 mmol/L, potassium >4.9 mmol/L, and blood urea nitrogen:creatinine ratio >29. In addition, ferritin >250 ng/mL and C-reactive protein (CRP) >4.6 mg/dL are added for the reassurance of ongoing systemic inflammation. In another study, the authors proposed that a diagnostic criterion including peripheral blood oxygen saturation to the fraction of inspired oxygen (SpO2/FiO2), CRP, ferritin, cytokines/chemokines, and neutrophil/lymphocyte ratio may have a strong diagnostic power for COVID-CS.102 Mehta et al.13 proposed that prospective screening for hyperinflammation using laboratory assays and the HScore should be performed in all severely ill COVID-19 patients to identify COVID-CS. Despite the requirement for further validation, these criteria indeed provide constructive suggestions for the development of officially recognized guidelines for COVID-CS.

COVID-CS is a complicated and dynamic inflammatory process caused by a group of cytokines from initiation, immune cell hyperactivation, to organ dysfunction. The development of precise therapeutic intervention in appropriate time is required to effectively control COVID-CS. In principle, the treatment strategy is to control ongoing inflammatory response by specifically or nonspecifically targeting inflammatory cytokines or related signaling pathways and to resume the host immunoregulatory system. Herein, we discuss the role of the key cytokines and associated signal pathways involved in COVID-CS (Fig. 5).

Fig. 5
figure 5

Inflammatory signaling cascades activated in COVID-CS. SARS-CoV-2 enters the host cells and is sensed by toll-like receptors (TLRs)3, 7, 8; RIG-I-like receptor, RIG-I or MDA5; and NOD-like receptor, NLRP3, that can also be directly activated by viral proteins or ROS released by apoptotic or inflamed cells. The downstream transcription factors IRF3/7 are activated to induce the production of IFN-I and related paradoxical hyperinflammation; NF-κB is activated to induce the production of pro-inflammatory cytokines; and NLRP3 inflammasome is activated to induce the production of mature IL-1β and IL-18. Pro-inflammatory cytokines such as IL-6, IL-2, TNF-α, and IFN-γ in turn activate the JAK-STAT or NF-κB signaling via binding to their receptors expressed on immune cells to induce more production of pro-inflammatory genes, forming a positive feedback to trigger the threshold of CS. Conversely, regulatory cytokines like IL-10 are compensatorily produced to antagonize immune hyperactivity

IL-6/JAK/STAT signaling

A retrospective, multicenter study including 150 patients from Wuhan, China showed significantly elevated levels of IL-6 in severe COVID-19 cases.103 A study from Germany showed that IL-6 >80 pg/mL in combination with CRP >97 mg/L presented a relatively high specificity and sensitivity to predict respiratory failure.104 In addition, other studies have also shown the remarkably increased serum levels of IL-6 in patients with severe COVID-19.68,105,106

IL-6, first produced by monocytes, macrophages, and DCs, serves as a prominent activator of the JAK/STAT3 pathway in the context of inflammation. Recent studies have determined that the IL-6–JAK–STAT3 axis is closely related to the severity of COVID-19,107,108 and the levels of phosphorylated STAT3 were higher in different subsets of leukocytes in COVID-19 patients than in healthy controls.109 IL-6 deploys two signaling pathways—classic cis-signaling and trans-signaling—to trigger the activation of downstream JAK/STAT3 signaling through the membrane-bound (mIL-6R) and soluble form of IL-6R (sIL-6R), respectively.110 In cis-signaling, IL-6 binds to mIL-6Rs that are restrictedly expressed on immune cells, forming an IL-6/IL-6R/gp130 complex to activate downstream JAK/STAT3, Akt/mTOR, and MAPK signaling. This exerts pleiotropic effects on immune cells, which are manifested as promoted differentiation of T-helper type 17 (Th17), CD8+ T, and B cells; increased migration of neutrophils; and reduced development of Tregs.111,112 These, in turn, induce increased secretion of IL-6 and aggravate inflammation. In trans-signaling, circulating IL-6 binds to sIL-6Rs to form a complex, then binds with the gp130 dimer that is expressed on almost all cell types. The resultant activation of the JAK–STAT3 signaling occurs in cells with absent expression of mIL-6R, such as endothelial cells and vascular smooth muscle cells (VSMCs). The overwhelming activation of the IL-6–IL-6R–JAK–STAT3 pathway triggers the secretion of various mediators, such as IL-6 itself, IL-8, MCP-1, and vascular endothelial growth factor (VEGF), and reduces the expression of E-cadherin expressed on endothelial cells.113 Several studies have shown that MCP-1 can facilitate the formation of atherogenesis,114 production of adhesion molecules,115,116 and proliferation and migration of VSMCs,Blockade of NF-κB signaling

The overwhelming expression of multiple pro-inflammatory proteins in COVID-CS has indicated the central roles of pro-inflammatory signaling pathways, and in particular, the NF-κB pathway. Immunomodulation of NF-κB activation has been proven effective in SARS-CoV-infected cells or mice.146 Thus, recent reviews have proposed that the NF-κB pathway represents a potential therapeutic target for critical COVID-19 illness.336,337,338

An in vitro study showed that phillyrin (KD-1), a well-studied anti-inflammatory and antioxidative agent, significantly reduced the replication of SARS-CoV-2 and expression of pro-inflammatory factors in Huh-7 cells via inhibition of the NF-κB signaling pathway.339 Another study showed that a novel pyrazole derivative remarkably reduced the expression of IL-6, TNFα, and IL-1β in LPS-stimulated RAW267.4 cells by inhibiting NF-κB signaling pathway activation.340 It has been suggested that blocking phosphorylation of the inhibitor of NF-κB kinase subunit beta, a primary downstream protein of NF-κB signaling, with pharmacological inhibitors may be an effective approach for COVID-CS treatment.339 Moreover, Liu Shen capsules, a traditional Chinese medicine, were also reported to exert anti-viral and anti-inflammatory effects in SARS-CoV-2-infected Huh-7 and Vero E6 cells, respectively, by suppressing the NF-κB signaling cascade.341 The effect of several anti-inflammatory or anti-viral drugs on COVID-19 such as dexamethasone,342 hydroxychloroquine,343 macrolide antibiotics,344,345 and N-acetylcysteine346,347 are also related to NF-κB cascade inhibition.

Despite the existence of various nonselective agents for NF-κB inhibition, develo** selective NF-κB inhibitors and a series of clinical trials are urgently required to further validate the clinical benefits.

Blockade of NLRP3 signaling

Several studies have shown that NLRP3 inflammasome is a potential therapeutic target for COVID-CS.164,348,349

NLRP3 signaling inhibition may be a potential mechanism of action for several anti-inflammatory drugs effective in COVID-19, such as colchicine. Studies have shown that colchicine can nonselectively inhibit NLRP3 inflammation by inhibiting the activation of P2X7 receptor or the interaction between NLRP3 protein and ASC.350,351 In addition, chloroquine352 and curcumin353 are also capable of inhibiting NLRP3 signaling. Several investigational agents capable of inhibiting NLRP3 activation, such as tranilast,354 dapansutrile (OLT1177, selective inhibitor),354 and thiazolo-alkenyl sulfonylurea derivative 7,355 are also being considered for COVID-CS treatment as reviewed by Freeman and Swartz.164 In addition, some nonselective or selective agents against NLRP3 inflammasome including melatonin (NCT04409522), OLT1177 (NCT04540120), açai palm berry extract (Euterpe oleracea) (NCT04404218), and ZYIL1 (selective inhibitor) (NCT04731324) are under investigation.

Interventions targeting multiple cytokines and pathways

Intravenous immunoglobulin (IVIg) therapy

IVIg is a natural immunoglobulin pool with a highly diverse antibody network and is administered to superimpose over a patient’s imbalanced repertoire caused by infections.356 It has been known for a while that a broad range of natural anti-cytokine autoantibodies such as those against IL-1, IL-6, and IFN-γ can be detected in the IVIg of healthy individuals.357,358,359,360 Although how the autoantibodies are induced is still poorly understood, it has been demonstrated that many of the anti-cytokine autoantibodies are neutralizing antibodies and may be responsible for the anti-inflammatory effect of IVIg in inflammatory and autoimmune disorders.

The potential efficacy of IVIg therapy was reported in SARS and the 2009 H1N1 influenza pandemic.361,362 In a randomized, controlled trial including 84 COVID-19 patients, IVIg treatment did not demonstrate any therapeutic benefits in severe cases; however, a significant positive relationship between the number of days from admission to IVIg treatment and the length of hospitalization was observed,363 which indicated the potential clinical benefit of IVIg administration during the early stage of COVID-19. In contrast, Suzuki et al.364 reported the potential efficacy of IVIg administration along with mechanical ventilation, methylprednisolone, favipiravir, ivermectin, and tocilizumab therapy in the late phase in an elderly patient with severe COVID-19.

In general, these studies were inconsistent with respect to the timing of IVIg administration, i.e., early phase vs. late phase; therefore, a larger number of clinical trials are warranted.

Corticosteroid treatment

Corticosteroids are one of the most commonly used anti-inflammatory drugs in the treatment of many inflammatory disorders. They exert immunoregulatory effects by inhibiting the expression of multiple pro-inflammatory cytokines and activation of various immune cells. At the beginning of 2020, corticosteroid treatment in COVID-19 was either contraindicated or not recommended,365 because of the statement that no clinical data indicated a benefit from corticosteroid treatment. Moreover, it even increased the mortality and secondary infection rates in SARS-CoV and MERS-CoV.366 In March 2020, the RECOVERY trials, one of the largest randomized, controlled trials for COVID-19 treatments including ~15% of all hospitalized COVID-19 patients in the UK, were launched. The dexamethasone arm enrolled 2104 patients receiving a low-to-moderate dexamethasone dose of 6 mg per day for 10 days and the control arm comprised 4321 patients receiving standard care. Compared to the control group, dexamethasone treatment reduced the 28-day mortality by one-third in mechanically ventilated patients and by one-fifth in patients receiving oxygen only, but not in patients with no need for ventilated support.367 Therefore, the UK government (https://www.gov.uk/government/news/world-first-coronavirus-treatment-approved-for-nhs-use-by-government/) and the National Institutes of Health in the United States (https://www.covid19treatmentguidelines.nih.gov/dexamethasone/) have authorized the standard use of dexamethasone in hospitalized COVID-19 patients who require oxygen. In addition, a meta-analysis of seven clinical trials showed that corticosteroid treatment was associated with a lower 28-day all-cause mortality in critically ill COVID-19 patients.225

Collectively, the clinical benefit of glucocorticoids in COVID-19 treatment is based on the selection of the correct dose, correct patient, and appropriate timing. Several studies have shown the temporal dynamics of viral shedding in SARS-CoV-2.368,369,370 After the replicative peaks, immunopathological factors may play a dominant role in the illness, while active viral replication may play a secondary one. Therefore, the viral load may serve as an indicator to determine the precise time of glucocorticoid treatment. Moreover, in contrast to other agents, dexamethasone is readily available worldwide at a low cost, which is beneficial in develo** countries with limited access to health care. However, although corticosteroid improves clinical syndromes in critically ill COVID-19 patients, its impact on CS is still unclear and requires further investigation.

Traditional Chinese medicine (TCM) treatment

Previous studies have reported the therapeutic effects of several TCMs such as Lizhong Decoction,371 Liujunzi Decoction,372 and Huanglian Jiedu Decoction373 on inflammatory diseases. Recently, TCM treatment has also been shown effective for COVID-19 and contributed substantially to control the pandemic in China. For example, a multicenter, randomized, controlled trial by Nanshan Zhong’ research team showed that Lianhuaqingwen capsule, a repurposed Chinese herb, can ameliorate clinical symptoms and shorten the recovery time in COVID-19 patients with no serious adverse effects.374

Several TCMs may play an immunosuppressive role to treat CS via multiple cytokines or pathways related to the CS. Yang et al.378 Previous studies have shown that blockade of CDK7 manipulated inflammation resolution by remodeling antitumor immunity378 and regulating cytokine secretion.379 A recent study by Wei et al.380 showed that small-molecule inhibitor, THZ1, mediated the blockade of CDK7 and thus mitigated hyperinflammatory states and CRS caused by CAR T cell therapy. Mechanistically, when CDK7 is blocked by THZ1, various pro-inflammatory genes, especially STAT1 and IL1 that are regulated by CDK7/RNA Pol II super-enhancers, are preferentially suppressed at the transcriptional level. This indicates that blockade of CDK7 may be a promising strategy to treat CRS. Considering similar cytokine profiles between COVID-CS and CRS, we can hypothesize that the strategy may also work in COVID-CS.

Conclusions

Clinical and basic research studies have identified and characterized COVID-CS, which has greatly enhanced our understanding of CS and related immunopathology in COVID-19. A full scenario of COVID-CS is now emerging, and it appears to be much larger in scale and contains more cytokines than the CS recognized in other conditions, and is therefore also more damaging. Although it is still unclear how the virus turns the protective cytokine profile into an inflammatory CS, the cytokines appear to be produced predominantly by innate cells because the lymphopenia was frequently reported in this condition. While the causative efficacy of individual cytokines on the development of certain immunopathogenic parameters in this condition is yet to be fully understood; it is obvious that COVID-CS as a whole is closely associated with the major pathogenic changes of COVID-19. To treat COVID-CS, several biologic interventions specifically targeting inflammatory cytokines or related signaling pathways have been clinically evaluated with promising results and many others are in the pipeline.

In principle, the treatment strategy should be to control ongoing inflammatory cytokine production or activity and resume the host’s homeostasis. However, we still lack safe and effective drugs to control the CS, and clinically, the treatment of CS has been proved difficult for several reasons: (1) many medical doctors are not aware of the condition and hence, clinical diagnosis and treatment guidelines are currently lacking; (2) it is a pharmaceutical challenge to simultaneously target multiple cytokines. It is therefore important to identify and target the key cytokines upstream and the cytokine induction network, or directly target the predominant cytokine-producing cells such as monocytes and macrophages. Alternatively, learning from the host immunoregulatory system and identifying more effective and safer anti-inflammatory factor/cytokines with a wide-spectrum inhibitory effect may provide a better option for therapeutic intervention; IL-37 has been suggested for this purpose;381,382,383,384 (3) it is difficult to balance CS and protective immunity in infectious diseases, as the appropriate level of inflammatory cytokines is protective against infections and inappropriately targeting inflammatory cytokines may lead to acquired immunodeficiency and subsequent infections; (4) differences among individuals with respect to age, immune status, and other comorbidities may result in virtual differences in the component and scale of the CS and treatment. Therefore, precise treatment is required. Of note, ideally targeting predominant cytokines or molecular pathways in a particular CS condition should be conducted first and in a timely manner.

COVID-19 has taught us a critical lesson regarding how to deal with natural pathogenic enemies. Knowledge and treatment options developed from COVID-CS will be invaluable, not just for this disease but also for other CS conditions.