Abstract
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is associated with diverse host response immunodynamics and variable inflammatory manifestations. Several immune-modulating risk factors can contribute to a more severe coronavirus disease 2019 (COVID-19) course with increased morbidity and mortality. The comparatively rare post-infectious multisystem inflammatory syndrome (MIS) can develop in formerly healthy individuals, with accelerated progression to life-threatening illness. A common trajectory of immune dysregulation forms a continuum of the COVID-19 spectrum and MIS; however, severity of COVID-19 or the development of MIS is dependent on distinct aetiological factors that produce variable host inflammatory responses to infection with different spatiotemporal manifestations, a comprehensive understanding of which is necessary to set better targeted therapeutic and preventative strategies for both.
Abstract
Development of severe COVID-19 versus multisystem inflammatory syndrome relies on distinct aetiological factors that lead to variable host immune responses and inflammatory manifestations, despite following a common trajectory of immune dysregulation. https://bit.ly/3Xtsa7P
Introduction
The management of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, causing the coronavirus disease 2019 (COVID-19) pandemic, has been a major new challenge to healthcare and research. SARS-CoV-2 infection is associated with diverse host response immunodynamics with variable inflammatory manifestations. The development of severe COVID-19 can typically be linked to a number of risk factors (e.g. age-related susceptibility, pre-existing chronic diseases or an immunosuppressed status, amongst others) that cause altered immune profiles [1–7], while another potentially fatal syndrome can develop in previously healthy individuals following SARS-CoV-2 infection, termed multisystem inflammatory syndrome (MIS) [8–10]. Despite a significant overlap in their dysregulated immune responses, these two hyperinflammatory disease phenotypes can be distinguished by a range of distinct spatiotemporal adaptations and clinical profiles. Understanding how different aetiological factors underpin the immunopathogenic pathways and timeline of events in the development of severe COVID-19 and MIS will guide patient stratification and more targeted therapeutic or vaccination strategies. Here, we pinpoint severe COVID-19 as an early acute hyperinflammatory response predominantly affecting the respiratory tract, but also with systemic involvement that is influenced by various risk factors to cause impairment of the primary innate response to infection, as well as dysregulation of downstream innate and adaptive defence mechanisms, and MIS as a delayed acute hyperinflammatory response to SARS-CoV-2 infection with systemic but minimal or no respiratory symptoms and low pulmonary viral loads, which might develop due to genetic susceptibility traits that trigger the activation of systemic autoimmune and hyperinflammatory pathways upon SARS-CoV-2 infection.
SARS-CoV-2 viral entry and primary host response to infection
SARS-CoV-2 binds the peptidase angiotensin converting enzyme 2 (ACE2) via its spike (S) glycoprotein [11–13] and SARS-CoV-2 enters cells through direct fusion of the virion and cell membranes via cleavage of S by the proteases transmembrane protease serine 2 (TMPRSS2) and furin [13–16] or via endocytosis of the SARS-CoV-2 virion in the absence of these proteases [13]. Expression of ACE2, TMPRSS2 and furin (and additional factors that may further support SARS-CoV-2 infectivity and tropism) varies widely according to tissue type, age and biological sex, with different disease conditions, pregnancy or with genetic differences, influencing host susceptibility to infection and viral replication [1, 7, 9, 17–22].
SARS-CoV-2 is a cytopathic virus that causes death and injury of infected cells. Local inflammatory responses extend to lymph nodes and are facilitated by local dendritic cells (DCs) that subsequently travel to draining lymph nodes for antigen presentation to T-cells there and possibly induce immune responses in the spleen [23–25]. Detection of viral RNA by DCs and local macrophages leads to their initiation of a prompt canonical antiviral type 1 interferon (IFN-1) response [26, 27]. IFN-1 signalling concludes with the induction of IFN-stimulated gene (ISG) transcription programmes that interfere with viral replication and activate host immune responses for resolution of infection [28, 29]. Human IFN-1s (including 13 IFN-α subtypes and a single IFN-β subtype) are typically produced at mucosal surfaces by local tissue-resident antigen-presenting cells, with downstream activation of innate and adaptive antipathogenic mechanisms. However, a late, noncanonical hyperinflammatory IFN-1 response can have deleterious immunomodulatory effects that promote viral replication and cause severe complications [28, 30–33]. The timing of IFN-1 signalling relative to peak virus replication is therefore a critical determinant of protective or pathogenic host immune responses [30, 34].
Clinical characteristics of the COVID-19 spectrum and MIS
The COVID-19 spectrum spans from mild–moderate disease to a severe or critical status. The disease course may be classified by four progressive and overlapping phases [35–37], although this classification remains to reach a fully stipulated consensus. First, there appears to be a post-exposure viral phase that may be asymptomatic or mild. During the second phase, detectable upper respiratory viral load decreases as the infection progresses from the upper to the lower respiratory tract, inducing viral pneumonia, which is accompanied by the generation of antibody responses. There is SARS-CoV-2 replication in the upper airways early in the disease course, followed by active replication in the lungs for up to 2 weeks [38]. However, recent data suggests that viral replication in the lower respiratory tract occurs at low levels due to scarce alveolar ACE2 expression and primarily results from a small number of alveolar epithelial type-2 (AT2) cells [39]. Most patients with efficient antiviral defence responses achieve viral clearance and recovery. In patients that develop severe/critical disease, incompetent immune response mechanisms during the second phase are coupled with elevated pro-inflammatory cytokine and acute phase marker release, but the extent to which the dysregulated inflammatory response directly correlates to an incompetent antiviral response is unclear [39, 40]. A third phase corresponds to a state of hypercoagulability in severely affected patients and a fourth phase of multiorgan involvement, damage and failure may follow [35].
A mild−moderate disease course consists of an upper respiratory tract infection with or without symptoms of fever, cough, malaise and possible rare gastrointestinal symptoms [41]. Severe COVID-19 is associated with pneumonia, pulmonary inflammation and injury with significant hypoxia, which leads to some features of acute respiratory distress syndrome (ARDS), hyperinflammation-mediated disruption of the epithelial–endothelial barrier leading to hypercoagulation, vasculopathy, multiorgan damage and circulatory and multiorgan failure [42–44]. Other hyperinflammatory manifestations can include myocardial injury, acute rheumatic manifestations, rhabdomyolysis, septic shock, acute hyperglycaemia, acute renal and hepatic injury, encephalopathy, stroke, and death [23, 41, 45] (figure 1). Excessively high levels of circulating pro-inflammatory cytokines, increased expression of acute phase markers, extensive lymphopenia (with decreasing lymphocyte counts corresponding to disease progression and severity), neutrophilia, coagulopathy and vasculopathy are used as markers of severe disease [26, 44, 46]. SARS-CoV-2 viral persistence is associated with poor outcomes [47].
MIS (originally described in children and adolescents <21 years as MIS-C and subsequently in adults >21 years as MIS-A) is a febrile multisystem hyperinflammatory syndrome that displays features of Kawasaki disease (an acquired paediatric vascular disease), toxic shock syndrome, cardiac dysfunction, acute gastrointestinal conditions and encephalopathy. It is associated with elevated pro-inflammatory cytokines, lymphopenia, neutrophilia, abnormal coagulation indices and multiorgan involvement. While MIS emulates several clinical characteristics of severe COVID-19, it evades severe respiratory illness but has prominent cardiovascular, gastrointestinal and haematological involvement and manifests as a delayed response 2–12 weeks following SARS-CoV-2 infection [8, 48–52] (figure 1).
Immunopathogenesis of SARS-CoV-2-induced disease phenotypes
Mild–moderate COVID-19 and asymptomatic SARS-CoV-2 infection
In patients with mild–moderate COVID-19, there is an early, transient IFN-α wave in the circulation (but an absence of IFN-β) [26, 27], albeit with a degree of viral antagonism by SARS-CoV-2 [53–55]. Resultant antiviral protection is induced via the initiation of prompt and robust innate and adaptive antiviral mechanisms [23], leading to viral clearance and recovery (figure 2a).
A novel SARS-CoV-2 human challenge model, with a low inoculum dose, in healthy young adults (aged 18–29) without any known immune-modulating risk factors, was used to establish viral kinetics over the course of primary infection with SARS-CoV-2 [56]. Such human challenge models can generate critical information via the controlled investigation of pathogenesis, linking early antiviral responses and inflammatory responses to both viral replication and host genetics, identification of host factors associated with protection in those who resist or recover well from infection, as well as testing the efficacy of vaccines and therapeutics.
SARS-CoV-2 infection can also be asymptomatic, with similarities in immunodynamics to mild disease [57–60]. However, despite having no clinical symptoms and normal chest radiography imaging, some of these asymptomatic individuals display subclinical lung abnormalities with characteristic pulmonary ground glass opacification on computed tomography imaging [57] that is indicative of significant pulmonary inflammation and could therefore have long-term harmful implications. Indeed, there are reports of patients suffering with diverse, prolonged multisystem involvement, persistent symptoms and significant disability, without having recovered by up to 7 months post-infection [61, 62]. These post-COVID-19 sequelae arise from prolonged inflammation and coagulopathy and are often referred to as long COVID, which is independent of initial disease severity [63–67]. This suggests that inflammatory responses can be marked and persistent despite early viral clearance, hinting at a potentially indirect correlation between the effectiveness or potency of an initial antiviral response and the subsequent inflammatory response induced.
An early acute response with hyperinflammatory pathology: severe COVID-19
Immune-modulating risk factors associated with severe COVID-19
Severe COVID-19 is associated with impaired host antiviral mechanisms, persistent blood viral loads and a disproportionate systemic pro-inflammatory response. Individuals with chronic inflammatory processes, prothrombotic states, atherogenic profiles with a reduced cardiorespiratory reserve or predisposing genetic defects may be susceptible to fatal outcomes due to altered immune response states that compromise host antiviral immune response mechanisms and/or activate hyperinflammatory responses [1, 7, 18–20, 28, 68] (table 1). Age-related susceptibility has been indicated as a major risk factor contributing to COVID-19 severity. As a result of numerous age-related phenomena (e.g. increased ACE2 expression patterns, weakened antiviral IFN-1 responses, inflammaging, immunosenescence and comorbidities), elderly adults are at particularly high risk of developing severe disease that requires admission to intensive care units and implementation of mechanical ventilation, and have a higher mortality risk [2, 3, 69]. Conversely, the majority of children without comorbidities have a mild disease course or asymptomatic infection, likely due to a combination of several factors (e.g. lower tissue ACE2 and TMPRSS2 expression in children, higher availability of naïve T-cells that can respond to new infections, and “trained immunity” due to frequent exposure to common viral respiratory pathogens and/or childhood vaccinations) [25, 70]. However, severe cases and deaths have also been reported in children, with infants being most vulnerable to severe disease and accounting for the highest proportion of hospitalisation. This may be linked to transiently increased ACE2 expression at birth, combined with poor IFN-1 responses upon viral infection, altered T-helper type 1 (Th1) function and low expression of cytotoxic and inflammatory mediators in neonates [25, 71]. Increased severity and mortality rates in all ages are also linked to comorbidities with pro-inflammatory diathesis (e.g. chronic respiratory conditions, cancer, obesity, diabetes, cardiovascular disease, hypertension, chronic kidney disease, sickle cell disease and autoimmune diseases), an immunosuppressed status and high occupational viral exposure [5–7, 44, 72, 73], regardless of age. Many of these chronic diseases are associated with higher ACE2 expression in the lungs, as well as IFN-1 dysregulation and inefficient innate and adaptive responses, which negatively affect antiviral host mechanisms and the ability to respond to new infections [7, 18, 28, 74]. Other factors that can modify host responses to infection include biological sex, pregnancy, certain ABO blood group antigens, as well as genetic variability [2, 4, 68, 75–77] (table 1). SARS-CoV-2 infection thus unmasks the host impairment of key antiviral defence mechanisms and altered regulation of inflammatory responses that would have otherwise been silent in many of these patients, emphasising the multifactorial causes of severe COVID-19. In this regard, SARS-CoV-2 appears to be particularly adept at revealing these traits, which have not been apparent following other infections, possibly due to the vast and accelerated spread of SARS-CoV-2 infection on a global scale within a very short time-frame.
Impaired host innate mechanisms in severe COVID-19
Impaired early IFN-1 signalling is a hallmark of severe COVID-19 and is associated with lower viral clearance. Low or no IFN-α response typically precedes clinical deterioration and transfer to intensive care units, characterising the most severe or critical cases requiring invasive ventilation, with a significant reduction in mean expression of six ISGs defining an IFN-1 signature, compared to mild–moderate disease with high IFN-α levels [26, 78]. Self-renewing tissue-resident alveolar macrophages (AMs), located within the airspace lumen in the lungs, provide the first line of defence against respiratory pathogens entering the respiratory system and, along with lung-resident plasmacytoid DCs (pDCs), produce large amounts of antiviral IFN-1α [79–81]. However, direct SARS-CoV-2 infection of AMs (as well as circulating blood monocytes) has been demonstrated [82–85]. While early in vitro studies suggest that SARS-CoV-2 infection of AMs may be abortive [86], subsequent in vivo studies indicate that infected AMs may be able to support SARS-CoV-2 viral replication, contingent on their polarisation [82, 83]. It is therefore possible that SARS-CoV-2 infection of pro-inflammatory AMs may actively facilitate viral spread, while the opposite might be true for alternatively activated AMs [83]. Notably, recent post mortem studies suggest that AMs are not a significant source of viral replication in the alveolus as they lack the receptors to support viral entry [39], while other studies have reported evidence that is indicative of viral replication in AMs in humanised mice or AMs in patients with severe SARS-CoV-2 [82, 87]. Differential ACE2 expression may explain these differences, which could be influenced by macrophage polarisation. The loss of tissue-resident AMs and recruitment of monocyte-derived inflammatory macrophages have been observed in post mortem lungs and humanised mouse models of COVID-19 [39, 87]. However, post mortem studies also suggest that pulmonary inflammation does not always directly correspond to the level of viral presence in the respiratory epithelium [88], indicating the presence of autonomous inflammatory circuits even following local viral clearance in the lung and viral translocation to other sites at the time of death. Indeed, there are reports of pyroptotic cell death of monocytes and AMs following antibody-mediated SARS-CoV-2 uptake, which contributes to the systemic hyperinflammation in severe COVID-19 [85]. Notably, IFN-1 antagonism, by SARS-CoV-2 proteins, in reprogrammed pro-inflammatory AMs found prevalently in the lungs of severely affected patients, as well as pyroptotic or apoptotic death of these resident AMs, can result in an impaired IFN-1 response [79, 89, 90]. Furthermore, pDCs, activated by the virus without productive infection [91], demonstrate diminished IFN-α production in severe COVID-19 [27]. Additional reports of IFN-1 dysregulation describe genetic errors that impede IFN-1 immunity [92, 93], autoantibodies that bind and functionally neutralise almost all IFN-1s [94, 95], as well as sustained abrogation of antiviral IFN-1 production [78], in patients with life-threatening COVID-19 compared to mild cases or asymptomatic subjects. IFN-1 dysregulation in severe COVID-19 thus occurs through a combination of different impaired host antiviral mechanisms and viral antagonism of IFN-1 induction [28, 31, 53, 55] . Further studies are needed, however, to define the exact cell types in which antiviral responses are defective, the roles of specific cell death paradigms in impairing these responses and the relative contributions of genetic factors and autoimmunity to impaired IFN-1 antiviral responses.
Aberrant IFN-1 responses have been further linked to a second wave of inflammatory cytokine release and the expression of factors that promote pulmonary intravascular coagulopathy and fibrin-based blood clots [96, 97], and a high incidence of multiorgan thrombosis is linked to severe COVID-19 with respiratory failure [98–101]. Excessive cytokine release may promote a hypercoagulable state by inducing endothelial dysfunction, the activation and aggregation of platelets with high thrombogenic capacity, as well as abnormal neutrophil activation, to promote tissue damage, vascular injury and immunothrombosis [102–106]. Endothelial cell activation affects vessel integrity, triggering the release of factors that increase platelet activation and adhesion, generating a procoagulative state [35]. Damaged endothelial cells can promote neutrophil activation and neutrophil extracellular trap (NET) formation [107]. NETs can trigger microvascular thrombosis, which is implicated in COVID-19-related ARDS, multiorgan dysfunction and death [104–106]. Antiphospholipid autoantibodies can also promote thrombosis in vascular beds through neutrophil, endothelial cell and platelet activation [108]. Activated platelets release platelet factor 4 (PF4), while activated endothelial cells release polyanionic proteoglycans (PGs), to form PF4-PG complexes, which can expose PF4 immunogenic epitopes to activate extrafollicular B-cells that secrete PF4 autoantibodies. The subsequent binding of PF4 autoantibodies to PF4-PG immune complexes on platelets and endothelial cells may stimulate their pro-coagulative activities [109]. Platelet activation can instigate their consumption by scavenging macrophages, leading to thrombocytopenia. The combination of thrombocytopenia and thromboembolic complications is associated with critical COVID-19 and increased mortality [110]. Autoantibodies targeting IFN-1s, antiphospholipids, PF4, as well as natural killer (NK) cells, CD8+ cytotoxic T-lymphocytes (CTLs), B-cells and macrophage-expressed proteins, have been associated with increasing COVID-19 severity [94, 95, 108, 111, 112]. Autoantibody production is typically genetically predetermined and can therefore shape clinical presentation following infection [113]. Further studies should correlate the level of these responses to COVID-19 severity.
Impaired host adaptive mechanisms in severe COVID-19
Canonical IFN-1 signalling contributes to the differentiation of T follicular helper (Tfh) cells, which are critical for germinal centre (GC) reactions, during which B-cells undergo developmental changes and acquire memory [31]. Severe COVID-19 cases exhibit defective Bcl-6+ Tfh cell differentiation and do not develop the lymph node and splenic GCs required for durable antiviral responses [24]. Abnormal Tfh cell differentiation may lead to extrafollicular B-cell activation in critically ill patients [114], with greater antibody titres corresponding to increasing COVID-19 severity and a pathogenic antibody response [114, 115]. This may be through antibody-dependent enhancement (ADE), whereby absent or sub-neutralising antibody concentrations relative to viral load (potentially due to impaired GC reactions) fail to neutralise SARS-CoV-2, thereby permitting virus–antibody complexes to bind to phagocyte Fcγ receptors (FcγRs) to enable viral uptake, replication and a rapidly increasing viral load [24, 85, 114–116]. However, while the FcγR-mediated uptake of antibody-coated SARS-CoV-2 by monocytes and macrophages triggers their pyroptotic death, thereby aborting viral replication, it nevertheless promotes systemic hyperinflammation, which contributes to acute lung injury, multiorgan damage, vascular leak and respiratory distress [85]. ADE might occur alone or in combination with impaired cellular responses [114, 117] that are associated with various risk factors (table 1).
Lymphocyte cytotoxic effector functions play protective roles that are critical for resolution of SARS-CoV-2 infection [26, 118, 119] and CD8+ CTLs may compensate for aberrant humoral immunity in COVID-19 [120–122]. However, functional exhaustion of CTLs and NK cells (while NK cells classically form part of the innate response, adaptive characteristics of NK cells following infection have been identified over the past decade), impaired lymphocyte cytotoxic capacity and progressive lymphopenia correlate with increasing disease severity and increased levels of circulating pro-inflammatory cytokine levels in SARS-CoV-2 infection [123, 124]. Pro-inflammatory cytokines can promote the functional exhaustion of NK cells, inhibiting their cytotoxicity, whilst simultaneously enhancing neutrophil infiltration and activity [125–128]. The S glycoprotein can also directly suppress NK cell killing via crosstalk with infected lung epithelial cells [129]. Finally, since IFN-1 signalling promotes CTL and NK cell effector functions [130–132], aberrant canonical IFN-1 signalling with inadequate induction of lymphocyte cytotoxic functions may promote inflammatory pathogenesis through ineffective viral clearance.
Innate-adaptive crosstalk in severe COVID-19
Impaired lymphocyte effector functions can promote viral spread within the body, thereby enhancing pathogenic inflammation. Damaged DNA released from infected host cells can be recognised by the adaptor molecule, STING (stimulator of IFN genes), which activates antiviral IFN regulatory factor 3 (IRF3) and/or inflammatory NF-κB pathways, terminating with IFN-1 induction. However, DNA release following SARS-CoV-2 infection activates NF-κB but not the antiviral IRF3 system, leading to exaggerated production of NF-κB-mediated cytokines, which may lead to the paradoxical upregualtion of ISGs in severe COVID-19 [133–135]. This noncanonical IFN-1 signalling later in the disease pathway represents an alternative pathway of response (in particular, it results in IFN-β expression, which is absent from all patients in early SARS-CoV-2 infection), and can result in inappropriate hyperinflammatory signalling without effective resolution of infection [26, 133, 136–139].
Pro-inflammatory cytokines can suppress lymphopoiesis while inducing myelopoiesis [140] and sustained release can lead to abnormal macrophage activation [97, 141–143]. Evidence of emergency myelopoiesis with production of excessive reactive oxygen species (ROS)- and nitric oxide synthase (NOS)-expressing myeloid cells has been implicated in severe COVID-19 with respiratory failure [141, 143, 144]. Key features include greater percentages of highly inflammatory monocyte and macrophage populations in peripheral blood and bronchoalveolar lavage fluid, markedly decreased HLA-DR (human leukocyte antigen–DR isotype; responsible for antigen presentation to T-cells) expression in CD14+ monocytes and lesions of histiocytic hyperplasia with haemophagocytosis and acute alveolar damage [141, 142, 144, 145]. Death of classically activated tissue-resident AMs can be triggered upon SARS-CoV-2 infection and there is recruitment of pro-inflammatory monocytes into the lungs and their transformation into pathogenic hyperinflammatory hyperferritinemic macrophages in patients with severe SARS-CoV-2 pneumonia [82, 145]. This altered macrophage composition within the alveolar compartment contributes to a detrimental loop of pro-inflammatory cytokine release, with a shift towards excessive ROS generation to cause widespread oxidative stress and tissue damage [79, 97, 145, 146]. Dysregulated macrophage activation is therefore a key contributor to the hyperinflammatory status in severe COVID-19. However, viral replication in the alveolar compartment appears to be limited to rare AT2 cells; and thus the promotion of inflammation by AMs may be more contingent on SARS-CoV-2 virions translocating to the alveoli from the upper airways, followed by endocytosis of virions/virally infected cells, driving the induction of specific pro-inflammatory phenotypes that can promote subsequent systemic responses [39]. Furthermore, any viral replication within the lungs could lead to increased infiltration of inflammatory immune cells [82, 147, 148]. Indeed, while the exact origin of the virus-promoting response has been debated, SARS-CoV-2-infected AMs produce T-cell chemoattractants, and upon recruitment and activation, T-cells induce inflammatory cytokine release from macrophages and further promote T-cell activation to form a positive feedback loop that drives persistent inflammation and tissue injury [82].
Notably, a late and persistent noncanonical IFN-1 response may further promote the infiltration of pathogenic hyperinflammatory monocyte-derived macrophages that replace tissue-resident AMs in the lungs, thereby enhancing lung immunopathology, vascular leakage and immune cell dysfunction [30, 32]. However, some studies have reported that, in patients with severe SARS-CoV-2 pneumonia, IFN-1 expression by cells in the alveolar compartment was not detected later in the clinical course [82], suggesting that the hyperinflammatory response may be generated systemically. Reasons for these differences in origin of responses remain unclear. Nevertheless, an IFN-1-associated prothrombotic neutrophil hyperinflammatory signature has been identified in COVID-19 ARDS [148]. As such, it appears as though severe COVID-19 cases lack early, mucosal, canonical (classical) antiviral IFN-1 signalling, leading to insufficient antiviral host immune responses and persistence of viral loads [26]. However, later on, a systemic, noncanonical IFN-1 pathway can be triggered to promote NF-κB-mediated hyperinflammation with uncontrolled pathogenic pro-inflammatory cytokine release that augments the dysregulation of host antiviral functional responses, whilst causing widespread tissue damage, coagulopathy and vasculopathy [133, 137, 149, 150] (figure 2b). The complications associated with severe COVID-19 may therefore be strongly related to prolonged exposure to circulating pathogenic inflammatory cytokines [82] and may become independent of viral replication, as we and others have suggested [39, 40].
Delayed acute systemic hyperinflammatory response: MIS
There is significant overlap of the immune perturbations in MIS with those observed in severe COVID-19, including impaired antigen presentation, abnormal B-cell and CD4+ T-cell responses, dysregulated cytotoxic lymphocyte effector functions with evidence of exhausted CTL and downregulated NK cell signatures, and hyperinflammatory macrophage activity [151–159] (figure 2c). However, the signatures of these two conditions illustrate dysregulation of inflammatory and antiviral immune defence mechanisms with distinct temporal patterns [9, 160]. While COVID-19 symptoms develop within a median of 6.57 days after viral exposure (with slightly shorter incubation with Omicron variants) and progress over the following week in severe disease [161, 162], MIS is diagnosed 2–12 weeks following initial SARS-CoV-2 infection, with accelerated progress to critical status [163]. MIS patients often demonstrate negative SARS-CoV-2 reverse-transcriptase PCR results with positive SARS-CoV-2 serology. A positive test result for SARS-CoV-2 infection within the preceding 12 weeks of presentation with severe illness requiring hospitalisation and laboratory evidence of severe inflammation [8–10] indicates a delayed hyperinflammatory response several weeks after initial asymptomatic/mild SARS-CoV-2 infection. MIS patients exhibit prominent systemic features with cardiovascular, gastrointestinal and haematological clinical manifestations whilst lacking severe localised respiratory illness and present with systemic disease and significant extrapulmonary organ dysfunction [8, 9, 49, 52, 156] (figure 1).
MIS patients often exhibit a lack of, or minimal, respiratory tract viral loads [8, 10] and develop a delayed immune activation syndrome driven by persistent antigen presence, with widespread viral persistence in extrapulmonary tissues [164]. In paediatric patients that develop MIS-C, the lack of severe respiratory illness could be attributed to fewer infected cells in the respiratory tract, due to lower ACE2 expression in the lungs in children [25, 157, 165, 166]. By extrapolation of the available evidence [8, 52, 160, 167], it is possible that initial low respiratory tract viral loads, and therefore lower immune activation, could explain the scarcity of severe pulmonary symptoms in MIS-A. Lower levels of cell lysis and release of virions may induce a sufficient antiviral host response to limit pulmonary disease [157] in both paediatric and adult cases. However, the inability of the host to achieve complete viral clearance from the respiratory tract [164] may allow extrapulmonary viral spread via the vascular system in predisposed individuals [156, 168, 169]. There is also evidence of direct SARS-CoV-2 infection of the vascular endothelium [38]. Furthermore, increased markers of immune cell activation and egress to the periphery have been identified in MIS-C patients [156], which may be accompanying viral migration and could contribute to the delayed post-infectious immune dysregulation in MIS. Since MIS was initially described in children, studies investigating its pathophysiology have largely focussed on MIS-C and therefore much of the evidence we provide here is based on studies investigating MIS-C. However, there appear to be no clear differences in case definitions of MIS in children and adults [8, 10, 170], likely indicating the same disease entity.
While the vast majority of healthy individuals without any pre-existing comorbidities have a mild COVID-19 disease course, we propose that healthy subjects possessing (to date, unidentified) intrinsic genetic susceptibility traits that trigger harmful maladaptations of key adaptive immune responses to SARS-CoV-2 infection may develop MIS. The resultant abnormal immune responses may permit viral spread to extrapulmonary tissues and promote systemic hyperinflammatory responses, driving MIS pathogenesis (figure 2c). This speculation is supported by several lines of evidence (albeit predominantly from small studies that may require further validation), including 1) dysregulation of several adaptive response pathways in MIS patients, 2) the autoimmune disease phenotype of MIS, 3) potential SARS-CoV-2 superantigenic activity being a contributor to its development, 4) the likely genetic components in the aetiology of hyperinflammatory syndromes such as Kawasaki disease (KD), with which MIS shares many features, and 5) the presence of vascular patrolling CTLs in MIS patients, which have previously been linked to inflammatory conditions with cardiovascular damage.
Dysregulation of various mechanisms that interfere with adaptive immune compartments have been described in MIS patients. These include abnormal CD4+ helper T-cell, cytotoxic lymphocyte and B-cell responses, which may promote viral spread and persistent antigen presence and lead to ADE, thereby prolonging inflammation [151–157, 164]. Furthermore, it is possible that self-DNA release from SARS-CoV-2-infected host cells could cause a STING-mediated late, noncanonical IFN-β response as seen in severe COVID-19 and KD [134, 137, 169] that leads to hypercoagulation, vasculopathy, multiorgan involvement and injury. Antigen-presenting cells responsible for antiviral IFN-1 production demonstrated augmented levels of phospho-signal transducer and activator of transcription 3 in MIS-C, which may be indicative of noncanonical IFN-1 signalling that can restrain antiviral responses [156, 171, 172]. Additional reports of sustained NF-κB and tumour necrosis factor-α activation, myocardial infiltration of hyperinflammatory macrophages and increased neutrophil activation in MIS-C patients with severe myocarditis [173] further support the likelihood of a late noncanonical hyperinflammatory IFN-1 response with subsequent NF-κB-mediated inflammatory pathogenesis in these patients [133, 134]. The immune dysregulation in MIS is therefore comparable to severe COVID-19, although the timing and localisation presentation represent a different disease phenotype with a delayed and systemic nature (figures 1 and 2).
MIS-C has been characterised as having an autoimmune disease phenotype with dysregulated B-cell responses and autoantibody production. This can lead to enhanced neutrophil activation and augmentation of complement and coagulation pathways, which are implicated in several systemic, autoimmune and inflammatory vascular diseases [174]. Multiple autoantibodies targeting endothelial cells, the gastrointestinal tract and immune mediators that implicate organ systems central to MIS-C pathology have been linked to its pathogenesis [152, 156] and autoantibody generation is often genetically predetermined [113]. These autoantibodies may trigger extensive immune complex formation that cannot be quickly eliminated [156]. Once deposited in tissues and perivascular spaces, these immune complexes may be capable of causing widespread inflammatory injury and vascular permeability via complement activation, Fc receptor-mediated responses and cytokine network dysregulation [169].
SARS-CoV-2 superantigenic activity, with the capacity to induce autoimmune pathways, has been indicated as another contributor to the development of MIS-C [175–177], representing a further pathomechanism that is shared with severe COVID-19. Superantigen-reactive hyperinflammatory CD4+ Th cells can promote macrophage hyperactivation to induce relentless pro-inflammatory cytokine secretion [169, 178]. Additional MIS manifestations such as conjunctivitis, oedema, rash and fever are symptoms observed in the context of superantigen-mediated responses, subsequent to elevated cytokine release [179, 180].
Hyperinflammatory syndromes mainly occur in genetically susceptible individuals, as demonstrated for the paediatric febrile vasculitis syndrome, KD, which is commonly associated with post-infectious epidemiology and with which MIS shares several resemblances [180, 181]. Indeed, it has been suggested that MIS-C could be a new presentation of KD that is triggered by SARS-CoV-2 infection [182], despite certain prominent differences between classical KD and MIS [151, 152, 156]. Furthermore, a few recent studies have suggested that rare genetic variants could be potential contributors to MIS pathogenesis, although larger studies would be required to interrogate and corroborate these findings [182–185].
Finally, vascular patrolling CX3CR1+ CTLs (associated with increased cardiovascular disease risk and implicated in inflammatory conditions with vascular damage) [186, 187] have been identified in MIS-C [164]. Since the hyperinflammatory response in MIS involves endothelial cell activation and associated coagulopathy, we presume the presence of the vascular patrolling CTLs to be secondary to endothelial cell activation, which we speculate may occur both directly as a result of SARS-CoV-2-mediated effects on the vascular endothelium [38], as well as a secondary response to systemic hyperinflammation [188]. The localisation of immune cells to the vascular endothelium is an important step in atherogenesis [186] and there is further evidence of endothelial injury and coronary artery immune cell infiltrates in pathological samples from MIS-C patients [180, 188]. Nonetheless, the precise mechanisms responsible for the spatiotemporal control of immune responses in MIS remain elusive and this topic requires further investigation.
Aetiology and immunological evolution of severe COVID-19 versus MIS
There appears to be a continuum of immune dysregulation across the COVID-19 spectrum and MIS, varying in extent, timing and localisation, but severity of COVID-19 or the development of MIS is dependent on a range of unique aetiological factors that can result in variable efficacy of host responses to infection. An early hyperinflammatory response with significant respiratory manifestations differentiates severe COVID-19 from MIS, the latter being a delayed hyperinflammatory response with pronounced systemic manifestations, but no significant respiratory illness (figures 1 and 2). COVID-19 and MIS therefore represent distinct disease phenotypes, with distinct spatiotemporal adaptations (figure 3). Severe COVID-19, associated with immune-modulating risk factors that affect host antiviral immune mechanisms [1, 2, 4–6] (table 1), is a consequence of perturbations in the primary innate response to infection, with further disruption of downstream innate and adaptive antiviral mechanisms that promote dysregulated inflammatory responses [26, 27, 133, 138, 139]. Conversely, MIS typically affects previously healthy individuals without any comorbidities [8]. It is possible that patients who develop MIS may have fewer infected respiratory tract cells (reflected in the diminutive/absent respiratory tract viral loads and lack of severe respiratory illness in these patients) [157, 169]. Primary innate response to infection is not impaired in MIS, but affected individuals most likely possess one or more intrinsic genetic susceptibility traits that remain inactive until triggered by SARS-CoV-2 infection [182] to cause harmful maladaptations of adaptive immune responses [151–157, 164]. This may potentiate viral translocation to extrapulmonary tissues via the vascular system, followed by activation of systemic autonomous hyperinflammatory loops to drive pathogenesis [133, 134, 156, 173] (figure 3).
Currently, a case-by-case approach is necessary for the management of both diseases. Generally, while glucocorticoids are used in both severe COVID-19 and MIS patients, high-risk patients can receive outpatient treatment with oral antivirals such as paxlovid, molnupiravir or intravenous remdesivir early after infection, while remdesivir (with or without dexamethasone and baricitinib) can be administered in hospitalised patients or those with hypoxemia [189–195]. This reflects the fact that epidemiological factors allow identification of high-risk patients and the rationale for interrupting earlier stages of viral replication is better established. Improved comprehension of the immunopathogenic pathways underpinning severe COVID-19 and MIS (caused by their respective unique etiological factors) may thus help to further stratify patients for more targeted treatment strategies. For example, identifying high-risk patients with mild–moderate disease, who have been hospitalised for a reason other than COVID-19, but may be at risk of moderate–severe lung disease and might benefit from treatment with monoclonal antibodies [78, 162]. For later presentations, anti-inflammatory agents (e.g. tocilizumab or baricitinib) are well established [196–198] and further stratification may allow selection of patients requiring treatment with various other immunomodulatory agents at later stages of disease. In MIS patients, anti-inflammatory strategies can be used, although with some differences, and intravenous immunoglobulin and thromboprophylaxis with low-dose aspirin is also considered [199–205]. Combinations of more directed anti-inflammatory and immunosuppressive therapies may also be beneficial for the treatment of MIS, given its autoimmune-like immunopathology.
Conclusion
Severe COVID-19 and MIS have distinct spatiotemporal profiles while sharing several immunological characteristics. The defects in host antiviral immune competence in both can contribute to dysregulated functional immune mechanisms and induce sustained hyperinflammation, coagulopathy, vasculopathy and multiorgan involvement. Given the epidemiological/aetiological factors underpinning the variable efficacy of host responses to infection and the timeline of events in the development of severe COVID-19 versus MIS, it will be important to determine the likely intrinsic defects in antiviral immunity against SARS-CoV-2 in populations at high risk of developing severe COVID-19 or MIS. It would also be important to establish whether vaccination can correct for any intrinsic defects in antiviral immunity against SARS-CoV-2 in some of these populations or whether the features of susceptibility persist in all high-risk groups despite vaccination, since some groups with altered immune profiles (e.g. those with an immunosuppressed status, obese subjects or the older population) have been shown to generate less effective responses to vaccination. A comprehensive understanding of the immunopathogenic pathways of severe COVID-19 and MIS, dictated by their respective aetiologies, can therefore enable the identification of superior targeted therapies and anti-inflammatory approaches that induce the most effective antiviral immune responses, as well as the development of improved vaccination strategies.
Points for clinical practice
Severe COVID-19 and MIS have distinct spatiotemporal profiles while sharing several clinical characteristics (including hyperinflammation, hypercoagulability, vasculopathy and severe multiorgan dysfunction).
Severe COVID-19 represents an early acute hyperinflammatory response to SARS-CoV-2 infection that develops within 1 week of viral exposure and is associated with several immune-modulating risk-factors including older age, noncommunicable chronic diseases, immunosuppression and pregnancy, among others.
Severe COVID-19 is associated with pneumonia, significant pulmonary damage and respiratory distress, and subsequent systemic complications.
Identification of high-risk patients with mild–moderate disease who have been hospitalised for a reason other than COVID-19 but may be at risk of severe pulmonary complications, may benefit from monoclonal antibody treatment and, for later presentations, patient stratification may allow selection of patients who might benefit from treatment with various other immunomodulatory agents.
MIS represents a delayed acute hyperinflammatory response to SARS-CoV-2 infection that develops 2–12 weeks following initial SARS-CoV-2 infection in previously healthy individuals who may be genetically predisposed.
MIS is a febrile hyperinflammatory syndrome without severe respiratory illness but prominent cardiovascular, gastrointestinal and haematological perturbations, and other diffuse systemic manifestations with multisystem involvement and an autoimmune-like immunopathological signature.
In MIS patients, directed anti-inflammatory and immunosuppressive therapies may be beneficial for the treatment of MIS, given its autoimmune-like immunopathology.
Questions for future research
Who is at high risk of developing long COVID-19 and why is it independent of initial disease severity?
What are the likely intrinsic defects in antiviral immunity against SARS-CoV-2 in populations at high risk of developing severe COVID-19 or MIS? Are there any characteristic differences between different susceptible demographic groups?
Can vaccination correct for intrinsic defects in antiviral immunity against SARS-CoV-2 in at least some of the populations at high risk of developing severe COVID-19?
What are the specific contributory genetic risk factors implicated in MIS development?
Can vaccination correct for intrinsic defects in antiviral immunity against SARS-CoV-2 in populations at high risk of developing MIS?
Acknowledgements
We are grateful to the members of the COVID-19 Disease Map project, a large-scale community effort to describe SARS-CoV-2 virus–host interaction mechanisms, and this review was prepared with a motivation to support this project.
Footnotes
Provenance: Submitted article, peer reviewed.
Author contributions: R. Fraser, A. Orta-Resendiz and A. Mazein developed the scope and focus of the review. R. Fraser wrote the article and prepared the figures and table. R. Fraser, A. Orta-Resendiz, D. Dockrell, M. Müller-Trutwin and A. Mazein critically appraised, revised and prepared the final version.
Conflict of interests: The authors declare no competing interests.
Support statement: A. Orta-Resendiz was supported by a doctoral fellowship from the University of Paris-Cité.
- Received October 14, 2022.
- Accepted December 31, 2022.
- Copyright ©The authors 2023
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org