Introduction

The first cases of coronavirus disease 2019 (COVID-19) likely occurred from a zoonotic transmission in China in December 2019, linked to a large seafood market that also traded in live wild animals. The causative virus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is capable of human-to-human transmission and spread rapidly to other parts of China and then to other locations. By 24 March 2020, SARS-CoV-2 had infected more than 381,000 people across 195 countries/regions and killed more than 16,000: a pandemic as declared by the World Health Organization1. Daily reports of sharp rises in the number of new cases continue to emerge from many countries/regions, but efforts to overcome the virus are hampered by a lack of knowledge of several important aspects of SARS-CoV-2 infection, ranging from pathogen biology to host response and treatment options. Therefore, there is an urgent need to better understand the host–pathogen biology of COVID-19 as this will offer important insights into treatment and management of the disease, including identification of new therapies. Here, we review the literature on SARS-CoV-2 pathophysiology, its interaction with target cells and the immune response to the virus, including the contribution of dysfunctional immune responses to disease progression. Specifically, we highlight the implications of specific features of the infection for promising therapeutic interventions that could target the virus or the dysfunctional immune response. Moreover, we discuss how studies focused on the adaptive immune response will be crucial in informing the development of vaccines and therapeutic monoclonal antibodies.

Pathogenesis of COVID-19

Coronaviruses are known to cause disease in humans and animals. Among these, four (human coronaviruses 229E, NL63, OC43 and HKU1) typically infect only the upper respiratory tract and cause relatively minor symptoms2. However, there are three coronaviruses (severe acute respiratory syndrome coronavirus (SARS-CoV), Middle East respiratory syndrome coronavirus (MERS-CoV) and SARS-CoV-2) that can replicate in the lower respiratory tract and cause pneumonia, which can be fatal. SARS-CoV-2 belongs to the betacoronavirus genus. Its closest relative among human coronaviruses is SARS-CoV, with 79% genetic similarity3. However, among all known coronavirus sequences, SARS-CoV-2 is most similar to bat coronavirus RaTG13, with 98% similarity4, and coronavirus sequences in the pangolin (a scaly anteater) also share high similarity5.

Like the other respiratory coronaviruses, SARS-CoV-2 is transmitted primarily via respiratory droplets, with a possible, but unproven, faecal–oral transmission route. On infection, the median incubation period is approximately 4–5 days before symptom onset6,7,8,9, with 97.5% of symptomatic patients developing symptoms within 11.5 days8. At the point of hospital admission, patients with COVID-19 typically exhibit a fever and dry cough; less commonly, patients also experience difficulty in breathing, muscle and/or joint pain, headache/dizziness, diarrhoea, nausea and the coughing up of blood6,10,11,12,13,14,15. Within 5–6 days of symptom onset, SARS-CoV-2 viral load reaches its peak — significantly earlier than that of the related SARS-CoV, where viral load peaks at about 10 days after symptom onset16,17,18,19. Severe COVID-19 cases progress to acute respiratory distress syndrome (ARDS), on average around 8–9 days after symptom onset11,20.

The pathophysiology of SARS-CoV-2 infection closely resembles that of SARS-CoV infection, with aggressive inflammatory responses strongly implicated in the resulting damage to the airways21. Therefore, disease severity in patients is due to not only the viral infection but also the host response. The pattern of increasing severity with age is also broadly consistent with the epidemiology of SARS-CoV and MERS-CoV6,11,14.

ARDS seen in severe COVID-19 is characterized by difficulty in breathing and low blood oxygen level22. As a result, some patients may succumb to secondary bacterial and fungal infections14. ARDS may lead directly to respiratory failure, which is the cause of death in 70% of fatal COVID-19 cases22. In addition, the vast release of cytokines by the immune system in response to the viral infection and/or secondary infections can result in a cytokine storm and symptoms of sepsis that are the cause of death in 28% of fatal COVID-19 cases22. In these cases, uncontrolled inflammation inflicts multi-organ damage leading to organ failure, especially of the cardiac, hepatic and renal systems (Fig. 1). Most patients with SARS-CoV infection who progressed to renal failure eventually died23.

Fig. 1: Chronology of events during SARS-CoV-2 infection.
figure 1

When severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infects cells expressing the surface receptors angiotensin-converting enzyme 2 (ACE2) and TMPRSS2, the active replication and release of the virus cause the host cell to undergo pyroptosis and release damage-associated molecular patterns, including ATP, nucleic acids and ASC oligomers. These are recognized by neighbouring epithelial cells, endothelial cells and alveolar macrophages, triggering the generation of pro-inflammatory cytokines and chemokines (including IL-6, IP-10, macrophage inflammatory protein 1α (MIP1α), MIP1β and MCP1). These proteins attract monocytes, macrophages and T cells to the site of infection, promoting further inflammation (with the addition of IFNγ produced by T cells) and establishing a pro-inflammatory feedback loop. In a defective immune response (left side) this may lead to further accumulation of immune cells in the lungs, causing overproduction of pro-inflammatory cytokines, which eventually damages the lung infrastructure. The resulting cytokine storm circulates to other organs, leading to multi-organ damage. In addition, non-neutralizing antibodies produced by B cells may enhance SARS-CoV-2 infection through antibody-dependent enhancement (ADE), further exacerbating organ damage. Alternatively, in a healthy immune response (right side), the initial inflammation attracts virus-specific T cells to the site of infection, where they can eliminate the infected cells before the virus spreads. Neutralizing antibodies in these individuals can block viral infection, and alveolar macrophages recognize neutralized viruses and apoptotic cells and clear them by phagocytosis. Altogether, these processes lead to clearance of the virus and minimal lung damage, resulting in recovery. G-CSF, granulocyte colony-stimulating factor; TNF, tumour necrosis factor.

Host cell infection and its prevention

The first step in infection is virus binding to a host cell through its target receptor. Earlier work on SARS-CoV demonstrated that this virus principally targets airway epithelial cells, alveolar epithelial cells, vascular endothelial cells and macrophages in the lung, all of which express the angiotensin-converting enzyme 2 (ACE2) host target receptor used by SARS-CoV24,25,26 (Fig. 2). As SARS-CoV-2 uses the same entry receptor, these cell subsets are likely targeted by this virus4,27,28. SARS-CoV infection reduces ACE2 expression in lung cells. Because loss of pulmonary ACE2 function is associated with acute lung injury, virus-induced ACE2 downregulation may be important for disease pathology29,30,31,32. ACE2 has been shown to regulate the renin–angiotensin system (RAS)32. Therefore, a reduction in ACE2 function after viral infection could result in a dysfunction of the RAS, which influences blood pressure and fluid/electrolyte balance, and enhance inflammation and vascular permeability in the airways.

Fig. 2: The structure of the trimeric spike protein of SARS-CoV-2.
figure 2

The receptor-binding domain (RBD) is shown interacting with its receptor, human angiotensin-converting enzyme 2 (ACE2). SARS-CoV-2, severe acute respiratory syndrome coronavirus 2. Adapted from Protein Data Bank IDs 6VSB42 and 6VW1 (ref.150).

COVID-19 shows a difference in fatality rate between males (2.8%) and females (1.7%)33. As ACE2 is located on the X chromosome, there may be alleles that confer resistance to COVID-19, explaining the lower fatality rate in females. Alternatively, the oestrogen and testosterone sex hormones have different immunoregulatory functions, which could influence immune protection or disease severity34.

SARS-CoV-2 shares 79% genome sequence identity with SARS-CoV4. The spike (S) protein is expressed on the surface of the virus particles, giving the characteristic ‘crown’ appearance. The S protein comprises two subunits: S1 and S2. The S1 subunit consists of an amino-terminal domain and a receptor-binding domain (RBD), which in SARS-CoV spans from amino acid residue 318 to amino acid residue 510 (refs35,36,37). The RBD binds to ACE2 as its host cell target receptor, which starts the infection process4. RBD binding to ACE2 triggers endocytosis of the SARS-CoV-2 virion and exposes it to endosomal proteases38. The S2 subunit consists of a fusion peptide (FP) region and two heptad repeat regions: HR1 and HR2 (refs39,40). Within the endosome, the S1 subunit is cleaved away, exposing the fusion peptide, which inserts into the host membrane. The S2 region then folds in on itself to bring the HR1 and HR2 regions together. This leads to membrane fusion and releases the viral package into the host cytoplasm.

There is 72% similarity in the amino acid sequence of the RBDs of SARS-CoV and SARS-CoV-2, with highly similar tertiary structures. Computational modelling and biophysical measurements indicate that the SARS-CoV-2 RBD binds to ACE2 with higher affinity than that of SARS-CoV41,42. In addition, the SARS-CoV-2 S protein contains a furin-like cleavage site, similarly to MERS-CoV and human coronavirus OC43, which is not found in SARS-CoV43. These characteristics could contribute to the increased infectivity of SARS-CoV-2 relative to SARS-CoV. In addition to furin precleavage, the cellular serine protease TMPRSS2 is also required to properly process the SARS-CoV-2 spike protein and facilitate host cell entry44.

One pathway for the development of therapeutics against SARS-CoV-2 is to block the host target ACE2 receptor or TMPRSS2 (Fig. 3). Currently, there are compounds that target these molecules that have been clinically approved for other indications. For example, machine learning algorithms predicted that baricitinib, a Janus kinase (JAK) inhibitor approved for treatment of rheumatoid arthritis, could inhibit ACE2-mediated endocytosis45. Another JAK inhibitor, ruxolitinib, will be tested in clinical trials for treatment of COVID-19 later this year46. An alternative strategy is to deliver high concentrations of a soluble form of ACE2 that could potentially reduce virus entry into target host cells. This principle is being tested with APN01, a recombinant form of ACE2 developed by APEIRON that is currently in clinical trials47. Monoclonal antibodies targeting the S protein may also inhibit virus entry or fusion and are further discussed in the section entitled B cell immunity. Nafamostat mesylate48,49 and camostat mesylate44 are known inhibitors of TMPRSS2 and are currently approved in several countries/regions to treat other conditions. While there are no clinical trials specifically testing these drugs against COVID-19 at the time of writing, when camostat mesylate was tested on SARS-CoV-2 isolated from a patient, it prevented entry of the virus into lung cells44,50. If this approach is validated, rapid repurposing of these drugs will be effective and timely in the fight against COVID-19.

Fig. 3: Potential therapeutic approaches against SARS-CoV-2.
figure 3

(1) Antibodies against the spike protein (raised through vaccination or by adoptive transfer) could block severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from interacting with the angiotensin-converting enzyme 2 (ACE2) receptor on host cells. (2) Protease inhibitors against the serine protease TMPRSS2 can prevent spike protein cleavage, which is necessary for viral fusion into the host cell. Blocking either ACE2 interaction or viral fusion could prevent the virus from infecting the host cell. (3) Virus-specific memory CD8+ T cells from a previous vaccination or infection can differentiate into effector cells during rechallenge. When they identify infected cells presenting virus-specific epitopes, they degranulate and kill infected cells before they can produce mature virions. (4) In a novel treatment method that targets the cytokine storm symptoms, the blood of patients with coronavirus disease 2019 (COVID-19) can be passed through customized columns that are specially designed to trap pro-inflammatory cytokines, before the purified blood is passed back into patients.

Inflammatory immunopathogenesis

SARS-CoV-2 infection and the destruction of lung cells triggers a local immune response, recruiting macrophages and monocytes that respond to the infection, release cytokines and prime adaptive T and B cell immune responses. In most cases, this process is capable of resolving the infection. However, in some cases, a dysfunctional immune response occurs, which can cause severe lung and even systemic pathology.

Cytopathic viruses, including SARS-CoV-2 (ref.51), induce death and injury of virus-infected cells and tissues as part of the virus replicative cycle. Viral infection and replication in airway epithelial cells52 could cause high levels of virus-linked pyroptosis with associated vascular leakage, as seen in patients with SARS-CoV53. Pyroptosis is a highly inflammatory form of programmed cell death that is commonly seen with cytopathic viruses54. This is a likely trigger for the subsequent inflammatory response55. IL-1β, an important cytokine released during pyroptosis, is elevated during SARS-CoV-2 infection11. Using a variety of pattern-recognition receptors (PRRs), alveolar epithelial cells and alveolar macrophages detect the released pathogen-associated molecular patterns (PAMPs), such as viral RNA, and damage-associated molecular patterns (DAMPs), including ATP, DNA and ASC oligomers. A wave of local inflammation ensues, involving increased secretion of the pro-inflammatory cytokines and chemokines IL-6, IFNγ, MCP1 and IP-10 into the blood of afflicted patients11,22. These cytokines are indicators of a T helper 1 (TH1) cell-polarized response, which parallels observations made for SARS-CoV and MERS-CoV56. Secretion of such cytokines and chemokines attracts immune cells, notably monocytes and T lymphocytes, but not neutrophils, from the blood into the infected site57,58. Pulmonary recruitment of immune cells from the blood and the infiltration of lymphocytes into the airways may explain the lymphopenia and increased neutrophil–lymphocyte ratio seen in around 80% of patients with SARS-CoV-2 infection6,59.

In most individuals, recruited cells clear the infection in the lung, the immune response recedes and patients recover. However, in some patients, a dysfunctional immune response occurs, which triggers a cytokine storm that mediates widespread lung inflammation. It was observed that patients with severe COVID-19, requiring intensive care in hospitals, exhibited higher blood plasma levels of IL-2, IL-7, IL-10, granulocyte colony-stimulating factor (G-CSF), IP-10, MCP1, macrophage inflammatory protein 1α (MIP1α) and tumour necrosis factor (TNF)11. IL-6 levels in these patients continue to increase over time and are relatively more elevated in non-survivors than survivors60. Notably, there exists a highly inflammatory monocyte-derived FCN1+ macrophage population in the bronchoalveolar lavage fluid of patients with severe but not mild COVID-19 (ref.61). Also, patients with severe disease show a significantly higher percentage of CD14+CD16+ inflammatory monocytes in peripheral blood than patients with mild disease62. These cells secrete inflammatory cytokines that contribute to the cytokine storm, including MCP1, IP-10 and MIP1α (Fig. 1).

The mechanisms by which SARS-CoV-2 subverts the body’s innate antiviral cytokine responses are yet to be studied, but research on SARS-CoV shows that multiple viral structural and non-structural proteins antagonize interferon responses. Antagonism occurs at various stages of the interferon signalling pathway, including by preventing PRR recognition of viral RNA63,64,65, by preventing PRR signalling through TBK1/inhibitor of nuclear factor-κB kinase subunit-ε (IKKε), TRAF3 and IRF3 (refs63,66), by preventing downstream interferon signalling through STAT1 (ref.67) and by promoting host mRNA degradation and inhibiting host protein translation68. It is very likely that at least some of these pathways are conserved in SARS-CoV-2. Antagonism of the interferon response aids viral replication, resulting in increased release of pyroptosis products that can further induce aberrant inflammatory responses.

Unrestrained inflammatory cell infiltration can itself mediate damage in the lung through excessive secretion of proteases and reactive oxygen species, in addition to the direct damage resulting from the virus. Together, these result in diffuse alveolar damage, including desquamation of alveolar cells, hyaline membrane formation and pulmonary oedema57,58. This limits the efficiency of gas exchange in the lung, causing difficulty in breathing and low blood oxygen levels. The lung also becomes more vulnerable to secondary infections.

In addition to local damage, cytokine storm also has ripple effects across the body. Elevated levels of cytokines such as TNF can cause septic shock and multi-organ failure. These may result in myocardial damage and circulatory failure observed in some patients69. Older people (those aged over 60 years) and people with co-morbidities are more likely to develop such a dysfunctional immune response that causes pathology and also fails to successfully eradicate the pathogen. The exact reasons for this are unclear, although one reason may be an ageing lung microenvironment causing altered dendritic cell maturation and migration to the lymphoid organs70, and thereby defective T cell activation. In contrast, children tend not to develop severe disease despite being capable of experiencing high viral titres71. Across all age groups younger than 18 years, more than 50% of children experienced mild symptoms or were asymptomatic, with less than 6% of children developing severe symptoms72. Thus, while the aforementioned studies represent important inroads, a full picture of the critical host immune factors that underlie the development of severer inflammatory responses in some patients remains poorly defined.

It remains controversial whether virus persistence is necessary to drive the ongoing damage. The peak of viral titres in respiratory tract samples might occur even before symptom onset of pneumonia in SARS-CoV and SARS-CoV-2 infections17,19. However, a large retrospective cohort study showed that viral RNA was detectable in non-survivors up until the point of death, suggesting a correlation between virus persistence and poor disease outcome60. As viral RNA may linger even after active infection, and is not representative of the infectivity of the virus, whether the poor disease outcome is directly due to large amounts of infectious particles is speculative at this moment. Furthermore, earlier studies of SARS-CoV found that the virus may infect other targets besides lung cells. Notably, virus was found in T lymphocytes73, macrophages74,75,76 and monocyte-derived dendritic cells77. Direct virus killing of lymphocytes could contribute to the observed lymphopenia in patients73. Viral infection in immune cells such as monocytes and macrophages can result in aberrant cytokine production, even if viral infection is not productive74,75,76,77. The degree to which SARS-CoV-2 targets these cells remains poorly defined. Understanding the precise drivers of immune dysfunction is crucial to guide the application of appropriate immunomodulatory treatments.

Several immunosuppressive therapies aimed at limiting immunomediated damage in COVID-19 are at various phases of development and are listed in Table 1. Currently, trials of corticosteroids for treatment of COVID-19 are under way78, although this class of treatment was not recommended during the 2003 SARS epidemic79,80. A clinical trial of the IL-6 antagonist tocilizumab is also under way to test its efficacy81, and sarilumab is also being explored82. Other clinical trials are also testing the effects of targeting granulocyte–macrophage colony-stimulating factor (GM-CSF), including the use of gimsilumab83, lenzilumab84 and namilumab85. Another novel adjunctive therapy is cytosorb86, which acts by absorbing a broad spectrum of cytokines, DAMPs and PAMPs in order to reduce their circulating levels and ameliorate immunopathology. Thalidomide, an agent with immunomodulatory properties, has also been successfully administered to a single patient with COVID-19 (ref.87). As a result, two clinical trials have now been initiated to test its potential to reduce lung injury88,89. TNF antagonism was suggested but not tested in the context of SARS-CoV infection, and it has not yet been tested in patients with COVID-19 (ref.90). A small open-label, non-randomized study suggested that a combination of hydroxychloroquine (a known antimalarial agent) and azithromycin (a common antibiotic) may be beneficial for treating patients with severe COVID-19 (ref.91). Although hydroxychloroquine’s effect on direct inhibition of the virus92 and its anti-inflammatory and immunomodulatory activities are known93, whether these mechanisms play a role against COVID-19 remains to be determined94.

Table 1 Summary of approved interventional clinical trials for COVID-19 as of March 2020

T cell immunity

Both T and B cell responses against SARS-CoV-2 are detected in the blood around 1 week after the onset of COVID-19 symptoms. CD8+ T cells are important for directly attacking and killing virus-infected cells, whereas CD4+ T cells are crucial to prime both CD8+ T cells and B cells. CD4+ T cells are also responsible for cytokine production to drive immune cell recruitment. The first autopsy of a patient with COVID-19 revealed an accumulation of mononuclear cells (likely monocytes and T cells) in the lungs, coupled with low levels of hyperactive T cells in the peripheral blood57. Together with reports of lymphopenia and reduced peripheral T cell levels in patients 6,95,96,97, these findings suggest that T cells are attracted away from the blood and into the infected site to control the viral infection. In patients with COVID-19, increased T cell exhaustion and reduced functional diversity predicted severe disease98. Despite the impaired response, patients who recovered from SARS-CoV infection developed coronavirus-specific memory T cells, which were found up to 2 years after recovery99,100.

SARS-CoV-specific CD4+ T cells express IFNγ, TNF and IL-2, which suggests that patients with SARS-CoV infection exhibit a TH1 cell response and mainly use cellular immunity to control the infection101,102. Although this pro-inflammatory profile may be an aggravating factor for immunopathogenesis, CD4+ T cells have been hypothesized to control SARS, as depletion of these cells in mice resulted in slower clearance of the virus from the host and severer lung inflammation103. With the use of a mouse-adapted strain of SARS-CoV, immunization with dendritic cells bearing SARS-CoV peptides resulted in higher numbers of virus-specific CD4+ and CD8+ T cells that accumulated in the lungs and increased survival104,105. Also, transfer of SARS-CoV-specific CD4+ and CD8+ T cells into immunodeficient mice resulted in better protection against a mouse-adapted strain of SARS-CoV105.

Despite evidence for an important role of T cells in controlling infection, several vaccine formulations against SARS-CoV previously tested in animal models showed signs of immunopathology associated with TH2 cell-mediated eosinophil infiltration106,107. In particular, aged mice that were vaccinated seemed to display increased immunopathology rather than protection108. Further study of the nature of protective versus detrimental T cell responses is critically needed to determine the optimal T cell engagement strategies for vaccines109.

Coronavirus-specific T cells are clearly important in eliminating the virus and controlling disease development and should be considered in vaccine strategies. However, whether T cell responses alone are capable of preventing infection in human settings remains to be investigated. This knowledge will be important for vaccine development.

B cell immunity

B cell responses in patients with COVID-19 occur concomitantly with T follicular helper cell responses, from around 1 week after symptom onset110. In patients with SARS-CoV infection, B cell responses typically arise first against the nucleocapsid (N) protein. Within 4–8 days after symptom onset, antibody responses to S protein are found111,112. Neutralizing antibody responses, likely to the S protein, begin to develop by week 2, and most patients develop neutralizing antibodies by week 3 (refs113,114). Given that viral titres peak earlier for SARS-CoV-2 than for SARS-CoV16,17,18,19, antibody responses may also arise earlier. It seems that a subset of patients may not develop long-lasting antibodies to SARS-CoV-2 (ref.115). It remains unknown whether these patients are susceptible to reinfection, of which there are sporadic reports116,117. Antibodies are likely to be effective against SARS-CoV-2: convalescent serum samples have been applied with apparently good clinical results in COVID-19 (ref.118) and were also previously used successfully in the treatment of SARS119,120,121.

While mechanistic correlates of protection have not yet been identified in humans, neutralization of the virus is presumed to be an important mechanism of action for antibodies, although the specific titre and specificity of the antibody repertoire required (for protection) remain undefined. In SARS-CoV, the primary target of neutralizing antibodies is the RBD122, comprising a 193 amino acid region (amino acids 318–510) in the S protein, which can independently bind to the host target ACE2 receptor35,36,37. Although a few previously identified monoclonal antibodies to SARS-CoV also bind to or neutralize SARS-CoV-2 (ref.123), the majority do not124. This could be due to significant differences in the RBDs of SARS-CoV-2 and SARS-CoV (Fig. 4). In particular, of the 33 amino acids in the region (amino acids 460–492) in the SARS-CoV S protein that contains the critical residues that contact ACE2 (ref.125), less than half (15/33) are conserved in SARS-CoV-2. Nevertheless, mouse antiserum raised against SARS-CoV protein can cross-neutralize SARS-CoV-2 pseudovirus, indicating overlapping neutralizing epitopes between the two viruses28,126.

Fig. 4: Sequence alignment and structural comparison of SARS-CoV and SARS-CoV-2 spike proteins.
figure 4

a | Sequence alignment of severe acute respiratory syndrome coronavirus (SARS-CoV) spike protein and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein, with conserved amino acid residues shown in black and non-conserved residues shown in colours. b | The 3D structure of SARS-CoV-2 (Protein Data Bank ID 6VSB42, peach ribbon) is superimposed on the SARS-CoV receptor-binding motif (RBM) complex with the neutralizing antibody (nAb; red ribbon) interfacing with the RBM (Protein Data Bank 2DD8 (ref.151), purple ribbon). Peach and purple spheres denote the RBMs of SARS-CoV-2 and SARS-CoV, respectively. Magenta spheres denote non-synonymous alterations in the SARS-CoV-2 spike protein that have been reported135.

In China, hospitals have initiated the use of convalescent plasma as a source of therapeutic polyclonal antibodies for treatment of COVID-19, and early data suggest a positive impact on respiratory viral load and mortality127,128. Efforts are under way to develop therapeutic monoclonal antibodies to SARS-CoV-2, using approaches including phage library display, traditional mouse immunization and hybridoma isolation, and cloning of B cell sequences from convalescent human patients129,130,131,132. SARS-CoV does not appear to have strong mechanisms to escape or prevent antibody neutralization, such as glycan shielding of the receptor-binding site against antibody binding133. This is further corroborated by the fact that patients with SARS-CoV infection were generally capable of developing neutralizing antibodies. A recombinant S protein fragment that included the RBD of SARS-CoV showed the highest immunogenicity relative to other recombinant S protein fragments tested, suggesting that the immune system is capable of targeting neutralizing epitopes effectively134. Thus, if SARS-CoV-2 behaves like SARS-CoV in this respect, it is likely that these efforts will be successful in developing neutralizing monoclonal antibodies.

It is possible that alterations in the S protein will render SARS-CoV-2 resistant to some monoclonal antibodies, especially as it spreads and mutates. As of now, the entire RBD remains conserved, and there are only four known rare non-synonymous alterations in the S protein: V483A, L455I, F456V and G476S135. The V483A alteration maps to a similar natural alteration found in MERS-CoV, I529T, where it reduced viral protein binding to the host receptor target and also increased resistance to antibody neutralization from serum samples from patients with MERS136. F456V and G476S alterations also map to similar alteration positions in SARS-CoV (L443R and D463G), which were found in a panel of neutralization escape mutants137.

However, the selection of therapeutic antibody candidates should include careful consideration of potential unwanted side effects. For example, pre-existing antibodies to other coronaviruses may exacerbate SARS-CoV infections through antibody-dependent enhancement138,139,140. Also, previous studies in animal models showed that in SARS-CoV infection, neutralizing antibodies to S protein can potentially augment severe lung injury by exacerbating inflammatory responses141. In addition, a correlation has been observed where development of ARDS coincides with antiviral IgG seroconversion in 80% of patients19. Patients who developed neutralizing antibodies to S protein earlier in infection had a higher rate of disease; it took an average of only 14.7 days for patients who died of infection to reach their peak levels of neutralizing antibody activity, as opposed to 20 days for patients who went on to recover142. Similarly, for MERS, patients with severer disease appear to have higher antibody titres than those with mild disease143,144, although one study argues that it is a delay in the development of antibody responses that is associated with disease145. The binding of antibody–virus immune complexes to activating Fc receptors on alveolar macrophages could induce the expression of pro-inflammatory factors, including IL-8 and MCP1, which add to the immunostimulatory milieu146. Such complexes may also activate the complement system and lead to further unwanted inflammation141. As a result, it is important to consider engineering therapeutic antibodies with little or no pro-inflammatory activity but that retain their virus-neutralizing capacity147. For instance, alterations could be made to the Fc region and/or its glycosylation to change its binding affinity for activating Fc receptors146,148.

Conclusion

This Review has presented the various mechanisms of SARS-CoV-2 infection and COVID-19 immunopathogenesis. Controlling the inflammatory response may be as important as targeting the virus. Therapies inhibiting viral infection and regulation of dysfunctional immune responses may synergize to block pathologies at multiple steps. At the same time, the association between immune dysfunction and outcome of disease severity in patients with COVID-19 should serve as a note of caution in vaccine development and evaluation. Further studies of the host immune response to SARS-CoV-2 are necessary, including a detailed investigation of the determinants of healthy versus dysfunctional outcomes. These will also help identify biomarkers to define immune correlates of protection and disease severity for effective triage of patients.