The dynamics of the antibody response against SARS-CoV-2 are currently under investigation, as antibodies may be considered potent diagnostic tools to complement RT-PCR based diagnosis.

SARS-CoV-triggered humoral S- and N-specific IgM response reached a peak within 4 weeks and was no more detectable 3 months post symptoms onset (PSO); the switch to IgG often occurred around day 14, and IgGs were detectable up to 36 months (22–24). A summary of the reports analyzing the dynamics of the antibody response during SARS-CoV-2 infection is reported in Table 1.

Xiao et al. showed that all 34 SARS-CoV-2 laboratory confirmed analyzed cases were positive for IgM and IgG at week 3-PSO. IgM levels decreased at week 4; 2 patients were negative at week 5, and additional 2 patients at the end of the observation (week 7). Therefore, in the majority of those patients, the acute phase of infection persisted for more than 1 month. Concomitantly to IgM decrease, IgG levels raised gradually from week 3 to week 7, indicating the activation of the humoral immune response against the virus (25). The authors speculated that the humoral response triggered by SARS-CoV-2 may be similar to that harbored by SARS-CoV.

An additional report on the dynamics of the antibody profile in COVID-19 patients showed that seroconversion appeared sequentially for total antibodies, IgM and IgG, with a median time of 11, 12, and 14 days. Total antibodies were detected by double recombinant antigens sandwich immunoassay (the RBD epitope of the S1 protein and the HRP-conjugated antigen), the IgM μ-chain capture method was used for IgM detection, and indirect ELISA kit based on recombinant NP antigen was used to detect IgG. The seroconversion rate was 93.1, 82.7, 64.7% for total antibodies, IgM and IgG, respectively, and no difference was observed between critical and non-critical patients. Importantly, the sensitivity of antibody detection was lower than the RNA test within 7 days from the onset of the disease (38.3% vs. 66.7%), but raised gradually since day 8 to day 39 PSO, overtaking that of RNA test. More importantly, detectable levels of total antibodies in the sera were found in those patients with undetectable levels of RNA in their respiratory tract samples. This evidence highlighted the extreme importance to combine molecular and serological tests for the accurate diagnosis of COVID-19 patients at different stages of the disease (26). In accordance with this study, Jin et al. reported that the specificity of serum IgM and IgG to detect SARS-CoV-2 infected patients was 90% compared to that of the molecular test. They also registered undetectable levels of any specific antibody up to day 8 PSO in 3 patients (27).

Guo et al. profiled the early antibody response to NP protein in two cohorts of SARS-CoV-2-infected patients. The 90.4% and the 93.3% of 208 patients harbored plasma IgM and IgA, respectively, and the 77.9% of plasma samples were positive for IgG. The median time for IgM and IgA detection was at day 5 PSO (IQR-3-6) and day 14 PSO (IQR 10–18) for IgG (28). The rapid and unexpected IgA seroconversion the authors observed might be an effect of the cytokines storm promoting the germline transcription of both the heavy chain constant α and μ genes. Alternatively, it has been found that T-cell-independent antibody responses stimulate a specialized B cell subset to produce both IgM and IgA during the infection of some pathogens (33). Although the T-cell-independent antibody response against viruses is controversial, some viruses can act in vivo as T-cell-independent antigens, eliciting protective, isotype-switched antibodies in the absence of conventional T cell help. Moreover, inactivated virus or virus-like particles can elicit IgM response, but factors induced during active virus infection seem necessary to induce the isotype switch leading to IgG or IgA responses (29).

Liu et al. analyzed a cohort of 214 COVID-19 patients. The 68.2% and the 70.1% of the patients were positive for rN-specific IgM and IgG, respectively; the 77.1% and the 74.3% were positive for rS-specific IgM and IgG, respectively. This data indicated that the detection of rS-specific-IgM was more sensitive compared to that of rN-spcific IgM, probably due to the higher immunogenicity of the S protein compared to that of the N protein. A bioinformatics analysis, indeed, predicted a higher number of B cells epitopes in the S protein than in the NP protein of SARS-CoV-2. In addition, the positive rates of IgM and IgG were low at early stages of the disease (0-10 DPO); conversely, IgM and/or IgG specific for rN and rS reached a peak at 11–15 DPO (30).

The sensitivity of the tests and the epitope on which the test is based on are relevant factors to take into account for the efficient detection of specific SARS-CoV-2 antibodies and timing the humoral response. Therefore, several tests are rapidly developing in many laboratories. Li et al. developed a point-care lateral flow immunoassay (LFIA) test based on the RBD antigen of the SARS-CoV-2 S1 protein that allowed the concomitant detection of IgM and IgG in human blood within 15 min with higher sensitivity than the individual IgG and IgM tests. However, the limit of detection of the test was not determined (34). Importantly, Amanat and collaborators developed sensitive and specific ELISA assays based on the recombinant full-length S protein and RBD epitope allowing the screening and detection of seroconversion upon SARS-CoV-2 infection 3 days PSO (35). Of note, no cross-reactivity from other human coronaviruses was detected, in accordance with another study highlighting that S1 is a specific antigen for SARS-CoV-2 diagnosis as cross-reactive antibodies against the S protein of MERS-CoV were not detected in a COVID-19 patient (31). In addition, strong IgA and IgM responses were uncovered and the IgG3 response was stronger than IgG1 (35).

The sensitivity of the test may pose challenges for the early detection of IgM. Indeed, some patients were more positive for IgG than IgM at the moment of hospitalization and 5 days later; moreover, they had an earlier IgG than IgM seroconversion (32).

SARS-CoV-2 specific antibodies were also detected in 6 sera of infants born from COVID-19 mothers. Five out of six infants and their mothers had high levels of IgG and two of them had high levels of IgM as well. Three out of six infants who had high levels of IgG had normal levels of IgM. However, two of their mothers showed high levels of IgM. How the newborns developed IgM needs further investigations. Indeed, due to their large size, IgM are not usually transferred through the placenta, unless it is affected by some pathology that compromises its structure. The newborn might get contact with the virus if the latter crosses the placenta; however, no virus was detected from RT-PCR analysis (13).

Some studies are investigating the correlation between antigen-specific antibodies and clinical characteristics of COVID-19 patients. Interestingly, patients with comorbidities had lower anti-RBD IgG, but not anti-NP IgM or IgG, than those without comorbidities, although the difference was not significant. No association with age was observed (36).

Neutralizing antibodies (NAbs) play critical roles in blocking viral infections, thus contributing to viral clearance during acute infection or controlling disease progression during chronic phase. These antibodies are, therefore, useful tools for the protection from viral infection and for the development of effective treatments.

NAbs in the plasma of recovered patients were successfully employed in the passive antibody therapy for SARS-CoV virus- (37), influenza virus- (38) and Ebola virus-infected subjects (39).

The S1 subunit of the S protein, particularly the 193 amino acid length RBD domain (N318-V510), is the main target for antibody-mediated neutralization, probably because it plays major roles during the early stages of infection (40). Studies from SARS-CoV and MERS-CoV demonstrated that many epitopes of the S protein, namely S1-NTD, RBD and S2 are highly immunogenic and can be targets to develop potent NAbs. Several human monoclonal antibodies targeting S1 were developed against SARS-CoV, demonstrating efficient blocking of the binding to the ACE2 receptor in both in vitro and animal models. They recognize different epitopes within the S1 subunit, and display different potency of neutralization, alone or in combination (39, 41–43).

Whether SARS-CoV Nabs bind or not SARS-CoV-2 is still controversial. Hoffman et al. demonstrated that the serum from a convalescent SARS-CoV patient neutralized SARS-CoV-2 entry in vitro (44). Some studies did not observed any binding (9, 45); however, the SARS-CoV CR3022 NAb bound to SARS-CoV-2 RBD, but it recognized an epitope that did not overlap with the ACE2 binding site within the RBD domain (45). This evidence may suggest a difference in the antigenicity of SARS-CoV and SARS-CoV-2, relevant for the cross-reactivity of NAbs and the design of specific therapeutics against SARS-CoV-2.

A cohort of 175 COVID-19 patients with mild symptoms developed high titers of SARS-CoV-2 specific NAbs targeting the S1, the RBD and the S2 domains of the S protein, with a peak at 10-15 DPO. Interestingly, these NAbs had cross-reactivity but not neutralizing activity against SARS-CoV (46). An unexplored relevant aspect emerged in this study. The titers of SARS-CoV-2 specific NAbs differed across patients and correlated with their age. Elderly and middle age patients displayed higher titers of NAbs than younger patients (46). Of note, Nabs titers in young patients varied, and 10 COVID-19 recovered patients showed Nabs titers below the limit of detection of the assay, although the molecular test confirmed they were SARS-CoV-2-infected. This correlation, recently confirmed by Wang et al. (47), was reported for SARS-CoV and MERS viruses as well; moreover, the strong humoral response observed in aged macaques infected with SARS-CoV related with a severe disease status (48–50). Therefore, age and disease severity may be considered covariates in relation to development of neutralizing antibodies. A rough estimate of the development of neutralizing antibodies after SARS-CoV-2 infection is reported in Figure 1.

No studies on the duration of SARS-CoV-2 specific NAbs have been reported so far. In a cohort study of 56 SARS-CoV convalescent patients, specific IgG and neutralizing antibodies were highly correlated, and persisted for 24 months, despite the decline of their titers (51). Another study showed the 74.2% and the 83.9% of the patients were positive for IgG and neutralizing antibodies 36 months PSO (24). An observational cohort study including 16 COVID-19 patients whose serum samples were collected 14 days PSO showed that the majority of patients harbored neutralizing IgM and IgG against both NP and RBD (36). NP is highly immunogenic, although smaller than S, lacks of glycosylation sites, and induces antibodies earlier than S during the infection, thus contributing to neutralization; therefore, anti-NP-specific antibodies might play a key role during the early stages of acute infection (52).

Discovering the epitopes enabling to elicit humoral responses against SARS-CoV-2 is relevant for the development of specific monoclonal antibodies for therapy or prophylaxis.

A bioinformatics analysis through the Immune Epitope Database and Analysis Resource (IEDB) revealed that the S protein had the highest number of predicted B cell epitopes, particularly the regions 491–505, 558–562, 703–704, 793–794, 810, 914, and 1,140–1,146; however, besides the S protein, the M protein and NP phospoprotein contained B cell immunodominat regions as well (53). Interestingly, the sequences of the B cell immunodominat regions of SARS-CoV were conserved in SARS-CoV-2. Of note, convalescent SARS-CoV patients harbored NAbs directed against the epitopes of five of these regions (54, 55). Moreover, two regions (1–25 and 131–152) within the M protein triggered high IgM and IgG responses (56, 57); the 156–175 region within the NP protein was reactive against sera from SARS patients and showed immunogenicity in a broad spectrum of species, including humans (58).

Plasma of convalescent COVID-19 patients were used to treat 10 severe cases of SARS-CoV-2 infections. The dose (200 mL) of plasma was well tolerated and viremia decreased in 7 days with the concomitant improvement of the clinical symptoms within 3 days (59). This data strongly suggests that the deeper characterization of plasma from recovered patients might give important information for the development of effective antibody-based therapies to treat COVID-19 patients.

An opening question rely on the huge difference in the severity of COVID-19 ranging from asymptomatic, low, mild and severe cases.

Tetro speculated that subjects who experienced the most severe forms of the disease might have been primed by one or more coronavirus exposure leading to the effects of Antibody Dependent Enhancement (ADE) of the SARS-CoV-2 infection (60). The antibodies elicited by a previous contact with a virus might not completely neutralize a second infection and, conversely, form complexes with the second virus or virus-activated complement components that interact with the Fc or complement receptors on susceptible cells, thereby facilitating viral entry (61, 62). In addition, ADE modulates the immune response, triggering inflammation, cytokine storm ad lymphopenia (60), responsible for the poor outcome of the disease.

ADE has been extensively investigated in dengue virus (63–65), and observed in HIV (66) and Ebola (62) infections as well. With respect to coronaviruses, antibodies induced by the SARS-CoV S protein enhanced the viral entry into the cells expressing the Fc receptor (67–69). Liu et al. showed that during acute SARS-CoV infection, anti–S-IgG altered macrophages functions by abrogating their wound-healing response, partially through FcγRs. Concomitantly, anti–S-IgG decreased TGF-β production, while inducing pro-inflammatory cytokine IL-8 and MCP1 production and inflammatory macrophage accumulation in the lung, finally leading to acute lung injury (70). Moreover, some non-neutralizing Abs targeting the non-RBD regions in the S protein may cause an antibody-dependent enhancement of SARS-CoV-2 infection, with consequent harmful immune response (71). Different from this study, Wan et al. showed that a MERS-CoV-specific neutralizing Mab targeting RBD mediated the entry of a MERS pseudovirus into Fc-expressing cells (72).

Importantly, some studies did not detect any cross-reactivity from other human coronaviruses (31, 35). Based on this observation, Amanat and collaborators excluded that the ADE from human coronaviruses might be the cause of the high pathogenicity of SARS-CoV-2 (35).

Further investigations are needed to understand the mechanism of ADE in facilitating viral infections and its putative role in COVID-19 onset and progression in order to address new viral vaccine design and antibody-based therapeutics.