Antibody Response and Therapy in COVID-19 Patients: Significance in Vaccine Development

The newly emerged severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has infected millions of people and caused tremendous morbidity and mortality worldwide. Effective treatment for coronavirus disease 2019 (COVID-19) due to SARS-CoV-2 infection is lacking and different therapeutic strategies are under testing. Host humoral and cellular immunity to SARSCoV-2 infection is a critical determinant for patients’ outcome. SARS-CoV-2 infection results in seroconversion and production of anti-SARS-CoV-2 antibodies. The antibodies may suppress viral replication through neutralization but also might also participate in COVID-19 pathogenesis through a process termed antibody-dependent enhancement. Rapid progress has been made in the research of antibody response and therapy in COVID-19 patients including characterization of the clinical features of antibody responses in different populations infected by SARS-CoV-2, treatment of COVID-19 patients with convalescent plasma and intravenous immunoglobin products, isolation and characterization of a large panel of monoclonal neutralizing antibodies, as well as preliminary clinical results from several COVID-19 vaccine candidates. In this review, we summarize the recent progress and discuss the implications of these findings in vaccine development.


Introduction
The newly emerged severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has infected more than 18.7 million people to cause coronavirus disease 2019  and led to ~705,000 deaths worldwide as of August 5, 2020(Johns Hopkins University, 2020). SARS-CoV-2 belongs to the  subgroup of the Coronaviridae family consisting of four genera ,  and  Zhou et al., 2020b). The  subgroup of the Coronaviridae family also includes two other highly pathogenic viruses, SARS-CoV and middle eastern respiratory syndrome (MERS-CoV) that have caused outbreaks in 2002-2003 and 2013-present, respectively . The SARS-CoV-2 RNA genome encodes four conserved structural proteins, spike (S), envelope (E), membrane (M), and nucleocapsid (N) as well as at least six accessory proteins (3a, 6, 7a, 7b, 8 and 10) (Kim et al., 2020). SARS-CoV-2 infects host cells through its S protein, which consists of the S1 subunit, S2 subunit, transmembrane, and cytoplasmic domains. The S1 subunit is further divided into N-terminal domain (NTD), receptor-binding domain (RBD), subdomain 1 (SD1), and subdomain 2(SD2). The RBD in the S1 subunit of SARS-CoV-2 binds to cellular receptor angiotensin-converting enzyme 2 (ACE2) Sungnak et al., 2020;Wang et al., 2020c;Yan et al., 2020) or CD147  and mediates virus entry. Upon RNA genome deposit into the cytoplasm, SARS-CoV-2 goes through cycles of intracellular replication and is released by exocytosis to infect other host cells (Hartenian et al., 2020).
The high morbidity and mortality of COVID-19 patients poses a tremendous health challenge and calls for urgent development of effective prevention and treatment strategies. The major pathology caused by SARS-CoV-2 infection is pneumonia. Severe and critically ill COVID-19 patients also have other organ damage (Robba et al., 2020). Moribund COVID-19 patients go through three distinct stages: initiation, amplification, and consummation . The initiation stage is characterized by rapid viral replication and early induction of predominant chemokines. In the absence of effective viral suppression by host humoral and cellular immunity, the patients enter the amplification stage in which both rampant production of more inflammatory mediators and massive recruitment of inflammatory cells to amplify immunopathological processes occur. Subsequently, COVID-19 patients succumb due to continuous increases of inflammatory mediators and widespread organ damage in the consummation stage . Antibody-mediated humoral immune responses to SARS-CoV-2 infection play essential roles in COVID-19 disease development. Upon infection, host B lymphocytes recognizing SARS-CoV-2 specific antigens are activated, and antibody production is initiated. Seroconversion of IgM, IgA, and IgG and the appearance of neutralizing antibodies to SARS-CoV-2 suppress viral replication. In the 6 months since the COVID-19 pandemic began, rapid progress has been made in the characterization of antibody responses including the dynamics and properties of anti-SARS-CoV-2 antibodies in infected populations. Furthermore, results from clinical studies using passive antibody treatment of COVID-19 patients with convalescent plasma and intravenous immunoglobulins (IVIG) have been reported. A large panel of neutralizing monoclonal antibodies have been isolated. Importantly, preliminary data from several vaccine trials has been reported. In this review, we summarize recent results on antibody response and therapy in COVID-19 patients. We further discuss the clinical significance of these results in the context of preliminary vaccine trial data.

Humoral responses to SARS-CoV-2 infection in COVID-19 patients.
The development of antibodies to SARS-CoV-2 has been assessed in COVID-19 patients at different disease stages and severities. Specific IgA, IgM, IgG, or total antibodies to SARS-CoV-2 antigens, including full length S protein, RBD, E, and N proteins, were assessed for binding activities in peripheral blood of a total of 4261 patients (Table 1) Li et al., 2020a;Long et al., 2020a;Lou et al., 2020;Lynch et al., 2020;Prevost et al., 2020;Wajnberg et al., 2020;Wang et al., 2020d;Zhang et al., 2020b;Zhao et al., 2020). Additional studies were also analyzed and presented in a systemic review (Deeks et al., 2020). Several major features appeared from these studies.
First, most of the COVID-19 patients developed one or more antigen-specific antibodies to SARS-CoV-2 during the course of the disease. The rates of antibody positivity ranged from 53-100% and were clearly time-dependent (Table 1). Pooled results showed that specific antibodies were detected in <30.1% of patients during the first week after disease onset (Deeks et al., 2020). The positive rates rapidly rose to 72.2% and 91.4% during the second and third weeks after the onset of disease, respectively, and reached 96.0% by the end of fifth week (Deeks et al., 2020). Thus, regardless of patient characteristics (age, sex, co-morbidities, and treatments), SARS-CoV-2 is a strongly immunogenic virus that can induce antigen-specific antibody production in most infected patients.
Second, the patterns of antibody seroconversion varied. Three types of seroconversion of antigen-specific IgM and IgG were observed: IgM seroconverted earlier than IgG, IgG seroconverted earlier than IgM, and synchronous IgM and IgG seroconversion (Long et al., 2020a). The pattern of seroconversion of IgG prior to IgM may be caused by some level of cross-reactivity between SARS-CoV-2 and other viruses previously infecting the hosts and suggest activation of memory B cell populations. This pattern of response may be clinically significant in COVID-19 pathogenesis (see discussion below). Nevertheless, the pattern of IgM seroconversion before IgG was most commonly observed in COVID-19 patients Lou et al., 2020;Lynch et al., 2020;Zhang et al., 2020b;Zhao et al., 2020). The median seroconversion time for IgM and IgG were 5-11 and 10-15 days after disease onset, respectively Lou et al., 2020;Lynch et al., 2020;Zhang et al., 2020b;Zhao et al., 2020). The levels of IgM reached peak earlier than IgG and declined after one month. In contrast, IgG elevated later but lasted longer . However, specific IgM antibody response was delayed in COVID-19 patients with severe progression . SARS-CoV-2-specific blood IgA had a similar median seroconversion time to that of IgM at 5 days after symptom onset . Secretory IgA was also measured in milk samples from SARS-CoV-2 infected donors and 80% of the samples exhibited both blood and secretory IgA reactivity, indicating efficient induction of secretory IgA (Fox et al., 2020).
Third, the levels of anti-SARS-CoV-2 antibodies are strongly correlated with COVID-19 disease severity ( Figure 1A). In a cohort of 173 patients, the antibody titers in critically ill COVID-19 patients were significantly higher than those in non-critically ill patients and were independent factors for disease severity classification . Critically ill COVID-19 patients had much higher IgM and/or IgG titers than mild cases Lynch et al., 2020). Higher neutralizing antibody titers were also associated with worse clinical classifications (Wang et al., 2020d). Furthermore, anti-SARS-CoV-2 (S, RBD, and N) antibodies in severe/critically ill COVID-19 patients developed one week later than those in mild/moderate patients but were 1.5-folder higher   (Figure 1A). These results raise the question of whether the increased levels of anti-SARS-CoV-2 antibodies play a pathological role in COVID-19 pathogenesis. Previous studies suggest that anti-SARS-CoV antibodies may mediate an antibody-dependent enhancement (ADE) effect (Peiris et al., 2003;Liu et al., 2019;Arvin et al., 2020).
Although SARS-CoV-2 induces specific antibody responses in the majority of infected humans, specific populations demonstrated defective antibody production. For example, renal transplant patients under immunosuppressive treatment failed to mount anti-SARS-CoV-2 IgM and IgG for a prolonged time Xia et al., 2020). Patients with chronic lymphocytic leukemia had delayed development of anti-SARS-CoV-2 total antibodies (Favresse et al., 2020). Interestingly, the patients developed antibodies after suspension of immunosuppressive treatment . These clinical manifestations indicate that immunosuppressed populations have defective humoral responses to SARS-CoV-2 infection.

Humoral responses to SARS-CoV-2 infection in asymptomatic populations
Anti-SARS-CoV-2 specific antibodies were also assessed in viral RNA confirmed asymptomatic patients, healthcare workers, and the general population. By studying seroconversion in these populations, it was shown that asymptomatically infected human subjects develop anti-SARS-CoV-2 antibodies with different kinetics compared to symptomatic COVID-19 patients. Furthermore, the seropositive prevalence in general populations is directly associated with local pandemic levels of SARS-CoV-2 infection.
Out of 200 front-line maternity health workers in London hospitals, 5.0% were found to be seropositive without any symptoms(Bampoe et al., 2020), while there was a 2.96% seropositivity in 1,282 healthcare workers in Minnesota, United States (Thomas et al., 2020). The prevalence of SARS-CoV-2 antibodies in populations were clearly related to local epidemic levels. Seropositivity was only 0.99% among 904 healthy blood donors in the low-incidence COVID-19 area of South Eastern Italy (Fiore et al., 2020). In contrast, a striking 23% seropositive rate was observed in 390 blood donors in the high COVID-19 Lodi Red Zone in Italy . In testing 11,933 residual clinical specimens collected from March 23 to May 3 at six sites in the United States, it was found that the seropositive rates were 1.1% in the Puget Sound, 1.9% in south Florida, 4.9% in Connecticut, and 6.9% in New York City (Havers et al., 2020). When compared to the locally reported confirmed numbers of COVID-19 cases during the same period, these seropositive rates suggest that >10 fold more SARS-CoV-2 infections occurred in these sites than the reported cases. In two other large-scale studies, 3.2-3.8% of a total of 35,759 human subjects in Wuhan, China tested positive for anti-SARS-CoV-2 IgG or IgM (Ling et al., 2020;Xu et al., 2020). The positive rates suggest a 3-4 times higher infection rate than the officially reported case numbers. Thus, these studies unequivocally demonstrate that a large fraction of SARS-CoV-2 infected humans were asymptomatic. It should be noted that these studies were based on samples largely collected in the first 4 months of the COVID-19 pandemic. As the pandemic rapidly expands both in case numbers and regions, the seropositive rates will undoubtedly change with the infections.
Humoral responses in SARS-CoV-2 infected asymptomatic individuals differed from those in the symptomatic group in several ways (Long et al., 2020b). They had a longer duration of viral shedding and lower virus-specific IgG levels in the acute phase. The asymptomatic individuals also had reduced IgG and neutralizing antibody levels during the early convalescent phase when compared to the symptomatic group. Furthermore, 40% of the asymptomatic individuals became seronegative, versus 12.9% in the symptomatic group (Long et al., 2020b). These results suggest that asymptomatic infected humans have weak immune responses to SARS-CoV-2 infection and raise a critical issue of whether these individuals have protective immunity against a second infection. A recent study also demonstrated that anti-RBD IgG rapidly decayed in mild COVID-19 patients (Ibarrondo et al., 2020). These observations are important in raising the question of how protective anti-SARS-CoV-2 immunity is in asymptomatic and mild cases of COVID-19, which account for the vast majority of infections. Furthermore, the observations suggest that vaccination to SARS-CoV-2 may need a strong response to induce lasting protective immunity.

Structural and neutralizing properties of anti-SARS-CoV-2 antibodies
Characterization of structural and neutralizing properties of anti-SARS-CoV-2 antibodies in COVID-19 patients has provided insights to potential mechanisms of antibody-mediated pathogenesis and virus-induced protective immunity. Antibody glycosylation is an important functional determinant for IgG. Serum IgG is highly fucosylated in healthy humans (de Haan et al., 2016). However, low fucosylation of anti-RhD IgG in pregnancy was observed as a predicting factor for hemolytic disease severity of the fetus and newborn (Kapur et al., 2014). Furthermore, nonfucosylated (afucosylated) IgG at its Fc-domain leads to enhanced antibody-dependent cellular cytotoxicity (ADCC) through increased Fc-receptor (FcRIIIa) binding and has been developed as antibody therapy (Shields et al., 2002;Ferrara et al., 2011;Pereira et al., 2018). Fucosylation in anti-S protein specific IgG1 in COVID-19 patients was analyzed and afucosylation was found in these specific antibodies (Chakraborty et al., 2020;Larsen et al., 2020). Importantly, afucosylated anti-S protein IgG1 was significantly increased in critically ill COVID-19 patients compared to that in asymptomatic and mild cases (Larsen et al., 2020). This finding suggests that afucosylated antigen specific IgG may have double roles: one to promote killing of SARS-CoV-2 infected target cells through ADCC and the other to mediate the ADE effect that exacerbates COVID-19 pathogenesis.
The kinetics of neutralizing activities of anti-SARS-CoV-2 antibodies in COVID-19 patients were measured in several studies using pseudotyped particles or native virus (Harvala et al., 2020;Ni et al., 2020;Prevost et al., 2020;Wang et al., 2020d;Wu et al., 2020a). Neutralizing antibodies developed within 2 weeks after disease onset and neutralizing activities significantly decreased after symptom resolution (Prevost et al., 2020). Notably, one study showed that 40% of convalescent patients did not have any neutralizing activity (Prevost et al., 2020), suggesting that neutralizing antibodies are not absolutely required for infection resolution. As expected, significant correlations between neutralizing activities and anti-S protein antibody titers were observed (Ni et al., 2020;Prevost et al., 2020;Salazar et al., 2020;Wu et al., 2020a). Neutralizing antibody titers were also strongly correlated with the numbers of virus-specific T cells (Ni et al., 2020), suggesting coordinated humoral and cellular immune responses. The anti-SARS-CoV-2 antibodies in COVID-19 patients exhibit crossreactivity and cross-neutralizing capability to SARS-CoV. Antibodies to S and N proteins of SARS-CoV-2 react to S and N antigens from SARS-CoV Long et al., 2020a;Prevost et al., 2020). Furthermore, anti-SARS-CoV-2 S protein antibodies neutralize SARS-CoV S protein pseudotyped viral particles (Prevost et al., 2020). Thus, neutralizing antibodies to SARS-CoV-2 were readily produced in COVID-19 patients and were cross-reactive and cross-neutralizing to SARS-CoV.
Did anti-SARS-CoV-2 antibodies pre-exist before the pandemic? Recent studies by Diez et al. characterized antibody reactivities and neutralizing capabilities to SARS-CoV-2, SARS-CoV, and MERS in commercial IVIG products (Diez et al., 2020a;Diez et al., 2020b). The IVIG products were generated with plasma collected in the US and European countries from March 2018 to October 2019. The IVIG products exhibited cross-reactivities to components of SARS-CoV-2, SARS-CoV, and MERS-CoV (Diez et al., 2020a). Interestingly, the IVIG preparations had neutralizing activities to SARS-CoV-2 and SARS-CoV but not MERS-CoV (Diez et al., 2020b). These studies indicate pre-existing crossreactive antibodies in the general population and provide a potential source of treatment for COVID-19 patients.

Are recovered COVID-19 patients protected from re-infection?
As discussed above, a majority of COVID-19 patients developed one or more antigen-specific antibodies to SARS-CoV-2. However, antibody titers were generally low in asymptomatic and mild cases as well as immunosuppressed patients. Furthermore, anti-SARS-CoV-2 antibodies appear to decay shortly after disease resolution in these cases. Thus, an important question regarding immune protection is whether the recovered COVID-19 patients are protected from re-infection. It was shown that rhesus macaques infected with SARS-CoV-2 were protected from re-infection . However, one report showed that 3 patients free of viral load but having low levels of anti-S and Nspecific antibodies were re-infected . In another report, 3 healthcare workers exhibited distinct episodes of recurrent symptoms after initial resolution and re-detection of SARS-CoV-2 RNA up to 60 days after disease onset (McGrath et al., 2020). Furthermore, 262 recovered and discharged COVID-19 patients were followed for re-occurrence (re-infection) for at least two weeks in Shenzhen, China and 38 of the patients became re-detectable positive for SARS-CoV-2 RNA . The authors attributed to the re-detection of SARS-CoV-2 RNA to long-term carriers of the virus, as shown by others Lan et al., 2020). However, there was no sufficient evidence to rule out reinfections of SARS-CoV-2 in these cases. Therefore, it is possible that patients with primary SARS-CoV-2 infection with weak and defective immune responses may not be protected from re-infection upon subsequent encounter.

Convalescent plasma and IVIG treatment of COVID-19 patients
Antibody-based therapy is an important strategy for infectious disease prevention and treatment. Given the lack of effective and specific SARS-CoV-2 inhibitors, neutralizing antibodies in convalescent plasma (CP) from recovered COVID-19 patients represent a potential treatment to reduce viremia and improve patients' conditions. This therapy has been approved by regulatory agencies in several countries. Limited previous clinical experiences with CP treatment in SARS-CoV, MERS, influenza, and Ebola suggest that CP may provide potential benefits to COVID-19 patients(Bloch et al., 2020;Devasenapathy et al., 2020). Many recent reports have documented clinical outcomes of CP treatment in severe and critically ill COVID-19 patients (Ahn et al., 2020;Duan et al., 2020;Jin et al., 2020;Keith et al., 2020;Li et al., 2020a;Li et al., 2020b;Liu et al., 2020a;Madariaga et al., 2020;Pei et al., 2020;Perotti et al., 2020;Putera and Hardianti, 2020;Shen et al., 2020a;Ye et al., 2020;Zhang et al., 2020a). Two categories of results are presented. First, CP treatment demonstrated clinical efficacy in severe and critically ill COVID-19 patients in most of these reports, including rapid conversion to negative viral tests, improved clinical symptoms such as fever and oxygen requirement, increased lymphocyte counts, decreased inflammation, and better survival for nonintubated patients. A second category of results show patients did not clinically benefit from CP treatment. Intubated and moribund critically ill COVID-19 patients did not have improved survival upon CP treatment Liu et al., 2020a).
Several caveats need to be considered in interpretation of these clinical efficacy reports. First, the strength of the data is limited by the sample size. Among the above 14 studies, 10 reported CP treatment in 10 or less COVID-19 patients (Ahn et al., 2020;Duan et al., 2020;Jin et al., 2020;Keith et al., 2020;Li et al., 2020a;Madariaga et al., 2020;Pei et al., 2020;Shen et al., 2020a;Ye et al., 2020;Zhang et al., 2020a). Three studies treated 37-46 patients Perotti et al., 2020;Putera and Hardianti, 2020). Only one randomized trial wherein 52 patients received CP treatment was reported and the results demonstrated that CP therapy plus standard treatment in severe or life-threatening COVID-19 patients did not result in significant clinical improvement . Thus, randomized controlled studies with sufficient power are needed to establish the clinical efficacy of CP treatment in COVID-19 patients. Second, the treatment procedures varied in different studies. These include variables such as times to collect plasma, definition of recovered patients, methods to process CP products, antibody titers and amounts of the re-infused CP, and patient selection criteria. The wide variations in these studies prevent comparison and solid conclusions on the efficacy of CP treatment in COVID-19 patients.
IVIG products represent another potential antibody treatment as these products exhibit crossreactivities and neutralizing capabilities to SARS-CoV-2 (Diez et al., 2020a;Diez et al., 2020b). Two studies reported clinical efficacies of IVIG infusion to treat severe and critically ill COVID-19 patients Shao et al., 2020). In one study, three severe COVID-19 patients receiving high-dose IVIG daily for 5 days had clinical improvement . In the second multicenter retrospective cohort study, 174 COVID-19 patients were treated with high-dose IVIG in comparison to 152 patients in the control group . Subgroup analysis demonstrated that critically ill patients had reduced mortality at 28 days, decreased inflammation, and improved organ functions. This study suggests that early treatment of COVID-19 patients within 7 days after admission with a high dose (>15 g/d) is beneficial and may be an effective treatment strategy. The advantage of using IVIG products to treat COVID-19 patients is the relatively large supply of commercial products with uniform quality.

Antibody-mediated pathogenesis in COVID-19 patients
On one hand, antigen-specific antibodies protect hosts from infection and enhance pathogen clearance. One the other hand, antibodies may contribute to disease progression ( Figure 1B). A potential mechanism of antibody-mediated pathogenesis in COVID-19 patients is ADE (Arvin et al., 2020;Eroshenko et al., 2020;Fierz and Walz, 2020). ADE refers to a process by which antigen-specific antibodies enhance viral entry and replication in host cells and increase infectivity and virulence. ADE has been observed in several types of viral infections including dengue virus, Zika virus, Ebola virus, and coronaviruses SARS-CoV and MERS-CoV (Peiris et al., 2003;Wang et al., 2014;Bardina et al., 2017;Katzelnick et al., 2017;Kuzmina et al., 2018;Liu et al., 2019;Wan et al., 2020b). To date, there is no report demonstrating a role for ADE in COVID-19 pathogenesis. However, several features of antibody responses in COVID-19 patients suggest that ADE may contribute to disease severity. First, as discussed above, the levels of anti-SARS-CoV-2 antibodies are correlated with COVID-19 disease severity Liu et al., 2020c;Lynch et al., 2020;Wang et al., 2020d;Zhao et al., 2020). Severe and critically ill COVID-19 patients had significantly higher anti-SARS-CoV-2 antibody titers than mild and asymptomatic cases. Although it was thought that ADE is primarily mediated through weak neutralizing or non-neutralizing antibodies(Arvin et al., 2020), a study showed that patients with worse clinical classification had higher neutralizing antibody titers (Wang et al., 2020d). Second, it was observed that a significant number of COVID-19 patients had IgG seroconversion earlier than IgM (Long et al., 2020a). This phenomenon may be caused by cross-reactive B cell responses to other coronavirus family members previously infecting the hosts(Tetro, 2020). Consistent with this notion, IVIG products contain cross-reactive antibodies to SARS-CoV, MERS-CoV, and SARS-CoV-2 (Diez et al., 2020a). Third, it was observed that children with COVID-19 have milder symptoms and better prognoses than adults (Ludvigsson, 2020), and age is a significant determinant in COVID-19 severity (Cummings et al., 2020;Yang et al., 2020;Zhou et al., 2020a). Together, these features of antibody responses support the hypothesis that rapid and increased induction of antigen-specific IgG upon SARS-CoV-2 infection in adults may result in ADE and contribute to COVID-19 pathogenesis (Fierz and Walz, 2020) (Figure 1B).
Another potential mechanism of antibody-mediated pathogenesis in COVID-19 patients may be the production of autoantibodies in severe and critically ill patients ( Figure 1B). Severe/critically ill patients with clinically significant coagulopathy were reported to have antiphospholipid (anticardiolipin (CL) and anti-2-glycoprotein I (2GPI)) IgA, IgM, and IgG antibodies (Sung and Anjum, 2020;Vlachoyiannopoulos et al., 2020;Zhang et al., 2020e). Furthermore, autoantibodies related to systemic autoimmune rheumatic diseases including anti-nuclear antibodies (ANA), anti-neutrophil cytoplasmic antibodies (ANCA), anti-extractable nuclear antigens (ENA), anti-dsDNAs, and anticyclic citrullinated peptide (CCP), as well as anti-phospholipid antibodies, were detected in 20/29 (68.7%) patients (Vlachoyiannopoulos et al., 2020). Importantly, none of these patients had a history of systemic autoimmune rheumatic disease. Similarly, rheumatoid factor, ANA, and ANCA were detected in another cohort of severe and critically ill COVID-19 patients (Mobini et al., 2020). These clinical findings suggest that SARS-CoV-2 infection may induce autoimmune activation. This is supported by a recent study showing that critically ill COVID-19 patients display lupus-like hallmarks of extrafollicular B cell activation (Woodruff et al., 2020). B cells from critically ill COVID-19 patients were enriched for extrafollicular responses and induced to generate large numbers of antibody secreting cells. These features of B cell responses were associated with high rates of mortality in COVID-19 patients (Woodruff et al., 2020). Thus, the above studies suggest that SARS-CoV-2 infection in some patients induces autoimmune activation and production of autoantibodies. The autoantibodies contribute to COVID-19 pathogenesis through antibody-mediated organ damage.
This rapid isolation of many neutralizing antibodies is a testament to the impressive worldwide effort in the search for means to treat and prevent COVID-19. Several important characteristics are reported about these neutralizing antibodies. The antibodies are selected from large pools of binding antibodies and have high affinity and potent neutralizing efficacy in blocking infections by pseudotyped virus and/or authentic SARS-CoV-2. Surprisingly, few antibodies isolated from COVID-19 patients cross-react to SARS-CoV antigens despite high homology between these two viruses. In contrast, efforts to screen cross-reactivities to SARS-CoV-2 in previously isolated neutralizing antibodies to SARS-CoV was fruitful (Huo et al., 2020;Pinto et al., 2020;Tai et al., 2020;Tian et al., 2020;Wec et al., 2020;Yuan et al., 2020b). Furthermore, most of these neutralizing antibodies bind to epitopes in the RBD; several also bind to the NTD domain ( Table 2). Epitope mapping and crystal structure studies of the antibodies revealed various binding sites in the RBD (Brouwer et al., 2020;Hansen et al., 2020;Huo et al., 2020;Ju et al., 2020;Pinto et al., 2020;Robbiani et al., 2020;Rogers et al., 2020;Seydoux et al., 2020;Shi et al., 2020;Wan et al., 2020a;Wu et al., 2020c;Zost et al., 2020). Some neutralizing antibodies bind to the ACE2 interaction sites on RBD, therefore sterically hindering viral engagement while others bind to non-ACE2 interacting areas on RBD. The antibodies' differential recognition of epitopes on RBD and NTD provides candidates for developing therapeutic antibody cocktails to prevent virus escape mutants. Indeed, a potent neutralizing antibody CC6.33 needed a 10-fold higher IC50 to neutralize a mutant (V367F) strain while another antibody CC6.29 completely lost its neutralizing capability to another mutant strain (G476S) compared to their neutralizing capability to the wildtype WA-1 strain (Rogers et al., 2020). Thus, combined use of various neutralizing mAbs is needed in future prevention and treatment.  (Hansen et al., 2020) 13. CR3022 Human* 0.114 6.3-115 N/A RBD Yes (Huo et al., 2020;Tian et al., 2020;Yuan et al., 2020b) 14.
S309 Human* 0.079 <0.001 N/A RBD Yes (Pinto et al., 2020) 1. Hu-mice: humanized mice. 2. Neutralizing efficacy was reported as IC50, ND50, or ID50. 3. N/A: not available. 4. * from SARS-CoV patient. 7. ** representative of 33 clones. 5. ***: representative of 19 clones Several neutralizing antibodies have been tested for their in vivo preventive and therapeutic efficacies in different SARS-CoV-2 infection animal models (Rogers et al., 2020;Shi et al., 2020;. The CB6 mAb was tested in rhesus macaques in both prophylactic and treatment settings . Administration of the mAb at 50 mg/kg reduced virus titers by ~3 log10 compared to PBS treated groups. Prophylactic injection of the mAb at 50 mg/kg significantly reduced viral infection . In the hACE2 transgenic mouse model, a single 25 mg/kg dose of neutralizing antibodies administered separately after virus challenge reduced the viral RNA copies in the lung of the infected mice by 2-3 log10 . In a third Syrian hamster model, administration of neutralizing mAbs at 4.2-16.5 kg/mg completely prevented SARS-CoV-2 induced bodyweight loss in the infected animals (Rogers et al., 2020). Together, these in vivo studies provide promising results for further development of the neutralizing mAbs into clinical drugs.

Progress of COVID-19 vaccine development
The S protein of SARS-CoV-2 is the primary target for COVID-19 vaccines due to its critical role in host infectivity. Many different forms of vaccines are under intensive development Zhang et al., 2020c). Some of the COVID-19 vaccine candidates have been tested in in vivo infected animal models. For example, a series of DNA vaccine candidates expressing different forms of the S protein showed protective effects by reducing SARS-CoV-2 by 3-4 log10 in nonhuman primates . A purified inactivated SARS-CoV-2 virus vaccine induced neutralizing antibodies in mice, rats, and nonhuman primates and provided partial or complete protection in macaques against virus challenge . Results from measles virus and vesicular stomatitis virus (VSV) based vaccine candidates in animal models were also reported (Horner et al., 2020;Yahalom-Ronen et al., 2020). Development of two major forms of COVID-19 vaccines, mRNAnanoparticle-based and adenovirus-vectored, have made rapid progress with both reporting results from human clinical trials.
In a phase I study, administration of mRNA-1273 vaccine encoding a stabilized prefusion S protein trimer to 45 healthy adults (18-55 years old) resulted in production of neutralizing antibodies in all participants at levels similar to or higher than those in convalescent sera (Jackson et al., 2020). The vaccine injected once or twice at three doses (25, 100, and 250 g) was safe with no serious adverse events. Furthermore, a Th-1-biased CD4 + T cell response was induced by mRNA-1273 vaccination. The phase I results support further development of this vaccine, which is currently in a phase II trial with 600 subjects, and a phase III trial begins in July 2020. In a second phase I study, BNT162b1, a lipid nanoparticle-formulated, nucleoside-modified mRNA vaccine encoding a trimerized RBD, was immunized twice in 45 18-55 year-old healthy adults at 3 different doses (10, 30, and 100g) (Mulligan et al., 2020). Neutralizing antibodies were induced to titers higher than those in convalescent plasma, while the vaccine was well tolerated with generally mild to moderate local reactions and systemic events. T cell responses were not reported in this study (Mulligan et al., 2020). The results support further clinical development of BNT162b1.
Two independent adenovirus-vectored SARS-CoV-2 vaccines have been tested in human trials. A human adenovirus type-5 (Ad5) vectored COVID-19 vaccine was administered to a total of 490 adults 18 years and older in phase I and II trials Zhu et al., 2020b). Ad5-vectored COVID-19 expressing a full-length S protein was injected intramuscularly once or twice at three doses of 5×10 10 , 1×10 11 , and 1·5×10 11 viral particles per ml. Most of the participants (96-97%) were seropositive at day 28 with significantly elevated neutralizing antibodies . Furthermore, antigen-specific T cell responses were observed in ~90% of the subjects. Although adverse events were reported in 77% of participants, no serious events were documented. However, some weaknesses for this vaccine candidate were noted. Humoral responses were lower in people over 55 years of age and 52% of participants with high pre-existing anti-Ad5 vector neutralizing antibody titers . A chimpanzee adenovirus-vectored vaccine termed ChAdOx1 COVID-19 was tested in a randomized controlled phase I/II trial in 543 participants aged 18-55 years with the same number of participants receiving meningococcal conjugate vaccine as controls (Folegatti et al., 2020). Seroconversion was observed in all participants. Neutralizing antibodies were positive in all 35 tested participants, and antigen-specific T cell responses were induced in all 43 tested participants, suggesting strong immunogenicity. Furthermore, ChAdOx1 COVID-19 vaccine was safe with no serious adverse events. The above studies support further phase III clinical trials for both adenovirus-vectored COVID-19 vaccine candidates.

Significance of antibody responses in vaccine development.
The above discussed antibody responses from recent clinical studies on COVID-19 patients have important implications in vaccine development ( Figure 1C). The first implication is in the safety of the vaccines. Both mRNA-nanoparticle and adenovirus-vectored vaccines have demonstrated overall acceptable safety profiles in phase I and II trials. The local and systemic events such as pain, fever, fatigue, chills, muscle ache, headache, and other symptoms were generally mild/modest and can be further alleviated with prophylactic paracetamol administration (Folegatti et al., 2020;Jackson et al., 2020;Mulligan et al., 2020;Zhu et al., 2020a;Zhu et al., 2020b). However, ADE of disease severity(Arvin et al., 2020), the critical adverse event that may be induced after vaccination, has yet to be determined in human population upon SARS-CoV-2 infection. Several studies have demonstrated ADE of disease severity upon immunization with different forms of SARS-CoV vaccines and re-challenge of SARS-CoV live virus in mice, ferrets, and non-human primates (Weingartl et al., 2004;Bolles et al., 2011;Tseng et al., 2012;Wang et al., 2016). The vaccine-induced immunopathology includes increased lung infiltration of eosinophils, neutrophils, and monocytes and enhanced hepatitis. Importantly, aged animals displayed more severe immune pathology and were not effectively protected against virus replication (Bolles et al., 2011). Thus, these observations raise important questions on the safety of COVID-19 vaccines: whether the immunized people develop ADE of disease severity and whether they have full protection against SARS-CoV-2 infection, especially in aged human populations. These critical questions will hopefully be adequately addressed in large scale phase III trials carried out in COVID-19 pandemic areas.
The second implication of the antibody responses in SARS-CoV-2 infected populations is related to the efficacy of COVID-19 vaccines. As discussed, anti-SARS-CoV-2 antibody responses in asymptomatic and mild COVID-19 patients are weak and the antibodies disappear rapidly. In light of these findings, the idea to develop herd immunity by wide-spread natural infection may not be feasible given that most infected persons are mild or asymptomatic. The clinical trial results showing that mRNA-and adenovirus-based COVID-19 vaccines induce antibody titers similar to or higher than those in CP of COVID-19 patients are encouraging (Folegatti et al., 2020;Jackson et al., 2020;Mulligan et al., 2020;Zhu et al., 2020a;Zhu et al., 2020b). However, these trials were conducted in mostly young and middle-aged adults. Thus, it remains unclear whether these vaccines can induce protective neutralizing antibodies in aged populations that most need protection.
The third implication of the antibody responses in COVID-19 patients relates to the efficacy of vaccines in neutralizing SARS-CoV-2 mutants. One of the major mutations in SARS-CoV-2 genome with biological significance is a point D to G change at position 614 in the S protein, now accounting for 97% of isolates worldwide (Korber et al., 2020). The D614G SARS-CoV-2 variant has several advantages over the wildtype virus including 3-10-fold more infectivity, growing to higher titers in vitro, and being more abundant in patient's respiratory tract Korber et al., 2020;Yurkovetskiy et al., 2020;Zhang et al., 2020d). Although some neutralizing mAbs retained reactivity to the D614G variant (Rogers et al., 2020;Yurkovetskiy et al., 2020), it was reported that 7% (3/41) of convalescent sera had decreased neutralizing activity against the G614 mutation . Furthermore, some of the neutralizing mAbs lost their capacity to inhibit viral replication of other point mutation variants (G476S, V367F) (Rogers et al., 2020). Thus, a critical issue in COVID-19 vaccine development is to maintain protective immunity against SARS-CoV-2 escape mutants.
In summary, the recent progress in characterization of antibody response and experimental treatment in COVID-19 patients as well as the isolation of a large panel of mAbs have shed light on the pathophysiology and immunobiology of SARS-CoV-2 infection. The preliminary results from clinical trials of several COVID-19 vaccine candidates are encouraging, but major issues need to be addressed before clinical approval.