Serological Assessment of COVID-19 Patients in Brazil: Levels, Avidity, and Subclasses of IgG Against RBD

SARS-CoV-2 is considered a global emergency, resulting in an exacerbated crisis in the health public in the world. Although there are advances in vaccine development, it is still not available for many countries. On the other hand, an immunological response that mediates protective immunity or indicates that predict disease outcome in SARS-CoV-2 infection remains undened. This work aimed to assess the antibody levels, avidity, and subclasses of IgG to RBD protein, in symptomatic patients with severe and mild forms of COVID-19 in Brazil using an adapted in-house RBD-IgG ELISA. The RDB IgG-ELISA showed 100% of specicity and 94.3% of sensibility on detecting antibodies in the sera of hospitalized patients. Patients who presented severe COVID-19 had higher anti-RBD IgG levels compared to patients with mild disease. Additionally, most patients analyzed displayed low antibody avidity, with 64.4% of the samples of patients who recovered from the disease and 84.6% of those who died in this avidity range. Our data also reveals an increase of IgG1 and IgG3 levels since the 8th day after symptoms onset, while IgG4 levels maintained less detectable during the study period. Surprisingly, patients who died during 8-14 and 15-21 days also showed higher anti-RBD IgG4 levels in comparison with the recovered (P < 0.05), suggesting that some life-threatening patients can elicit IgG4 to RBD antibody response in the rst weeks of symptoms onset. Our ndings constitute the effort to clarify IgG antibodies' kinetics, avidity, and subclasses against SARS-CoV-2 RDB in symptomatic patients with COVID-19 in Brazil, highlighting the importance of IgG antibody avidity in association with IgG4 detection as tool laboratory in the follow-up of hospitalized patients with more signicant potential for life-threatening. the effort to clarify the kinetics of IgG antibodies, avidity, and subclasses against SARS-CoV-2 RDB in symptomatic patients with COVID-19 in Brazil, highlighting the importance of IgG antibody avidity in association with IgG4 detection as a laboratory tool in the follow-up of hospitalized patients with more signicant potential for life-threatening conditions in the population analyzed.


Introduction
COVID-19 (Coronavirus disease 2019), the most recent pandemic caused by severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2), resulting in an exacerbated crisis in the health public, declared as a global emergency by World Health Organization (WHO) 1 in March 2020. SARS-CoV-2 was rst reported in Wuhan, China, in December 2019, resulting in a large number of individuals presenting symptoms such as fever, dry cough, dyspnea, body pain, loss of taste and smell, and sometimes atypical pneumonia that might be fatal in a small percentage of cases (around 5%) 2,3,4 . Brazil became the epicenter of COVID-19 5 in June of 2020, averaging about 1.000 deaths per day 6 . After a gradual reduction in the number of deaths from COVID-19 7 , nowadays Brazil has experienced a rise in the number of cases and deaths associated with SARS-CoV-2 infection. This scenario may impose new challenges to health services shortly, including the requirement for novel rapid diagnostic tools to interrupt the COVID-19 epidemiological chain 8 .
SARS-CoV-2 is a novel human-infecting Betacoronavirus that is enveloped, non-segmented, positivesense single-stranded RNA of around 65-125 nm diameter with crown-like spikes proteins on its outer surface 9 . The new coronavirus genome encodes nonstructural proteins from two open reading frames 1a and 1b (Open reading frame 1 -ORF1), and structural proteins, such as spike (S), small envelope (E), membrane (M), and nucleocapsid (N) glycoprotein 10 . The S glycoprotein is a transmembrane protein with a molecular weight of 150KDa that binds to the angiotensin-converting enzyme 2 (ACE2) or CD147 receptor expressed on the host cells surface through receptor binding protein (RBD) presented in the S1 subunit of S protein 11,12 .
Although the gold-standard diagnosis test is based on detecting viral nucleic acid by real-time reverse transcription-polymerase chain reaction (RT-PCR) assay, serological tests as anti-COVID-19 IgG/IgM rapid tests have been applied mainly as an essential diagnostic tool to evaluate the seroprevalence of SARS-CoV2. Nevertheless, their accuracy is inferior to 13,14,15 . Therefore, the development of new serological tests is an urgent need because RT-PCR, used for disease diagnosis with high sensitivity and speci city, can only detect virus from pharyngeal and saliva specimens samples from the rst week up to 10 days after symptoms onset 16,17 , and asymptomatic subjects can be missed from this screening. Also, manipulations of potentially infectious materials, which may cause splashes, droplets, or aerosols, are highly recommended to be done in a biosafety level − 2 (BSL-2) structure at least, besides RT-PCR is a laborious and expensive technique 18,19,20,21 .
Studies reveal that the higher infectivity capacity of SARS-CoV-2 compared to other coronavirus is due to genomic mutations on the RBD domain from the S protein that plays an essential role in virus attachment and invasion to the host cell 22 . On the other hand, because the RBD domain is highly immunogenic and induces IgG antibody response in acutely infected patients, it is considered a potential target for serological assays [23][24][25][26][27][28][29] and vaccine development 10,30,31 . Although there are advances in vaccine development 32,33 , it is still not available for many countries. On the other hand, the immunological response in SARS-CoV-2 infection remains unde ned, with several data still emerging in the scienti c literature concerning this new coronavirus pandemic 34,35,36 While virus nucleic acid is detected between 3-10 days after infection, antibody production can only be detected at least 7 days after symptoms onset, similar to SARS-CoV-1 infection 37  such as an association between antibody neutralizing titers and antibody avidity 42 . Antibodies produced by patients that recovered from coronavirus disease 2019 (COVID-19) have been used as a "molecular framework" to design antibody-based therapies 43 . Thus, the study aimed to evaluate the kinetic, level, and avidity of antibodies IgG RBD-speci c and the relationship of speci c IgG subclasses to the severity of the disease.

Results
Demographic data, clinical aspects, and comorbidities.  Otherwise, from the second week on after onset symptoms, the presence of IgG RDB-speci c antibodies was observed among 88.6% of patient's samples, reaching 100% of positivity in the next time-point, highlighting that the sensitivity of RBD-based ELISA improves according to the timing after the onset of clinical symptoms ( Fig. 1b-h). The general immunoreactivity of RDB ELISA (including rst and second week) resulted in 100% of speci city and 94.3% of sensibility in the population analysed ( Fig. 1).
IgG antibody avidity to SARS-CoV-2 RBD Samples collected from 15 days after onset of symptoms from patients with COVID-19 who symptoms were evaluated for binding strength of IgG antibodies to RBD protein. The majority of patients analyzed displayed low antibody avidity, with 64.4% of the samples of patients who recovered from the disease and 84.6% of those who died in this avidity range (Fig. 3a). Intermediate antibody avidity was observed in 33.3% and 15.4% in the recovered or died group, respectively. Remarkably, one sample from a patient that recovered displays high avidity to SARS-CoV2 RBD, despite a relatively short period (weeks) posts the onset of the symptoms. When the samples were analyzed in parallel, in each avidity range (low, intermediate, high), no differences were detected between patients who recovered or died from covid-19 (Fig. 3a). Similarly, when samples from patients who recovered or died were analyzed each week (7-day intervals) comparatively, no statistically signi cant difference was observed (Fig. 3b).

Antibody subclasses response in symptomatic patients with COVID-19
IgG subclasses maybe become relevant in clinical conditions COVID-19, considering that IgG subclasses to SARS-CoV-2 are a key to a better clinical condition, with IgG1 and IgG3 be more abundant in patients that are in the mild case and do not die. Our data reveal an increase of IgG1 and IgG3 levels since the 8th day after symptoms onset ( Fig. 4a), while IgG4 levels maintained less detectable during the study period.
Anti-RBD IgG1 positivity oscillated from 66.6-100%, reaching the highest values in the third and fourth week of analysis. The positivity of anti-RBD IgG3 ranged from 66.6-90.9%, whereas IgG4 presented positivity from 66.6 to 46.1%, with higher values for IgG3 and IgG4 positivity observed in the third and fourth weeks, respectively.
When patients who died and who recovered were analyzed in parallel, it was possible to identify a higher anti-RBD IgG1 response in patients who died compared to those who recovered during 8-14 days after symptoms onset (P < 0.05), as illustrates in the Fig. 4b. No statistical differences were observed in IgG3 levels from recovered and dead patients (Fig. 4c). Surprisingly, patients who died during 8-14 and 15-21 days also showed higher anti-RBD IgG4 levels in comparison with the recovered (P < 0.05) (Fig. 4d), suggesting that some life-threatening patients can elicit IgG4 to RBD response in the rst weeks of symptoms onset.  43 , suggesting an association between illness severity and antibody production.
Interestingly, in our study, survivors who developed the severe form of illness displayed higher anti-RBD IgG levels compared to patients with mild disease, also lighting that clinical presentation of disease may produce substantial differences in IgG responses. For this purpose, we also investigated the anti-RBD IgG levels in recovered patients compared to those who died. Although no statistical differences were observed, a slightly lower anti-RBD IgG level was observed in patients who died compared to the recovered ones. The divergence observed in our study may be due to differences in the number of serum samples analyzed and the time of blood collection in our study since a limited number of samples was obtained from critically ill patients who quickly progressed to death by COVID-19.
Besides the comorbidities observed in patients who died, obesity was more frequently observed in our study. The high percentage of obese patients who died with COVID-19 is in concordance with previous studies 48, 49 . Obesity is a factor that directly associated chronic activation of innate immune system cells and consequent local and systemic in ammation 50 . B cells from obese patients express leptin induced-activation markers (TNF-α, TLR4, micro-RNAs) that correlate reduced B-cell functions 49,51,52 . Therefore, obesity and COVID-19 share common elements of the in ammatory process (and possibly also metabolic disturbances), exacerbating SARS-CoV-2 infection in the obese 53 , leading these individuals to severe COVID-19, even to death.
In the present study, high levels of anti-RBD IgG were detected in both groups of patients who died or survived. However, these ndings are not enough to support the hypothesis that these individuals displayed either an extended-lasting protective humoral response against SARS-CoV-2 or neutralizing antibodies in convalescent plasma. Our study has some limitations in this context since it was not possible to assess the presence and neutralizing antibodies. Interestingly, a possible association between SARS-CoV-2 spike antibody avidity with neutralizing IgG titter, as a potential screening parameter for identifying optimal convalescent plasma donors was proposed 42 . Likewise, high-avidity antibodies toward another virus capable of blocking receptor binding were protective and promoted virus neutralization 54 , indicating that antibody avidity maturation could be, at least in part, associated with the production of protective neutralizing antibodies in viral infections. Although our study has limited data on temporal dynamics (< 45 days) to correlates SARS-CoV-2 antibody avidity with the illness severity, it was observed that the majority of patients who had symptoms showed low avidity. Our data is by what was reported to SARS-CoV-2 infection with low IgG antibody avidity during the 50 days after symptoms onset 42 , however, evaluating antibodies against nucleoprotein 55 , and to SARS-CoV-2 anti-spike and antinucleoprotein, that reached the avidity peak at 21 after days of symptoms onset, during a study period median of 45days 42 .
Similarly, low antibody avidity was also observed early infection and augmented within the rst month of symptom onset in SARS outbreaks 56 . It is noteworthy that one-third of the patients who recovered had intermediate avidity of IgG antibodies to RDB, despite a relatively short period (weeks) post the onset of the symptoms. On the other hand, approximately one-sixth of the patients who died produced intermediate avidity antibody responses, suggesting that IgG avidity may be useful for monitoring hospitalized patients with COVID-19 in association with other serological markers. As was expected, IgG antibody avidity was low during initial infection and increased with time, although no statistical differences were observed between patients who died and recovered in the time-points post the onset of the symptoms.
The IgG1 and IgG3 subclasses represent the predominant antibody responses to several viral diseases 57,41 , and recently it was also associated with the new SARS-CoV − 2 infection 29,58 . IgG1 and IgG3 responses are related to immune functions such as viral neutralization, opsonization, and complement activation in viral respiratory infection 41 . Thus, to further analyze the antibody response, we also analyzed IgG to SARS-CoV-2 RDB in sera from patients with COVID-19. As it was expected, a robust antibody response of IgG1 and IgG3 speci c to SARS-CoV-2 RDB occurred predominantly in comparison with minor IgG4 responses. Likewise, Suthar et al.; 2020 demonstrated that COVID-19 patients analysed in USA produced RBD-speci c IgG1 and IgG3 early during acute infection, with no detectable IgG2 or IgG4 29 . Similar antibody responses were also reported by Mazzini et al.; in Italy, with a strong reactivity for IgG1 and IgG3 in sera from positive patients for SARS-CoV-2 infection 58 .
The comparative analysis of IgG subclasses in serum samples from COVID-19 patients who died revealed a higher level of RDB-speci c IgG1 when compared to those who survived during 0-8 days after symptoms onset. However, this difference was not maintained in more advanced times of the onset of the symptoms. Although the production of RDB-speci c IgG1 is consistent with activation of type 1 T helper lymphocytes (Th1) 59 , this focal difference in the rst week cannot be explained simply by balancing different subpopulations of T helper cells but maybe involve other factors, including sample bias, differences in individual immune responses, and/or early viral load. Also, no statistical differences were observed in IgG3 levels between patients who died and recovered. Surprisingly, we also noticed higher levels of RBD-speci c IgG4 in sera from patients who died when compared to survivors in the second and third weeks of analysis. In our analysis, 55% (10 out of 18) serum samples of patients who progressed to death showed early positivity to RDB-speci c IgG4 antibodies, whereas patients who recovered from COVID-19 were IgG4 negative to SARS-CoV-2 RDB in the same window of time. The IgG4 biosynthesis is known to be induced under conditions of increased IL-10 cytokine 60 having as a primary source several immune cells, including Th2 cells, regulatory T cells (Treg), or even regulatory B cells (Breg). Patients with severe COVID-19 display sustained in ammation and continued production of various anti-and pro-in ammatory cytokines (cytokine storm syndrome) 36, including the IL-10 production that may be associated with induction of IgG4 antibodies in severe COVID-19. Although substantial knowledge about the antibody response has already been generated nowadays for COVID-19, further studies are necessary to understand the role of IgG4 antibodies in COVID-19 pathophysiology.
In conclusion, the present study constitutes the effort to clarify the kinetics of IgG antibodies, avidity, and subclasses against SARS-CoV-2 RDB in symptomatic patients with COVID-19 in Brazil, highlighting the importance of IgG antibody avidity in association with IgG4 detection as a laboratory tool in the follow-up of hospitalized patients with more signi cant potential for life-threatening conditions in the population analyzed.

Study design
Forty-seven symptomatic patients tested positive for SARS-CoV2 infection by RT-PCR were admitted at Institute of Infectology Emilio Ribas (IIER) São Paulo, Brazil, between March and June 2020 were enrolled in this study. All patients who presented typical symptoms of illness such as fever, dry cough, dyspnea, myalgia, etc., were classi ed as mild or severe, according to IIER protocols previously established for COVID-19. Blood samples were taken at different time points until either the patient was discharged or died. A total of 294 serum samples were used from patients (136 collected from patients who were discharged and 40 from patients who died) and 118 serum samples SARS-CoV2 negative, collected before September 2019 and selected from Institute Adolfo Lutz (IAL) routine, were analyzed (Fig. 5). SARS-CoV2 negative serum samples have a documented history of other viral infections (HIV-1, HIV-2, Hepatitis B, Hepatitis C, Dengue, Chikungunya, Yellow fever) non-related with any coronavirus and bacterial infections (Treponema pallidum, Mycoplasmas pneumoniae).

Ethical Approvements
The Committee for Ethics approved clinical research from both Institute of Infectology Emilio Ribas and Institute Adolfo Lutz, CAAE number: 35589320.6.3001.0075. All methods were performed in accordance with the relevant guidelines and regulations. Human samples used in this study correspond to discarded peripheral blood, collected previously for monitoring hematological characteristics of patients with suspected SARS-CoV2 infection. Demographic data, Clinical and hematological conditions at rst attendance were obtained for all patient included in the study.
Indirect ELISA for detection of SARS-CoV-2 RBD-speci c IgG antibodies (ELISA-RBD) IgG RBD-speci c antibodies were detected by Enzyme-Linked Immunosorbent Assay (ELISA) using an adapted protocol previously described by Stadlbauer et al. 24 . Brie y, high-binding 96-well plates (Nunc MaxiSorp™ at-bottom) were coated with 50 µl per well of 2.5 µg per ml of RBD protein diluted in PBS 1x at 4ºC overnight. The next morning, plates were washed four times with PBS 1x supplement with 0.01% Tween 20 (PBST). All wash steps were performed using an ELISA plate washer (Washwell plate, Robonik, Thane, India). After, 200 µl per well of 5% skim milk powder diluted in PBST was added to the plates and incubated for 2 h at room temperature as a blocking solution. After blocking, plates were washed four times with PBST and incubated with 200 µl per well of each sera sample, in duplicate, diluted 1:200 in PBST containing 1% of skin milk for 1 h at room temperature. Next, plates were washed four times with PBST and 50 µl of a 1:15000 dilution of goat anti-human IgG (whole molecule) − Horseradish Peroxidase (HRP) antibody (Sigma-Aldrich) diluted in PBST containing 1% skim milk was added to wells, and the plates incubated for 1 h at room temperature. Plates were rewashed with PBST and incubated for 10 minutes with 100 µl of One Step-TMB (3,3',5,5'-tetramethylbenzidine) (Scienco, Santa Catarina, Brazil). The reaction was stopped by the addition of 50 µl per well of 1 N sulfuric acids. The optical density at 450 nm (OD450) was measured using a Multiskan MS plate reader (Labsystems). The cut-off value was established based on the maximum sensitivity and speci city using a two-graph receiver operating characteristic (TG-ROC) analysis as previously described 61 . Antibody titers were expressed as ELISA index (EI), according to the following formula: EI = OD sample/cut-off. Samples with EI values > 1.0 were considered positive.
Evaluation of IgG antibody avidity to SARS-CoV-2 RDB Serum samples were submitted to IgG avidity ELISA using potassium thiocyanate (KSCN) as a chaotropic chemical reagent as previously described 62 to assess the interaction between them IgG antibodies and RBD. Following the above described ELISA-RBD, an extra step was performed after incubation with serum. Brie y, after serum incubation, plates were washed four times with PBST, and wells were treated incubated in the presence or absence of KSCN 1.5M (200 µl/well) for 20 minutes at room temperature. After, plates were washed four times and. Avidity index (AI%) was expressed as follows: AI%= (OD mean value from KSCN treated sample divided by the OD mean value from the nontreated) multiplied by 100. AI values above 50% were considered high antibody avidity; between 31-49%, intermediate avidity, and below 30%, low avidity 63 . Measurement of IgG speci c to SARS-CoV-2 RDB ELISA was used to detect IgG1, IgG3, IgG4 using a protocol previously described 64 . Brie y, 96-well plates (Nunc MaxiSorp™ at-bottom) were coated with 50 µl per well of 2.5 µg per ml of RBD protein diluted in PBS 1x at 4ºC overnight. Plates were then washed four times with PBS 1x with 0.01% Tween 20 (PBST) using a Washwell Plate (Robonik). Samples were blocked with 100 µl per well of 1% Bovine Serum Albumin (BSA) diluted in PBST for one h at 37º C. After blocking, plates were washed four times with PBST and incubated with serum diluted 1:50 in PBST-0,1% BSA, in duplicate, for IgG1 detection. For IgG3 and IgG4 subclasses, serum was diluted 1:5 in the same solution, and the assay was also performed in duplicate. Thus, samples were incubated for 2 h at 37ºC, washed four times, and incubated with the respective biotinylated secondary antibodies (Sigma): goat anti-human IgG1 (1:1000), anti-human IgG3 (1:1000), or anti-human IgG4 (1:1000) diluted in PBST − 0.1% BSA for 1 h at 37 •C. Plates were rewashed four times and incubated with 50 µl of 1:500 streptavidin-peroxidase (Sigma/Merck) diluted in PBST − 0.1% BSA (Sigma) for 30 min at 37ºC. After the nal washing step (four times), samples were revealed with ABTS (Sigma/Merck). The optical density at 405 nm (OD405) was measured using a Multiskan MS plate reader (Labsystems). Cut-off of reaction was calculated using optical density values of negative pools plus three standard deviations as described 59 . Antibody titers were expressed as ELISA index (EI), and values > 1.0 were considered positive.

Statistical Analyses
The data were evaluated for normal distribution by D'Agostino & Pearson, Shapiro-Wilk, Kolmogorov-Smirnov normality tests. Statistically, differences among antibody IgG levels, antibody avidity, and IgG subclasses to SARS-CoV-2 RB were determined by Kruskal-Wallis and Dunn's multiple comparisons test or Mann-Whitney test when appropriate. P values < 0.05 were considered statistically signi cant 65 . Statistical analyses and graphics were performed using the GraphPad Prism v. 8.0 (GraphPad Software, San Diego, USA).   Statistically, differences between groups were determined by the Mann-Whitney test each time (* P < 0,05).