Asymptomatic or mild symptomatic SARS-CoV-2 infection elicits durable neutralizing antibody responses in children and adolescents

As SARS-CoV-2 continues to spread globally, questions have emerged regarding the strength and durability of immune responses in specific populations. In this study, we evaluated humoral immune responses in 69 children and adolescents with asymptomatic or mild symptomatic SARS-CoV-2 infection. We detected robust IgM, IgG, and IgA antibody responses to a broad array of SARS-CoV-2 antigens at the time of acute infection and 2 and 4 months after acute infection in all participants. Notably, these antibody responses were associated with virus neutralizing activity that was still detectable 4 months after acute infection in 94% of children. Moreover, antibody responses and neutralizing activity in sera from children and adolescents were comparable or superior to those observed in sera from 24 adults with mild symptomatic infection. Taken together, these findings indicate children and adolescents with mild or asymptomatic SARS-CoV-2 infection generate robust and durable humoral immune responses that are likely to protect from reinfection.


INTRODUCTION
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the etiological agent of coronavirus disease 2019 (COVID- 19), has caused more than 140 million infections and nearly three million deaths globally. 1 The effectiveness and durability of the immune responses induced by SARS-CoV-2 infection has major implications for the risk of reinfection and the establishment of herd immunity. Studies conducted among adult populations suggest that there is substantial variability in SARS-CoV-2 humoral immune responses based on patient factors and characteristics of the initial illness. In a study of more than 30,000 individuals in New York who tested positive for SARS-CoV-2 infection by PCR, more than 90% had detectable IgG antibodies to the viral spike protein within 30 days of SARS CoV-2 acute infection, irrespective of the severity of the initial illness. 2 More recent studies have shown that antibody responses in adults are durable, with reports of detectable SARS-CoV-2-specific antibodies up to 6-8 months after acute infection. 3 Several prior studies reported an association between the durability of antibody responses and the severity of the acute infection. For example, Long and colleagues found that 40% of the asymptomatic adults and 13% of the symptomatic adults with initial antibody responses seroreverted within two months after acute infection. 4 Conversely, Kong and colleagues reported that subjects with mild or moderate symptoms had earlier and more robust antibody responses than subjects with severe illness during the 40 days after symptom onset. 5 Taken together, these studies demonstrate the marked variability of immune responses within different populations and highlight a potential association between illness severity and the development of immune responses to SARS-CoV-2.
Epidemiological data from around the world indicate that children and adolescents with SARS-CoV-2 infection typically have milder illness courses than adults. [6][7][8] Notably, SARS-CoV-2 infection in children and adolescents is often asymptomatic or associated with such minor symptoms that children infrequently come to medical attention. 9,10 The varied clinical manifestations of SARS-CoV-2 infection among children and adults suggest that age may modify the host response to SARS-CoV-2, as has previously been demonstrated for several other viruses. [11][12][13][14][15][16][17] To date, studies of SARS-CoV-2 immune responses in pediatric populations have focused primarily on children hospitalized for severe COVID-19 or who developed multisystem inflammatory syndrome in children (MIS-C), a potentially life-threatening inflammatory condition that can occur after SARS-CoV-2 infection. [18][19][20][21][22] While these studies provide important insights into the immune responses of children and adolescents who develop these rare manifestations of SARS-CoV-2 infection, surprisingly little is known about the immune responses of the much larger population of children with asymptomatic or mild symptomatic SARS-CoV-2 infection.
In this study, we evaluated the temporal evolution of SARS-CoV-2-specific humoral immune responses in a cohort of children and adolescents with asymptomatic or mildly symptomatic SARS-CoV-2 infection. We report clinical characteristics, serum titers of antibodies specific to a panel of SARS-CoV-2 antigens, and virus neutralizing antibody responses over a period of up to 4 months after acute infection. Further, we compare the antibody responses of these children and adolescents to those of a group of adults with mild symptomatic SARS-CoV-2 infection. Knowledge of the effectiveness and durability of SARS-CoV-2specific immune responses in children is critical to the development of pediatric vaccination strategies and approaches to mitigate SARS-CoV-2 transmission in schools and other congregate childcare settings.

Patient characteristics
We evaluated SARS-CoV-2-specific humoral immune responses in 69 children and adolescents (<21 years of age) who participated in a study of acute SARS-CoV-2 infection. 6 Clinical data and sera were collected from these participants during the acute phase of infection and approximately 2 and 4 months after acute infection. Median [interquartile range (IQR)] age was 11.5 (5.2, 16.5) years, and 51% were female ( Table  1). The most common chronic medical conditions were obesity (29%) and asthma (6%). Fifty-five (80%) participants reported one or more symptoms associated with acute SARS-CoV-2 infection with a median (IQR) symptom duration of 3 (2, 7) days. The most common symptoms reported were fever (49%), cough (36%), and headache (29%). The remaining 14 (20%) participants reported no symptoms associated with acute infection (2 weeks prior to SARS-CoV-2 diagnosis and up to 28 days after study enrollment) and were classified as asymptomatic. Sera from acute infection was collected at a median (IQR) of 9 (6,12) days from symptom onset or SARS-CoV-2 diagnosis in asymptomatic subjects. Sera from 2 months and 4 months after acute infection were collected at a median (IQR) of 57 (50, 71) days and 109 (101, 120) days from symptom onset or SARS-CoV-2 diagnosis, respectively. To compare humoral immune responses across the full spectrum of age, we similarly evaluated humoral immune responses to SARS-CoV-2 in 24 adults (21 years of age or older). Median (IQR) age was 43 (31, 57) years, and 58% were female ( Table 1).
All adult participants reported one or more symptoms associated with acute SARS-CoV-2 infection. No pediatric or adult subjects required hospitalization for SARS-CoV-2 infection.

SARS-CoV-2-specific antibodies in sera from children and adolescents
We used Luminex-based binding antibody multiplex assays to measure SARS CoV-2-specific antibodies targeting whole spike, subunit 1 (S1), receptor binding domain (RBD), N-terminal domain (NTD), subunit 2 (S2), nucleocapsid (NC), and membrane (M) proteins in sera from children and adolescents during acute infection (n=69), 2 months after acute infection (n=56), and 4 months after acute infection (n=50). At the time of acute infection, all children and adolescents had detectable levels of IgM and IgG antibodies to one or more SARS-CoV-2 antigens, and 68 (99%) had a detectable IgA response to one or more SARS-CoV-2 antigens. Levels of SARS-CoV-2-specific IgM and IgA antibodies were highest during acute infection and declined at 2 and 4 months after acute infection, while levels of IgG antibodies generally increased from acute infection to 2 months after infection before decreasing between 2 and 4 months after acute infection (Figure 1, Supplemental Figure 1, and Supplemental Table 1). We also measured levels of IgG subclasses to SARS-CoV-2 antigens at these same time points (Supplemental Figure 2a). Between acute infection and 2 months after acute infection, we observed significant increases in the ratios of IgG1:IgG3 antibodies to spike (p=0.04), S1 (p<0.0001), RBD (p<0.0001), S2 (p=0.03), and NC (p<0.0001); these differences persisted comparing acute infection to 4 months after acute infection for these same antigens (spike: p<0.0001; S1: p<0.0001; RBD: p<0.0001; S2: p<0.0001; NC: p<0.0001). No differences in the ratios of IgG1:IgG3 antibodies to NTD or M were observed over time (Supplemental Figure 2b).

SARS-CoV-2-infected children and adolescents generate durable neutralizing antibody responses
We next measured neutralizing activity of sera from these children and adolescents using a luciferase-based SARS-CoV-2 pseudovirus (614G) assay, as previously described. 23 Neutralizing antibodies at a 50% activity measured in this pseudovirus assay was most strongly correlated with levels of antibodies to spike, S1, and RBD (Supplemental Figure 3).

SARS-CoV-2-specific antibody responses are similar in asymptomatic and mildly symptomatic children
Given that prior studies in adults suggested that humoral immune responses to SARS-CoV-2 infection may differ based on symptom severity, 5,24-29 we compared antibody responses among children and adolescents with asymptomatic and mild symptomatic SARS-CoV-2 infection. We observed no significant differences in the levels of IgM, IgG, or IgA antibodies specific to any SARS-CoV-2 antigen during acute infection or at 2 or 4 months after acute infection in asymptomatic and mildly symptomatic children (Figure 3a-c).
Similarly, the degree of pseudovirus neutralization activity in sera did not differ based on the presence of symptoms at any of these time points (Figure 3d).  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Neutralizing antibody responses, which are believed to correlate most closely with protection from reinfection, 36,37 are reportedly stable for at least 3 months after infection in adults. 38,39 Similarly, seroconversion occurred in all children and adolescents evaluated in our study. Neutralizing activity was more likely to be detectable in sera collected from both child and adults participants who were slightly further into the course of their acute infection, indicating a slight lag in the development of an effective neutralizing antibody response. Importantly, SARS-CoV-2-specific humoral immune and virus neutralizing responses were detectable in the vast majority of subjects 4 months after acute infection, demonstrating that neutralizing antibody responses in children are similar in duration to those in adults.

Comparisons of SARS-CoV-2 antibody responses in children, adolescents, and adults
To date, most studies of the immune responses of children to SARS-CoV-2 have been cross-sectional, have focused on patients hospitalized for severe COVID-19 or MIS-C, or have assessed immunity only during acute infection. 22 Weisberg and colleagues evaluated SARS-CoV-2-specific antibody levels and neutralizing activity in non-hospitalized adults with COVID-19, hospitalized adults with COVID-19 acute respiratory distress syndrome, children with MIS-C, and SARS-CoV-2-infected children, half of whom All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted April 20, 2021. ; https://doi.org/10.1101/2021.04.17.21255663 doi: medRxiv preprint were asymptomatic. 18 They found no significant differences in antibody levels or neutralizing activity between the two groups of children; however, both groups of adults exhibited higher SARS-CoV-2-specific antibody levels and neutralizing activity than either group of children. 18 Similarly, Pierce and colleagues found that children who were hospitalized with COVID-19 or MIS-C had lower levels of neutralizing activity than hospitalized adults. 21 In our study, we found that SARS-CoV-2-specific antibody levels and virus neutralizing activity among children and adolescents with asymptomatic or mild symptomatic SARS-CoV-2 infection were generally similar to those of adults with mild symptomatic infection. Importantly, levels of SARS-CoV-2-specific IgG and serum neutralizing activity were similar at the time of acute infection but generally higher in children and adolescents than adults at 2 and 4 months after acute infection, suggesting that SARS-CoV-2-specific IgG responses may decline more slowly in children and adolescents.
In contrast, we observed higher levels of SARS-CoV-2-specific IgM and IgA antibodies with increasing age, particularly 4 months after acute infection, suggesting that young children generate less robust and durable responses for these SARS-CoV-2-specific antibody isotypes. Overall, our findings indicate that children and adolescents have a similar degree of protective immunity as adults after mild or asymptomatic SARS-CoV-2 infection. Given similarities in the response to natural infection in children and adults, it is likely that vaccination against SARS-CoV-2 will also elicit a similar degree of protection across the full spectrum of age, as has recently been reported for the Pfizer-BioNTech vaccine in children 12-15 years of age. 40 Though we cannot directly compare our results to the neutralizing antibody titers reported in vaccine trial studies, that younger age may be associated with greater neutralizing antibody responses. Together with our results, these clinical trial results demonstrated the ability of children to elicit robust SARS CoV-2-specific antibody responses. Future studies will need to directly evaluate associations between age and SARS-CoV-2 vaccine responses.
There are conflicting data regarding the impact of illness severity on immune responses to SARS-CoV-2 infection. Lau and colleagues found that the neutralizing activity of sera from 195 adults and children with All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Our study has several limitations. First, we focused on pediatric and adult populations with asymptomatic and mild infections, but did not include participants with severe COVID-19 or children with MIS-C; thus, our findings are generalizable only to individuals with asymptomatic or mild symptomatic SARS-CoV-2 infection. Second, we did not evaluate cellular immunity, which is likely required for long-term immunological memory. 3 Third, while several studies in adults have measured antibody responses out to 8 months after acute infection, we currently only have data for children and adolescents up to 4 months after infection. Finally, additional studies will be needed to assess the impact of emerging SARS-CoV-2 variants on viral-specific immune responses in children and adolescents and define if immune responses induced by prior infections are able to protect against these variants.
In conclusion, we found that children and adolescents with asymptomatic or mild symptomatic SARS-CoV-2 infection mount broad, effective, and durable antibody responses that exhibit robust viral neutralizing activity at least 4 months after acute infection. Notably, these responses were largely similar or superior to All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted April 20, 2021. ; https://doi.org/10.1101/2021.04.17.21255663 doi: medRxiv preprint those observed in adults with symptomatic SARS-CoV-2 infection who did not require hospitalization. Our findings suggest that children and adolescents develop effective humoral immune responses irrespective of illness severity that are likely to provide protection against reinfection, thereby contributing to the establishment of herd immunity.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Study Design
The Duke Biospecimens from RespirAtory Virus-Exposed Kids (BRAVE Kids) study is a prospective cohort study of children and adolescents (<21 years of age) with confirmed SARS-CoV-2 infection or close contact with an individual with confirmed SARS-CoV-2 infection. 6 This study is being conducted within the Duke University Health System (DUHS) in Raleigh-Durham, North Carolina. DUHS is a large, integrated health system consisting of three hospitals and over 100 outpatient clinics. This study was approved by the DUHS Institutional Review Board. Informed consent was obtained from all study participants or their legal guardians, with written approval obtained using an electronic consent document.

Study Procedures
The analyses presented herein were limited to participants with SARS-CoV-2 infection diagnosed by PCR between May 1, 2020 and July 31, 2020. SARS-CoV-2 was detected from nasopharyngeal or nasal swabs through PCR testing performed for clinical care or through a quantitative real-time PCR assay, as previously described. 6 We collected exposure, sociodemographic, and clinical data at the time of enrollment through review of electronic medical records and a directed caregiver questionnaire conducted by telephone. In addition, we conducted a home visit to collect whole blood from participants via venipuncture and to obtain other biospecimens. Serum was isolated from whole blood via centrifugation and frozen to -80ºC prior to analysis. We conducted follow-up study visits at home or at a research clinic site approximately 2 and 4 months after acute infection.

Adult Study Participants
We obtained previously collected sera from subjects enrolled in the Duke University Molecular and Epidemiological Study of Suspected Infection (MESSI) 42 . Participants in this prospective cohort are identified via community-enrollment, DUHS, or the Durham Veterans Affairs Health System (DVAHS) as All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Informed consent was obtained from all study participants, with written approval obtained using an electronic consent document. Mild disease was defined as any PCR-confirmed infection that did not require hospitalization.

Measurement of serum SARS-CoV-2-specific antibodies
We measured serum antibodies using a customized binding antibody multiplex assay to the following SARS-CoV-2 antigens: whole spike (S), subunit 1 (S1), subunit 2 (S2), receptor binding domain (RBD), N-terminal domain (NTD), nucleocapsid (NC), and membrane (M) proteins (all from Sinobiological except from M, from MyBiosource). SARS-CoV-2 antigens were covalently coupled to magnetic fluorescent beads (MagPlex biospheres, Luminex). Unconjugated (blank) beads were included to monitor non-specific binding. After a pilot assay to identify the optimal serum dilution, antigen-coupled beads were incubated with a 1:400 serum dilution for measurement of IgG and IgG subclasses and 1:100 serum dilution for measurement of IgA and IgM. Antibody binding to the bead-coupled antigens was then detected with phycoerythrin (PE)-conjugated mouse anti-human IgG, IgM or IgA (Southern Biotech) at 2 μg/ml, using a Bio-Plex 200 instrument (Bio-Rad Laboratories), which rendered a mean fluorescent intensity (MFI) for each sample. For measurement of SARS CoV-2-specific IgG1 and IgG3, a biotinylated mouse anti-human IgG-1 or IgG-3 followed by PE-conjugated streptavidin was used for detection. Sera from ten individuals collected before the COVID-19 pandemic (2013-2014) was used to define the assay positivity threshold for each antigen (mean MFI plus 3 standard deviations). A prescreened pooled serum sample of two unrelated SARS-CoV-2 infected donors was used as positive control in all assays to ensure reproducibility between assays and to ensure detection of antibodies against all antigens tested. Criteria for accepting results included ≤20% coefficient of variation of the two duplicates with a bead count of ≥100 for each sample. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  After a 10-minute incubation at room temperature, 100 µl of Bright-Glo luciferase reagent (Promega) was added to all wells. After 2 minutes, 110 µl of the cell lysate was transferred to a black/white plate.

SARS-CoV-2 neutralization was measured with spike-pseudotyped viruses in HEK
Luminescence was measured using a PerkinElmer Life Sciences, Model Victor3 luminometer. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Data Analysis
We described characteristics of the study population using frequencies and percentages for categorical variables and medians and interquartile ranges (IQR) for continuous variables. We used chi-squared or Fisher's exact tests for comparisons of categorical variables. To compare humoral immune responses at specific time points based on the presence of symptoms and across age categories, we used Wilcoxon ranksum tests or ANOVA based on Gaussian distribution of a variable. To account for repeated measurements from subjects, we used Wilcoxon signed-rank tests to compare humoral immune responses in paired serum samples collected from individuals across specific time points. We adjusted for multiple comparisons using the Benjamini-Hochberg procedure. 43 Study data were managed using REDCap electronic data capture tools hosted at Duke University. 44 Analyses were performed using GraphPad Prism (GraphPad Software, Inc., CA, US) and R version 4.0.3. 45 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  IQR, interquartile range All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted April 20, 2021. ;   (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Wilcoxon signed-rank tests. * p<0.05; ** p<0.01; *** p<0.005; **** p<0.0001 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.