If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Li Ka Shing Institute of Health Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, ChinaThe Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong SAR, China
Department of Medicine and Therapeutics, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, PR ChinaLi Ka Shing Institute of Health Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China
HKU-Pasteur Research Pole, School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, ChinaCarl R. Woese Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, USA
Carl R. Woese Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, USADepartment of Biochemistry, University of Illinois at Urbana-Champaign, Urbana, USACenter for Biophysics and Computational Biology, University of Illinois at Urbana-Champaign, Urbana, USACarle Illinois College of Medicine, University of Illinois at Urbana-Champaign, Urbana, USA
HKU-Pasteur Research Pole, School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, ChinaLi Ka Shing Institute of Health Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, ChinaThe Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong SAR, China
Li Ka Shing Institute of Health Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, ChinaThe Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong SAR, China
Seroprevalences of the four human coronaviruses (HCoVs) in children showed similar trajectories.
•
Immunoglobulin G levels of the four HCoVs in mothers and their neonates were correlated.
•
The immunoglobulin G in the neonates then dropped significantly until their 1 year of age.
•
Majority of children were seropositive to the four HCoVs before 18 years old.
Abstract
Objectives
Four seasonal coronaviruses, including human coronavirus (HCoV)-229E and HCoV-OC43, HCoV-NL63, and HCoV-HKU1 cause approximately 15-30% of common colds in adults. However, the full landscape of the immune trajectory to these viruses that covers the whole childhood period is still not well understood.
Methods
We evaluated the serological responses against the four seasonal coronaviruses in 1886 children aged under 18 years by using enzyme-linked immunosorbent assay. The optical density values against each HCoV were determined from each sample. Generalized additive models were constructed to determine the relationship between age and seroprevalence throughout the whole childhood period. The specific antibody levels against the four seasonal coronaviruses were also tested from the plasma samples of 485 pairs of postpartum women and their newborn babies.
Results
The immunoglobulin (Ig) G levels of the four seasonal coronaviruses in the mother and the newborn babies were highly correlated (229E: r = 0.63; OC43: r = 0.65; NL63: r = 0.69; HKU1: r = 0.63). The seroprevalences in children showed a similar trajectory in that the levels of IgG in the neonates dropped significantly and reached the lowest level after the age of around 1 year (229E: 1.18 years; OC43: 0.97 years; NL63: 1.01 years; HKU1: 1.02 years) and then resurgence in the children who aged older than 1 year. Using the lowest level from the generalized additive models as our cutoff, the seroprevalences for HCoV-229E, HCoV-OC43, HCoV-NL63, and HCoV-HKU1 were 98.11%, 96.23%, 96.23% and 94.34% at the age of 16-18 years.
Conclusion
Mothers share HCoV-specific IgGs with their newborn babies and the level of maternal IgGs waned at around 1 year after birth. The resurgence of the HCoV-specific IgGs was found thereafter with the increase in age suggesting repeated infection occurred in children.
SARS-CoV-2 has become highly transmissible since it was discovered in 2019 and is now persistent in the human population. It is reasonable to expect that most people will be exposed to the virus for the first time during their childhood. Understanding the development of acquired immunity against the seasonal coronaviruses (human coronavirus [HCoV]-NL63 and HCoV-229E, HCoV-OC43 and HCoV-HKU1) in the young age group will thus give us a clue on the impact of SARS-CoV-2 on humans in the post–COVID-19 era. The HCoVs have been circulating in the human population for many years and are accounted for approximately 15-30% of upper respiratory tract infections [
]. Immunoglobulin (Ig) G antibodies in neonates are transferred from their mothers so as to provide a transient immune barrier against the potential infection [
]. This transferred immunity plays a protective role before the infants establish their own specific adaptive immunity to the same pathogen. So far, there is a paucity of data to describe the transition period from transferred immunity to acquired immunity for the seasonal coronavirus in children. Moreover, the accumulation of immune response to seasonal coronaviruses in children is also not yet well understood. A longitudinal study showed that adults are repeatedly infected by the seasonal HCoV every 12 months [
]. Although it was found that the induction of antibodies after each infection is short-lasting, frequent reinfections lead to persistent levels of antibodies to the four seasonal coronaviruses in most adults [
]. These pre-existing antibodies against seasonal coronaviruses were recently found to be associated with the neutralizing antibody response against SARS-CoV-2 that may mitigate disease manifestations from SARS-CoV-2 infection [
]. In this study, we determined the serological response against four seasonal coronaviruses in the plasma samples of children and modeled the seroprevalence trajectories of the four viral subtypes during childhood.
Methods
Sample collection
1886 pediatric patients aged under 18 years and without signs of influenza-like illness in non-respiratory diseases wards were recruited from January and March 2020 in our study. All plasma samples were obtained from the ethylenediaminetetraacetic acid anti-coagulated peripheral blood samples in the Guangdong Women and Children Hospital, Guangzhou, China. Peripheral blood samples were centrifuged at 3000 x g for 10 minutes at room temperature for plasma collection. All plasma samples were kept at -80°C until used. Moreover, 485 plasma samples from healthy postpartum women were collected between January and March 2020 in the same hospital, with paired plasma samples collected from their healthy newborn babies. Of note, the blood samples, which were from the pediatric patients recruited from different departments or the paired maternal and infant, were collected for routine examination in the Department of Clinical Laboratory. Neonate was defined as those who were 4 weeks old or younger and the definition of children covers neonates and male/female under 18-year-old. All study procedures were performed after informed consent. The study was approved by the Human Research Ethics Committee at the Guangdong Women and Children Hospital (Approval number: 202101231).
Enzyme-linked immunosorbent assay
The S1 subunits of spike protein (His tag) of HCoV-229E (Seattle/USA/SC1073/2016), HCoV-HKU1 (Hong Kong/isolate N5/2006), HCoV-NL63 (Florida/UF-2/2015) and the hemagglutinin-esterase (HE) protein (His Tag) of HCoV-OC43 (Seattle/USA/SC9741/2016) were purchased from Sino Biological (China). A 96-well enzyme-linked immunosorbent assay (ELISA) plate (Nunc MaxiSorp, Thermo Fisher Scientific) was first coated overnight with 100 ng per well of purified recombinant protein in phosphate-buffered saline (PBS) buffer. On the next day, plates were washed three times with PBS containing 0.1% Tween 20 first. The plates were then blocked with 100 μl of Chonblock blocking/sample dilution ELISA buffer (Chondrex Inc, Redmon, US) and incubated at room temperature for 1 hour. Each human plasma sample was diluted to 1:100 in Chonblock blocking/sample dilution ELISA buffer and then added into the ELISA plates for a 2-hour incubation at 37°C. After three times of washing with PBS containing 0.1% Tween 20, each well in the plate was further incubated with the anti-human IgG secondary antibody (1:5000, Thermo Fisher Scientific) for 1 hour at 37°C. The ELISA plates were then washed five times with PBS containing 0.1% Tween 20. Subsequently, 100 μl of Goat anti-Human secondary IgG (horseradish peroxidase) substrate (Ncm TMB One; New Cell and Molecular Biotech Co. Ltd, Suzhou, China) was added into each well. After 10 minutes of incubation, the reaction was stopped by adding 50 μl of 2 M H2SO4 solution and analyzed on a Sunrise (Tecan, Männedorf, Switzerland) absorbance microplate reader at 450 nm wavelength.
Modeling
Generalized additive model (GAM) was fitted to investigate the association between age and the ELISA results. The restricted cubic splines (smooth curve) with five knots were used to construct the model [
]. Of note, percentile places knots at five spaced percentiles of the explanatory variable, which are the 5th, 27.5th, 50th, 72.5th, and 95th percentile. R version 4.0.4 was used for the analysis.
Statistical analysis
Significance between the two groups was determined by the Mann-Whitney test, with a P-value >0.05 being considered statistically significant. Correlation between plasma samples were evaluated by using Pearson's correlation coefficients.
Results
We tested the seroprevalence to the four seasonal coronaviruses by the ELISA using the plasma samples collected from 1886 children (Female: 43.9%) with age ranging from 0 (Neonates) to 18 years old in Guangzhou, China between January and March 2020. Among our cohort, 259 participants were >6 months old, 161 were between >6 months to <12 months, 278 were >1 to <3 years old, 603 were >3 to <7 years old, 466 were >7 to <12 years old, 66 were >12 to <16 years old, 53 were >16 to <18 years old (Table 1). The S1 domains of the spike protein were used for measuring the serological response to 229E, NL63, and HKU1. Although the homology of the S1 protein between HKU1 and OC43 is low, they both bind to receptors that carry 9-O-acetylated sialic acid. HE but not S1 of OC43 was thus used for the ELISA assay to further reduce the cross-reactive signal. HE acts as a receptor-destroying enzyme that facilitates the release of viral progeny from infected cells [
The IgG levels to the four seasonal coronaviruses were determined from each plasma sample (Supplementary Figures S1 and S2). The association between the IgG level and the age in each seasonal coronavirus was constructed by GAM (Figure 1) [
]. The restricted cubic splines (smooth curve) with five knots were used to visualize the association. We found that the seroprevalences of the four seasonal coronaviruses showed a similar trajectory from the GAMs. Compared with the entire childhood period, the levels of IgG in the neonates dropped significantly and reached the lowest level of the GAM after the age of 1 year (1.18 years: HCoV-229E; 0.97 years: HCoV-OC43; 1.01 years: HCoV-NL63; 1.02 years: HCoV-HKU1) (P <0.001). The levels of IgG then increased and accumulated when the children became older. The IgG levels against HCoV-OC43, HCoV-NL63, and HCoV-HKU1 increased to the comparable levels in children at the age of 8, 9, and 6 years respectively. However, it was intriguing to find that the IgG to the HCoV-229E increased slowly compared with other seasonal coronaviruses and it reached the comparable level in children at the age of 16 years. The serological results of each coronavirus were further stratified into two sex groups (male/female) and were further compared (Figure 2). Importantly, we found that the IgG waning of all four seasonal coronaviruses in male children were much faster than that in female. The time required for dropping the IgG of each coronavirus to their lowest level of the GAMs in male children were 1.95 (HCoV-229E: 0.75[M] vs 1.46[F]), 1.84 (HCoV-OC43: 0.63[M] vs 1.16[F]), 1.69 (HCoV-NL63: 0.68[M] vs 1.15[F]), 1.71 (HCoV-HKU1: 0.67[M] vs 1.15[F]) folds faster than that of the female children (P <0.001).
Figure 1Seroprevalence trajectory of the four seasonal coronaviruses in children. The plasma samples were collected from 1886 children who aged from 0 (neonates) to 18 years old. Each sample was tested by enzyme-linked immunosorbent assay against either S1 (HCoV-229E, HCoV-NL63 or HCoV-HKU1) or hemagglutinin-esterase (HCoV-OC43) protein. GAMs was used to model the association between the serological data and the age. The black lines showed the fitted values and gray areas showed the 95% confidence intervals. Each sample was tested in duplicate, and the results were represented by the mean of the two values. The solid line represents the cutoff of the negative control (PBS). The dashed lines represent the cutoff of the lowest point in GAMs and the solid lines represent the background (PBS) of the assay.
Figure 2Antibody levels against the four seasonal coronaviruses in different genders. The 1886 plasma samples which were collected from children were further stratified into female (n = 828 samples) and male (n = 1058 samples) for analysis. Each sample was tested by enzyme-linked immunosorbent assay against either S1 (a: HCoV-229E, c: HCoV-NL63 or d: HCoV-HKU1) or hemagglutinin-esterase (b: HCoV-OC43) protein. GAMs was used to model the association between the serological data and the age. The black lines showed the fitted values and gray areas showed the 95% confidence intervals. Each sample was tested in duplicate, and the results were represented by the mean of the two values. The dashed lines represent the cutoff of the lowest point in GAMs and the solid lines represent the background (phosphate-buffered saline) of the assay.
The relatively high levels of IgG antibody against four seasonal coronaviruses in the children under 1-year-old suggested a vertical transfer of the maternal immune response. It has been recently shown that the passive immunity against SARS-CoV-2 of children was contributed by their mothers [
]. We collected plasma samples from 485 pairs of postpartum women and their newborn babies for testing the levels of their IgG to the four seasonal coronaviruses using similar serological assays. We found that the maternal IgG level was linearly associated with their neonatal IgG levels in each seasonal coronavirus: HCoV-229E (r = 0.63, 95% CI: 0.57-0.68, P <0.0001), HCoV-OC43 (r = 0.65, 95% CI: 0.60-0.70, P <0.0001), HCoV-NL63 (r = 0.69, 95% CI: 0.64-0.74, P <0.0001), HCoV-HKU1 (r = 0.63, 95% CI: 0.58-0.69, P <0.0001) (Figure 3). Interestingly, when we compared the antibody levels between the mothers and their newborn babies using Wilcoxon pairwise test, the newborn babies showed a lower level of 229E antibody compared with their corresponding mothers (Supplementary Figure S3). However, no significant differences were found in HCoV-OC43, HcoV-NL63, or HcoV-HKU1. While comparing with the previous report that maternally derived antibodies against SARS-CoV-2 could persist up to 6 months of age in their infant [
], our results indicated that the passive transferred immunity against the seasonal coronaviruses in children can maintain longer time (1.25 years: HcoV-229E; 1 years: HcoV-OC43; 1.08 years: HcoV-NL63; 1.08 years: HcoV-HKU1) (Figure 1).
Figure 3Correlation between the maternal and neonatal IgG levels of the four seasonal coronaviruses. 485 paired of maternal and neonatal plasma samples were collected and tested by enzyme-linked immunosorbent assay. Antibody levels against (a) HCoV-229E-S1, (b) HCoV-NL63-S1, (c) HCoV-OC43-HE, and (d) HCoV-HKU1-S1 were determined and the correlations between the paired samples in the four seasonal coronavirus groups were shown. The black lines showed the fitted values and gray areas showed the 95% confidence intervals. The r represented the correlation coefficient.
Prevalence of the seasonal coronaviruses in children is determined either by detecting the specific nucleic acids from the respiratory specimen or through serology test. However, it is difficult to define and collect true negative reference samples because the seasonal coronaviruses are highly circulating in children. Previous studies adopted an approach in which the cutoffs were determined from a small subset of reference samples who the children were between 1-2 years old, and the tested samples were defined as positive if the results were above the mean of the references [
]. Here, we estimated the prevalence of the seasonal coronaviruses by using the lowest level in the GAMs as our negative reference (Table 1). As shown in the figure, the lowest point in GAMs for four seasonal coronaviruses are optical density 0.13 (HCoV-229E), 0.18 (HCoV-OC43), 0.27 (HCoV-NL63) and 0.17 (HCoV-HKU1) respectively. We assumed that children with IgG level above this point indicate infection of the corresponding seasonal coronaviruses and thus defined it as seropositive. 91.12%, 82.24%, 79.92%, and 84.17% of seropositivity to HCoV-229E, HCoV-OC43, HCoV-NL63 and HCoV-HKU1 respectively were found in those under 6 months of age. In infants with the age between 6 months and 12 months, the seropositive rates dropped to 44.72% (HCoV-229E), 43.48% (HCoV-OC43), 45.96% (HCoV-NL63), and 45.96% (HCoV-HKU1) (P <0.001). The seropositivity of each seasonal coronavirus increased with age and was over 64.51% of prevalence in the children at their preschool age (3-6 years). The seroprevalences for HCoV-229E, HCoV-OC43, HCoV-NL63, and HCoV-HKU1 kept increasing and were 98.11%, 96.23%, 96.23%, and 94.34%, respectively, at the age of 16-18 years.
Discussion
Our study described the transition from passive to acquired immunity against seasonal coronaviruses in children. The established approach here provides a view to identify the waning period of immunity against coronavirus after birth, which will be useful to apply on SARS-CoV-2. The best timing to receive COVID-19 vaccine is still being debated. Though US CDC suggested that COVID-19 vaccination is recommended for children aged 6 months or older, it is mainly based on the safety concerns rather than aiming for better protection. Defining the waning period in SARS-CoV-2 using our approach will provide scientific evidence to determine the vaccination window for children in post-COVID era.
Full spike protein was used in the previous study to detect the antibody against the four seasonal coronaviruses in 218 children [
]. It is known that there is 63-98% of sequence similarity at the S2 among the seven HCoVs. Binding from the cross-reactive antibodies at the S2 domain may lead to an overestimation of the seroprevalence. The homology of the S1 amino acid sequences among the four HCoVs is between 40-68% only. We expected that using S1 as the antigens for detection should represent the specific serological response. However, as HKU1 and OC43 both bind to receptors carrying 9-O-acetylated sialic acid, we used HE to replace the S1 in the assay for measuring the antibody to OC43. The homology of S1 and HE between the two viruses is 68% and 66% respectively. Interestingly, despite using S1 or HE domain to evaluate the specific antibody response for the seasonal coronaviruses, the trajectory of the seroprevalence of the four subtypes were still very similar.
Although seasonal coronaviruses are responsible for only 4-6% of acute respiratory tract infections in children [
Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method.
], our serology study showed that the infections are indeed very common in this age group. The low detection rate may be because most of the infections were either asymptomatic or very mild. We observed that the levels of IgG to the seasonal coronaviruses increased with the age of children. Among the four subtypes, the level of IgG to HCoV-229E rises comparatively slower than the HCoV-OC43, HCoV-NL63, and HCoV-HKU1. It was reported that HCoV-229E was less frequently detected in humans including children [
] and the previous human-challenge study showed that individuals infected with HCoV-229E were resistant to reinfection with the same strain but partially susceptible to an antigenicity different strain [
]. In this study, we first reported that high levels of IgG to the four seasonal coronaviruses in neonates are linearly correlated with their corresponding maternal IgG levels and this suggested that the vertical transmission of coronavirus-specific antibodies from mothers to neonates. It drops within the first year after the babies are born and male children showed a faster decline in the antibody production compared with female children. However, further investigation is needed to determine whether male children may have a higher risk for serious infection if they catch seasonal coronavirus during their early life. On the other hand, the previous results showed that though maternal antibody provide protection for infant, maternal antibody may also suppress the B cell response by epitope masking or inhibition of infant B cell activation by Fcγ-receptor mediated signaling [
]. Therefore, future studies should explore the positive and negative effect of the maternal antibodies, which will provide important insights into the antibody response with pre-existing maternal antibodies.
There were some limitations in our study. Firstly, the trajectories were illustrated using cross-sessional samples from population age groups, not in the longitudinal cohort. Secondly, the seroprevalences from our cohort were determined by ELISA only. The neutralizing effect to the seasonal coronaviruses was not evaluated. Thirdly, although the children were recruited from the nonrespiratory ward or routine body check center, we did not collect their clinical background for analysis in this study.
In conclusion, we described that IgG antibody against four seasonal coronaviruses could be transferred from mother to their infant by conducting a large-scale cohort. Importantly, we reported that this transferred immunity waned for one year after birth and children could acquire immunity against four seasonal coronaviruses with the increase in age. Overall, these results provide a comprehensive analysis of the antibody dynamic in the early life of the children.
Declaration of competing interest
The authors have no competing interests to declare.
Funding
This work was supported by The Medical Scientific Research Foundation of Guangdong Province of China (A2022196), Calmette and Yersin scholarship from the Pasteur International Network Association (H.L.), Guangdong-Hong Kong-Macau Joint Laboratory of Respiratory Infectious Disease (20191205) (C.K.P.M.) and Emergency Key Program of Guangzhou Laboratory (Grant No. EKPG22-30-6) (C.K.P.M.), RGC's Collaborative Research Fund (C6036-21GF) (C.K.P.M), and visiting scientist scheme from Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore (C.K.P.M.).
Author contributions
H.L., N.C.W. and C.K.P.M. conceived the research idea and designed the study. Y.L., Y.S. C.L., Y.D., B.L. and X.M coordinated and carried out cohort recruitment. H.L., S.Z., K.K., C.K.P.M. and H.M.T., analyzed the data. Y.L., H.L., C.C., W.L, Q.W.T, R.T.Y.S., Y.L, Z.D., J.Z., D.Z. and J.F. performed the experiments. H.L., R.B., H.M.T., and C.K.P.M. wrote the manuscript.
Figure S1. Serological response of the four seasonal coronaviruses in children. The plasma samples were collected from 1886 children who aged from 0 (neonates) to 18 years old. Each sample was tested by enzyme-linked immunosorbent assay against either S1 (HCoV-229E, HCoV-NL63 or HCoV-HKU1) or hemagglutinin-esterase (HCoV-OC43) protein. Each sample was tested in duplicate, and the results were represented by the mean of the two values. HCoV, human coronavirus; OD, optical density.
Figure S2. Serological response of the four seasonal coronaviruses in different genders of children. The 1886 plasma samples which were collected from children were further stratified into female (n = 828 samples) and male (n = 1058 samples) for analysis. Each sample was tested by enzyme-linked immunosorbent assay against either S1 ([a]: HCoV-229E, [c]: HCoV-NL63 or [d]: HCoV-HKU1) or hemagglutinin-esterase ([b]: HCoV-OC43) protein. Each sample was tested in duplicate, and the results were represented by the mean of the two values. HCoV, human coronavirus; OD, optical density.
Figure S3. Correlation between the maternal and neonatal IgG levels of the four seasonal coronaviruses. 485 paired of maternal and neonatal plasma samples were collected and tested by enzyme-linked immunosorbent assay. Antibody levels against (a) HCoV-229E-S1, (b) HCoV-NL63-S1, (c) HCoV-OC43-HE, and (d) HCoV-HKU1-S1 were determined and the correlations between the paired samples in the four seasonal coronavirus groups were shown. Each pair of samples is linked with dashed line. Wilcoxon matched pairs signed rank test is used to compares the median of the maternal and neonatal antibody. P-values were calculated using a two-tailed t-test (**** P <0.001, ns [not significant]). HCoV, human coronavirus; OD, optical density.
References
Desforges M
Le Coupanec A
Dubeau P
Bourgouin A
Lajoie L
Dubé M
et al.
Human coronaviruses and other respiratory viruses: underestimated opportunistic pathogens of the central nervous system?.
Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method.
00636-1&id=gr1.jpg){kind=link}
00636-1&id=gr2.jpg){kind=link}
00636-1&id=gr3.jpg){kind=link}