The severity and morbidity of COVID-19 differs significantly between children and adults, and this represents one of the mysteries of SARS-CoV-2 infection. This phenomenon is quite intriguing, as for most other respiratory viruses young children account for the most vulnerable along with older adults, whilst COIVID-19 is more one sided for advanced age. The underlying mechanisms of this different impact of SARS-CoV-2 in children and adults are largely unknown. Understanding how the immune system responds to the virus in these two populations is one clue, especially how antibodies, the proteins that fight in a specific and adaptative mode against the virus, differ in quantity and what they target from the virus.
Antibodies are produced by the immune system after infection or vaccination and can be measured in the blood to determine if someone seroconverted and made a successful response against exposure. Serology surveys of blood antibodies are crucial for determining infection attack rates in the population and for assessing the response to current vaccines and impact of prevention measures. Most COVID-19 surveys study seroconversion by the measurement of antibodies targeting two structural proteins: anti-Spike or anti-Nucleocapsid antibodies, but not to other proteins of the virus. Indeed, SARS-CoV-2 virions are composed of 4 structural proteins (Spike, Nucleocapsid, Membrane and Envelope), but during virus infection, additional non-structural and accessory proteins can be expressed. In the case of vaccination, antibodies will be made against the proteins that are present in the vaccine : for instance mRNA and viral vectored vaccines encode the Spike, whilst whole virion vaccines encode structural proteins, both Spike and Nucleocapsid. Whereas in the case of infection, antibodies can potentially target all the proteins of the virus that it produces as it replicates and makes more copies of itself, therefore non-structural proteins can be used to differentiate infection in any vaccinated individuals. Furthermore, children mount a lower response to infection than adults based on current serological tests for anti-Spike and anti-Nucleocapsid antibodies therefore new serological markers could be needed for the pediatric population.
In our paper SARS-CoV-2 accessory proteins reveal distinct serological signatures in children, we characterized the antibody response for many viral proteins in children in order to find potential new serologic markers of SARS-CoV-2 infection. We asked first, is the antibody response to COVID-19 in kids similar to what is observed in adults? Second, if different, what antibody targets are specific to the pediatric populations? Indeed, these differences could impact the long-term protection acquired by children after SARS-CoV-2 infection, and could reveal some functions of the accessory proteins for the virus infection cycle in children. To address these issues we used infected adults and children’s samples collected over the first year of the pandemic in Hong Kong, before the roll-out of COVID-19 vaccines. Hong Kong had intensive contact tracing and case-finding measures that allowed us to include in our study some asymptomatic pediatric cases with RT-PCR confirmed infections, that are a rare entity in most studies.
To characterize the SARS-CoV-2 antibody landscape in children we employed an unbiased and quantitative approach using fluorescently tagged viral proteins to pull down antibodies, known as the luciferase immunoprecipitation system (LIPS). The LIPS assay is a liquid phase immune-assay allowing the quantitation of antibodies by measuring luminescence emitted by the reporter enzyme Renilla luciferase (Ruc) fused to an antigen of interest, expressed by the pREN2 vector in mammalian cells. This approach allowed us to assess the serological response to a panel of 14 SARS-CoV-2 antigens representing the structural, non-structural (representative NSP1) and accessory viral (Open reading frames (ORF)3a, 3d, 6, 7a, 7b, 8 and 10) proteins in SARS-CoV-2 infected children, covering all the ORFs of the virus.
We first determined the levels of antibodies to structural proteins of the virus and found that infected children have lower levels of Spike and Membrane antibodies than adults. Amongst the structural proteins, Nucleocapsid and Envelope were the best-performing antigens for diagnostics (based on a cut-off of the negative mean + 3x standard deviations) in the pediatrics population with 65% sensitivity and 100% specificity for Nucleocapsid and 78% sensitivity and 100% specificity for Envelope. Next, amongst antibodies directed against the accessory proteins of the virus, we found that ORF8 antibody levels were significantly elevated in pediatric COVID-19 samples compared to adults. The antibody distribution of the overall specificity of pediatrics with COVID-19 was also found different from the adult one, revealing a potential link with the pathogenicity of the disease.
We then analyzed the whole dataset by two different types of methods. First, using a cluster of points that depicts each individual sample by a combination of three different parameters taken together and that highlights the relevant relations of these parameters. This revealed that the Nucleocapsid, ORF3d, ORF8 antibodies combination can accurately allow the positive discrimination of the pediatric COVID-19 cases from the negatives, which reflects the potential interest of these markers in serology surveys. Second, using a Principal component analysis (PCA) we showed that antibodies to structural proteins do not solely drive the main responses, and that ORF3d, NSP1, ORF3a, ORF8 highly contribute to variance between children and adults. Taken together, both these analyses highlighted the importance of accessory viral proteins as targets of the COVID-19 humoral response. No differences in antibody responses between asymptomatic versus mild COVID-19 children were found for all the 14 antigens, but longitudinal samples allowed us to assess the stability of these 14 antibodies. We found that antibodies to non-structural proteins increased or were stable with time, except ORF7b antibodies that decayed longitudinally at a slow rate. Finally, cytokines, soluble molecules that play a role in shaping the adaptive immune priming, were measured in early time-point samples (<day 7) to determine the relationship between inflammation and seroconversion. Two of the cytokines tested, IL-8 and MCP-1 that are determinants in the COVID-19 cytokine storm, correlated with antibodies to accessory proteins including ORF3d and ORF8, reflecting the potential role of these targets in the host-pathogen response.
We believe in the interest of COVID-19 antibodies to non-neutralising and internal proteins particularly in the pediatric population for serology surveys but also for the distinction between SARS-CoV-2 past infection and vaccination in specific populations, especially with the recent roll-out of pediatric vaccines. The function of these antibodies needs to be further studied as they could impact viral infection through interactions during virus entry, fusion, replication and egress, but also through Fc mediated effector functions, and open new avenues on the biological relevance of the SARS-CoV-2 accessory proteins.
Figure legend: Children and adults display diverse antibody responses after SARS-CoV-2 infection. Antibodies to SARS-CoV-2 accessory proteins represent a potential interesting tool for the diagnosis of infected children and for the distinction between SARS-CoV-2 infection and vaccination, as accessory proteins are not present in vaccines. The figure was made with the Biorender software.