Samples used in this study derive from three different studies and were all approved by the Medical-Ethical Review Committee of the University Medical Center Utrecht (SARS-CoV-2 transmission study: NL13529.041.06, NVI-255 study: NL29241.000.09, and IMMFACT study: NL46795.094.13). Written informed consent was obtained from all participants and all trial-related activities were conducted according to Good Clinical Practice, including the provisions of the Declaration of Helsinki.

Participants of the previously described SARS-CoV-2 transmission study (14) included persons 18 years and older with a positive SARS-CoV-2 PCR test who had at least one child in their household. All household contacts (adults and children) were enrolled in the study, except for contacts below one year of age. This study was carried out in the Netherlands early in the COVID-19 pandemic in the period of March-April 2020. During home visits blood samples were collected for serum and PBMC isolation. In order to study the early immune responses, index case participants were selected with a first blood collection home visit (T1), 3 to 7 days, after testing PCR+ during routine COVID19 surveillance ( Table 1 ). At time the index case participant was sampled, their household members also provided samples. Household members who tested PCR+ in either nose, throat, saliva, or feces were included in this study. The two follow-up time points took place 2-3 weeks (T2), 4-6 weeks (T3) after T1. For children, no PBMCs were collected at T3. PBMCS from healthy adults (NVI-255 study) (15) and healthy children (IMMFACT study) (16), both collected several years before the COVID-19 pandemic, were also included as SARS-CoV-2 unexposed controls.

As previously described (14) severity of COVID-19 was classified into mild, moderate and severe, which were indicated by severity indexes 1, 2 and 3, respectively. Symptoms reported by participants included sore throat, cough, respiratory difficulties, fever, chills, headache, muscle pain, joint ache, diarrhea, nausea, vomiting, loss of appetite or fatigue. Participants with mild disease presented any of the clinical symptoms described above. Those with moderate COVID-19 reported having pneumonia, including dyspnea, and those with severe disease reported dyspnea and have consulted a healthcare professional or have been admitted to the hospital. The overall disease severity for each participant was calculated, by adding the severity indexes of each participant during the first three home visits which resulted in: No Symptoms (0), Mild (1-2) Moderate (3-4) and Severe disease (>5).

A SARS-CoV-2 heat-inactivated viral suspension (iSARS-CoV-2) was used in the present study. The SARS-CoV-2 isolate (hCoV-19/Netherlands/ZuidHolland_10004/2020) was obtained from a Dutch patient (Centre for Infectious Disease Control, RIVM). The virus was grown on VERO-E6 cells as previously described (17). Heat-inactivation was performed by incubating the virus at 60°C for 2 hours after which the samples were stored at −80°C.

PBMCS were isolated from whole blood using Ficoll–Paque density centrifugation and cryopreserved following standardized protocols. Thawed PBMCS (2x10⁵ cells/well) were stimulated for 24 h in the presence of inactivated (i)SARS-CoV-2 (MOI=3) or different PRR-ligands: 10⁷ cells/mL heat killed Listeria monocytogenes (HKLM, TLR2 ligand), 2.5 mg/mL R848 (TLR7/8 ligand) or 25 ng/mL Lipopolysaccharide (TLR4 ligand) (Invivogen).

After 24 h, supernatants were collected and analyzed for cytokine production by Legendplex, measuring IL-1β, IL-6, TNFα, CXCL10, IFNλ1, IL-8, IFNα2, GM-CSF, IFNβ, IL-10, IFNγ, IL-12-p70, IFNλ2/3 (LEGENDplex™ Human anti-Virus response Panel (13-plex), BioLegend) following manufacturer’s instructions. IL-12-p70 levels were below lower limit of detection and were therefore excluded from further analysis.

After collection of the supernatants, the stimulated cells were analyzed for different monocyte subset markers by flow cytometry. For flow cytometry, cells were washed with 0.5%BSA/PBS and incubated 5 min with Trustain (10x in 0.5%BSA/PBS, Biolegend), thereafter a mixture of antibodies (all from Biolegend, unless stated otherwise) was added for 20 min: CD14-FITC (clone HCD14), CD16-APC (clone 3G8), CD3-AlexaFluor700 (clone HIT3a), CD4-AlexaFluor700 (clone RPA-T4), CD19-AlexaFluor700 (clone HIB19), CD56-AlexaFluor700 (clone HCD56), CD15-AlexaFluor700 (clone HI98), NKp56-AlexaFluor700 clone 9E2), Fixable Viability Stain780 (BD Biosciences), HLA-DR-BrilliantViolet510 (clone L243), CCR5-BrilliantViolet421 (clone J418F1), CCR2-BrilliantViolet605 (clone 036C2), CX3CR1-PE/Dazzle594 (clone 2A9-1), CD11c-PE (clone 3.9) and CD11b-BrilliantUV737 (clone M1/70, BD Biosciences). After washing, the cells were acquired on a FACS Symphony (BD) and total monocytes were gated using FlowJo Software (V10.6.2). To identify the specific monocyte populations, preprocessing and clustering of the FACS datasets was performed in R v4.2.0 (18) and RStudio (v2022) with packages flowCore (19) and CATALYST (20, 21) ( Supplementary Figure 1 ).

SARS-CoV-2 multiplex immunoassays (MIA) were carried out to determine specific serum antibody IgG levels towards SARS-CoV-2 Spike S1 (Sino Biological, 40591-V08H) as described previously (22). In short, the serum was incubated with antigen-coupled beads for 45 minutes in the dark, followed by a 30-minute incubation with Goat-anti-human IgG. Samples were incubated in SM01 (Surmodics) and washing steps after each incubation were carried out with PBS. Antibody binding to antigen-coupled beads was determined with FM3D (Luminex) and antibody levels were expressed as binding antibody units (BAU)/ml for S1. The threshold for seropositivity was set at 10.1 BAU/mL for Spike S1, as previously standardized for the Wuhan (vaccine) strain against the NIBSC/WHO COVID-19 reference serum 20/136 (23).

Statistical analysis was performed using non-parametric Kruskal-Wallis test for unpaired analysis or Friedmans test for paired samples over time, both followed by Dunn’s multiple comparison test, or multiple Mann-Whitney test with Bonferroni-Dunn multiple comparison test, as indicated, all in Graphpad Prism version 9.5.1. Differences between groups were considered statistically significant at adjusted p-value <0.05. Spearman’s correlation analysis was used between cytokine levels and the proportion of monocyte subsets, or disease severity. A correlation was considered significant at adjusted p-value <0.05. Kruskal-Wallis Chi-squared test was used to analyze proportion of monocyte subsets with disease severity, p-value <0.05 was considered significant.

PBCMs taken from children and adults within the first week after testing PCR+ for SARS-CoV-2 early in the pandemic were stimulated in vitro with the iSARS-CoV-2 preparation and the cytokine levels were measured 24 hours after incubation. PBMCs from the PCR+ children secreted significantly higher levels of IFNβ and CXCL10 compared to the PBMCs from PCR+ adults, after iSARS-CoV-2 stimulation ( Figure 1 ). Compared to the corresponding unexposed healthy controls, stimulated PBMCs from the PCR+ children produced significant higher levels of IL-10 and GM-CSF ( Figure 1 ). In contrast, PBMCs from the PCR+ adults secreted lower levels of IL-8 compared those from adult healthy controls ( Figure 1 ). No significant differences were observed between children and adults in the healthy control groups. These data indicate different innate immune responses between PCR+ children and adults as well as between age-matched PCR+ and healthy individuals.

The development of specific antibody response after primary SARS-CoV-2 infection was measured at the different time points <7 days (T1), 2-3 weeks (T2) and 4-6 weeks (T3) after testing PCR+. For 22/24 children there were T2 and/or T3 serum sample available to determine SARS-CoV-2 Spike-specific IgG antibody levels. Four out of the 22 children did not show Spike-specific IgG levels at T2 and/or T3 ( Figure 2A ), whereas all the adults did develop Spike-specific IgG antibodies at T3 ( Figure 2B ). The Spike-specific antibody levels did not significantly differ between children and adults at any given timepoint. Next, we assessed whether the innate responsiveness of PBMCs at T1 stimulated with iSARS-CoV-2 could predict the development of the spike-specific IgG titers at T2 and T3. However, we did not find any significant association between the levels of cytokines and chemokines secreted by PBMCs after iSARS-CoV-2 stimulation and the Spike-specific IgG levels at T2 and T3.

Since the levels of some of the cytokine and chemokines secreted by PBMCs from PCR+ adults and children were highly variable after iSARS-CoV-2 stimulation, we questioned whether this could be related to the clinical status of the participants. Indeed, correlation analysis indicated a relation between the levels of IL-8 and the clinical status in adults (r=-0,527 p=0.0016). In children the levels of GM-CSF and IL-10 correlated with symptom severity (r=0.466, p=0.0036; r=0.621, p<0.001, respectively). Compared to healthy adults, the levels of IL-8 were significantly lower for adults with moderate symptoms ( Figure 3A ). No significant correlation was found between the levels of IFN-β or CXCL10 and symptom severity for adults (data not shown). For PBMCs from children with mild symptoms significantly higher levels of GM-CSF and IL-10 were found to be secreted after iSARS-CoV-2 stimulation, compared to their healthy counterparts ( Figure 3B ). Furthermore, the levels of IL-10 were also significantly higher for children with moderate symptoms ( Figure 3B ). These findings indicate distinct associations between cytokine responses and symptom severity in children and adults.

Monocytes are the most abundant PRR expressing cells in the PBMCs compartment and may determine the outcome of the PBMCs interaction with iSARS-COV-2 (24). Phenotypic analysis indicates that upon in vitro iSARS-CoV-2 stimulation, the classical monocytes in PBMCs of PCR+ adults were the most abundant subset, which was significantly higher compared to PBMCs from PCR+ children ( Figure 4 ). In contrast, iSARS-CoV-2 stimulated PBMCs from PCR+ children presented significantly larger proportion of intermediate and non-classical monocytes compared to those in adults ( Figure 4 ). In adults the proportion of non-classical monocytes was significantly smaller in PCR+ compared to healthy individuals ( Figure 4 ). Over time the proportion of monocyte subsets in PCR+ adults changed to levels similar to those of healthy controls with smaller proportion of classical, and larger proportion of non-classical monocytes at T2 and T3 ( Supplementary Figure 2 ).

To assess whether changes in monocyte subset proportions relate to cytokine and chemokine production after iSARS-CoV-2 stimulation early after infection (T1) we performed correlation analysis. The proportion of classical and intermediate monocytes in PBMCs from PCR+ adults showed a significant correlation with IL-1β, GM-CSF and IL-10 secretion ( Table 2 ). The non-classical monocyte subset correlated with interferon (IFNλ1, IFNα2, and IFNβ), CXCL10, and IL-8 secretion. Proportion of non-classical monocytes in PBMCs from PCR+ children also showed a significant correlation with interferons (IFNα2 and IFNβ), CXCL10, and IL-8, albeit less strong as those in adults ( Table 2 ). Comparing the proportion of non-classical monocyte between healthy control and PCR+ individuals we observed significantly lower values in adults, but not in children, with moderate symptoms (p-adj = 0.0088). The correlation between non-classical monocytes, symptom severity, and the production of interferons, CXCL10 and IL-8 after iSARS-CoV-2 stimulation indicates important differences in anti-viral innate immune response between PCR+ children and adults early after infection.

Several studies have highlighted the importance of PRR activation including that of TLR2, TLR4 and TLR7/8 in the innate immune response to SARS-CoV-2 (25–27). Therefore, we next measured the responsiveness of PBMCS from the PCR+ children and adults stimulated with pathogen-associated molecular patterns activating these PRR. Results indicate that PBMCS from the PCR+ children secrete significantly higher levels of IL-1β, IL-6, IFNλ1, IFNβ, TNFα, CXCL10 and GM-CSF after stimulation with TLR7/8 ligand R848 compared to adults ( Figure 5 ). In addition, we observed that activation of TLR2 by HKLM resulted in significantly higher levels of IL-1β, IFNλ1, IFNγ, TNFα and GM-CSF for the PBMCS from the PCR+ children ( Figure 5 ). Activation of TLR4 with LPS did not result in significant differences although a trend was observed for increased levels of several of these cytokines for the PCR+ children ( Figure 5 ). Interestingly, when PBMCS from the PCR+ individuals were cultured overnight in medium only (Mock) the levels of secreted IL-6, IL-8, IL-10, and specially CXCL10 were already significantly higher for PCR+ children compared to PCR+ adults ( Figure 5 ). For mock-stimulated PBMCs from healthy controls, higher levels of CXCL10 and IFNβ were also observed in children, but no increased levels of IL-6, IL-10 and IL-8 were found ( Supplementary Figure 3 ). Also worth noticing is that while in healthy controls the levels of IL-8 after activation with the different TLR ligands were significantly higher for adults compared to children, this difference was reversed for the PCR+ adults. Altogether, these findings indicate that TLR activation, particularly TLR7/8, shows pronounced differences in the cytokine secretion profile between PCR+ children versus adults and that mock stimulated PBMCs already show some differences in the response between these two age groups.

To determine whether the cytokine secretion profile observed within a few days after testing PCR positive (T1) changed in time, we tested the responsiveness to iSARS-CoV-2 by PBMCs from the same adult individuals 2-3 weeks (T2) and 4-6 weeks (T3) later. The levels of CXCL10 after iSARS-CoV-2 stimulation of adult PBMCs were significantly higher at T2 and T3 compared to T1, and the levels of IFNγ and IFNα2 were significantly higher at T3 compared to T1 ( Figure 6A ). Notably, the levels of IL-8 remained lower compared to those of healthy controls up to six weeks post infection ( Figures 1 , 6A ). Furthermore, adult PBMCs samples also produced higher levels of IFNγ at T2 and T3, compared to T1, after TLR2 stimulation ( Figure 6B ), TLR4 ( Figure 6C ), and TLR7/8 ( Figure 6D ) stimulation. In addition, TLR2 and TLR4, but not TLR7/8, stimulation also resulted in higher production of IL-1β in T2 and T3 samples compared to T1 samples ( Figures 6B–D ). The response to iSARS-CoV-2 or TLR stimulating agents by PBMCs from PCR+ children collected at T2 did not differ from T1 (data not shown). Together, these data indicate that at 2 and 6 weeks post infection adult PBMCs increase their responsiveness to iSARS-CoV-2 or PRR stimuli, whereas children retained stable response at T2.