All of the patients recovered and were discharged from the hospital. The median age of the enrolled patients was 9 years and nine (47%) of them were male. As shown in Table 1, the initial clinical signs were cough (47%), fever (37%) and pneumonia (58%). Laboratory tests of the patients were performed within 2 days after admission. The counts of white blood cells (WBC), neutrophils and lymphocytes were relatively normal, with only one patient having lymphopenia and three having neutropenia. As for inflammatory indicators, two patients (11%) had increased erythrocyte sedimentation rate (ESR) and three (16%) had increased procalcitonin (PCT), C-reactive protein (CRP) and interleukin (IL)-6 respectively. Notably, there was only one patient who had all higher ESR (34 mm/h), PCT (0.07 ng/dL), CRP (15 mg/L) and IL-6 (24.77 pg/mL). Total globulin was higher in 8 patients (42%) and 6 of 13 patients showed higher total IgE. Besides, 8 of 12 patients had increased D-dimer levels.
Value Local normal range Age-Median(range) 9 year (7 m–13 year) – Sex-Male/Female 9/10 – Clinical features-No. (%) Fever 7 (37%) – Cough 9 (47%) – Sore throat 5 (26%) – Rhinorrhea 7 (37%) – Pneumonia (by X-ray or CT) 11 (58%) – Laboratory findings-Median(range) WBCs × 109/L 5.5 (3.2–16) – 29 d–3 y 7.95 (6.4–8.7) 6–12 > 3 y 4.8 (3.2–16) 4–10 Hemoglobin (g/dL) 134 (4.7–168) 110–160 Platelets × 109/L 265 (128–404) 100–400 Neutrophils × 109/L 2.9 (1.15–11.2) – 29 d–3 y 2.22 (1.67–2.90) 2.4–4.8 > 3 y 3.51 (1.15–11.2) 2.0–7.0 Neutrophils (%) 47.1 (25.3–75) – 29 d–3 y 28.2 (25.2–43.4) 30–40 > 3 y 50.2 (27.3–75) 50–70 Lymphocytes × 109/L 2.34 (1.17–5.9) – 29 d–3 y 5.05 (4.5–5.9) 4.0–8.4 > 3 y 2 (1.17–4) 1.2–4.0 Lymphocytes (%) 42 (17.5–69.8) – 29 d–3 y 63.25 (48.5–69.8) 50–70 > 3 y 41 (17.5–66.9) 30–40 Procalcitonin (ng/dL) 0.04 (0.02–0.12) < 0.05 C-reactive protein (mg/L) 5.11 (0.5–35) < 8 IL-6 (pg/mL) 4.465 (0.91–24.77) < 7 Erythrocyte sedimentation rate (mm/h) 14 (4–39) 0–26 Alanine transaminase (U/L) 16.5 (6.9–100) 7–50 Aspartate aminotransferase (U/L) 28.9 (15.4–142) 13–40 Urea (mmol/L) 3.6 (1.97–5.7) 2.8–7.6 Creatine kinase-MB (U/L) 21.4 (12.3–51.9) < 25 Lactate dehydrogenase (U/L) 220 (156–383) 110–430 Myohemoglobin (μg/L) 9 (6–11) 10–46 Total protein (g/L) 74.3 (59–85.6) 60–80 Albumin (g/L) 44.9 (38.8–51.1) 40–55 Globulin (g/L) 28.6 (17.9–35.1) 20–30 IgG (g/L) 11.8 (9.9–14.5) 3.5–14.26 IgA (g/L) 1.4 (0.44–2.56) 0.06–2.5 IgM (g/L) 1.28 (0.76–2.35) 0.36–2.16 Total-IgE (IU/ml) 73.12 (7.91–1134.01) < 100 D-dimer (mg/L) 0.58 (0.33–1.02) 0–0.5 COVID-19, Coronavirus Disease 2019; CT, computed tomography
Table 1. Clinical features and laboratory findings of children with COVID-19.
Considering that most of children recovered within 2 weeks after disease onset, we divided the disease course into acute phase (0–3 days after onset), middle phase (4–10 days after onset) and convalescent phase (11–27 days after onset). As shown in Fig. 1, the IFN-γ-induced protein 10 (IP10) was increased markedly at the onset of the disease and mostly went back to normal as early as in middle phase (Fig. 1A). The other three chemokines, monocyte chemoattractant protein 1 (MCP1), IL-7 and IL-8, didn't show significant increase in patients (Fig. 1B–1D). A peculiar chemokine, IL-16, which was found to be related with virus infection (Sampson et al. 2017), was strikingly increased in patients during acute phase and didn't show a downward trend even in convalescent phase (Fig. 1E).
Figure 1. Serum cytokine/chemokine levels of children with COVID-19. Serum cytokine/chemokine expression of children with COVID-19 during acute phase (n = 19), middle phase (n = 10) and convalescent phase (n = 14) were detected by CBA assay or ELISA. The horizontal line represented medians with interquartile range. Results were analyzed by Mann–Whitney U test. P < 0.05 was considered of significant difference.
Pro-inflammatory cytokines including TNF, IFN-γ, IL-6 and IL-1β didn't change prominently in all three phases (Fig. 1F–1I). But IL-10, which is an anti-inflammatory cytokine, was obviously elevated in acute phase and remained higher than the healthy controls in convalescent phase (Fig. 1J). No significant changes were observed in another Th2 cytokine, IL-4, and the T cell growth factor, IL-2 (Fig. 1K, 1L). Other cytokines (IL-3, IL-5 and IL-13) in sera were lower than the detection limit (data not shown).
Although no remarkable changes were observed in the surface markers of T cells (CD3, CD4, CD8) in acute phase (Fig. 2A, 2B), we found a statistically significant increase in T helper (Th) 2, as well as mild elevations in Th1 and Th17 (no statistical meaning), suggesting a Th2-dominant immune response shortly after disease onset (Fig. 2C). Additionally, regulatory T cell (Treg), which confers immunosuppressive role, was apparently suppressed in patients during acute phase (Fig. 2C). In convalescent phase, we observed a diminished percentage of CD8+ T cells and CD3+ T cells (Fig. 2E). Moreover, after the major elevation of Th2 in acute phase, the Th subsets tended to be on a downward trend with Th1 being reduced mostly (Fig. 2F). Besides, Both Th2 and Treg returned to normal in convalescent phase (Fig. 2F). Unexpectedly, the cytotoxic lymphocytes (CTL) response didn't seem to be activated since the TNF-α-producing and IFN-γ-producing CD8+ T cells in acute (Fig. 2D) and convalescent phases (Fig. 2G) were both comparable to the healthy controls.
Figure 2. T cell immune response of children with COVID-19. A The gating strategy for CD4+ T cell subsets and CD8+ T cell cytokine production were showed. B–D The surface markers for T cells (CD3, CD4, CD8) (B), the Th1(CD4+ IFN-γ+)/Th2(CD4+ IL-4+)/Th17(CD4+ IL-17+)/Treg (CD4 + CD25+ CD127−) subsets (C), the IFN-γ-producing and TNF-α-producing CD8+ T cells (D) of the patients in acute phase were showed (n = 9). E–G The surface markers for T cells (E), the Th1/Th2/Th17/Treg subsets (F), the cytokine production in CD8+ T cells (G) of the patients in convalescent phase were displayed (n = 14). Data were presented as mean ± SEM. Student's t test was performed and P < 0.05 was considered of significant difference.
Additionally, we analyzed the memory T cell subsets including naïve T cells (NT), central memory T cells (TCM), effector memory T cells (TEM) and CD45RA+ effector memory T cells (TEMRA) during convalescent phase (Fig. 3A). Consequently, no significant changes were found in all the memory T cell subsets in CD4+ T (Fig. 3B) and CD8+ T cells (Fig. 3C), suggesting that no perturbation was developed in the homeostasis of memory T cells.
Figure 3. Memory T cell subsets in children with COVID-19 during convalescent phase. A The gating strategy for memory T cell subsets including the naïve T cells (NT) (CD45RA+ CCR7+), the central memory T cells (TCM) (CD45RA− CCR7+), the effector memory T cells (TEM) (CD45RA− CCR7−) and the CD45RA+ effector memory T cells (TEMRA) (CD45RA+ CCR7−) were showed (n = 7). (B–C) The proportion of memory T cell subsets in CD4+ T cells (B) and CD8+ T cells (C) in the patients during convalescent phase were displayed (n = 7). Data were shown as mean ± SEM. Student's t test was performed and P < 0.05 was considered of significant difference.
S-RBD or N-specific total antibodies or IgG all displayed a prominent increasing trend and mostly mounted a strikingly high level in middle and convalescent phase (Fig. 4A–4D). There was only one N IgG negative sample (5.89 × 103 RLU/1 μL) in convalescent phase, while her S-RBD IgG was remarkably high (1.14 × 105 RLU/1 μL).
Figure 4. Kinetics of antibody production of children with COVID-19 in different phases of disease onset. (A–H) Sera from 19 COVID-19 patients were analyzed on SARS-CoV-2 S-RBD or N-specific total antibody (A, B), IgG (C, D), IgM (E, F) and IgA (G, H) during acute phase (n = 19), middle phase (n = 10) and convalescent phase (n = 14) respectively. The doted horizontal lines indicate the cut-off values which were set at the triple means of the RLU values of healthy controls. (I, J) The kinetics of S-RBD-specific IgG, IgM and IgA in 2 cases of the children with COVID-19. RLU, relative light unit; S/CO ratio, the ratio of sample RLU value to cut-off RLU value.
A total of 10 (53%) patients were considered as S-RBD IgM positive, including 6 of 19 samples in acute phase, 1 of 10 in middle phase and 3 of 14 in convalescent phase (Fig. 4E). Nine (47%) patients were defined as N IgM positive, with 3 of 19 samples in acute phase, 5 of 10 in middle phase and 9 of 14 in convalescent phase (Fig. 4F). Altogether, 11 patients (58%) were determined as S-RBD or N IgM positive, and 9 (47%) were detected during acute phase.
S-RBD seemed to be a better antigen to induce IgA than N protein, since 14 patients (74%) were defined as S-RBD IgA positive (Fig. 4G), while only 2 (11%) were N IgA positive (Fig. 4H). S-RBD IgA was on a clearly increasing trend from acute to middle phase, with 2 of 19 samples in acute phase and 7 of 10 in middle phase were considered as positive. Notably, 3 of the 7 S-RBD IgA positive patients in middle phase showed a decreasing trend afterwards.
Furthermore, to clarify the antibody kinetics after disease onset, the only two patients who had both consecutive samples and early S-RBD IgM response were chosen to show the pattern of antibody production individually. Typically, after a short presence in acute phase, S-RBD IgM declined rapidly to undetectable levels, while S-RBD IgA reached its peak level in middle phase but decreased afterwards. S-RBD IgG were elicited later than IgM and IgA, but continued to grow in convalescent phase (Fig. 4I, 4J).