Abstract

Graphical abstract

Open in a separate window

Introduction

Data on the persistence of the SARS-CoV-2 virus and immune response of the infected host might guide efforts for the prevention and control of its expansion, as well as treatment of COVID-19 cases. Therefore, we studied the kinetics of viral load in a variety body fluids, IgA, IgG, IgM and neutralizing antibodies against SARS-CoV-2 by monitoring mild and severe COVID-19 patients from designated hospitals in Guangdong, China. Healthy close contacts were used as the control group.

Material and methods

Results

We recruited 35 COVID-19 cases in this study, including 28 mild and seven severe cases. In total, 618 clinical specimens were collected, including 105 nasopharyngeal swabs and 84 throats swabs, and 160 faeces, 60 sputum and 209 serum samples. We found that the viral loads in severe cases was higher than those in mild cases in nasopharyngeal (1.3 × 10⁶ vs 6.4 × 10⁴, p < 0.05; 7~8 weeks) and throat samples (6.9 × 10⁶ vs 2.9 × 10⁵, p < 0.05), but similar in sputum samples (5.5 × 10⁶ vs 0.9 × 10⁶, p < 0.05) (Fig. S1). In both groups, the viral load in upper respiratory tract samples (nasopharyngeal, throat swabs) reached the peak level at the first week after illness onset and declined quickly within the following 2~4 weeks (Figs. 1 A,B). Prolonged shedding of virus RNA could be detected in nasopharyngeal swabs of each mild and severe case at 6~8 weeks after illness onset, while the viral load was extremely low and close to the detection limit of real time RT-PCR (1.7~6.0 × 10³ copies/mL). In sputum (Fig. 1C) and faeces samples (only measured in mild cases, Fig. 1D), the viral load was somehow higher than that in upper respiratory tract samples, and reached to 5.6 × 10⁷ copies/mL in peak level at the first week after illness onset, but declined to 1.6 × 10³ copies/mL within 2~3 weeks. Viraemia was only detected in one of seven severe cases, and sustained for 5~6 days (Fig. 1E).

Open in a separate window

Fig. 1

Kinetics of viral load in 35 COVID-19 cases. The copies of SARS like coronavirus 2 (SARS-CoV-2) RNA in nasopharyngeal swab (A), throat swab (B), sputum (C), faeces (D) and serum specimens (E) were estimated by means of real-time reverse-transcriptase polymerase chain reaction. The dashed curve draws with exponential growth model and indicates the dynamic tendency of viral load in variety body fluids. Each dot of viral load present as mean and standard deviation. Purple symbols denote severe cases, and cyan symbols denote mild cases.

IgA in serum samples was detected in 1 of 35 cases at the first week after the date of illness onset (Fig. 2 A). IgA levels were already high at day 11 in severe cases, but still low in mild cases. The kinetics of IgA showed an increasing trend in mild cases, but declining in severe cases (Fig. 2A). The production of IgM remained at a low level in both the severe and the mild group (Fig. 2B). Surprisingly, IgG was detected in one mild case at the first week after illness onset. At the third week, 68.9% (n = 20/29) patients seroconverted in mild cases, while it was 100% (n = 7/7) in severe cases (Fig. 2C). A similar increasing trend of IgG was observed both in severe and in mild cases (Fig. 2C).

Open in a separate window

Fig. 2

Kinetics of IgA (A), Ig M (B), Ig G (C) and neutralizing antibodies (D) to SARS-CoV-2 over time in 35 COVID-19 cases. The dashed curve draws with experiential growth model and indicates the dynamic tendency of antibodies. Each dot of ELISA presents as mean and standard deviation, and titres of neutralizing antibodies present geometric mean and standard deviation. Purple symbols denote severe cases, and cyan symbols denote mild cases.

None of the 35 cases produced neutralizing antibodies at the first week after the date of illness onset (Fig. 1E). At the second week, seven of 28 mild cases produced neutralizing antibodies with titres of 1:32 to 1:128, and the remaining cases showed no neutralizing antibodies. At the third week, all 35 cases produced neutralizing antibodies, reaching titres of 1:512 in two of 29 mild cases and one of seven severe cases. Importantly, we found no differences in the increasing trends of neutralizing antibodies between mild and severe cases. All 21 healthy close contacts tested negative for both virus RNA and neutralizing antibodies against SARS-CoV-2.

Additionally, we determined the consistency among neutralizing antibodies and IgA, IgM and IgG by calculating the Spearman correlation coefficient and kappa value between mild and severe case groups. Neutralizing antibodies correlated with IgG (r _s = 0.85, p < 0.05; kappa = 0.85) and IgA (r _s = 0.64, p < 0.05; kappa = 0.61) in severe but not in mild cases (IgG, r _s = 0.42, kappa = 0.33; IgA, r _s = 0.32, kappa = 0.22). No correlation with IgM antibodies in either severe (r _s = 0.17, kappa = 0.06) or mild cases (r _s = 0.27, kappa = 0.15) was seen (Table S2).

Discussion

Our data showed that the course of virus shedding and antibody development triggered by SARS-CoV-2 resembles that of SARS-CoV and MERS-CoV [4,5]. The major difference was prolonged virus shedding in upper respiratory tract samples, e.g. nasopharyngeal swab samples, which seems to mimic the manifestation of seasonal influenza [6]. Although we found higher viral load in sputum and faeces than in upper respiratory tract samples, the heterogeneity of the specimens limits its significance in explaining the differences of the ability for onward transmission by either sample type. The early production of IgG antibodies showed a similar increasing trend with neutralizing antibodies in both mild and severe cases. However, the seroconversion rate of IgM was low in the acute stage of infection in both severe and mild cases. This phenomenon was also encountered by other groups [[7], [8], [9]]. Thus, the early detection of COVID-19 through determination of IgM may not be an ideal strategy. Surprisingly, we found two cases of early secretion of IgA in serum at the first week after illness onset in the mild group, which highlights an alternative choice of target antibody for early detection of COVID-19 in clinic [10]. Moreover, we found a higher proportion of seroconversion for IgA in severe cases, which may serve as a candidate indicator of prognosis of COVID-19 in future [10]. The correlation between neutralizing antibodies and IgG suggests that IgG could be used as a marker for the production of neutralizing antibodies [11]. However, we found the titres of neutralizing antibodies in most cases remained at a low level, which may indicate an inability to clear the infection. Indeed, we found neutralizing antibodies in all severe cases. The potential mechanisms, because of either low titres or lower host immune cell-mediated response, need further investigation [12]. In conclusion, our findings may be of significance in interpreting the kinetics of viral load and antibodies to SARS-CoV-2 in both severe and mild cases, which may aid the current and future prevention and control of the global pandemic of COVID-19.

Author contributions

S.J., S.R. and K.C. designed the study. T.X., B.R. and T.J. collected the epidemiological and clinical data. L.C., Z.L., H.L., R.Y., Z.P., H.X., X.Q., P.J., C.F., K.B., S.J., L.Z. and L.J. carried out the investigations. S.J. drafted the manuscript, S.R. and K.C. revised the final manuscript.

Transparency declaration

This work was supported by grants from the Guangdong Provincial Novel Coronavirus Scientific and Technological Project (2020111107001) and Guangzhou Novel Coronavirus Scientific and Technological Project (202008040004). All authors report no conflicts of interest relevant to this article.

Notes

Editor: J. Rodriguez-Baño

Footnotes

References