Kinetics of viral load and antibody response in relation to COVID-19 severity

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent for coronavirus 2019 (COVID-19) pneumonia. Little is known about the kinetics, tissue distribution, cross-reactivity, and neutralization antibody response in patients with COVID-19. Two groups of patients with RT-PCR-confirmed COVID-19 were enrolled in this study: 12 severely ill patients in intensive care units who needed mechanical ventilation and 11 mildly ill patients in isolation wards. Serial clinical samples were collected for laboratory detection. Results showed that most of the severely ill patients had viral shedding in a variety of tissues for 20-40 days after onset of disease (8/12, 66.7%), while the majority of mildly ill patients had viral shedding restricted to the respiratory tract and had no detectable virus RNA 10 days after onset (9/11, 81.8%). Mildly ill patients showed significantly lower IgM response compared with that of the severe group. IgG responses were detected in most patients in both the severe and mild groups at 9 days after onset, and remained at a high level throughout the study. Antibodies cross-reactive to SARS-CoV and SARS-CoV-2 were detected in patients with COVID-19 but not in patients with MERS. High levels of neutralizing antibodies were induced after about 10 days after onset in both severely and mildly ill patients which were higher in the severe group. SARS-CoV-2 pseudotype neutralization test and focus reduction neutralization test with authentic virus showed consistent results. Sera from patients with COVID-19 inhibited SARS-CoV-2 entry. Sera from convalescent patients with SARS or Middle East respiratory syndrome (MERS) did not. Anti-SARS-CoV-2 S and N IgG levels exhibited a moderate correlation with neutralization titers in patients' plasma. This study improves our understanding of immune response in humans after SARS-CoV-2 infection.

Keywords: Adaptive immunity; Infectious disease; Molecular biology; Molecular diagnosis; Virology.

Figure 1. Temporal profile of serial viral load from different tissue samples.

Viral loads in patients in the ICU (PT1–PT12) and patients with mild disease (PT13–PT23) as measured by nasal swabs (A), pharyngeal swabs (B), sputum (C), feces (D), urine (E), and blood (F). The x axis indicates the number of days after onset, the y axis indicates patient numbers. Heatmap of Ct values of viral loads were shown. A Ct value less than 37 indicates the presence of SARS-CoV-2 nucleic acid in the sample. Each square represents 1 sample detected and the gray squares indicate that the sample was viral nucleotide acid–negative.

Figure 2. Kinetics of IgM and IgG responses against SARS-CoV-2 in severely and mildly ill patients.

IgM (A) and IgG (B) antibody responses against the N protein of SARS-CoV-2 in plasma were detected. Serial plasma samples were collected from 12 severely ill and 11 mildly ill patients infected with SARS-CoV-2. Forty-eight plasma samples previously collected from healthy volunteer donors in 2017–2018 were used as a healthy donor group (HD). Positive (PC) and negative (NC) controls provided by detection kit were included to ensure test validity.

Figure 3. Kinetics of IgM and IgG responses against SARS-CoV-2 in different tissues.

Urine (A), sputum (B), feces (C), BALF, and pleural effusion (D) specimens from patients with COVID-19 were detected for the presence of IgM and IgG antibodies against the N protein of SARS-CoV-2. Positive (PC) and negative (NC) controls provided by detection kit were included to ensure test validity. Plasma from 48 HDs was also included.

Figure 4. IgG antibody response against different SARS-CoV-2 proteins or fragments.

Plasma samples collected at different time points after admission were used for IgG detection in different protein-coated ELISAs: S (1209 aa) (A), S1 (681 aa) (B), RBD (457 aa) (C), S2 (539 aa) (D), and N (430 aa) (E). Eleven plasma samples from HDs were used as controls. The correlations among IgG levels against different viral proteins were analyzed and summarized. Pearson’s correlation coefficient was used to assess the relationship among antiviral IgG levels of different proteins (F). A Student’s t test was used to analyze differences in mean values between groups A–E. A P value less than 0.05 was considered to be statistically significant. **P ≤ 0.01.

Figure 5. IgG cross-reactivity analysis between the other 6 human CoVs and SARS-CoV-2.

Spike (S) and nucleoprotein (N) of the other 6 human CoVs were used as coated target antigens to establish an in-house ELISA to detect IgG antibody for HCoV-229E (A), HCoV-NL63 (B), HCoV-HKU1 (C), HCoV-OC43 (D), SARS-CoV (E), and MERS-CoV (F). Plasma from 96 HDs and 23 SARS-CoV-2–infected patients were used (A–F). Severe indicates a severely ill patient with COVID-19; mild indicates a mildly ill patient with COVID-19; HD indicates healthy donors. Plasma samples from 18 SARS-convalescent (E) and 12 MERS-convalescent (F) patients were used as controls, respectively. A Student’s t test was used to analyze differences in mean values between groups (A–F). Experiments for each virus were independently carried out. Multiple comparisons following 1-way ANOVA and Kruskal-Wallis test were performed for statistical analysis. Bonferroni’s correction was used to avoid inflation of experiment-wise Type I error. In A–D, a difference was considered statistically significant when the P value was lower than 0.0167 (0.05/3); *P ≤ 0.0167, **P ≤ 0.0033, ***P ≤ 0.00033, ****P ≤ 0.000033. In E and F, a difference was considered statistically significant when the P value was lower than 0.0083 (0.05/6); ^†P ≤ 0.0083, ^††P ≤ 0.0017, ^‡P ≤ 0.00017, ^‡‡P ≤0.000017.

Figure 6. Neutralizing and cross-protection of antibody response against SARS-CoV-2 in severely and mildly ill patients.

Serial plasma samples were collected from severely ill (A) and mildly ill (B) patients infected with SARS-CoV-2, and used for authentic SARS-CoV2 neutralizing test FRNT₅₀ to evaluate kinetics of neutralizing antibodies in SARS-CoV-2 infected patients. Plasma samples collected 3 weeks after onset were used to compare cross-neutralizing antibodies between severely ill and mildly ill patients with SARS-CoV-2 and SARS-CoV–convalescent patients using SARS-CoV-2 pseudotype (C) and authentic virus (D) at a fixed dilution (1:40). A Student’s t test was used to analyze differences in mean values between groups. Experiments for each virus were independently carried out. Multiple comparisons following 1-way ANOVA and Kruskal-Wallis tests were performed for statistical analysis. Bonferroni’s correction was used to avoid inflation of experiment-wise Type I error. There were a total of 10 pairwise comparisons among 5 groups. Hence, a difference was considered statistically significant when the P value was lower than 0.005 (0.05/10). ****P ≤ 0.0001 (C and D). Pearson’s correlation coefficient was used to assess the relationship between neutralizing titer and S- and N-specific IgG levels (E and F); viral loads of respiratory specimens (G) were analyzed.