Initial COVID-19 severity influenced by SARS-CoV-2-specific T cells imprints T-cell memory and inversely affects reinfection

Abstract

The immunoprotective components control COVID-19 disease severity, as well as long-term adaptive immunity maintenance and subsequent reinfection risk discrepancies across initial COVID-19 severity, remain unclarified. Here, we longitudinally analyzed SARS-CoV-2-specific immune effectors during the acute infection and convalescent phases of 165 patients with COVID-19 categorized by severity. We found that early and robust SARS-CoV-2-specific CD4⁺ and CD8⁺ T cell responses ameliorate disease progression and shortened hospital stay, while delayed and attenuated virus-specific CD8⁺ T cell responses are prominent severe COVID-19 features. Delayed antiviral antibody generation rather than titer level associates with severe outcomes. Conversely, initial COVID-19 severity imprints the long-term maintenance of SARS-CoV-2-specific adaptive immunity, demonstrating that severe convalescents exhibited more sustained virus-specific antibodies and memory T cell responses compared to mild/moderate counterparts. Moreover, initial COVID-19 severity inversely correlates with SARS-CoV-2 reinfection risk. Overall, our study unravels the complicated interaction between temporal characteristics of virus-specific T cell responses and COVID-19 severity to guide future SARS-CoV-2 wave management.

Fig. 1. Longitudinal COVID-19 cohort and dynamics of viral load.

a Schematic of study participants split by severity categories, and longitudinal sample collection including every 7 days during acute COVID-19 symptom onset and at 4- and 7-months post-recovery. b, c Distribution of age and number of comorbidities between moderate (n = 47), severe (n = 45) and critical (n = 30) groups. d Proportion of patients with positive (purple) and negative (blue) nucleic acid tests, divided by a CT value threshold of 40, among the three groups (moderate, 7 dps n = 7, 14 dps n = 14, 21 dps n = 14, 28 dps n = 13, 35 dps n = 10; severe, 7 dps n = 9, 14 dps n = 11, 21 dps n = 12, 28 dps n = 12, 35 dps n = 9, 42 dps n = 5 ; critical, 7 dps n = 6, 14 dps n = 11, 21 dps n = 10, 28 dps n = 12, 35 dps n = 8, 42 dps n = 6). e Comparison of SARS-CoV-2 viral loads in the moderate, severe, and critical groups in the early (7–14 dps), middle (21–28 dps), and late (35–42 dps) stages of acute infection performed by t-test. f The kinetics of SARS-CoV-2 RNA clearance among patient groups using a polynomial nonlinear regression model with 95% confidence intervals visually represented by the shaded colors. The dashed line represents the threshold for the virus load at 0. Gender, comorbidity, clinical outcomes between groups were assessed using chi-square test. Each dot represents one donor. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 2. Virus-specific T cell response following acute infection across COVID-19 severity.

a Representative flow cytometry plots showing virus-specific (IFNγ⁺) CD4⁺ and CD8⁺ T cell responses during the acute infection period in the moderate (23 dps), severe (21 dps), and critical (26 dps) groups. b, d Longitudinal analysis of absolute counts of virus-specific CD4⁺ (b) and CD8⁺ (d) T cells in peripheral blood (PB) samples per mL in the three groups (moderate, 7 dps n = 6, 14 dps n = 13, 21 dps n = 11, 28 dps n = 13, 35 dps n = 9; severe, 7 dps n = 8, 14 dps n = 9, 21 dps n = 12, 28 dps n = 11, 35 dps n = 11, 42 dps n = 6 ; critical, 7 dps n = 5, 14 dps n = 13, 21 dps n = 13, 28 dps n = 12, 35 dps n = 7, 42 dps n = 6). The black lines within the violin plots represent median and percentiles. c, e Kinetics of virus-specific CD4⁺ (c) and CD8⁺ (e) T cell responses in the moderate, severe, and critical groups following acute infection, using a polynomial nonlinear regression model with 95% confidence intervals visually represented by the shaded colors. The peak points are marked with colored dashed lines. f, g Comparison of virus-specific CD4⁺ (f) and CD8⁺ (g) T cell responses between the short-term (length of stay ≤ 12 days) and extended hospitalization groups (length of stay >12 days) (short-term, 7–14 dps n = 31, 21–28 dps n = 32, 35–42 dps n = 18; extended, 7–14 dps n = 22, 21–28 dps n = 39, 35–42 dps n = 21). Each dot represents one donor. Comparisons were performed using Mann–Whitney tests. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 3. Acute antivirus antibody responses across COVID-19 severity.

a, c, e Kinetics of serum BA.5 Nab (a), anti-S + N IgG (c), and XBB.1.9 Nab (e) in the moderate, severe and critical groups following acute infection, using a polynomial nonlinear regression model with 95% confidence intervals visually represented by the shaded colors. The peak points are marked with colored dashed lines. b, d, f Comparison of geometric mean titers (GMTs) of anti-viral antibodies in longitudinal serum samples, including BA.5 Nab (b), anti-S + N IgG (d), and XBB.1.9 Nab (f) in the moderate, severe, and critical groups (moderate, 7 dps n = 7, 14 dps n = 12, 21 dps n = 13, 28 dps n = 13, 35 dps n = 10; severe, 7 dps n = 9, 14 dps n = 10, 21 dps n = 12, 28 dps n = 11, 35 dps n = 11, 42 dps n = 7; critical, 7 dps n = 5, 14 dps n = 13, 21 dps n = 13, 28 dps n = 12, 35 dps n = 7, 42 dps n = 6). Titers presented by GMT with 95%CI. g Comparison of BA.5 Nab GMT between the short-term (length of stay ≤ 12 days) and extended hospitalization groups (length of stay > 12 days) (short-term, 7–14 dps n = 33, 21–28 dps n = 34, 35–42 dps n = 20; extended, 7–14 dps n = 23, 21–28 dps n = 40, 35–42 dps n = 21). Titers presented by GMT with 95%CI. h, i Correlation between BA.5 Nab (Log10) and the number (h) or frequency (i) of IFNγ⁺CD4⁺ T cells ⁽Log10) in the moderate, severe, and critical groups. Frequency values are first multiplied by 1000 and then Log10 transformed. Spearman’s correlation coefficients and significance are shown accordingly. Each dot represents one donor. Comparisons were performed using Mann–Whitney tests. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 4. Decay and maintenance of antiviral antibodies in the COVID-19 recovery individuals.

a Timeline diagram of the different SARS-CoV-2 subvariants pandemic in China and follow-up of the COVID-19 convalescent cohort. b–e Comparison of paired plasma antibody titers from the acute infection period to 4 months after recovery, including BA.5 Nab (b), WT Nab (c), XBB.1.9 Nab (d), and anti-S + N IgG (e) in the mild (n = 13), moderate (n = 25) and severe groups (n = 24). Each dot represents one donor. Comparisons for paired samples were performed using the Wilcoxon matched-pairs signed rank test, with the fold decline marked in black font at the top. Comparisons between groups were performed using Mann–Whitney tests. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 5. Memory T cell maintenance and differentiation across initial COVID-19 severity.

a Representative flow cytometry plots showing virus-specific CD4⁺ and CD8⁺ T cell responses (IFNγ⁺) at 4 months after COVID-19 recovery in the mild, moderate and severe groups. b Comparison of the number of virus-specific memory CD4⁺ and CD8⁺ T cell responses in 4 M between individuals recovered from initial mild (n = 13), moderate (n = 25) and severe (n = 24) infection. The black lines within the box plots represent median. c Comparison of the number change of IFNγ⁺CD4⁺ and CD8⁺ T cells from the acute phase to 4 M in the moderate (n = 25) and severe (n = 24) groups. d Comparison of the numbers of bulk CD4⁺ and CD8⁺ T cells from the acute phase to 4 M in the moderate (n = 25), severe (n = 24) and mild (n = 13) groups. e Analysis of the multi-functionality of virus-specific memory CD4⁺ and CD8⁺ T cells in the mild, moderate, and severe groups at 4 M. f Representative overlay flow cytometry plots showing the bulk (black) and virus-specific (blue) CD4⁺ and CD8⁺ memory T cell phenotypes divided by CD45RA and CCR7 in the mild, moderate, and severe groups. g Frequency of different virus-specific memory T cell phenotypes in the mild (n = 13), moderate (n = 25) and severe (n = 24) groups in 4 M. Each dot represents one patient. Comparisons of paired samples were performed using the Wilcoxon matched-pairs signed rank test, with fold increase marked in black font at the top. Comparisons between groups were performed using Mann–Whitney tests. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 6. Subsequent reinfection risk and long-term immune maintenance among individuals with distinct initial COVID-19 severity.

a, b Analysis of the symptomatic SARS-CoV-2 reinfection rate (a) and the serologically confirmed (b) reinfection rate (4-fold increase in XBB.1.9 Nab between 4M and 7M) in the mild, moderate and severe convalescent groups at 7 months post acute infection. c–e Comparison of the number of virus-specific memory CD4⁺ (c) and CD8⁺ (d) T cell responses and XBB.1.9 Nab titers (e) between non-reinfected individual of mild (n = 6), moderate (n = 18) and severe (n = 17) groups in 7 M. The black lines within the violin plots represent median and percentiles. Reinfection risk between groups were assessed using chi-square test. Comparisons for paired samples were performed using the Wilcoxon matched-pairs signed rank test, with fold increase marked in black font at the top. Comparisons between groups were performed using Mann–Whitney tests. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 7. Schematic representation of acute and long-term immune effectors kinetics, and reinfection risk, across initial COVID-19 severity.

a Kinetics representation of virus-specific immune effectors and virus loads in relation to clinical severity during acute SARS-CoV-2 infection. b The profile of virus-specific T-cell memory response and neutralizing antibody maintenance across initial COVID-19 severity in 4-month post infection. c The reinfection risk and immune recall response post reinfection in 7-month post initial COVID-19 across severity