Remodeling of T Cell Dynamics During Long COVID Is Dependent on Severity of SARS-CoV-2 Infection

Abstract

Several COVID-19 convalescents suffer from the post-acute COVID-syndrome (PACS)/long COVID, with symptoms that include fatigue, dyspnea, pulmonary fibrosis, cognitive dysfunctions or even stroke. Given the scale of the worldwide infections, the long-term recovery and the integrative health-care in the nearest future, it is critical to understand the cellular and molecular mechanisms as well as possible predictors of the longitudinal post-COVID-19 responses in convalescent individuals. The immune system and T cell alterations are proposed as drivers of post-acute COVID syndrome. However, despite the number of studies on COVID-19, many of them addressed only the severe convalescents or the short-term responses. Here, we performed longitudinal studies of mild, moderate and severe COVID-19-convalescent patients, at two time points (3 and 6 months from the infection), to assess the dynamics of T cells immune landscape, integrated with patients-reported symptoms. We show that alterations among T cell subsets exhibit different, severity- and time-dependent dynamics, that in severe convalescents result in a polarization towards an exhausted/senescent state of CD4+ and CD8+ T cells and perturbances in CD4+ Tregs. In particular, CD8+ T cells exhibit a high proportion of CD57+ terminal effector cells, together with significant decrease of naïve cell population, augmented granzyme B and IFN-γ production and unresolved inflammation 6 months after infection. Mild convalescents showed increased naïve, and decreased central memory and effector memory CD4+ Treg subsets. Patients from all severity groups can be predisposed to the long COVID symptoms, and fatigue and cognitive dysfunctions are not necessarily related to exhausted/senescent state and T cell dysfunctions, as well as unresolved inflammation that was found only in severe convalescents. In conclusion, the post-COVID-19 functional remodeling of T cells could be seen as a two-step process, leading to distinct convalescent immune states at 6 months after infection. Our data imply that attenuation of the functional polarization together with blocking granzyme B and IFN-γ in CD8+ cells might influence post-COVID alterations in severe convalescents. However, either the search for long COVID predictors or any treatment to prevent PACS and further complications is mandatory in all patients with SARS-CoV-2 infection, and not only in those suffering from severe COVID-19.

Keywords: COVID-19; T cell exhaustion/senescence; convalescents; full spectral cytometry; immune system; inflammation resolution; long COVID; post-acute COVID-syndrome (PACS).

Figure 1

(A) The experimental pipeline. (B) IgG antibodies against SARS-Cov-2 S1/S2 antigens in COVID-19 convalescent individuals. Type of severity (mild, moderate and severe) as well as time-points (I or II) are indicated. Data shown in the log2 scale represent individual values, median +/- 95% CI (confidence interval). Statistics analysis by non-parametric T-test, *p ≤0,05; **p ≤ 0,01; ***p≤ 0,001; ****p≤ 0,0001.

Figure 2

Plasma level of cytokines and chemokines from COVID-19 convalescents and controls. Quantification of cytokines and other mediators in plasma obtained from COVID-19 convalescents after severe (time-point I, n=20, time-point II, n=11), moderate (time-point I, n=20, time-point II, n=12) and mild disease (time-point I, n = 15, time-point II, n=13) at two time points after disease confirmation and from healthy controls (n = 13). Data represent individual values, mean (centre bar) ± SD (upper and lower bars). Statistical analysis by one-way ANOVA test, *p ≤ 0,05; **p ≤ 0,01; ***p ≤ 0,001; ****p ≤ 0,0001.

Figure 3

Unsupervised analysis of CD4+ T cells and their characterization. (A) Uniform Manifold Approximation and Projection (UMAP) UMAP representation of CD4+ T cell landscape. (B) Heat map representing different clusters identified by FlowSOM, with relative identity and percentages in healthy controls and convalescent patients. The color in the heat map represents the median of the arcsinh, 0–1 transformed marker expression calculated over cells from all the samples, varying from blue for lower expression to red for higher expression. Each cluster has a unique color assigned (bar on the left). Barplots along the rows (clusters) and values on the right indicate the relative sizes of clusters. (C) Differential analysis of all severity groups of COVID-19 convalescent patients (mild, moderate, severe), as well as healthy donors at the time points I and II. The heat represents arcsine-square-root transformed cell frequencies that were subsequently normalized per cluster (rows) to mean zero and standard deviation of one. The color of the heat varies from dark blue indicating relative under-representation to red indicating relative over-representation. Bars at the right indicate significantly differentially abundant clusters (green). (D) Differential proportion of selected clusters presented as % of CD4+ cells. *p ≤ 0,05; **p ≤ 0,01; ***p ≤ 0,001

Figure 4

Unsupervised analysis of Treg cells and their characterization. (A) Uniform Manifold Approximation and Projection (UMAP) UMAP representation of Treg cell landscape. (B) Heat map representing different clusters identified by FlowSOM, with relative identity and percentages in healthy controls and convalescent patients. The color in the heat map represents the median of the arcsinh, 0–1 transformed marker expression calculated over cells from all the samples, varying from blue for lower expression to red for higher expression. Each cluster has a unique color assigned (bar on the left). Barplots along the rows (clusters) and values on the right indicate the relative sizes of clusters. (C) Differential analysis of all severity groups of COVID-19 convalescent patients (mild, moderate, severe), as well as healthy donors at the time points I and II. The heat represents arcsine-square-root transformed cell frequencies that were subsequently normalized per cluster (rows) to mean zero and standard deviation of one. The color of the heat varies from dark blue indicating relative under-representation to red indicating relative over-representation. Bars at the right indicate significantly differentially abundant clusters (green). (D) Differential proportion of selected clusters presented as % of Treg cells. *p ≤ 0,05; **p ≤ 0,01; ***p ≤ 0,001.

Figure 5

Unsupervised analysis of CD8+ T cells and their characterization. (A) Uniform Manifold Approximation and Projection (UMAP) UMAP representation of CD8+ T cell landscape. (B) Heat map representing different clusters identified by FlowSOM, with relative identity and percentages in healthy controls and convalescent patients. The color in the heat map represents the median of the arcsinh, 0–1 transformed marker expression calculated over cells from all the samples, varying from blue for lower expression to red for higher expression. Each cluster has a unique color assigned (bar on the left). Barplots along the rows (clusters) and values on the right indicate the relative sizes of clusters. (C) Differential analysis of all severity groups of COVID-19 convalescent patients (mild, moderate, severe), as well as healthy donors at the time points I and II. The heat represents arcsine-square-root transformed cell frequencies that were subsequently normalized per cluster (rows) to mean zero and standard deviation of one. The color of the heat varies from dark blue indicating relative under-representation to red indicating relative over-representation. Bars at the right indicate significantly differentially abundant clusters (green). (D) Differential proportion of selected clusters presented as % of CD8+ cells. *p ≤ 0,05; ** p ≤ 0,01; *** p ≤ 0,001

Figure 6

Unsupervised analysis of cytokine production by CD4+ (A), Treg (B) and CD8+ (C) cells from convalescent patients after mild, moderate and severe COVID-19 as well as healthy donors analyzed at time I and time II. Upper panels - Heatmaps showing cell clusters identified by FlowSOM with differential analysis of all severity groups of COVID-19 convalescent patients (mild, moderate, severe) as well as healthy controls at the time points I and II. The heat represents arcsine-square-root transformed cell frequencies that were subsequently normalized per cluster (rows) to mean of zero and standard deviation of one. The color of the heat varies from dark blue indicating relative under-representation to red indicating relative over-representation. Bars at the right indicate signi ficant differentially abundant clusters (green). The population without any activity (negative for all markers) has been marked as “All-”. Lower panels – differential proportion of selected clusters presented as % of analyzed cells. *p ≤ 0,05; **p ≤ 0,01; ***p ≤ 0,001.