Improvement of immune dysregulation in individuals with long COVID at 24-months following SARS-CoV-2 infection

Abstract

This study investigates the humoral and cellular immune responses and health-related quality of life measures in individuals with mild to moderate long COVID (LC) compared to age and gender matched recovered COVID-19 controls (MC) over 24 months. LC participants show elevated nucleocapsid IgG levels at 3 months, and higher neutralizing capacity up to 8 months post-infection. Increased spike-specific and nucleocapsid-specific CD4⁺ T cells, PD-1, and TIM-3 expression on CD4⁺ and CD8⁺ T cells were observed at 3 and 8 months, but these differences do not persist at 24 months. Some LC participants had detectable IFN-γ and IFN-β, that was attributed to reinfection and antigen re-exposure. Single-cell RNA sequencing at the 24 month timepoint shows similar immune cell proportions and reconstitution of naïve T and B cell subsets in LC and MC. No significant differences in exhaustion scores or antigen-specific T cell clones are observed. These findings suggest resolution of immune activation in LC and return to comparable immune responses between LC and MC over time. Improvement in self-reported health-related quality of life at 24 months was also evident in the majority of LC (62%). PTX3, CRP levels and platelet count are associated with improvements in health-related quality of life.

Fig. 1. Humoral response in participants with long COVID following SARS-CoV-2 infection.

A Anti-spike IgG levels were elevated in LC at earlier timepoints. B Higher anti-nucleocapsid were significantly higher at 3-months in LC. C Neutralization tired remained higher in LC up to 8-months post-infection. There was no difference following vaccination at 12- or 24-months. D Similar frequencies of bulk CD19 + B-cells between the two groups. E Representative dot plots showing spike tetramer binding from memory (IgD-) B cells. Increased frequencies of spike-specific B cells evident following vaccination. F Representative dot plots showing nucleocapsid tetramer binding from memory (IgD-) B cells. Data shown as medians with interquartile ranges. LC= Long COVID, MC= matched controls, M= months. Mann–Whitney U test (two-sided) was used for un-paired analysis, p < 0.05 (*) were considered significant. Data points represent n = 31 biologically independent samples per group. Source data are provided as a Source Data file.

Fig. 2. T cells responses and Inhibitory marker expression in participants with long COVID.

A, B Spike-specific and nucleocapsid-specific CD4⁺T cell responses were higher in LC at 3- and 8-months. No difference at 24-months. C, D Sustained spike-specific and nucleocapsid-specific CD8 + T cell responses across all 3 timepoints. E No difference in PD-1 expression on CD4⁺ T cells. F Elevated PD-1 expression on CD8⁺ T cells at 3- and 8-months. G, H Increased TIM-3 expression on CD4 and CD8 T cells at earlier timepoints. Data shown as medians with interquartile ranges. LC= Long COVID, MC= matched controls. Mann–Whitney U test (two-sided) was used for un-paired analysis, p < 0.05 (*), < 0.01 (**) were considered significant. Data points represent n = 31 biologically independent samples per group. Source data are provided as a Source Data file.

Fig. 3. Reduction of immune activation at 24-months.

A Significantly elevated IFN-γ and IFN-β in LC at 24-months. B Higher frequencies of activated monocytes at 3- and 8-month, but not 24-months. C Lower percentages of activated myeloid dendritic cells (mDC) at 24-months. D Higher percentages of activated plasmacytoid dendritic cells (pDC) at 3- and 8-month that decreased 24-months. Data shown as medians with interquartile ranges. LC Long COVID, MC matched controls. Mann–Whitney U test (two-sided) was used for un-paired analysis, p < 0.05 (*), < 0.01 (**) were considered significant. Data points represent n = 31 biologically independent samples per group. Source data are provided as a Source Data file.

Fig. 4. Reconstitution of immune cell subsets at 24-months.

A Longitudinal plot showing changes in cell subset proportions overtime (3-, 8- and 24-months) in MC (left) and LC (right) (B) UMAP showing cellular composition single-cell RNAseq data; all (combined cells [45,988 cells]) then separated into LC (n = 10) and MC (n = 10) at 24-months. C Composition of cell subset frequencies between LC and MC with p values and false discovery rate (FDR). D UMAP assessing only naïve B and T cell subsets, with no clear difference between LC or MC. E Box and whiskers graph (median with IQR) showing interferon response scores (IRS) in innate cell subsets. Elevated IRS in CD14+ monocytes. F Exhaustion score in T cell subsets. Dots represent individual cells (outliers) for each subset with weighted scores (median with IQR). LC Long COVID, MC matched controls. Data shown as mean scores. p < 0.05 (*), < 0.0001(****) were considered significant. Data represent n = 10 biologically independent samples per group. Source data are provided as a Source Data file.

Fig. 5. Changes in health-related quality of life by EQ-5D-5L index score.

EQ-5D-5L index score at 4-month visit are ordered ascending on the x-axis for Matched Control (A) and Long COVID (B) participants. Vertical lines connect the participants’ initial EQ-5D-5L score (4-month) and the last available EQ-5D-5L score (12- or 24-month). Crosses = EQ-5D-5L index score at 4-month visit, Triangle symbol = EQ-5D-5L index score at 12-month visit, Circle symbol = EQ-5D-5L index score at 24-month visit, Closed symbol = no COVID-19 reinfection during follow-up, Open symbol = COVID-19 reinfection during follow-up. C Median (dark dashed line) and participant (pale solid line) trajectories of EQ−5D-5L index score over study follow-up. D Participant-reported functional status at 24-months post-infection, related to full recovery from COVID, return to COVID-19 work, return to usual activities, and return to exercise level. LC Long COVID, MC matched controls.

Fig. 6. Blood parameters associated with improvement in health-related quality of life at 24-months.

A Representative bar graph of log-linear model, showing frequency of features highly associated with recovery. B Table summarizing accuracy and F1 score for top 2 and top 3 most highly associated features. CI = 95% confidence interval. C Left-panel: 3-dimensional scatter plot of recovered vs unrecovered participant with concentration values of 3 markers (PTX3, CRP and platelets). Right-panel: 2D projections of PTX3 vs platelets (upper) and PTX3 vs. CRP (lower) with line representing the decision boundary. Recovered refers to improvement in health-related quality of life, unrecovered= no improvements, MC matched controls. Source data are provided as a Source Data file.