Vagus nerve inflammation contributes to dysautonomia in COVID-19

Abstract

Dysautonomia has substantially impacted acute COVID-19 severity as well as symptom burden after recovery from COVID-19 (long COVID), yet the underlying causes remain unknown. Here, we hypothesized that vagus nerves are affected in COVID-19 which might contribute to autonomic dysfunction. We performed a histopathological characterization of postmortem vagus nerves from COVID-19 patients and controls, and detected SARS-CoV-2 RNA together with inflammatory cell infiltration composed primarily of monocytes. Furthermore, we performed RNA sequencing which revealed a strong inflammatory response of neurons, endothelial cells, and Schwann cells which correlated with SARS-CoV-2 RNA load. Lastly, we screened a clinical cohort of 323 patients to detect a clinical phenotype of vagus nerve affection and found a decreased respiratory rate in non-survivors of critical COVID-19. Our data suggest that SARS-CoV-2 induces vagus nerve inflammation followed by autonomic dysfunction which contributes to critical disease courses and might contribute to dysautonomia observed in long COVID.

Keywords: COVID-19; Dysautonomia; Neuroinflammation; Vagus nerve.

Fig. 1

Inflammation of the vagus nerve is commonly observed in COVID-19. a H&E staining (left) and HLA-DR staining (middle) with arrow indicating the entry of the vagal nerve. Scale bar shows 2 mm. Higher magnification of vagal nerve entry into the brain stem of HLA-DR (middle, top) and CD8 (middle, bottom) staining and quantification. Scale bar shows 100 µm. b Graphical summary of experimental setup. c Volcano plot depicting all identified genes. Genes with FDR-adjusted P value < 0.05 and log2 foldchange > 1 are labeled in red and < − 1 in blue. Top 5 most up-regulated and down-regulated genes are labeled. d GO term analysis of top 200 differentially up-regulated genes. Color shows Padj and size number of genes included in the respective GO term. e GO term analysis of top 200 differentially down-regulated genes. Color shows FDR-adjusted P value and size number of genes included in the respective GO term. f SARS-CoV-2 RNA levels determined by qPCR in 23 available vagus nerve samples grouped into low, intermediate, and high groups. Logarithmic scale is shown on y-axis. g–i Volcano plots of differential expression analyses between low (g), intermediate (h), and high (i) viral load groups against controls. j Heatmaps of GO terms that are up- or down-regulated independent of SARS-CoV-2 viral load (top) or correlate with SARS-CoV-2 viral load (down) determined by generalized linear model corrected for age and sex. k GO terms that are significantly enriched independently of SARS-CoV-2 viral load. l GO terms that are positively associated with SARS-CoV-2 viral load. m GO terms that are negatively associated with SARS-CoV-2 viral load

Fig. 2

Cell-type-dependent inflammatory responses in vagus nerves of COVID-19 patients. a Network heatmap plot of weighted gene-correlated network analysis (WGCNA) for module identification in all samples. b Modules separated by clearly identified cell types through enrichment analysis with PanglaoDB. Modularity between control and COVID-19 samples were compared within each module. P < 0.05 shows significantly stronger enrichment of COVID-19 samples. Dashed line represents P = 0.05. Color represents module. c Heatmap depicting the enrichment of interferon signaling and response GO terms in all identified modules. Cell types and modules are separately labeled as individual colors. Color range shows row z-score. Non-significant enrichment is depicted as gray tile. d–g Emap plots of GO term enrichment analyses in the orange monocyte (d), turquoise endothelial cell (e), brown Schwann cell (f), and saddle-brown neuronal (g) modules. Interferon signaling GO terms were excluded. Color shows FDR-adjusted P value. Size shows number of genes included in the GO term. h Representative images of neurofilament, CD8, and CD68 in vagus nerves of control (top) and COVID-19 (bottom) samples. Scale bar shows 50 µm. i Comparison of neurofilament (DAB/tissue intensity ratio), CD8⁺ T cells per mm² and CD68 (DAB/tissue intensity ratio) between control (n = 8, except for CD68 where 1 outlier was removed since it was > 3.5 × z-score of all samples) and COVID-19 samples (n = 16). Student’s t-test was used for statistical comparison. *P < 0.05

Fig. 3

Maladaptive respiratory rate predicts lethal outcome in COVID-19. a–d Comparisons of oxygen saturation (in percent; a), C-reactive protein (mg L^–1; b), white blood cell count (billion mL^–1; c), and respiratory rate (per minute; d) in patients with mild, moderate, critical, and lethal COVID-19. One-way ANOVA was performed for testing group differences. T test with FDR correction for multiple comparisons against mild COVID-19 group was performed. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. e Odds ratio analysis corrected for age and sex to predict lethal outcome in critically ill patients. pH, venous carbon dioxide levels (pCO₂), peripheral oxygen saturation (sO₂), white blood cell count (WBC), lactate, interleukin-6 (IL-6), C-reactive protein (CRP), and respiratory rate (RR) were tested as predictors. Odds ratio and 95% confidence intervals are shown. *P < 0.05. f–i Correlation analyses of z-transformed respiratory rate and z-transformed venous oxygen level (pO₂; f), venous carbon dioxide level (pCO₂; g), pH (h), and C-reactive protein (CRP; i). Spearman correlation analysis was performed. Correlation coefficients and P values are provided in the figure. j Graphical summary of vagus nerve inflammation in COVID-19