SARS-CoV-2-associated T-cell infiltration in the central nervous system

Abstract

Objectives: Infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes coronavirus disease 2019 (COVID-19). Although an acute SARS-CoV-2 infection mainly presents with respiratory illness, neurologic symptoms and sequelae are increasingly recognised in the long-term treatment of COVID-19 patients. The pathophysiology and the neuropathogenesis behind neurologic complications of COVID-19 remain poorly understood, but mounting evidence points to endothelial dysfunction either directly caused by viral infection or indirectly by inflammatory cytokines, followed by a local immune response that may include virus-specific T cells. However, the type and role of central nervous system-infiltrating T cells in COVID-19 are complex and not fully understood.

Methods: We analysed distinct anatomical brain regions of patients who had deceased as a result of COVID-19-associated pneumonia or complications thereof and performed T cell receptor Vβ repertoire sequencing. Clonotypes were analysed for SARS-CoV-2 association using public TCR repertoire data.

Results: Our descriptive study demonstrates that SARS-CoV-2-associated T cells are found in almost all brain areas of patients with fatal COVID-19 courses. The olfactory bulb, medulla and cerebellum were brain regions showing the most SARS-CoV-2 specific sequence patterns. Neuropathological workup demonstrated primary CD8⁺ T-cell infiltration with a perivascular infiltration pattern.

Conclusion: Future research is needed to better define the relationship between T-cell infiltration and neurological symptoms and its long-term impact on patients' cognitive and mental health.

Keywords: COVID‐19; COVID‐19‐specific T cell; SARS‐CoV‐2; TCR sequencing; central nervous system; neuroinflammation.

Figure 1

TCR Vβ sequencing of distinct anatomic brain regions in patients deceased from COVID‐19. (a) Schematic overview of brain biopsy regions of patients who died from severe SARS‐CoV‐2 virus infections (COVID‐19). (b) T‐cell receptor Vβ (TRB) sequencing from gDNA was performed on eight locations. (c) Number of brain‐infiltrating T‐cell receptor (TCR) clonotypes per sample of COVID‐19 patients and uninfected controls. One TCR clonotype is defined by a unique CDR3 nucleotide sequence of the rearranged TRB chain. (d) Total counts of indicated clonotypes in different brain regions of patient 8. (e) Total counts and location of clonotypes shared between individuals. (f) TRBV gene usage of brain‐infiltrating TCR clones of COVID‐19 patients and uninfected controls as a bar plot.

Figure 2

Representation of brain T‐cell clones from deceased COVID‐19 patients across different brain regions. (a–c) Search of brain‐infiltrating clones (n = 103 clones in total) derived from deceased COVID‐19 patients (n = 10 patients) in blood repertoires of unrelated COVID‐19 patients with acute infection (n = 140) or in blood of healthy controls (n = 140). The set of 103 brain‐infiltrating clones derived from deceased COVID‐19 patients was divided into eight sets of clones according to the brain area where the clone was detected: Frontal cortex (n = 9), olfactory bulb (n = 18), hippocampus (n = 6), basal ganglia (n = 11), medulla 1 (n = 17), medulla 2 (n = 18), cerebellum (n = 14), corpus callosum and gyrus cinguli (n = 10). A hit is considered a TCR clone with an identical CDR3 amino acid sequence of the rearranged TRB chain. (a) Mean hits per repertoire. Error bars represent SEM. * and **** indicate P‐values < 0.05 and < 0.0001, two‐way ANOVA. (b) Number of blood TCR repertoires which contained at least one of the brain‐infiltrating clones of the corresponding set. (c) Percentage of brain‐infiltrating clones derived from deceased COVID‐19 patients which were found in at least one blood TCR repertoire of unrelated COVID‐19 patients or healthy controls. (d) Bubble plots of brain‐infiltrating TRB clones derived from 10 deceased COVID‐19 patients grouped by brain area of origin. Brain repertoires of two uninfected control patients are shown as a comparison. One bubble represents one clone, which is defined by a unique CDR3 nucleotide sequence of the TRB chain. The area size of the bubbles corresponds to the clonal fraction within the repertoires. TRBV genes are coded by fill colour. (e) Listing of brain‐derived clonotypes that showed unique similarity matches (based on indicated Levenshtein distance) to CDR3 sequences derived from the VDJdb database with verified SARS‐CoV‐2‐reactive TCRs (n = 5609).

Figure 3

Histopathological and immunohistochemical analysis of different brain regions from COVID‐19 deceased patients. (a) Haematoxylin & eosin‐ (HE) and Kluever stainings to assess general morphology and myelination status (first two columns). The remaining 5 columns show the results of the immunohistochemistry for GFAP, CD4⁺ and CD8⁺ T cells, MHC class II expression (HLA‐DR) and microglia/macrophage infiltration (CD68) to assess the degree of immune infiltration and local neuroinflammation. In comparison with the hindbrain, reactive astrogliosis was more pronounced in the forebrain. Generally, perivascular CD4⁺ cells were very sparse, while both perivascular and parenchymal CD8⁺ cells were seen more frequently, especially in the medulla oblongata. The degree of microgliosis varied both between patients and regions, changing between diffuse patterns and microglial nodules (see HLA‐DR of upper medulla and corpus callosum, as well as CD68 of lower medulla). (b) Subsequent semiquantitative analysis of immune cell infiltration and astrogliosis in grey and white matter. For statistical testing, ANOVA was used.