Distinct Omicron longitudinal memory T cell profile and T cell receptor repertoire associated with COVID-19 hospitalisation

Abstract

SARS-CoV-2 has claimed more than 7 million lives worldwide and has been associated with prolonged inflammation, immune dysregulation and persistence of symptoms following severe infection. Understanding the T cell mediated immune response and factors impacting development and continuity of SARS-CoV-2 specific memory T cells is pivotal for developing better therapeutic and monitoring strategies for those most at risk from COVID-19. Here we present a comprehensive analysis of memory T cells in a convalescent cohort (n=20), three months post Omicron infection. Utilising flow cytometry to investigate CD4⁺CD45RO⁺ and CD8⁺CD45RO⁺ memory T cell IL-2 expression following Omicron (B.1.1.529/BA.1) peptide pool stimulation, alongside T cell receptor repertoire profiling and RNA-Seq analysis, we have identified several immunological features associated with hospitalised status. We observed that while there was no significant difference in median CD4⁺CD45RO⁺ IL-2⁺ and CD8⁺ CD45RO⁺ IL-2⁺ memory T cell count between subgroups, the hospitalised subgroup expressed significantly more IL-2 per cell following Omicron peptide pool exposure in the CD8⁺CD45RO⁺ population (p <0.03) and trended towards significance in CD4⁺CD45RO⁺ cells (p <0.06). T cell receptor repertoire analysis found that the non-hospitalised subgroup had a much higher number of circulating clonotypes, targeting a wider range of predominantly MHC-I epitopes across the SARS-CoV-2 genome. Several immunodominant epitopes, conserved between both subgroups, were observed, however hospitalised individuals were less likely to express putative HLA alleles responsible for pMHC presentation which may impact TCR affinity. We observed a bias towards shorter CDR3 segments in TCRβ repertoire analysis within the hospitalised subgroup, alongside lower rates of repertoire overlap in CDR3 sequences compared to the non-hospitalised subgroup. We found a significant proportion of TCRs targeted epitopes along the SARS-CoV-2 genome including non-structural proteins, responsible for viral replication and immune evasion. These findings highlight how the continuity of T cell based protective immunity is impacted by both the viral replication cycle of SARS-CoV-2 upon intracellular and innate immune responses, and HLA-type upon TCR affinity and clonotype formation. Our novel Epitope Target Analysis Pipeline (Epi-TAP) could prove beneficial in development of new therapeutic strategies through rapid identification of shared immunodominant epitopes across non-hospitalised and hospitalised subgroups.

Keywords: COVID-19; SARS-CoV-2; T cell immunity; T cell receptor (TCR) recognition; adaptive immunity; antigen presentation; immuno-informatics.

Figure 1

Flowchart of Epi-TAP: Epitope Target Analysis Pipeline (Version 1.0). Oval shapes represent start/stop of the pipeline. Parallelogram boxes represent input/output. Rectangular boxes represent processing steps. Servers/tools/software and databases used are mentioned in parenthesis.

Figure 2

Gating strategy for the identification of IL-2 and CD158b positive CD4⁺CD45RO⁺ and CD8⁺CD45RO⁺ T. Black arrows show the sequence of gates applied to cell subpopulations. Lymphocytes were gated based on the low SSC-A and FSC-A subpopulation within PBMCs. The singlet lymphocytes were gated by a sequence of FSC and SSC sub-gates. CD3⁺ T lymphocyte-specific marker was used to gate for CD3⁺CD4⁺ and CD3⁺CD8⁺ cells. Activated cells were identified by further gating on the CD3⁺CD4⁺CD45RO⁺ and CD3⁺CD8⁺CD45RO⁺ cell subpopulations. CD3⁺CD4⁺CD45RO⁺ and CD3⁺CD8⁺CD45RO⁺ T cells were further sub-gated with IL-2 and CD158b expression.

Figure 3

IL-2 and 158b expression in CD4⁺CD45RO⁺ T cells. (A) Median number of CD4⁺CD45RO⁺ T cells. (B) Median number of CD4⁺CD45RO⁺IL-2⁺ T cell events. (C) GeoMean of IL-2 expression in CD4⁺CD45RO⁺IL-2⁺ T cells. (D) Median number of CD4⁺CD45RO⁺158b⁺ T cell events. (E) GeoMean of 158b expression in CD4⁺CD45RO⁺158b⁺ T cells. All data are presented as Delta values with median and IQR for a given number of observations (n=10 non-hospitalised, n=10 hospitalised). Welch’s t-tests were used for statistical analysis. Data was considered to be significant if p <0.05.

Figure 4

IL-2 and 158b expression in CD8⁺CD45RO⁺ T cells. (A) Median number of CD8⁺CD45RO⁺ T cells. (B) Median number of CD8⁺CD45RO⁺IL-2⁺ T cell events. (C) GeoMean of IL-2 expression in CD8⁺CD45RO⁺IL-2⁺ T cells. (D) Median number of CD8⁺CD45RO⁺158b⁺ T cell events. (E) GeoMean of 158b expression in CD8⁺CD45RO⁺158b⁺ T cells. All data are presented as Delta values with median and IQR for a given number of observations (n=10 non-hospitalised, n=10 hospitalised). Welch’s t-tests were used for statistical analysis. Data was considered to be significant if p <0.05.

Figure 5

T cell receptor (TCR) repertoire overview. (A) Average abundance of SARS-CoV-2 specific clones following immune repertoire analysis. (B) Average diversity of SARS-CoV-2 Clonotypes (defined by unique V/J genes CDR3 sequence) (C) TCR repertoire overlap (Jaccard similarity index) (D) Distribution of CDR3 amino acid length.

Figure 6

SARS-CoV-2 epitopes targeted by T cell receptor (TCR) repertoire. SARS-CoV-2 epitopes targeted by TCRα and TCRβ chains for Non-Hospitalised (Top) and Hospitalised (Bottom) subgroups. Epitopes are ordered by position along the genome (left to right) and colour-coded by region. The number of clonotypes targeting each epitope are labelled within each bar, alongside their residing structural domain (where applicable). Immunodominant epitopes (>5% of total repertoire from either subgroup) are labelled with a dot above their respective bar.

Figure 7

Immunodominant epitope V-J distribution and pairing. Chain distribution and pairing for immunodominant epitopes in Non-Hospitalised and Hospitalised subgroups.