New academic paper: Using TCR sequencing and AI to shed light on the intricate dynamics of our immune system during SARS-CoV-2 infection

A new study co-authored by ImmuneWatch’s co-founders, Pieter Meysman, Benson Ogunjimi and Kris Laukens, analysed the role of T cells in the severity of COVID-19 and the specificity of their response to SARS-CoV-2, demonstrating the power of AI in unravelling this complex immune response.

The researchers used novel data to train the T cell receptor (TCR) epitope model, TCRex, to predict T cell receptor interactions with viral epitopes. To understand the various immunological responses associated with COVID-19, the technology was paired with experimental data.

They made the following observations:

1. CD8+ T cells targeting common epitopes shared by different coronaviruses were found in both critical and non-critical patients in the first week after symptom onset, indicating the presence of pre-existing cross-reactive memory T cells. This suggests that people who have previously been exposed to coronaviruses may have a quick immune response to SARS-CoV-2.
   
2. Critical patients had a drop in the breadth of TCR repertoires by the second week following symptom start, most likely due to lymphopoenia (low lymphocyte count). Non-critical patients, on the other hand, had T lymphocytes that were activated and expanded in response to distinct SARS-CoV-2 epitopes. This suggests that delayed or dysfunctional activation of SARS-CoV-2-specific T cells may contribute to disease severity.
   
3. As CD8+ T cells recognise unique SARS-CoV-2 epitopes, newly expanding T cell clones may play a larger role in the immune response than pre-existing cross-reactive T cell clones.
   
4. Individuals with the same disease severity may have varied T cell immune responses due to interpersonal variance in the CD8+ T-cell response.
   

A limitation of the study was suggested to be the potential underestimate of the magnitude of CD8+ T cell responses by the current TCR prediction models. Improved experimental data collecting and modelling may result in more accurate forecasts.

The key takeaways:

• Tracking TCR repertoires and T cell responses may aid in determining the efficiency of therapies and providing protection against developing SARS-CoV-2 variants.
  
  
• Investigating the progression of T cell responses to SARS-CoV-2 and other viruses may provide light on the post-acute COVID-19 symptoms.
  

Overall, this work emphasises the value of AI models in offering insights into illness aetiology, creating personalised treatments, and assessing intervention success.

The whole manuscript is available here.