Mathematical modelling and transcriptional characterisation of intratumoral regulatory T cell sub-populations

Thesis

Regulatory T cells (Tregs) suppress harmful immune responses and help prevent autoimmune reactions. However, high densities of Tregs accumulate in solid tumours and contribute to the suppression of anti-tumor responses. The targeting of intratumoral Tregs via antibody-mediated blockade/depletion is a promising cancer treatment strategy but remains ineffective and induces adverse events. Some of these challenges can be addressed by conducting extensive transcriptional profiling of intratumoral Tregs and quantitatively assessing Treg migration dynamics to identify appropriate therapeutic targets and guide the development of safe and effective Treg-targeting therapeutics.

We hypothesised that tumours harbour resident and transient Tregs with distinct transcriptional and migratory profiles. To test our hypothesis, we used photoconvertible Kaede mice to label and quantify Tregs in MC38 and CT26 tumours. We developed a mathematical model to investigate whether the presence of resident and transient tumour Tregs aligns with the observed data and applied Bayesian parameter estimation to infer the rates of intratumoral Treg migration. In both tumours, we confirmed the presence of resident and transient Tregs and found that Tregs egressed from CT26 tumours to lymphoid tissue to a larger extent than MC38 tumours. Single cell analysis of intratumoral Tregs revealed that resident Tregs upregulate Lag3, Tnfrsf18 and Tnfrsf9. To investigate similarities between murine and human intratumoral Tregs, we constructed Treg meta-atlases using public single cell datasets. In mice, we observed that Lag3 and Ccr8 are unique to tumour Tregs and exhausted CD8 T cells. In humans, CCR8 expression was confined to tumour Tregs whereas LAG3 was mainly expressed on exhausted CD8 T cells.

Collectively, our findings outlined the presence of kinetically and transcriptionally distinct intratumoral Treg subpopulations. Furthermore, our Treg meta-atlases exposed key differences in CCR8 and LAG3 expression between mouse models and humans and supported the potential of CCR8 as a target for depleting human intratumoral Tregs.

Authors: Muzhingi, I

• DOI: 10.5287/ora-2a0eegv1d
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: single cell meta-atlas; CCR8; immunotherapy; LAG3; mathematical modelling; tumour Tregs
• Subjects: Computational Biology; Computational Immunology; Immuno-oncology