Repository logo

A novel antigen-receptor signalling reporter mouse reveals clonotype-independent intratumoral trapping as a spatial feature of CD8+ T cell exhaustion



Change log


So, Tsz 


Interrogation of multi-clonal antigen-specific populations of lymphocytes can reveal fundamental principles about lymphocyte behaviour, but this has been difficult to achieve with existing technologies. A precise in vivo tool is required not only for the unbiased selection of antigen-specific lymphocytes against target antigens in an immune reaction, but also for the defined time-controlled sampling of their time-dependent fates. We have developed a novel mouse model, termed Antigen-Receptor Signalling Reporter (AgRSR) mouse, that uses the specificity of Nur77 (NR4A1) for antigen receptor activation to drive expression of a tamoxifen-inducible CreERT2, enabling time-stamping of TCR-activated lymphocytes with EYFP expression through a Cre-LoxP mechanism. This allows multiple clonal lymphocytes and their progeny to be permanently labelled. Combining isolation of the EYFP+ cells with further technologies would enable in-depth analysis of their cell states, clonotype, tissue location, and labelling localization.

Validating the AgRSR model, I first showed that the system is robust in its dependencies for TCR signalling and tamoxifen administration for EYFP induction in T lymphocytes. These experiments also revealed a surprising capacity to label endogenous antigens in the absence of exogenous antigen stimulus or adjuvant in the AgRSR mouse. It prompted further investigations into how the TCR signal threshold controls Nur77 reporter activity. in vitro T cell stimulation assays demonstrate that the reporter activity not only had a broad dynamic range of signalling reporter activity but also a correlation with TCR signal strength in vitro; a finding corroborated using a dual reporter version of the AgRSR mouse generated specifically to label T cells in receipt of high and low TCR signals, both in vitro and in vivo. Collectively, these results highlight the AgRSR mouse's capacity to label T cells reactive to diverse antigens of a broad avidity range, previously elusive to traditional methods. As such, it represents an advancement in our technological armamentarium and has the potential to contribute to understanding a variety of disease settings.

Amongst disease, cancer has been characterized as having T cell responses driven by unknown and weak antigenic signals. Exploiting the unique advantages of a system that both assesses T cell immunity across a broad dynamic range of antigenic strength and T cell immunity to unknown antigens, I investigated the tumour immune response. I applied the AgRSR system to a model of melanoma (YUMMER1.7) to understand how clonal dysfunction is achieved for CD8+ T cells in tumours. It was first demonstrated that, unlike bystander T cells marked before tumour implantation, tumour-relevant T cells marked after tumour implantation showed significant intratumoral expansion. Remarkably, an enrichment of EYFP+ T cells that are PD1+ in the spleen was also observed. This led us to hypothesize the presence of tumour-reactive T cells in the spleen and lymph node tissue that may be clonally related. Kinetic assessment of the EYFP+ cells later revealed a similar kinetic trend between the tissue sites, suggesting a similar dynamics of expansion and contraction exist between both systems.

With the hypothesis that the CD8+ T cells in both spatial compartments of the tumour and secondary lymphoid tissues may contain identical clonotypes, I set out to develop workflows that link the selectivity of the AgRSR mouse to in-depth analysis of their transcriptome and clonality using single cell RNA sequencing (scRNAseq) to enable analysis of T cells at a clonal level. First, I generated robust protocols for the isolation and processing of EYFP+ cells to scRNAseq pipeline to overcome the challenge of preserving rare cells under time constraints. With these protocols established, I moved on to characterize these cells using combined scRNAseq and VDJseq by sampling the T cells at day 8 and day 18 after tamoxifen administration, at their respective peak and plateau of expansion. Bioinformatic analysis revealed the presence of T cell clones confined exclusively to the tumour and T cell clones distributed between the tumour and secondary lymphoid organs. These clones differed in their overall transcriptional state: with exhausted members compartmentalized in the tumour and functional effectors remaining in the secondary organs. Further comparison of the dynamics between day 8 and day 18 samples revealed a loss in the extent of clones distributed across the tumour and secondary lymphoid organs, suggesting that the supply of rejuvenating functional lymphoid tissues are lost over time.

Finally, to establish directionality in the migratory relationship between the spatially confined effector T cells in the secondary lymphoid tissue and exhausted T cells in the tumour, I developed robust methods for intratumoral delivery of 4-hydroxytamoxifen to specifically label intratumoral CD8+ T cells and the isolation of those cells for scRNAseq analysis. This revealed an intriguing property of exhausted CD8+ T cells, that they do not egress from the tumour, in contrast to Treg cells that are freely detected in both tissues. This adds a spatial characteristic to the definition of T cell exhaustion.

Taken together, this thesis established the utility of the AgRSR mouse in characterizing CD8+ T cell behaviours in tumours, permitting us to identify clonotype-independent invariant properties of T cell dysfunction. From the immunotherapeutic perspective, it can be concluded that a focus on rejuvenating the functional clones in the secondary lymphoid tissues may overcome current limitations in immunotherapies and cancer treatment. More broadly, the use of AgRSR mouse as a tool can pave the way for further interrogation of clonal principles towards other disease settings, fundamental principles regulating clone-dependent, and T-cell intrinsic behaviours can be distinguished.





Thaventhiran, James


mouse model


Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge
MRC (2385240)
MRC Toxicology Unit CRUK Cambridge Institute