Final Report Summary - QUANTI (Quantitative T cell Immunology)
With the enthusiastic participation of all academic and industrial partners and associated partners, the network recruited and trained fourteen fellows. Training, management and scientific events were hosted by academic and industrial partners in the UK, Ireland and Germany. The inaugural meeting was hosted by NUI Maynooth, the midterm review meeting by DKFZ Heidelberg and the closing meeting by the University of Leeds. Complementary Skills Training events and Summer schools were hosted by NUI Galway, University of Leeds, Bayer, Unilever and Microsoft Research Cambridge.
The research in the different work packages (WPs) was strengthened by secondments, including industrial placements at Unilever and Astra-Zeneca and academic placements at MIT and Imperial College. The Network benefited from the participation of scientists and entrepreneurs from Associated Partners, such as Phil Hodgkin, Jose Faro and Paul Jonsen, invited speakers such as Arup Chakraborty, Lucy Walker, Mark Coles, Sarah Teichmann and Hyun Park, and ad hoc training from Martin Meier-Schellersheim, Peter Hunter and Mario Castro. Fellows have benefitted from specialist summer schools, such as Q-bio in 2016 and participated in the organisation of dedicated sessions at international conferences.
WP1 was a collaboration between UCMU (Utrecht), Imperial College London and Bayer (Germany). Work on comparing different stable-isotope labelling techniques has been completed and published by UMCU and Imperial. Deuterium-labelling experiments in young and old goats have been performed by Mariona Baliu Pique, to be able to compare T-cell dynamics at different sites in the body, including the bone marrow, blood and lymph nodes, and to study how these rates of turnover change during healthy ageing. After establishment and application of a physiologically based pharmacokinetic (PBPK) model, deuterium labelling data was studied using computational models of the dynamics of stem cell-like memory T cells (TSCM) and CD57+ memory T cells. Analysis of deuterium labelling data together with telomere length data suggested that the average half-life of cell of TSCM cells is about 2 years. The half-life of human neutrophils in blood was analysed using deuterium and glucose labelling data and a computational model representing blood neutrophils as well as mitotic neutrophil precursors. A half-life in blood of 13h to around 19 h was obtained.
The experimental-theoretical partnership in WP2 developed Antonio Freitas’ “quorum-sensing” hypothesis, an indirect feedback loop where IL-2 is produced by a sub-population of CD4 T cells and detected by regulatory T cells. The data from Sary El Daker’s experiments at Institut Pasteur was analysed by Luis de la Higuera and Carmen Molina-París at the University of Leeds, and a mathematical model devised to assess contributions of the constitutive and the ligand-induced synthesis pathways under different signalling hypotheses.
In their work for WP3, Pedro Goncalves and Benedita Rocha (INSERM) have performed an analysis of the diversity of CD8+ TCR repertoire in naive cells and its modification during LCMV immune responses. Mathematical analysis by Marco Ferrarini and Grant Lythe (Leeds) suggests that the majority of TCR clonotypes in a mouse consist of only one cell. The focus of the research at NUI Galway within WP3 was on the development of bioinformatics tools for the analysis of TCR repertoires and repertoire diversity from high-throughput sequence data and the analysis of data generated by experimental collaborators on this work package. They developed LymAnalyzer, a portable software package to map short-read sequences from TCRs to their constituent germline genes and retrieve and cluster CDR3 sequences. As a consequence of the development of LymAnalyzer, it became clear that the standard Cathal Seoighe and Yaxuan Yu made a database (called Lym1K) of novel putative alleles recovered in this way available to the research community.
Naive T cells differentiate into short-lived effectors and memory T cells upon antigen stimulation. The timing and mechanisms underlying the differentiation of naive cells into effector and memory T cells is not yet clear. Within WP4, Rob de Boer and Aridaman Pandit developed mathematical models of CD8+ T cell fate commitment, performed Gillespie simulations to compare the variance observed in real datasets with that obtained using different proposed mechanisms. They also developed a branching process model to address the stochasticity demonstrated in the experiments published by NKI, Amsterdam and other groups. In collaboration with Christoph Niederalt at Bayer, a mechanistic computational model was developed to describe the T cell response during blockade of a checkpoint inhibitor, with the PD-1/PD-L1 system chosen as example. The model describes the dynamics of active and inactive T cells depending on the PD-L1 expression on tumour cells and the dynamics of PD-L1 positive tumour cells depending on IFN-gamma secretion.
The 2016 paper "T cell stimuli independently sum to regulate an inherited clonal division fate" revealed basic biological facets of clonal cellular proliferation programming, and essential elements of how stimuli are integrated. These findings have ramifications for our interrogation of adaptive immune response data and our quantitative understanding of immune intervention. The methodology introduced, which allows high through-put clonal division tracking even for adherent cells, has broader application and is in the process of being adopted for cell systems beyond immunology by other labs. With Ken Duffy at NUI Maynooth, Harry Tideswell and Giulio Prevedello have contributed significantly to the design and development of the novel mathematical and statistical framework required to analyse primary data from a new clonal multiplex experimental system created in collaboration with the QuanTI associated partner, the Hodgkin lab at the Walter and Elisa Hall Institute of Medical Research.
Melania Barile, working in WP6 at the Thomas Höfer group at DKFZ Heidelberg, has developed a deterministic model for cell population dynamics in order to quantitatively characterise adult hematopoiesis under steady state and perturbed conditions upon combining in vivo time course data and deterministic mathematical modelling. The goal is to unravel the molecular mechanisms behind individual cytokine memory of cytotoxic and of other subsets of helper T-cells. Our cytokine memory study reveals that quantitative heterogeneity of IFN-ɣ expression in T cell populations do not rise from random noise of gene expression, but individual cells commit to and retain distinct levels of cytokine gene expression.
Grant Lythe, Network Coordinator
quanti@maths.leeds.ac.uk
http://www1.maths.leeds.ac.uk/applied/QuanTI/
The research in the different work packages (WPs) was strengthened by secondments, including industrial placements at Unilever and Astra-Zeneca and academic placements at MIT and Imperial College. The Network benefited from the participation of scientists and entrepreneurs from Associated Partners, such as Phil Hodgkin, Jose Faro and Paul Jonsen, invited speakers such as Arup Chakraborty, Lucy Walker, Mark Coles, Sarah Teichmann and Hyun Park, and ad hoc training from Martin Meier-Schellersheim, Peter Hunter and Mario Castro. Fellows have benefitted from specialist summer schools, such as Q-bio in 2016 and participated in the organisation of dedicated sessions at international conferences.
WP1 was a collaboration between UCMU (Utrecht), Imperial College London and Bayer (Germany). Work on comparing different stable-isotope labelling techniques has been completed and published by UMCU and Imperial. Deuterium-labelling experiments in young and old goats have been performed by Mariona Baliu Pique, to be able to compare T-cell dynamics at different sites in the body, including the bone marrow, blood and lymph nodes, and to study how these rates of turnover change during healthy ageing. After establishment and application of a physiologically based pharmacokinetic (PBPK) model, deuterium labelling data was studied using computational models of the dynamics of stem cell-like memory T cells (TSCM) and CD57+ memory T cells. Analysis of deuterium labelling data together with telomere length data suggested that the average half-life of cell of TSCM cells is about 2 years. The half-life of human neutrophils in blood was analysed using deuterium and glucose labelling data and a computational model representing blood neutrophils as well as mitotic neutrophil precursors. A half-life in blood of 13h to around 19 h was obtained.
The experimental-theoretical partnership in WP2 developed Antonio Freitas’ “quorum-sensing” hypothesis, an indirect feedback loop where IL-2 is produced by a sub-population of CD4 T cells and detected by regulatory T cells. The data from Sary El Daker’s experiments at Institut Pasteur was analysed by Luis de la Higuera and Carmen Molina-París at the University of Leeds, and a mathematical model devised to assess contributions of the constitutive and the ligand-induced synthesis pathways under different signalling hypotheses.
In their work for WP3, Pedro Goncalves and Benedita Rocha (INSERM) have performed an analysis of the diversity of CD8+ TCR repertoire in naive cells and its modification during LCMV immune responses. Mathematical analysis by Marco Ferrarini and Grant Lythe (Leeds) suggests that the majority of TCR clonotypes in a mouse consist of only one cell. The focus of the research at NUI Galway within WP3 was on the development of bioinformatics tools for the analysis of TCR repertoires and repertoire diversity from high-throughput sequence data and the analysis of data generated by experimental collaborators on this work package. They developed LymAnalyzer, a portable software package to map short-read sequences from TCRs to their constituent germline genes and retrieve and cluster CDR3 sequences. As a consequence of the development of LymAnalyzer, it became clear that the standard Cathal Seoighe and Yaxuan Yu made a database (called Lym1K) of novel putative alleles recovered in this way available to the research community.
Naive T cells differentiate into short-lived effectors and memory T cells upon antigen stimulation. The timing and mechanisms underlying the differentiation of naive cells into effector and memory T cells is not yet clear. Within WP4, Rob de Boer and Aridaman Pandit developed mathematical models of CD8+ T cell fate commitment, performed Gillespie simulations to compare the variance observed in real datasets with that obtained using different proposed mechanisms. They also developed a branching process model to address the stochasticity demonstrated in the experiments published by NKI, Amsterdam and other groups. In collaboration with Christoph Niederalt at Bayer, a mechanistic computational model was developed to describe the T cell response during blockade of a checkpoint inhibitor, with the PD-1/PD-L1 system chosen as example. The model describes the dynamics of active and inactive T cells depending on the PD-L1 expression on tumour cells and the dynamics of PD-L1 positive tumour cells depending on IFN-gamma secretion.
The 2016 paper "T cell stimuli independently sum to regulate an inherited clonal division fate" revealed basic biological facets of clonal cellular proliferation programming, and essential elements of how stimuli are integrated. These findings have ramifications for our interrogation of adaptive immune response data and our quantitative understanding of immune intervention. The methodology introduced, which allows high through-put clonal division tracking even for adherent cells, has broader application and is in the process of being adopted for cell systems beyond immunology by other labs. With Ken Duffy at NUI Maynooth, Harry Tideswell and Giulio Prevedello have contributed significantly to the design and development of the novel mathematical and statistical framework required to analyse primary data from a new clonal multiplex experimental system created in collaboration with the QuanTI associated partner, the Hodgkin lab at the Walter and Elisa Hall Institute of Medical Research.
Melania Barile, working in WP6 at the Thomas Höfer group at DKFZ Heidelberg, has developed a deterministic model for cell population dynamics in order to quantitatively characterise adult hematopoiesis under steady state and perturbed conditions upon combining in vivo time course data and deterministic mathematical modelling. The goal is to unravel the molecular mechanisms behind individual cytokine memory of cytotoxic and of other subsets of helper T-cells. Our cytokine memory study reveals that quantitative heterogeneity of IFN-ɣ expression in T cell populations do not rise from random noise of gene expression, but individual cells commit to and retain distinct levels of cytokine gene expression.
Grant Lythe, Network Coordinator
quanti@maths.leeds.ac.uk
http://www1.maths.leeds.ac.uk/applied/QuanTI/