Hoare, RL;
(2016)
Modelling Immune Reconstitution following Paediatric Haematopoietic Stem Cell Transplantation and in HIV-Infected Children.
Doctoral thesis , UCL (University College London).
Preview |
Text
Thesis_final_RLHoare.pdf - Accepted Version Download (14MB) | Preview |
Abstract
Mechanistic mathematical modelling can be used to understand the fundamental drivers of the immune system and how the system is affected by medical interventions. Key to this understanding in children is the interplay between age and treatment-related effects. This thesis focusses on immune reconstitution following paediatric haematopoietic stem cell transplantation (HSCT) and following the start of antiretroviral therapy (ART) in children infected with human immunodeficiency virus (HIV). Since quantitative reconstitution is only one aspect of immune function, in the final chapter I develop a model to explore the dynamics of T cell receptor diversity. Following HSCT, reconstitution of neutrophils and platelets was modelled using a previous mechanistic model. For CD4 T cell reconstitution, a novel mechanistic model was constructed that included age-related changes in T cell dynamics, the delay to thymic output after HSCT and competition for resources. In HIV-infected children starting ART, a simplified previous model for CD4 T cell and HIV dynamics was adapted to include mechanistic elements for multi-phasic viral load decline, age-related changes in T cell dynamics and competition for resources. Using nonlinear mixed-effects modelling with these deterministic models allowed parameters to be estimated with the uneven and often sparse data available. The models were then used to find factors that affect reconstitution. The model for CD4 reconstitution following HSCT was then used to make verifiable predictions of reconstitution in a new cohort of paediatric patients. T cell receptor diversity dynamics were investigated with a stochastic model in which all T cells compete equally for a global resource. The model was simple enough that numerical simulations could be performed with large numbers of cells and clonotypes, and the model could be characterised analytically. Equations were obtained for long-term mean T cell numbers, clonotype numbers, clonotype size distributions and the Gini coefficient as a measure of dispersion. The model was then extended to model host-donor CD8 memory T cell dynamics in bone marrow transplanted mice, showing that biologically simple assumptions could explain the observed dynamics.
Archive Staff Only
 |
View Item |
