Multi-omic time-series analysis of T-cells as a model for identification of biomarkers, treatments and upstream disease regulators

Sammanfattning: CD4+ T-cell function and their process of differentiation is a central piece of the puzzle in a multitude of diseases. CD4+ T-cells are part of the adaptive immune system and function by directing other immune cells to the site of infection and instructing B-cells to produce antibodies, among many other functions. CD4+ T-cells may differentiate into several different sub-types, such as T-helper 1, 2 and 17, with differing functions within the immune system. T-helper 1 (Th1) cells are most closely associated with the elimination of viral infections but are also associated with autoimmune diseases such as multiple sclerosis (MS) and rheumatoid arthritis (RA). T-cells develop in the thymus first as double-negative T-cells, that express neither CD4 nor CD8, going through multiple development stages before becoming double-positive T-cell that express both CD4 and CD8, before eventually giving rise to single positive CD4+ and CD8+ T-cells. This process of development is under tight control and if this control fails, cancer may result. Once CD4+ T-cells are fully developed, they may specialize as outlined above and if said process is not properly controlled, autoimmunity may result. As such, the proper understanding of these control mechanisms is of great importance for the understanding of diseases of the immune system and the discovery of biomarkers and treatments against said diseases. These control processes are often studied in a singular fashion using one omic technique, e.g., RNA sequencing (RNA-seq), with the assumption that a signal in one omic layer will be reflected in another. Recent studies attempting to integrate multiple omics have however cast doubt on this and it is becoming increasingly apparent that to gain a complete understanding of a system, the system needs to be studied at multiple levels of regulation, i.e., multiple omics.The aim of this thesis was to use multi-omics to investigate the development and differentiation process of CD4+ T-helper cells and relate it to disease mechanisms. To start, we studied T-cell development through the model of T-cell acute lymphoblastic leukaemia (T-ALL). More specifically, we studied the TET2 gene and investigated its importance in T-ALL for treatment susceptibility and mechanism in vitro. TET2 is a demethylase and functions through the removal of cytosine methylation on the DNA, a marker of gene silencing. Through treatment with decitabine, an inhibitor of DNA-methylation, and Vitamin C, a co-factor for TET2, we showed that TET2 deficient cancer cell lines were more vulnerable to treatment targeting DNA methylation and investigated the mechanistic effects of said treatment by RNA sequencing. We then moved on to study primary human naïve CD4+ T-cells and their differentiation into Th1-cells. First, we focused on T-cell activation and its importance to MS to understand the role of T-cells in mediating the lowered disease activity usually observed during pregnancy in MS. This showed that the major pregnancy hormone progesterone significantly dampens T-cell activation, providing a possible explanation for the beneficial effects of pregnancy on MS. Then, using ATAC sequencing (ATAC-seq), RNA-seq and proteomics we studied Th1-differentiation as a time series to elucidate regulatory events throughout the differentiation process and to study their implications for MS with the inclusion of progesterone treatment.  The integration of several omic techniques presents unique challenges as one does not necessarily directly translate to the other. As such, we first focused on the integration of RNA-seq and proteomics by designing a model for the prediction of protein abundance from RNA-seq and validated it through biomarker discovery. Next, we focused on the integration of ATAC-seq and RNA-seq using correlation between time series of the two techniques. This thesis provides a thorough investigation of Th1-cell differentiation and its potential involvement in disease. Time series datasets were produced to study gene regulation (ATAC-seq), gene expression (RNA-seq) and protein expression (mass spectrometry) and the work focused on their integration. This profoundly showed that through combining multiple omic techniques it was possible to gain new insights that were not possible to discover with one or the other. Multi-omic analyses are becoming more and more common in medicine today as their power to produce new insight into the complexity of complex diseases is being increasingly recognized. As such, this work forms an important foundation for future discovery of biomarkers and treatments in such diseases.