Locus specific DNA methylation in human immunological responses

Abstract

All cells in the human body contains the same genome. Yet, there are hundreds of different cell types, with widely different phenotype, and function. The differential gene expression leading to this diversity is tightly regulated by epigenetics, i.e. modifications to the genome without interfering with the sequence. There are a number of epigenetic modifications. In this thesis, I have studied DNA methylation as a readout for lineage specificity or effector function in cells from the immune system, in clinical settings.

In paper I we develop a method based on DNA methylation analysis to determine the lineage commitment of the CD4+ T helper cells. First, we establish a regulatory site in the promotor of the IL17A gene. Next, we use signature genes IFNG, IL13, IL17A and FOXP3, to determine lineage commitment towards corresponding T helper cell subsets Th1, Th2, Th17 and Tregs. We call this method EILA (Epigenetic immune lineage assay), and demonstrate that it is usable in clinical samples from rheumatoid arthritis and multiple sclerosis.

In paper II, we take advantage of the same CD4+ T cell specific loci to investigate the adaptive immune response in and around the tumour microenvironment in specimens derived from patients with urinary bladder cancer (UBC). By sorting and examining CD4+ T cells from blood, tumour and regional lymph nodes, we conclude that patients with higher degree of lineage commitment have a better prognosis. Furthermore, we demonstrate that patients responding to neoadjuvant chemotherapy have a larger proportion of commitment cells, post treatment, compared with the non-responders.

In paper III we further examine the same patient material from UBC patients, but instead focus on the cytotoxic features of CD8+ T lymphocytes. We establish a methylation pattern in the perforin gene PRF1 predictive for protein expression and deploy this locus as a readout for cytotoxic functionality. We demonstrate that the tumour infiltrating CD8+ T cells are pre- dispositioned to be cytotoxic through PRF1 demethylation, but that they lack corresponding protein expression, and show signs of exhaustion. The cells demonstrating a TRM phenotype, can still be woken anew, upon in vitro re-stimulation, demonstrating that they are not terminally exhausted.

In paper IV we investigate whole blood leucocytes and the DNA methylation status of the glucocorticoid gene NR3C1. In contrast to literature studies on healthy volunteers, our cohort of surgical patients demonstrate a homogeneous pattern of demethylation in the previously described CpG island of NR3C1. As opposed to our hypothesis, we found no correlation between methylation and clinical outcome post-surgery in this patient cohort. Nevertheless, when employing multifactorial analysis to investigating the impact of genotype we found four single nucleotide polymorphisms that influenced the outcome.

In conclusion, this thesis demonstrate multiple ways in which DNA methylation analysis can be used to read the immune system, but also that the loci selected for investigation has to be carefully chosen following thorough functional investigations. The results presented herein can contribute to further development of treatments to a variety of clinical conditions.