Abstract:
The overall aim of this study was to develop a computational model of uterine natural killer (uNK) cells’ intracellular signalling to predict the response to placental cells (trophoblasts) during pregnancy. Pregnancy success is critically dependent on the remodelling of maternal uterine arteries (spiral arteries) by trophoblasts. There is strong evidence that this process is modulated by immune cells, specifically uNK cells. Natural killer (NK) cells’ usual action on aberrant cells by the secretion of cytotoxins is locally inhibited during pregnancy to promote trophoblast implantation. The mother’s ’tolerant’ response to invading trophoblasts may be the key to a successful pregnancy, but this immune tolerance is incompletely understood. We hypothesize that differential expression of key activation and inhibition receptors in uNK cells compared with peripheral natural killer cells (pbNK) may play an important role in this process. Receptors on uNK cells are stimulated by human leukocyte antigen (HLA) class 1b genes expressed on trophoblasts’ membrane surfaces. The unique class of HLA expression on trophoblasts may help trophoblasts to actively avoid immune attack. In this work, we predicted the contribution of NK cell receptors and intracellular molecules to cytokine secretions. This study uses mathematical modelling to investigate the downstream effects of NK cells’ activation and inhibition receptors after stimulation by key ligands on uNK cells. Development of our model is based on a comprehensive qualitative description of pbNKs’ intracellular signalling pathways leading to chemokine and cytotoxin secretion, obtained from the KEGG database (https://www.genome.jp/kegg-bin/show_pathway?hsa04650). From this qualitative description we built a novel quantitative model, reusing existing curated models where possible and implementing new models as needed. This approach was enabled through the availability of these models in standard formats in publicly accessible repositories. This is the first model to incorporate a comprehensive description of the pathways between activation and inhibition receptors and the secretion of IFNγ, TNFα and GM-CSF cytokines, which aims to predict pbNK function with the necessary detail to match existing literature data. We then assessed the behaviours of the pbNK model derived, to investigate whether alternate solutions that matched pbNK function but were not ’best fits’ exist, and whether changes in relative receptor expression between pbNK and uNK cells may be sufficient to change cell behaviour as expected in pregnancy. While the model was not able to be fully parameterised to uNK cell function, the model assessment reveals the relative importance of different reactions for molecules in the NK intracellular pathways on secretion of IFNγ, TNFα and GM-CSF cytokines for both types of NK cells. This model provides important steps toward understanding how differential receptor expression between these two types of NK cells impacts their function. In particular, we demonstrate model predictions of how the different levels of ligands (HLA-C, -E, and -G) and associated receptors’ expression of uNKs contribute to altered cytokine secretion in uNKs. Ultimately, this new knowledge will help us to define the interaction between placental trophoblasts and uNKs in humans.