Modelling the uterine circulation to understand the causes and consequences of inadequate spiral artery remodelling

Abstract

The placenta is a fetal exchange organ that transfers nutrients and oxygen from mother to baby during pregnancy. Establishment of an adequate blood supply to the placenta is critical for pregnancy success. To do this, specialised placental cells, called trophoblasts, invade into the mother’s uterus and act to remodel the mother’s most distal uterine blood vessels, the spiral arteries (SAs). In the first trimester, the trophoblasts have been shown to at least partially occlude the SAs, restricting oxygenated blood flow to the placenta until approximately 10 weeks of gestation. In doing so, they create a physiologically normal low oxygen environment that is key for adequate placental developments but also significantly change the haemodynamic properties within the SA. Timing for SA unblocking by trophoblasts is important but it is not clear how it occurs. Trophoblasts also migrate along the SA length (against blood flow) and remodel them from tight muscular spirals into wide non vasoactive conduits that maximise volumetric blood flow for the remainder of pregnancy. Thus, beyond the first trimester, the volume of maternal blood delivered to the placenta is high, but its velocity is relatively low, avoiding damage to delicate placental tissue and maximising exchange between the maternal and fetal circulations. Disruptions to this remodelling process can lead to poor pregnancy outcomes, but the impact of SA remodelling on local blood flow rates, shear stress and subsequent SA remodelling is difficult to observe in vivo and largely unknown. To find out more about this delicate process of SA plugging by trophoblasts and its remodelling, more investigations in vivo and/or in vitro is required but doing that is also limited by obvious ethical constraints. Therefore, in silico approaches are more suitable to address these challenges. The aim of this thesis was to develop computational models of the utero-placental circulation throughout pregnancy and to link these models to trophoblast migration behaviours. First, we consider the first trimester of pregnancy, when the SA is plugged by trophoblasts. We present an analytically solvable model of the plugged SA and an alternative pathway for uterine blood flow, arteriovenous anastomoses. This model shows the existence of a physiological trophoblast plug prevents blood flow into the IVS, and decreases shear stress on the vessel wall upstream of the plug to a level that would generate permissible conditions for trophoblast migration and SA remodelling. The model provides estimates for the range of flow velocities through the SA in early pregnancy (which are not consistently observable by ultrasound) and the shear stress sensed by trophoblasts that are migrating in the SA. We then analyse the impact of fluid flow on trophoblast migration, by analysing an in vitro dataset of trophoblast migration in micro-fluidic channels in the context of the model of the plugged SA, with the aim of incorporating these data into an agent based model of trophoblast migration in the first trimester. This novel cell-based modelling framework was developed to simulate trophoblast migration within a SA under known environmental stimuli including blood flow and chemotaxis. Although chemotaxis sources and concentrations in the SA are still not well defined, we are able to parameterise the agent based model to data showing trophoblast responses to shear stress, and then to assess the impact of varying unknown parameters relating to chemotaxis in the model. The model suggests that chemotaxis and cell-cell interaction forces have a significant effect on the nature of trophoblast migration and plug dislodgment. We show that the plug could potentially dislodge due to cell migration from the plug to the vessel wall, or due to the formation of channels of high flow in the plug which occur due to asymmetries in the system or high flow velocity regions. Finally, we move beyond the first trimester and present a computational model of blood flow from remodelled SA openings into the intervillous space. This model shows that the structure of the placental tissue beyond the SA significantly impacts on ultrasound measured "jets" of flow. The model predicts that jets of flow observed by ultrasound are likely correlated with increased tissue porosity near the SA mouth and is proposed that observed mega-jets (flow penetrating more than half the placental thickness) are only possible when SAs open to regions of the placenta with very sparse tissue structures. Therefore, it is postulated that placental tissue density must decrease at the SA mouth through gestation, supporting the hypothesis that blood flow from SAs influences placental development. Computational models developed in this study provide a new approach to understanding the process of trophoblast migration behaviours, SA remodelling and implications of inadequate remodelling. The presented modelling framework in this thesis provides a platform which can be customised to incorporate more realistic structural and functional data on the utero-placental circulation as it becomes available.