Dynamical Modeling of Preeclampsia
Motivation
As a serious complication of pregnancy, preeclampsia, affects 7% of all pregnancies and is a major source of maternal and fetal morbidity and mortality. Worldwide, PE is responsible for approximately 75,000 maternal deaths and 500,000 infant deaths annually (WHO and UNICEF, The Eclampsia Trial Collaborative Group). In addition, 25% of babies born to mothers with PE are growth restricted and 33% are premature; and PE is responsible for 8% of all preterm births and 20% of admissions to the neonatal intensive care unit (Ananth, et al). Further, women who have suffered PE are at increased risk of developing stroke, cardiovascular disease, and diabetes in later life (Triche, et al), and the offspring of PE pregnancies are developmentally programmed for cardiovascular, metabolic, and neurodevelopmental disorders.
Despite decades of investigation and clinical trials we cannot predict or prevent preeclampsia, nor can we treat preeclampsia other than by delivery of the placenta and baby. There have been several multicenter prediction studies measuring a range of first trimester biomarkers (Kenny, et al; Myatt, et al; Myatt, et al) , but they have failed to yield clinical utility. The consensus reached by researchers and clinicians is that preeclampsia is a heterogeneous condition variably involving different organ systems that no single treatment can prevent and for which no single biomarker or combination of biomarkers can predict the different phenotypes that display the syndrome. The consensus has been reinforced by the recent transcriptomic analysis of several differing placental phenotypes of women who are diagnosed with preeclampsia (Leavey, et al; Leavey, et al; Than, et al). There exists then a need for novel approach to decipher the heterogeneity, identify the underlying pathways and use that data to define individualized therapeutic approaches.
Dynamical modeling
Mathematical modeling plays an increasingly important role in understanding the behavior and the mechanisms governing diverse and complex biological systems. Dynamical systems modeling uses differential equations to describe the behavior of a system as its interacting components change with time. In contrast to statistical modeling, dynamical modeling allows computation of precise time trajectories for all model variables. When the mechanisms governing development of a system are not well understood, a dynamical model can be useful to test various hypotheses about those mechanisms by providing predictions following from a set of hypotheses. Once a satisfactory dynamical model is created, interventions can be simulated in silico, including those not tested in vivo, which is ideal for identifying novel candidate therapeutic targets. Simulation of various treatment protocols relating to the same therapeutic agent as well as simulation of combined effect of multiple therapeutic agents administered together can also be done.
A dynamic modeling framework has been applied to research various types of cancer, diabetes, arthritis, stroke, cardiovascular, metabolic, hematologic, autoimmune, neurodegenerative, and even ophthalmological diseases. As a clear sign of acceptance of dynamical modeling, there are multiple cases where the FDA accepted dynamical models as a substitute for pre-clinical animal testing of new treatment strategies, for example: UVA/PADOVA Type 1 Diabetes Simulator (Man, et al) and Cooper Discovery2 HIV simulator. Since one major obstacle in introducing novel pharmaceutical interventions to improve pregnancy outcomes is based on the general fear of inflicting potential harm, particularly to the fetus (Ilekis, et al), clinical trial simulations (aka in silico trials) can be especially valuable in the search for efficient preeclampsia treatment or prevention. Dynamical modeling is especially well suited for developing personalized treatment strategies in the context of heterogeneous conditions because it allows for systematic investigation of individual and combined effects of patient characteristics, environmental conditions, and the effects of various interventions on the course of the disease.
Another advantage of the dynamical systems is that they allow modeling of the feedback loops that are essential to maintain homeostasis, and both normal and disrupted homeostasis can be modeled. With one noticeable exception (Brubaker, et al) where interactions between progesterone receptor transcriptional activity and inflammation in prediction of the contractile status of human myometrium at parturition have been modeled, pregnancy complications have not yet been studied by means of dynamical modeling. We believe there are multiple reasons why modeling of pregnancy complications is lagging behind modeling efforts dedicated to other disorders. One of them is complexity: different maternal organ systems impact pregnancy development and a comprehensive model should account for at least major interactions between these systems. Therefore, a variety of quantitative datasets relating to various aspects of pregnancy will be used to formulate and calibrate a successful model.
Longitudinal studies
Here are the links to various studies and review papers dedicated to preeclampsia.
Sovio, U., et al., 4-Hydroxyglutamate is a novel predictor of pre-eclampsia. International journal of epidemiology, 2020. 49(1): p. 301-311.
Sovio, U., et al., Screening for fetal growth restriction with universal third trimester ultrasonography in nulliparous women in the Pregnancy Outcome Prediction (POP) study: a prospective cohort study. Lancet (London, England), 2015. 386(10008): p. 2089-2097.
Erez, O., et al., The prediction of late-onset preeclampsia: Results from a longitudinal proteomics study. PLoS One, 2017. 12(7): p. e0181468.
Tarca, A.L., et al., The prediction of early preeclampsia: Results from a longitudinal proteomics study. PLoS One, 2019. 14(6): p. e0217273.
Konrad, E., et al., Correlation of elevated levels of lipoprotein(a), high-density lipoprotein and low-density lipoprotein with severity of preeclampsia: a prospective longitudinal study. Journal of Obstetrics and Gynaecology, 2020. 40(1): p. 53-58.
Tarca, A.L., et al., Maternal whole blood mRNA signatures identify women at risk of early preeclampsia: a longitudinal study. The Journal of Maternal-Fetal & Neonatal Medicine, 2020: p. 1-12.
Murtoniemi, K., et al., Longitudinal changes in plasma hemopexin and alpha-1-microglobulin concentrations in women with and without clinical risk factors for pre-eclampsia. PLOS ONE, 2019. 14(12): p. e0226520.
Pasyar, S., et al., Investigating the diagnostic capacity of uric acid in the occurrence of preeclampsia. Pregnancy Hypertension, 2020. 19: p. 106-111.
Evers, K.S., et al., Neurofilament as Neuronal Injury Blood Marker in Preeclampsia. Hypertension, 2018. 71(6): p. 1178-1184.
Wikström, A.-K., et al., Plasma levels of S100B during pregnancy in women developing pre-eclampsia. Pregnancy Hypertension: An International Journal of Women's Cardiovascular Health, 2012. 2(4): p. 398-402.
Bergman, L. and H. Åkerud, Plasma Levels of the Cerebral Biomarker, Neuron-Specific Enolase, are Elevated During Pregnancy in Women Developing Preeclampsia. Reproductive Sciences, 2016; 23(3): p. 395-400.
Bergman, L., et al., Blood-based cerebral biomarkers in preeclampsia: Plasma concentrations of NfL, tau, S100B and NSE during pregnancy in women who later develop preeclampsia - A nested case control study. PLOS ONE, 2018. 13(5): p. e0196025.
Kelly, C.B., et al., Circulating adipokines are associated with pre-eclampsia in women with type 1 diabetes. Diabetologia, 2017. 60(12): p. 2514-2524.
Hinkle, S.N., et al., Maternal adipokines longitudinally measured across pregnancy and their associations with neonatal size, length, and adiposity. International Journal of Obesity, 2019; 43(7): p. 1422-1434.
Zerón, H.M., et al. Hyperleptinemia as a prognostic Factor For preeclampsia: a cohort study. Acta Medica (Hradec Kralove), 2012; 55(4): p. 165-71.
Romero, R., et al., Maternal plasma-soluble ST2 concentrations are elevated prior to the development of early and late onset preeclampsia – a longitudinal study. The Journal of Maternal-Fetal & Neonatal Medicine, 2018. 31(4): p. 418-432.
Brosnihan, K.B., et al., Longitudinal study of angiotensin peptides in normal and pre-eclamptic pregnancy. Endocrine, 2020. 69(2): p. 410-419
Khalil, A., et al., Longitudinal changes in maternal serum placental growth factor and soluble fms-like tyrosine kinase-1 in women at increased risk of pre-eclampsia. Ultrasound in Obstetrics & Gynecology, 2016. 47(3): p. 324-331.
Zhu, J., et al., Angiogenic factors during pregnancy in Asian women with elevated blood pressure in early pregnancy and the risk of preeclampsia: a longitudinal cohort study. BMJ Open, 2019. 9(11): p. e032237.
Herraiz, I., et al., Longitudinal change of sFlt-1/PlGF ratio in singleton pregnancy with early-onset fetal growth restriction. Ultrasound in Obstetrics & Gynecology, 2018. 52(5): p. 631-638.
Khalil, A., et al., Longitudinal changes in maternal soluble endoglin and angiopoietin-2 in women at risk for pre-eclampsia. Ultrasound in Obstetrics & Gynecology, 2014. 44(4): p. 402-410.
Moutquin, J.-M., et al., Do prostacyclin and thromboxane contribute to the "Protective Effect" of pregnancies with chronic hypertension? A preliminary prospective longitudinal study. American Journal of Obstetrics & Gynecology, 1997. 177(6): p. 1483-1490.
Agudelo-Zapata, Y., et al., Serum 25-hydroxyvitamin D levels throughout pregnancy: a longitudinal study in healthy and preeclamptic pregnant women. Endocrine connections, 2018. 7(5): p. 698-707.
Mayer-Pickel, K., et al., Effect of Low-Dose Aspirin on Soluble FMS-Like Tyrosine Kinase 1/Placental Growth Factor (sFlt-1/PlGF Ratio) in Pregnancies at High Risk for the Development of Preeclampsia. Journal of Clinical Medicine, 2019. 8(9): p. 1429.