New Technologies and Perinatal Medicine Prediction and Prevention of Pregnancy Complications Series in Maternal-Fetal Medicine About the Series Published in association with the Journal of Maternal Fetal and Neonatal Medicine, the series in Maternal Fetal Medicine keeps readers up to date with the latest clinical therapies to improve the health of pregnant patients and ensure a successful birth. Each volume in the series is prepared separately and typically focuses on a topical theme. Volumes are published on an occasional basis, depending on the emergence of new developments. Textbook of Diabetes and Pregnancy, Third Edition Moshe Hod, Lois G. Jovanovic, Gian Carlo Di Renzo, Alberto De Leiva, Oded Langer Cesarean Delivery: A Comprehensive Illustrated Practical Guide Gian Carlo Di Renzo, Antonio Malvasi Obstetric Evidence Based Guidelines, Third Edition Vincenzo Berghella Maternal-Fetal Evidence Based Guidelines, Third Edition Vincenzo Berghella Maternal-Fetal and Obstetric Evidence Based Guidelines, Two Volume Set, Third Edition Vincenzo Berghella The Long-Term Impact of Medical Complications in Pregnancy: A Window into Maternal and Fetal Future Health Eyal Sheiner Operative Obstetrics, 4E Joseph J. Apuzzio, Anthony M. Vintzileos, Vincenzo Berghella, Jesus R. Alvarez-Perez Placenta Accreta Syndrome Robert M. Silver Neurology and Pregnancy: Clinical Management Michael S. Marsh, Lina Nashef, Peter Brex Fetal Cardiology: Embryology, Genetics, Physiology, Echocardiographic Evaluation, Diagnosis, and Perinatal Management of Cardiac Diseases, Third Edition Simcha Yagel, Norman H. Silverman, Ulrich Gembruch New Technologies and Perinatal Medicine: Prediction and Prevention of Pregnancy Complications Moshe Hod, Vincenzo Berghella, Mary E. D’Alton, Gian Carlo Di Renzo, Eduard Gratacós, Vassilios Fanos For more information about this series please visit: https://www.crcpress.com/Series-in-Maternal-Fetal-Medicine/ book-series/CRCSERMATFET New Technologies and Perinatal Medicine Prediction and Prevention of Pregnancy Complications Edited by Moshe Hod md Director, Mor Comprehensive Women’s Health Care Center Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel Vincenzo Berghella md facog Director, Division of Maternal-Fetal Medicine Professor, Department of Obstetrics and Gynecology Sidney Kimmel Medical College of Thomas Jefferson University Philadelphia, Pennsylvania, USA Mary E. D’Alton md mb bs Chair, Department of Obstetrics & Gynecology Willard C. Rappleye Professor of Obstetrics & Gynecology Columbia University Irving Medical Center Director of Services, Sloane Hospital for Women New York-Presbyterian, New York City, New York, USA Gian Carlo Di Renzo md phd Professor and Chairman, Department of Obstetrics and Gynecology Director, Perinatal and Reproductive Medicine Center and Midwifery School, University Hospital, Perugia, Italy Director, Permanent International and European School of Perinatal and Reproductive Medicine (PREIS), Florence, Italy Eduard Gratacós md Director and Professor, BCNatal, Hospital Clínic and Hospital Sant Joan de Deu, University of Barcelona, Barcelona, Spain Vassilios Fanos md Professor of Pediatrics Director, Neonatal Intensive Care Unit University of Cagliari, Cagliari, Italy CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2020 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed on acid-free paper International Standard Book Number-13: 978-1-138-70614-9 (Hardback) This book contains information obtained from authentic and highly regarded sources. While all reasonable efforts have been made to publish reliable data and information, neither the author[s] nor the publisher can accept any legal responsibility or liability for any errors or omissions that may be made. The publishers wish to make clear that any views or opinions expressed in this book by individual editors, authors or contributors are personal to them and do not necessarily reflect the views/opinions of the publishers. The information or guidance contained in this book is intended for use by medical, scientific or health-care professionals and is provided strictly as a supplement to the medical or other professional’s own judgement, their knowledge of the patient’s medical history, relevant manufacturer’s instructions and the appropriate best practice guide- lines. Because of the rapid advances in medical science, any information or advice on dosages, procedures or diagnoses should be independently verified. The reader is strongly urged to consult the relevant national drug formulary and the drug companies’ and device or material manufactur- ers’ printed instructions, and their websites, before administering or utilizing any of the drugs, devices or materials mentioned in this book. This book does not indicate whether a particular treatment is appropriate or suitable for a particular individual. Ultimately it is the sole responsibility of the medical professional to make his or her own professional judgements, so as to advise and treat patients appropriately. The authors and publishers have also attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any elec- tronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com Contents Introduction: Why do we need omics and systems biology? vii Contributors xi Section i PReGnAncY coMPLicAtionS: SettinG tHe Scene 1 The mother: Adaptation to pregnancy and normal metabolism 2 Francesca Parisi, Alice Zavatta, Roberta Milazzo, and Irene Cetin 2 Maternal and fetal normal and abnormal nutrition 7 Sarah Louise Killeen, Eilleen C. O’Brien, and Fionnuala M. McAuliffe 3 Great obstetrical syndromes: It’s all in the placenta 12 Martin Gauster and Gernot Desoye 4 Normal and abnormal fetal growth 18 Javier Caradeux, Eduard Gratacós, and Francesc Figueras 5 Preterm labor and birth 23 Vincenzo Berghella and Eduardo da Fonseca 6 Gestational diabetes mellitus 28 Silvia Vannuccini and Federico Mecacci 7 Preeclampsia 34 Jon Hyett and Liona C. Poon 8 Maternal obesity 40 Tahir A. Mahmood and Rohan Chodankar 9 Maternal health: Immediate, short-, and long-term complications following pregnancy 46 Gil Gutvirtz, Omri Zamstein, and Eyal Sheiner 10 The fetus and the neonate: Immediate, short-, and long-term impact 56 Umberto Simeoni and Elie Saliba 11 Cost of pregnancy complications related to noncommunicable diseases and cost effectiveness of interventions to address them 60 Anil Kapur, Jon Hyett, and H. David McIntyre Section ii toWARDS PReDiction AnD PReVention 12 Integrated system biology approaches to fetal medicine problems 74 Jezid Miranda, Fátima Crispi, and Eduard Gratacós 13 Omics and female reproduction 81 Galia Oron 14 Maternal genome and pregnancy outcomes 85 Nagendra K. Monangi, Ge Zhang, Mikko Hallman, Kari Teramo, Bo Jacobsson, and Louis J. Muglia 15 Placental development and omics 90 Sylvie Hauguel-de Mouzon, Gernot Desoye, and Silvija Cvitic 16 Placental metabolomics in obese pregnancies 95 Irene Cetin, Chiara Novielli, and Chiara Mandò 17 Methylome and epigenetic markers 98 Skevi Kyriakou, Marios Ioannides, George Koumbaris, and Philippos Patsalis 18 Microbiome and pregnancy complications 102 Maria Carmen Collado and Omry Koren 19 Small noncoding RNAs as biomarkers for pregnancy complications 107 Liron Yoffe, Meitar Grad, Avital Luba Polsky, Moshe Hod, and Noam Shomron 20 Urine metabolomics and proteomics in prenatal health 112 Daniela Duarte, Maria do Céu Almeida, Pedro Domingues, and Ana M. Gil 21 Metabolomics and perinatal complications 125 Flaminia Bardanzellu, Moshe Hod, and Vassilios Fanos 22 Metabolomics in normal and pathologic pregnancies 133 Antonio Ragusa, Alessandro Svelato, and Sara D’Avino v vi Contents 23 Metabolomics in amniotic fluid 139 Alexandra-Maria Michaelidou, Foteini Tsakoumaki, Maria Fotiou, Charikleia Kyrkou, and Apostolos P. Athanasiadis 24 Omics and coagulation disorders in pregnancy 149 Sara Ornaghi and Michael J. Paidas 25 Omics and perinatal medicine: Preeclampsia 156 Piya Chaemsaithong and Liona C. Poon 26 Single nucleotide polymorphisms and pregnancy complications 172 Federica Tarquini, Giuliana Coata, Elena Picchiassi, and Gian Carlo Di Renzo 27 Metabolomics and perinatal cardiology 174 Roberta Pintus, Angelica Dessì, and Vassilios Fanos 28 Metabolomics and human breast milk: A unique and inimitable food for infants 177 Flamina Cesare Marincola, Sara Corbu, Roberta Pintus, Angelica Dessì, and Vassilios Fanos 29 Neurodevelopment and placental omics 181 Despina D. Briana and Ariadne Malamitsi-Puchner 30 Early life complications, placental genomics, and risk for neurodevelopmental disorders in offspring 185 Pasquale Di Carlo, Giovanna Punzi, and Gianluca Ursini 31 Metabolomics and perinatal asphyxia 192 Ernesto d’Aloja, Emanuela Locci, Antonio Noto, Matteo Nioi, Giovanni Bazzano, and Vassilios Fanos 32 Environment, pregnancy complications, and omics 204 Chen Ben David and Ido Solt 33 Sleep and pregnancy complications 209 Orna Sever and Riva Tauman 34 Maternal plasma cell-free DNA screening: Basic science and applications 214 Peter Benn and Howard Cuckle 35 Maternal plasma cell-free DNA screening: Integration into clinical practice 218 Howard Cuckle and Peter Benn 36 Microarrays 223 Melissa Stosic, Jessica L. Giordano, Brynn Levy, and Ronald Wapner 37 Whole exome and whole genome sequencing 230 Mary E. Norton Index 239 Introduction Why do we need omics and systems biology? Omics implies many different approaches are concurrently In obstetrics, the prevalence of preeclampsia is applied to a seemingly singular problem. The presumption 5%–7%. We know its epidemiology (primigravid is that the plethora of concurrent data will lead to causal women; multiparous women with new partners). We elucidations and treatment of pregnant woman and fetuses diagnosed on the basis of elevated blood pressure, not otherwise possible. Those of us who have practiced and hyperreflexia, edema, and proteinuria. We knew that conducted investigations for decades could be forgiven a woman whose pre-pregnancy blood pressure was for wondering just why this unfamiliar and vastly more 100/70 need not reach 140/90 to be considered to have complex assignment has suddenly become necessary. After preeclampsia. Diagnosis, prevention, and treatment all, we have long been conducting randomized clinical trials have not fundamentally changed for many years— (RCTs), often happy with results, regardless of whether nor has progress. The current elixir—aspirin—might resolution had been reached on an arguable issue. RCTs benefit 20% of women. This means that 80% would funded by the National Institutes of Health (NIH), Medical not be “cured.” Despite this noninfectious condition Research Council (MRC), and other agencies seemed well affecting 5% of the population, we persisted in thinking constructed, took into account potential confounding there was usually only one etiology. Yet, heterogeneity variables that could be adjusted (multivariate analysis), exists; thus, something else is necessary for progress. and delivered crisp conclusions on the primary outcomes. Omics and systems biology become candidates. Why, then, must we current investigators embark on II. All common disorders are etiologically heterogeneous: a much more complex strategy? The simple answer is If almost all noninfectious medical disorders of that the disorders we once believed to be a homogenous incidence 2%–5% are heterogeneous in etiology, what condition are rarely so. These disorders are heterogeneous. is their underlying basis? Illustrative is that medical This volume illustrates well the approaches that are thus geneticists had the advantage of knowing that heritable needed. In turn, we are in need of hypotheses derived in inborn errors of metabolism were heterogeneous. agnostic (“out of the box”) fashion, not just derivative of Mutations in different enzymes in a common pathway our preconceived beliefs. could have the same end result, for example, elevated Let us, then, briefly consider why omics and systems cholesterol or decreased cortisol. This was the result biology will open doors that one could say are locked at of defects in one of the sequential enzymes in a present. single biosynthetic pathway. That each disorder was due to a different gene (etiologic heterogeneity) was I. Common disorders were once considered homogenous evident, even if at least one component of a pleiotropic in etiology: Perusing textbooks written perhaps 30–40 phenotype was similar. Differentiation was also years ago leaves one with the conclusion that most possible as a result of different profiles of metabolites. common perinatal diseases were of a single etiology. A familiar example is phenylketonuria (PKU). PKU Phenotypes of affected patients might differ in degree is an autosomal recessive disorder due to deficiency of of expression, as for example shown by chronological the enzyme phenylalanine hydroxylase (PAH), which age of onset or gestational week of manifestation. is necessary to convert phenylalanine to tyrosine. However, guidelines were fundamentally predicated Diagnosis of PKU was initially made not on the basis on the assumption that a given disorder was singular. of directly measuring excess phenylalanine, but by Systemic hypertension in adults was diagnosed detecting an increase in the by-product metabolite (then and now) if blood pressure was greater than phenylpyruvic acid. Phenylpyruvic acid accumulated 140/90 mm Hg. Simply put, below that value was secondarily to elevated levels of phenylalanine. But normal. True, professional societies frequently even the uncommon disorder PKU is heterogeneous. explored different thresholds. Yes, we knew rare Classical, variant, and benign phenotypes exist, all causes existed, such as Cushing syndrome or adrenal due to different perturbations in the PAH enzyme. biosynthetic disorders. However, these “zebras” Perturbations exist in still other genes—PCR and were considered exceptional, merely to be excluded PHPR. All contribute to clinical PKU. initially for completeness of our differential diagnosis. We in obstetrics-gynecology (OB-GYN) have lacked Most cases would eventually prove to be of the same analogous examples like inborn errors of metabolism. etiology and, hence, treated similarly. Aggressiveness Global conferences are convened to codify diagnostic of therapy (e.g., dosage) might differ, but qualitatively criteria (e.g., Rotterdam criteria), but classifications the strategy would not. tend to be quantitative variants on the theme of vii viii Introduction follicles, hirsutism, and anovulation. Yet over a dozen individuals. Let us assume that any of a set of genes (A, different genes are associated with polycystic ovarian B, C, D, or E) could if perturbed produce a histologically syndrome (PCOS) and characterized by robust identical cancer. Assume further that three different statistical significance (1,2). chemotherapeutic agents confer benefit in patients Fortunately, increased appreciation of genetic and having a tumor of common histologic type. Yet no drug etiological heterogeneity is occurring in many common alone is universally efficacious. Chemotherapeutic obstetrical conditions. Progress in differentiating agents 1 and 3 might be efficacious in certain patients, hyperglycemias is an example. Once, only binary whereas only agent 2 might provide benefit in other stratification into childhood (insulin-dependent) and patients. Higher doses of agent 2 could be delivered if 1 adult-onset (insulin resistant) diabetes mellitus (DM) and 3 are not concurrently administered. The converse existed. This categorization later became disrupted also is applicable. We can predict that analogous with the recognition of maturity onset diabetes of strategies will evolve for detecting and treating youth (MODY). Its age of onset (late teens; early third preeclampsia and preterm birth. decade) was different. Heritability was different (often IV. Precision medicine integrates omics, clinical data, a single gene). Later, MODY began to be stratified environmental exposures, and social determinants: into different types on the basis of genes and mode of Marketing teams often tout an institution for its inheritance. “precision medicine.” Applicability to obstetrics III. Phenotypically similar but etiologically distinct disorders is not a new idea (3). However, do we in 2019 offer require different diagnostic criteria and different “precision medicine in obstetrics”? Not really. To date, treatments: Adrenal hyperplasia is the result of one only limited diagnostic options exist for women with of a series of enzyme defects in the biosynthetic path obstetrical disorders. We can measure the presence that results in cortisol. Negative feedback inhibition or absence of an associated protein (recall that based on cortisol levels modulates rate of synthesis. enzymes are gene products) or metabolites indicative If an enzyme block exists in the pathway, precursors of an altered pathway. But usually this information accumulate prior to the block, leading secondarily to overlaps with those characteristics of other conditions excesses of other steroids. A dozen distinct defects thus not being qualitatively unique. By contrast, a exist. Many differ considerably from one another in novel footprint might exist if we interrogated DNA phenotype, but others show only nuanced differences. sequences in a targeted region from among the 21,000 The adrenal biosynthetic pathway includes the protein-coding genes. We can compare a promising enzymes 21-hydroxylase and 11-β-hydroxylase. sequence in a patient-derived sample to the normal 21-Hydroxylase is required to convert 17α-OH reference genome. Sequencing is no longer difficult, progesterone to 11-deoxycortisol, a penultimate step in expensive, or time-consuming. Benign variants can be the synthesis of cortisol. Genital ambiguity occurs in distinguished readily from pathogenic variants. affected females because when 17α-OH progesterone Imagine the power of expanding and applying our accumulates, conversion to androgens occurs. Salt nascent genomic knowledge to obstetrical conditions. wasting also occurs because 11-deoxycortisol, which Of the protein-coding genes, function is known in facilitates salt retention, is not synthesized in normal only one-third. A portion of the remaining two-thirds amounts. The situation differs in part with the can be expected to play pivotal roles in embryonic/ next enzyme in the adrenal biosynthetic pathway: fetal differentiations or placentation. The sentinel 11-β-hydroxylase, whose deficiency results in excess genomewide association study (GWAS) by Zhang, 11-deoxycortisol that again leads to deficiencies in Muglia, and colleagues is a shining example (4). A cortisol. Like 21-hydroxylase, genital ambiguity total of 43,568 women of European ancestry had occurs as a result of accumulation of excess androgens. self-reported their experience with preterm birth. However, no longer does salt retention occur. Instead, This discovery group was compared to replicates of the salt-retaining steroid 11-deoxycortisol accumulates, three Nordic data sets comprising 8,643 women. Six resulting in hypertension due to hypervolemia. Thus, significant sequences were found. Their contiguous salt wasting occurs in 21-hydroxylase deficiency, genes were thus associated with preterm birth, whereas salt retention occurs in 11β-hydroxylase gestational length, or both. Several genes had not been deficiency. Treatment with cortisol abrogates genital appreciated or received attention previously. Hallman, ambiguity in both. Were genital ambiguity the only Zhang, Muglia, and colleagues further describe how feature, phenotypic differences would not be so well GWAS is only the start (5). appreciated. Protein-coding genes (exome) constitute only 1.5% Differentially treating phenotypically similar but of DNA; 98.5% of the human genome is not protein- actually nonidentical conditions is a major principle coding genes, carrying out regulatory functions. underlying improvement in oncology treatment. Regulatory regions govern the degree of gene Genetic testing often reveals a series of different expression. Expression of a given gene is not necessarily mutations in histologically similar tumors in different on or off, but quantifiable in degree. Imagine a Introduction ix temperature thermostat, in which a single setting is might be traditional, or a novel idea hypothesized not expected and does not need to be constant. Gene by an investigator. Despite immense software and expression can be increased, decreased, or turned off computer prowess needed, recall that this approach without altering the DNA sequence. The mechanism fundamentally requires a human. Computers are very is presumed to involve DNA methylation. One efficient at processing massive data sets to recognize can thus track gene expression by RNA expression patterns; however, someone must program the endpoint (transcription) or by DNA suppression (methylome). If or threshold that justifies an algorithmic response. To a gene ordinarily not expressed (imprinted) suddenly illustrate, let us fancifully try to identify members of a becomes expressed, its regulatory sequence might terrorist cell. As the investigator, one might enumerate have become demethylated and, hence, transcribed a plethora of nefarious characteristics hypothesized (transcriptome). to distinguish terrorist from nonterrorist. This list From study of the genome, transcriptome, and would doubtless be based on our prior assumption methylome, one can concurrently study proteins of behavior expected by a terrorist. For example, one (proteome) and metabolites (metabolome). This might predict gun purchase, proclivity for fast cars, approach is well illustrated by the integrated system and social isolation. A machine learning approach biology approach of Than et al. (6) on preeclampsia. An could thus sort through massive data in hopes of integrated approach was applied by Tsang et al. (7) to identifying recurring patterns indicative of suspect determine DNA or RNA in individual placental cells, individuals. Law enforcement could then focus on correlating with histology, metabolic information, individuals identified. In medicine, we can apply and single cell function and morphology via cell machine learning to fine-tune disease predictions or separation. There was shown to be 18 different types stratify given a known set of associated characteristics. of placental cells. In addition, one can incorporate AI is agnostic. There is no a priori hypothesis. The information on environment and interaction with principle involves interrogating all relationships, not just genome (epigenome), measures of social determination those logically deduced by humans but those seemingly (e.g., nutrition), and clinical status. Examples of the illogical. AI would thus explore all hypotheses. In fact, latter are clinical risk factors and sociodemographic we already use agnostic approaches to locate gene- findings gathered from individual patient data on 4.1 associated sequences. Zhang et al. (4) had no restrictions million women in four countries and California (8). on genes likely to be associated with gestational length These data were used to generate population-based or preterm birth, but found six associated protein- odds ratios (ORs) for risk factors. These differ from coding genes. Not all were anticipated. A human- ORs for an individual’s risk factor. A high OR (e.g., 6.0) derived hypothesis might have been restrictive and for a woman having had a prior preterm birth (PTB) perhaps counterproductive; humans can imagine only would be good counsel, but because PTB occurs only so much, and in doing so subliminally filter out the in 8%–10% of deliveries the population-attributable implausible. Unimaginable relationships would never risks would not be great. By contrast, a lower OR (OR be explored. Let us return to our search for terrorists. AI 1.3) would confer a greater attributable risk for the can proceed agnostically without prior hypotheses and general population if the frequency of a risk factor without unwitting restriction of a hypothesis. Perhaps were more than 8%–10%. Thus, nulliparity and male attending a certain movie theater on the Saturday before sex actually confer a greater overall (population) effect a planned terrorist action will unwittingly pinpoint than preeclampsia or PTB. In aggregate, both nature terrorist members. Perhaps the members also frequent a and nurture can be concurrently assessed. particular ice cream parlor and eat pistachio ice cream. V. Data interrogation and clinical conclusion: Perhaps members unexpectedly enjoy a television show Accumulating an enormous amount of data is of unexpected motif—not an action movie but a fashion attractive but presents a problem. Data accumulation show. AI can interrogate and correlate all this, without and ongoing monitoring are necessary. The bias. A set of characteristics could evolve that would not integrated repository being created for the March of be imaginable by us. Dimes Prematurity Research Centers is exemplary Machine learning and AI are being utilized for (9). How, then, can one utilize these data to system biology, and “multiomic” approaches have determine causative significance and therapy? How resulted in novel observations, especially relating can one extract truly novel information? There are to adaptations during pregnancies using patterns of two conceptual approaches for data exploration fetal gene expression. Predictions can be made on and mining: (1) machine learning and (2) artificial gestational length—not by ultrasound but by analysis intelligence (AI). Tools are similar and overlap, but of maternal blood. Maternal (and fetal) cell free DNA expectations differ. correlated with gestational age predicts gestational Machine learning implies brute-force iterative age within 2 weeks (10). A Stanford team is pursuing queries, using computers and bioinformatics to the concept of gestational clocks: immunologic (11), accomplish such searches. The underlying hypothesis proteomic (12), and a truly agnostic “multiomic” (13).