The MPRINT Hub Data, Model, Knowledge and Research Coordination Center: Bridging the gap in maternal-pediatric therapeutics research through data integration and pharmacometrics

Abstract

Maternal and pediatric populations have historically been considered “therapeutic orphans” due to their limited inclusion in clinical trials. Physiologic changes during pregnancy and lactation and growth and maturation of children alter pharmacokinetics (PK) and pharmacodynamics (PD) of drugs. Precision therapy in these populations requires knowledge of these effects. Efforts to enhance maternal and pediatric participation in clinical studies have increased over the past few decades. However, studies supporting precision therapeutics in these populations are often small and, in isolation, may have limited impact. Integration of data from various studies, for example through physiologically-based pharmacokinetic/pharmacodynamic (PBPK/PD) modeling or bioinformatics approaches, can augment the value of data from these studies, and help identify gaps in understanding. To catalyze research in maternal and pediatric precision therapeutics, the Obstetric and Pediatric Pharmacology and Therapeutics Branch of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) established the Maternal and Pediatric Precision in Therapeutics (MPRINT) Hub. Herein, we provide an overview of the status of maternal-pediatric therapeutics research and introduce the Indiana University-Ohio State University MPRINT Hub Data, Model, Knowledge and Research Coordination Center (DMKRCC), which aims to facilitate research in maternal and pediatric precision therapeutics through the integration and assessment of existing knowledge, supporting pharmacometrics and clinical trials design, development of new real-world evidence resources, educational initiatives, and building collaborations among public and private partners, including other NICHD-funded networks. By fostering use of existing data and resources, the DMKRCC will identify critical gaps in knowledge and support efforts to overcome these gaps to enhance maternal-pediatric precision therapeutics.

Introduction

With 3.6 million births occurring annually in the United States¹ and 22.4% of the population under 18 years of age², maternal and pediatric populations constitute a major portion of the population. Yet, pregnant individuals and children have historically been excluded from clinical studies, leaving much unknown with respect to optimizing medical therapy in these groups.

However, they are not immune to chronic disease (e.g. asthma, seizure disorders) or acute illnesses (e.g. viral infections or injuries). Additionally, pregnant individuals suffer from conditions associated with pregnancy (e.g. gestational diabetes or preeclampsia). Over 68% of individuals report taking at least one prescription drug, excluding vitamins, during their pregnancy, with a large proportion taking four or more drugs³. In a prospective, longitudinal study of 9,546 nulliparous pregnant individuals, 73.4% took a medication other than prenatal vitamins, multivitamins, additional iron, and folic acid, during pregnancy with 55.1% taking medications in the first trimester. Additionally, 13% took more than five medications during pregnancy ⁴. Based on data from the 2011-2014 National Health and Nutrition Examination Survey (NHANES) survey, 21.9% of US children and adolescents aged 0 to 19 years used prescription medications in the 30 days prior to the survey⁵. The most used medications were for asthma, antibiotics, attention-deficient/hyperactivity disorder (ADHD) treatments, topical agents, and antihistamines. In 2019, 7.2% of prescription medications were dispensed to US children 0-19 years of age, with almost 33 million prescriptions dispensed to children aged 0-2 years of age⁶.

The objective of this work is to describe the current landscape of maternal and pediatric pharmacology knowledge and research. We also describe the Indiana University-Ohio State University Maternal and Pediatric Precision in Therapeutics (MPRINT) Hub Data, Model, Knowledge and Research Coordination Center (DMKRCC) and its goals of integrating data relating to maternal and pediatric therapeutics from literature, real-world evidence, and other studies using integrative bioinformatic and pharmacometric modeling approaches.

Maternal and pediatric populations differ from “healthy volunteers”

Physiologically, infants and children are far different from adults. Dr. Abraham Jacobi, the father of American pediatrics, recognized the importance of developmental pharmacotherapy, stating that “Pediatrics does not deal with miniature men and women, with reduced doses and the same class of disease in smaller bodies, but … has its own independent range and horizon.”⁷ Drug distribution is influenced by relatively large extracellular and total body water space in neonates and young infants compared with adults⁸. Ontogeny of drug transporters also impacts drug distribution. For instance, hepatic expression of P-glycoprotein (ABCB1) at birth is 30% to 50% of adults⁹. In drug metabolism, the salient example is the developmental expression of the phase I enzyme Cytochrome P450 (CYP) 3A¹⁰. CYP3A7 is the predominant CYP isoform in fetal liver, protecting the fetus by detoxifying dehydroepiandrosterone sulfate and potentially teratogenic derivatives of retinoic acid. CYP3A7 function peaks at birth, and then declines rapidly afterward. Meanwhile, CYP3A4 and CYP3A5 are not expressed in the fetus, but activation rapidly increases after birth. Although developmental changes in pharmacokinetic-related properties are fairly well understood, there is sparse data regarding ontogeny of drug receptors, leading to pharmacodynamic variability.

Gender differences in pharmacology and toxicology have been noted in both animal¹¹ and human studies¹². Pregnancy extends and amplifies the impact of gender differences on absorption, distribution, metabolism, and elimination¹³. Briefly, the bioavailability of some drugs could be affected due to nausea and vomiting in early pregnancy and delayed gastric emptying and an increase of gastric pH as pregnancy progresses^14–16. For drug distribution, the decrease of albumin concentrations that starts from the second trimester and continues throughout pregnancy will affect pharmacological effect of high extraction ratio drugs with a narrow therapeutic window. Changes in net protein binding, as well as increased total body water and fat stores could theoretically change the volume of distribution of many drugs. Pregnancy brings about variable effects on drug clearance. The activity of some CYP isoenzymes (CYP2A6, CYP2C9, CYP2D6 and CYP3A4) and uridine diphosphate glucuronosyltransferase (UGT) enzymes are increased, while others, such as CYP1A2 and CYP2C19, are reduced. Renal excretion of drugs may be altered by 30-50% during the pregnancy because glomerular filtration rate increases during pregnancy. However, very little knowledge is known regarding the effect of pregnancy on tubular secretion and/or reabsorption that can also affect drug renal excretion^14–16. Although many animal studies unveiled drug toxicity and efficacy, the results from clinical studies are still limited.

As described above, physiology differs between women and men, between nonpregnant and pregnant individuals, and between adults and children, leading to differences in drug distribution and response ^14–17. The underrepresentation of women and children in clinical research, has led to their designation as therapeutic orphans¹⁸. The National Institutes of Health (NIH) Revitalization Act of 1993 required that “women and minorities” be included in NIH-funded studies unless it is inappropriate to the health of the subject or purpose of the research¹⁹. The Institute of Medicine’s committee on Women and Health Research unanimously asserted that “pregnant women be presumed to be eligible for participation in clinical studies.”¹⁸ Yet, with no incentive to enroll pregnant individuals in studies, and ethics surrounding perceived risk to the fetus, pregnant individuals remain therapeutic orphans. Sadly, this has led to limited data regarding changes in pharmacokinetics (PK) and pharmacodynamics (PD) of drugs during pregnancy.

Although there has been growth in maternal and pediatric therapeutics over the past decade, much work remains to be done. A search of PubMed for “pharmacokinetics” AND “pregnancy” clinical trials returns less than 40 publications a year. Likewise, the number of lactation studies published yearly averages in the single digits. Clinicaltrials.gov lists 63 active or planned studies and 38 completed studies including the terms “pregnancy” AND “pharmacokinetics”, with less than 30% of studies receiving support from pharmaceutical industry²⁰. Yet, as noted above, between 68-73% of pregnant individuals have taken at least one prescription drug ^{3, 4}. With a large majority of pregnant individuals taking medications, it is beyond critical that we understand how pregnancy alters the PK of drugs both for optimized treatment of maternal conditions and safety of the developing fetus. The advantages of breastfeeding are numerous ²¹, yet breastfeeding is often discontinued due to fear of how a drug may affect their child ²². These concerns with regard to breastfeeding may be unfounded, as the relative infant dose of drug through breast milk may be sub-therapeutic. However, with the paucity of studies providing information on drug exposure to breastfed infants, patients often err on the side of apparent “safety”, while inadvertently withholding a valuable resource from their infant. Although the importance of ontogeny of drug metabolizing enzymes and transporters has become widely recognized ^23–31, there is little knowledge on ontogeny of drug receptors or mechanism of inter-individual variability in ontogeny functions.

Drug Development & Regulatory Changes

A series of initiatives have increased clinical pharmacology studies in pregnant individuals and children with limited success. These initiatives included: the National Institutes of Health Revitalization Act of 1993¹⁹; the 1999 Institute of Medicine’s committee on Women and Health Research ¹⁸; the 1994 Pediatric Rule³²; the 1997 FDA Modernization Act (FDAMA)³³; 1998 Pediatric Rule³⁴; the 2002 Best Pharmaceuticals for Children Act (BPCA)³⁵; and the 2003 Pediatric Research Equity Act (PREA)³⁶. PREA, in particular, enabled the United States Food and Drug Administration (FDA) to require pediatric studies and provided incentives for pharmaceutical companies for studies in pediatric populations. However, even with all these encouragements to the research community, understanding of drug disposition and response in these groups continues to lag. It remains true that nearly half of the unique drugs studied for pediatric indications have failed development ³⁷. Some reasons for failure of pediatric studies include incorrect dosing, differences in the underlying etiology between adult and pediatric disease, enhanced placebo response in pediatric patients, and incorrect study design. Employing pharmacometric modeling approaches, including physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) models incorporating pediatric ontogeny and mechanisms of underlying disease, in clinical trial design may have averted these failures.

Although these guidances have enhanced drug research in pediatrics, no such incentives are available for studies of drugs in pregnant or lactating individuals. During clinical development of most drugs and biological products, pregnant individuals are actively excluded from trials with the concern that an experimental medication might harm a fetus. And if pregnancy does occur during a trial, the usual procedure is to discontinue treatment and monitor the individual to assess pregnancy outcomes. Consequently, at the time of a drug or biological product’s initial marketing, except for drugs and biological products developed to treat conditions unique to pregnancy, there are little-to-no human data to inform the safety of a drug or biological product taken during pregnancy³⁸. Similarly, research in lactation is lacking. The FDA has drafted two guidances for industry regarding study considerations in pregnant and lactating populations. In one, the FDA recommends pharmacovigilance, pregnancy registries, and complementary data sources as the post-market surveillance approaches during pregnancy³⁸. In the other, it is recommended that industry should conduct clinical lactation studies when “a drug under review for approval is expected to be used in women of reproductive age”, or “if a new indication is being sought for an approved drug and there is evidence of use or anticipated use of the drug by lactating women”³⁹.

Limited inclusion of pregnant and lactating individuals during drug development may lead to unintended, and potentially harmful, consequences. HMG-CoA reductase inhibitors (statins) are one example where pregnant individuals were summarily excluded from study during drug development. With the theoretical risk that reduced cholesterol synthesis may have on the developing fetal brain, and the perceived lack of benefit from using statins in pregnancy, pregnant women were excluded from studies of statins. All statins were given a Category X label, despite their physiochemical differences, mainly due to the lack of potential benefit to outweigh any risk from using the drug in pregnancy. However, it was recognized that endothelial dysfunction and angiogenic imbalance occurring in preeclampsia could be reversed by treatment with statins, as shown in both animal and in-vitro studies^{40, 41}. Studies of pravastatin in pregnant individuals have demonstrated safety and potential efficacy in the treatment of preeclampsia ^{42, 43}.

Although the benefits of breastfeeding are multifold, reducing both maternal and infant morbidity and mortality ^{21, 44–46}, only 47% of infants in the United States are exclusively breastfed for at least 3 months ⁴⁷. Of 575 drugs approved by the FDA from 2015-2017, only 15% had any data regarding lactation ⁴⁸. Additionally, less than 50% of drugs in Lactmed have any data in breastfeeding⁴⁸. More worrisome, is that patients often stop taking medications, or do not seek treatment for an acute condition, because of the lack of information on the drug during pregnancy and lactation. Up to 10% of respondents listed medication use as a factor in their decision to avoid or stop breastfeeding⁴⁹. This number is even higher for those with chronic conditions, such as rheumatoid arthritis, where 58% stopped breastfeeding due to medication use⁵⁰. Study of drugs during lactation is complex, due to variability in milk composition and consumption volume ⁵¹. The FDA recently released guidelines on conducting PK studies in lactation⁵², and recent efforts have provided guidance for PK study design and analysis using pharmacometric modeling approaches⁵³.

The BPCA and the PREA provide incentives for pharmaceutical companies to conduct studies in pediatric populations ^{35, 36}. The FDA supports a database that records new pediatric labeling information ⁵⁴. As of March 2022, it reports 916 drugs with new pediatric studies. However, as of 2017, there were only 46 drugs studied in neonates, 35 neonatal label changes, and 20 drugs indicated for neonates. Yet only 33.8% of mandatory pediatric post-marketing studies were completed (median follow-up 6.8 years); at the time of approval, only 18 of 114 drugs (15.8%) had any pediatric efficacy, safety, or dosing information in their label ⁵⁵. This lack of information results in up to 85% of pediatric patients being treated off-label⁵⁶. Although great strides have been made in understanding the ontogeny of drug metabolizing enzymes and transporters in pediatric patients ^{57, 58}, additional data is needed to fully understand maturation of absorption, distribution, metabolism, and elimination of drugs. Importantly, we have limited knowledge regarding mechanisms controlling maturation processes which may lead to sources of inter-individual variation. It is largely recognized that “children are not small adults.” We are well-aware that developmental changes impact drugs’ PK, but differences in disease etiology and the ontogeny of PD response are less recognized.

Due to ethical and practical limitations, large studies of maternal and pediatric patients are limited. Thus, integration of knowledge across multiple sources is critical for optimizing pharmacotherapy in maternal and pediatric populations. The overarching goal of the MPRINT DMKRCC is to develop a centralized knowledgebase and pharmacometrics modeling resource for maternal and pediatric therapeutics, including data pertinent to understanding variability of PK and PD, to support drug development and personalized therapy in maternal and pediatric populations.

The MPRINT DMKRCC – integrating data, modeling, and real-world evidence to enhance maternal-pediatric therapeutics research

The Obstetric and Pediatric Pharmacology and Therapeutics Branch of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) established the MPRINT Hub to serve as a national resource for conducting and fostering research in pregnant, lactating, and pediatric patients. It is the objective of the Indiana University-Ohio State University MPRINT DMKRCC to address the knowledge deficits and support novel research in maternal and pediatric therapeutics by knowledge aggregation, application of pharmacometric modeling, and dissemination. The DMKRCC functions collaboratively with the two MPRINT Centers of Excellence in Therapeutics (CETs), at the University of California San Diego and Vanderbilt University Medical Center, to serve as a national resource to facilitate and catalyze research and drug development in maternal and pediatric therapeutics (FIGURE 1).

Figure 1. The Maternal and Pediatric Precision in Therapeutics (MPRINT) Hub.

consists of the Indiana University-The Ohio State University Data, Model, Knowledge and Research Coordination Center (DMKRCC) and Centers of Excellence in Therapeutics (CET) at University of California San Diego (UCSD) and Vanderbilt University Medical Center (VUMC). The MPRINT Hub brings together data, knowledge, resources, and expertise through collaborations with various networks and public-private partnerships to enhance maternal-pediatric precision therapeutics.

The MPRINT DMKRCC consists of three research cores: The Knowledgebase and Portal Core, The Real-World Evidence Core, and The Pharmacometrics and Clinical Trial Design Core (Pharmacometrics Core); and two supporting cores: The Outreach and Education Core and a Logistics Core (FIGURE 2). These cores work in concert with each other and the CETs to provide resources and educational opportunities to enhance maternal and pediatric precision therapeutics research.

Figure 2. The MPRINT Hub Data, Model, Knowledge, and Research Coordination Center.

includes a Knowledgebase and Portal Core, Real-World Evidence Core, Pharmacometrics and Clinical Trial Design Core, and Outreach and Education Core, which function together to integrate and utilize information relevant to maternal-pediatric therapeutics to enhance research and precision therapy in these populations.

MPRINT, Maternal and Pediatric Precision in Therapeutics; NLP, Natural Language Processing; EHR, Electronic Health Record; PK/PD, pharmacokinetic/pharmacodynamic; PBPK-PD, physiologically based pharmacokinetic/pharmacodynamic; MIDD, model-informed drug discovery

The MPRINT Hub meets regularly with NICHD Program Officers and External Program Consultants with expertise in data sciences, clinical pharmacology, pediatric subspecialties, maternal health, clinical trial design and analysis, developmental biology, and regulatory sciences, to review the functioning and progress of the Hub and ensure the Hub is operating optimally and efficiently.

Below, we describe how the DMKRCC Knowledgebase and Portal Core, Real-World Evidence Core, and Pharmacometrics Core function to identify and organize data relating to maternal and pediatric therapeutics into a knowledgebase to support development of pharmacometric models aimed at enhancing clinical trial design and pharmacotherapy in these populations.

Pharmacology knowledgebase development

The primary goals of the Knowledgebase and Portal Core are to develop and disseminate a maternal and pediatric therapeutics knowledgebase for the MPRINT Hub. This knowledgebase will include data on maternal and pediatric physiology and pharmacology, including data on PK, PD, pharmacogenomics (PG), and pharmacoepidemiology (PE) from in vitro studies through phase 4 clinical trials. Although a number of databases and resources provide information relevant to the drug PK, PD, PG, PE, drug targets, and clinical trials, few provide sufficient information relating to maternal and pediatric populations (Table 1). However, information on maternal and pediatric physiology and their impact on pharmacology are limited and often fragmented, with no single source containing integrated PK, PD, PG, and PE data in these populations.

Table 1:

Public domain drug data sources

	Maternal	Pediatric	PK	PD	PG	PE	Targets	Clinical Trials
PubMed	X	X	X	X	X	X	X	X
Drug Labels	X	X	X	X	X	X	X	X
Goodman & Gilman’s: The Pharmacological Basis of Therapeutics 59			X
UCSF-FDA TransPortal 60			X		X
PharmGKB 61					X	X	X	X
FDA Adverse Event Reporting System (FAERS)62	X	X				X
DrugBank 63			X		X		X
ClinicalTrials.gov 20	X	X						X

PK, pharmacokinetic; PD, pharmacodynamic, PG, pharmacogenomic; PE, pharmacoepidemiology; FDA, U.S. Food and Drug Administration; UCSF, University of California San Francisco

The scientific complexity behind pediatric ontogeny and maternal physiology is undeniable. Elucidation of many mechanisms and further understanding of drug disposition in maternal and pediatric patients is not only hampered by studies with small sample sizes or the lack of any clinical investigation altogether, but also by the conflicting or inconclusive findings for the studies that are published in literature. The Knowledgebase and Portal Core is collating data from literature sources using natural language processing to support text mining methods. This effort includes expanding previously developed ontologies for PK and PG ⁶⁴ to support the complexities associated with maternal and pediatric populations. We have previously developed a PK ontology, which included Gene Ontology, pharmacology genes including metabolism enzymes and transporters and their functional variants, single drug and drug interaction parameters PK, in both preclinical experiments and clinical studies, and patient populations data ⁶⁴. The MPRINT DMKRCC is developing a patient-centric ontology to describe seven primary annotation categories of physiology focusing on pediatric ontogeny and the physiological changes that occur with and around pregnancy and lactation: disease, drug, PK, PG, PD, drug toxicity, and drug efficacy.

A maternal and pediatric pharmacotherapy knowledgebase will be developed to integrate various data sources for PK, PD, PE, PG, and clinical trials in children and pregnant/lactating individuals. The knowledgebase will be used to facilitate assessment of the quality, reliability, and gaps in evidence and support translational research in maternal and pediatric precision therapeutics. Literature or existing public domain data sources will be integrated into the MPRINT Hub Knowledgebase, including information from PubMed, drug labels, and PharmGKB⁶¹. Table 2 shows the data annotation scheme for PK, PD, PE, PG, and Clinical trials which share the data elements pertinent to study design, demographics, and physiology and disease states for children, and pregnant, and lactating individuals. Data will be curated from unstructured text (e.g. abstracts, manuscripts, drug labels) using both machine learning and manual curation.

Table 2:

Data annotations for pharmacology research in children and pregnant/lactating individuals

Evidence type	Shared data elements	Specific data elements
PK	Physiology of children and pregnant/lactating individuals, diseases, demographics, data sources	Single drug PK: CL, CLR, CLH, AUC, t1/2, fm, fe, Cmax; Drug interaction PK AUCR or CLR, In vitro studies PK: Vmax, Km, fu, KI, IC50, EC50, Emax
PD		drug effects on biomarkers.
PG		Pharmaco-genes, genetic variants, genetic effects on PK, PD, or clinical outcomes.
PE		Efficacy and adverse event definition in the PE study, and drug effects.
Clinical trials		Efficacy and adverse event definition, drug effects, study outcomes

PK, pharmacokinetic; PD, pharmacodynamic, PG, pharmacogenomic; PE, pharmacoepidemiology; CL, clearance; CL_R, renal clearance; CL_H, hepatic clearance; AUC, area under the concentration-time curve; t_1/2, half-life; fm, fraction metabolized; fe, fraction excreted unchanged in the urine; C_max, maximum plasma concentration; V_max, maximum velocity; Km, Michaelis constant; f_u, fraction unbound; K_I, inhibition rate constant; IC50, half maximal inhibitory concentration; EC50, half-maximal effective concentration; E_max, maximum drug effect

A relational database is being developed to integrate various data resources regarding PK, PD, PG, PE, and clinical trials, with links to the raw data. The database will be searchable by generic drug name or chemical name, combined with demographic and disease information (children or pregnant/lactating individuals) as the primary key. All data will link back to the primary resource (e.g. Pubmed Identifier or Digital Object Identifier (DOI)). The database will be accessible through a Knowledgebase Portal on the MPRINT Hub website (mprint.org).

The data portal will be designed to accommodate a wide range of users, including maternal and pediatric clinicians, basic science researchers, and pharmacologists. Data products on the knowledgebase will be available to view through web applications and as downloadable datasets to ensure maximum flexibility to fit users’ needs. The User Interface will be developed based on user feedback to be purposeful, comprehensive, easy to navigate, and informative. The queried information will be organized and displayed based on the evidence type (PK, PG, PD, PE, and clinical trials) in individual sections. Drop-down and auto-complete features for drug selection are available. A graphical interface will enable users to visually identify available data and gaps in knowledge for a given query.

Real-World Evidence Core

Historically, the evidence basis of pediatric and maternal treatments has lagged behind those in other patient populations. A key aspect of this is the lack of pediatric or maternal data, which originates from the logistic, ethical, and legal challenges of performing clinical investigations in children and pregnant or lactating individuals. The collection of real-world data through electronic health records (EHRs) has sharply increased in the last decade, which opens an unprecedented potential for data collection with more subjects, more variables, and lower costs ⁶⁵. However, the primary barriers have been the accessibility of the maternal and pediatric data in the EHR, and lack of informatics resource or expertise in designing the clinical trials using EHR data ⁶⁶. Real-world evidence is available but not yet easily accessible in maternal and pediatric pharmacotherapy research ⁶⁷.

Indiana University, through Regenstrief Institute, and the partnership of Nationwide Children’s Hospital (pediatric) and the Wexner Medical Center at The Ohio State University (adult), have developed integrative EHRs and biobank efforts that can be a rich resource for research in maternal and pediatric therapeutics. Maternal health and well-being, including medication use, have been associated with long-term outcomes in offspring.

However, as mothers and their infants may be treated at more than one hospital, or even hospital system, it becomes difficult to longitudinally track maternal-child dyads from pregnancy through childhood and into adulthood. For example, infants delivered at The Ohio State University Medical Center may be transferred to Nationwide Children’s Hospital for specialty care. Similarly, in Indianapolis, babies delivered at Eskenazi Hospital in need of a level IV care are typically transferred to Indiana University Health Riley Hospital. The Real-World Evidence Core is developing a linkage algorithm for maternal-infant records that will enable researchers to evaluate associations between maternal exposure and long-term childhood outcomes and to gconnect maternal-child data across health systems.

Ultimately, we plan to establish a curated maternal-pediatric data mart that incorporates EHR and other real-world data, with linkage to biospecimens that have been collected through biobank studies at both Indiana University and Nationwide Children’s Hospital. Data relevant to understanding maternal and pediatric physiology and pharmacology will be identified and integrated into the maternal-pediatric data mart to support studies to enhance knowledge and understanding of drug disposition and response, and to support drug development and therapeutics in maternal and pediatric populations. The first step in this process, developing data governance to enable data sharing and dissemination, is currently underway.

The Pharmacometrics and Clinical Trial Design Core

Although pharmacometric approaches have become more widespread in pediatrics ^{23, 25, 27, 68, 69}, there is still limited use in obstetric populations ^70–73 and even fewer examples in lactation ^{53, 74, 75}. Even within the pediatric literature, PBPK models focus on “healthy” patients, with limited integration of comorbid conditions and disabilities. Meanwhile, population PK models (popPK) may be developed within a limited subset of patients, for instance surgery patients, which can limit extrapolation to other sub-populations. Additionally, there are few examples of pharmacometric models that integrate maternal and neonatal exposure. Understanding the overall exposure of a neonate to drug requires understanding in utero exposure through both placenta and amniotic fluid, neonatal intake through breastmilk, and exogenous dosing of drug to the infant. In addition to various sources of drug exposure, one must also consider the ontogeny of drug metabolizing enzymes and transporters, as well as maturation of renal clearance and other processes that influence drug disposition.

Pharmacometrics, systems pharmacology, artificial intelligence, and other novel quantitative approaches provide important tools to explore the PK and PD of therapeutic agents employed in maternal and pediatric populations. Quantitative systems pharmacology, including PBPK/PD modeling, allows the integration of data from disparate sources to develop predictive PK/PD models which augment the often sparse data available from clinical studies in maternal and pediatric populations. Modeling and simulation approaches are more commonly used in drug development, with several examples in pediatrics including bridging studies ^{76, 77}.

The Pharmacometrics Core utilizes pharmacometric modeling and simulation to improve understanding of PK and PD in maternal and pediatric populations in order to enhance drug development and therapeutic outcomes. The Pharmacometrics Core brings together experts in pharmacometrics and clinical pharmacology to support development of guidelines for data quality and assist in analyzing data from maternal and pediatric studies using cutting-edge pharmacometric methods; develop PBPK/PD models encompassing dynamic changes in maternal and pediatric physiology and ontogeny and incorporating sources of inter- and intra-individual variability such as PG, obesity, concomitant disease and disabilities; and support model informed drug development, clinical study design, and personalized pharmacotherapy through the maternal and pediatric lifespan.

Several recent reviews have described physiologic changes in pregnancy that impact PK, including changes in body composition, plasma binding proteins, increased glomerular filtration, and alterations in drug metabolizing enzymes. ^{14–16, 78, 79} Others have described PBPK models of pregnancy^{71, 73}. Although PBPK models are being used to understand and predict PK changes of drugs in pregnancy, less is known about effects of pregnancy on drug receptors and other aspects influencing pharmacodynamic response, including insulin sensitivity ⁸⁰, immune function ⁸¹, and opioid receptors ⁸². Although PK samples are often collected at delivery, very few studies have evaluated the impact of labor and treatments associated with delivery (e.g. induction agents, cesarean section), and blood loss on the concentrations observed in these studies. Additionally, little is known about postpartum changes in PK/PD. Literature data on postpartum physiologic changes are limited and lack granularity essential for understanding the rate of these changes. For instance, data are not available on postpartum changes in blood flow and tissue volume of the gastrointestinal track, which could greatly influence oral absorption⁷⁹. Although it is commonly assumed that maternal physiology returns to the pre-pregnancy state by 4-6 weeks postpartum, it is also recognized that pregnancy results in long-term changes to maternal physiology ⁸³. Lactation may also induce changes in maternal physiology, leading to additional variation in post-partum PK ⁷⁹.

Although the importance of incorporating maturational changes into pediatric PBPK models is well-recognized ^{29, 84}, we still have a limited understanding of ontogeny of drug metabolizing enzymes and transporters, and the factors that regulate it. Use of advanced proteomic methods have greatly improved our understanding of ontogeny drug transporters in the liver and kidney ^{9, 85}. However, we still have limited understanding of transporters in other tissues, including the gastrointestinal tract and brain ⁸⁶. Even less is known about maturational changes in receptors and other factors that influence PD response in pediatric patients.

There is also evidence that disease and disability impact PK as well as PD. Children with obesity appear to have an increased risk for multiple sclerosis and increased relapses on first-line treatment with interferon beta and glatiramer acetate, necessitating the need for second-line therapy⁸⁷. Spinal cord injuries, one cause of disability in pediatric patients, result in multiple sequelae leading to changes in neuromuscular, cardiovascular, respiratory, gastrointestinal, and urinary function that can impact PK of drugs including reduced bioavailability of theophylline and increased elimination of vancomycin⁸⁸. Additionally, individuals with disabilities may be taking medications that induce or inhibit metabolism of concomitant drugs. For instance, the anti-epileptic drugs carbamazepine and phenytoin are strong inducers of CYP450 3A4, 2B6, and other drug metabolizing enzymes and transporters⁸⁹. With increasing rates of childhood obesity, it is also important to consider the impact that body composition has on pediatric (and maternal) PK⁹⁰. For example, altered midazolam clearance and distribution have been reported in morbidly obese adolescents⁹¹. Although some efforts have been made to capture covariates (e.g. due to comorbid conditions⁹²), most maternal and pediatric PBPK-PD models assume a “healthy” population, limiting extrapolation to support use in complex patients, even in the simple case without considering polypharmacy.

Although much progress has been made in recent years to extend PBPK modeling to maternal and pediatric populations, there remain knowledge gaps that prevent the full application of PBPK/PD modeling to support precision dosing in these populations. As noted in a recent review, additional data are needed to inform maternal-fetal PBPK models, including longitudinal changes in hepatic and extrahepatic drug metabolizing enzymes and transporters; biologics; comorbid conditions; alterations in PD response; and factors influencing placental and fetal transport, metabolism, and PD⁷³. Similarly, additional information on enzyme and transporter ontogeny, comorbid conditions, and enhanced understanding of PD response are needed to further support pediatric PBPK/PD models.

With the difficulties of conducting clinical studies in maternal and pediatric populations, often data is sparse and of poor quality. For instance, the majority of PK data on mental health drugs in pregnancy come from case reports, small case series, or opportunistic sampling studies. Many of these studies only reported maternal and cord plasma concentrations at delivery. Some studies included patients taking a wide range of doses yet did not report dose-corrected concentrations ^93–95. One study assessed concentrations of drugs and metabolites in hair as a non-invasive method to estimate PK changes across gestation ⁹⁶. However, it is unclear how variation in rate of hair growth, hair type, and other factors may impact drug uptake into hair follicles ^{97, 98}. Inclusion of inappropriately designed studies into PBPK-PD models may lead to inaccurate conclusions. Therefore, the Pharmacometrics Core will support development of guidelines for assessment of clinical studies, including assessment of drug dosing, timing of sample collection, description of study population, sensitivity and specificity of drug assays, and appropriateness of PK-PD analyses.

There is large variability in PBPK model structures of pregnancy where models may represent the fetal-placental unit as a single compartment ^{70, 99}, as separate fetal and placental compartments¹⁰⁰, or a more extensive, multi-compartment fetal PBPK model, enabling prediction of fetal organ concentrations¹⁰¹. Additionally, pharmacometric model quality, verification, and reporting standards have varied widely. A systematic review reported that 56% of published PBPK models did not set a priori criteria for validation ¹⁰². Although the International Consortium for Innovation & Quality in Pharmaceutical Development IQ PBPK working group suggests that criteria should be predefined regarding model evaluation, no consensus is available on what comprises those criteria ¹⁰³. In addition to data quality standards, the Pharmacometrics Core will support development and utilization of standards for model development, qualification, and reporting.

The future of maternal-pediatric precision therapeutics

As described above, the primary objective of the MPRINT DMKRCC is to compile, interpret, and augment data related to maternal and pediatric pharmacology through bioinformatics tools and pharmacometric modeling to enhance development of precision therapeutics. The DMKRCC cores work in concert with one-another to support this goal. For instance, the Real-World Evidence Core can inform population parameters for development of PBPK models or data that can be leveraged to model disease response by the Pharmacometrics Core ^{104, 105}. The Knowledgebase will provide critical information for validation of PBPK/PD models and biomarkers to support clinical trial design.

In addition, the MPRINT DMKRCC will integrate the novel findings and tools developed by the MPRINT CETs to enhance understanding of the acute and long-term effects of antibiotic and opioid exposure during pregnancy and lactation on maternal and infant health and well-being. As shown in Figure 1, the MPRINT Hub is building partnerships with other networks and researchers to develop cooperative efforts to further catalyze research in maternal and pediatric precision therapeutics. Through this integrative approach, the MPRINT DMKRCC will help identify gaps in knowledge necessary for precision pharmacotherapy in maternal and pediatric populations, support well-designed research studies in these critical areas, and further the understanding of maternal and pediatric precision therapeutics. Through NIH-backed open data initiatives, we hope to optimally leverage the limited data to further optimize therapy in maternal and pediatric populations. Through the education, collaboration, and support of both existing and novel methodologies focused on critical gaps in knowledge provided by the MPRINT Hub, we hope to bring synergistic growth to the understanding and application of maternal-pediatric precision therapy.