Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism Stephanie C

University of Connecticut OpenCommons@UConn

Doctoral Dissertations University of Connecticut Graduate School

5-4-2018 Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism Stephanie C. Piekos University of Connecticut - Storrs, [email protected]

Follow this and additional works at: https://opencommons.uconn.edu/dissertations

Recommended Citation Piekos, Stephanie C., "Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism" (2018). Doctoral Dissertations. 1850. https://opencommons.uconn.edu/dissertations/1850 Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism Stephanie C. Piekos, PhD University of Connecticut, 2018

Adverse reactions or lack of therapeutic efficacy are commonly encountered challenges when prescribing pharmaceutical drugs to patients. These undesirable outcomes highlight the large amount of interindividual variability in drug responses, and can often be attributed to differences in drug metabolism and clearance rates among patients. Patient age and current medication usage are two variables that can increase or decrease the expression and activity of drug metabolism by cytochrome P450 enzymes (CYPs), thus impacting responses to other drugs. Neonates, infants, and children, express CYP enzymes at different levels than adults due to immature liver functions, and this causes drug responses to differ in pediatric populations. Be- cause infants and children are not typically included in clinical drug trials, there is a significant knowledge gap in understanding how CYP expression and liver function may be impacted by drug treatment in early postnatal life. Antiepileptic drugs (AEDs) are commonly administered to newborns, infants, and children suffering from acute seizures or chronic epilepsy, however these medications are known to induce the expression of drug-metabolizing CYPs. This work has explored the consequences of exposure to two AEDs, phenobarbital and phenytoin, on the expression and activity of drug-metabolizing CYPs in the liver during postnatal development in mice. Short-term, as well as long-term, effects on gene and protein expression profiles and enzymatic activities were evaluated. Our results demonstrate that responses to AED treatment differ significantly in neonatal mice compared to adults in terms of CYP inducibility and liver functions that were altered. Additionally, exposure to AEDs at neonatal ages may cause permanent effects on gene and protein expression in the liver in adulthood, which could impact drug efficacy and susceptibility to liver diseases later in life. These differential effects may be due to epigenetic modifications that are susceptible to alteration by drug treatment in the maturing liver that does not occur in the fully differentiated adult liver. Overall, these studies highlight a critical need for a better understanding of pharmacokinetic regulation in pediatric populations, as drug responses in adults cannot be used for accurate prediction or extrapolation of drug responses in infants and children. Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism

Stephanie C. Piekos

B.S., University of Connecticut, 2013

A Dissertation Submitted in Partial Fulﬁllment of the Requirements for the Degree of

Doctor of Philosophy at the University of Connecticut

2018 APPROVAL PAGE

Doctor of Philosophy Dissertation

Consequences of Antiepileptic Drug Exposure During Postnatal Liver Development on Cytochrome P450 Expression and Metabolism

Presented by Stephanie C. Piekos, B.S.

Major Advisor: Xiao-bo Zhong

Associate Advisor: Charles Giardina

Associate Advisor: Warren Ku

Associate Advisor: Jose Manautou

University of Connecticut 2018

ii Acknowledgements

First and foremost, I will be eternally grateful for my adviser, Dr. Xiao-bo Zhong. Five years ago, he provided me with an opportunity to pursue research, and has never given up on me. His guidance throughout my graduate work is something I will always be thankful for, and he has equipped me with invaluable skills for a successful journey into my scientiﬁc career. He taught me how to think critically, how to connect with others, and how to tell a story with science that can leave a lasting impact. He has been understanding and compassionate through difﬁcult times, and he is someone I will always appreciate and look up to.

I would like to thank my doctoral advisory committee, Dr. Jose Manautou, Dr. Charles Gia- rdina, and Dr. Warren Ku for their generosity in providing their time and thoughtful insights into my research. Each one of them has been uniquely inﬂuential on my graduate career and I appreciate them all so much.

To all past and current members of our laboratory, especially Dr. Yun-Chen ”Erica” Tien, Dr. Lai Peng, Dr. Chad Pope, Parimal Pande, Liming Chen, Yifan Bao, Xueyan Shao, Pei Wang, Austin Ferrara, Sharon Yaqoob, and Pramika Stephen, I will always be thankful for all that you have taught me and assisted me with, and for sharing your ideas. And to all of my peers in Pharm/Tox and Med Chem, I am so grateful for your friendship, help, support, and encouragement.

I would also like to thank the Pharmacology/Toxicology faculty for all of their guidance and insight through their courses and seminars. Also thank you to Leslie LeBel for the countless things she has helped me with and for her unending support. This is truly a wonderful department that has provided me with so much opportunity. In addition, I would like to thank Dr. Dong-Guk Shin in the Computer Science and Engineering Department, and all of his lab members, for the opportunity to be involved in their research and for teaching me indispensable bioinformatics skills.

I am very thankful for our collaborators, Dr. Xiaochao Ma and Dr. Pengcheng Wang at the University of Pittsburgh, and Dr. Hao-jie Zhu and Dr. Jian Shi at the University of Michigan, for their generous help with enzymatic activity analyses and protein quantiﬁcation, respectively. I would also like to acknowledge our funding from the NIH, 1R01GM118367, and thank the American Foundation for Pharmaceutical Education for their support through their Pre-Doctoral Fellowship.

iii Finally, I would not be where I am today without the overwhelming support from my family. I owe them everything. They have picked me up when I thought I could not go on, and have encouraged me to never give up. They have instilled the values of hard work and dedication into me, and I always strive to make them proud. I also am so grateful for my wonderful friends that have stuck by me through all of this. I will never be able to thank them enough for their support and understanding and for the happiness they bring me.

iv Contents

Approval Page ii

Acknowledgements iii

List of Figures viii

Abbreviationsx

1 Introduction1 1.1 Drug Responses in Pediatric Populations...... 2 1.2 Drug-Drug Interactions...... 4 1.3 Aim of the Project...... 6

2 Literature Review and Background8 2.1 Pharmacokinetics and Drug Responses in Adults...... 8 2.1.1 Hepatic Drug Metabolism...... 9 2.1.2 Cytochrome P450 Enzymes...... 11 2.1.3 Cytochrome P450s in Drug-Drug Interactions...... 14 2.2 Pharmacokinetics and Drug Responses in Infants and Children...... 15 2.2.1 Postnatal Liver Maturation...... 17 2.2.2 Ontogeny of Cytochrome P450s...... 19 2.2.3 Induction of Cytochrome P450s in Pediatric Populations...... 22 2.2.4 Use of Antiepileptic Drugs in Pediatric Patients...... 23 2.3 Mechanisms of Cytochrome P450 Regulation...... 27 2.3.1 Genetic Polymorphisms...... 27 2.3.2 Nuclear Receptors...... 30 2.3.3 Epigenetic Modiﬁcations...... 32 2.3.4 Hormones...... 35 2.4 Regulation of Cytochrome P450s During Postnatal Maturation...... 37 2.4.1 Neonatal Phenobarbital Exposure and Permanent Changes to Cytochrome P450 Expression...... 38

v 2.4.2 Neonatal Exposure to Other Drugs and Permanent Changes to Liver Function...... 40 2.4.3 Nuclear Receptors and Neonatal Imprinting...... 43 2.5 The Mouse as an In Vivo Model for Studying Drug Metabolism...... 49 2.5.1 Human and Mouse Orthologous Cytochrome P450 Genes...... 50 2.5.2 Inter-Species Age and Dose Extrapolation...... 56

3 Neonatal Phenobarbital Exposure and Drug Interactions in Adulthood 59 3.1 Introduction...... 59 3.2 Materials and Methods...... 62 3.3 Results...... 64 3.3.1 Omeprazole inhibits the gastric proton-pump to increase pH in the stomach of adult mice...... 64 3.3.2 Concurrent administration of phenobarbital and omeprazole temporar- ily reduces efﬁcacy of omeprazole in adult mice...... 65 3.3.3 Neonatal administration of phenobarbital reduces the efﬁcacy of omeprazole in adulthood...... 68 3.4 Discussion...... 71

4 Long-term Effects on the Liver Transcriptome Following Neonatal Phenobarbital Exposure 75 4.1 Introduction...... 75 4.2 Materials and Methods...... 78 4.3 Results...... 81 4.3.1 Long-term effects on the liver transcriptome following neonatal phenobarbital exposure...... 81 4.3.2 Long-term effects on drug metabolism following neonatal phenobarbital exposure...... 85 4.3.3 Long-term effects on liver functions following neonatal phenobarbital exposure...... 88 4.4 Discussion...... 94

5 Long-term Effects of Neonatal Phenytoin Exposure on Cytochrome P450 Expres- sion 100 5.1 Introduction...... 100 5.2 Materials and Methods...... 102 5.3 Results...... 108 5.3.1 Long-term effects of neonatal phenytoin exposure on the basal expression of cytochrome P450s...... 108 5.4 Discussion...... 113

6 Short-term Effects of Phenytoin Exposure on Cytochrome P450 Expression Through- out Postnatal Development 117 6.1 Introduction...... 117

vi 6.2 Materials and Methods...... 119 6.3 Results...... 121 6.3.1 Cyp2b10 induction and function following phenytoin exposure..... 122 6.3.2 Cyp2c29 induction and function following phenytoin exposure..... 124 6.3.3 Cyp3a11 induction and function following phenytoin exposure..... 124 6.3.4 Cyp3a16 repression following phenytoin exposure...... 126 6.3.5 Phenytoin represses H19/Igf2 expression during early postnatal life... 129 6.4 Discussion...... 130

7 The Differential Effects of Phenytoin Treatment on the Liver Proteome in Neonates and Adults 135 7.1 Introduction...... 135 7.2 Materials and Methods...... 137 7.3 Results...... 138 7.3.1 Short-term effects on the adult liver proteome following phenytoin exposure...... 139 7.3.2 Short-term effects on the neonatal liver proteome following phenytoin exposure...... 141 7.3.3 Long-term effects on the adult liver following phenytoin exposure at the neonatal age...... 144 7.4 Discussion...... 146

8 Summary and Future Directions 152

References 156

vii List of Figures

2.1 Chemical structures of phenobarbital and phenytoin...... 27 2.2 Direct and indirect pathways leading to CAR activation by xenobiotics in mouse. 28 2.3 Summary of studies on long-term alterations to CYPs...... 41 2.4 Drugs administered to pediatric patients...... 48 2.5 Human and mouse orthologous CYP enzymes and their substrates...... 52 2.6 Ontogeny of murine drug-metabolizing CYPs during postnatal liver maturation. 53 2.7 Repression of murine fetal CYP isoform Cyp3a16 during postnatal liver maturation...... 55 2.8 Ontogeny of murine nuclear receptors Car and Pxr during postnatal liver maturation...... 56

3.1 Omeprazole increases gastric pH without altering CYP expression in adult mice. 66 3.2 Concurrent administration of phenobarbital and omeprazole results in a drug- drug interaction in adult mice...... 67 3.3 Neonatal administration of phenobarbital causes a long-term drug-drug interaction and reduces efﬁcacy of omeprazole in adult mice...... 70 3.3 Neonatal administration of phenobarbital causes a long-term drug-drug interaction and reduces efﬁcacy of omeprazole in adult mice (cont.)...... 71

4.1 Correlation matrix of transcriptome proﬁles among adult mice treated with phenobarbital or control at the neonatal age...... 82 4.2 Summary of RNA-seq results...... 83 4.3 Comparison of the expression of all cytochrome P450 enzymes in adult liver following neonatal phenobarbital treatment...... 84 4.3 Comparison of the expression of all cytochrome P450 enzymes in adult liver following neonatal phenobarbital treatment (cont.)...... 85 4.4 Comparison of expression of phase I enzymes in adult liver following neonatal phenobarbital exposure...... 87 4.4 Comparison of expression of phase I enzymes in adult liver following neonatal phenobarbital exposure (cont)...... 88 4.5 Comparison of expression of phase II enzymes in adult liver following neonatal phenobarbital exposure...... 89 4.5 Comparison of expression of phase II enzymes in adult liver following neonatal phenobarbital exposure (cont.)...... 90

viii 4.6 Comparison of the expression of phase III ABC transporters in adult liver following neonatal phenobarbital exposure...... 90 4.7 Expression of phase III SLC transporters in adult liver following neonatal phenobarbital exposure...... 91 4.7 Comparison of expression of phase III SLC transporters in adult liver following neonatal phenobarbital exposure (cont.)...... 92 4.8 Gene ontology analysis of signiﬁcantly differentially expressed genes in adult mice following neonatal exposure to phenobarbital...... 93

5.1 Expression and activity of Cyp2b10 throughout postnatal development following phenytoin exposure at neonatal age...... 110 5.2 Expression of Cyp2c29 throughout postnatal development following phenytoin exposure at the neonatal age...... 111 5.3 Expression and activity of Cyp3a11 throughout postnatal development following phenytoin exposure at the neonatal age...... 112

6.1 Short-term effects of phenytoin treatment on the expression and activity of Cyp2b10 at different ages during postnatal development...... 123 6.2 Short-term effects of phenytoin treatment on the expression of Cyp2c29 at different ages during postnatal development...... 125 6.3 Short-term effects of phenytoin treatment on the expression and activity of Cyp3a11 at different ages during postnatal development...... 127 6.4 Repression of Cyp3a16 expression following phenytoin treatment at different ages during postnatal development...... 128 6.5 Repression of H19 and Igf2 gene expression following phenytoin treatment during postnatal development...... 129

7.1 Summary of differential protein expression in the liver following phenytoin treatment in neonates and adults...... 140 7.2 Ontology analysis of differentially expressed proteins in response to short-term phenytoin treatment in adults...... 142 7.3 Ontology analysis of differentially expressed proteins in response to short-term phenytoin treatment in neonates...... 145 7.4 Ontology analysis of differentially expressed proteins in adult mice in response to long-term phenytoin treatment at the neonatal age...... 147

ix Abbreviations

ADR Adverse drug reaction DDI Drug-drug interaction PBPK Physiologically based pharmacokinetic modeling CYP Cytochrome P450 enzyme SNP Single nucleotide polymorphism FMO Flavin-containing monooxygenase UGT UDP-glucoronosyltransferase SULT Sulfotransferase NAT N-acetyltransferase GST Glutathione S-transferase ABC ATP-binding cassette SLC Solute carrier NAFLD Non-alcoholic fatty liver disease FDA Food and Drug Administration IGF Insulin-like growth factor C/EBP CCAAT-enhancer binding protein HNF Hepatocyte nuclear factor PXR Pregnane X receptor CAR Constitutive androstane receptor AhR Aryl hydrocarbon receptor LncRNA Long non-coding RNA

x AED Antiepileptic drug CNV Copy number variant PPI Proton-pump inhibitor RXR Retinoid X receptor SRC Steroid receptor co-activator PGC PPARγ co-activator PPAR Peroxisome proliferator-activated receptor TCDD 2,3,7,8-tetrachlorodibenzo-p-dioxin GR Glucocorticoid receptor VDR Vitamin D receptor DNMT DNA methyltransferase HAT Histone acetyltransferase HDAC Histone deacetylase EZH Enhancer of zeste GHR Growth hormone receptor ER Estrogen receptor BPA Bisphenol A SHP Small heterodimer protein TCPOBOP 1,4-bis-[2-(3,5-dichloropyridyloxy)]benzene PCN Pregnenolone 16 α-carbonitrile CITCO 6-(4-Chlorophenyl)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde- O-(3,4-dichlorobenzyl)oxime PBS Phosphate-buffered saline OME Omeprazole PB Phenobarbital GAPDH Glyceraldehyde 3-phosphate dehydrogenase DME Drug metabolizing enzyme GO Gene ontology CTCF CCCTC-binding factor EGFR Epidermal growth factor receptor

xi PP2A Protein phosphatase 2A PHY Phenytoin TFA Triﬂuoroacetic acid DTT Dithiolthreitol IAA Iodoacetamide BSA Bovine serum albumin SWATH-MS Sequential window acquisition of all theoretical mass spectra IDA Information-dependent acquisition MS Mass spectrometry

xii To my sister and best friend, Jennifer. And to Mom and Dad:

My family unit, my unrelenting strength, and the reason for everything I have become.

xiii Chapter 1

Introduction

In the current, traditional model of clinical healthcare, patients that are diagnosed with a specific disease or condition are often all prescribed the same medication at the same dosage, with little regard for individual factors and differences. Whereas this system of medical intervention is successful for some patients, an even greater proportion of patients experience either no therapeutic efficacy or potentially dangerous adverse side effects. It is estimated that most major drugs are efficacious only in 25-60% of patients [1]. Additionally, more than two million cases of adverse drug reactions (ADRs) are reported each year and account for nearly 5% of total hospital admissions in the United States. Reactions to drugs can be life threatening, and cause 100,000 deaths annually [2]. Many genetic, physiological, and environmental factors synchronously play a role to determine how a patient will respond to drug treatment. Under- standing these individual factors and implementing them into a more personalized model of therapeutic medical intervention could improve the number of patients that respond positively and without side effects to drug treatment.

1 Chapter 1 2

1.1 Drug Responses in Pediatric Populations

Age is a significant factor known to affect drug response, yet the majority of therapeutic medications are developed and subjected to clinical trials in adult patients only to characterize their efficacy and safety. Due to ethical and technical constraints, drugs simply are not usually tested in pediatric or geriatric populations before being approved and reaching the market. Infants and children, in particular, often demonstrate highly differential effects from adults to the same drug treatment due to immature renal and hepatic function [3]. Because of the lack of sufficient testing in pediatric populations, clinicians are left to extrapolate adult efficacy and toxicity data to their young patients. However, dose extrapolation from adults to neonates and children is not straight-forward and presents a major clinical challenge. Pediatric patients are not just small versions of adult patients, and dosing cannot be adjusted based solely upon bodyweight. The composition of bodily water and fat, plasma proteins, drug clearance rates, and hormone levels change rapidly throughout postnatal development and play a significant role in drug disposition

[4]. All of these factors should be considered collectively when selecting a dosing regimen for a pediatric patient, however they are largely overlooked in today's clinical practice.

Due to insufﬁcient clinical testing in neonates and children, approximately 45% of drugs administered to pediatric patients are used off-label, without proper dosing guidelines and disposition proﬁles. Off-label drug use is associated with unforeseen side effects and a higher rate of adverse reactions. ADRs are unfortunately common in infants and children, and nearly half of total reported ADRs from populations of all ages occur in children under the age of four [5].

Infants younger than the age of one are considered to be the most susceptible to ADRs [6].

Whereas drug reactions in younger patients tend to be non-life threatening, with skin rash and Chapter 1 3 gastrointestinal disturbance being the most common, up to 10% of reported pediatric adverse effects are serious enough to cause hospitalization [7]. The classes of drugs most frequently associated with adverse side effects include antibiotics, antipyretics, and non-steroidal anti- inﬂammatories. Anticonvulsants and antineoplastic agents, while not as commonly utilized, are known to cause the more severe ADRs in neonates and children [8].

Off-label drug usage in children increases the risk of unanticipated and unpredictable side effects. Drug responses in pediatric patients often differ greatly from responses in adults, com- plicating the extrapolation of adult drug data to children. The toxicity and efficacy profile of a given drug can vary depending on the age of the patient. For example, at toxic doses, the antibiotic chloramphenicol will cause aplastic anemia in adults. In newborns, however, chloramphenicol toxicity presents as Gray baby syndrome, and causes vomiting, gray coloring of the skin, and hypotension [9]. On the contrary, younger patients may also be resistant to the effects of a drug because of immature drug clearance function. Neonates treated with acetaminophen for pain management and fever reduction, for example, appear to be more resistant to toxicity than adults. Acetaminophen is notorious for causing hepatotoxicity if taken in large enough doses due to accumulation of the toxic drug metabolite, however this adverse event is rarely reported for infant populations [10]. The goal of drug therapy is the same for pediatrics as it is for adults: to maximize therapeutic efficacy while minimizing toxicity. However, efficacy and toxicity varies drastically between children and adults, even when they are treated with the same drug at the same dose. There is a serious need to further understand the mechanisms that underlie differential drug responses based upon age and postnatal developmental stage.

Identifying the physiologic factors and cellular mechanisms that contribute to drug responses Chapter 1 4 in children is imperative to decrease ADRs and develop a more personalized model of pediatric medicine.

1.2 Drug-Drug Interactions

A drug-drug interaction (DDI) occurs when the administration of one drug alters the efficacy or toxicity of another drug taken concurrently. In an increasingly medicated population where taking several therapeutic drugs at once is increasingly common, DDIs are a significant cause of variability in drug responses between people. Polypharmacy, or the use of multiple medications simultaneously, is especially prevalent in geriatric and chronically ill populations due to comorbidities, and significantly increases DDI risk in these patients [11]. The adverse effects of drug interactions can range from mild, such as gastrointestinal disruption, to severe, including arrhythmia, hepatoxicity, hemorrhage, and even death [12]. Approximately 1% of total hospital admissions are due to adverse effects caused by DDIs [13]. Additionally, drug interactions present a clinical challenge if they attenuate the efficacy of another medication. Many medications that are known to interact with the action of other drugs are well-documented and physicians are advised to avoid certain drug combinations. For example, the interaction between the anticoagulant warfarin and many antibiotics is widely known. Antibiotics are commonly prescribed medications, and can alter the metabolism of warfarin as well as the gut microflora that produce vitamin K. Both of these factors increase the efficacy of warfarin, and can cause dangerous excessive bleeding in patients. Accordingly, the dose of warfarin is typically decreased in patients that are taking an antibiotic concomitantly [14]. Despite our knowledge on Chapter 1 5 drug combinations that can cause adverse effects and efforts to avoid them, many patients still experience unexpected and undesirable drug responses as a result of DDIs.

Pediatric patients are also highly susceptible to adverse events caused by DDIs, perhaps even more so than adults. Although polypharmacy tends to be less common in younger populations due to fewer comorbidities, 50-75% of hospitalized children are exposed to potential DDIs [15].

The most common types of prescription drugs indicated in pediatric drug interactions are opioids, antibiotics, neurologic agents, gastrointestinal agents, and cardiovascular agents. In children that experience clinically serious DDIs, the most frequently reported adverse events are respiratory depression, bleeding risk, QT interval prolongation, central nervous system depression, and hyperkalemia [16]. Signiﬁcant drug interactions can also occur with over-the-counter medications, such as ibuprofen, which are routinely given to children without physician consult.

Due to the lack of clinical drug testing in pediatric populations, physicians are left to extrapolate adult dosing and DDI data to infants and children. Even though multiple studies indicate pediatric patients cannot be treated as ’small adults’ in terms of drug pharmacokinetics, most methods of pediatric dose adjustments are based on simple algorithms that extrapolate doses from adult data based on body weight, height, or body surface area. However, one algorithm is not appropriate to use for determining equivalent doses for all stages during postnatal development. Additionally, utilizing more sophisticated physiologically based pharmacokinetic modeling (PBPK) requires robust input data, which is typically unavailable in sufﬁcient amounts from infant and children populations [4, 17]. There is still a critical need to understand other factors that contribute to enzyme induction, DDI development, and pharmacokinetic differences in newborn, infant, and child populations to better determine drug selection and dosage. Chapter 1 6

1.3 Aim of the Project

There is a serious lack of adequate clinical drug testing in pediatric populations. However, it is unlikely that all clinical trials will involve more infants or children due to ethical, technical, and financial constraints. Sufficient data on pediatric responses to existing drugs is also usually limited or not even available. The extrapolation of adult clinical data and dose selection to infants and children is currently flawed, and pediatric drug responses can be unpredictable and unique from adult responses. Developing a deeper understanding of the mechanisms that cause pediatric drug responses to differ from adult drug responses is key to preventing ADRs and DDIs in younger patients in the absence of thorough clinical testing. Making informed, developmental stage-appropriate decisions for drug choices and dosing regimens can improve precision medicine in pediatric patient populations.

The goal of this research was to examine the induction of hepatic CYP enzyme expression and activity in response to drug treatment during postnatal maturation. A mouse model was utilized to investigate the long-term alterations to drug metabolism and CYP expression following exposure to two pharmaceutical compounds, phenobarbital and phenytoin, at the neonatal age. Acute CYP induction in response to drug treatment was compared across a range of ages beginning at the neonatal stage and extending through adulthood. The differences in short-term enzymatic responses between neonates and adults were also examined. To better understand the molecular mechanisms that contribute to the regulation of CYP expression at different ages during postnatal maturation, a proteomic study was implemented and changes in post-transcriptional modiﬁcations were revealed. Chapter 1 7

The results generated from this research will provide fundamental data to developmental pharmacological researchers, as well as pediatric clinicians, on how drug responses can differ depending on age and stage of postnatal maturation. Although the majority of this work was performed in a mouse model, the information can be utilized as proof of a concept that may one day be validated in humans. Chapter 2

Literature Review and Background

2.1 Pharmacokinetics and Drug Responses in Adults

The pharmacokinetic and pharmacodynamic properties of a drug determine its effect in the body. Pharmacokinetics involves characterizing the absorption, distribution, metabolism, and excretion of drug compounds in relation to their plasma concentration over time. Pharmacody- namics is focused on studying the physiologic effects of a drug on the body, such as receptor binding and biological outcomes [18]. Genetic, physiological, and environmental factors can all have an impact on a drugs pharmacokinetic and pharmacodynamic profile in a particular patient. Genetic factors, such as mutations or single nucleotide polymorphisms (SNPs), may affect the expression of proteins that are either drug targets, or enzymes involved in drug clearance. Physiologic factors including age, sex, and comorbidity, and environmental factors including additional medication usage, toxicants, and diet, can also influence concentrations of a drug in the circulation and impact drug response [19, 20]. To create the most favorable response 8 Chapter 2 9 to a therapeutic agent while minimizing adverse effects, it would be beneficial to incorporate all factors that may impact pharmacodynamic and pharmacokinetic properties when selecting a drug and dose for an individual patient.

The biological processes involved in pharmacokinetics are particularly variable among patients

[21]. In order for drugs to be therapeutically efﬁcacious while also minimally toxic, their plasma concentrations must be maintained within a therapeutic range. Alterations to the absorption, distribution, metabolism, or excretion of a drug will affect bioavailability and, ultimately, patient response. Absorption can be inﬂuenced by food intake and the permeability of the intestinal wall. Distribution is largely dependent on a persons body weight and composition.

Kidney function and the expression of drug transporters contribute to determining the rate of elimination. Drug metabolism, however, can be considered the process that has the widest range of interindividual variation. The liver is responsible for the majority of metabolism and is often regarded as the most important organ in biotransformation, although the kidneys and intestines also play roles in metabolizing ingested compounds. Hepatic drug metabolism is heavily in-

ﬂuenced by factors such as sex, age, diet, disease state, circadian rhythm, and polypharmacy, which all signiﬁcantly contribute to drug disposition [22].

2.1.1 Hepatic Drug Metabolism

The majority of xenobiotic metabolism, or biotransformation, occurs in the liver. This process is intended to protect the body from potentially harmful exogenous chemicals by increasing their solubility for efficient excretion in the urine. Endogenous compounds, such as steroids, Chapter 2 10 also undergo metabolism in the liver to allow the conversion into their active forms or elimination via degradation. Nearly all oral pharmaceutical drugs will undergo biotransformation by enzymes in the liver to facilitate their detoxification into drug metabolites, and this is referred to as first-pass metabolism [23]. The overall process of drug metabolism is divided into three phases that ultimately enhance the water solubility of a compound in order for it to be efficiently excreted out of the body. Phase I of drug metabolism is catalyzed mainly by cytochrome P450 enzymes (CYPs), which belong to a family of oxidases that introduce reactive or polar groups to the chemical structure of their substrate compounds via oxidation, reduction, or hydrolysis reactions. Reductases, flavin-containing monooxygenases (FMOs), perox- idases, monoamine oxidases, dehydrogenases, and carboxylesterases also contribute to phase

I of metabolism. Phase II enzymes, including UDP-glucoronosyltransferases (UGTs), sulfotransferases (SULTs), N-acetyltransferases (NATs), and glutathione S-transferases (GSTs), are responsible for conjugating the modiﬁed drug compound to other polar compounds to further increase molecular polarity. During phase III, the now hydrophilic drug metabolites are elimi- nated from the liver via proteins belonging to the ATP-binding cassette (ABC) and solute carrier

(SLC) families of transporters. Metabolites and any unmodiﬁed drug compounds can then be excreted from the body in the urine [24].

The efficacy or toxicity of a drug is largely dependent on the rate of hepatic drug metabolism in a patient as this process determines bioavailability. Slower rates of metabolism cause an increased concentration of the parent drug compound in circulation and a delay in excretion of drug metabolites. High concentrations of active drug compound in the blood are important for Chapter 2 11 its therapeutic efficacy, however excessive levels can lead to toxicity. On the other hand, elevated rates of metabolism decrease the concentration of parent drug compound in the circulation while increasing the amount of drug metabolites produced. This becomes problematic if the parent compound is the active form of the drug, because plasma concentrations will fall below the therapeutic range and a patient might experience a lack of efficacy. The buildup of metabolites can also lead to adverse side effects if they are toxic. In some cases, drug metabolites are responsible for eliciting the therapeutic effect and drug compounds must undergo biotransformation to reach their active forms. If so, decreased rates of metabolism could cause attenuated drug efficacy while increased rates could cause the drugs effects to be excessive [25].

2.1.2 Cytochrome P450 Enzymes

The hepatic CYPs are oxidase enzymes that utilize heme as a cofactor to catalyze monooxygenase and hydroxylation reactions. They are responsible for the metabolism of exogenous drugs, environmental toxins, and dietary components, as well as the biosynthesis and degradation of endogenous steroids, lipids, and vitamins. Most life forms, including primates, rodents, ca- nines, plants, fungi, and bacteria, express CYP enzymes in some amount, and approximately

60 CYP genes have been identiﬁed in humans. The individual enzyme isoforms are classi-

ﬁed according to their family number and a subfamily letter, followed by a number to identify the speciﬁc enzyme. All mammals share the same 18 CYP gene families, but the number of subfamilies and enzymes in each family may differ depending on the organism [26].

Enzymes belonging to the CYP1, CYP2, and CYP3 families contribute most signiﬁcantly to the metabolism of pharmaceutical drugs. Approximately 80% of currently marketed drugs are Chapter 2 12 substrates for CYP enzymes and will undergo phase I metabolism, which accounts for nearly

50% of overall drug elimination [27]. Drug compounds may be substrates for multiple CYP isoforms, and one isoform may be responsible for metabolizing hundreds of different substrate drugs. In contrast, other isoforms have stringent speciﬁcity and only bind one or two substrates.

CYP3A4 is often considered to be the most important CYP in terms of drug metabolism due to its involvement in the biotransformation of more than 50% of medications on the market, from antidepressants to chemotherapeutic agents to statins. It is also the most abundant isoform in the liver, accounting for 30-40% of the total hepatic CYP abundance . Approximately

25% of clinically utilized drugs are metabolized by CYP2D6 and include antipsychotics, beta- blockers, opioids, and anti-arrhythmics. CYP2C19 metabolizes up to 10% of current clinical medications, most notably antiepileptic drugs, antidepressants, and proton-pump inhibitors.

Also a member of the CYP2C subfamily, the enzyme CYP2C9 is known to metabolize up to

100 marketed drugs, including non-steroidal anti-inﬂammatories (NSAIDs) and the anticoagulant warfarin. CYP2B6 has far fewer substrates than the previously mentioned CYPs, but this isoform plays a signiﬁcant role in the metabolism of the antidepressant bupropion, the antiretroviral efavirenz, and even nicotine [28].

Environmental toxicants and dietary compounds can also undergo metabolism by CYP enzymes. CYP2E1 is responsible for the majority of the metabolism of acetaminophen and anesthetic medications, but also plays a large role in the detoxiﬁcation processes of ingested caffeine and ethanol. CYP1A2 is also known to metabolize caffeine as well as various other drugs including antidepressants [28]. Chapter 2 13

CYP enzymes also play important roles in the biosynthesis and degradation of many endogenous compounds. Whereas CYP2D6 appears to metabolize drugs exclusively, most drug- metabolizing CYPs are also responsible for the metabolism of steroid hormones, fatty acids, retinoids, prostaglandins, cholesterol, and bile acids produced by the body [29]. Bile acids, which are products of cholesterol catabolism necessary for fat digestion and absorption, are metabolized by CYP3A4 to prevent their accumulation and associated toxicity Chen et al.,

2014). On the other hand, CYP7A1, CYP8B1, and CYP27A1 are responsible for synthesizing bile acids from cholesterol, although these enzymes do not play a major role in drug metabolism

[30]. Endogenous steroids also undergo extensive metabolism by CYP enzymes either for their activation or degradation. CYP2B6 catalyzes testosterone 16- and 16-hydroxylation, and has shown to be differentially expressed in male and female liver samples [31]. CYP3A4, 2C9, and

2C19 also metabolize testosterone, along with various other hormones including progesterone and estradiol [32]. Members of the CYP4 family are responsible for fatty acid metabolism and regulate lipid catabolism and storage in the liver [33]. Taken together, CYP enzymes are key players in maintaining the homeostasis of energy utilization, hormone signaling, and nutrient uptake within the liver and throughout the rest of the body. However, their activity varies widely among patients [34].

Due to their roles in the metabolism of endogenous compounds and essential nutrients, alterations to the activity of certain CYPs have been implicated in various diseases. Because CYPs are so prevalent in the liver, it makes sense that many CYPs are involved in the pathogenesis of hepatic disorders such as alcoholic liver disease, non-alcoholic fatty liver disease (NAFLD), steatohepatitis, and hepatocellular carcinoma. Enhanced CYP2E1 activity, in particular, has Chapter 2 14 been associated with with alcoholic liver disease as well as NAFLD, due to the creation of reactive oxygen species that damage cellular membranes and mitochondria [35]. CYPs are also highly expressed in the kidney, brain, cardiovascular system, and lungs, and their altered expression has also been implicated in diseases of these organs as well [36–38]. CYP expression activity levels not only determine patient responses to drug treatment, but they may also play a large role in the etiology of the diseases themselves.

2.1.3 Cytochrome P450s in Drug-Drug Interactions

While CYP enzymes are responsible for the detoxiﬁcation of exogenous drug compounds, their gene expression and protein function can simultaneously be modiﬁed by the administration or ingestion of other xenobiotics. Certain medications, as well as dietary components and other environmental chemicals, have the ability to induce or inhibit CYP enzymatic function.

Because of this, concomitant usage of therapeutic drugs with the ability to alter CYP activity is a major contributor to the cause of DDIs. Compounds known as inducers enhance CYP activity by increasing the amount of CYP enzymes present via the up-regulation of gene transcription.

Inducers tend to induce speciﬁc CYP isoforms rather than CYP activity as a whole. DDIs occur when an inducer medication is administered at the same time as a victim drug that is a substrate for the induced CYP. The increased rate of metabolism of the victim drug can cause attenuated therapeutic efﬁcacy due to increased active drug compound clearance. Adverse side effects may also occur due to a buildup of drug metabolites that are toxic. On the other hand, certain medications require bioactivation by CYP enzymes to achieve their therapeutically active form. Chapter 2 15

Induced CYP activity, in this case, would enhance therapeutic efﬁcacy and could also lead to unwanted side effects [39].

CYP activity can also be decreased by drug administration via protein inhibition mechanisms.

In this case, metabolism of substrate drugs taken simultaneously would be down-regulated.

Increases in toxic side effects are possible due to excessive circulating concentrations of the active drug compound and slowed clearance. Therapeutic efﬁcacy may also be attenuated if the drug requires bioactivation by CYPs to reach its biologically functional form [40].

2.2 Pharmacokinetics and Drug Responses in Infants and

Children

The process of selecting medications and dosing regimens for pediatric patients must differ from that of adult patients due to significant alterations in drug clearance during postnatal development. The mechanisms involved in drug clearance vary drastically during the first few years of life and must be taken into consideration. It is also necessary to evaluate drug clearance based on a pediatric patients age, rather than categorizing all infants and children into one pediatric category [41]. The population of pediatric patients is further broken down into specific age groups based upon FDA Guidance from 1998 (Guidance for Industry: General Considera- tions for Pediatric Pharmacokinetic Studies for Drugs and Biological Products). Neonates are classified as ranging from birth to 1 month, infants are 1 month to 2 years, developing children are 2-12 years, and adolescents are 12-16 years. Overall growth and the development and maturation of organs occurs rapidly during the first two years of life, while body composition Chapter 2 16 continues to change throughout childhood and adolescence. Body size, the volume of distribution, and drug clearance processes have a large impact in determining a drugs pharmacokinetic profile and are therefore dependent upon age and stage of development. Expression of proteins that are pharmacological drug targets and disease pathogenesis may also change throughout postnatal childhood development, and could therefore affect a compounds pharmacodynamics as well. All of these factors contribute to determining drug response and therapeutic efficacy at specific pediatric ages, however the effects are poorly understood [42].

All aspects of drug clearance involving the absorption, distribution, metabolism, and excretion of ingested compounds have been proven to be altered in pediatric populations in comparison to adults. Absorption determines the onset of drug action as well as the effective dose and can be inﬂuenced by gastric pH and intestinal permeability, both of which are altered in infants [43].

Neonates and infants have signiﬁcantly higher gastric pH levels than adults that can increase the bioavailability of acid-labile drugs, while reducing the bioavailability of drugs of weaker acidity, such as phenobarbital and phenytoin [44]. Differences in intestinal permeability and the expression of intestinal transporters that allow the passage of drug compounds from the intestinal lumen into circulation also affect the bioavailability of many medications in newborns and infants [3]. Following absorption, compounds are distributed to different compartments within the body according to physiochemical properties. Drug distribution is impacted by protein binding in the plasma, and neonate and infants tend to have lower concentrations of proteins, such as albumin, in their blood than adults. This decreases protein binding and increases the amount of free, unbound drug compounds in the circulation [45]. Changes in body composition also affect distribution. Infants have higher volumes of total body weight and lower volumes of Chapter 2 17 total body fat than adults, and this alters the volumes of distribution for water- and fat-soluble drugs [3].

Metabolism, as previously discussed, has a signiﬁcant impact on drug efﬁcacy and toxicity.

Members of phase I and phase II metabolic enzymes are not be fully expressed in neonates and infant livers, and therefore drug metabolism capabilities may be reduced. The ontogenic expression of P450 drug metabolizing enzymes, in particular, during postnatal ages will be outlined in more detail in later sections. There are also perinatal forms of drug-metabolizing enzymes that are enriched specifically in the fetus and immediately after birth, while absent in adults. These metabolic differences affect the plasma concentration of administered drugs, along with their metabolites, which could produce age-specific toxicities [46]. Finally, excretion is largely dependent upon kidney function along with renal blood flow. Renal excretion is highly variable in neonates and infants, and can occur at rates higher or lower than those of adults depending on the drug [47].

2.2.1 Postnatal Liver Maturation

At birth, the liver does not possess the fully mature metabolic capabilities that are seen in adults.

The fetal and neonatal liver acts as a hematopoietic organ, generating blood cells. During the

first several years of life, the neonatal liver undergoes a postnatal developmental process that allows it to transform from a proliferating, hematopoietic organ into a differentiated, metabolic organ. The complete process takes up to five years in humans, but occurs most rapidly during the first two years of life [48]. During the beginning of this period of postnatal growth, the liver is responsible for producing hematopoietic cells and is in a state of rapid cellular proliferation Chapter 2 18 so it can grow in size along with the rest of the body. The hepatocytes also differentiate simultaneously at this time to achieve mature gene expression profiles and enzymatic activities.

Because of this, neonatal and infant livers have limited metabolic capabilities and are often more sensitive to drug treatment than adults due to lowered rates of drug clearance [49].

This process of postnatal liver maturation is remarkable in that cell proliferation and differentiation occur simultaneously. However, the cellular mechanisms regulating this process are relatively understudied. There is limited research regarding the biological pathways, transcription factors, and signaling events that orchestrate both hepatic growth and maturation during the postnatal stage. The Wnt/-catenin signaling pathway has been shown to be involved in postnatal liver growth and cell proliferation, and is necessary for the liver to increase in size

[50]. The yes-associated protein (Yap) and the hippo kinase pathway have also been proven to regulate postnatal liver growth, cell proliferation, and the expression of genes involved in hepatic bile acid and retinoic acid metabolism [51, 52]. Several studies have also demonstrated that the Notch signaling pathway, along with the ligand Jag, play a critical role in hepatic biliary duct formation and differentiation during early postnatal life [53, 54]. On the other hand, transcription factors E2F7 and E2F8 have been shown to regulate liver growth by controlling excessive cell proliferation and suppressing hepatic tumor formation during postnatal development [55]. Insulin signaling is known to be an important pathway in mammalian organ growth and development in general, and studies have demonstrated the varying expression of insulin- like growth factors 1 and 2 (IGF1; IGF2) during postnatal liver development, possibly due to regulation by the CCAAT-enhancer binding protein (C/EBP ) [56]. The liver is responsible for producing circulating IGFs, and IGF2 is considered to be most abundant in fetal and Chapter 2 19 early postnatal life, while IGF1 is believed to be the predominant form in adults [57]. Several nuclear receptors, including hepatocyte nuclear factor 4 (HNF), pregnane X receptor (PXR), constitutive androstane receptor (CAR), and aryl hydrocarbon receptor (AhR) are shown to be activated at different points during postnatal development and contribute to liver differentiation by inducing the transcription of mature hepatocyte-speciﬁc genes, including drug-metabolizing enzymes, xenobiotic transporters, and albumin.

The expression of cellular components involved in epigenetic gene transcription regulation also change in the liver throughout postnatal development. Long non-coding RNAs (lncRNAs) have recently been acknowledged for their roles in regulating gene and protein expression. Many hepatic lncRNAs were shown to be enriched at different stages during postnatal development, and their expression levels were highly correlated with the expression of their neighboring coding genes involved in regulating hematopoiesis, cell proliferation, and metabolism [58, 59].

2.2.2 Ontogeny of Cytochrome P450s

The majority of CYP enzymes are expressed at significantly lower levels at birth and during the first few years of life than in adults. In fact, the fetal and neonatal liver is considered to only have 30-60% of the total CYP content of adults. Therefore, newborns and infants are considered to have overall significant reductions in CYP-mediated drug metabolism. During postnatal liver maturation, each CYP isoform appears to follow a distinct pattern of ontogenic regulation in which its transcription, expression, and activity varies depending upon age. Three groups of CYP gene expression patterns appear and consist of the CYPs that are enriched during Chapter 2 20 the neonatal period but become undetectable in adults, the CYPs that increase in expression quickly after birth, and the CYPs that become up-regulated during adolescence [60, 61].

As previously stated, most hepatic CYPs involved in drug metabolism are expressed at very low levels in newborns at birth. During the ﬁrst two years of life, expression of these enzymes increases rapidly with age. Expression may then steadily increase throughout childhood and adolescence until adult expression levels and full enzymatic activity are achieved. CYPs that follow this postnatal developmental pattern include CYP1A2, 3A4, 2C9, and 2C19. All of these isoforms are nearly undetectable in the liver in the fetus and immediately after birth. CYP1A2 becomes detectable between one and three months of age, and is expressed at 30% of adult levels at age one. Between two and three years of age, enzymatic activity is comparable to that of adults. The CYP2C isoforms increase in expression more rapidly after birth, reaching 30% of adult values by one month of age, and 50% of adult values by ﬁve months of age. Expression then slowly rises to adult levels over the next few years and may not be fully mature until age

10 [62, 63]. Like CYP1A2, CYP2E1 is expressed at around 40% of adult levels by age one, but also might not reach levels comparable to adults until age 10 [64]. CYP3A4 expression rapidly increases during the ﬁrst month after birth as well, and reaches 50% of adult values between 6 and 12 months of age. Interestingly, the literature also suggests that CYP3A4 activity might be higher in infants after one year of age than in adults [65]. The expression pattern of CYP2B6 is unique in that it appears to be expressed more highly immediately following the perinatal stage during the ﬁrst one to three months of life than in adults. In infants aged one year or older, levels decrease and become more comparable to those of adults [66].

Fetal forms of CYP enzymes also exist in which the highest levels of expression occur during Chapter 2 21 gestation in the fetal liver and immediately after birth. CYP3A7 is the best-studied example of this. This isoform is considered to be the fetal form of the CYP3A family of enzymes. Its expression is highest in the fetal liver and in newborns, but quickly decreases to undetectable levels within the ﬁrst few months of life. This decrease in CYP3A7 expression correlates well with the increase of the expression of the adult isoforms CYP3A4 and CYP3A5 during infant development, and the CYP3A postnatal switch has been well documented in the literature.

In mice, the Cyp3a switch is associated with a change in epigenetic histone modiﬁcations.

Suppression of the fetal Cyp3a16 during postnatal liver maturation coincided with increased repressive H3K27me3 and decreased H3K4me2 in the gene. Simultaneously, the expression of the adult isoform Cyp3a11 increased during postnatal maturation and coincided with increased levels of H3K4me2 within the gene [67]. This suggests that epigenetics and play a signiﬁcant role in regulating the developmental expression patterns of CYP enzymes.

Several genetic and environmental factors have been shown to affect the natural ontogeny of

CYP activity during postnatal development. For example, formula-fed infants were shown to gain CYP1A2 and CYP3A4 metabolic activity more rapidly than infants that were breastfed

(Blake et al., 2006). At six months of age, infants that are formula-fed clear caffeine from their systems more rapidly than infants that are breastfed [68]. The speciﬁc components of infant formula also seem to have an impact on CYP ontogeny, with infants fed soy-based formulas having higher levels of CYP3A activity [69]. Depletion of the nuclear farnesoid X receptor

(FXR) in mice appeared to delay or attenuate the expression of many CYPs during postnatal maturation, indicating a possible role of bile acids in regulating CYP ontogeny [70]. A further discussion of the endogenous factors that contribute to orchestrating the changes in postnatal Chapter 2 22 expression of CYP enzymes and activity are discussed in later sections. l

2.2.3 Induction of Cytochrome P450s in Pediatric Populations

The process of CYP enzymatic induction is normally studied in the context of the fully mature, adult liver. In this case, induction is a process that is temporary and reversible, and induced CYP expression returns to basal levels once the inducer drug is discontinued and completely cleared from circulation. Induction in the context of the neonatal, infant, and child liver during postnatal maturation is an understudied area and not well understood. However, pediatric patients are still exposed to a wide range of drugs that have capabilities to induce CYP expression, and the outcomes of enzyme induction during stages of postnatal liver growth and differentiation are relatively unknown.

Very few studies have investigated whether the inducibility of CYPs varies with age, especially during postnatal development. One study showed that the inducibility of hepatic drug- processing genes during mouse postnatal development was age-dependent and correlated with the basal expression of the enzyme. However, this study only examined induction at the mRNA level, and produced transcriptional induction via administration of direct nuclear receptor ligands, not pharmaceutical drugs [71]. Another study compared the inducibility of hepatic carboxylesterases, which also play a role in drug metabolism, between neonatal and adult mice.

In response to phenobarbital treatment, the carboxylesterases tested were induced to a greater extent at the mRNA, protein, and catalytic levels in neonates than in adults. The induction at the neonatal age also caused more rapid hydrolyzation of clopidogrel and the accumulation of plasma lipids, indicating significant functional consequences [72]. There is a clear need to Chapter 2 23 further examine the inducibility of specific CYP isoforms at various ages during postnatal development in response to drug treatment, as this may greatly affect the efficacy and toxicity of other medications that are administered concurrently to pediatric patients.

2.2.4 Use of Antiepileptic Drugs in Pediatric Patients

Neonatal seizures occurring in newborns immediately after delivery are a fairly common occurrence, with a frequency of three cases per 1,000 live births in the United States. Child- hood epilepsy is also a common disease affecting pediatric populations at all ages of postnatal development. Treatment of these disorders involves the administration of one or multiple antiepileptic drugs (AEDs) to prevent and control seizures that could be detrimental to brain development [73]. A wide variety of AEDs can be prescribed to children suffering from either chronic epilepsy or an acute seizure episode and include the ﬁrst generation AEDs (carbamazepine, phenobarbital, phenytoin, clonazepam, and valproic acid), second generation AEDs

(felbamate, gabapentin, lamotrigine, and topiramate), and the most recently developed third generation AEDs (lacosamide, perampanel, retigabine, stiripentol). However, most second and third generation AEDs are not approved for pediatric use and are prescribed to infants and children off-label [74, 75]. Neonates generally clear AEDs at much slower rates than older patients, presumably due to low levels of hepatic CYPs. On the contrary, infants and younger children have higher clearance rates of AEDs than adults and often require dosage adjustments, which we can hypothesize is due to the changing expression of CYPs in the liver during this period of postnatal development [76]. Additionally, many AEDs are well known to cause CYP induction and are implicated in common DDIs, which becomes problematic as epileptic patients are Chapter 2 24 frequently treated with multiple medications to control seizures and other comorbidities [77].

AEDs also have narrow therapeutic plasma concentration windows and are prone to causing adverse side effects in patients. Overall, the usage of AEDs comes with several risks, and the chance of adverse effects increases even more signiﬁcantly in pediatric patients [78].

AEDs that are considered ﬁrst-line therapy for treating neonatal seizures include phenobarbital, phenytoin, diazepam, and lorazepam, with phenobarbital and phenytoin being the most commonly used clinically [79]. Simple and partial seizures in infants and children are also commonly treated with phenobarbital and phenytoin, and these drugs are considered ﬁrst-line therapy for treating generalized and infant febrile seizures as well. However, newer AEDs are beginning to be preferred due to the lower occurrence of adverse effects. Unfortunately, second and third generation AEDs are more expensive, and families without health insurance or those living in poorer countries are often left to use the more inexpensive AEDs such as phenytoin and phenobarbital to treat their children. Additionally, phenobarbital appears to remain the most effective treatment option [80].

Phenobarbital (trade name Luminal) was ﬁrst used clinically in 1904 and is on the World Health

Organizations List of Essential Medicines. It is a barbiturate drug that prevents convulsions and induces sedation by increasing the activity of the inhibitory neurotransmitter GABA and blocking glutamate excitatory signaling. Treatment with phenobarbital begins with a loading dose of 20 mg/kg bodyweight, followed by subsequent maintenance doses of 3-5 mg/kg. Serum concentrations of 40 g/ml are desirable for therapeutic efﬁcacy in adults, however it has been shown that concentrations up to 100 g/ml are more desirable to prevent seizures in infants.

Phenobarbital is metabolized primarily by CYP2C9 in the liver, with smaller contributions Chapter 2 25 by CYP2C19 and CYP2E1. In adults and neonates, the half-life of the drug is around 100 hours, however in young infants, the half-life is decreased and may only be around 60 hours.

Phenobarbital is also a potent inducer of CYPs, speciﬁcally CYP2B6, via the activation of CAR.

Enzymes in the CYP2C and CYP3A families are also induced by phenobarbital treatment.

Because of this, phenobarbital is implicated in many DDIs and often causes adverse effects in patients [81].

Phenytoin (brand name Dilantin) is often better tolerated than phenobarbital for seizure treatment and is frequently preferred over phenobarbital for the treatment of neonatal seizures because of this. It is also on the World Health Organizations List of Essential Medicines and is inexpensive compared to other AEDs. Phenytoin is also occasionally used for the treatment of arrhythmias. It exerts its therapeutic effects through a different mechanism of action from phenobarbital by blocking voltage gated sodium channels to prevent repetitive ﬁring of action potentials in neurons in the brain. A loading dose of around 20 mg/kg is also recommended for pediatric patients [82]. Phenytoin is metabolized by CYP2C9 and CYP2C19 in the liver, although it has a much shorter half-life (10-15 hours) than phenobarbital. Because CYP2C9 and 2C19 are barely expressed in the newborn patient, the half-life of phenytoin is prolonged relative to older infants and adults. However, the elimination rate of the drug is higher in infants and children than it is in adults, perhaps due to the lack of CYP2C9-mediated metabolism or variability in the inducibility of the enzyme. Because of this, phenytoin doses in pediatric patients must be 50-100% higher than in adult patients to achieve comparable therapeutic plasma concentrations [83]. Like phenobarbital, though, phenytoin is also a potent inducer of CYPs in the liver, particularly CYP1A2, CYP2B6, CYP2C9, CYP2C19, and CYP3A4. The mechanism Chapter 2 26 by which phenytoin causes enzyme induction, however, is not well understood, although it is thought that CAR activation is also involved [77, 84].

Because phenobarbital and phenytoin are potent inducers of general CYP activity in the liver, they are implicated in many DDIs, especially with co-administered AEDs. In hospitalized children with epilepsy, phenobarbital and phenytoin are among the most common causative agents of DDIs. Phenobarbital and phenytoin treatment can reduce serum concentrations of the commonly co-administered AEDs valproic acid and lamotrigine by over 50%, which may attenuate efﬁcacy and warrant dosage adjustments. In children, other DDIs with pain-reducing agents, antimicrobials, and immunosuppressants are most frequently reported [85].

Both phenobarbital and phenytoin are considered ﬁrst-line therapies for the treatment of seizures in newborns, and are commonly used for the treatment of epilepsy in children. Because of this, these are the AEDs we have focused on for the duration of this research. Both drugs also activate the CAR nuclear receptor indirectly to induce CYP expression and drug metabolism in the liver. A previous study has shown direct CAR activation at the neonatal age can cause permanent changes to drug metabolism in the liver of mice [86] via epigenetic memory, so it is critical to investigate the consequences of CAR-activating drug exposure during postnatal life.

Whereas they are similar in the fact that they can both activate CAR to induce CYP activity in the liver, they still have unique pharmacokinetic proﬁles, chemical structures (Figure 2.1), and mechanisms of action. Phenobarbital is also known to activate CAR via its action on the epidermal growth factor receptor (EGFR) (Figure 2.2), while the mechanism for CAR activation by phenytoin is not understood. It would therefore be beneﬁcial to study them separately to better understand whether they have different effects on CYP induction during postnatal development. Chapter 2 27

FIGURE 2.1: Chemical structures of phenobarbital and phenytoin. Obtained from Pub- Chem (pubchem.ncbi.nih.gov).

2.3 Mechanisms of Cytochrome P450 Regulation

2.3.1 Genetic Polymorphisms

Heritable changes to the genetic sequence of drug-metabolizing CYP genes can both enhance as well as suppress their enzymatic activity. Copy number variants (CNVs) and single nucleotide polymorphisms (SNPs) within the gene promoter typically encompass gain-of-function variants that increase the number of functional gene copies, thus up-regulating the rate of substrate metabolism. On the other hand, loss-of-function polymorphisms typically affect CYP gene splicing and expression, rather than transcription, and act to down-regulate enzymatic activity.

The phenotypes associated with speciﬁc polymorphisms are referred to using the terms poor metabolizer, intermediate metabolizer, extensive metabolizer, and ultrarapid metabolizer. Ex- tensive metabolizers are considered the normal phenotype and represent the majority of the patient population. Poor metabolizers carry polymorphic alleles that do not code for a functional Chapter 2 28

FIGURE 2.2: Direct and indirect pathways leading to CAR activation by xenobiotics in mouse. Upon activation, CAR dissociates from heatshock protein 90 (hsp90) and cytoplas- mic CAR retention protein (CCRP) in the cytoplasm and translocates to the nucleus where it associates with transcriptional coactivator CREB-binding protein (CBP)/p300 (not pictured). The complex binds to response elements (PBREM) in the promoter of target genes, including Cyp2b10, to up-regulate transcription. Activation of CAR can either be direct, as through the direct binding of the ligand TCPOBOP, or indirect, as through the binding of phenobarbital (PB) to the epidermal growth factor receptor (EGFR). PB antagonizes EGFR, which leads to activation of RACK-1 and protein phosphatase 2A (PP2A), which dephosphorylates CAR, leading to its activation and translocation into the nucleus [87]. Chapter 2 29

CYP protein, and these patients typically require smaller doses of a drug. Intermediate metabolizers, while similar to poor metabolizers, usually carry deﬁcient alleles that attenuate metabolic capacity rather than abolish it completely. Ultrarapid metabolizers carry gain-of-function polymorphic alleles and have a greater capacity for metabolic activity and sometimes require higher drug doses to achieve therapeutic effect. However, if the drug requires metabolism to receive its active form (i.e. prodrugs), ultrarapid metabolizers may require lower doses to avoid toxicity, while poor metabolizers may require higher doses to achieve therapeutic efﬁcacy [28].

Many SNPs in drug-metabolizing CYP genes have been identiﬁed and their resulting metabolic phenotypes are known and occasionally utilized clinically. CYP2B6 is highly polymorphic, and the CYP2B6*6 allele, which is associated with a 50-75% decrease in hepatic protein, has the greatest frequency in patient populations. This variant causes decreased clearance of efavirenz, nevarapine, and 5-methadone, which is associated with resulting adverse effects due to toxicity and drug discontinuation [88]. On the other hand, CYP2B6*4 and *22 variants, while less common, cause increased CYP2B6 activity, which leads to up-regulated clearance and decreased efﬁcacy of buproprion, efavirenz, and selgiline [31].

Strong phenotype-genotype correlations have been found for the CYP2C19 gene, in which the *2 and *3 null alleles abolish metabolic activity. Proton-pump inhibitors (PPIs) such as omeprazole and lansoprazole, which increase pH in the stomach, depend on the genotype of the CYP2C19 gene for their efﬁcacy. PPIs actually tend to work better in poor metabolizers, as the drug remains in the circulation for longer and increases gastric pH more signiﬁcantly [89].

On the contrary, the antiplatelet drug clopidogrel requires bioactivation by CYP2C19 for its therapeutic action. Poor metabolizers, in this case, have a higher risk of adverse cardiovascular Chapter 2 30 events due to attenuated efﬁcacy. A black box warning has even been given by the FDA for this drug warning about decreased efﬁcacy in patients with the loss-of-function allele variants [? ].

CYP polymorphisms have also been shown to make an impact on AED metabolism, efﬁcacy, and toxicity. Patients treated with phenytoin that carry variations in the CYP2C9 allele associated with decreased activity are at greater risk of adverse skin and nervous system effects due to higher plasma concentrations of the drug [90]. Genetic testing prior to initiating AED treatment is not typical, however. Instead, physicians are encouraged to routinely monitor plasam levels of both phenytoin and phenobarbital and adjust dosing accordingly [91]. Whether this is a safe and effective method for dosing pediatric epileptic patients is unknown.

Many genetic polymorphisms in CYP genes have been identiﬁed and produce clinical consequences recognized to make an impact in drug responses. However, these genetic variations only account for 10-30% of the inter-individual variability observed in CYP metabolism among patients [92]. Other factors beyond the underlying DNA sequences in CYP genes must be con- tributing to determine the extent of CYP enzyme activity in a patient.

2.3.2 Nuclear Receptors

Nuclear receptors are unique signaling proteins in that they act as both sensors of chemical ligands as well as transcription factors that directly affect the activation or repression of gene transcription. All nuclear receptors share a common structure that consists of a ligand binding domain and a DNA binding domain. Various nuclear receptors have been well-characterized for their roles in the regulation of CYP gene transcription. Nuclear receptors tend to remain in Chapter 2 31 the cytoplasm of the cell until they become activated by an endogenous or exogenous chemical ligand. Following activation, they translocate into the nucleus and bind to speciﬁc response elements in the promoters of the target genes. Nuclear receptors are able to recruit various other co-activator or co-repressor proteins that can then manipulate chromatin architecture to facilitate or suppress transcription of their target genes. Enzyme induction occurs predominantly via the action of nuclear receptors, however basal enzyme expression may also be under the control of constitutively active nuclear receptors that remain bound to the promoter regions of the target genes.

The pregnane X nuclear receptor (PXR), encoded by the NR1I2 gene, and the constitutive androstane nuclear receptor (CAR), encoded by the NR1I3 gene, in particular, control the induced expression of multiple CYPs involved in drug metabolism and have been implicated in many

DDIs. Both form a heterodimer with the retinoid X receptor (RXR) upon activation, then bind to response elements within the DNA to up-regulate the transcription of target CYP genes. PXR can be activated by a variety of xenobiotics, including synthetic glucocorticoids, antiglucocor- ticoids, antibiotics (e.g. rifampicin), antifungals, and herbal extracts (e.g. St. John’s Wort), as well as endogenous compounds such as steroids. This receptor recruits several co-activators, including steroid receptor co-activators (SRC-1 and SRC-2) and PPARγ co-activator (PGC-1α) to facilitate up-regulated transcription of target genes (e.g. CYP3A4) [93]. CAR is activated by several xenobiotics such as AEDs (e.g. phenobarbital) and endogenous steroids such as androstanes. This receptor has also been shown to have high constitutive activity even in the absence of ligands, although CAR is still typically thought to be important in CYP induction in response to drug treatment. CYP2B6 is most commonly used as an example of a CAR target Chapter 2 32 gene, however CAR can also inﬂuence the transcription of many other drug-metabolizing enzymes and transporters. CAR also plays an important role in the regulation of gluconeogenesis, lipid metabolism, and bile acid homeostasis within the liver [94]. Many CYPs, such as CYP2B,

2C, and 3A enzymes, have shown both PXR- and CAR-dependent induction, while others, such as CYP1A enzymes, are only induced by CAR activation [95].

While PXR and CAR are perhaps the best studied nuclear receptors in terms of CYP-mediated drug metabolism, several other nuclear receptors play roles in regulating CYP expression in the liver. The aryl hydrocarbon receptor (AhR) can be activated by enivornmental hydrocarbons such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) to up-regulate the expression of CYP1A and 1B enzymes. The glucocorticoid receptor (GR), which also has target genes including

CYP1A and 1B enzymes, can be activated by the drug dexamethasone. The vitamin D receptor

(VDR), activated by vitamin D3 and lithocholic acid, has been shown to affect the expression of many drug metabolizing genes, including CYP3A4, 2B6, and 2C9. The hepatocyte nuclear factors (HNF1 and HNF4) have also been considered master regulators of CYP gene expression, especially under basal conditions [96]. These receptors have also been implicated in regulating the expression of other hepatic nuclear receptors involved in drug metabolism regulation, including PXR and CAR, suggesting the expression of all nuclear receptors are coordinated in a type of cross-talk network that cooperatively orchestrates overall liver function [97].

2.3.3 Epigenetic Modiﬁcations

Modiﬁcations to the chromatin landscape of CYP genes can up- or down-regulation their transcription, and thus have signiﬁcant affect on the amount of CYP mRNA and protein present Chapter 2 33

in the liver. This often leads to clinically relevant changes to CYP activity, drug clearance,

and drug response. DNA methylation, histone methylation, and histone acetylation have all

been shown to play extensive roles in regulating the expression of relevant drug-metabolizing

CYPs, and understanding the mechanisms offers insight into how the expression of CYPs may

be controlled under different circumstances [98].

DNA methylation is perhaps the best deﬁned epigenetic process known to regulate the expres-

sion of many CYP and other drug-metabolizing genes. DNA methylation is typically associated

with transcriptional repression and the down-regulation of target genes. The mechanism in-

volves the addition of a methyl group to cytosine pyrimidine rings in CpG dinucleotides in the

promoter region or gene body, and is catalyzed by DNA methyltransferase enzymes (DNMTs).

This facilitates long-term gene silencing via the formation of heterochromatin, which prevents

the binding of transcription factors and allows the binding of certain co-repressors [99]. DNA

methylation has been shown to regulate the expression of certain phase I, phase II, and phase III

transporter genes, and is associated with determining the transcription of CYP1A2, CYP2D6, and CYP2E1. Hypermethylation of CYP2D6 and CYP2E1 was associated with increased liver injury rates due to a decreased ability to clear toxic metabolites of anti-tuberculosis drugs [100].

The acetylation of histones, catalyzed by histone acetyltransferase enzymes (HATs), is conversely associated with active gene transcription, as the attached acetyl groups facilitate open chromatin structure that allows the binding of transcription factors. Histone acetylation has been implicated as a mechanism in the induction of CYPs in response to drug treatment, particularly involving the activation of the CAR nuclear receptor. Chemically inhibiting histone deacetylase enzymes (HDACs), which remove acetyl groups thus repressing gene transcription, Chapter 2 34 have been shown to up-regulate the expression of several CYPs, including CYP2B1, CYP2B2,

CYP2B6, and CYP3A4, through the recruitment of CAR [101, 102].

Histone methylation may have either activating or repressing effects on transcription. This process involves the addition of methyl groups to certain lysine residues on histone tails, and depending on the lysine that is methylated, transcription of the target gene may be up- or down- regulated. For example, histone 3 lysine 4 methylation (H3K4me3), is associated with increases in transcription and the reaction is catalyzed by the Su(var)3-9, Enhancer of zeste, Trithorax- containing (SET) family of histone methyltransferases. On the other hand, histone 3 lysine

27 methylation (H3K27me3) is associated with transcriptional repression, and this reaction is catalyzed by the Enhancer of zeste 2 (EZH2) methyltransferase. Both of these marks are implicated in regulating the expression of the CYP3A4 gene. In response to rifampicin treatment, which activates PXR, increases of H3K4me3 and decreases in H3K27me3 are observed in the promoter region of CYP3A4, and correlates with up-regulated expression of the gene [103].

Taken together, epigenetic modﬁcations to the chromatin of drug metabolizing enzymes work synchronously to affect the expression of CYPs within the liver. Understanding the pathways that contribute to altering hepatic DNA methylation, histone acetylation, and histone methylation will enable a better understanding of CYP regulation and how factors that alter epigenetics may affect drug responses. Chapter 2 35

2.3.4 Hormones

Endogenous factors, such as steroid hormones, can also inﬂuence the expression and function of CYPs by altering transcriptional regulation via endocrine signaling and the activation of nuclear receptors. Certain pharmaceutical drugs may act as direct hormone agonists or synthetic analogs, whereas other drugs can modulate the bodys endocrine function and change the levels of circulating hormones. In either case, it is important to understand the effects steroid hormones have on drug-metabolizing CYP activity and the pathways through which they alter

CYP expression. Alterations to CYP activity can then not only change the rate of exogenous drug metabolism, but also affect the rate of subsequent steroid hormone metabolism and energy homeostasis.

Differences in the expression of several drug-metabolizing enzymes exist between males and females, and this can be attributed to the role of sex hormones in determining constitutive levels of CYP gene expression. The female hormone 17β-estradiol is known to enhance CYP2B6 and

CYP3A4 expression and this explains why the expression of these enzymes is typically higher in females than in males, and is increased even more signiﬁcantly during pregnancy [104].

Progesterone, another predominantly female hormone, has also been shown to up-regulate the expression of CYP2B6 and CYP3A4 in the liver [105]. Additionally, since many CYPs also play roles in hormone synthesis and metabolism, this provides a feedback mechanism which regulates hormone homeostasis within the body [106].

Growth hormone, also known as somatostatin, exerts its effect by binding to the membrane growth hormone receptor (GHR), and is also partially responsible for the differences in CYP Chapter 2 36 expression between males and females. In the adult male, the secretion of growth hormone from the pituitary gland occurs in an episodic manner, which is characterized by a few large bursts of growth hormone secretion throughout the day that are separated by longer intervals devoid of the circulating hormone. Females, on the other hand, have circulating growth hormone proﬁles that are more continuous, characterized by numerous, smaller bursts of growth hormone secretion throughout the day. In both sexes, a similar amount of growth hormone is released per day in total, but the difference lies in the pattern of secretion. These patterns play a role in determining basal and induced levels of drug-metabolizing CYPs in the liver. Growth hormone appears to have a repressive effect on CYP expression, and has been shown to suppress CYP2B induction in response to phenobarbital treatment [107].

Glucocorticoids, which include the stress hormone cortisol, also have an effect on the basal and induced expression of CYPs in the liver that can provide an explanation as to why CYPs exhibit circadian patterns of expression. The level of circulating cortisol follows circadian rhythm and is normally highest in the early morning and lowest around midnight. The cellular effects of cortisol and other glucocorticoids are exerted through binding to the glucocorticoid nuclear receptor (GR), which is then able to bind to the promoter of target CYP genes to induce their transcription. The GR appears to be necessary for the maximal induction response of CYP2B,

CYP2C, and CY3A enzymes in response in inducer treatment, as well as for determining constitutive expression levels. It has also been reported that activation of GR also enhances the action of PXR and CAR in inducing the expression of CYP2B, CYP2C, and CYP3A enzymes.

Because of this, GR has been described as a master regulator of CYP expression due to its ability to directly affect the transcription of drug-metabolizing CYPs after its activation, as well as Chapter 2 37 regulate the functions of other nuclear receptors involved in CYP induction [108].

2.4 Regulation of Cytochrome P450s During Postnatal Mat-

uration

As previously mentioned, CYP enzyme induction is considered a temporary and reversible event, and metabolic activity returns to constitutive levels once the inducing drug is ceased and fully cleared from the body [77]. While this is a documented fact in adult patients, it is not thoroughly understood whether CYP induction during neonatal or infant ages causes permanent alterations to enzyme expression and activity. Numerous studies have suggested that inducing CYP expression at young ages may have lasting effects on the expression and function of these enzymes in adulthood. The liver is undergoing rapid cellular proliferation and differentiation in the years following birth, which is accompanied by signiﬁcant changes to the epigenetic landscape. Manipulating epigenetic changes during this time by administering inducer drugs may have permanent consequences on liver function. Therefore, it is pertinent to gain a better understanding of how CYP induction in neonates, infants, and children can affect drug metabolism in the long-term.

Histone modifications have also been implicated in regulating the basal expression of CYPs in the liver during the period of postnatal maturation. As previously discussed, specific CYP enzymes follow dinstict ontogenic patterns during the first several years of life. Adult CYP isoforms increase in expression over time, while fetal and neonatal isoforms simultaneously decrease in abundance. Associations between the enrichment of certain histone marks in the Chapter 2 38 promoters of CYP genes and CYP mRNA expression levels have been observed during postnatal liver maturation in mouse. During the first five days following birth, the fetal isoform of Cyp3a has high mRNA expression and is associated with high levels of the active histone mark H3K4me2 in its promoter. By the adult age, mRNA of the fetal enzyme is practically undetectable, and enrichment of the repressive histone mark H3K27me3 becomes enriched in the promoter region instead. These cellular events occur correlate with low mRNA expression of the adult Cyp3a isoform during the neonatal age and no enrichment of the activating histone mark in this gene. By the adult age, however, the adult Cyp3a increases in expression by 6- fold, which correlates with a 5-fold increase in H3K4me3 enrichment in the promoter region of the gene [67]. Since this study was performed in the absence of any drug treatment, this data shows that epigenetic modifications in CYP genes are highly plastic during postnatal maturation, and are changing in response to endogenous signaling. Similar findings have also been found for the role of histone modifications in the regulation of Cyp2d enzyme otogeny [109].

The consequences of disrupting or interfering with the endogenous regulation of CYP ontogeny by histone modiﬁcations have not been thoroughly studied and warrant future investigation.

2.4.1 Neonatal Phenobarbital Exposure and Permanent Changes to Cy-

tochrome P450 Expression

Beginning as early as 1981, many studies have documented the potential for early postnatal exposure to specific drugs, dietary components, and environmental toxicants to alter the metabolic and detoxification processes in the adult liver. Initial evidence of this phenomenon was first observed in the enzymatic activities of hepatic microsomes prepared from the livers of adult rats Chapter 2 39 that were exposed to various compounds during the first five days following birth. The administration of phenobarbital, methadone, or testosterone to the neonatal pups altered their hepatic monooxygenase capabilities and increased CYP activity in adulthood. These landmark studies suggested that CYP-dependent metabolism may be subjected to permanent manipulation by exogenous compounds early in life [110–114].

With the discovery of individual CYP isoforms, subsequent research throughout the 1990s validated the previous observations by focusing on alterations in the expression and activity of speciﬁc enzymes following neonatal exposure to various compounds. Phenobarbital has been the most commonly utilized inducing agent, and the permanent effects on CYP activity following early life exposure to this drug have been reported in multiple studies. The expression and activity of hepatic CYP2B and CYP2C enzymes were found to be higher in adult rats that were exposed to phenobarbital as neonates, which then had a functional effect on steroid hormone metabolism [115, 116]. This phenomenon was thought to be a consequence of altered growth hormone secretion during development caused by phenobarbital, which ultimately affected the expression of CYPs in the adult liver [117]. Adult animals that were exposed to phenobarbital early in life also had greater extents of CYP induction when re-challenged with phenobarbital in adulthood, suggesting neonatal exposure increases the sensitivity of the adult liver to enzyme induction [118, 119]. Aside from changes in CYP expression, rats that received phenobarbital as neonates also demonstrated higher rates of liver tumorigenesis and a reduced lifespan compared to control animals [119].

A more recent study from our own laboratory has also conﬁrmed that neonatal phenobarbital exposure in mice is capable of causing permanent alterations to CYP expression in adulthood. Chapter 2 40

In addition, this work examined the role of the dose of phenobarbital and the age of exposure in producing long-term changes to Cyp2b10, Cyp2c29, and Cyp3a11 expression in the liver. Fol- lowing a single dose of phenobarbital ﬁve days after birth, all three CYP isoforms were overexpressed in both male and female mice at 60 days after birth when compared to control animals.

This effect was found to be dose-dependent in that the greater the dose of phenobarbital given, the larger the extent of CYP overexpression. The effect was also found to be age-dependent, in that the earlier the age of phenobarbital exposure, the larger the extent of CYP overexpression in adulthood. If mice were given phenobarbital at ages older than day 15, signiﬁcant alterations to CYP expression in adulthood were no longer apparent, suggesting a window of sensitivity exists in which the neonatal liver is susceptible to enzymatic programming [120].

2.4.2 Neonatal Exposure to Other Drugs and Permanent Changes to Liver

Function

A few additional xenobiotics have also been tested for their potential to cause permanent effects on adult CYP-mediated metabolism if administered during postnatal development. Lindane, an organochlorine compound used for the treatment of lice and as an insecticide, caused the overexpression of CYP1A and CYP2B enzymes in adult rodents following neonatal treatment

[121]. Neonatal exposure to tamoxifen, an estrogen receptor modulator used for breast cancer therapy, caused increased levels of CYP2A1 in adult male rats only. In females, however, a lowered expression of CYP3A9 was observed [122]. Exposing neonatal rodents to diethylbisterol, a synthetic estrogen once used in various female hormonal therapies, caused decreased CYP3A expression and increased CYP1A and CYP2B expression in adult males only [123]. Taken Chapter 2 41

Drug P450s Permanently Expression ↑ or Level of Species Reference administered Altered ↓ Quantification Rat (male) Phenobarbital 2B ↑ Protein

Rat (male) Diethylstilbestrol 2C6 ↑ Protein (Zangar et al., 1993) Rat (male) Diethylstilbestrol 3A2 ↓ Protein

(Agrawal Rat Phenobarbital 2C7 ↑ RNA, protein, activity and Shapiro, 2000)

(Agrawal Rat Phenobarbital 2C6, 3A1, 3A2 ↑ RNA, protein and Shapiro, 2003)

(Agrawal 2C7, 2C6, 2B1, 2B2, Rat Phenobarbital ↑ RNA, protein, activity and Shapiro, 3A1, 3A2 2005)

(Agrawal Rat Phenobarbital 2B1, 2B2 ↑ RNA, protein, activity and Shapiro, 1996a)

(Tien et al., Mouse Phenobarbital 2B10, 2C29, 3A11 ↑ RNA, protein, activity 2015) (Johri et al., Rat Lindane 2B1, 2B2, 1A1, 1A2 ↑ RNA, protein 2008) Rat (male) Tamoxifen 2A1 ↑ Protein, activity (Kawai et al., Rat (female) Tamoxifen 3A9 ↓ RNA 1999) Rat (male) Tamoxifen 2C11 ↓ Protein, activity (Murakami Rat 17β-estradiol 3A ↑ Activity et al., 2004)

FIGURE 2.3: Summary of studies that demonstrate long-term alterations in hepatic CYP expression, activity, and/or induction following neonatal administration of speciﬁc drugs. together, these studies suggests clear gender-dependent effects following neonatal exposure to various xenobiotics. Additionally, most of the xenobiotics that have been tested are either synthetic steroid hormones, or resemble steroid hormones in their chemical structure. This also implies a signiﬁcant role in circulating hormone levels in regulating CYP ontogeny and expression during postnatal development, which can then have lasting effects on liver metabolism into adulthood. A table summarizing all rodent studies that found altered CYP expression in adulthood following neonatal drug exposure can be found in Figure 2.3. Chapter 2 42

Although the focus of this research is on drug-metabolizing CYP enzymes in the liver, there are several other studies that examined long-term alterations to CYPs involved in other metabolic processes or located in other organs. CYP8B1 is responsible for regulating cholesterol metabolism through its involvement in bile acid synthesis. Following neonatal exposure to the glucocorticoid drug dexamethasone in rats, CYP8B1 levels were lowered and resulted in impaired hepatic lipid secretion once the animals reached adulthood [124]. CYP activity in the brain has also been shown to be modiﬁed by neonatal drug exposure. Cerebral expression of CYP1A and

CYP2B enzymes were up-regulated in adults that received treatment with lindane during the perinatal period [121]. Treatment with the selective serotonin reuptake inhibitor, fluoxetine, for the first 16 days following birth caused rats to have elevated CYP4A expression in their adult brains, which altered cerebral arachidonic acid metabolism [125, 126]. A summary of all the discussed work regarding neonatal drug exposure and permanent alterations to CYP expression can be found in Figure 2.3. Together, these studies indicate that CYP regulation in general is quite sensitive during early postnatal life and may be particularly susceptible to exogenous modification with permanent effects. The mechanisms behind this phenomenon are just beginning to become elucidated, however they undoubtedly involve underpinning changes to the epigenetic landscape. Chapter 2 43

2.4.3 Nuclear Receptors and Neonatal Imprinting

The term “imprinting” can have two different meanings in the field of biology. Genomic imprinting refers to the phenomenon that causes certain genes to be expressed in a parent-of- origin-specific manner, as a result of particular DNA or histone methylation patterns that repress gene transcription on one allele [127]. This project, however, will refer to the concept of neonatal imprinting, which describes a phenomenon of developmental programming that produces delayed effects as a result of neonatal exposure to a specific exogenous compound or environmental stimulus. This type of imprinting requires a specific time window of sensitivity and organ plasticity during early postnatal, or even prenatal, life in which exposure must occur

[128]. During this window of sensitivity, organogenesis occurs and tissues undergo extensive differentiation as epigenetic markers are installed on the genome to facilitate specialized gene expression. Exposure to compounds that alter how the epigenetic modiﬁcations are placed can cause lifelong disruptions to gene expression, causing programmed alterations to organ function [129].

The mechanisms that underlie neonatal imprinting involve influencing the proteins involved in creating histone modifications and DNA methylation during postnatal windows of sensitivity in which organs still exhibit plasticity. As discussed in previous sections, DNA methylation, histone acetylation, and histone methylation have strong influence in determining whether a gene in transcriptionally repressed or activated. In adulthood, when organs are fully developed and differentiated, epigenetic modifications that result from an environmental exposure are usually transient and short-lived. However, during periods of tissue growth and maturation when organs still exhibit plasticity, epigenetic modifications have the possibility of becoming permanent, to Chapter 2 44 offer a type of cellular memory for later gene expression and exposure response later in life

[130]. Upstream of the specific epigenetic modifications, however, are the transcription factors that bind to target genes and recruit the co-activator proteins to facilitate the addition or removal of the epigenetic modifications. It is still not well understood how certain compounds that affect transcription factor binding influence neonatal imprinting during postnatal development.

The concept of neonatal imprinting, although still elusive, has been described in various organs in response to a variety of environmental exposures. Several studies have demonstrated permanent changes in chromatin structure and histone modifications in different tissues following postnatal exposure to specific xenobiotics that cause the activation of nuclear receptors, which also act as transcription factors. Endocrine disruptors, in particular, have been shown to play a significant role in neonatal imprinting in multiple organs due to their ability to activate the nuclear estrogen receptor (ER) [131]. Rats exposed to bisphenol A (BPA), an endocrine disruptor found in plastic, during the first five days of life exhibited markedly up-regulated expression of secretaglobin (Scgb2a1), a secreted protein involved in inflammation, tissue repair, and tumorigenesis, in their prostate at age day 70. The increase in gene expression correlated with persistently enriched H3K9 acetylation and H3K4 methylation and decreased methylation in the CpG island upstream of the genes transcription start site, which facilitated increased transcription. The overexpression and enhanced secretion of secretaglobin in adulthood could subsequently have implications in prostate cancer risk and also affect the efficacy of drugs that bind to the protein [132]. The expression of other genes in the prostate cancer pathway were not found to be altered in adults following neonatal BPA exposure until the adult animals Chapter 2 45 were re-challenged with a stimulus. In this case, when testosterone was administered in adulthood, other genes within the prostate cancer pathway exhibited an exaggerated response to the hormone due to neonatal BPA exposure, even though there was no difference in their basal expression in the absence of testosterone administration [133].

Neonatal exposure to genistein, an endocrine disruptor found in dietary soy, and diethylbisterol, both of which bind the ER, also caused permanent changes to gene expression and inductive responses in adulthood by altering the epigenetic landscape. The promoters of genes in the uterus remained permanently hypomethylated following treatment, and lead to increased gene transcription in adults. Responses to re-challenges with estrogen in adulthood were exaggerated, and included up-regulated secretion of lactoferrin, a protein found in breastmilk [134, 135]. To- gether, these studies on neonatal programming following endocrine disruptor exposure had several mechanistic factors in common: 1) the activator of a nuclear receptor, 2) altered response to hormone stimulation, and 3) up-regulated protein secretion. Whether this phenomenon can occur with the activation of other nuclear receptors that play a role in hormone signaling is worth investigating.

Endocrine disruptors and the ER have also been shown to play a role in the neonatal programming of gene expression in the liver through permanent epigenetic alterations. Diethyl- stilbestrol administration at the neonatal age activated the ER and permanently altered hepatic gene expression through the activity of small heterodimer partner (SHP), which plays a role in transcriptional repression. Adult mice exhibited decreased Cyp7a1 and Cyp8a1 expression and increased Cyp7b1 expression, which altered bile acid synthesis and homeostasis permanently. Additionally, neonatal exposure to diethylstilbestrol caused decreased levels of CAR Chapter 2 46 and PXR in adult livers, which in turn reduced the expression of Cyp3a4, Cyp2b10, and other hepatic transporters [136]. In terms of neonatal imprinting in the liver and permanent effects on

CYP-mediated drug metabolism, no studies have established the role of pharmaceutical drugs that are able to activate PXR or CAR, however. Limited research utilizing non-pharmaceutical compounds that directly bind nuclear receptors to cause their activation has shown that CAR and PXR activation during postnatal liver development can permanently alter hepatic gene expression in adulthood via persisting epigenetic modiﬁcations. A landmark study using a mouse model was able to establish persistent induction of Cyp2b10 and Cyp2c37 in the livers of adult mice that received treatment with TCPOBOP, a direct murine CAR ligand, at the neonatal age.

As a result, the adult mice showed a decrease in efﬁcacy of zolaxamine, a compound that induces sleep, due to enhanced drug metabolism. Permanent increased enrichment of H3K4 methylation and decreased H3K9 methylation in the promoter region of the Cyp2b10 gene accompanied the long-term changes in CYP mRNA expression. The permanent overexpression of CYPs and associated histone modiﬁcations were abolished, however, in knockout mice that lacked CAR, suggesting this nuclear receptor was necessary to produce the lasting effects of neonatal imprinting by TCPOBOP [86].

Since the study with TCPOBOP and permanent alterations to CYP2B10 and CYP2C37 expression in 2012, only one other project has looked at neonatal nuclear receptor activation and the effects on CYP expression in the liver in adulthood. In this case, mice were treated with either

PCN, a direct murine PXR ligand, or TCPOBOP, in the neonatal stage, and the mRNA expressions of all genes in the liver were measured and compared in adulthood. Neonatal exposure to TCPOBOP resulted in persisting up-regulation of Cyp2b10 and Cyp2c29 along with other Chapter 2 47 xenobiotic-processing enzymes and transporters. On the other hand, Cyp4a enzymes were sig- niﬁcantly down-regulated. Neonatal exposure to PCN resulted in persisting up-regulation of

Cyp2b10, Cyp2b13, Cyp2a, and Cyp2c55 and various other enzymes involved in metabolism, and a decrease in Cyp4a expression as well [71]. This suggests that there can be overlap in nuclear receptor signaling and the subsequent permanent changes to gene expression that result. However, this also indicates that there may also be speciﬁcity in which genes become imprinted as a result of the speciﬁc nuclear receptor that becomes activated at the neonatal age. Additionally, this study found that both PCN and TCPOBOP treatment during early postnatal development caused permanent down-regulation of the nuclear peroxisome proliferator- activated receptor (PPAR) in adult livers. PPAR regulates hepatic lipid metabolism and induces expression of CYP4A enzymes when activated [71]. This also suggests that nuclear receptor activation in the neonatal age has the potential to impact the expression of target genes, as well as other nuclear receptors and transcription factors, permanently, which can affect organ function in its entirety.

Phenobarbital appears to be the prototypical, clinically relevant drug administered to neonatal animals to elicit the persistent up-regulation of drug-metabolizing P450s. Other studies have utilized compounds that directly bind nuclear receptors to also elicit permanent changes to the same enzymes. However, the link between postnatal nuclear receptor activation by pharmaceutical drugs and altered CYP-mediated drug metabolism in adulthood has not yet been well- established. Pediatric patients are regularly exposed to many other drugs besides phenobarbital that produce CYP enzyme induction for indications such HIV, tuberculosis, and seizures, as summarized in Figure 2.4. Chapter 2 48

FIGURE 2.4: Drugs administered to pediatric patients with the ability to activate nuclear receptors to induce CYP activity. GCR = glucocorticoid receptor

It is also possible for mothers that are taking medications that cause enzyme induction to pass through breastmilk to her infant. Additionally, all of the discussed studies were concepts only shown in rodent models, which will eventually need to be put to the test in humans. One study has shown that formula-fed infants have higher levels of CYP1A2 activity and clear caffeine from their systems more rapidly than breastfed infants at six months of age, and the authors concluded this was due to components of the formula activating the AhR, which regulates

CYP1A2 induction [137]. Therefore, it may be possible for nuclear receptor activation caused by pharmaceutical drugs administered to pediatric patients to program, and permanently alter,

CYP expression and drug metabolism in humans. However, future studies utilizing additional pharmaceutical inducers are absolutely necessary to generalize this phenomenon. Chapter 2 49

2.5 The Mouse as an In Vivo Model for Studying Drug Metabolism

So little is known about the regulation of drug metabolism in pediatric patients simply because they are not usually included in clinical trials that determine pharmacokinetic profiles. This is due to obvious ethical, financial, and technical constraints, and it is not likely that pediatric patients will be used regularly in clinical trials in the future. Epidemiological studies in infants and children that compare different lifestyle, environmental, and genetic factors to alterations in CYP activity exist, however it is impossible to eliminate all confounding variables. Liver biopsies from healthy children are also extremely difficult to obtain, and an in vitro cell model representative of a true infant or child hepatocyte does not exist. Therefore, studies in animals remain as our best option for understanding the regulation of CYP expression in response to drug induction during early postnatal development.

This study, in particular, utilizes the C57Bl/6 mouse as a model organism. The mouse is an ideal model for postnatal developmental studies for several reasons. First, the short lifespan enables us to examine the effects of early life drug exposure on metabolism in adulthood over a relatively short period of time. Mice also give birth to large litters than usually contain between six and twelve pups, in our experience. This allows us to include both vehicle control- and test compound-treated pups within the same litter, which decreases inter-litter variability and helps to diminish confounding factors. Finally, mouse genetics are well-defined and widely manipu- lated as evidenced by the seemingly infinite amount of genetically modified mouse models that have been engineered. Mice that lack functional nuclear receptors, such as PXR and CAR, are readily available to further study the effects of CYP induction at postnatal ages in the absence of these key signaling molecules. Chimeric mouse models that express the humanized forms Chapter 2 50 of PXR and CAR also exist, and can be utilized to study compounds that are specific to the activation of human nuclear receptors.

While phenobarbital is an indirect activator of both human and murine CAR, the nuclear receptor is activated by different compounds and drugs in each respective species, with only a few that are common. Unfortunately, this is a regular problem when attempting to use the mouse as a model to study human drug metabolism. Many compounds that activate the murine form of a nuclear receptor do not activate the human form of a nuclear receptor, and vice versa. There are also differences in the enzymatic substrates between human and murine CYP orthologs. How- ever, no model organism or system is a perfect representation of the human body. This section will discuss the similarities in CYP-mediated drug metabolism between mice and humans, as well as the challenges that will need to be considered when extrapolating the data from mouse drug metabolic studies to clinically relevant conclusions in humans.

2.5.1 Human and Mouse Orthologous Cytochrome P450 Genes

CYP enzymes have been genetically conserved across all species and appear to have originated from a single ancestral gene approximately 1.36 billion years ago. All gene members of the

CYP superfamily contain highly conserved regions of amino acids, with smaller variations in primary amino acid sequences between species. This gives rise to orthologous CYPs, in which isoforms from different species may be designated as belonging to the same subfamily due to conserved regions, while the slight differences in the rest of the sequence are acknowledged by assigning a unique isoform number to the gene name [138]. Chapter 2 51

Mice, in general, possess higher relative amounts of hepatic enzymes than humans when measured as the concentration of CYP enzymes per gram of bodyweight. Thus, drug metabolism and clearance is considered to be more rapid in the mouse than in the human. Mice also have a greater number of total CYP genes than humans (93 versus 57, respectively). Forty pairs of orthologous CYP genes between the two species have been identiﬁed, and they encode for enzymes that perform similar functions and share common substrates [139, 140]. Differences between the activities of orthologous CYPs can exist however, and in some cases drug substrates are unique to each species. This becomes a challenge when utilizing a mouse model to characterize a compounds predicted pharmacokinetic proﬁle and identify the enzymes responsible for its metabolism and clearance in a human. This project, however, is not attempting to characterize the metabolism or clearance of any compounds. We are attempting to establish a proof of concept, which argues that CYP regulation and induction during postnatal develop differ from that in later ages of adulthood. Therefore, we believe the use of a mouse model is appropriate at this stage of research. The concept, of course, will need to be validated eventually in humans. A list of human drug-metabolizing CYPs and their corresponding orthologs in mouse that we will use for our studies have been listed (Figure 2.5), along with probe substrate drugs that are clinically used to measure enzymatic activity.

In order to extrapolate the data from this project to relevant human scenarios, we ﬁrst must understand the postnatal ontogeny of CYP expression in mice as it relates to humans. As described previously, individual CYP isoforms follow distinct patterns of expression throughout the course of postnatal liver maturation in humans. Our laboratory has previously performed

RNA-sequencing (RNA-seq) on mouse liver samples collected at various ages of postnatal Chapter 2 52

FIGURE 2.5: Human and mouse orthologous CYP enzymes and their substrates. Human drug-metabolizing CYPs have been listed with their corresponding accepted murine orthologous enzyme. Probe substrates for each enzyme typically utilized to evaluate enzymatic activity rates (based on FDA guidance) have also been listed. development (Figure 2.6). This allowed us to quantitatively measure the absolute amounts of each CYP isoform in the liver across the neonatal, infant, child, adolescent, and young adult stages [141]. Like human CYP expression patterns, the ontogenies of murine CYP expression also appear to belong to one of three categories: fetal/neonatal enriched, adolescent enriched, and adult enriched [60].

In general, most drug-metabolizing CYPs belonging to the subfamilies Cyp1, Cyp2, and Cyp3 are expressed at low levels before and immediately following birth, similar to humans. Through- out postnatal maturation, the expression of these CYPs increases, with expression levels comparable to those of adult mice attained by day 30 of age. Further examination of the expression of individual murine CYPs throughout postnatal development also reveals similarity to postnatal

CYP ontogeny in humans. Murine Cyp2c29, 2e1, and 3a11 enzymes are the most abundant of the drug-metabolizing CYPs expressed in the adult mouse liver, which correlates with relative Chapter 2 53

C y p 1 a 2 C y p 2 b 1 0

1 0 0 0 6 0 n

n 8 0 0

s s

s 4 0

)

) e

e 6 0 0

F 4 0 0

(

( A

A 2 0

R R

2 0 0

m m

0 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0

C y p 2 c 2 9 C y p 2 d 2 2

2 0 0 0 1 5 0

o i

i 1 5 0 0

s s

s 1 0 0

)

K x

x 1 0 0 0

(

( A

A 5 0

N R

R 5 0 0

m m

0 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0

C y p 2 e 1 C y p 3 a 1 1

8 0 0 0 4 0 0 0

o i

i 6 0 0 0 3 0 0 0

)

K x

x 4 0 0 0 2 0 0 0

(

N R

R 2 0 0 0 1 0 0 0

m m

0 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0

FIGURE 2.6: Ontogeny of murine drug-metabolizing CYPs following during postnatal liver maturation. RNA-seq was performed on livers from male C57Bl/6 mice collected on different days following birth (x-axis) to evaluate gene expression throughout postnatal maturation. Data are displayed as the mean of three liver samples ± S.E.M for each time point Chapter 2 54 abundance of CYPs in humans. The enzymes Cyp2d22 and Cyp3a11 appear to reach their maximal expression at the age of day 30, however the other enzymes shown in the ﬁgure (Cyp1a2,

2b10, 2c29, and 2e1) actually appear to be expressed at higher levels in late adolescent stages between the ages of day 20 to day 45, than in fully mature adults. This would suggest that drug metabolism catalyzed by these enzymes might occur at faster rates in adolescent stages than in adults, at least in mice. However, this also correlates well with the fact that the half-lives of certain drugs are actually lower in human children than in adults, suggesting faster drug clearance during childhood [3]. Both phenobarbital and phenytoin, which are primarily metabolized by

CYP2C enzymes in humans, have shorter half-lives in infants and children compared to adults, suggesting higher activity of CYP2C enzymes in pediatric populations.

During postnatal liver maturation in the mouse, there is also a switch in the expression of the Cyp3a family of enzymes, similar to what occurs in humans. In humans, CYP3A7 is considered the fetal/neonatal enzyme of the subfamily, and is most highly expressed in the fetal liver during gestation and for several months after birth. As the infant develops, CYP3A7 expression decreases while CYP3A4, the predominant adult enzyme, increases in expression.

A similar pattern is seen in the developing mouse liver. Cyp3a16 is considered the murine fetal/neonatal Cyp3a enzyme, and it is expressed most highly during gestation and for the ﬁrst

20 days followning birth (Figure 2.7). At later ages, its expression of Cyp3a16 is repressed as it becomes practically undetectable in the adult liver, while the Cyp3a11 enzyme increases in expression and becomes the predominantly expressed Cyp3a enzyme once the liver is fully matured. Because of this, high expression of Cyp3a16 in the mouse liver is a sign that the organ is still undergoing the maturation process [142]. Chapter 2 55

C y p 3 a 1 6

1 5 0

i s

s 1 0 0

)

F (

A 5 0

R m

0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0

FIGURE 2.7: Repression of murine fetal CYP isoform Cyp3a16 during postnatal liver maturation. Ontogeny of Cyp3a16 expression throughout postnatal development is shown at various days following birth (x-axis). Data are displayed as the mean of three liver samples ± S.E.M.

The nuclear receptors CAR and PXR that contribute to the regulating CYP expression in response to induction may also be expressed at slightly higher levels during the human postnatal development phase compared to adult [? ]. However, the variability among liver samples from pediatric patients is quite large and it is difficult to draw concrete conclusions. In mice, however, we can observe slightly higher expression of Car and Pxr during neonatal ages before day 20 (Figure 2.8). Whether this is biologically significant and/or contributes to the ontogenic regulation of CYPs is relatively unknown. It is difficult to extrapolate murine data on Car and

Pxr function to relevant conclusions for human CAR and PXR, though, due to very distinct differences in receptor ligands between the species. This results from differences in the amino Chapter 2 56

C a r P x r

8 0 3 0

o i

6 0 i

s s

s 2 0

)

K x

4 0 x

(

( A

A 1 0

N N

R 2 0

m m

0 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0 0 1 3 5 1 0 1 5 2 0 2 5 3 0 4 5 6 0

FIGURE 2.8: Ontogeny of murine nuclear receptors Car and Pxr during postnatal liver maturation. Expression of Pxr and Car in the livers of mice collected at different days following birth (x-axis) determined by RNA-seq. Data are presented as the mean of three livers ± S.E.M. acid sequences in the ligand binding domains in each species. For example, human PXR is activated by the administration of the drug rifampicin to induce CYP3A4 expression, while murine

PXR is not activated by rifampicin. Instead, prenenolone 16 α-carbontrile (PCN) is considered a direct murine-speciﬁc Pxr ligand used to induce Cyp3a11 expression. In terms of CAR, the chemical CITCO is used as a direct human CAR ligand to induce CYP2B6 expression. While this compound does not activate Car in mouse, the chemical TCPOBOP is often utilized as a direct murine Car ligand [143] to induce Cyp2b10 expression.

2.5.2 Inter-Species Age and Dose Extrapolation

In terms of studying developmental pharmacology utilizing the mouse as a model, it is critical to understand how different ages in mouse correlate with ages in humans, and how doses of a Chapter 2 57 compound compare between the species. Mouse and human ages cannot be directly extrapolated, as mice develop at an accelerated rate compared to humans.

On average, the lifespan of a mouse is approximately two years, and we consider the average lifespan of a human as 80 years. When taking the entire lifespan of each species into account, nine mouse days is roughly equivalent to one human year. This conversion is not very useful during postnatal development, however a variety of factors can be evaluated to estimate how a certain age in mice might correlate to humans. First, young mice are typically weaned from their mother between 21-28 days of age, while human babies are usually weaned between the ages of 6-12 months. This indicates that human infants are weaned much earlier in life in comparison to mice and this must be considered when studying postnatal development. During the period of postnatal development prior to weaning, one human year is equivalent to roughly

57 mouse days. Conversely, mice reach puberty much sooner in life following weaning than human children. By age day 42, mice are considered to have attained puberty, however puberty does not occur in humans until 12-16 years of age. Based on this, one human year is equivalent to only three mouse days approximately during the pubertal phase. Finally, mice at the age of day 60 are considered fully matured adults, while humans are typically considered fully matured (growth-wise) around age 20 [144]. Based on these developmental landmarks, we have utilized the following mouse ages as representation of specific phases of human postnatal development: day 5-10: neonatal; day 15: infant; day 20: child; day 30: adolescent; day 45: young adult; day 60: adult. In terms of CYP ontogeny, the most drastic changes occur within the first year of life in humans, which would correlate with the first 10 days of life in mice.

Drug dosage also must be adjusted in mice to make valid conclusions based on clinical doses Chapter 2 58 utilized in humans. Mice are considered to metabolize xenobiotics more quickly than humans, so doses based on bodyweight are typically higher in mice to achieve the same efﬁcacy or toxicity. The FDA offers a simple equation to convert human doses to equivalent doses in a variety of species. To estimate the dosage in mouse, the human dose is simply multiplied by

12.3 [145]. So for example, a dose of 20 mg/kg bodyweight of a drug in a human would be equivalent to a dose of roughly 246 mg/kg bodyweight of the same drug in a mouse. This equation is only validated for adult human dose conversions, however [146]. In the following studies, we converted suggested pediatric doses in humans to developmental mouse equivalent doses using the same formula since a pediatric-speciﬁc inter-species dose equivalence formula does not exist. Chapter 3

Neonatal Phenobarbital Exposure and

Drug Interactions in Adulthood

3.1 Introduction

Polypharmacy-induced DDIs signiﬁcantly raise the risks for decreasing therapuetic efﬁcacy and increasing adverse reactions, with particularly higher risks for the elderly, children, and female populations [147]. Approximately 50% of the population aged over 65 years now takes at least

ﬁve different medications, with 35-60% of these elderly patients exposed to a potential DDI, and 5-15% suffering clinically signiﬁcant adverse reactions [148]. Additionally, it is estimated that close to 50% of hospitalized patients under the age of 21 are also exposed to potential DDIs

[16].

59 Chapter 3 60

On the basis of their mechanisms of action, DDIs are classiﬁed into two main categories (pharmacokinetic or pharmacodynamic) based on their origin. Pharmacokinetic DDIs occur when one drug alters the absorption, distribution, metabolism, or excretion of another drug administered concomitantly. Altering any of these processes can increase or decrease the concentrations of a drug or its metabolites in the circulation. Certain drugs possess the ability to up- or down-regulate the rate at which another drug is metabolized, leading to signiﬁcantly lower or higher plasma concentrations of the drug or its metabolites. These types of interactions are common, owing to the induction of many CYPs responsible for the biotransformation of 80% of prescription drugs.

Induction of CYPs in response to drug treatment occurs shortly after the inducing drug is taken, with a delay depending on the half-life of the drug. Induction is considered to be temporary in adults, in whom CYP expression levels will return to basal levels once the inducing drug is discontinued and excreted [149]. However, a large body of work from Dr. Bernard

Shapiro’s laboratory has demonstrated that neonatal exposure to the AED phenobarbital may cause lifelong alterations to the basal and induced levels of CYP enzyme expressions and activities [116, 117, 119]. Recently, research in our laboratory has demonstrated that neonatal exposure to phenobarbital can cause a permanent elevation of the activities of several CYPs in the adult mouse liver and that this effect is dependent upon two key factors: the age at which the mouse is exposed to the drug in early life, and the dose at which phenobarbital is administered

[120]. Phenobarbital is still considered the ﬁrst drug of choice for treating acute seizures in newborns, which occurs in two-three out of every 1,000 live births in the United States [150], and is widely administered to babies and children across the globe. It is also a known inducer Chapter 3 61 of CYP2B6, 2C9, 2C19, and 3A4 in humans, and Cyp2b10, 2c29, and 3a11 in mice [151].

Treatment at earlier ages with high doses of phenobarbital produced a permanent induction of

CYP enzymes in adults in mice, however whether the efﬁcacy of other drugs administered in adulthood is impacted has not yet been studied.

This study aims to determine whether treatment with the CYP-inducing drug, phenobarbital, early in life can affect the efﬁcacy of a drug taken later in life. Omeprazole (brand name

Prilosec OTC) is a proton-pump inhibitor (PPI) commonly utilized to treat gastric ulcers, gas- troesophageal reﬂux disease (GERD), and heartburn by increasing the pH of the stomach. PPIs are some of the most commonly prescribed drugs and are usually available over the counter

[152]. Omeprazole was chosen as a probe drug to test for changes in efﬁcacy in adult mice following neonatal exposure to phenobarbital for two reasons. First, its efﬁcacy can be easily assessed by measuring the pH of gastric contents following a daily dosing regimen. Secondly, omeprazole is known to be primarily metabolized by CYP2C19 and 3A4 in humans [153, 154].

The metabolism of omeprazole has also been shown to be altered when co-administered with phenobarbital [155]. Although the CYP enzymes that mediate the primary metabolism of omeprazole in mice have not yet been deﬁned, we are hypothesizing that omeprazole is metabolized by Cyp2c and Cyp3a enzymes in the mouse as well, particularly Cyp2c29 and Cyp3a11, both of which are induced by phenobarbital. Investigating whether neonatal phenobarbital exposure affects the ability of omeprazole to increase stomach pH in adults can give insight into how drug treatment in early life can impact drug interactions later in life, even if two drugs are not administered at the same time. This knowledge may prompt a re-evaluation of how DDIs in the clinic are viewed and predicted. In order for a DDI to occur, it may not be necessary for Chapter 3 62 two drugs to be taken concomitantly.

3.2 Materials and Methods

Chemicals. Phenobarbital sodium salt, omeprazole, and Tween-20 were purchased from Sigma-

Aldrich (St. Louis, MO).

Animals. The use of animals in this study was approved by the Institutional Animal Care and

Use Committee at the University of Connecticut. C57Bl/6 mice (Jackson Laboratories, Bar

Harbor, ME) were bred and housed under standard conditions in the Animal Resources Facility at the University of Connecticut according to the animal care guidelines provided by the Amer- ican Association for Laboratory Animal Science. The speciﬁc timing of drug treatment for each experiment is outlined in the ﬁgures. Phenobarbital was dissolved in phosphate-buffered saline (PBS) and administered intraperitoneally to animals at a dose of 200 mg/kg. This dose of phenobarbital was chosen based on our previous publication showing that a single dose of

200 mg/kg phenobarbital could permanently induce CYP enzymes in mouse [120]. Three consecutive doses of omeprazole dissolved in PBS supplemented with 1% Tween-20 were also administered to mice intraperitoneally at a dose of 150 mg/kg. This dose was selected based on previously described efﬁcacy to inhibit the gastric H+/K+-ATPase pump in mice [156]. Only male mice were used in this study to avoid variability in CYP expression based on estrous cycles in adult females.

Gastric pH Measurement in Stomach. One hour after the ﬁnal dose of omeprazole (or vehicle control) was administered, adult mice were anesthetized with isoﬂurane and an incision was Chapter 3 63 made into the stomach. An Orion 9863BN micro-pH electrode was placed in the middle of the stomach and the pH of gastric liquid was measured using a previously described protocol [157].

Mice were then sacriﬁced and livers were harvested and stored at -80 °C.

Real-Time Quantitative PCR. Total RNAs were isolated from liver tissue without gall blad- ders using TRIzol reagent according to the manufacturers protocol (Life Technologies, Guil- ford, CT). RNA concentrations were measured using a NanoDrop spectrophotometer (Nan- oDrop Technologies, Wilmington, DE) at a wavelength of 260 nm. The integrity of RNA was evaluated using an Agilent 2200 Tape Station (Agilent Technologies, Santa Clara, CA). To obtain cDNA, reverse transcription was performed using iScript Reverse Transcription Supermix according to the manufacturers protocol (Bio-Rad, Hercules, CA). RT-PCR was performed using a CFX-96 thermocycler system (Bio-Rad) with TaqMan gene expression assays for Gapdh,

Cyp2c29, and Cyp3a11 (Thermo Fisher Scientiﬁc, Waltham, MA). Relative gene expression was quantiﬁed using the Ct method with normalization to Gapdh.

Western Blotting. Total cellular proteins were isolated from liver tissue using RIPA buffer according to the manufacturers protocol (Life Technologies). Protein concentration was determined by the Qubit Protein Assay Kit (Life Technologies). Extracted proteins were diluted to a concentration of 40 g/mL and were separated on a 12% SDS-PAGE gel using an electrophore- sis system (Bio-Rad), then electro-blotted onto a PVDF membrane with a transfer system (Bio-

Rad). The membranes were blocked with 5% non-fat milk at room temperature for one hour followed by incubation with the primary antibody against Cyp3a11 (1:1000 dilution) (Millipore

Corporation, Bedford, MA). Gapdh antibody (Sigma-Aldrich) was used as a loading control for normalization (1:3000 dilution). Immunoreactive bands were detected by chemiluminescence Chapter 3 64

using corresponding horseradish peroxidase-conjugated secondary antibodies (1:3000 dilution

(Sigma-Aldrich)).

Statistical Analysis. Data are presented as mean ± S.D. with n = 3 for each treatment group.

Statistical analyses were performed using GraphPad Prism 6 software. Comparisons between control and treatment groups were evaluated using Students t-test and a p-value of < 0.05 was considered statistically signiﬁcant, provided a power of 80% if the mean between control and treatment groups were greater than 2-fold and standard deviation was less than 30%.

3.3 Results

3.3.1 Omeprazole inhibits the gastric proton-pump to increase pH in the

stomach of adult mice.

A control experiment (Figure 3.1A) was performed to examine the efﬁcacy of omeprazole in inhibiting proton-pump activity in the adult mouse stomach. Adult mice that received PBS as a vehicle control had a gastric pH of 2.4 ± 0.2 in an unfasted condition. Adult mice that received omeprazole for three days had a gastric pH of 4.5 ± 0.2 also in an unfasted state

following the ﬁnal treatment (Figure 3.1B). This indicates that omeprazole efﬁciently blocked

the activity of gastric proton-pumps to reduce the amount of acid released into the stomach

and signiﬁcant raised the pH (**p < 0.01). Expression of Cyp2c29 and Cyp3a11 mRNAs and

Cyp3a11 protein in livers were also were also measured by RT-qPCR and Western blotting, respectively (Figure 3.1C, D, E). Because an antibody speciﬁc to Cyp2c29 was not available, Chapter 3 65 the expression of Cyp2c29 protein could not be determined. No differences in gene expression of either Cyp3a11 or Cyp2c29 were observed following omeprazole treatment. These results show that treatment with omeprazole does not induce or repress Cyp3a11 and Cyp2c29 gene transcription.

3.3.2 Concurrent administration of phenobarbital and omeprazole tem-

porarily reduces efﬁcacy of omeprazole in adult mice.

A potential interaction between phenobarbital and omeprazole in adult mice was investigated utilizing the experimental design illustrated in Figure 3.2A. Mice in the omeprazole control group had a gastric pH of 4.0 ± 0.1, which was comparable to our earlier results. Mice in the co-treatment group that received phenobarbital and omeprazole concurrently had a gastric pH of 3.5 ± 0.1, which was signiﬁcantly lower (***p < 0.001) than the control group. This indicates that administering phenobarbital at the same time as omeprazole results in a DDI that attenuates the efﬁcacy of omeprazole in increasing gastric pH in the stomach (Figure 3.2B).

However, administering phenobarbital several days before beginning treatment with omeprazole does not adversely affect the action of omeprazole. Mice in this post-treatment group had a gastric pH of 4.2 ± 0.1 following the final dose of omeprazole, which showed no statistically significant difference to the gastric pH in the control group. This suggests that in adult mice, the efficacy of proton-pump inhibition by omeprazole is not affected by a previous administration of phenobarbital, and phenobarbitals potential to cause drug interactions may only be transient. Chapter 3 66

FIGURE 3.1: Omeprazole increases gastric pH without altering CYP expression in adult mice. (A) A schematic of animal treatment. Adult mice were treated with either vehicle control (PBS, n = 3) or 150 mg/kg omeprazole (OME, n = 3) for three consecutive days at the ages of days 57, 58, and 59 after birth. At age day 60, gastric pH was measured and livers were collected for gene expression analysis. (B) Gastric pH with mean S.D. in the mouse stomach 1 hour after the ﬁnal treatment of PBS or OME. (C) Relative fold changes of mRNA of CYP2C29, (D) mRNA of CYP3A11, and (E) protein of CYP3A11 in the liver of mice treated with either vehicle or omeprazole. **p < 0.01. GAPDH; glyceraldehyde 3-phosphate dehydrogenase. Chapter 3 67

FIGURE 3.2: Concurrent administration of phenobarbital (PB) and omeprazole (OME) results in a drug-drug interaction and reduces omeprazole efficacy in adult mice. (A) A schematic of animal treatment. Adult mice in the OME control group (n = 3) were treated with three consecutive doses of 150 mg/kg/day omeprazole (OME) at ages day 57, 58, and 59 after birth. Adult mice in the PB/OME co-treatment group (n = 3) received a single dose of 200 mg/kg phenobarbital together with a dose of 150 mg/kg omeprazole at day 57, followed by two more treatments with just omeprazole on days 58 and 59. Mice in the PB/OME post- treatment group (n = 3) received a single dose of 200 mg/kg phenobarbital at age day 57, then three consecutive doses of 150 mg/kg/day omeprazole were initiated three days later, beginning at age day 60. In all cases, gastric pH and liver collection were performed 1 hour following the final treatment of omeprazole. (B) Gastric pH represented as mean ± S.D. in the mouse stomach at 1 hour after the final omeprazole treatment. (C) Relative fold changes of mRNA of CYP2C29 and (D) CYP3A11 and (E) protein expression of CYP3A11 in the mouse liver. ***p < 0.001 Chapter 3 68

To exclude an effect on proton-pump activity by phenobarbital treatment, a group of mice

was treated with phenobarbital only and then gastric pH was measured. These mice had a

gastric pH of 2.7 ± 0.3 in an unfasted condition, and this was not a statistically signiﬁcant

difference from the gastric pH of the control group. Gene expression of Cyp2c29 and Cyp3a11

in the liver was also determined by RT-qPCR following the completion of omeprazole treatment

(Figure 3.2C, D, E). Compared with the omeprazole control group, expression of both Cyp2c29

and Cyp3a11 mRNAs were signiﬁcantly induced to 7.3 ± 0.9− (***p < 0.001) and 12.6 ±

0.8− (***p < 0.001) fold higher, respectively, in the co-treatment group of mice that received phenobarbital and omeprazole concurrently. Mice that received phenobarbital three days prior to beginning omeprazole treatment had no signiﬁcant induction in either Cyp3a11 or Cyp2c29 mRNA expression following the ﬁnal dose of omeprazole. Correlating changes in Cyp3a11 protein were also observed. These results indicate that phenobarbital-mediated CYP induction in adult mice is not long-term, and induced levels of Cyp3a11 and Cyp2c29 mRNAs will return to normal at least six days after phenobarbital treatment is ceased.

3.3.3 Neonatal administration of phenobarbital reduces the efﬁcacy of

omeprazole in adulthood.

For these experiments, neonatal mouse pups were administered phenobarbital at day 5 of age and were then treated with omeprazole in adulthood as previously described in the prior experiments (Figure 3.3A). Neonatal treatment with vehicle control (neo vehicle/adult OME) did not affect omeprazoles ability to inhibit proton-pump activity and gastric pH in adult stomach was consistently 4.5 ± 0.2. Mice that only received phenobarbital as neonates and no omeprazole Chapter 3 69 in adulthood (Neo single-PB/adult vehicle) had gastric pH levels of 2.3 ± 0.8, consistent with un-fasted high stomach acidity in adult mice from the previous experiments. This indicates that neonatal exposure to phenobarbital does not affect gastric pH in adulthood. However, mice that were treated with phenobarbital as neonates then challenged with omeprazole as adults (neo single-PB/adult OME), however, had signiﬁcantly lower gastric pH levels (3.6 ± 0.1, **p <

0.01) than adult mice that received omeprazole and the vehicle control as neonates (4.5 ± 0.2).

Mice that received multiple doses of phenobarbital as neonates (Neo multi-PB/adult OME) exhibited even lower gastric pH levels after omeprazole treatment in adulthood (3.1 ± 0.2, **p

< 0.01) (Figure 3.3B). These observations suggest that phenobarbital exposure at the neonatal stage reduces the efficacy of omeprazole in adulthood to inhibit gastric proton-pumps to increase the pH of the stomach. This indicates a type of unconventional DDI, in which prior phenobarbital treatment is still able to decrease the efficacy of omeprazole as if they were administered concomitantly, when in reality significant time elapsed between the administrations of both drugs.

Based on our hypothesis, the decrease in omeprazole efﬁcacy in adulthood could be caused by increased rates of drug metabolism due to the permanent induction of CYPs caused by phenobarbital exposure at the neonatal age. Therefore, we measured the expression of Cyp3a11 and

Cyp2c29 in the livers of these mice as well. The administration of phenobarbital to neonates caused the expression of Cyp3a11 and Cyp2c29 mRNA to be up-regulated in adulthood following both vehicle and omeprazole treatment (Figure 3.3C, D). Cyp3a11 protein was also overexpressed in these animals (Fig. 3.3E). Administering multiple doses of phenobarbital to neonates also appeared to increase the variability of expression of the CYPs in adulthood Chapter 3 70

FIGURE 3.3 Chapter 3 71

FIGURE 3.3: Neonatal administration of phenobarbital causes a long-term drug-drug interaction and reduces efficacy of omeprazole in adult mice. (A) A schematic of animal treatment. Mice in the neo vehicle/adult OME control group (n = 3) received vehicle (PBS) at the neonatal age of day 5 and three consecutive doses of 150 mg/kg/day omeprazole (OME) at the adult ages of day 57, 58, and 59. The mice in the neo single PB/adult vehicle group (n = 3) received a single dose of 200 mg/kg phenobarbital (PB) at age day 5 and three consecutive doses of vehicle PBS at ages day 57, 58, and 59. Mice in the neo single PB/adult OME group (n = 3) received a single dose of 200 mg/kg phenobarbital at age day 5 and three consecutive doses of 150 mg/kg/day omeprazole at ages day 57, 58, and 59. Mice in the neo multi PB/adult OME group (n = 3) received three consecutive doses of 200 mg/kg/day phenobarbital at ages day 5, 6, and 7 followed by three consecutive doses of 150 mg/kg/day omeprazole at ages day 57, 58, and 59. In all groups, gastric pH and liver collection were performed 1 hour following the final treatment of omeprazole. (B) Gastric pH represented as mean S.D. in the mouse stomach 1 hour after final treatment of omeprazole. (C) Relative fold changes of mRNA of CYP2C29 and (D) CYP3A11 and (E) protein expression of CYP3A11 in the livers. *p < 0.05 and **p < 0.01. compared to mice that only received a single dose of phenobarbital as neonates.

3.4 Discussion

This study presents a case in which drug treatment at the neonatal age was able to affect the activity and efficacy of a different drug administered much later in adulthood. Omeprazole is a medication utilized to inhibit gastric proton-pumps to decrease stomach acidity and raise pH levels. When phenobarbital and omeprazole are administered concurrently in adulthood, the efficacy of omeprazole is attenuated and gastric pH levels do not increase as significantly as they do in the absence of phenobarbital. When phenobarbital is administered during the neonatal stage, there still appears to be an interaction between the two drugs even when omeprazole is administered much later in adulthood. This observation challenges the current definition of a

DDI, and suggests that perhaps two drugs do not need to be administered concurrently for an interaction to occur if one of the drugs is administered at the neonatal age. We hypothesize Chapter 3 72 that this is due to the inducer drug, phenobarbital, permanently up-regulating the expression and activity of Cyp3a11 and Cyp2c29 after treatment during neonatal life, which alters the metabolism of omeprazole in adulthood, increasing the rate of its clearance and thus lowering its efﬁcacy.

Phenobarbital, along with several other first-generation antiepileptic and sedative drugs, is known to cause induction of several different CYP enzymes in the liver of both rodents and humans [78]. Because of this, physicians are currently aware of the risk of DDIs with many commons medications in patients prescribed phenobarbital [85]. However, the current clinical practice only takes DDIs into considerations when multiple drugs are administered at the same time. Historical usage of drugs in previous weeks, months, or years is not a consideration for DDI risk, as most cases of drug-mediated CYP induction are temporary and expression of the enzymes return to normal once the inducer drug is discontinued. Whereas this is true for adults, our findings suggest that CYP induction at the neonatal age can lead to permanent consequences on drug metabolism later in life. This could eventually lead to a change in how DDI risk is evaluated and significantly impact the clinical practice.

Studies from Dr. Bernard Shapiros laboratory have demonstrated that permanent induction of

CYP expression in the adult rat liver can be achieved when neonates are exposed to phenobarbital [116, 117, 119]. A previous study from our own laboratory confirmed this observation in mice, and also further illustrated that the phenomenon was dependent on both the dose of phenobarbital administered, and the age at which the drug was given [120]. In our current study, we additionally showed that the permanent induction of CYPs resulting from neonatal Chapter 3 73 phenobarbital exposure may be associated with a significant decrease in the efficacy of gastric proton-pump inhibition by omeprazole administered in adult life. Other prior studies have shown that neonatal treatment with compounds that induce drug-metabolizing CYPs may affect the action of other compounds in adulthood. For example, neonatal phenobarbital exposure decreased sleep time in vivo in adult rats treated with hexobarbital [119], and neonatal exposure to the constitutive androstane receptor ligand TCPOBOP decreased paralysis time in adult mice treated with zoxazolamine [86]. However, neither hexobarbital nor zoxazolamine are currently used as pharmaceutical medications. Omeprazole, on the other hand, is a very commonly utilized drug across patient populations, and our study has shown that its efficacy can also be affected by prior phenobarbital exposure at the neonatal age.

This study presents a set of observational experiments used to establish a concept in a rodent model. This concept of long-term, unconventional DDIs resulting from neonatal drug exposure will need to be conﬁrmed in additional models with a wider range of inducer and victim pharmaceutical compounds. The efﬁcacy of midazolam, another sleep-inducing agent, could easily be measured and is metabolized primarily by CYP3A enzymes. Other drugs that induce

CYP expression and activity, such as phenytoin and dexamethasone, could also be administered at the neonatal age to determine whether this phenomenon is restricted to phenobarbital or if other inducers are capable of producing similar effects in adult drug metabolism. Effective dose ranges and exposure-sensitivity windows will also need to be established for each drug.

Conversely, if neonatal exposure to additional inducer drugs does not permanently alter CYP expression into adulthood, it may be worth investigating what makes phenobarbital unique and able to produce permanent changes to hepatic metabolic function. Chapter 3 74

Furthermore, the underlying molecular mechanisms also need to be explored for an explanation of how neonatal drug exposure is capable of permanently altering the expression and metabolic activity of CYP enzymes. Epigenetic mechanisms, including DNA methylation, histone modiﬁcations, and microRNAs, have been implicated in regulating the expression of drug-metabolizing enzymes in both neonatal and adult livers [92, 98, 158]. Additionally, phenobarbital treatment does not only cause changes in the gene expression of CYP enzymes.

Many drug-metabolizing enzymes are transporters are altered following phenobarbital treatment, however it is unknown whether all of these are permanently impacted following neonatal exposure as well. Chapter 4

Long-term Effects on the Liver

Transcriptome Following Neonatal

Phenobarbital Exposure

4.1 Introduction

Drug metabolizing enzymes (DMEs) and transporters are expressed at high levels in the liver in both constitutive and induced states. DMEs are responsible for modifying xenobiotic substrates that enter the body through metabolic reactions that increase the solubility of the compound to facilitate eventual excretion. Phase I enzymes catalyze hydrolysis, reduction, and

75 Chapter 4 76 oxidation reactions, whereas Phase II enzymes conjugate water-soluble groups to the modified compounds. Phase III transporters provide systems of both influx and efflux to transport compounds in and out of hepatocytes. The cytochrome P450 family of enzymes (CYPs) are monooxygenases that make up the majority of enzymes responsible for phase I oxidative metabolism. Their substrates consist of a wide variety of both endogenous and exogenous substances in addition to pharmaceutical drugs, including environmental toxicants, dietary components, and endogenous hormones. Following phase I metabolism, phase II enzymes carry out conjugation reactions to facilitate detoxification and consist mainly of transferases, including

UDP-glucuronyltransferases (UGTs), sulfotransferases (SULTs), N-acetyltransferases (NATs), glutathione S-transferases (GSTs), and methyltransferases. Finally, the modified compounds can be effluxed out of the hepatocyte via ATP-binding cassette (ABC) and solute carrier (SLC) phase III transporters. The expression of phase I and II enzymes and phase III transporters can be induced by xenobiotic exposure, which causes an increase in their activity and ultimately affects the clearance and efficacy of the drug and other co-administered compounds [68].

The expression and activity of the phase I CYP enzymes have a large influence on the clearance rate of drugs and can affect how slowly or quickly a compound is excreted by the body as this is the first metabolic step an exogenous compound will undergo. The exposure to certain xenobiotics and pharmaceutical drugs can cause the induction of CYP genes, which up-regulates their activity by increasing the rate of their transcription. Identifying commonly used pharmaceutical drugs that induce CYP activity is critical to avoid drug-drug interactions (DDIs), which can decrease the efficacy or increase the toxicity of additional medications taken at the same time as an inducing drug. Medications such as phenobarbital, phenytoin, rifampicin, and Chapter 4 77 dexamethasone are classic examples of CYP inducers that are implicated in many DDIs. CYP induction is traditionally only studied in the context of adults, and it is assumed mechanisms of enzyme induction are similar in infants and children. In adults, CYP induction is a temporary and reversible event, and enzyme expression returns to basal levels once the inducer drug is discontinued and fully excreted. However, our research has demonstrated that exposure to phenobarbital during the neonatal stage of development may permanently up-regulate the expression of several hepatic CYPs in adult life, as explained in Chapter Three [159]. We therefore contemplated whether neonatal phenobarbital exposure permanently impacts additional genes important to liver function in adulthood.

Phenobarbital is known to up-regulate the expression of many other genes in addition to CYP enzymes. Several in vitro studies in human hepatocyte cell lines have examined transcriptomic and proteomic changes in response to phenobarbital treatment. In HepaRG cells, exposure to phenobarbital caused signiﬁcant induction of multiple CYPs and other genes involved in drug metabolism, including phase II transferases and phase III transporters. Additionally, dozens of other genes involved in lipid biosynthesis, hormone metabolism, digestion, carboxylic acid metabolism were also induced as a result of phenobarbital exposure. Suppression occurred in genes relating to cell adhesion, defense responses, and the regulation of cell proliferation

[160, 161]. It is hypothesized that phenobarbital is capable of affecting the expression of so many genes due to its ability to activate both the CAR and PXR nuclear receptors, which act as transcription factors for many target genes in the liver beyond CYP enzymes. However, it is unknown whether neonatal exposure to phenobarbital permanently alters the expression of other genes known to be induced by the drug. Chapter 4 78

This study utilizes RNA-sequencing (RNA-seq) in a mouse model to characterize the long- term consequences on gene expression in the adult liver following phenobarbital exposure at the neonatal age in vivo. It is not feasible to study the developmental effects of phenobarbital exposure in a cell model, so an animal model must be utilized. There are known differences in the effects of the activation of murine CAR versus human CAR, and species differences in the response to phenobarbital [162]. Therefore, this study will serve as proof of a concept that exposure to phenobarbital at the neonatal age may cause permanent transcriptomic changes that can impact the function of the liver through adulthood.

4.2 Materials and Methods

Animals. Eight-week-old C57Bl/6 mice were purchased from Jackson Laboratories (Bar Har- bor, ME) and set up into breeding pairs. Mice were housed according to the Animal Care

Guidelines provided by the American Association for Animal Laboratory Sciences and were bred under standard conditions in the Animal Care Facility at the University of Connecticut.

The use of these animals was approved by the Institutional Animal Care and Use Committee.

On day 5 after birth, male mice were administered either saline vehicle control (PBS) or 200 mg/kg phenobarbital (phenobarbital sodium salt) intraperitoneally (n = 3 for each group). At age day 60 after birth, all mice were sacriﬁced and livers were collected and snap frozen in liquid nitrogen. Samples were stored at -80 °C for further analyses.

Total RNA Extraction. Total RNAs were isolated from liver tissue using TRIzol reagent as previously described in Chapter Three. The integrity of RNAs was evaluated on an Agilent Chapter 4 79

2200 Tape Station (Agilent Technologies, Santa Clara, CA), and samples with RIN integrity values of greater than 8.0 were acceptable for sequencing library construction. cDNA Library Construction. cDNA libraries from all samples were prepared using an Illu- mina TruSeq RNA Sample Prep Kit (Illumina, San Diego, CA) with 3 g of total RNA used for RNA input. Messenger RNAs were isolated from total RNAs by poly(A) selection using poly(T) primers. RNA fragmentation, ﬁrst-, and second-strand cDNA syntheses, end repair, adapter ligation, and PCR ampliﬁcation were performed according to the manufacturers protocol. Fragments of the cDNA library ranged from 220 to 500 bp with a peak size of 280 bp

(including the 120-bp adapter sequence). Quality of the cDNA libraries was evaluated using an

Agilent 2100 Bioanalyzer before sequencing.

RNA-sequencing. RNA-seq was performed on an Illumina HiSeq 2000 instrument (Perkin

Elmer, Branford, CT). Clusters of the cDNA libraries were generated on a TruSeq ﬂow cell and sequenced for 100-bp paired end reads (2 x 100) with a TruSeq 200-cycle SBS kit (Illumina).

A PhiX bacteriophage genome and a universal human reference RNA were used as controls and sequences in parallel with other samples to ensure that the data generated for each run were accurately calibrated during the image and data analysis. In addition, the PhiX was spiked onto each cDNA sample at approximately 1% as an internal quality control.

RNA-seq Data Analysis. After the sequencing images were generated by the platform, the pixel-level raw data collection, image analysis, and base calling were performed using Illu- mina Real Time Analysis software. The output .bcl files were converted to qseq files by Illu- mina BCL Converter 1.7 software and subsequently converted to .FASTQ files for downstream analyses. The RNA-seq reads from the FASTQ files were mapped to the mouse reference Chapter 4 80

genome (NCBI37/mm9) and splice junctions were identiﬁed by TopHat 1.2. The output ﬁles in

.BAM format were analyzed by Cufﬂinks 1.0.3 to estimate transcript abundance. The mRNA

abundance was expressed as the number of fragments per kilobase of exon per millions reads

mapped (FPKM).

Statistical Analysis. Differentially expressed genes in the phenobarbital-treated animals were identiﬁed using the program Cuffdiff. Genes with an adjusted P-value 0.05 determined by one- way ANOVA, FDR < 0.001, and log2 fold change ≥ 2 or ≤ −2 were considered signiﬁcantly

differentially expressed. CummeRbund was utilized to visualize and integrate data produced

by Cuffdiff analysis.

Gene Ontology Analysis. Gene ontology (GO) was performed using the Database for Annota-

tion, Visualization, and Integrated Discovery (DAVID, v6.8). GO terms with a p-value < 0.05

and a FDR < 1 were considered statistically signiﬁcantly enriched.

Transcription Factor Enrichment. Enriched transcription factors known to bind to the pro-

moter regions of selected genes were identiﬁed using the ENCODE ChIP-Seq Signiﬁcance

Tool found at encodeqt.simple-encode.org. Lists of signiﬁcantly up-regulated and signiﬁcantly

down-regulated genes were input into the program and analyzed against the mm10 background

for transcription factors experimentally shown to bind within 2,000 bp upstream or 500 bp

downstream of the TSS 5‘ end. Q-values were calculated using the hypergeometric test accord-

ing to Benjamini-Hochberg. Chapter 4 81

4.3 Results

4.3.1 Long-term effects on the liver transcriptome following neonatal phe-

nobarbital exposure.

Mice were treated with a single dose of either PBS vehicle control of 200 mg/kg phenobarbital at the age of day 5 after birth. At age day 60 after birth, the mice were sacriﬁced and livers were collected for transcriptome analysis by RNA-seq. A total of 10,499 genes were found to be expressed in the liver. The similarity of the transcriptomes between samples were evaluated by calculating correlations (Figure 4.1). The transcriptomes of control mice were more similar to each other, while neonatal phenobarbital appeared to increase the variability in transcriptomes of treated mice, as evidenced by the lower r-values of correlation between the treated mice.

We identiﬁed differentially expressed genes between the two groups of mice with the program

Cuffdiff. Genes with an adjusted p-value of < 0.05 and a FDR < 0.001 were considered signiﬁcantly differentially expressed. Based on this criteria, a total of 1,649 genes were found to be differentially expressed in adult livers of mice treated with phenobarbital at the neonatal age.

Of these differentially expressed genes, 926 were up-regulated and 723 were down-regulated in comparison to vehicle-treated mice. However, we narrowed these genes down further by selecting for those that had a fold-change ≥ 2 or ≤ −2. Of these, 166 genes were up-regulated by at least two-fold, and 57 were down-regulated by at least two-fold (Figure 4.2).

The most signiﬁcantly up-regulated genes included CYPs, SULTs, and other genes involved in detoxiﬁcation such as Fmo3 and drug transporters. The CYPs with the greatest fold-changes, Chapter 4 82

FIGURE 4.1: Correlation matrix of transcripte proﬁles among adult mice treated with phenobarbital or control at the neonatal age. Overall correlation for the expression of each gene expressed in the liver (in FPKM) was calculated for adult mice (n = 3) that received either PBS (C 1, C 2, and C 3) or 200 mg/kg phenobarbital (PB 1, PB 2, and PB 3) at day 5 of age. Blue represents r = 1, while red would represent r = -1. Chapter 4 83

FIGURE 4.2: Summary of RNA-seq results. Summary of the results obtained from RNA-seq analyses on adult livers following day 5 treatment with 200 mg/kg phenobarbital, compared to age-matched mice that received vehicle control (PBS) on day 5 (n = 3 for both groups). however, were genes not typically associated with drug metabolism in the mouse, and included Cyp11b1, Cyp2b13, and Cyp2b9. In fact, these genes were not even expressed in vehicle-treated mice. Interestingly, the gene with the highest fold-increase was the hydroxyl- delta-5-steroid-dehydrogenase, 3 beta- and steroid delta-isomerase 1 (Hsd3b1). This enzyme plays an important role in the biosynthesis of all hormonal steroids, including the catalysis of pregnenolone to progesterone and DHEA to 4-androstenedione, along with the metabolism of testosterone. This gene was not expressed in control males, but was up-regulated over 10- fold in phenobarbital-treated mice. The gene with the greatest degree of suppression of was

Serpina4, also known as kallistatin. The protein functions as an inhibitor of kallikrein, which has amdiolytic and kininogenase activity, to protect tissues against inﬂammation, ﬁbrosis, and oxidative stress [163]. Chapter 4 84

Cytochrome P450s

−2 0 1 2 Row Z−Score

Cyp3a16 Cyp2c70 Cyp21a1 Cyp3a44 Cyp2c39 Cyp2g1 Cyp3a13 Cyp2c69 Cyp2b9 Cyp2b13 Cyp2a4 Cyp2a22 Cyp39a1 Cyp2c68 Cyp4f16 Cyp17a1 Cyp3a41a.1 Cyp3a41a Cyp2c38 Cyp3a11 Cyp2c55 Cyp2c54 Cyp2c50 Cyp2a12 Cyp2c29 Cyp2b10 Cyp4f15 Cyp2d22 Cyp2a5 Cyp3a59 Cyp3a25 Cyp8b1 Cyp2d13 Cyp4a31 Cyp4a10 Cyp20a1 Cyp4a14 Cyp26a1 Cyp1a2 Cyp2j6 Cyp2c53.ps Cyp2d11 Cyp2d10 Cyp2e1 Cyp2d26 Cyp4b1 Cyp2d12.1 Cyp2d12 Cyp7b1 Cyp4a12b Cyp2u1 Cyp4a12a Cyp2d9 Cyp2d40 Cyp4v3 Cyp4f13 Cyp26b1 Cyp2f2 Cyp2c67 Cyp4f17 Cyp2r1 Cyp2c44 Cyp27a1 Cyp7a1 Cyp2j9 Cyp4f14 Cyp2d37.ps C1 C2 C3 PB1 PB2 PB3

FIGURE 4.3 Chapter 4 85

FIGURE 4.3: Comparison of the expression of all expressed cytochrome P450 enzymes (FPKM > 1) in adult liver following neonatal phenobarbital treatment. Heatmap of expression proﬁles for all expressed CYPs (FPKM > 1) in adult livers.

4.3.2 Long-term effects on drug metabolism following neonatal pheno-

barbital exposure.

To evaluate global changes to genes involved in drug metabolism and transport, heatmaps were created that included phase I and phase II enzymes and phase III transporters. When visualiz- ing the data, red coloring represents gene expression higher than the average of all the samples, while green coloring represents gene expression lower than the average of all the samples. Of all phase I enzymes, 75 were signiﬁcantly affected in adults by neonatal phenobarbital exposure (Figure 4.3, Figure 4.4). Most were up-regulated, although a few genes, such as Aldh4,

Cyp2d40, and Pon2, were significantly down-regulated. There are also inconsistencies among up- and down-regulated genes between the phenobarbital-treated samples. Specifically, the PB1 sample shows a more distint gene expression profile from the PB2 and PB3 samples, which are more similar. There are a cluster of CYPs in the PB1 sample that are not up-regulated compared to control mice, as in PB2 and PB3 samples. CYPs that were induced in all three samples do include Cyp2b10, which was up-regulated 8-fold, and Cyp2c29 and Cyp3a11 which were up-regulated 3- and 1.5-fold, respectively. While this confirms our observation from earlier

RT-PCR experiments that showed neonatal phenobarbital up-regulates CYP expression in adult life, it does not reﬂect the large extent of enzyme induction that was observed in RT-PCR gene expression evaluation. The Cyp4 enzymes also appear to be commonly down-regulated among Chapter 4 86 all three phenobarbital-treated mice. Other phase I drug-metabolizing enzymes that were sig- niﬁcantly up-regulated included Cyp1a2 (0.7-fold), Aldh1a1 (1.3-fold), Fmo3 (8-fold), and

Ces2a (2.5-fold). There are also inconsistences in the rest of the phase I enzymes changed by phenobarbital treatment, as many enzymes up-regulated in samples PB2 and PB3 are not up-regulated in PB1. These data suggest that while phenobarbital treatment at the early age deﬁnitely alters phase I drug-metabolizing enzymes in adult life, the changes in gene expression may not always be the same in all animals. Therefore, neonatal phenobarbital treatment can increase variability in gene expression in adult life.

Additionally, 32 phase II enzymes were also signiﬁcantly differentially expressed in phenobarbital- treated mice (Figure 4.5). Again, most genes were up-regulated, but a few genes including

Ugt2a1 and Ugt2a3, were down-regulated. Many of the SULTs had the largest fold-increases in expression, including Sult3a1, Sult2a1, Sult1e1, and Sult2a2, which had fold-changes of 9.6,

8.8, 8.7, and 8.6, respectively. Again, there was variability in the up- or down-regulation of specific phase II enzymes in response to neonatal phenobarbital treatment. Phase III transporters were also impacted, with 27 showing significant differential expression in phenobarbital-treated mice (Figure 4.6, Figure 4.7). However, the differences in expression were not as dramatic as with phase I and phase II enzymes. SLC transporters appeared to be more significantly impacted than ABC transporters, with Slc22a26 having the greatest fold-increase (6.5-fold). Chapter 4 87

Phase I enzymes (excluding CYPs)

−2 0 1 2 Row Z−Score

Adh7 Adh5 Aldh3b1 Akr1b8 Aldh16a1 Akr1e1 Ces2e Akr1c14 Aox1.1 Aox1 Sdr9c7 Ces3a Adh4 Txnrd1 Ces3b Adh6.ps1 Aox3.1 Aox3 Ephx2 Ces4a Ces1f Akr1c12 Akr7a5 Txnrd3 Cyb5b Pon2 Sdr39u1 Cyb5r1 Cyb5r3 Ces1e Pon3 Aldh5a1 Txnrd2 Akr1c19 Akr1c13 Dhdh Sdr42e1 Adhfe1 Cyb5 Aldh1a1 Mocos Ces2a Pon1 Ces1g Aldh9a1 Cyb5d1 Ces2g Akr1b3 Fmo3 Cybb Aldh6a1 Ces1b Xdh Aldh1a7 Akr1d1 Nqo1 Fmo2 Fmo1 Ces1c Nqo2 Aldh4a1 Fmo4 Akr1b7 Akr1b10 Cyba Aldh1l1 Aldh1b1 Cyb5r4 Cyb5d2 Cyb561d2 Cyb5rl Ces2d.ps Ces2c Gphn Aldh7a1 Fmo5 Dpyd Adh1 Akr1c20 Ces1d Aldh3a2 Aldh2 Ephx1 Akr1a1 C_1 C_2 C_3 PB_1 PB_2 PB_3

FIGURE 4.4 Chapter 4 88

FIGURE 4.4: Comparison of expression of phase I enzymes in adult liver following neonatal phenobarbital exposure. Comparison of the expression of the rest of expressed phase I enzymes (FPKM > 1, excluding CYPs) in adult livers following neonatal treatment with PBS (C 1, C 2, C 3) or phenobarbital (PB 1, PB 2, PB 3).

4.3.3 Long-term effects on liver functions following neonatal phenobar-

bital exposure.

To further understand the major hepatic functions and pathways altered by neonatal phenobarbital treatment, we submitted the lists of differentially expressed genes for ontology analysis by the program DAVID. For the genes up-regulated in adult liver by phenobarbital treatment

(Figure 4.8A), the most significantly enriched biological processes affected were the epoxygenase P450 pathway, oxidation-reduction process, digestion, steroid metabolic process, and proteolysis. Many other processes the program identified were involved in lipid metabolism, drug metabolism, detoxification, and hormone biosynthesis. Most of these genes in this list had the molecular functions of steroid hydroxylase activity, iron ion binding, arachidonic acid epoxygenase activity, and aromatase activity. The up-regulated genes were found to be significantly involved in the steroid hormone biosynthesis, chemical carcinogenesis, pancreatic secretion, and arachidonic acid metabolism KEGG pathways. We also examined the biological processes that were enriched in genes that were down-regulated by neonatal phenobarbital exposure (Figure 4.8B). Interestingly, the most significant process suppressed was circadian rhythm and its regulation and effects on gene expression. The circadian rhythm and MAPK signaling KEGG pathways were also identified as significantly enriched. Other processes that were enriched were involved in gene transcription, cell differentiation, apoptosis, and liver development. These genes involved in these processes mainly had transcription factor activity or Chapter 4 89

Phase II enzymes

−1 1 Row Z−Score

Ugt2b36 Ugt2b34 Tpmt Nat2 Ugt2b37 Ugt1a6b Ugt2b1 Ugt3a2 Ugt2b35 Nat8 Nat10 Gstcd Ephx1 Ugt1a6a Ephx2 Ugt2b5 Ugt2b38 Gstp1 Ugt3a1 Comt Ugt2a3 Nat6 Gstp2 Sult1a1 Gstm7 Sult2a5 Sult2a2 Sult2a7 Sult2a1 Sult2a3 Sult1e1 Gstz1 Gstm5 Gstk1 Nat9 Nat1 Ugt1a1 Gstt3 Sult5a1 Gstm3 Gsta1 Sult1d1 Gstt1 Gstm1 Gstm6 Gstm2 Sult1b1 Gsta4 Gsta2 Gstm4 Sult1c2 Ugt1a9 Gsta3 Ugt1a5 Ugt1a2 Gstt2 Gsto1 C_1 C_2 C_3 PB_1 PB_2 PB_3

FIGURE 4.5 Chapter 4 90

FIGURE 4.5: Comparison of expression of phase II enzymes in adult liver following neonatal phenobarbital exposure. Comparison of the expression of all expressed phase II enzymes (FPKM > 1) in adult livers following neonatal treatment with PBS (C 1, C 2, C 3) or phenobarbital (PB 1, PB 2, PB 3).

Phase III transporters (ABCs)

−1 1 Row Z−Score

Abcf3 Abcd4 Abcc6 Abcc2 Abcb11 Abca8b Abcg8 Abcg5 Abca3 Abcd2 Abcc4 Abcd1 Abcc3 Abcb8 Abcc9 Abca8a Abcc10 Abca2 Abcd3 Abcb7 Abcb4 Abcb6 Abca6 Abca7 Abcf1 Abca1 Abcf2 Abce1 Abcg2 Abcb10 C_1 C_2 C_3 PB_1 PB_2 PB_3

FIGURE 4.6: Comparison of the expression of phase III ABC transporters in adult liver following neonatal phenobarbital exposure. Comparison of the expression of all expressed phase III transporters belonging to the ABC family (FPKM > 1) in adult livers following neonatal treatment with PBS (C 1, C 2, C 3) or phenobarbital (PB 1, PB 2, PB 3). Chapter 4 91

Phase III transporters (SLCs)

−2 0 1 2 Row Z−Score

Slc25a47 Slc22a1 Slc17a4 Slc25a42 Slc6a12 Slc22a27 Slc22a26 Slc48a1 Slc24a6 Slc23a2 Slc23a1 Slc47a1 Slc22a18 Slc25a51 Slc16a7 Slco1a4 Slco2b1 Slc29a1 Slc27a5 Slc31a1 Slc38a10 Slc16a12 Slc46a1 Slc35f5 Slc35c2 Slc3a1 Slc13a3 Slc39a14 Slc16a10 Slc9a3r1 Slc16a1 Slc39a1 Slc25a5 Slc10a1 Slc25a3 Slc26a1 Slc1a2 Slc37a4 Slc27a2 Slc39a7 Slc22a7 Slc25a45 Slc35b1 Slc41a2 Slc25a25 Slc12a7 Slc40a1 Slc19a2 Slc17a2 Slc25a30 Slc22a23 Slco1a1 Slc25a17 Slc25a16 Slc17a5 Slc46a3 Slc35a3 Slc20a1 Slc2a2 Slc25a10 Slc25a15 Slc38a3 Slc25a44 Slc7a2 Slc25a1 Slc38a4 Slc38a2 Slc16a2 Slco2a1 Slc39a4 Slc25a23 Slco1b2 Slc25a13 Slc6a13 Slc10a5 Slc35d1 Slc25a20 Slc30a9 Slc25a22 Slc20a2 Slc22a30 Slc22a28 Slc33a1 Slc30a5 Slc17a3 Slc25a39 Slc25a11 C_1 C_2 C_3 PB_1 PB_2 PB_3

FIGURE 4.7 Chapter 4 92

FIGURE 4.7: Expression of phase III SLC transporters in adult liver following phenobarbital exposure at the neonatal age. Comparison of the expression of phase III transporters belonging to the SLC family following neonatal treatment with PBS (C 1, C 2, C 3) or phenobarbital (PB 1, PB 2, PB 3). Due to the large number of SLC transporters expressed, only genes with FPKM > 15 are shown. were involved in histone deacetylase and DNA binding functions.

We also submitted up- and down-regulated gene lists for transcription factor enrichment analysis (Figure 4.8C). This program identiﬁes common transcription factors validated by chromatin immunoprecipitation experiments known to bind to the promoter regions of the genes submitted. For up-regulated genes, signﬁcant common transcription factors included p300 and CCCTC-binding factor (CTCF), which were experimentally shown to bind to over 50 of the genes were submitted. Of the down-regulated genes, common transcription factors also included p300 and CTCF, along with ETS proto-oncogene 1 (ETS1), negative elonga- tion factor complex member 3 (NELFe), MYC associated factor X (Max), and Max interactor

1 (Mxi1). Interestingly, even though fewer genes were signiﬁcantly down-regulated by phenobarbital treatment, there was a greater number of common transcription factors shown to bind to these genes. There was also overlap in the transcription factors that bound to both up- and down-regulated genes, suggesting common mechanisms that activate as well as repress gene transcription in response to phenobarbital. Chapter 4 93

A Number of Biological Process P-value FDR genes Epoxygenase P450 pathways 14 3.15E-20 4.74E-17 Oxidation-reduction process 21 4.21E-07 6.34E-04 Digestion 6 5.92E-06 8.91E-03 Steroid metabolic process 7 5.61E-05 8.43E-02 Proteolysis 15 2.32E-04 3.48E-01 Lipid metabolic process 13 3.12E-04 4.68E-01 Sulfation 4 5.70E-05 8.57E-02

B Number of Biological Process P-value FDR genes Circadian rhythm 6 7.81E-06 1.09E-02 Circadian regulation of gene expression 5 1.72E-05 2.41E-02 Rhythmic process 6 1.79E-05 2.50E-02 Positive regulation of transcription 11 1.62E-04 2.26E-01 Regulation of transcription, DNA-templated 16 3.40E-04 4.75E-01

C Transcription Number of Transcription Number of Q-value Q-value factor genes factor genes Pol2 109 5.58E-05 p300 35 4.95E-11 CTCF 40 8.61E-11 p300 56 6.10E-05 Pol2 51 3.74E-10 CTCF 66 1.20E-03 ETS1 23 1.62E-05 SRF 10 2.68E-04 TCF12 12 1.06E-02 BHLHE40 24 9.28E-04 MyoD 9 3.88E-02 Max 28 1.03E-03 MAZ 25 2.08E-03 CHD2 21 2.36E-03 CEBPB 7 4.27E-03 Mxi1 26 6.28E-03 NELFe 26 9.18E-03 E2F4 13 1.05E-02 GATA2 4 1.05E-02 JunD 6 1.11E-02 USF 7 1.11E-02

FIGURE 4.8: Gene ontology analysis of significantly differentially expressed genes in adult mice following neonatal exposure to phenobarbital. Gene ontology analysis was performed on significantly up-regulated genes (A) and down-regulated genes (B) in adult mice following neonatal exposure to phenobarbital. Enriched biological processes with p < 0.05 and FDR < 0.1 are listed. Enriched transcription factors known to bind to the promoter region of differentially expressed genes were also identfied (C) for up-regulated genes (left) and down-regulated genes (right). Transcription factors with Q < 0.05 are listed. Chapter 4 94

4.4 Discussion

Over the past 30 years, phenobarbital has been repeatedly shown to affect adult hepatic CYP activity in rodents after it is administered at the neonatal stage [115, 119, 120]. The permanent increases to CYP expression resulting from neonatal programming have also been shown to affect the efﬁcacy of other drugs and compounds administered in adulthood, such as omeprazole and zolaxamine [86, 159], presumably due to increased clearance rates. Due to the ability of phenobarbital to activate the CAR nuclear receptor and affect the circulating concentrations of various hormones, we hypothesized that neonatal phenobarbital exposure causes overall global alterations to adult liver gene expression and function in addition to the permanent up-regulation of drug-metabolizing CYPs. Therefore, we performed RNA-seq on liver samples from adult mice that received a single dose of phenobarbital at the neonatal age, and compared gene expression to control mice to identify signiﬁcantly differentially expressed transcripts. Gene ontology analysis revealed multiple biological processes and pathways altered by neonatal phenobarbital exposure in addition to all phases of drug metabolism. These results demonstrate that exposure to a single dose of phenobarbital early in life may have the potential to alter various functions of the liver in adulthood, including lipid metabolism, energy homeostasis, and steroid hormone biosynthesis in adition to xenobiotic biotransformation.

Phenobarbital is known to cause the induction of multiple drug-metabolizing CYP enzymes via the activation of the nuclear receptor CAR [164]. CAR is well-known for its role in regulating the expression of CYPs and other drug-metabolizing enzymes, which is evidenced by the large amount of phase I, phase II, and transporter genes we found to be affected by neonatal phenobarbital treatment. In addition, CAR has also been characterized for its involvement in Chapter 4 95 glucose and lipid metabolism [165] and the promotion of hepatocyte differentiation [86]. This would explain why we also found genes not related to drug metabolism, but rather involved in digestion, lipid metabolism, and proteolysis, to be affected by phenobarbital treatment. CAR activation, particularly by phenobarbital, has also been shown to promote liver tumorigenesis in mice by altering the expression of genes involved in regulating cellular responses to DNA damage and tumor promotion [166], suggesting CAR also plays an important role in the regulation of cell growth and proliferation.

Upon its activation, CAR is translocated into the nuclear to promote the transcription of its target genes. However, our common transcription factor analyses did not reveal CAR as a signiﬁcantly enriched transcription factor shown to bind to the genes we identiﬁed as significantly affected by neonatal phenobarbital exposure. However CAR is also known to recruit a variety of co-activator proteins to the promoter of target genes to affect gene transcription.

Steroid receptor co-activator 3 (SRC-3) has been identified as a prominent co-activator that is recruited with CAR to promote hepatocyte proliferation and induction of drug metabolism upon TCPOBOP treatment [167], but was also not identified in our analyses. However, p300 a significant common transcription factor we identified, has been shown to be a key regulator of various liver functions, including apoptosis, cell proliferation, gluconeogenesis, and liver regeneration [168, 169]. p300 is also heavily involved in proper development including regulating hormone metabolism and cell differentiation [170], which are two processes also identified by our gene ontology analyses. CTCF is also important for regulating epigenetic landscapes during liver development and maturation to promote proper gene expression and function, and was identified as an enriched transcription factor altered by neonatal phenobarbital treatment Chapter 4 96

[171]. Perhaps changes in gene expression resulting from neonatal phenobarbital exposure are due to other factors beyond CAR activation that are involved in more global liver maturation regulation.

Phenobarbital does not directly bind CAR, but rather causes receptor activation through an indirect mechanism that leads to its translocation into the nucleus. This mechanism by which phenobarbital produces CAR activation was elusive for decades, until recently where it has been shown that phenobarbital binds and antagonizes the epidermal growth factor receptor (EGFR), which ultimately causes CAR activation [172]. Following EGFR binding and thus inhibition, phosphorylation on the tyrosine of the scaffold protein RACK1 is abolished. Dephosphorylated

RACK1 is then able to activate protein phosphatase 2A (PP2A). The activated PP2A then dephosphorylates CAR, which causes the translocation of the nuclear receptor into the nucleus where it is able to alter the transcription of its target genes, with the classic example being

CYP2B enzymes. However, there are many other signaling pathways that exist downstream from EGFR activation or antagonism that can be affected by phenobarbital binding apart from

CAR activation [173–175]. It is likely that the permanent alterations to multiple hepatic biological pathways following neonatal phenobarbital exposure are also due to changes in EGFR signaling, rather than just the activation of CAR.

EGFR, also known as ErbB-1, is a membrane receptor known for its role in promoting cell proliferation, cell migration, and tumorigenesis. Following its activation by growth factor ligands, downstream Ras-Raf-MEK-ERK1/2-STAT3/5 and PI3K-Akt-mTOR pathways become Chapter 4 97 up-regulated, which leads to DNA synthesis and subsequent cell proliferation and the promotion of cell survival. EGFR has been shown to be indispensible for proper gestational development in rodents, and EGFR knockout mice are exhibit severe growth retardation [176]. The abolishment of EGFR in hepatocytes speciﬁcally also causes whole body growth retardation and body size reduction, indicating hepatic EGFR plays a role in energy homeostasis that controls the growth of the entire body. EGFR is most highly expressed in the adult liver and is required for proper liver function and tissue regeneration following hepatic injury [177]. During liver regeneration, growth hormone signaling can also indirectly regulate EGFR expression and function. Down-regulation of the growth hormone receptor accompanies decreased expression of EGFR on hepatocytes, which negatively affects proper hepatocyte proliferation and organ growth [178]. Interestingly, neonatal phenobarbital treatment has been shown to permanently alter growth hormone secretion in rodents [117]. Perhaps neonatal phenobarbital exposure either permanently alters EGFR signaling, or the relationship between circualting growth hormone and EGFR, which promotes the differentiation of hepatocytes during development, which would explain the large number of drug-, lipid- and steroid-metabolizing enzymes affected by phenobarbital treatment. Also in support of EGFR signaling modulation, EGFR inhibitors have been shown to up-regulate p300 activity in cancer cells [179].

The regulation of circadian rhythm was the most signiﬁcantly enriched process associated with the down-regulated genes in phenobarbital-exposed mice. It is thought that circadian rhythm evolved due to the increased production of reactive oxygen species (ROS) with the development of redox reactions in the presence of daylight and oxygen. Therefore, it makes sense that the Chapter 4 98 up-regulation of CYPs and other hepatic enzymes by phenobarbital treatment would affect circadian rhythm via the increased production of ROS through reactions involved in metabolism.

Glucose, lipid, and bile acid metabolism have also been shown to follow circadian patterns

[180]. Certain epigenetic modiﬁcations are also controlled by circadian rhythm, and it is possible that epigenetic alterations caused by phenobarbital treatment in neonates become permanent and persist into adulthood, altering gene expression. However, the opposite might be true, in which phenobarbital affects circadian rhythms, and this is in turn altering the expression of the many genes involved in hepatic drug and energy metabolism. The circadian rhythm and time of day of drug administration has been repeatedly shown to signiﬁcantly affect patient pharmacokinetics [181].

Most of the adverse effects associated with phenobarbital use are behavioral or cognitive and include attention deﬁcits, problems with memory, impairment of judgment, sedation, and slowed learning [182]. Growth defects have also been seen in children treated with phenbarbital, suggesting a potential role of EGFR and/or growth hormone disturbances [183]. Long-term use of phenobarbital is also associated with increased risks of liver cancer, decreased bone density, metabolic issues, decreased serum concentrations of thyroid hormones, and even increased risk of early death in adults [184–186]. It is interesting that many of the biological processes and pathways enriched in the livers of mice neonatally exposed to phenobarbital are associated with bone differentiation, energy metabolism, and hormone biosynthesis. Due to the propensity of phenobarbital to cause adverse effects that are not well-tolerated, however, the medication is usually not prescribed for long-term use. In light of this, our study highlights the possibility of permanent effects on the liver following only a single dose of phenobarbital. Future studies Chapter 4 99 will be necessary to determine whether the signiﬁcantly increased or decreased gene transcripts resulting from phenobarbital treatment are actually translated into protein and have functional consequences in the liver, and whether these effects occur in humans. Chapter 5

Long-term Effects of Neonatal Phenytoin

Exposure on Cytochrome P450 Expression

5.1 Introduction

Seizures, whether isolated or due to chronic epileptic conditions, are unfortunately a relatively common occurrence in pediatric patients. Between two and four out of every 1,000 newborns suffer from seizures following birth. It is crucial to control seizures immediately, utilizing therapeutic interventions to prevent brain damage and other complications in young, developing patients. Many AEDs, however, are known to induce CYP expression via the activation of hepatic nuclear receptors, namely CAR and PXR. Following their activation, nuclear receptors translocate into the nucleus to alter gene transcription by the recruitement of co-activator

100 Chapter 5 101 proteins, which alter the epigenetic landscape surrounding promoters. Increased levels of enzymes involved in catalyzing the addition or removal of epigenetic modifications to induce and suppress gene transcription exist in the liver during early postnatal ages, which facilitates the maturation of developing organs and proper programming of gene expression [187]. This may also cause developing organs to be especially susceptible to alterations in normal gene expression ontogeny by xenobiotics that induce transcription via epigenetic modifications. It is also possible for epigenetic modifications made during postnatal organ growth to become permanent and alter cellular responses in adulthood [188].

We have already demonstrated in Chapters Three and Four that neonatal exposure to phenobarbital can permanently alter the expression of hundreds of genes in the liver in adult mice. In particular, CYP genes involved in drug metabolism are up-regulated in response to phenobarbital exposure in early life, and this may affect the efﬁcacy or toxicity of other drugs administered in adulthood [120, 159]. It is now critical to investigate whether this phenomenon occurs following exposure to additional drugs at the neonatal age. Because neonates are more commonly exposed to AEDs rather than other inducer drugs due to the frequency of infant seizures, we decided to investigate the long-term effects of a different AED, phenytoin. Phenytoin is also considered a ﬁrst-line therapy for treating seizures and convulsions in pediatric patients, and is even additionally used to treat arrhythmias.

Phenytoin is known to induce enzymes belonging to the CYP2B, CYP2C, and CYP3A families in both rodents and humans, at least at the adult age. The induction of CYP2B6 by phenytoin has been shown to be exclusively mediated by CAR [84]. The induction is temporary and the Chapter 5 102 expression of these enzymes returns to basal levels once the inducing agent is ceased. Addi- tionally, a study has shown that longer term, chronic administration of phenytoin even abolishes inductive effects of CYP2B and CYP2C enzymes in rats [189]. However, no studies have examined whether neonatal exposure to phenytoin has any effect on CYP expression in the liver in adulthood. Because we have already established that neonatal phenobarbital exposure results in altered expression of drug-metabolizing CYPs in adult life that can diminish the efﬁcacy of other administered drugs, we hypothesized that neonatal phenytoin exposure would result in the same consequences in adults. This is due to the common ability for both phenobarbital and phenytoin to activate the nuclear receptor CAR to up-regulate CYP expression and drug metabolism. We utilized the same mouse model and treated neonatal mice at the age of day 5 with a single dose of phenytoin, then measured the expression of drug-metabolizing CYPs in adulthood at the mRNA, protein, and enzymatic activity levels.

5.2 Materials and Methods

Chemicals and Reagents. Phosphate buffered saline (PBS), 5,5-diphenylhydantoin salt (phenytoin, PHY), triﬂuoroacetic acid (TFA), formic acid, and acetonitrile were purchased from Sigma-

Aldrich (St. Louis, MO). Urea and dithiolthreitol (DTT) were obtained from Fisher Scientiﬁc

Co. (Pittsburgh, PA). Iodoacetamide (IAA) and ammonium bicarbonate were purchased from

Acros Organics (Morris Plains, NJ). TPCK-treated trypsin was obtained from Worthington Bio- chemical Corporation (Freehold, NJ). Bovine serum albumin (BSA) standard was purchased from Thermo Fisher Scientiﬁc (Waltham, MA). Chapter 5 103

Animals. C57Bl/6 mice were housed in accordance with the animal care guidelines outlined by the American Association for Animal Laboratory Sciences and were bred under standard conditions in the Animal Resources Facility at the University of Connecticut. The use of these animals was approved by the University of Connecticuts Institutional Animal Care and Use

Committee. At 12 weeks of age, mice were set up into breeding pairs. To study the long-term effects of drug treatment at the neonatal age, mice were administered a single dose of phenytoin (100 mg/kg) or PBS vehicle control intraperitoneally at age day 5 after birth. Following treatment, mice were sacriﬁced at different ages (day 10, 15, 20, 30, or 60 after birth) and livers without gallbladders were collected and stored at -80 °C. Re-challenge experiments were also performed in which mice received phenytoin treatment at age day 5, and were then re- challenged with a dose of phenytoin at day 60. For each age group, 4-6 livers from different mice were collected and both male and female mice were included in this study.

Total RNA Extraction, Reverse Transcription, and RT-PCR. Total RNAs were isolated from whole livers using TRIzol reagent, and reverse transcription was performed using iScript Re- verse Transcription Supermix as previously described. RT-PCR was performed using a CFX-96 thermocycler system with TaqMan gene expression assays for Cyp2b10, Cyp2c29, Cyp3a11, and Gapdh as an endogenous control.

Total Sample Preparation and LC-MS/MS-based Quantification. Liver microsomes were obtained from S9 liver homogenates prepared from whole liver tissue samples. Protein concentration was determined using a Pierce BCA protein assay kit (Thermo Fisher Scientific). Pro- tein digestion was conducted according to a previously reported Lyc-C/Trypsin combinatorial digestion protocol with minor modifications. Aliquots of 80 µg protein from liver microsomes Chapter 5 104

were mixed with 0.2 µg of bovine serum albumin (BSA) internal standard. After adding ace- tone, the mixture was incubated and centrifuged, then precipitated proteins were isolated and washed with ethanol. Proteins were re-suspended in 4 mM DDT in 8 M urea solution containing 100 mM ammonium bicarbonate. Samples were sonicated, then incubated at 37 °C for 45 min. A solution of 20 mM IAA in 8 M urea/100 mM ammonium bicarbonate was then added.

Following an incubation for alkylation, the urea concentration was adjusted to 6 M by adding ammonium bicarbonate. Lysyl endopeptidase (Wako Chemicals, Richmond, VA) was added in a ratio of 100:1 to facilitate protein digestion for 6 hours at 37 °C. Samples were then diluted to

1.6 M urea and a second digestion step was carried out with trypsin at a ratio of 50:1 overnight at 37 °C. Digestion was terminated by adding TFA.

Digested protein peptides were extracted and puriﬁed using Waters Oasis HLB columns (Wa- ters Corporation, Milford, MA) according to the manufacturers instructions. Eluted peptides were dried in a Speed Vac SPD1010 (Thermo Scientiﬁc, Hudson, NH) and reconstituted in

3% acetonitrile with 0.1% formic acid. The peptide samples were centrifuged and half of the supernatant was collected and supplemented with synthetic iRT standard solutions (Biognosys

AG, Cambridge, MA) prior to LC-MS/MS analysis.

Digested protein peptides were analyzed on a TripleTOF 5600+ mass spectrometer (AB Sciex,

Framingham, MA) coupled with an Eksigent 2D plus LC system (Eksigent Technologies,

Dublin, CA). LC separation was performed via a trap-elute conﬁguration including a trapping column (ChromXP C18-CL, 120 A,˚ 150 x 0.3 mm, 5 µm, Eksigent Technologies, Dublin,

CA). The mobile phased consisted of water with 0.1% formic acid (phase A) and acetonitrile

with 0.1% formic acid (phase B) (Avantor, Center Valley, PA). A total of 6 µg of protein was Chapter 5 105 injected for analysis. Peptides were trapped and cleaned on the trapping column with phase A delivered at a ﬂow rate of 10 L/min for 3 min before being separated on the analytical column with a gradient elution at a ﬂow rate of 5 L/min. The gradient time program was set as follows for phase B: 0 to 68 min: 3% to 30%, 68 to 73 min: 30% to 40%, 73 to 75 min: 40% to 80%,

75 min to 78 min: 80%, 78 to 79 min: 5% to 3%, and ﬁnally 79 to 90 min at 3% for column equilibration. The TripleTOF instrument was operated in a positive ion mode with an ion spray voltage ﬂoating at 5500 V, ion source gas one at 28 psi, ion source gas two at 16 psi, curtain gas at 25 psi, and source temperature at 280 °C.

To generate the reference spectral library for SWATH (Sequential Window Acquisition of all

THeoretical mass spectra) data analysis, Information-Dependent Acquisition (IDA) was performed on three pooled mouse liver microsomes samples of equal amounts. The IDA method was set up with a 250 ms TOF-MS scan from 400 to 1250 Da, followed by an MS/MS scan in high sensitivity mode from 100 to 1500 Da (50 ms accumulation time, 10 ppm mass toler- ance, charge state from +2 to +5, rolling collision energy and dynamic accumulation) of the top 30 precursor ions from the TOF-MS scan. The IDA data from the pooled mice liver microsomes were searched by MaxQuant (version 1.5.3.30, Max Planck Institute of Biochemistry,

Germany). The mouse FASTA database with 16,911 protein entries downloaded from Uniport on 8/29/2016 was used at the reference sequences for the search. Tryspin/P was used as the protease. Peptide length was between 7 and 25 residues with up to two missed cleavage sites allowed. Carbamidomethyl (C) was set as a fixed modification. A false discovery rate of 0.01 was used as the cutoff for both peptide and protein identification. Chapter 5 106

All microsomes samples were analyzing using a SWATH method comprised of a 250 ms TOF-

MS scan from 400 to 1250 Da, followed by MS/MS scans from 100 to 1500 Da performed

on all precursors in a cyclic manner. The accumulation time was 50 ms per isolation window,

resulting in a total cycle time of 2.8 s. The spectral alignment and target data extraction of the

SWATH data were performed on Skyline-daily (version 3.7.1.11271, University of Washington,

Seattle, WA) with the reference spectral library generated from the IDA searches. The isolation

scheme in Skyline-daily was set as SWATH (VW 100). The MS1 and MS/MS ﬁltering were

both set as TOF mass analyzer with the resolution power of 30,000 and 15,000, respectively.

The retention time prediction was based on the auto-calculate regression implemented in the

iRT calculator. Proper peak selection was checked manually with the automated assistance

of Skyline-daily. The surrogate peptides used for the quantiﬁcation of Cyp210, 2c29, 3a11,

and 3a16 are listed in Appendix A. These peptides were selected based on the uniqueness and

chromatographic performance. The peak areas of the top 3 to 5 fragment ions were summed up

and normalized to the internal standard BSA. The BSA-normalized peak area of each peptide

was further divided by the average of the 30 samples to calculate the relative abundance of the

peptide. The average of relative abundance of all surrogate peptides of a protein was used to

determine the relative abundance of the protein.

Evaluation of Enzymatic Activities. Liver tissues were homogenized in ice-cold mM phosphate buffered and S9 fractions were obtained by centrifugation. Protein concentrations were determined using a Modiﬁed Lowry Protein Assay Kit (Thermo Fisher Scientiﬁc, Rockford,

IL). Pentoxyresoruﬁn O-dealkylation, paclitaxel 6-hydroxylation, and midazolam 1-hydroxylation were used as probes for Cyp2b10, 2c29, and 3a11 activities, respectively [120]. Incbuations Chapter 5 107 were carried out in 1X phosphate-buffered saline (pH 7.4), containing 0.1 mg of mouse liver S9 protein and 30 µM substrate for a ﬁnal volume of 95 µL. The reactions were initiated by adding

5 µL of 20 mM NADPH and continued for 30 min for pentoxyresoruﬁn and paclitaxel or 10 min for midazolam. Incubations were terminated by adding 100 µL of ice-cold acetonitrile.

After centrifugation at 15,000 rpm, a 1.0 µL aliquot was injected into a Waters Synapt G2-

S QTOFMS system (Milford, MA) for metabolic quantiﬁcation analysis. Chromatographic separation of metabolites was performed on an Acquity UPLC BEH C18 column (2.1 x 50 mm,

1.7 µm, Waters). The mobile phase A (MPA) consisted of 0.1% formic acid in water and mobile phase B (MPB) was 0.1% formic acid in acetonitrile. The gradient for aqueous extraction began at 5% MPB and held for 0.5 min, followed by 5 min linear gradient to 95% MPB, held for 2 min, and decreased 5% MPB for column equilibration. The ﬂow rate of the mobile phase was 0.50 mL/min at the column temperature was maintained at 50 °C. The G2-S QTOFMS system was operated in resolution mode (resolution ∼ 20, 000) with electrospray ionization. The source and desolvation temperatures were set at 150 and 500 °C, respectively. Nitrogen was applied as the cone gas (50 L/hr) and desolvation gas (800 L/hr). The capillary and cone voltages were set at 0.8 kV and 40 V. The data were acquired in positive ionization mode. QTOFMS was calibrated with sodium formate and monitored by the intermittent injection of lock mass leucine encephalin (m/z = 556.2771) in real time. Quanlynx software (Waters, Milford, MA) was used for the quantiﬁcation of the concentration of metabolites.

Statistical Analysis. Data are presented as the mean ± S.E.M. Statistical analyses were performed using GraphPad Prism 7 software (GraphPad Software, Inc., La Jolla, CA). Compar- isons between groups were done using Students t-test and discovery was determined using the Chapter 5 108

two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli with Q = 1%. A p- value of less than 0.05 was considered signiﬁcant.

5.3 Results

5.3.1 Long-term effects of neonatal phenytoin exposure on the basal ex-

pression of cytochrome P450s.

At the age of 5 days following birth, both male and female mice were treated with a single dose of either vehicle control (PBS) or 100 mg/kg phenytoin. These mice were allowed to grow to speciﬁed ages (day 10, 15, 20, 30, or 60), at which point they were then sacriﬁced and livers were harvested to investigate the long-term effects on CYP expression and activity. For RT-

PCR analyses, mice that were sacriﬁced before reaching the age of day 60 were not separated based on sex, so both male and female samples were included for fold-change calculations.

Analyses performed on mice at the age of day 60 were executed separately for each sex, due to mature hormonal inﬂuences on CYP expression. For protein quantiﬁcation and enzymatic activity assays, only male mice were utilized to ensure as little variability as possible.

The expressions and activities of three CYPs (Cyp2b10, Cyp2c29, and Cyp3a11) that contribute significantly to murine drug metabolism and known to be induced by phenytoin treatment were evaluated. Following treatment with phenytoin at age day 5, Cyp2b10 mRNA remained up- regulated by roughly 25-fold compared to control mice at age day 10 (Figure 5.1A). At ages day 15, 20, and 30, however, Cyp2b10 mRNA expression was not statistically significant from Chapter 5 109 that of control mice. At age day 60, Cyp2b10 mRNA was up-regulated approximately 3-fold in male mice that received phenytoin at age day 5, however there was no statistically significant differences in Cyp2b10 expression in adult females. To determine whether the significant increase in Cyp2b10 mRNA expression in males translated into significant differences in protein expression, Cyp2b10 protein was quantified in male livers at the age of day 60 by SWATH-

MS method (n = 5). No statistically significant differences were found between adult control and treated mice (Figure 5.1B). These results were confirmed by enzymatic activity assays in adult male mice (n = 5). Again, there was no statistically significant difference in the rate of pentoxyresorufin O-depentylation of microsomes isolated from control and treated male livers (Figure 5.1C). This indicates the slight up-regulation of Cyp2b10 mRNA at the adult age following neonatal phenytoin treatment probably has no consequences on the function of the

Cyp2b10 protein in adulthood.

The expression of Cyp2c29 was also examined following phenytoin exposure at age day 5. At ages day 20 and day 30, Cyp2c29 mRNA appeared to be slightly significantly up-regulated compared to age-matched control mice, however any significant mRNA over-expression dis- appears at age day 60 in both males and females (Figure 5.2A). Quantification of Cyp2c29 protein in adult males livers (n = 5) unexpectedly showed significantly down-regulated expression in phenytoin-treated mice (Figure 5.2B). The biological relevance of this observed down-regulation of Cyp2c29 protein could not be validated with enzymatic activity analyses as

Cyp2c29 activity was too low in prepared microsomes.

The long-term effects on the expression and activity of Cyp3a11 were ﬁnally examined in response to neonatal phenytoin exposure. Drug treatment did not appear to signiﬁcantly impact Chapter 5 110

FIGURE 5.1: Expression and activity of Cyp2b10 throughout postnatal development following phenytoin exposure at neonatal age. (A) Results of RT-PCR for Cyp2b10 expression at different ages following treatment with 100 mg/kg phenytoin (PHY) at age day 5 after birth. Fold-changes were calculated against age-matched mice that received PBS treatment on day 5. Males and females were evaluated collectively at ages before day 60 (left), while males and females were evaluated separately at age day 60 (right). (B) Quantiﬁcation of Cyp2b10 protein at age day 60 following day 5 treatment with either PBS or 100 mg/kg phenytoin. (C) Rates of pentoxyresoruﬁn O-depentylation at age day 60 following day 5 treatment with either PBS or PHY. *p < 0.05 **p < 0.01. Chapter 5 111

FIGURE 5.2: Expression of Cyp2c29 throughout postnatal development following phenytoin exposure at the neonatal age.(A) Results of RT-PCR for Cyp2c29 gene expression at different ages following treatment with 100 mg/kg phenytoin (PHY) at age day 5 after birth. Fold-changes were calculated against age-matched mice that received PBS treatment on day 5. Males and females were evaluated collectively at ages before day 60 (left), while males and females were evaluated separately at age day 60 (right). (B) Quantiﬁcation of Cyp2c29 protein at age day 60 following day 5 treatment with either PBS or PHY. *p < 0.05 the expression of Cyp3a11 mRNA at any age evaluated, in neither males nor females (Figure

5.3A). This observation was also verified by protein quantification method (Figure 5.3B) and midazolam 1'-hydroxylation enzymatic activity (Figure 5.3C), both of which showed no statistically significant up-regulation in treated adult male mice compared to control (n = 5 for both assays). However, Cyp3a11 protein and the enzymatic rate of midazolam metabolism do Chapter 5 112

FIGURE 5.3: Expression and activity of Cyp3a11 throughotu postnatal development following phenytoin exposure at the neonatal age. (A) Results of RT-PCR for Cyp3a11 gene expression at different ages following treatment with 100 mg/kg phenytoin (PHY) at age day 5. Fold-changes were calculated against age-matched mice that received PBS treatment on day 5. Males and females were evaluated collectively before age day 60 (left), while males and females were evaluated separately at age day 60 (right). (B) Quantification of Cyp3a11 protein at age day 60 following day 5 treatment with either PBS or PHY. (C) Rates of midazolam 1’-hydroxylation at age day 60 following day 5 treatment with either PBS or PHY. *p < 0.05 appear to be slightly down-regulated in treated mice, however not significantly, which is interesting due to the fact that Cyp2c29 protein is also significantly down-regulated in adults. In all

CYP enzymes evaluated, phenytoin treatment at the neonatal age does increase the variability represented by the standard error of their expression at later ages compared to age-matched control mice. Chapter 5 113

5.4 Discussion

In this study, we aimed to examine the possibility of any permanent effects on CYP expression and activity in adulthood resulting from neonatal exposure to a single dose of phenytoin.

We hypothesized that, like phenobarbital, administration of phenytoin to mice at age day 5 would cause persistent up-regulation of the drug-metabolizing enzymes Cyp2b10, Cyp2c29, and Cyp3a11 in adult mice. However, this was not the case. No signiﬁcant up-regulation of

Cyp2c29 nor Cyp3a11 mRNA was observed at the adult age of day 60, and protein expression appeared to be even down-regulated, if anything. Cyp2b10 mRNA was signiﬁcantly up- regulated at age day 60 only in males, however there were no increases apparent in either protein or enzymatic activity measurements. It is also interesting that both Cyp3a11 and Cyp2c29 protein were slightly down-regulated at age day 60, even though the decrease is only signiﬁcant for

Cyp2c29. There was no signiﬁcant down-regulation of Cyp2c29 mRNA, suggesting possible enzyme inhibition perhaps at the protein, or post-translational, level. This mechanism will have to be further investigated as well, as there are no studies that show decreased CYP transcription or protein inhibition in response to phenytoin treatment.

We also measured CYP expression at various ages during postnatal development following day

5 phenytoin treatment. There was no mRNA induction greater than 3-fold found in any of the CYPs at any postnatal ages tested, except for Cyp2b10 expression at age day 10, which was found to be induced by an average of approximately 25-fold. The mechanism leading to a ﬁve-day duration of Cyp2b10 induction will need to be further investigated, as the half-life of phenytoin in mice is only around 16 hours, although this might not hold true in neonates

[190]. Cyp2c29 was observed to be at least slightly induced at ages day 20 and day 30 at the Chapter 5 114 mRNA level, however whether the induction would translate into meaningful up-regulation of enzymatic activity is left to be questioned.

Additional studies will be necessary to further determine why neonatal exposure to phenobarbital produces overexpression of Cyp2b10, 2c29, and 3a11 mRNA and protein, with conse- quential up-regulation of activity, in adulthood, while neonatal exposure to phenytoin does not.

Phenobarbital up-regulates the transcription of CYPs through its indirect action on the nuclear receptor CAR [191]. Previous work has shown that CAR activation using TCPOBOP, a murine- specific CAR ligand, at the neonatal age also produces permanent alterations to CYP expression and function in adulthood, which altered the efficacy of zoxazolamine adminstered later in life [86]. Phenobarbital exposure at the neonatal age also caused a decrease in the efficacy of omeprazole adminstered to adult mice, suggesting its role in causing a non-conventional type of

DDI [159]. Due to the ability of phenytoin to also activate CAR, we hypothesized that neonatal exposure to this drug would also cause a similar permanent alteration to CYP expression; however our hypothesis was incorrect. Unlike TCPOBOP, both phenobarbital and phenytoin produce CAR activation indirectly. Whereas they are both classiﬁed as AEDs, these chemicals each have distinct mechanisms of action for treating seizures. It is plausible that phenobarbital and phenytoin produce indirect CAR activation via separate mechanisms or pathways. A pathway further upstream of CAR activation may be what is permanently affected following neonatal phenobarbital treatment that is not impacted by phenytoin treatment.

Phenobarbital is known to indirectly activate CAR via its binding and antagonism of EGFR

[172]. The mechanism by which phenytoin activates CAR is not as clearly understood. Human

CAR has been shown to be necessary to induce CYP2B6 in response to phenytoin treatment in Chapter 5 115 liver cells [84], and that murine CAR is necessary for induction of CYp2b10, 2c29, and 2c37 in mice [192, 193]. Beyond the fact that CAR is necessary to produce the induction of CYPs in response to phenytoin, nothing else is known about how exactly phenytoin causes CAR activation. However, it is also possible that CAR is not responsible for producing permanent effects in response to neonatal phenobarbital exposure. Phenobarbital is also known to activate several other hepatic nuclear receptors, including PXR and HNF4a, which all have roles in CYP regulation [194], and perhaps these are the mechanisms producing permanent alterations to

CYP expression. However, another study administered the direct PXR ligand PCN to neonatal mice and did not observe as drastic induction of drug-metabolizing CYPs as our previous study has shown [71].

Further studies are necessary to determine exactly how phenytoin activates CAR, and how this mechanism may differ from phenobarbital. It is also possible that the window of sensitivity for neonatal programming may differ depending on the drug, and perhaps day 5 was not the optimal age to treat mice with phenytoin to cause permanent alterations to CYP expression. Ex- periments that expose mice to phenytoin at younger or older ages during postnatal development will be critical. Additionally, only one dose of phenytoin was investigated in this particular study. Perhaps multiple-day dosing of phenytoin is necessary to cause changes to CYP expression that exist into adulthood. It is also possible our chosen dose of 100 mg/kg bodyweight was not sufﬁcient to produce permanent effects, as our study with phenobarbital demonstrated that the phenomenon is dose-dependent. However, 100 mg/kg was the highest dose we were able to administer to neonatal mice without lethal effects. A dose of 100 mg/kg in mice is equivalent to a human dose of 8 mg/kg, according to the FDA Guidance for Industry for species dose Chapter 5 116 extrapolation. This is representative of a maintenance dose of phenytoin in humans, and it is likely that newborns suffering from seizures would be treated with a dose closer to 20 mg/kg.

There may be species differences in tolerability to phenytoin, and administering a dose more representative of what is used in the clinic may produce permanent changes to CYP expression in mice.

Based on this data, we could suggest that phenytoin may be a better therapeutic choice for treating neonatal seizures as it does not appear to permanently affect CYP expression at this point. This could have a major impact in the clinic, as phenobarbital is currently the most commonly used drug for treating neonatal seizures. Further animal studies and clinical human testing will need to be performed in the future to validate this suggestion, however this study could provide a basis and a proof-of-concept for those studies. Chapter 6

Short-term Effects of Phenytoin Exposure on Cytochrome P450 Expression

Throughout Postnatal Development

6.1 Introduction

The postnatal ontogeny of CYPs during human development has been established and corresponds well to the postnatal ontogeny of orthologous CYP genes in rodent models [60, 141].

However, the induction of CYPs in response to drug treatment has not been thoroughly studied in pediatric populations at different stages of early development. It is largely unknown to what extent CYP enzymatic activity can be induced in neonates and children in comparison to adults. Whether the capacity for enzymatic induction varies at different stages during postnatal

117 Chapter 6 118

liver maturation is also not completely understood. Understanding CYP induction in pediatric

populations is critical to prevent DDIs and ADRs in infants and children. Many young patients

are exposed to multiple medications at once, especially those that are hospitalized or critically

ill [15]. However, DDIs in the context of pediatric patients are not as well understood as DDIs

in adults. This is mainly due to the lack of appropriate in vitro models and the fact that clinical trials do not usually include infants and children [195]. Neonatal patients appear to be particularly susceptible to ADRs due to diminished clearance capabilities during the ﬁrst two years of life [196]. It is well-established that the neonatal liver is not fully developed at birth and undergoes a period of postnatal maturation, during which metabolic function is not comparable to that of adults. Drug metabolism by CYPs, in particular, varies greatly during the ﬁrst two years after birth [197]. Most drug-metabolizing CYPs, including CYp3A4, 2C9, and 2C19, are expressed at low levels during gestation, at birth, and during early postnatal life, and do not reach expression levels comparable to those of adults until at least two years of age [62, 65].

This causes an attenuated capacity for the metabolism of many drugs and is implicated in numerous ADRS in infants [198]. Other enzymes, such as CYP3A7, are only expressed in the fetal and early postnatal liver and become undetectable once the liver is fully matured [199].

How induction in response to xenobiotics corresponds to the drug-metabolizing CYP ontogenic expression patterns is unknown.

The goal of this study was to examine the capacity for the postnatal liver to induce drug- metabolizing CYP enzymes in response to inducer treatment at different ages spanning neonatal, infant, child, and adolescent stages in mouse. The short-term effects of exposure to a single dose of phenytoin were investigated at multiple levels of CYP expression, including utilizing Chapter 6 119 the novel quantitative proteomics technology sequential window acquisition of all theoretical mass spectra (SWATH-MS) to determine CYP protein concentrations. Phenytoin was selected as a model inducer drug for this study because, like phenobarbital, it is an antiepileptic drug commonly used to treat neonatal seizures and is a potent inducer of drug-metabolizing CYPs in humans and rodents. We also wanted to investigate whether phenytoin induces CYP enzymes at young ages, because neonatal phenytoin treatment did not cause permanent alterations to

CYP expression as discussed in Chapter Five. Based on our previous work, we hypothesized that the neonatal liver is particularly sensitive to CYP induction by drug treatment in early life, and enzyme responses differ from those in adults.

6.2 Materials and Methods

Chemicals and Reagents. Phosphate buffered saline (PBS), 5,5-diphenylhydantoin salt (phenytoin, PHY), triﬂuoroacetic acid (TFA), formic acid, and acetonitrile were purchased from Sigma-

Aldrich (St. Louis, MO). Urea and dithiolthreitol (DTT) were obtained from Fisher Scientiﬁc

Co. (Pittsburgh, PA). Iodoacetamide (IAA) and ammonium bicarbonate were purchased from

Animals. C57Bl/6 mice (Jackson Laboratories, Bar Harbor, ME) were housed in accordance with the animal care guidelines outlined by the American Association for Animal Laboratory Chapter 6 120

Sciences and were bred under standard conditions in the Animal Resources Facility at the Uni-

versity of Connecticut. The use of these animals was approved by the University of Connecti-

cuts Institutional Animal Care and Use Committee. At 12 weeks of age, mice were set up

into breeding pairs. To study the short-term effects of drug treatment at different ages during

postnatal development, mice were administered a single dose of phenytoin (100 mg/kg) or PBS

vehicle control intraperitoneally at ages day 5, 10, 15, 20, 30, or 60 following birth. Twenty

fours after the treatment, mice were sacriﬁced and livers with gallbladders removed were col-

lected. Samples were snap frozen in liquid nitrogen then stored at -80 °C. For both control and

treatment groups, 4-6 mice were used for sufﬁcient statistical power and both male and female

mice were included.

Total RNA Extraction, Reverse Transcription and RT-PCR. Total RNAs were isolated from

whole livers using TRIzol reagent, and reverse transcription was performed using iScript Re-

verse Transcription Supermix as previously described. RT-PCR was performed using a CFX-96

thermocycler system with TaqMan gene expression assays for Cyp2b10, Cyp2c29, Cyp2d22,

Cyp2e1, Cyp3a11, Cyp3a16, and Gapdh as an endogenous control.

Protein Sample Preparation and LC-MS/MS-based Quantiﬁcation. The same procedures

used for protein quantiﬁcation in Chapter Five were used in this study.

Evaluation of Enzymatic Activities. The same procedures used for enzymatic activity quan- tiﬁcation in Chapter Five were used in this study. Chapter 6 121

Statistical Analysis. Data are presented as the mean ± S.E.M. Statistical analyses were per-

formed using GraphPad Prism 7 software (GraphPad Software, Inc., La Jolla, CA). Compar-

isons between groups were done using Students t-test and discovery was determined using the

two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli with Q = 1%. A p- value of less than 0.05 was considered signiﬁcant.

6.3 Results

To evaluate the impact of phenytoin exposure on the expression and function of drug-metabolizing

CYPs throughout postnatal maturation, we continued to use the C57Bl/6 mouse model. Short- term effects of drug treatment were evaluated by administering a single dose of phenytoin then collected livers 24 hours later at different time points after birth. The expression of hepatic murine Cyp2b10, 2c29, 2d22, 2e1, 3a11, and 3a16 isoforms were evaluated because of their orthologous equivalence to human CYP2B6, 2C9/19, 2D6, 2E1, 3A4, and 3A7, respectively.

The effect of drug treatment on each enzyme was measured at the mRNA level using RT-PCR, at the protein level using SWATH-MS quantiﬁcation, and at the enzymatic activity level using

MS/MS to conﬁrm function. Both male and female mice were evaluated collectively for data presented at ages younger than day 30, however male and female data were evaluated separately at age of day 60 due to hormonal inﬂuences on CYP expression. At each time point, mice that received drug treatment were compared against age-matched mice that received vehicle control. Chapter 6 122

6.3.1 Cyp2b10 induction and function following phenytoin exposure.

The extent of Cyp2b10 mRNA induction 24 hours following phenytoin administration was the greatest of all CYPs measured, presumably however this may be due to established low constitutive Cyp2b10 expression levels that inﬂate induction fold-changes. Mice at age of day 10 showed the greatest fold-change in mRNA expression when compared to control, age-matched mice, whereas fold-changes at other ages were more comparable to those observed in adult mice following phenytoin treatment. Adult males also had higher fold-changes than females at day 60, however adult male mice are known to have lower basal Cyp2b10 expression than females (Figure 6.1A). When comparing gene expression to Gapdh, which was unaffected by phenytoin treatment, Cyp2b10 mRNA was much less abundant than Gapdh in control mice but reached similar expression levels following drug exposure (Figure 6.1B). The surge in Cyp2b10 expression at age day 30 was also apparent and seemed to cause a corresponding peak in the inducibility of the enzyme.

Cyp2b10 protein was also quantiﬁed 24 hours after treatment in neonatal mice at the age of day 5 and in adult male mice at the age of day 60 (Figure 6.1C). Following vehicle treatment,

Cyp2b10 protein was nearly undetectable in both neonates and adults, however significant induction of protein was evident at both ages following phenytoin treatment. The inductive potential was far greater in adults than in neonates, however induced levels of Cyp2b10 in the neonate significantly exceeded basal levels in the adult. This data also correlated well with re- sorufin O-depentylation by microsomes prepared from neonatal and adult liver samples (Figure

6.1D), indicating functional consequences of induction by phenytoin at both ages tested. Chapter 6 123

FIGURE 6.1: Short-term effects of phenytoin treatment on the expression and activity of Cyp2b10 at different ages during postnatal development. Male and female mice were treated with either vehicle (PBS) or 100 mg/kg phenytoin (PHY) at different ages following birth and livers were collected 24 hours following treatment. Changes in Cyp2b10 gene expression were evaluated at the mRNA level using RT-PCR by calculating fold changes relative to mice that received vehicle control (A). Both males and females were evaluated collectively at ages day 30 and younger (left), but calculations were performed with males and females separate at age day 60 (right). Fold changes of mRNA were also calculated against the expression of Gapdh, with the dotted line representing Gapdh expression set to 1 (B). Cyp2b10 protein was quantiﬁed using an LC-MS/MS method in males following treatment at ages day 5 and day 60 (C). Enzymatic activity was evaluated by measuring the rate of resoruﬁn O-depentylation in microsomes prepared from male livers sample from mice treated at age day 5 and day 60 (D). *p < 0.05, **p < 0.01, ***p < 0.001. Chapter 6 124

6.3.2 Cyp2c29 induction and function following phenytoin exposure.

Induction of Cyp2c29 mRNA and protein expression 24 hours following phenytoin treatment was also examined at different ages. The amount of mRNA was signiﬁcantly greater in mice that received phenytoin compared to mice that received vehicle at all ages tested, with age of day 10 appearing to again by the most sensitive to enzyme induction (Figure 6.2A). When evaluating Cyp2c29 mRNA expression relative to GAPDH expression, Cyp2c29 mRNA was much less abundant in neonates at the age of day 5 in PBS treated control mice, but steadily increased to levels comparable to GAPDH at day 20 (Figure 6.2B). Induced levels of Cyp2c29 mRNA exceeded GAPDH expression by around 10-fold at all ages except day 5, where it remained minimally expressed even in the presence of phenytoin. This was also reﬂected in

Cyp2c29 protein quantiﬁed at age day 5 following 24 hour exposure duration to phenytoin

(Figure 6.2C). Phenytoin-mediated protein induction at age day 5 was significant, but minimal compared to induction at the adult age of day 60. Induced levels of Cyp2c29 protein in the neonate were significantly lower than even basal levels of the protein in adults. Enzymatic activity of Cyp2c29 could not be evaluated in these experiments due to insufficient metabolic activity detected from prepared microsomes.

6.3.3 Cyp3a11 induction and function following phenytoin exposure.

The short-term effects on Cyp3a11 mRNA induction following phenytoin treatment are evident across all ages studied during postnatal development. The greatest fold-change between Chapter 6 125

FIGURE 6.2: Short-term effects of phenytoin treatment on the expression of Cyp2c29 at different ages during postnatal developmentMale and female mice were treated with either vehicle (PBS) or 100 mg/kg phenytoin (PHY) at different ages following birth and livers were collected 24 hours following treatment. Changes in hepatic Cyp2c29 gene expression were evaluated at the mRNA level using RT-PCR by calculating fold changes relative to mice that received vehicle control (A). Both males and females were evaluated collectively at ages prior to day 60 (left), while males and females were evaluated separately at age day 60 (right). Fold changes of mRNA were also calculated against the expression of Gapdh, with the dotted line representing Gapdh gene expression set to 1 (B). Cyp2c29 protein was quantified using an LC- MS/MS method in males following treatment at ages day 5 and day 60 (C). *p < 0.05, **p < 0.01, ***p < 0.001. Chapter 6 126 phenytoin-treated mice and control mice occurred in neonates at age day 5, presumably because this is when the basal expression of Cyp3a11 is the lowest (Fig. ). As mice matured following birth, basal expression of Cyp3a11 mRNA increased and correlated with increased induced levels of mRNA with age (Fig. ). Maximal induction occurred at adult ages (day 30 and 60), which correlated with protein expression (Fig. ). Protein quantification revealed that induced levels of Cyp3a11 in the neonate at age day 5 were not statistically different from basal levels of Cyp3a11 protein in the adult. This was also confirmed by enzymatic activity analyses with midazolam in liver microsomes (Fig. ). At age of day 5, mice treated with phenytoin had similar rates of 1-hydroxymidazolam metabolite formation as adult mice treated with vehicle control, while adult mice treated with phenytoin produced the most metabolite overall.

6.3.4 Cyp3a16 repression following phenytoin exposure.

Cyp3a16 is considered to be the fetal isoform of the CYP3A subfamiliy of CYP enzymes.

Whereas it is highly expressed immediately after birth, its expression declines steadily throughout postnatal maturation and is undetectable in the adult liver. Phenytoin treatment also appeared to signiﬁcantly affect the expression of Cyp3a16 in the short-term at different ages during postnatal development. Twenty four hours following phenytoin administration, Cyp3a16 mRNA was signiﬁcantly suppressed in mice at all ages examined (Figure 6.4A). The fold- change in the decrease of expression was greatest at age day 20 and this correlated to the age at which Cyp3a16 mRNA because undetectable in phenytoin-treated mice (Figure 6.4B).

Cyp3a16 mRNA was not detected by RT-PCR in mice at the ages of day 30 nor day 60, so the Chapter 6 127

FIGURE 6.3: Short-term effects of phenytoin treatment on the expression and activity of Cyp3a11 at different ages during postnatal development. Male and female mice were treated with either vehicle (PBS) or 100 mg/kg phenytoin (PHY) at different ages following birth, and livers were collected 24 hours following treatment. Changes in hepatic Cyp3a11 gene expression were evaluated at the mRNA level using RT-PCR by calculating fold changes relative to mice that received vehicle control (A). Both males and females were evaluated collectively at ages prior to day 60 (left), while males and females were evaluated separately at age day 60 (right). Fold changes of Cyp3a11 mRNA were also calculated against the expression of Gapdh, with the dotted line representing Gapdh gene expression set to 1 (B). Cyp3a11 protein was quantiﬁed using an LC-MS/MS method in male livers following treatment at ages day 5 and day 60 (C). Cyp3a11 enzymatic activity was evaluated by measuring the rate of midazolam 1'-hydroxylation of microsome isolated from male livers treated at day 5 and day 60 of age (D). *p < 0.05, **p < 0.01, ***p < 0.001. Chapter 6 128

FIGURE 6.4: Repression of Cyp3a16 expression following phenytoin treatment at different ages during postnatal development. Male and female mice were treated with either vehicle (PBS) or 100 mg/kg phenytoin (PHY) at different ages following birth, and livers were collected 24 hours following treatment. Changes in hepatic Cyp3a16 gene expression were evaluated at the mRNA level using RT-PCR by calculating fold changes relative to mice that received vehicle control (A). Fold decreases of Cyp3a16 expression were also calculated against Gapdh, with the dotted line representing Gapdh gene expression set to 1 (B). Cyp3a16 mRNA was not detected at ages day 30 or day 60, so data is not shown. Cyp3a16 protein was quantiﬁed using an LC-MS/MS method in male livers following treatment at ages day 5 and day 60 (C). **p < 0.01, ***p < 0.001. data is not shown. The decrease in Cyp3a16 in response to phenytoin treatment was also con-

firmed by protein quantification (Figure 6.4C). At age of day 5, Cyp3a16 protein was decreased following drug exposure, however the different was not statistically significant. Cyp3a16 protein was not detectable in adults regardless of phenytoin administration. Chapter 6 129

FIGURE 6.5: Repression of H19 and Igf2 gene expression following phenytoin treatment during postnatal development. Male and female mice were treated with either vehicle (PBS) or 100 mg/kg phenytoin (PHY) at different ages following birth, then livers were collected 24 hours after treatment. Changes in gene expression of H19 and Igf2 were measured using RT-PCR and fold changes were calculated against PBS control mice.

6.3.5 Phenytoin represses H19/Igf2 expression during early postnatal life.

Due to the ability of phenytoin treatment to repress the expression of the fetal/neonatal-enriched enzyme Cyp3a16, we also wondered whether phenytoin was able to repress the expression of other liver genes that are enriched in early postnatal life. The insulin like growth-factor (Igf2) is an important fetal growth factor produced by the liver in high amounts during fetal and neonatal life, and becomes repressed in adulthood. The cis-neighboring gene, H19, is a long non-coding

RNA expressed highly in fetal and neonatal livers, and is repressed in adulthood unless liver regeneration or hepatocellular carcinoma is occurring [59]. Treatment with phenytoin at the ages of day 5, 10, and 15 repressed the gene expression of both H19 and Igf2 signiﬁcantly compared to age-matched control animals (Figure 6.5). At age day 20, there were no signiﬁcant differences in the expression of either gene between control and treated animals. No further ages are shown due to low gene expression. Chapter 6 130

6.4 Discussion

This study revealed the immediate effects on the expression and function of several drug- metabolizing CYP enzymes following exposure to the AED phenytoin during murine postnatal development. Short-term induction of Cyp2b10, 2c29, and 3a11 in response to phenytoin administration appeared to vary in extent based on age, sex, and particular CYP enzyme. Younger pups, such as those at ages of day 5 and day 10, had higher fold-changes relative to control for

CYP induction; however, these values are probably inﬂated due to the low basal expression of

CYPs in the neonatal liver. As basal expression of CYPs became greater with age, the extent of induction was less extreme. For this reason, we also chose to evaluate mRNA expression of each isoform in relation to Gapdh expression. Gapdh expression was not altered by drug treatment and remained relatively consistent throughout postnatal life. Analyzing gene expression in relation to Gapdh allowed the ontogeny of each CYP to be observed in basal and induced conditons at each age. Induction of each enzyme (with the exception of Cyp3a16), appeared to be proportional to the basal expression level of the enzyme; the higher the mRNA expression of a CYP at a certain age, the greater its capacity for induction. On the other hand, Cyp3a16 appeared to be repressed to greater extents at older ages in postnatal development, which correlated with decreasing basal expression.

For both Cyp2b10 and 2c29, day 10 appeared to be the age most sensitive to the inductive effects of phenytoin as evidenced by the greatest fold-changes in mRNA expression compared to control mice. Interestingly, this corresponds with our previous data that shows a surge of

Cyp2b10 and 2c29 expression beginning at the same age [141]. Both of these enzymes are primarily under the transcriptional control of CAR, while Cyp3a11 is known to be regulated Chapter 6 131 primarily by PXR. The expression of PXR and CAR actually appear to be the highest during early postnatal life in both mouse and infant humans [200]. The consequences of elevated expression of these nuclear receptors at the neonatal stage on the expression and regulation of CYPs are not understood. Perhaps, increased levels of CAR and/or PXR in the neonatal liver contribute to the increased inducibility of Cyp2b10 and Cyp2c29 observed at this stage of postnatal development.

The inducibility of other hepatic drug-processing genes utilizing direct activation of PXR and

CAR has also been shown to be age-dependent in mouse. Greater fold-changes in mRNA expression following xenobiotic treatment have been typically observed at the neonatal age. How- ever, our study shows that despite the apparent greater extent of mRNA induction at younger ages, protein expression and enzymatic activity following phenytoin treatment is still greater in adults than in neonates. The expression of Cyp2b10, 2c29, and 3a11 proteins is several magni- tudes higher in adults than in neonates following phenytoin treatment, suggesting the adult liver has a greater capacity for enzyme induction following drug treatment than the infant liver. In the case of Cyp2b10, protein expression in the neonate following phenytoin exposure exceeded that of adults in basal conditions following vehicle treatment. The expression of Cyp2c29 protein following phenytoin treatment in the neonate, however, did not reach adult expression levels. This data suggests that individual CYP enzymes are under unique transcriptional control during postnatal development and their regulation must be evaluated independently. Only liver samples from day 5 were used for protein and enzymatic activity analyses during early life, so additional ages will need to be examined for protein and functional consequences in the future. Chapter 6 132

The decrease in expression of Cyp3a16 in response to phenytoin treatment suggests drug treatment is causing overall differentiation of the neonatal liver, rather than strictly induction of drug-metabolizing enzymes. Cyp3a16, the murine equivalent of human CYP3A7, is expressed only in fetal and neonatal livers, and signiﬁes immature metabolic function. CYP3A7 has reduced metabolic capability compared to CYP3A4 (and CYP3A5), and does not contribute to midazolam metabolism, as was tested for Cyp3a11 enzymatic activity [201]. The expression of CYP3A7 in the developing liver is shown to be induced by glucocorticoids, including dexamethasone, cortisol, and cortisone, all of which activate the glucocorticoid receptor [202].

In contrast, our data shows that treatment with phenytoin, a CAR activator, has the ability to down-regulate Cyp3a16 in mice. Additional studies have shown that the expression of Cyp3a16 is correlated with changes to histone modiﬁcations. As the mRNA expression of Cyp3a16 in the liver decreases over the course of postnatal development, H3K27 methylation increases in the genes promoter, which signiﬁes transcriptional repression [67]. It is possible that phenytoin is modifying the chromatin, in our experiments, which up-regulates the mature adult CYP enzyme, while down-regulating fetal CYP enzymes. There may, however, be species differences in the regulation of the fetal forms of CYP3A enzymes.

The repression of H19 and Igf2 gene expression also indicate changes to epigenetic modiﬁca- tions as a result of phenytoin treatment. Both H19 and Igf2 are part of an imprinted genomic region, meaning that the genes are only transcribed from one allele in a parent-of-origin fashion.

The transcription of these genes is regulated by differential DNA methylation in the imprinting control region (ICR) of the domain. An unmethylated ICR allows the binding of CTCF, thus impacting the transcription of the genes [203]. Proper methylation of this region is critical for Chapter 6 133 proper fetal growth and organ development. Repression of both of these genes by phenytoin suggests alterations to DNA methylation and CTCF binding resulting from drug treatment. This could have body-wide impacts on organ growth and development.

The expression of CYPs in the human neonatal and infant liver follow distinct ontogenic patterns that may be specific to each individual enzyme. Previous work from our laboratory has defined the ontogenic profiles of all CYPs in the mouse liver from two days before birth through age day 60 after birth [141]. The expression pattern of the murine enzymes investigated in this study follow similar ontogenic trends to that of their human orthologous counterparts. It is therefore plausible that human CYP enzymatic induction might also vary in extent at different ages during early developmental periods following birth. Whereas most drugs currently have recommendations for pediatric or infant dosing that is separate from adult regimens, it might be necessary to further define pediatric dosing standards for specific ages based on the

CYP-mediated metabolism of a drug to best avoid ADRs and DDIs in young patients. Other studies have also highlighted the need for pediatric pharmacokinetics to be evaluated in age- and maturation-dependent models, rather than grouping all neonatal and infant guidelines into a single pediatric category [204].

Phenytoin has a narrow therapeutic index with a large propensity to produce ADRs due to patient interindividual variability. Adverse side effects of phenytoin are common and often include rashes, headache, behavioral changes, Steven-Johnson syndrome, cardiovascular col- lapse, and arrhythmias, especially with rapid intravenous administration, in both children and adults [205, 206]. Many pediatric patients with severe epilepsy are often treated with poly- therapy, and interactions between phenytoin and other AEDs and the resulting adverse effects Chapter 6 134 have been documented. For example, concurrent treatment with phenytoin and valproic acid produced phenytoin toxicity in an adolescent epileptic patient, despite adhering to pediatric dosing guidelines for the medications [207]. SNPs in CYP2C9 and CYP2C19 have also been shown to alter phenytoin metabolism in pediatric patients, leading to adverse reactions as well

[208]. However, genetic variation in CYP genes cannot explain the large amount of patients that experience serious adverse effects while taking phenytoin [209]. A better understanding of the additional mechanisms behind phenytoin-mediated CYP induction via CAR activation, especially throughout postnatal development, could help reduce the high-risk of adverse reactions that commonly accompany phenytoin treatment. Chapter 7

The Differential Effects of Phenytoin

Treatment on the Liver Proteome in

Neonates and Adults

7.1 Introduction

Phenytoin is one of the most common antiepileptic drugs (AEDs) administered to infants, children, and adults for the treatment of acute seizures and to control chronic epilepsy. It is preferred over benzodiazepines and barbiturates due to its reduced sedative effect [210]. However, phenytoin use is still associated with a high risk of adverse effects, and is the most common of all AEDs in causing ADRs [211]. This is due to its narrow therapeutic window and wide range of interindividual variation in patient responses. Minor to severe cutaneous reactions

135 Chapter 7 136

[212] are widely reported in adults. In pediatric patients, adverse reactions commonly reported include poor scholastic performance, gum hypertrophy, headache, behavioral problems, drowsi- ness [205]. There are also ample case reports of idiosyncratic adverse reactions to individual pediatric patients despite adhering to dosing guidelines. Additionally, phenytoin treatment is associated with hypersensitivty syndrome that includes fever, rash, lymphadenopathy, and hep- atitis, suggesting an immune response or potential liver damage [213]. Phenytoin is also involved in many DDIs due to its ability to induce drug-metabolizing enzymes in the liver [78].

Despite the extensive body of evidence that implicates many problems with phenytoin therapy, the mechanisms by which phenytoin may cause changes to gene expression and liver function are unknown.

Our previous study identified significant differences in the extent of CYP enzyme induction in the liver following phenytoin treatment between neonatal and adult mice. Therefore, we hypothesized that phenytoin may have differential effects on more proteins in the liver between neonates and adults. Additionally, we have previously shown that the AED phenobarbital has permanent effects on the expression of CYPs in the liver, as well the expression of many other genes involved in various liver functions. Because of this, we also hypothesized that neonatal phenytoin treatment would significantly affect the adult liver proteome. The following study utilized the novel mass-spectrometry (MS)-based proteomics method known as

SWATH-MS to quantify the expression of all proteins known to be expressed in the mouse liver.

SWATH-MS uses data-independent acquisition (DIA) to measure the concentration of unqiue peptide fragments that correspond to proteins within a known library. This enables unbiased and targeted data extractions that alleviates most constraints of current proteomic methods [214]. Chapter 7 137

Whereas SWATH-MS has been widely used to compare protein expression and alterations to the epigenome, particularly in hepatocellular carcinoma [215], this method has not yet been utilized to characterize the differences in drug response between neonates and adults in the liver.

7.2 Materials and Methods

Chemicals and Reagents. Phosphate buffered saline (PBS), 5,5-diphenylhydantoin salt (phenytoin, PHY), triﬂuoroacetic acid (TFA), formic acid, and acetonitrile were purchased from Sigma-

Aldrich (St. Louis, MO). Urea and dithiolthreitol (DTT) were obtained from Fisher Scientiﬁc

Co. (Pittsburgh, PA). Iodoacetamide (IAA) and ammonium bicarbonate were purchased from

Animals. C57Bl/6 mice (Jackson Laboratories, Bar Harbor, ME) were housed in accordance with the animal care guidelines outlined by the American Association for Animal Laboratory

Sciences and were bred under standard conditions in the Animal Resources Facility at the Uni- versity of Connecticut. The use of these animals was approved by the University of Connecti- cuts Institutional Animal Care and Use Committee. At 12 weeks of age, mice were set up into breeding pairs. To study the short-term effects of drug treatment at different ages during postnatal development, mice were administered a single dose of phenytoin (100 mg/kg) or PBS vehicle control intraperitoneally at age day 5 or age day 60 following birth. Twenty fours after Chapter 7 138 the treatment, mice were sacriﬁced and livers with gallbladders removed were collected. Sam- ples were snap frozen in liquid nitrogen then stored at -80 °C. To study the long-term effects of phenytoin treatment, a single dose of 100 mg/kg was administered intraperitoneally at age day

5, then livers were collected in the same manner at age day 60.

Protein Sample Preparation and LC-MS/MS-based Quantiﬁcation. The same procedures used for protein sample preparation and quantiﬁcation in Chapter Five were used in this study.

However, for this experiment, unique peptide sequences of all known liver proteins were used for quantiﬁcation.

Protein Ontology Analysis. Ontology was performed using the Database for Annotation,

Visualization, and Integrated Discovery (DAVID, v6.8). Biological processes and molecular function terms with a p-value < 0.05 and a FDR < 1 were considered statistically signiﬁcantly enriched.

7.3 Results

To investigate global changes to liver function following phenytoin treatment, we performed proteomics analysis rather than RNA-seq. This provided us with better insight into the functional consequences of drug treatment, especially since earlier differences in gene expression of CYPs based on RT-PCR did not always correlate with protein measurements, as seen in

Chapters Five and Six. We examined all proteins expressed in the liver of adult and neonatal mice 24 hours after treatment with either vehicle control or 100 mg/kg phenytoin utilizing a

SWATH-MS mass spectrometry method. This allowed us to analyze the short-term effects of Chapter 7 139

phenytoin treatment on the liver proteome. We also investigated long-term effects following

neonatal phenytoin exposure by treating mice at day 5 of age, then collecting livers for pro-

teome analysis at age day 60. Treatment schemes are illustrated in Figure 7.1A. Signiﬁcantly

differentially expressed proteins were identiﬁed based on a p-value <0.05 and a fold-change

≥ 0.5 or ≤ −0.5 in comparison to age-matched mice that received PBS vehicle control. Follow- ing phenytoin treatment, 98 proteins were differentially expressed in the adult liver compared to control. Of these, 57 were up-regulated and 41 were down-regulated. Following phenytoin treatment in the neonate, however, 252 proteins were differentially expressed. Of these, only

69 were up-regulated while 183 were down-regulated. Our ﬁndings are summarized in Figure

7.1B. In total, there were only 25 common differentially expressed proteins between neonate and adults following phenytoin treatment. The rest were speciﬁc to either adult or neonatal livers. Additionally, of the 252 proteins affected by phenytoin treatment in neonates in the short-term, only 12 of these proteins were commonly affected in the long-term by phenytoin treatment at day 5. The other 42 proteins signiﬁcantly differentially expressed in the long-term were unique and did not overlap with neither neonate or adult short-term groups (Figure 7.1C).

7.3.1 Short-term effects on the adult liver proteome following phenytoin

exposure.

First, we verified that drug-metabolizing CYPs were identified and up-regulated in response to 24-hour exposure to phenytoin in adults. Cyp2b10 was the most significantly up-regulated protein in general, with a fold-increase of 5.4 compared to control mice. Cyp3a11 and Cyp2c29 were also significantly increased, with fold-changes of 2.3 and 1.3, respectively. In total, 15 Chapter 7 140

A Neonate short-term PHY

Adult short-term PHY

Neonate long-term PHY

B Total number of Up-regulated Down-regulated Animal Group proteins proteins proteins

Adult short-term PHY 98 57 41 Neonate short-term PHY 252 69 183

Neonate long-term PHY 55 39 16

Short-term PHY, 73 - - adult-specific

Short-term PHY, neonate- 228 - - specific

Short-term PHY, common 25 - -

FIGURE 7.1: Summary of differential protein expression in the liver following phenytoin treatment in neonates and adults The groups of animal treatment schemes are illustrated in (A). Short-term responses to a single dose of 100 mg/kg phenytoin (PHY) in neonates were evaluated by collecting livers 24 hours following treatment and comparing the proteome to mice that received PBS control. Short-term responses to a single dose of PHY in adults were evaluated by collecting livers 24 hours following treatment and comparing the proteome to mice that received PBS control. Long-term effects of a single dose of PHY were evaluated by treating mice at day 5 of age, then collecting livers at day 60 of age, and comparing the proteome to mice that received PBS control at day 5. The number of differentially expressed proteins in PHY-treated groups compared to PBS-treated controls are summarized in (B). The overlap in speciﬁc differentially expressed proteins between each treatment group are represented in (C). Chapter 7 141

CYP proteins were differentially expressed. Most were up-regulated in response to phenytoin, and consisted of Cyp3a25, Cyp2c54, Cyp3a13, Cyp2c50, Cyp4a10, Cyp27a1, Cyp2b19,

Cyp8b1, and Cyp2c39 in addition. Three CYPs were down-regulated, including Cyp7b1,

Cyp51a1, and Cyp2c70. Other phase I enzymes such as Por, Ces1c, Ces1e, and Ces2a were also signiﬁcantly up-regulated in response to phenytoin treatment. Very few phase II enzymes or transporters were affected by phenytoin treatment.

To gain a better understanding of the overall immediate effect of phenytoin treatment on liver function, we also performed ontology analysis using the lists of differentially expressed proteins. The up-regulated proteins were signiﬁcantly enriched in biological processes such as oxidation-reduction pathways, epoxygenase P450 pathway, fatty acid metabolism, and mito- chondrial respiration and electron transport. Most of these proteins had the molecular function of oxidoreductase activity, aromatase activity, heme binding, and other types of steroid hormone metabolic activity. Down-regulated proteins were signiﬁcantly enriched in the biological processes of cholesterol biosynthesis, lipid biosynthesis, and receptor-mediated endocytosis.

The majority of these proteins functioned as molecular protein binders.

7.3.2 Short-term effects on the neonatal liver proteome following pheny-

toin exposure.

We observed strikingly different proteomic responses to phenytoin in neonatal mice compared to adults. Our ﬁrst step was to examine the expression of CYP proteins that were altered by drug treatment. Cy2b10 and Cyp3a11 were again found to be up-regulated, but only by 1.3- Chapter 7 142

A Number of Molecular Number of Biological process P-value FDR P-value FDR proteins Function proteins

Oxidation-reduction Oxidoreductase 28 3.34E-25 4.20E-22 24 1.06E-19 1.25E-16 pathway activity

Epoxygenase P450 Aromatase activity 10 1.35E-15 1.58E-12 7 2.36E-10 2.97E-07 pathway Heme binding 14 6.71E-15 7.86E-12 Fatty acid metabolic 6 8.40E-05 1.05E-01 process Monooxygenase 12 2.89E-14 3.42E-11 activity Transport 16 1.28E-04 1.61E-01 Arachidonic acid Lipid metabolic 8 3.23E-04 4.05E-01 epoxygenase 8 1.19E-10 1.40E-07 process activity

Respiratory electron Steroid 3 1.34E-03 7.76E-01 transport chain hydroxylase 8 3.29E-10 3.88E-07 activity Electron carrier 6 6.61E-07 7.80E-04 activity Flavin adenine dinucleotide 5 7.76E-05 0.09 binding

Number of Molecular Number of Biological process P-value FDR P-value FDR proteins function proteins

Cholesterol Protein 5 1.40E-06 0.0019 27 1.87E-02 19.70 biosynthetic process binding

Isoprenoid Actin filament 3 4.23E-04 0.56 3 3.38E-02 32.90 biosynthetic process binding

Lipid biosynthetic Poly(A) RNA 3 4.23E-04 0.56 7 3.46E-02 33.50 process binding

Receptor-mediated SnoRNA 4 0.0079 10.02 2 4.51E-02 41.42 endocytosis binding

FIGURE 7.2: Ontology analysis of differentially expressed proteins in response to short- term phenytoin treatment in adults. Protein ontology analyses were performed on both sig- niﬁcantly up-regulated (A) and down-regulated (B) proteins in response to phenytoin exposure in adult male mice. Enriched biological processes and molecular functions with a p < 0.05 and FDR < 1 are listed. In the case of down-regulated proteins, all enriched terms are listed regardless of FDR. Chapter 7 143 and 1.5-fold, respectively. Cyp2c29 was not found to be up-regulated in the neonatal liver.

A total of 14 CYP proteins were found to be differentially expressed, however the extent of

CYP induction was not as great as in the adult liver. There appeared to be no difference in the amount of CYPs either induced or repressed, with Cyp3a25, Cyp2d11, Cyp27a1 Cyp1a2,

Cyp3a13 being additionally up-regulated proteins, and Cyp2c40, Cyp3a16, Cyp4a10, Cyp4a12,

Cyp51a1, and Cyp2d26, being down-regulated. Additional phase I enzymes that exhibited differential expression in response to phenytoin included Aldh5a1, Fmo1, Pon3, and Aox1. No

Ces proteins were signiﬁcantly altered, and only a few phase II enzymes and transporters were affected (Sult2a2, Sult1a1, Gstt3, Ugt16, Slc27a2). The protein with the greatest fold-induction was tryptophan 2,3-dioxygenase (Tdo2), which is not involved in drug metabolism.

To gain a better understanding of the overall effect of phenytoin on proteins in the developing neonatal liver, we also performed ontology analyses on these differentially expressed proteins lists (Figure 7.3). The biological processes enriched in up-regulated proteins were similar to those of the adult samples and included oxidation-reduction, steroid metabolism, lipid metabolism, and drug metabolism biological processes. However, additional processes that were not identified in adults included ATP biosynthesis, cellular responses to glucagon, blood coagulation, hemostasis, and fibrinolysis were also identified. The up-regulated proteins had functions that were consistent with oxidoreductase activity, iron binding, and energy transport.

The majority of differentially expressed proteins in the neonatal samples were, however, down- regulated. The protein with the greatest fold-reduction was thymidine kinase 1 (Tk1), which was down-regulated by over 2-fold. The biological processes enriched for these proteins were Chapter 7 144 quite different from the down-regulated proteins in adults and consisted of chromatin silencing, nucleosome assembly, translation, RNA splicing, and mRNA processing (Figure 7.3B).

The processes of steroid biosynthesis, oxidation-reduction reactions, cholesterol biosynthesis, and lipid metabolism were also similarly enriched for repression. The molecular functions of down-regulated proteins in the neonate showed enrichment for poly(A) RNA binding, ribosomal structure, RNA binding, DNA binding, and oxidoreductase activity functions. Strikingly, many histone proteins (25 variants) were down-regulated in response to phenytoin neonates, belonging to histone 1 and histone 2 submfamilies, and were identiﬁed for roles in chromatin silencing and nucleosome assembly. This effect was absent in adults.

7.3.3 Long-term effects on the adult liver following phenytoin exposure

at the neonatal age.

Following neonatal exposure to phenytoin at age day 5, only 55 proteins were significantly altered in the adult liver at age day 60. Only 12 proteins that were up-regulated in the short- term by phenytoin were also significantly altered in these adult mice, and included apoptotic chromatin condenstion inducer (Acin1), fibrinogen (Fga, Fgb), keratins (Krt6, Krt16), several splicing factors (Srsfs), and small nuclear ribonuclear proteins (Snrps). The remainder of the proteins altered in adults following neonatal exposure were unique to this group. The most up-regulated protein was DExH-Box Helicase 9 (Dhx9), which had a 1.3-fold increase. Several other heterogeneous ribonuclear proteins were also up-regulated. The most down-regulated protein was glucosamine-6-phosphate deaminase 1 (Gnpda1), which had a 1.7-fold decrease. Chapter 7 145

A Number of Number of Biological process P-value FDR Molecular function P-value FDR proteins proteins Oxidation-reduction Oxidoreductase 21 1.74E-13 2.48E-10 20 6.84E-13 8.75E-10 process activity Steroid metabolic Iron ion binding 12 4.66E-10 5.96E-07 7 6.10E-07 8.68E-04 process Monooxygenase ATP biosynthetic 8 1.83E-07 2.35E-04 5 1.36E-06 1.94E-03 activity process Proton-transport ATP 5 1.93E-07 2.47E-04 Drug metabolic synthase activity 4 3.64E-05 5.19E-02 process Aromatase activity 6 2.47E-07 3.15E-04 Lipid metabolic process 10 4.15E-05 5.91E-02 Amino acid binding 5 5.05E-06 6.46E-03 Cellular response to 3 1.30E-04 1.85E-01 Electron carrier glucagon 5 5.05E-05 6.45E-02 activity Protein 5 1.54E-04 2.19E-01 homotetramerization Catalytic activity 10 8.55E-05 1.09E-01

Blood coagulation 5 2.51E-04 3.57E-01 ATPase activity 7 1.17E-04 1.50E-01

Hemostasis 4 6.30E-04 8.94E-01 Metal ion binding 27 1.54E-04 1.96E-01 Flavin adenine 5 1.68E-04 2.14E-01 dinucleotide binding B Number of Number of Biological process P-value FDR Molecular function P-value FDR proteins proteins

Chromatin silencing 16 1.46E-16 1.67E-13 Poly(A) RNA binding 51 8.37E-18 1.13E-14

Nucleosome Structural constituent 17 2.26E-14 3.43E-11 26 2.43E-16 3.00E-13 assembly of ribosome Oxidoreductase Translation 25 1.18E-11 1.79E-08 27 4.32E-09 5.81E-06 activity Steroid biosynthetic 12 5.75E-11 8.70E-08 Large ribosomal process 6 3.14E-07 4.22E-04 subunit rRNA binding Oxidation-reduction 18 4.92E-09 7.45E-06 process RNA binding 27 6.86E-07 9.24E-04 RNA splicing 17 8.67E-09 1.31E-05 U1 snRNP binding 5 1.04E-06 1.41E-03 Cholesterol 8 4.23E-08 6.39E-05 biosynthetic process DNA binding 44 2.58E-06 3.47E-03 mRNA processing 18 8.61E-08 1.30E-04 Lipid metabolic Enzyme binding 17 7.35E-06 9.90E-03 17 4.22E-05 6.39E-02 process Estradiol 17-beta- Negative regulation of 12 3.21E-03 4.75 dehydrogenase 4 3.68E-04 4.94E-01 cell proliferation activity Histone H3-K27 Monooxygenase 3 3.14E-03 4.64 7 1.38E-03 1.84 trimethylation activity Defense response to Structural molecule 10 1.44E-03 1.92 Gram-positive 7 5.24E-04 0.79 activity bacterium

FIGURE 7.3: Ontology analysis of differentially expressed proteins in response to short- term phenytoin treatment in neonates. Protein ontology analyses were performed on both signiﬁcantly up-regulated (A) and down-regulated (B) proteins in response to phenytoin exposure in neonatal mice. Enriched biological processes and molecular functions with p < 0.05 and FDR < 1 (in most cases) are listed. Chapter 7 146

Overall, the long-term neonatal treated group had less drastic fold-changes in protein expression compared to the other two groups.

Ontology analyses for the up-regulated proteins revealed biological processes such as RNA splicing, mRNA processing, and mRNA transport as being enriched (Figure 7.4A). This suggests possible permanent changes to transcriptional and translational regulation in the liver following neonatal phenytoin treatment. This corresponds to the enriched molecular functions of RNA binding, nucleotide binding, and structural molecule activity that were also identiﬁed.

Additionally, blood coagulation, platetel activation, fibrinolysis, and plasminogen activation were processes identified to be significantly altered. No terms relating to drug metabolism,

CYP activity, or energy metabolism were found to be enriched. Signiﬁcantly enriched biological processes of down-regulated proteins (Figure 7.4B) included nucleotide biosynthesis, cellular response to lipids, cellular response to testosterone, heat generation, and regulation of glucose metabolism. The molecular functions of these proteins that were enriched also included small molecule binding and insulin-activated receptor activity. No terms or functions relating to chromatin silencing, drug metabolism, or nucleosome assembly were identiﬁed, as in the short-term neonatal group.

7.4 Discussion

This study revealed important differences in liver protein responses to phenytoin treatment between neonatal and adult mice, and also identiﬁed many proteins that may be permanently altered in adult mice following neonatal exposure to the drug. The use of the novel proteomics Chapter 7 147

A Number of Number of Biological Process P-value FDR Molecular function P-value FDR proteins proteins RNA splicing 14 7.95E-16 9.77E-13 RNA binding 20 5.20E-16 6.22E-13

mRNA processing 15 1.00E-15 1.25E-12 Poly(A) RNA binding 22 1.41E-15 1.62E-12

mRNA splicing 7 9.81E-08 1.24E-04 Nucleotide binding 21 6.22E-10 6.97E-07 Structural molecule mRNA transport 5 3.06E-05 0.04 7 1.29E-05 0.014 activity Blood coagulation 3 4.06E-05 0.05 mRNA binding 6 1.53E-05 0.017 Cell adhesion molecule Platelet activation 4 6.85E-05 0.09 4 5.31E-04 0.59 binding Osteoblast differentiation 5 8.14E-05 0.10 Positive regulation of 3 8.50E-05 0.11 peptide hormone secretion Plasminogen activation 3 1.81E-04 0.23

Protein polymerization 3 3.13E-04 0.39

Fibrinolysis 3 4.21E-04 0.53 B Number of Number of Biological process P-value FDR Molecular function P-value FDR proteins proteins Nucleotide biosynthesis 3 3.29E-05 0.038 Small molecule binding 4 3.75E-06 3.70E-03 Ribose phosphate Cellular response to lipids 3 7.66E-05 0.088 3 1.34E-05 1.30E-02 diphosphokinase activity Cellular response to Insulin-activated receptor 3 4.82E-05 0.055 3 6.93E-05 6.84E-02 testosterone activity Heat generation 3 7.66E-05 0.088 Pheromone binding 4 1.56E-04 0.15 Positive regulation of 3 1.53E-04 0.18 glucose metabolism Locomotor rhythm 3 2.36E-04 0.27 Mitochondria 3 3.59E-04 0.41 morphogenesis

FIGURE 7.4: Ontology analysis of differentially expressed proteins in adult mice in response to long-term phenytoin treatment at the neonatal age. Protein ontology analyses were performed on both significantly up-regulated (A) and down-regulated (B) proteins identified in the livers of adult mice following exposure to phenytoin at the neonatal age. Enriched biological processes and molecular functions with p < 0.05 and FDR < 1 are listed. method, SWATH-MS, allowed us to identify significant differences among all of the proteins known to be expressed in the mouse liver between control and treatment groups. This is one of the first studies that has looked at global changes to liver protein expression following drug treatment in the mouse. This is also one of the only studies that has compared drug responses in the adult mouse to drug responses in the neonatal mouse. This data supports the argument Chapter 7 148 that drug responses vary considerably between pediatric and adult patients, and adult drug data cannot be used to extrapolate drug responses to infants and children.

Alterations to proteins in the liver of adult mice exposed to phenytoin for 24 hours were as expected and included proteins involved in oxidation-reduction pathways, CYP metabolism, cholesterol metabolism, and lipid metabolism. Phenytoin, as previously discussed, is well known to induce the expression of many drug-metabolizing enzymes in the liver, including

CYPs. Phenytoin has also been associated with changes in glucose and lipid metabolism, which is supported by our data [216]. Long-term phenytoin treatment has also been associated with changes in serum triglyceride and cholesterol levels [217]. Because phenytoin is known as an enzyme inducer, it is also interesting to note the large amount of proteins acts to down-regulate, which is an important consideration when predicting drug responses.

Alterations to proteins in the liver of neonatal mice exposed to phenytoin for 24 hours were much more surprising. Biological processes found to be enriched included oxidation-reduction, drug metabolism, and lipid metabolism, similar to adults, however ATP biosynthesis, steroid biosynthesis, blood coagulation, and cellular response to glucagon were also up-regulated, which was not observed in adult mice. Phenytoin has been shown to modulate estrogen receptor (ER) activity [218], as well as testosterone metabolism in the brain of rodents [219], which makes sense due to the capacity for CYP enzymes to also metabolize steroid hormones.

Phenytoin has also been shown to increase serum glucagon levels [220], which may also be related to the enriched ATP biosynthetic process our analyses identiﬁed due to altered energy demands within liver cells. Hemorrhage has also been loosely linked to phenytoin exposure in the newborn due to a vitamin K deﬁency that results from the up-regulation of liver enzymes Chapter 7 149

[221]. This could explain the fact that blood coagulation was also a signiﬁcantly altered process resulting from phenytoin in our data.

The most striking differences in the response to phenytoin in the neonate was in the large number of down-regulated proteins, however. According to ontology analysis, biological processes such as chromatin silencing, nucleosome assembly, translation, RNA splicing, steroid biosynthesis, lipid biosynthesis, and histone modification were significantly down-regulated. None of these terms were enriched in down-regulated proteins identified in our adult mice. This suggests that phenytoin has significant effect on transcriptional and translational regulatory processes in young mice only. This could also explain why a greater number of proteins were affected following short-term phenytoin treatment in neonates, since general epigenetic changes, such as chromatin silencing and mRNA processing, could affect the transcription and thus protein expression of many target genes. A down-regulation of chromatin silencing would suggest an increase in gene activation and transcription, and this could serve as a mechanism for enzyme induction in neonates. At this point, no studies have identified an association between phenytoin treatment and chromatin modifications. A down-regulation of nucleosome assembly also suggests diminished DNA synthesis and thus cell proliferation, which might affect hepatocyte growth and liver development during postnatal maturation. In utero exposure to phenytoin is known to cause a condition known as Fetal Hydantoin Syndrome, which is characterized by growth retardation and alterations to developmental pathways [222]. It is possible that phenytoin treatment might still significantly interfere with developmental processes during postnatal maturation as well.

Due to the propensity of phenobarbital to cause permanent changes to the liver transcriptome Chapter 7 150 following neonatal exposure, we wondered if phenytoin treatment at the neonatal age might also produce lasting effects on liver proteins. In adult mice that received phenytoin on day 5 of age, 55 proteins were found to be differentially expressed compared to control mice. Only 12 of these proteins were also affected by short-term phenytoin treatment in neonates, suggesting cellular mechanisms that occur later in life which alter the transcription or translation of genes programmed by neonatal phenytoin treatment. Interestingly, proteins that are up-regulated in adult mice include keratins, which are a classic marker for liver tumors and hepatic injury

[223]. This suggests that damage to the liver caused by neonatal phenytoin administration may persist to adulthood and increase susceptibility for liver disease. Blood coagulation, a biological processes identiﬁed as enriched in our mice, also plays a role in liver disease as the liver is responsible for clearing coagulation factors from the blood [224]. This also supports the idea that neonatal phenytoin treatment may increase the risk of liver disease or damage in later life.

It would be interesting to test susceptibility of adult mice to hepatotoxicants following neonatal phenytoin treatment and compare liver damage to control animals. There are also permanent changes to the expression of proteins involved in mRNA splicing, processing, and transport following neonatal phenytoin exposure, which could impact global transcription regulation.

Perhaps this would affect enzyme induction in adulthood following exposure to inducing drugs.

However, this study showed no changes to adult basal drug metabolic processes following neonatal phenytoin exposure. Lipid and glucose metabolism, however, were both observed to be potentially down-regulated in adult livers, as well as cellular responses to testosterone.

Disturbances in the regulation of sex hormones, particularly testosterone, have been identiﬁed in male epileptic patients taking AEDs [225], so it would be interesting to further explore the link between neonatal AED exposure and potential hormonal imbalances later in life. Chapter 7 151

Future studies will be necessary to determine whether the identified changes in protein expression following phenytoin exposure correspond to significant changes in liver function in both neonates and adults. However, protein expression is typically more highly correlated to enzyme activity, function, and biological consequences than mRNA expression [226]. Also, our study only measured changes to protein expression following a single dosing concentration and only at one timepoint. It would be interesting to see if there are any differences in liver protein profiles following exposure to phenytoin at higher or lower doses than 100 mg/kg. It would also be of value to measure protein expression following longer exposure, perhaps 48 hours, since it requires more time to observe significant changes in protein expression than mRNA expression. In light of this, the present study highlights the fact that drug response can vary drastically between neonates and adults, and the neonatal liver may be more susceptible to epigenetic changes resulting from exposure to inducing drugs. Additionally, while neonatal pheytoin exposure does not appear to signficantly affect drug-metabolizing proteins as phenobarbital dose, neonatal exposure to phenytoin may affect adult susceptibility to liver damage and disease. Chapter 8

Summary and Future Directions

The studies discussed in this dissertation highlight the significant differences between neonatal and adult responses to drug treatment, in particular the AEDs phenobarbital and phenytoin, especially in regards to CYP expression and hepatic drug metabolism. Additionally, this work also demonstrates the capability for neonatal drug exposure to affect gene and protein expression in the liver in adulthood, which could potentially affect drug efficacy and susceptibility to liver diseases later in life. Because these studies were completed in a mouse model, future studies will need to be completed in humans to verify our findings and determine inter-species differences in the responses to these drugs. In light of this, our studies provide a solid basis for the argument that adult drug response data cannot be directly extrapolated to predict drug responses in infants and children. There is a significant need for pharmacokinetic evaluation of specific drugs in infant and children populations, and to characterize CYP induction and drug metabolic pathways for all drug compounds at multiple stages of postnatal development.

152 Chapter 8 153

Our work has demonstrated that neonatal phenobarbital treatment up-regulates the expression of many drug-metabolizing enzymes, CYPs, and transporters in the adult mouse liver, which can then affect drug efﬁcacy later in life. Many other genes involved in hepatic processes such as steroid metabolism, lipid metabolism, proteolysis, and circadian rhythm are also signiﬁ- cantly impacted. We hypothesized that this phenomenon was due to neonatal programming of the CAR nuclear receptor, based on other work that described the necessity of CAR to produce permanent changes to CYP expression following neonatal treatment with the CAR agonist

TCPOBOP [86]. Future studies with Car-null mice and neonatal phenobarbital exposure would be imperative to prove this speculation. Regardless, we subsequently predicted that neonatal phenytoin exposure would also induce permanent up-regulation to CYP expression and drug metabolism in adulthood due to the ability of the drug to activate CAR as well. However, this did not occur. This suggests that CAR activation may not be the key factor enabling permanent induction of CYP enzymes following drug treatment at the neonatal age. Whereas the mechanism for the indirect activation of CAR by phenobarbital has been elucidated [172], the mechanism of CAR activation by phenytoin is still unknown. There may be fundamental differences in the way the two AEDs induce CYP expression, which warrants further examination into the involvement of additional nuclear receptors and/or hormonal regulation. In addition, it is also likely that different drugs may have speciﬁc windows of sensitivity and require higher/chronic doses in order to program gene expression in early postnatal life.

Neonatal exposure to phenytoin possibly induces differentiation of hepatocytes during postnatal liver maturation. In addition to up-regulating several CYPs and drug-metabolizing enzymes, phenytoin administration to neonatal mice also increases the expression of proteins involved in Chapter 8 154 mature liver functions such as lipid metabolism, steroid metabolism, and energy homeostasis.

Phenytoin treatment also repressed the expression of fetal-speciﬁc genes such as Cyp3a16,

H19, and Igf2. Chromatin silencing via epigenetic modiﬁcations such as histone methylation are known to be critical to maintaining proper gene expression during postnatal development, and phenytoin treatment was also shown to down-regulate these processes in the neonatal liver.

Future studies on the impact of inducer treatment in early life on the proper differentiation and maturation of the liver could provide valuable information on mechanisms underlying postnatal liver development. Previous worked from our laboratory has shown that the loss of the nuclear receptor Fxr in mice signiﬁcantly affects the maturation of liver gene expression throughout postnatal development [227]. It would be interesting to understand whether over-expressing certain nuclear receptors, or the constant activation of nuclear receptors, in the liver would affect the normal course of organ development and maturation.

Finally, consequences on functional processes and responses in the adult liver following neonatal drug exposure should be examined. It is possible that neonatal treatment with phenytoin programs the liver to have exaggerated responses to chemical challenges later in life, as demonstrated by Dr. Cherly Lyn Walker [129]. Adult re-challenges with other clinically utilized drugs could be used to measure CYP induction, efﬁcacy, and toxicity in relation to control animals.

Neonatal phenytoin exposure also caused up-regulation of various protein markers for liver disease in adulthood. It would be interesting to investigate whether these mice are more susceptible to liver damage by xenobiotics (e.g. acetaminophen, carbon tetrachloride) or to developing liver diseases (e.g. NAFLD, ﬁbrosis, hepatocellular carcinoma). Phenytoin also affects energy Chapter 155 metabolism in additional to drug metabolism in the liver. It would also be valuable to better understand if neonatal phenytoin exposure has permanent effects on lipid or glucose metabolism in adulthood and risk for developing obesity or diabetes. References

[1] B. B. Spear, M. Heath-Chiozzi, and J. Huff. Clinical applications of pharmacogenetics. Trends in molecular medicine, 7:201204, 2001. doi: 10.5772/1277.

[2] Jason Lazarou, Bruce H. Pomeranz, and Paul N. Corey. Incidence of adverse drug reactions in hospitalized patients. Jama, 279(15):1200, 1998. doi: 10.1001/jama.279.15. 1200.

[3] Eva Fernandez, Raul Perez, Alfredo Hernandez, Pilar Tejada, Marta Arteta, and Jose Ramos. Factors and mechanisms for pharmacokinetic differences between pediatric population and adults. Pharmaceutics, 3(1):5372, Jul 2011. doi: 10.3390/ pharmaceutics3010053.

[4] N. Nwobodo Ndubuisi and A. Obu Herbert. Overview of clinical pharmacokinetics in pediatrics: Possible implications in therapy. Biomedical Pharmacology Journal, 7(1): 169174, 2014. doi: 10.13005/bpj/468.

[5] Ettore Napoleone. Children and adrs (adverse drug reactions). Italian Journal of Pedi- atrics, 36(1):4, 2010. doi: 10.1186/1824-7288-36-4.

[6] R Priyadharsini, C Adithan, Firojkumar Sahoo, A Surendiran, and S Sreenivasan. A study of adverse drug reactions in pediatric patients. Journal of Pharmacology and Pharmacotherapeutics, 2(4):277, 2011. doi: 10.4103/0976-500x.85957.

[7] Rebecca Mary Diane Smyth, Elizabeth Gargon, Jamie Kirkham, Lynne Cresswell, Su Golder, Rosalind Smyth, and Paula Williamson. Adverse drug reactions in childrena systematic review. PLoS ONE, 7(3), May 2012. doi: 10.1371/journal.pone.0024061.

[8] J. Le, T. Nguyen, A. V. Law, and J. Hodding. Adverse drug reactions among children over a 10-year period. Pediatrics, 118(2):555562, Jan 2006. doi: 10.1542/peds. 2005-2429.

[9] A Mulhall, J De Louvois, and R Hurley. Chloramphenicol toxicity in neonates: its incidence and prevention. Bmj, 287(6403):14241427, Dec 1983. doi: 10.1136/bmj.287. 6403.1424.

[10] A Penna and N Buchanan. Paracetamol poisoning in children and hepatotoxicity. British Journal of Clinical Pharmacology, 32(2):143149, 1991. doi: 10.1111/j.1365-2125.1991. tb03873.x.

156 References 157

[11] Mara Espinosa-Bosch, Bernardo Santos-Ramos, Mara Victoria Gil-Navarro, Mara Do- lores Santos-Rubio, Roberto Marn-Gil, and Paloma Villacorta-Linaza. Prevalence of drug interactions in hospital healthcare. International Journal of Clinical Pharmacy, 34 (6):807817, Nov 2012. doi: 10.1007/s11096-012-9697-0.

[12] Supinya Dechanont, Sirada Maphanta, Bodin Butthum, and Chuenjid Kongkaew. Hos- pital admissions/visits associated with drug-drug interactions: a systematic review and meta-analysis. Pharmacoepidemiology and Drug Safety, 23(5):489497, Oct 2014. doi: 10.1002/pds.3592.

[13] Tai-Yin Wu, Min-Hua Jen, Alex Bottle, Mariam Molokhia, Paul Aylin, Derek Bell, and Azeem Majeed. Ten-year trends in hospital admissions for adverse drug reactions in england 19992009. Journal of the Royal Society of Medicine, 103(6):239250, 2010. doi: 10.1258/jrsm.2010.100113.

[14] David J. Greenblatt and Lisa L. Von Moltke. Interaction of warfarin with drugs, natural substances, and foods. The Journal of Clinical Pharmacology, 45(2):127132, 2005. doi: 10.1177/0091270004271404.

[15] Dingwei Dai, James A. Feinstein, Wynne Morrison, Athena F. Zuppa, and Chris Feudt- ner. Epidemiology of polypharmacy and potential drugdrug interactions among pediatric patients in icus of u.s. childrens hospitals*. Pediatric Critical Care Medicine, 17(5), 2016. doi: 10.1097/pcc.0000000000000684.

[16] J. Feinstein, D. Dai, W. Zhong, J. Freedman, and C. Feudtner. Potential drug-drug interactions in infant, child, and adolescent patients in childrens hospitals. Pediatrics, 135 (1), 2014. doi: 10.1542/peds.2014-2015.

[17] Hannah Katharine Batchelor and John Francis Marriott. Paediatric pharmacokinetics: key considerations. British Journal of Clinical Pharmacology, 79(3):395404, 2015. doi: 10.1111/bcp.12267.

[18] Tracy L. Sandritter, Matthew Mclaughlin, Michael Artman, and Jennifer Lowry. The interplay between pharmacokinetics and pharmacodynamics. Pediatrics in Review, 38 (5):195206, 2017. doi: 10.1542/pir.2016-0101.

[19] Elliot S. Vesell. Genetic and environmental factors causing variation in drug response. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 247(2): 241257, 1991. doi: 10.1016/0027-5107(91)90020-o.

[20] Keith Bailey. Physiological factors affecting drug toxicity. Regulatory Toxicology and Pharmacology, 3(4):389398, 1983. doi: 10.1016/0273-2300(83)90009-0.

[21] Cecilie Johannessen Landmark, Svein I. Johannessen, and Torbjrn Tomson. Host factors affecting antiepileptic drug deliverypharmacokinetic variability. Advanced Drug Deliv- ery Reviews, 64(10):896910, 2012. doi: 10.1016/j.addr.2011.10.003. References 158

[22] Mara Jos Gmez-Lechn, Jos Vicente Castell, and Mara Teresa Donato. Hepatocytes- the choice to investigate drug metabolism and toxicity in man: In vitro variability as a reﬂection of in vivo. Chemico-Biological Interactions, 168(1):3050, 2007. doi: 10.1016/j.cbi.2006.10.013.

[23] Susan M. Pond and Thomas N. Tozer. First-pass elimination. Clinical Pharmacokinetics, 9(1):125, 1984. doi: 10.2165/00003088-198409010-00001.

[24] O. A. Almazroo, M. K. Miah, and R. Venkataramanan. Drug metabolism in the liver. Clinics in Liver Disease, 21(1):120, Feb 2017.

[25] Grant R. Wilkinson. Drug metabolism and variability among patients in drug response. New England Journal of Medicine, 352(21):22112221, 2005. doi: 10.1056/ nejmra032424.

[26] Daniel W. Nebert and Timothy P. Dalton. The role of cytochrome p450 enzymes in endogenous signalling pathways and environmental carcinogenesis. Nature Reviews Can- cer, 6(12):947960, 2006. doi: 10.1038/nrc2015.

[27] Santosh Kumar. Plasma exosomes and drug metabolic cytochrome p450 enzymes. Jour- nal of Drug Metabolism Toxicology, 06(01), 2015. doi: 10.4172/2157-7609.1000e124.

[28] Ulrich M. Zanger and Matthias Schwab. Cytochrome p450 enzymes in drug metabolism: Regulation of gene expression, enzyme activities, and impact of genetic variation. Phar- macology Therapeutics, 138(1):103141, 2013. doi: 10.1016/j.pharmthera.2012.12.007.

[29] G. W. M. Chang and P. C. A. Kam. The physiological and pharmacological roles of cytochrome p450 isoenzymes. Anaesthesia, 54(1):4250, 1999. doi: 10.1046/j.1365-2044. 1999.00602.x.

[30] David W. Russell. The enzymes, regulation, and genetics of bile acid synthesis. Annual Review of Biochemistry, 72(1):137174, 2003. doi: 10.1146/annurev.biochem.72.121801. 161712.

[31] Ulrich M. Zanger and Kathrin Klein. Pharmacogenetics of cytochrome p450 2b6 (cyp2b6): advances on polymorphisms, mechanisms, and clinical relevance. Frontiers in Genetics, 4, 2013. doi: 10.3389/fgene.2013.00024.

[32] Hiroshi Yamazaki and Tsutomu Shimada. Progesterone and testosterone hydroxylation by cytochromes p450 2c19, 2c9, and 3a4 in human liver microsomes. Archives of Bio- chemistry and Biophysics, 346(1):161169, 1997. doi: 10.1006/abbi.1997.0302.

[33] James P. Hardwick. Cytochrome p450 omega hydroxylase (cyp4) function in fatty acid metabolism and metabolic diseases. Biochemical Pharmacology, 75(12):22632275, 2008. doi: 10.1016/j.bcp.2008.03.004.

[34] Anne M. Mcdonnell and Cathyyen H. Dang. Basic review of the cytochrome p450 system. Journal of the Advanced Practitioner in Oncology, 4(4), Jan 2013. doi: 10. 6004/jadpro.2013.4.4.7. References 159

[35] J.-P. Villeneuve and V. Pichette. Cytochrome p450 and liver diseases. Current Drug Metabolism, 5(3):273282, Jan 2004. doi: 10.2174/1389200043335531. [36] Charmaine S. Ferguson and Rachel F. Tyndale. Cytochrome p450 enzymes in the brain: emerging evidence of biological signiﬁcance. Trends in Pharmacological Sciences, 32 (12):708714, 2011. doi: 10.1016/j.tips.2011.08.005. [37] Catherine M. Davis, Xuehong Liu, and Nabil J. Alkayed. Cytochrome p450 eicosanoids in cerebrovascular function and disease. Pharmacology Therapeutics, 179:3146, 2017. doi: 10.1016/j.pharmthera.2017.05.004. [38] Ari Hirvonen, Kirsti Husgafvel-Pursiainen, Sisko Anttila, Antti Karjalainen, Marja Sorsa, and Harri Vainio. Metabolic cytochrome p450 genotypes and assessment of individual susceptibility to lung cancer. Pharmacogenetics, 2(6):259263, 1992. doi: 10.1097/00008571-199212000-00003. [39] Jiunn H. Lin and Jianjun Guo. Clinical implications of cyp induction-mediated drug- drug interactions. Encyclopedia of Drug Metabolism and Interactions, 2012. doi: 10. 1002/9780470921920.edm084. [40] Karoly Tihanyi, Monika L. Bakk, Eva Hellinger, and Monika Vastag. Cyp inhibition- mediated drug-drug interactions. Current Enzyme Inhibition, 6(3):130145, Jan 2010. doi: 10.2174/157340810793384106. [41] D R Abernethy and G J Burckart. Pediatric dose selection. Clinical Pharmacology Therapeutics, 87(3):270271, 2010. doi: 10.1038/clpt.2009.292. [42] Terence Stephenson. How childrens responses to drugs differ from adults. British Jour- nal of Clinical Pharmacology, 59(6):670673, 2005. doi: 10.1111/j.1365-2125.2005. 02445.x. [43] Mona Agunod, Nobuo Yamaguchi, Rafael Lopez, A. Leonard Luhby, and George B. Jerzy Glass. Correlative study of hydrochloric acid, pepsin, and intrinsic factor secretion in newborns and infants. The American Journal of Digestive Diseases, 14(6): 400414, 1969. doi: 10.1007/bf02239360. [44] H. Lu and S. Rosenbaum. Developmental pharmacokinetics in pediatric populations. The journal of pediatric pharmacology and therapeutics: JPPT, 19:262276, 2014. [45] H. Kurz, A. Mauser-Ganshorn, and H. H. Stickel. Differences in the binding of drugs to plasma proteins from newborn and adult man. i. European Journal of Clinical Pharma- cology, 11(6):463467, 1977. doi: 10.1007/bf00562940. [46] Gail D Anderson and Anne M Lynn. Optimizing pediatric dosing: A developmental pharmacologic approach. Pharmacotherapy, 29(6):680690, 2009. doi: 10.1592/phco. 29.6.680. [47] Jane Alcorn and Patrick J. Mcnamara. Ontogeny of hepatic and renal systemic clearance pathways in infants. Clinical Pharmacokinetics, 41(13):10771094, 2002. doi: 10.2165/ 00003088-200241130-00005. References 160

[48] Karen Walthall, Gregg D. Cappon, Mark E. Hurtt, and Tracey Zoetis. Postnatal development of the gastrointestinal system: A species comparison. Birth Defects Re- search Part B: Developmental and Reproductive Toxicology, 74(2):132156, 2005. doi: 10.1002/bdrb.20040.

[49] V. M. Pineiro-Carrero and E. O. Pineiro. Liver. Pediatrics, 113:10971106, 2004.

[50] Udayan Apte, Gang Zeng, Michael D. Thompson, Peggy Muller, Amanda Micsenyi, Benjamin Cieply, Klaus H. Kaestner, and Satdarshan P. S. Monga. -catenin is critical for early postnatal liver growth. American Journal of Physiology-Gastrointestinal and Liver Physiology, 292(6), 2007. doi: 10.1152/ajpgi.00359.2006.

[51] Bin Zhao, Qun-Ying Lei, and Kun-Liang Guan. The hippoyap pathway: new connec- tions between regulation of organ size and cancer. Current Opinion in Cell Biology, 20 (6):638646, 2008. doi: 10.1016/j.ceb.2008.10.001.

[52] Seth Septer, Genea Edwards, Sumedha Gunewardena, Andy Wolfe, Hua Li, James Daniel, and Udayan Apte. Yes-associated protein is involved in proliferation and differentiation during postnatal liver development. American Journal of Physiology- Gastrointestinal and Liver Physiology, 302(5), 2012. doi: 10.1152/ajpgi.00056.2011.

[53] N. Tanimizu. Notch signaling controls hepatoblast differentiation by altering the expression of liver-enriched transcription factors. Journal of Cell Science, 117(15):31653174, Jan 2004. doi: 10.1242/jcs.01169.

[54] Y. Zong, A. Panikkar, J. Xu, A. Antoniou, P. Raynaud, F. Lemaigre, and B. Z. Stanger. Notch signaling controls liver development by regulating biliary differentiation. Devel- opment, 136(10):17271739, 2009. doi: 10.1242/dev.029140.

[55] Lindsey N. Kent, Jessica B. Rakijas, Shusil K. Pandit, Bart Westendorp, Hui-Zi Chen, Justin T. Huntington, Xing Tang, Sooin Bae, Arunima Srivastava, Shantibhusan Senap- ati, and et al. E2f8 mediates tumor suppression in postnatal liver development. Journal of Clinical Investigation, 126(8):29552969, 2016. doi: 10.1172/jci85506.

[56] Yiting Tang, Kai Xiong, Ming Shen, Yulian Mu, Kui Li, and Honglin Liu. Ccaat- enhancer binding protein (c/ebp) regulates insulin-like growth factor (igf) 1 expression in porcine liver during prenatal and postnatal development. Molecular and Cellular Bio- chemistry, 401(1-2):209218, 2015. doi: 10.1007/s11010-014-2308-8.

[57] S. M. Davies. Developmental regulation of genomic imprinting of the igf2 gene in human liver. Cancer research, 54:25602562, 1994.

[58] Lai Peng, Ariel Paulson, Hua Li, Stephanie Piekos, Xi He, Linheng Li, and Xiao-Bo Zhong. Developmental programming of long non-coding rnas during postnatal liver maturation in mice. PLoS ONE, 9(12), Nov 2014. doi: 10.1371/journal.pone.0114917.

[59] Chad Pope, Stephanie C. Piekos, Liming Chen, Shashank Mishra, and Xiao-Bo Zhong. The role of h19, a long non-coding rna, in mouse liver postnatal maturation. Plos One, 12(11), Mar 2017. doi: 10.1371/journal.pone.0187557. References 161

[60] S. N. Hart, Y. Cui, C. D. Klaassen, and X.-B. Zhong. Three patterns of cytochrome p450 gene expression during liver maturation in mice. Drug Metabolism and Disposition, 37 (1):116121, Sep 2008. doi: 10.1124/dmd.108.023812.

[61] J. Y. Cui, H. J. Renaud, and C. D. Klaassen. Ontogeny of novel cytochrome p450 gene isoforms during postnatal liver maturation in mice. Drug Metabolism and Disposition, 40(6):12261237, 2012. doi: 10.1124/dmd.111.042697.

[62] Jean-Marc Treluyer, Gildas Gueret, G??rard Cheron, Michelle Sonnier, and Thierry Cresteil. Developmental expression of cyp2c and cyp2c-dependent activities in the human liver: in-vivo/in-vitro correlation and inducibility. Pharmacogenetics, 7(6):441452, 1997. doi: 10.1097/00008571-199712000-00002.

[63] S. B. Koukouritaki, J. R. Manro, S. A. Marsh, J. C. Stevens, A. E. Rettie, D. G. Mc- Carver, and R. N. Hines. Developmental expression of human hepatic cyp2c9 and cyp2c19. Journal of Pharmacology and Experimental Therapeutics, 308(3):965974, 2004. doi: 10.1124/jpet.103.060137.

[64] Jessica R. Oesterheld. A review of developmental aspects of cytochrome p450. Journal of Child and Adolescent Psychopharmacology, 8(3):161174, 1998. doi: 10.1089/cap. 1998.8.161.

[65] Saskia N. De Wildt, Gregory L. Kearns, J Steven Leeder, and John N. Van Den Anker. Cytochrome p450 3a. Clinical Pharmacokinetics, 37(6):485505, 1999. doi: 10.2165/ 00003088-199937060-00004.

[66] Edward L. Croom, Jeffrey C. Stevens, Ronald N. Hines, Andrew D. Wallace, and Ernest Hodgson. Human hepatic cyp2b6 developmental expression: The impact of age and genotype. Biochemical Pharmacology, 78(2):184190, 2009. doi: 10.1016/j.bcp.2009. 03.029.

[67] Y. Li, Y. Cui, S. N. Hart, C. D. Klaassen, and X.-B. Zhong. Dynamic patterns of histone methylation are associated with ontogenic expression of the cyp3a genes during mouse liver maturation. Molecular Pharmacology, 75(5):11711179, Feb 2009. doi: 10.1124/ mol.108.052993.

[68] Changjiang Xu, Christina Yong-Tao Li, and Ah-Ng Tony Kong. Induction of phase i, ii and iii drug metabolism/transport by xenobiotics. Archives of Pharmacal Research, 28 (3):249268, 2005. doi: 10.1007/bf02977789.

[69] Martin J. J. Ronis, Ying Chen, Kartik Shankar, Horatio Gomez-Acevedo, Mario A. Cleves, Jamie Badeaux, Michael L. Blackburn, and Thomas M. Badger. Formula feeding alters hepatic gene expression signature, iron and cholesterol homeostasis in the neonatal pig. Physiological Genomics, 43(23):12811293, 2011. doi: 10.1152/physiolgenomics. 00055.2011.

[70] Lai Peng, Stephanie Piekos, Grace L. Guo, and Xiao-Bo Zhong. Role of farnesoid x receptor in establishment of ontogeny of phase-i drug metabolizing enzyme genes in References 162

mouse liver. Acta Pharmaceutica Sinica B, 6(5):453459, 2016. doi: 10.1016/j.apsb. 2016.07.015.

[71] Cindy Yanfei Li, Sunny Lihua Cheng, Theo K. Bammler, and Julia Yue Cui. Editors highlight: Neonatal activation of the xenobiotic-sensors pxr and car results in acute and persistent down-regulation of ppar-signaling in mouse liver. Toxicological Sciences, 153 (2):282302, 2016. doi: 10.1093/toxsci/kfw127.

[72] Da Xiao, Yi-Tzai Chen, Dongfang Yang, and Bingfang Yan. Age-related inducibility of carboxylesterases by the antiepileptic agent phenobarbital and implications in drug metabolism and lipid accumulation. Biochemical Pharmacology, 84(2):232239, 2012. doi: 10.1016/j.bcp.2012.04.002.

[73] Yin-Hsuan Lai, Che-Sheng Ho, Nan-Chang Chiu, Chih-Fan Tseng, and Yuan-Ling Huang. Prognostic factors of developmental outcome in neonatal seizures in term infants. Pediatrics Neonatology, 54(3):166172, 2013. doi: 10.1016/j.pedneo.2013.01.001.

[74] Blaise F. D. Bourgeois. Antiepileptic drugs in pediatric practice. Epilepsia, 36(s2), 1995. doi: 10.1111/j.1528-1157.1995.tb05998.x.

[75] Anna Rosati, Salvatore De Masi, and Renzo Guerrini. Antiepileptic drug treatment in children with epilepsy. CNS Drugs, 29(10):847863, 2015. doi: 10.1007/ s40263-015-0281-8.

[76] Ilo E. Leppik. Metabolism of antiepileptic medication: Newborn to elderly. Epilepsia, 33:3240, 1992. doi: 10.1111/j.1528-1157.1992.tb06225.x.

[77] Martin J. Brodie, Scott Mintzer, Alison M. Pack, Barry E. Gidal, Charles J. Vecht, and Dieter Schmidt. Enzyme induction with antiepileptic drugs: Cause for concern? Epilep- sia, 54(1):1127, 2012. doi: 10.1111/j.1528-1167.2012.03671.x.

[78] Emilio Perucca. Clinically relevant drug interactions with antiepileptic drugs. British Journal of Clinical Pharmacology, 61(3):246255, 2006. doi: 10.1111/j.1365-2125.2005. 02529.x.

[79] Michael J. Painter, Mark S. Scher, Aryeh D. Stein, Stacey Armatti, Zhiming Wang, Joseph C. Gardiner, Nigel Paneth, Beth Minnigh, and John Alvin. Phenobarbital compared with phenytoin for the treatment of neonatal seizures. New England Journal of Medicine, 341(7):485489, Dec 1999. doi: 10.1056/nejm199908123410704.

[80] Shama Perveen, Abhishek Singh, Amit Upadhyay, Neha Singh, and Rashmi Chauhan. A randomized controlled trial on comparison of phenobarbitone and levetiracetam for the treatment of neonatal seizures: pilot study. International Journal of Research in Medical Sciences, page 20732078, 2016. doi: 10.18203/2320-6012.ijrms20161763.

[81] Gian M. Paciﬁci. Clinical pharmacology of phenobarbital in neonates: Effects, metabolism and pharmacokinetics. Current Pediatric Reviews, 12(1):4854, Apr 2016. doi: 10.2174/1573397111666151026223914. References 163

[82] J. Piper, D. Hawcutt, G. Verghese, S. Spinty, P. Newland, and R. Appleton. Phenytoin dosing and serum concentrations in paediatric patients requiring 20 mg/kg intravenous loading dose. Archives of Disease in Childhood, 99(Suppl 1), Jan 2014. doi: 10.1136/ archdischild-2014-306237.318.

[83] P M Loughnan, A Greenwald, W W Purton, J V Aranda, G Watters, and A H Neims. Pharmacokinetic observations of phenytoin disposition in the newborn and young infant. Archives of Disease in Childhood, 52(4):302309, Jan 1977. doi: 10.1136/adc.52.4.302.

[84] Hongbing Wang, Stephanie Faucette, Rick Moore, Tatsuya Sueyoshi, Masahiko Negishi, and Edward Lecluyse. Human constitutive androstane receptor mediates induction ofcyp2b6gene expression by phenytoin. Journal of Biological Chemistry, 279(28): 2929529301, 2004. doi: 10.1074/jbc.m400580200.

[85] M. B. Lebowitz, K. L. Olson, M. Burns, M. B. Harper, and F. Bourgeois. Drug-drug interactions among hospitalized children receiving chronic antiepileptic drug therapy. Hospital Pediatrics, 6(5):282289, 2016. doi: 10.1542/hpeds.2015-0249.

[86] Wei-Dong Chen, Xianghui Fu, Bingning Dong, Yan-Dong Wang, Steven Shiah, David D. Moore, and Wendong Huang. Neonatal activation of the nuclear receptor car results in epigenetic memory and permanent change of drug metabolism in mouse liver. Hepatology, 56(4):14991508, 2012. doi: 10.1002/hep.25766.

[87] Lisa Mausi. File:activation mechanisms of car.png. URL https://commons. wikimedia.org/w/index.php?curid=37381421.

[88] Evan D. Kharasch, Karen J. Regina, Jane Blood, and Christina Friedel. Methadone pharmacogenetics. Anesthesiology, 123(5):11421153, 2015. doi: 10.1097/aln. 0000000000000867.

[89] Manuel Romn, Dolores Ochoa, Sergio Daniel Snchez-Rojas, Maria Talegn, Rocio Prieto-Prez, ngela Rivas, Francisco Abad-Santos, and Teresa Cabaleiro. Evaluation of the relationship between polymorphisms in cyp2c19 and the pharmacokinetics of omeprazole, pantoprazole and rabeprazole. Pharmacogenomics, 15(15):18931901, 2014. doi: 10.2217/pgs.14.141.

[90] Valentina Franco and Emilio Perucca. Cyp2c9 polymorphisms and phenytoin metabolism: implications for adverse effects. Expert Opinion on Drug Metabolism Tox- icology, 11(8):12691279, Mar 2015. doi: 10.1517/17425255.2015.1053463.

[91] Ulrich Klotz. The role of pharmacogenetics in the metabolism of antiepileptic drugs. Clinical Pharmacokinetics, 46(4):271279, 2007. doi: 10.2165/ 00003088-200746040-00001.

[92] Marc Bonder, Silva Kasela, Mart Kals, Riin Tamm, Kaie Lokk, Isabel Barragan, Wim A Buurman, Patrick Deelen, Jan-Willem Greve, Maxim Ivanov, and et al. Genetic and epigenetic regulation of gene expression in fetal and adult human livers. BMC Genomics, 15(1):860, 2014. doi: 10.1186/1471-2164-15-860. References 164

[93] Aneta Novotna, Aneta Doricakova, Petr Pavek, and Zdenek Dvorak. Construction and characterization of peroxisome proliferator-activated receptor-gamma co-activator 1 al- pha (pgc-1 over-expressing cell line derived from human hepatocyte carcinoma hepg2 cells). Biomedical Papers, May 2012. doi: 10.5507/bp.2012.075.

[94] Xiaojuan Chai, Su Zeng, and Wen Xie. Nuclear receptors pxr and car: implications for drug metabolism regulation, pharmacogenomics and beyond. Expert Opinion on Drug Metabolism Toxicology, 9(3):253266, 2013. doi: 10.1517/17425255.2013.754010.

[95] Hiroki Tojima, Satoru Kakizaki, Yuichi Yamazaki, Daichi Takizawa, Norio Horiguchi, Ken Sato, and Masatomo Mori. Ligand dependent hepatic gene expression proﬁles of nuclear receptors car and pxr. Toxicology Letters, 212(3):288297, 2012. doi: 10.1016/j. toxlet.2012.06.001.

[96] Paavo Honkakoski and Masahiko Negishi. Regulation of cytochrome p450 (cyp) genes by nuclear receptors. Biochemical Journal, 347(2):321, 2000. doi: 10.1042/0264-6021: 3470321.

[97] Chandra Prakash, Baltazar Zuniga, Chung Seog Song, Shoulei Jiang, Jodie Cropper, Sulgi Park, and Bandana Chatterjee. Nuclear receptors in drug metabolism, drug response and drug interactions. Nuclear Receptor Research, 2, 2015. doi: 10.11131/2015/ 101178.

[98] M. Ingelman-Sundberg, X.-B. Zhong, O. Hankinson, S. Beedanagari, A.-M. Yu, L. Peng, and Y. Osawa. Potential role of epigenetic mechanisms in the regulation of drug metabolism and transport. Drug Metabolism and Disposition, 41(10):17251731, May 2013. doi: 10.1124/dmd.113.053157.

[99] Tina Branscombe Miranda and Peter A. Jones. Dna methylation: The nuts and bolts of repression. Journal of Cellular Physiology, 213(2):384390, 2007. doi: 10.1002/jcp. 21224.

[100] Jinling Zhang, Xuebin Zhu, Yuhong Li, Lingyan Zhu, Shiming Li, Guoying Zheng, Qi Ren, Yonghong Xiao, and Fumin Feng. Correlation of cpg island methylation of the cytochrome p450 2e1/2d6 genes with liver injury induced by anti-tuberculosis drugs: A nested case-control study. International Journal of Environmental Research and Public Health, 13(8):776, Jan 2016. doi: 10.3390/ijerph13080776.

[101] L. Cerveny, L. Svecova, E. Anzenbacherova, R. Vrzal, F. Staud, Z. Dvorak, J. Ulri- chova, P. Anzenbacher, and P. Pavek. Valproic acid induces cyp3a4 and mdr1 gene expression by activation of constitutive androstane receptor and pregnane x receptor pathways. Drug Metabolism and Disposition, 35(7):10321041, Apr 2007. doi: 10.1124/dmd.106.014456.

[102] D. Takizawa, S. Kakizaki, N. Horiguchi, H. Tojima, Y. Yamazaki, T. Ichikawa, K. Sato, and M. Mori. Histone deacetylase inhibitors induce cytochrome p450 2b by activating nuclear receptor constitutive androstane receptor. Drug Metabolism and Disposition, 38 (9):14931498, Jan 2010. doi: 10.1124/dmd.110.032854. References 165

[103] Liang Yan, Yiting Wang, Jingyang Liu, Yali Nie, Xiao-Bo Zhong, Quancheng Kan, and Lirong Zhang. Alterations of histone modiﬁcations contribute to pregnane x receptor- mediated induction of cyp3a4 by rifampicin. Molecular Pharmacology, 92(2):113123, 2017. doi: 10.1124/mol.117.108225.

[104] I. Papageorgiou, S. Grepper, and J. D. Unadkat. Induction of hepatic cyp3a enzymes by pregnancy-related hormones: Studies in human hepatocytes and hepatic cell lines. Drug Metabolism and Disposition, 41(2):281290, Jun 2012. doi: 10.1124/dmd.112.049015.

[105] S.-Y. Choi, K. H. Koh, and H. Jeong. Isoform-speciﬁc regulation of cytochromes p450 expression by estradiol and progesterone. Drug Metabolism and Disposition, 41(2): 263269, 2012. doi: 10.1124/dmd.112.046276.

[106] Yuki Tsuchiya, Miki Nakajima, and Tsuyoshi Yokoi. Cytochrome p450-mediated metabolism of estrogens and its regulation in human. Cancer Letters, 227(2):115124, 2005. doi: 10.1016/j.canlet.2004.10.007.

[107] David J. Waxman and Caitlin OConnor. Growth hormone regulation of sex-dependent liver gene expression. Molecular Endocrinology, 20(11):26132629, 2006. doi: 10.1210/ me.2006-0007.

[108] Zdenek Dvorak and Peotr Pavek. Regulation of drug-metabolizing cytochrome p450 enzymes by glucocorticoids. Drug Metabolism Reviews, 42(4):621635, 2010. doi: 10. 3109/03602532.2010.484462.

[109] Ye Li and Xiao-Bo Zhong. Epigenetic regulation of developmental expression of cyp2d genes in mouse liver. Acta Pharmaceutica Sinica B, 2(2):146158, 2012. doi: 10.1016/j. apsb.2012.01.001.

[110] R. Faris and T. Campbell. Exposure of newborn rats to pharmacologically active compounds may permanently alter carcinogen metabolism. Science, 211(4483):719721, 1981. doi: 10.1126/science.7455708.

[111] E. M. Lui, J. Gregson, and G. W. Lucier. Altered sex differentiation of hepatic ethylmor- phine n-demethylation in the male rat following neonatal methadone exposure. Pediatric pharmacology, 1:187195, 1981.

[112] D.m. Bagley and J.r. Hayes. Neonatal phenobarbital administration results in increased cytochrome p450-dependent monooxygenase activity in adult male and female rats. Bio- chemical and Biophysical Research Communications, 114(3):11321137, 1983. doi: 10.1016/0006-291x(83)90680-0.

[113] Daniel M. Bagley and Johnnie R. Hayes. Neonatal phenobarbital imprinting of hepatic microsomal enzymes in adult rats: Modulation by neonatal testosterone presence. Tox- icology and Applied Pharmacology, 79(2):227235, 1985. doi: 10.1016/0041-008x(85) 90344-8. References 166

[114] J. M. Haake and S. H. Safe. Neonatal phenobarbital imprinting of rat hepatic microsomal testosterone hydroxylations. Journal of Biochemical Toxicology, 3(4):309319, 1988. doi: 10.1002/jbt.2570030408.

[115] A. K. Agrawal and B. H. Shapiro. Phenobarbital induction of hepatic cyp2b1 and cyp2b2: pretranscriptional and post-transcriptional effects of gender, adult age, and phenobarbital dose. Molecular pharmacology, 49:523531, 1996.

[116] A. K. Agrawal and B. H. Shapiro. Latent overexpression of hepatic cyp2c7 in adult male and female rats neonatally exposed to phenobarbital: a developmental proﬁle of gender- dependent p450s. The journal of pediatric pharmacology and therapeutics: JPPT, 293: 10271033, 2000.

[117] A. K. Agrawal, N. A. Pampori, and B. H. Shapiro. Neonatal phenobarbital-induced defects in age- and sex-speciﬁc growth hormone proﬁles regulating monooxygenases. American Journal of Physiology-Endocrinology and Metabolism, 268(3), 1995. doi: 10.1152/ajpendo.1995.268.3.e439.

[118] Arun K. Agrawal and Bernard H. Shapiro. Phenobarbital-imprinted overinduction of adult constituent cyp isoforms. Pharmacology, 68(4):204215, 2003. doi: 10.1159/ 000070460.

[119] Arun K. Agrawal and Bernard H. Shapiro. Neonatal phenobarbital imprints overexpression of cytochromes p450 with associated increase in tumorigenesis and reduced life span. The FASEB Journal, 19(3):470472, 2005. doi: 10.1096/fj.04-2550fje.

[120] Y.-C. Tien, K. Liu, C. Pope, P. Wang, X. Ma, and X.-B. Zhong. Dose of phenobarbital and age of treatment at early life are two key factors for the persistent induction of cytochrome p450 enzymes in adult mouse liver. Drug Metabolism and Disposition, 43 (12):19381945, 2015. doi: 10.1124/dmd.115.066316.

[121] Ashu Johri, Sanjay Yadav, Alok Dhawan, and Devendra Parmar. Responsiveness of cerebral and hepatic cytochrome p450s in rat offspring prenatally exposed to lindane. Toxicology and Applied Pharmacology, 231(1):1016, 2008. doi: 10.1016/j.taap.2008. 03.019.

[122] M. Kawai, S. M. Bandiera, T. K. Chang, F. M. Poulet, P. M. Vancutsem, and G. D. Bell- ward. Modulation of hepatic cyp2a1, cyp2c11, and cyp3a9 expression in adult rats by neonatal administration of tamoxifen. Drug metabolism and disposition, 27:13921398, 1999.

[123] Richard C. Zangar, David L. Springer, and Donald R. Buhler. Alterations in cytochrome p450 levels in adult rats following neonatal exposure to xenobiotics. Journal of Toxicol- ogy and Environmental Health, 38(1):4355, 1993. doi: 10.1080/15287399309531699.

[124] Yan Liu, Rick Havinga, Feike R Van Der Leij, Renze Boverhof, Pieter J J Sauer, Folkert Kuipers, and Frans Stellaard. Dexamethasone exposure of neonatal rats modulates biliary lipid secretion and hepatic expression of genes controlling bile acid metabolism in References 167

adulthood without interfering with primary bile acid kinetics. Pediatric Research, 63(4): 375381, 2008. doi: 10.1203/pdr.0b013e318165b8af. [125] Epolia Ramadan, Helene Blanchard, Yewon Cheon, Meredith A. Fox, Lisa Chang, Mei Chen, Kaizong Ma, Stanley I. Rapoport, and Mireille Basselin. Transient postnatal fluoxetine leads to decreased brain arachidonic acid metabolism and cytochrome p450 4a in adult mice. Prostaglandins, Leukotrienes and Essential Fatty Acids (PLEFA), 90(5): 191197, 2014. doi: 10.1016/j.plefa.2014.01.003. [126] Zhi-Xin Yuan and Stanley I. Rapoport. Transient postnatal fluoxetine decreases brain concentrations of 20-hete and 15-epi-lxa4, arachidonic acid metabolites in adult mice. Prostaglandins, Leukotrienes and Essential Fatty Acids (PLEFA), 101:914, 2015. doi: 10.1016/j.plefa.2015.07.002. [127] Wolf Reik and Jrn Walter. Genomic imprinting: parental influence on the genome. Na- ture Reviews Genetics, 2(1):2132, 2001. doi: 10.1038/35047554. [128] Y. Omoto, O. Imamov, M. Warner, and J.-A. Gustafsson. Estrogen receptor and imprinting of the neonatal mouse ventral prostate by estrogen. Proceedings of the National Academy of Sciences, 102(5):14841489, 2005. doi: 10.1073/pnas.0409168102. [129] Cheryl Lyn Walker. Minireview: Epigenomic plasticity and vulnerability to edc exposures. Molecular Endocrinology, 30(8):848855, 2016. doi: 10.1210/me.2016-1086. [130] Donato F. Romagnolo, Janos Zempleni, and Ornella I. Selmin. Nuclear receptors and epigenetic regulation: Opportunities for nutritional targeting and disease prevention. Ad- vances in Nutrition, 5(4):373385, Jan 2014. doi: 10.3945/an.114.005868. [131] Lauren Prusinski, Ayman Al-Hendy, and Qiwei Yang. Developmental exposure to endocrine disrupting chemicals alters the epigenome: Identification of reprogrammed targets. Gynecology and Obstetrics Research - Open Journal, 3(1):16, 2016. doi: 10.17140/goroj-3-127. [132] Rebecca Lee Yean Wong, Quan Wang, Lindsey S. Trevio, Maarten C. Bosland, Jing Chen, Mario Medvedovic, Gail S. Prins, Kurunthachalam Kannan, Shuk-Mei Ho, and Cheryl Lyn Walker. Identification of secretaglobin scgb2a1as a target for developmental reprogramming by bpa in the rat prostate. Epigenetics, 10(2):127134, 2015. doi: 10. 1080/15592294.2015.1009768. [133] Quan Wang, Lindsey S. Trevino, Rebecca Lee Yean Wong, Mario Medvedovic, Jing Chen, Shuk-Mei Ho, Jianjun Shen, Charles E. Foulds, Cristian Coarfa, Bert W. Omalley, and et al. Reprogramming of the epigenome by mll1 links early-life environmental exposures to prostate cancer risk. Molecular Endocrinology, 30(8):856871, 2016. doi: 10.1210/me.2015-1310. [134] S. Li, K. A. Washburn, R. Moore, T. Uno, C. Teng, R. R. Newbold, J. A. McLachlan, and M. Negishi. Developmental exposure to diethylstilbestrol elicits demethylation of estrogen-responsive lactoferrin gene in mouse uterus. Cancer research, 57:43564359, 1997. References 168

[135] Wan-Yee Tang, Retha Newbold, Katerina Mardilovich, Wendy Jefferson, Robert Y. S. Cheng, Mario Medvedovic, and Shuk-Mei Ho. Persistent hypomethylation in the promoter ofnucleosomal binding protein1 (nsbp1) correlates with overexpression ofnsbp1 in mouse uteri neonatally exposed to diethylstilbestrol or genistein. Endocrinology, 149 (12):59225931, 2008. doi: 10.1210/en.2008-0682. [136] Aurlie Vega, Marine Baptissart, Emmanuelle Martinot, Jean-Paul Saru, Silvre Baron, Kristina Schoonjans, and David H Volle. Hepatotoxicity induced by neonatal exposure to diethylstilbestrol is maintained throughout adulthood via the nuclear receptor shp. Expert Opinion on Therapeutic Targets, page 110, 2014. doi: 10.1517/14728222.2014.964209. [137] Haibo Xu, Ratheishan Rajesan, Patricia Harper, Richard B Kim, Bo Lonnerdal, Ming- dong Yang, Satoko Uematsu, Janine Hutson, Jo Watson-Macdonell, Shinya Ito, and et al. Induction of cytochrome p450 1a by cow milk-based formula: a comparative study between human milk and formula. British Journal of Pharmacology, 146(2):296305, 2009. doi: 10.1038/sj.bjp.0706319. [138] J. H. Lin. Species similarities and differences in pharmacokinetics. Drug metabolism and disposition, 23:10081021, 1995. [139] Dharamainder Choudhary, Ingela Jansson, Ivaylo Stoilov, Mansoor Sarfarazi, and John B. Schenkman. Expression patterns of mouse and human cyp orthologs (families 14) during development and in different adult tissues. Archives of Biochemistry and Biophysics, 436(1):5061, 2005. doi: 10.1016/j.abb.2005.02.001. [140] E. Hrycay and S. Bandiera. Expression, function and regulation of mouse cytochrome p450 enzymes: Comparison with human cytochrome p450 enzymes. Current Drug Metabolism, 10(10):11511183, Jan 2009. doi: 10.2174/138920009790820138. [141] L. Peng, B. Yoo, S. S. Gunewardena, H. Lu, C. D. Klaassen, and X.-B. Zhong. Rna sequencing reveals dynamic changes of mrna abundance of cytochromes p450 and their alternative transcripts during mouse liver development. Drug Metabolism and Disposi- tion, 40(6):11981209, 2012. doi: 10.1124/dmd.112.045088. [142] Kazuo Nakayama, Yukiko Sudo, Yasumasa Sasaki, Hiroshi Iwata, Minoru Takahashi, and Tetsuya Kamataki. Studies on transcriptional regulation of cyp3a16 gene in mouse livers by application of direct dna injection method. Biochemical and Biophysical Re- search Communications, 287(4):820824, 2001. doi: 10.1006/bbrc.2001.5663. [143] Yue-Ming Wang, Su Sien Ong, Sergio C Chai, and Taosheng Chen. Role of car and pxr in xenobiotic sensing and metabolism. Expert Opinion on Drug Metabolism Toxicology, 8(7):803817, Mar 2012. doi: 10.1517/17425255.2012.685237. [144] Sulagna Dutta and Pallav Sengupta. Men and mice: Relating their ages. Life Sciences, 152:244248, 2016. doi: 10.1016/j.lfs.2015.10.025. [145] Anroopb Nair and Shery Jacob. A simple practice guide for dose conversion between animals and human. Journal of Basic and Clinical Pharmacy, 7(2):27, 2016. doi: 10. 4103/0976-0105.177703. References 169

[146] Vijay Sharma and John H Mcneill. To scale or not to scale: the principles of dose extrapolation. British Journal of Pharmacology, 157(6):907921, 2009. doi: 10.1111/j. 1476-5381.2009.00267.x. [147] H. Sharifi, M. Hasanloei, and J. Mahmoudi. Polypharmacy-induced drug-drug interactions; threats to patient safety. Drug Research, 64(12):633637, May 2014. doi: 10.1055/s-0033-1363965. [148] Lara Magro, Ugo Moretti, and Roberto Leone. Epidemiology and characteristics of adverse drug reactions caused by drugdrug interactions. Expert Opinion on Drug Safety, 11(1):8394, 2012. doi: 10.1517/14740338.2012.631910. [149] T Lynch and A Price. The effect of cytochrome p450 metabolism on drug response, interactions, and adverse effects. Am Fam Physician, 76(3):391396, Aug 2007. [150] Laurel A. Slaughter, Anup D. Patel, and Jonathan L. Slaughter. Pharmacological treatment of neonatal seizures. Journal of Child Neurology, 28(3):351364, 2013. doi: 10.1177/0883073812470734. [151] P Czekaj. Phenobarbital-induced expression of cytochrome p450 genes. Acta Biochim Pol, 47(4):10931105, 2000. [152] Gabriella Der. An overview of proton pump inhibitors. Gastroenterology Nursing, 26 (5):182190, 2003. doi: 10.1097/00001610-200309000-00003. [153] Tommy Andersson. Pharmacokinetics, metabolism and interactions of acid pump inhibitors. Clinical Pharmacokinetics, 31(1):928, 1996. doi: 10.2165/ 00003088-199631010-00002. [154] T Andersson, Jo Miners, Me Veronese, and Dj Birkett. Identification of human liver cytochrome p450 isoforms mediating secondary omeprazole metabolism. British Jour- nal of Clinical Pharmacology, 37(6):597604, 1994. doi: 10.1111/j.1365-2125.1994. tb04310.x. [155] Eun-Ja Park, Hea-Young Cho, and Yong-Bok Lee. Effect of cimetidine and phenobarbital on metabolite kinetics of omeprazole in rats. Archives of Pharmacal Research, 28 (10):11961202, 2005. doi: 10.1007/bf02972986. [156] Lina Panayiota Aristoteli, Jani L. Orourke, Stephen Danon, Hakan Larsson, Bjorn Mell- gard, Hazel Mitchell, and Adrian Lee. Urea, fluorofamide, and omeprazole treatments alter helicobacter colonization in the mouse gastric mucosa. Helicobacter, 11(5):460468, 2006. doi: 10.1111/j.1523-5378.2006.00439.x. [157] Karen E. Brenneman, Crystal Willingham, Jacquelyn A. Kilbourne, Roy Curtiss 3Rd, and Kenneth L. Roland. A low gastric ph mouse model to evaluate live attenuated bac- terial vaccines. PLoS ONE, 9(1), 2014. doi: 10.1371/journal.pone.0087411. [158] Marina Kacevska, Maxim Ivanov, and Magnus Ingelman-Sundberg. Epigenetic- dependent regulation of drug transport and metabolism: an update. Pharmacogenomics, 13(12):13731385, 2012. doi: 10.2217/pgs.12.119. References 170

[159] Yun-Chen Tien, Stephanie C. Piekos, Chad Pope, and Xiao-Bo Zhong. Phenobarbital treatment at a neonatal age results in decreased efﬁcacy of omeprazole in adult mice. Drug Metabolism and Disposition, 45(3):330335, Jun 2017. doi: 10.1124/dmd.116. 073601.

[160] Carine B. Lambert, Catherine Spire, Nancy Claude, and Andre Guillouzo. Dose- and time-dependent effects of phenobarbital on gene expression proﬁling in human hepatoma heparg cells. Toxicology and Applied Pharmacology, 234(3):345360, 2009. doi: 10. 1016/j.taap.2008.11.008.

[161] Daochuan Li, Bryan Mackowiak, Timothy G. Brayman, Michael Mitchell, Lei Zhang, Shiew-Mei Huang, and Hongbing Wang. Genome-wide analysis of human constitutive androstane receptor (car) transcriptome in wild-type and car-knockout heparg cells. Bio- chemical Pharmacology, 98(1):190202, 2015. doi: 10.1016/j.bcp.2015.08.087.

[162] Raphalle Luisier, Harri Lempiinen, Nina Scherbichler, Albert Braeuning, Miriam Geissler, Valerie Dubost, Arne Mller, Nico Scheer, Salah-Dine Chibout, Hisanori Hara, and et al. Phenobarbital induces cell cycle transcriptional responses in mouse liver humanized for constitutive androstane and pregnane x receptors. Toxicological Sciences, 139(2):501511, Jan 2014. doi: 10.1093/toxsci/kfu038.

[163] Zhiyun Cheng, Yinghui Lv, Suqiu Pang, Ruyu Bai, Mingxi Wang, Shuyu Lin, Tianwen Xu, Duncan Spalding, Nagy Habib, Ruian Xu, and et al. Kallistatin, a new and reli- able biomarker for the diagnosis of liver cirrhosis. Acta Pharmaceutica Sinica B, 5(3): 194200, 2015. doi: 10.1016/j.apsb.2015.02.003.

[164] Tatsuya Sueyoshi, Takeshi Kawamoto, Igor Zelko, Paavo Honkakoski, and Masahiko Negishi. The repressed nuclear receptor car responds to phenobarbital in activating the humancyp2b6gene. Journal of Biological Chemistry, 274(10):60436046, May 1999. doi: 10.1074/jbc.274.10.6043.

[165] Jiong Yan, Baian Chen, Jing Lu, and Wen Xie. Deciphering the roles of the constitutive androstane receptor in energy metabolism. Acta Pharmacologica Sinica, 36(1):6270, 2014. doi: 10.1038/aps.2014.102.

[166] Hui Yang and Hongbing Wang. Signaling control of the constitutive androstane receptor (car). Protein Cell, 5(2):113123, 2014. doi: 10.1007/s13238-013-0013-0.

[167] Tenghui Chen, Qiang Chen, Yixiang Xu, Qiling Zhou, Jingwei Zhu, Hao Zhang, Qiao Wu, Jianming Xu, and Chundong Yu. Src-3 is required for car-regulated hepatocyte proliferation and drug metabolism. Journal of Hepatology, 56(1):210217, 2012. doi: 10.1016/j.jhep.2011.07.015.

[168] Meghan Breaux, Kyle Lewis, Leila Valanejad, Polina Iakova, Fengju Chen, Qianx- ing Mo, Estela Medrano, Lubov Timchenko, and Nikolai Timchenko. p300 regulates liver functions by controlling p53 and c/ebp family proteins through multiple signaling pathways. Molecular and Cellular Biology, 35(17):30053016, 2015. doi: 10.1128/mcb.00421-15. References 171

[169] Ling He, Karuna Naik, Shumei Meng, Jia Cao, Aniket R. Sidhaye, Anlin Ma, Sally Radovick, and Fredric E. Wondisford. Transcriptional co-activator p300 maintains basal hepatic gluconeogenesis. Journal of Biological Chemistry, 287(38):3206932077, 2012. doi: 10.1074/jbc.m112.385864.

[170] Jianneng Li, Wangsheng Wang, Chao Liu, Weihua Wang, Wenjiao Li, Qun Shu, Zi-Jiang Chen, and Kang Sun. Critical role of histone acetylation by p300 in human placental 11- hsd2 expression. The Journal of Clinical Endocrinology Metabolism, 98(7), 2013. doi: 10.1210/jc.2012-4291.

[171] Youwen Yang, Ji-Fan Hu, Gary A. Ulaner, Tao Li, Xiaoming Yao, Thanh H. Vu, and Andrew R. Hoffman. Epigenetic regulation of igf2/h19 imprinting at ctcf insulator binding sites. Journal of Cellular Biochemistry, 90(5):10381055, Dec 2003. doi: 10.1002/jcb.10684.

[172] S. Mutoh, M. Sobhany, R. Moore, L. Perera, L. Pedersen, T. Sueyoshi, and M. Negishi. Phenobarbital indirectly activates the constitutive active androstane receptor (car) by inhibition of epidermal growth factor receptor signaling. Science Signaling, 6(274), Jul 2013. doi: 10.1126/scisignal.2003705.

[173] Wee Ping and Zhixiang Wang. Epidermal growth factor receptor cell proliferation signaling pathways. Cancers, 9(12):52, 2017. doi: 10.3390/cancers9050052.

[174] Hideki Makinoshima, Masahiro Takita, Shingo Matsumoto, Atsushi Yagishita, Satoshi Owada, Hiroyasu Esumi, and Katsuya Tsuchihara. Epidermal growth factor receptor (egfr) signaling regulates global metabolic pathways in egfr-mutated lung ade- nocarcinoma. Journal of Biological Chemistry, 289(30):2081320823, 2014. doi: 10.1074/jbc.m114.575464.

[175] Karin Komposch and Maria Sibilia. Egfr signaling in liver diseases. International Jour- nal of Molecular Sciences, 17(1):30, 2015. doi: 10.3390/ijms17010030.

[176] Maria Sibilia, Renate Kroismayr, Beate M. Lichtenberger, Anuradha Natarajan, Manfred Hecking, and Martin Holcmann. The epidermal growth factor receptor: from development to tumorigenesis. Differentiation, 75(9):770787, 2007. doi: 10.1111/j.1432-0436. 2007.00238.x.

[177] Carmen Berasain and Matas A. Avila. The egfr signalling system in the liver: from hepatoprotection to hepatocarcinogenesis. Journal of Gastroenterology, 49(1):923, Oct 2013. doi: 10.1007/s00535-013-0907-x.

[178] A. Collin De Lhortet, A. Zerrad-Saadi, C. Prip-Buus, V. Fauveau, N. Helmy, M. Ziol, C. Vons, K. Billot, V. Baud, Hlne Gilgenkrantz, and et al. Gh administration rescues fatty liver regeneration impairment by restoring gh/egfr pathway deﬁciency. Endocrinology, 155(7):25452554, 2014. doi: 10.1210/en.2014-1010.

[179] Jasmine Kaur and Kulbhushan Tikoo. p300/cbp dependent hyperacetylation of histone potentiates anticancer activity of geﬁtinib nanoparticles. Biochimica et Biophysica Acta References 172

(BBA) - Molecular Cell Research, 1833(5):10281040, 2013. doi: 10.1016/j.bbamcr. 2013.01.029.

[180] Jessica M. Ferrell and John Y.l. Chiang. Circadian rhythms in liver metabolism and disease. Acta Pharmaceutica Sinica B, 5(2):113122, 2015. doi: 10.1016/j.apsb.2015.01. 003.

[181] Robert Dallmann, Alper Okyar, and Francis Lvi. Dosing-time makes the poison: Cir- cadian regulation and pharmacotherapy. Trends in Molecular Medicine, 22(5):430445, 2016. doi: 10.1016/j.molmed.2016.03.004.

[182] Carol S. Camﬁeld, Sheila. Chaplin, Anna-Beth. Doyle, Stanley H. Shapiro, Carl. Cum- mings, and Peter R. Camﬁeld. Side effects of phenobarbital in toddlers; behavioral and cognitive aspects. The Journal of Pediatrics, 95(3):361365, 1979. doi: 10.1016/s0022-3476(79)80507-7.

[183] Peter Vestergaard. Effects of antiepileptic drugs on bone health and growth potential in children with epilepsy. Pediatric Drugs, 17(2):141150, Aug 2015. doi: 10.1007/ s40272-014-0115-z.

[184] Carlo La Vecchia and Eva Negri. A review of epidemiological data on epilepsy, phenobarbital, and risk of liver cancer. European Journal of Cancer Prevention, 23(1):17, 2014. doi: 10.1097/cej.0b013e32836014c8.

[185] Peter B. Mller, Karen J. Wolfsheimer, Joseph Taboada, Giselle Hosgood, Beth P. Part- ington, and Frdric P. Gaschen. Effects of long-term phenobarbital treatment on the thyroid and adrenal axis and adrenal function tests in dogs. Journal of Veterinary Internal Medicine, 14(2):157, 2000. doi: 10.1892/0891-6640(2000)014h0157:eolptoi2.3.co;2.

[186] Eirs J, Novo-Rodrguez Mai, Ro M Del, P Meseguer, Ro M.c Del, and M Castro-Gago. The effects on lipid and apolipoprotein serum levels of long-term carbamazepine, valproic acid and phenobarbital therapy in children with epilepsy. Epilepsy Research, 41 (1):17, 2000. doi: 10.1016/s0920-1211(00)00119-4.

[187] Hong Lu, Julia Cui, Sumedha Gunewardena, Byunggil Yoo, Xiao-Bo Zhong, and Curtis Klaassen. Hepatic ontogeny and tissue distribution of mrnas of epigenetic modiﬁers in mice using rna-sequencing. Epigenetics, 7(8):914929, 2012. doi: 10.4161/epi.21113.

[188] Stephanie Piekos, Chad Pope, Austin Ferrara, and Xiao-Bo Zhong. Impact of drug treatment at neonatal ages on variability of drug metabolism and drug-drug interactions in adult life. Current Pharmacology Reports, 3(1):19, Mar 2017. doi: 10.1007/ s40495-016-0078-6.

[189] Atsushi Kawase, Hiroyuki Tanaka, Toru Otori, Kenji Matsuyama, and Masahiro Iwaki. Effects of duration of phenytoin administration on mrna expression of cytochrome p450 and p-glycoprotein in the liver and small intestine of rats. Asian Journal of Pharmaceu- tical Sciences, 11(5):662667, 2016. doi: 10.1016/j.ajps.2016.04.003. References 173

[190] Geoffrey J. Markowitz, Shilpa D. Kadam, Dawn M. Boothe, Natasha D. Irving, and Anne M. Comi. The pharmacokinetics of commonly used antiepileptic drugs in immature cd1 mice. NeuroReport, 21(6):452456, 2010. doi: 10.1097/wnr.0b013e328338ba18. [191] Igor Zelko and Masahiko Negishi. Phenobarbital-elicited activation of nuclear receptor car in induction of cytochrome p450 genes. Biochemical and Biophysical Research Communications, 277(1):16, 2000. doi: 10.1006/bbrc.2000.3557. [192] J. P. Jackson, S. S. Ferguson, M. Negishi, and J. A. Goldstein. Phenytoin induction of the cyp2c37 gene is mediated by the constitutive androstane receptor. Drug Metabolism and Disposition, 34(12):20032010, 2006. doi: 10.1124/dmd.106.012005. [193] J. P. Jackson, S. S. Ferguson, R. Moore, M. Negishi, and J. A. Goldstein. The constitutive active/androstane receptor regulates phenytoin induction of cyp2c29. Molecular Pharmacology, 65(6):13971404, Jan 2004. doi: 10.1124/mol.65.6.1397. [194] Aaron W. Bell and George K. Michalopoulos. Phenobarbital regulates nuclear expression of hnf-4 in mouse and rat hepatocytes independent of car and pxr. Hepatology, 44 (1):186194, 2006. doi: 10.1002/hep.21234. [195] Sven Bjorkman. Prediction of cytochrome p450-mediated hepatic drug clearance in neonates, infants and children. Clinical Pharmacokinetics, 45(1):111, 2006. doi: 10. 2165/00003088-200645010-00001. [196] Valentina Fabiano, Chiara Mameli, and Gian Vincenzo Zuccotti. Adverse drug reactions in newborns, infants and toddlers: pediatric pharmacovigilance between present and future. Expert Opinion on Drug Safety, 11(1):95105, Sep 2011. doi: 10.1517/14740338. 2011.584531. [197] Ronald N. Hines. The ontogeny of drug metabolism enzymes and implications for adverse drug events. Pharmacology Therapeutics, 118(2):250267, 2008. doi: 10.1016/j. pharmthera.2008.02.005. [198] R.n. Hines. Developmental expression of drug metabolizing enzymes: Impact on disposition in neonates and young children. International Journal of Pharmaceutics, 452 (1-2):37, 2013. doi: 10.1016/j.ijpharm.2012.05.079. [199] Dan Lacroix, Michelle Sonnier, Arlette Moncion, Gerard Cheron, and Thierry Cresteil. Expression of cyp3a in the human liver - evidence that the shift between cyp3a7 and cyp3a4 occurs immediately after birth. European Journal of Biochemistry, 247(2): 625634, 1997. doi: 10.1111/j.1432-1033.1997.00625.x. [200] C. A. Vyhlidal, R. Gaedigk, and J. S. Leeder. Nuclear receptor expression in fetal and pediatric liver: Correlation with cyp3a expression. Drug Metabolism and Disposition, 34(1):131137, 2005. doi: 10.1124/dmd.105.005967. [201] J. A. Williams, B. J. Ring, V. E. Cantrell, D. R. Jones, J. Eckstein, K. Ruterbories, M. A. Hamman, S. D. Hall, and S. A. Wrighton. Comparative metabolic capabilities of cyp3a4, cyp3a5, and cyp3a7. Drug Metabolism and Disposition, 30(8):883891, Jan 2002. doi: 10.1124/dmd.30.8.883. References 174

[202] Xiao-Yan Pang, Jie Cheng, Jung-Hwan Kim, Tsutomu Matsubara, Kristopher W. Krausz, and Frank J. Gonzalez. Expression and regulation of human fetal-speciﬁc cyp3a7 in mice. Endocrinology, 153(3):14531463, 2012. doi: 10.1210/en.2011-1020. [203] Purnima Singh, Dong-Hoon Lee, and Piroska E. Szab. More than insulator: multiple roles of ctcf at the h19-igf2 imprinted domain. Frontiers in Genetics, 3, 2012. doi: 10.3389/fgene.2012.00214. [204] Khaled Abduljalil, Masoud Jamei, Amin Rostami-Hodjegan, and Trevor N. Johnson. Changes in individual drug-independent system parameters during virtual paediatric pharmacokinetic trials: Introducing time-varying physiology into a paediatric pbpk model. The AAPS Journal, 16(3):568576, Apr 2014. doi: 10.1208/s12248-014-9592-9. [205] Dipika Bansal, Chandrika Azad, Manpreet Kaur, Neelima Rudroju, Pravallika Vepa, and Vishal Guglani. Adverse effects of antiepileptic drugs in north indian pediatric outpatients. Clinical Neuropharmacology, 36(4):107113, 2013. doi: 10.1097/wnf. 0b013e31829a498d. [206] Ipek Polat, Pakize Karaoglu, Muge Ayanoglu, Uluc Yis, and Semra Hiz. Life-threatening and rare adverse effects of phenytoin. Pediatric Emergency Care, 31(7), 2015. doi: 10.1097/pec.0000000000000495. [207] Indira Valadł Carvalho, Renata Cavalcanti Carnevale, Marlia Berlofa Visacri, Priscila Gava Mazzola, Rosiane De Ftima Lopes Ambrsio, Marcelo Conrado Dos Reis, Rachel Alvarenga De Queiroz, and Patricia Moriel. Drug interaction between phenytoin and valproic acid in a child with refractory epilepsy. Journal of Pharmacy Practice, 27 (2):214216, 2013. doi: 10.1177/0897190013515708. [208] Yoshiaki Yamamoto, Yukitoshi Takahashi, Katsumi Imai, Kou Miyakawa, Hiroko Ikeda, Yuki Ueda, Tokito Yamaguchi, Hirosato Nasu, Hideyuki Ohtani, Hideo Shigematsu, and et al. Individualized phenytoin therapy for japanese pediatric patients with epilepsy based on cyp2c9 and cyp2c19 genotypes. Therapeutic Drug Monitoring, 37(2):229235, 2015. doi: 10.1097/ftd.0000000000000128. [209] Kittika Yampayon, Chonlaphat Sukasem, Chanin Limwongse, Yotin Chinvarun, Therd- pong Tempark, Ticha Rerkpattanapipat, and Pornpimol Kijsanayotin. Inﬂuence of genetic and non-genetic factors on phenytoin-induced severe cutaneous adverse drug reactions. European Journal of Clinical Pharmacology, 73(7):855865, Aug 2017. doi: 10.1007/s00228-017-2250-2. [210] B. Guldiken, J. Rmi, and Soheyl Noachtar. Cardiovascular adverse effects of phenytoin. Journal of Neurology, 263(5):861870, Aug 2015. doi: 10.1007/s00415-015-7967-1. [211] Jose L. Herranz, Juan A. Armijo, and Rosa Arteaga. Clinical side effects of phenobarbital, primidone, phenytoin, carbamazepine, and valproate during monotherapy in children. Epilepsia, 29(6):794804, 1988. doi: 10.1111/j.1528-1157.1988.tb04237.x. [212] Wen-Hung Chung and Shuen-Iu Hung. Genetic factors associated with severe cutaneous adverse reactionsreply. Jama, 312(20):2166, 2014. doi: 10.1001/jama.2014.14148. References 175

[213] J. Steven Leeder, Xiaoli Lu, Yoav Timsit, and Andrea Gaedigk. Non-monooxygenase cytochromes p450 as potential human autoantigens in anticonvulsant hypersen- sitivity reactions. Pharmacogenetics, 8(3):211226, 1998. doi: 10.1097/ 00008571-199806000-00004.

[214] Ludovic C. Gillet, Pedro Navarro, Stephen Tate, Hannes Rst, Nathalie Selevsek, Lukas Reiter, Ron Bonner, and Ruedi Aebersold. Targeted data extraction of the ms/ms spectra generated by data-independent acquisition: A new concept for consistent and accurate proteome analysis. Molecular Cellular Proteomics, 11(6), 2012. doi: 10.1074/mcp. o111.016717.

[215] Yanyan Gao, Xinzheng Wang, Zhihong Sang, Zongcheng Li, Feng Liu, Jie Mao, Dan Yan, Yongqiang Zhao, Hongli Wang, Ping Li, and et al. Quantitative proteomics by swath-ms reveals sophisticated metabolic reprogramming in hepatocellular carcinoma tissues. Scientiﬁc Reports, 7:45913, May 2017. doi: 10.1038/srep45913.

[216] Itemobong Ekaidem, Itoro Usoh, and Friday Uboh. Phenytoin induced changes in glucose and lipid metabolism is related to increased urate synthesis. British Journal of Medicine and Medical Research, 16(7):110, Oct 2016. doi: 10.9734/bjmmr/2016/24851.

[217] P. V. Luoma, M. I. Reunanen, and E. A. Sotaniemi. Changes in serum triglyceride and cholesterol levels during long-term phenytoin treatment for epilepsy. Acta Medica Scandinavica, 206(1-6):229231, 1979. doi: 10.1111/j.0954-6820.1979.tb13500.x.

[218] A. Fadiel, J. Song, D. Tivon, A. Hamza, T. Cardozo, and Frederick Naftolin. Phenytoin is an estrogen receptor -selective modulator that interacts with helix 12. Reproductive Sciences, 22(2):146155, 2014. doi: 10.1177/1933719114549853.

[219] Rosenbrock, Hagemeyer, Singe, Knoth, and Volk. Testosterone metabolism in rat brain is differentially enhanced by phenytoin-inducible cytochrome p450 isoforms. Journal of Neuroendocrinology, 11(8):597604, 2001. doi: 10.1046/j.1365-2826.1999.00371.x.

[220] William Banner, David G. Johnson, Philip D. Walson, and Donald Jung. Effects of single large doses of phenytoin on glucose homeostasis-a preliminary report. The Journal of Clinical Pharmacology, 22(2-3):7981, Mar 1982. doi: 10.1002/j.1552-4604.1982. tb02652.x.

[221] David A Keith and Paul M Gallop. Phenytoin, hemorrhage, skeletal defects and vitamin k in the newborn. Medical Hypotheses, 5(12):13471351, 1979. doi: 10.1016/ 0306-9877(79)90105-1.

[222] Janee Gelineau-Van Waes, Gregory D. Bennett, and Richard H. Finnell. Phenytoin- induced alterations in craniofacial gene expression. Teratology, 59(1):2334, 1999. doi: 10.1002/(sici)1096-9926(199901)59:1h23::aid-tera7i3.0.co;2-m.

[223] Pavel Strnad, Stephan Paschke, Kwi-Hoon Jang, and Nam-On Ku. Keratins: Markers and modulators of liver disease. Current Opinion in Gastroenterology, 28(3):209216, 2012. doi: 10.1097/mog.0b013e3283525cb8. References 176

[224] James F. Trotter. Coagulation abnormalities in patients who have liver disease. Clinics in Liver Disease, 10(3):665678, 2006. doi: 10.1016/j.cld.2006.08.006.

[225] Alberto Verrotti, Giulia Loiacono, Melissa Laus, Giangennaro Coppola, Francesco Chiarelli, and Gian Mario Tiboni. Hormonal and reproductive disturbances in epileptic male patients: Emerging issues. Reproductive Toxicology, 31(4):519527, 2011. doi: 10.1016/j.reprotox.2011.02.002.

[226] Christine Vogel and Edward M. Marcotte. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nature Reviews Genetics, 13(4): 227232, 2012. doi: 10.1038/nrg3185.

[227] Lai Peng, Stephanie C. Piekos, Grace L. Guo, and Xiao-Bo Zhong. Role of farnesoid x receptor in the determination of liver transcriptome during postnatal maturation in mice. Nuclear Receptor Research, 4, 2017. doi: 10.11131/2017/101308.