Biomarkers of Bisphenol a Exposure and Metabolism

BIOMARKERS OF BISPHENOL A EXPOSURE AND METABOLISM

IN A POPULATION OF NEONATES IN BALTIMORE, MARYLAND

by Rebecca Massa Nachman, MPH

A dissertation submitted to Johns Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy

Baltimore, Maryland October, 2013

Abstract

Bisphenol A (BPA) exposure is widespread in the general population in countries around the world. Human health concerns arise from its estrogenic properties which may confer a range of health effects across the life span. Following ingestion, the main route of exposure, BPA undergoes first pass metabolism to BPA glucuronide, a rendering it biologically inactive. In newborns and young infants, glucuronidation capacity is low compared to that of adults and older children. The purpose of this research was to quantify the effect of low glucuronidation capacity on the internal dose of BPA in newborns and young infants.

First, a review of the literature on metabolism of BPA in newborns and infants was conducted. Human studies, animal studies and PBPK models were included in the review. Both free BPA and BPA glucuronide were detected frequently in the urine of premature infants in a NICU but less frequently in populations of infants age 1 month and older. BPA glucuronidation was less complete in orally dosed neonatal rats and mice compared with adults, but this effect was not observed in primates. PBPK models predicted elevated serum free BPA concentrations in human infants age 0-3 months.

Second, free BPA and BPA glucuronide were measured in the urine of 12 healthy, full-term neonates and young infants age at age 7-24 days using a high performance liquid chromatography with tandem mass spectrometry method that reduces contamination of samples with BPA from background sources. BPA exposure was confirmed in all 12 subjects, but free BPA was detected in only a single sample for which the replicate was negative.

Lastly, free BPA and BPA glucuronide were measured in urine from 44 neonates collected at two separate time points before and after the first week of life. BPA exposure was confirmed in 71% of the study population, but free BPA was not detected in any urine sample collected from neonates in either age group.

These data are the first human BPA biomarker measurements ever recorded for healthy full term neonates and fundamentally challenge our prior assumptions of the toxicology of BPA.

Dissertation Readers

Peter S. J. Lees, PhD (Advisor) Professor, Department of Environmental Health Sciences

Janet DiPietro, PhD (Chair) Professor, Department of Population, Family, and Reproductive Health

John D. Groopman, PhD Professor, Department of Environmental Health Sciences

W. Christopher Golden, MD Assistant Professor of Pediatrics, School of Medicine

iii

In memory of Dr. Alison S. Geyh

Preface

This dissertation is organized in manuscript format. First, the specific aims of the research and an overview of the study, which was conducted in two phases, are presented. The first chapter is a systematic review of the literature on Bisphenol A (BPA) metabolism in infants. In the second chapter are the published results of Phase I of the study, which include measured concentrations of two compounds, free BPA and BPA- glucuronide in urine collected from a cohort of neonates and young infants at the age of

1-6 weeks in Baltimore Maryland. In the third chapter, are the results of Phase II of the study, in which we report free BPA and BPA-glucuronide concentrations in urine collected at two separate time points from neonates age 3-28 days. The final chapter is a summarization and discussion of the findings of the study and its implications for public health and future research.

Acknowledgments

I would like to acknowledge several people who advised and mentored me throughout my time conducting this research. I am grateful to my advisor, Dr. Peter Lees whom I have been able to count on for advice at many points along the way to bringing this research to fruition. Also, I thank Dr. John Groopman, whose support and expertise were critical to making this research possible. I am thankful for the support of Dr. W.

Christopher Golden, Medical Director of the full term nursery at the Johns Hopkins

Hospital, who has been an engaged advisor throughout the many stages of this research.

Special thanks to Dr. Ana Navas-Acien who has generously taken the time on many occasions to discuss my research with me. I also would like to acknowledge Dr. Erica

Sibinga of the Harriet Lane Pediatric Clinic, whose support made this research possible, as well as Dr. Janet DiPietro, who has been a valuable member of my research committee. Most importantly, this study would not have been possible without the enthusiasm and dedication of the late Dr. Alison Geyh. Her enduring spirit has guided us all along the way.

Special thanks to Steve Fox, analytical chemist at the Frederick National

Laboratory for Cancer Research, whose contribution was a crucial component of this research. In addition, I would like to acknowledge the researchers in the Groopman lab who generously shared their time and expertise, including Patricia Egner, Marlin Friesen,

Jamie Johnson, Kevin Kensler, Natalie Johnson, and Patrick Dolan.

This research would not have been possible without the dedication of the pediatric nurse practitioners at the Johns Hopkins Hospital Full Term Nursery, who screened patients for eligibility: Christa Bay, Kristen Byrnes, Krista Kline, Carol Long, Krystina

Mints, Alice Rice, Suzanne Rubin, Joanne Schwartz, Patti Smouse, and Linda

Weisburger. I am also grateful to nurse manager Meghan D’Angelo and the triage nurses,

Sharon Adams, Ebony Edwards, Karon Harley and Roslyn Sample at the Harriet Lane

Pediatric Clinic. It was a pleasure to work with them.

I thank my husband, Keeve Nachman, for his unwavering support of my endeavors, as well as my inspirations, Maximilian Nachman and Ike Nachman. Also I am grateful to Dr. and Mrs. Dennis and Patricia Massa and all the Massa-Welches. My father, Dr. Massa, contributed to this research by tutoring me in the basics of BPA synthesis and its commercial applications. I thank the late Donna Massa, dedicated mother and nurse. I also thank Mr. and Mrs. Kenneth and Cheryl Nachman and Rosalie

Nachman. I would not be here without the support of all these family members.

As Center for a Livable Future Learner Fellow, I am grateful for funding provided by the center and for the guidance of Dr. Robert Lawrence, Shawn McKenzie, Dr. Roni

Neff, and Dave Love. In addition, I am grateful for funding received through the Wendy

Klag Memorial Award, the National Institutes of Environmental Health Sciences

Training Grant, the National Institute for Occupational Safety and Health, the Dr. C. W.

Krusé Memorial Fund, and the Department of Environmental Sciences student development fund.

A number of people have created an intellectually stimulating, fun and creative work environment. These include Jennifer Hartle, Jillian Fry, Patti Truant, Patrick Baron,

Jesse Berman, Beth Feingold, Talia Abott Chalew, Rachel Zamoiski, Rebecca Adler,

Pam Dopart, Ben Davis, Sut Soneja, Melissa Opryszko, Meghan Davis, Stacey Woods,

Chrissy Torrey, Nina Kulacki, Kay Castleberry, Ruth Quinn, and Ada Simari. Special

vii thanks to alumni of the Department of Environmental Health Sciences, Dr. Amy Sapkota and Dr. Amir Sapkota, for their advice over the years. Finally, thank you to the department Chair, Dr. Marsha Wills-Karp, for her dedication to students.

viii

Table of Contents

Abstract ...... ii

Preface...... v

Acknowledgments...... vi

Table of Contents ...... ix

List of Tables ...... xii

List of Figures ...... xiii

Abbreviations ...... xiv

INTRODUCTION AND SPECIFIC AIMS...... 1

CHAPTER 1: Early Life Metabolism of Bisphenol A: A Review of the Literature ...... 7

Abstract ...... 8

Introduction ...... 10

Methods...... 12

Human Biomarker Studies ...... 13

Toxicokinetic Studies In Animals ...... 17

PBPK Models...... 19

Discussion ...... 21

Summary of Findings ...... 21

Future Research Needs ...... 24

CHAPTER 2: Urinary Free Bisphenol A and Bisphenol A Glucuronide Concentrations in

Newborns ...... 30

Abstract ...... 32

Introduction ...... 32

Methods...... 33

Results ...... 34

Discussion ...... 35

CHAPTER 3: Urinary Free Bisphenol and Bisphenol A Glucuronide Concentrations in a

Population of Neonates in Baltimore, Maryland ...... 39

Abstract ...... 40

Introduction ...... 41

Methods...... 43

Population and Study Design...... 43

Research Visit 1: Participant Recruitment Protocol ...... 44

Research Visits 2 and 3: Collection of Urine Samples and Administration of

Questionnaires ...... 45

Laboratory Analysis ...... 47

Data Analysis ...... 48

Results ...... 52

Study Population Characteristics and Attrition ...... 52

Urinary Free BPA and BPA Glucuronide Concentrations ...... 53

BPA Intake Rates Back-Calculated from Urinary ug BPA/L in Urine based on

BPA Glucuronide Concentrations ...... 54

Results of Regression Analysis of Potential Predictors of BPA-Glucuronide

Concentrations ...... 55

Discussion ...... 57

Conclusion ...... 62

CHAPTER 4: Conclusions ...... 81

Summary of Findings ...... 81

Future Research and Implications ...... 83

Final Conclusions...... 85

REFERENCES ...... 87

Specific Aims ...... 87

Chapter 1 ...... 87

Chapter 2 ...... 91

Chapter 3 ...... 92

Conclusions ...... 96

APPENDIX I: Exclusion Criteria ...... 98

APPENDIX II: Sample Collection Form and Questionnaires ...... 100

APPENDIX III: Reducing Opportunities for Sample Contamination during Sample

Handling, Collection and Storage ...... 108

APPENDIX IV: Use of Pediatric Urine Collection Bags for the Collection of Urine

Samples from Neonates and Young Infants ...... 113

APPENDIX V: Quality Control Sample Preparation Procedure for the Measurement of

BPA and BPA Conjugates in Infant Urine ...... 116

CURRICULUM VITAE ...... 126

List of Tables

CHAPTER 1

Table 1. Measurements of Urinary Free BPA, BPA Glucuronide, and Total BPA in

Humans from Birth to One Year of Age...... 27

Table 2. Studies of Postnatal Metabolism in Animal Models...... 28

Table 3. PBPK models of Infant Internal Dose of BPA ...... 29

CHAPTER 2

Table 1. Urinary Concentrations of Free BPA and BPA-glucuronide for 12 Infants ...... 37

CHAPTER 3

Table 1. Population Characteristics...... 64

Table 2. Breast Milk and Formula Intake in the Study Population ...... 65

Table 3. Concentrations of Free BPA and BPA Glucuronide in the Urine

of 44 Neonates ...... 66

Table 4. Intake Rates in Newborns Back-Calculated from Urinary BPA-Glucuronide

Concentrations for 44 Neonates ...... 67

Table 5. Intake Rates Back-Calculated from Urinary BPA-Glucuronide Concentrations

for 39 Neonates Age 7 – 27 days ...... 68

Table 6. European Food Safety Authority (EFSA) BPA Intake Rates and Corresponding

Modeled Serum or Plasma Concentrations of Free BPA...... 69

Table 7. Association of Individual Factors with Urinary BPA Glucuronide

Concentration ...... 70

Table 8. Geometric Mean Urinary BPA Glucuronide Concentrations by Age Group

and Feeding Status ...... 71

xii

List of Figures

INTRODUCTION AND SPECIFIC AIMS

Figure 1. Phase 1 Sampling Design...... 6

Figure 2. Phase 2 Sampling Design ...... 6

CHAPTER 2

Figure 1. Representative HPLC-MS/MS Chromatogram of Neonatal Urine Sample ...... 38

CHAPTER 3

Figure 1. Representative Chromatogram for a Synthetic Urine Rinsate Blank...... 72

Figure 2. Chromatogram for Calibration Standard H , BPAG

and BPA at 0.1ng/ml (LOQ) ...... 73

Figure 3. Eligible Participants Screen, Enrolled, and Followed Up ...... 74

Figure 4. Age Distribution of Study Population at the Time of Sample Collection ...... 75

Figure 5. Urinary BPA glucuronide Concentrations in Neonates by Subject

at Visits 1 and 2 by Subject...... 76

Figure 6a. Representative chromatogram for urine sample from the study population

with a BPA glucuronide concentration of 0.8 ug/L ...... 77

Figure 6b. Representative chromatogram for urine sample from the study population

with a BPA glucuronide concentration of 8.6 ug/L ...... 78

Figure 7. Repeated Urinary BPA-glucuronide Concentrations in Neonates at Visit 1

(Age 2-6 Days) and Visit 2 (Age 7-28 Days) with Lines Connecting

Measurements from the Same Individual ...... 79

Figure 8. Urinary BPA Glucuronide Concentrations by Visit and Feeding Type ...... 80

xiii

Abbreviations

13C BPA Carbon-13 bisphenol A (stable isotope)

ANOVA Analysis of variance

BPA Bisphenol A

CI Confidence Interval

D6-BPA Deuterium 6 bisphenol A (stable isotope)

D16-BPA Deuterium 16 bisphenol A (stable isotope)

EFSA European Food Safety Authority

HPLC-MS/MS High performance liquid chromotography with tandem mass

spectrometry

HLC Harriet Lane Clinic

HLPC Harriet Lane Pediatric Clinic

JHH Johns Hopkins Hospital

LOD Limit of detection

LOQ Limit of quantitation

NICU Neonatal intensive care unit

NTP-CERHR National Toxicology Program Center for the Evaluation of Risks to

Human Reproduction

PBPK Physically based pharmacokinetic (model)

PC Polycarbonate

PND Postnatal day

PNP Pediatric nurse practitioner

QC Quality control

xiv

UGT Uridine diphosphate glucuronosyl transferase

INTRODUCTION AND SPECIFIC AIMS

Bisphenol A (BPA) is a high production volume chemical found in baby bottles, food can linings, plastic dishes, electronics, car parts, dental sealants and recycled paper.1

BPA interferes with the functions of the endocrine system by binding with receptors for endogenous hormones such as estrogen.1 The main route of human exposure to BPA is the diet due to leaching from BPA-containing food contact surfaces.2 Upon ingestion,

BPA undergoes metabolism in the liver to BPA glucuronide.3 A critical issue for the assessment of health risks from BPA exposure is the balance of unconjugated free BPA and its glucuronide conjugate. The presumptive estrogenic effects of BPA are only mediated by the availability of free BPA, with BPA glucuronide viewed as a biologically inactive derivative. The absence of quantitative data on BPA and its metabolites in human samples in the first year of life have hindered public health-based judgments of risk in this vulnerable subpopulation.1 The goal of this research was to determine whether the known deficiency in hepatic glucuronidation capacity at birth adversely impacts the conjugation and therefore detoxication of BPA in newborns and young infants. To achieve our research objective, a study was conducted in two phases, each with its own specific aims. The results of Phase 1 were used to develop the aims for Phase 2.

Phase 1: Optimization of sample collection, preparation and analysis methods and quantification biomarkers of BPA exposure and BPA metabolism in infants from the source population.

Aim 1a: Optimize sample collection, preparation and analysis to identify and quantify

BPA glucuronide and free BPA in urine collected from infants.

Aim 1b: Identify and quantify BPA glucuronide and free BPA in urine samples collected from a cohort of infants aged 1 to 6 weeks in Baltimore, MD.

Aim 1c: Utilize Phase 1 data to design Phase 2, a longitudinal study to investigate age- dependent changes in BPA metabolism in infants.

The Specific Aims for Phase 1 were designed to establish successful protocols in the field and laboratory and to generate data on BPA exposure and metabolism in our source population which could be used to plan subsequent research. It was determined that the samples would be analyzed by high performance liquid chromotography with tandem mass spectrometry according to a method recently published by colleagues from the Frederick National Laboratory for Cancer Research who agreed to be study collaborators.4 Phase 1 tested the hypothesis that both free BPA and BPA glucuronide would be detectable in the urine of urine and young infants age 1 – 6 weeks living in

Baltimore.

The first aim was optimization of sample collection protocols via testing of equipment used for the collection, handling, and storage of urine samples for the presence of BPA (Aim 1a).

Next, single spot urine samples were collected from a small cohort of approximately 10 neonates and young infants (Aim 1b). The subjects were recruited from

2 the Full Term Nursery at the Johns Hopkins Hospital. Eligible patients were screened by pediatric nurse practitioners (PNPs) from the nursery, who relayed the names to investigators so that mothers of newborns who met the inclusion criteria could be approached in their hospital rooms for recruitment during their postpartum hospital stay.

The exclusion criteria are listed in Appendix I. We chose to restrict recruitment to patients who would receive follow-up pediatric care at the Harriet Lane Clinic (HLC) located at the Johns Hopkins Hospital. Samples were collected during the participants’ pediatric well-visits at the clinic occurring between the ages of 1 and 6 weeks. (Figure 1)

This age range was chosen because it was young enough that we expected glucuronidation to be limited, but old enough for circulating BPA glucuronide from in utero exposure would have been cleared resulting in measurements reflecting postnatal exposure and metabolism. The proximity of the sites where recruitment, sample collection and laboratory processing of samples for storage would take place enabled these activities to be conducted concurrently by a single individual, the student investigator.

For the final Aim of Phase 1, (Aim 1c) a second stage of the study was planned after an analysis of the data from Phase 1, which included data from a brief questionnaire

(Appendix II).

Phase 2: Investigation of the determinants of BPA exposure and metabolism in infants using measurements of free BPA and BPA glucuronide.

Aim 2a: Quantify concentrations of free BPA and BPA glucuronide in urine collected from neonates at age 2-6 days.

Aim 2b: Quantify concentrations of free BPA and BPA glucuronide in urine collected from the same neonates in Aim 2a, this time at age 7 – 28 days.

Aim 2c: Using data on free BPA and BPA glucuronide concentrations in urine from neonates, and data from questionnaires, assess potential individual determinants of BPA exposure in neonates.

Aim 2d: Back-calculate BPA intake rates using measured concentrations of free BPA and

BPA glucuronide in urine of the study population.

In Phase 2, the study sample size was expanded with the goal of recruiting approximately 50 newborns, and repeated samples were collected from the same individual at two time points in the neonatal period, first at age 2-6 days, and again at age

7-10 days (Aims 2a and 2b). We hypothesized that BPA glucuronide would be detectable in the urine of neonates at both ages, but free BPA would only be detectable in the urine of neonates in the first week of life reflecting limited glucuronidation capacity that increases by the second week of life such that free BPA is no longer detected in urine in the second and subsequent weeks of the neonatal period.

The target age for collection of the first sample was 3-5 days, when serum bilirubin concentrations usually peak, indicating lower hepatic glucuronidation capacity

4 and possible competition for or induction of glucuronidation enzymes. Newborns who receive pediatric care at the HLC usually have their first appointment between the ages of

3 and 6 days, most often on day 4, facilitating collection at the age of interest. The second sample collection took place between ages 7-28 days, a period which also coincides with a neonatal well-visit. The same screening, recruitment, and sampling protocols that were used in Phase 1 were also used in Phase 2 with a few minor changes. As in Phase 1, names of eligible participants were provided by nurse practitioners, but in addition, patients were asked in advance whether they would be willing to speak with investigators and allow their medical information to be shared. For those who enrolled in the study and were consented, the date and time of birth, serum bilirubin concentrations, and compatibility between the maternal and newborn’s blood type were provided to investigators for use along with data from a questionnaire administered at the time of sample collection (Appendix II) in the identification of factors associated with BPA exposure and metabolism (Aim 2c).

Figure 1. Phase 1 Sampling Design: Recruitment

Sampling Period

0 1 2 3 4 5 6 Age (Weeks)

Figure 2. Phase 2 Sampling Design: Recruitment

Sample Sample 2 1 0 1 2 3 4 5 6 Age (Weeks)

CHAPTER 1:

Early Life Metabolism of Bisphenol A: A Review of the Literature

Rebecca Nachman,1 Peter S. J. Lees,1 John D. Groopman1

1Department of Environmental Health Sciences, Johns Hopkins University Bloomberg

School of Public Health, Baltimore, MD

Address Correspondence to:

Rebecca Nachman, MPH

Johns Hopkins Bloomberg School of Public Health

615 N. Wolfe Street, Room E7402

Baltimore, Maryland 21205

Email: [email protected]

Sources of Funding: This research was supported in part with a grant from The Johns

Hopkins Center for a Livable Future, the Wendy Klag Memorial Fund, and by grants from the NIH (P01 ES006052, P30 ES003819, and training grant T32 ES007141). The authors declare no conflicts of interest.

Abstract

In the 2008 report on the reproductive and developmental toxicity of BPA, the National

Toxicology Program Center for the Evaluation of Risks to Human Reproduction expert panel highlighted the need for more research on role of metabolism in determining internal dose in the newborn. At the time the report was published, diet had been established as the main source of BPA exposure and glucuronidation was known to be the main pathway by which BPA was deactivated and cleared in urine. However, little data were available to assess the extent to which poor glucuronidation capacity in newborns impacts internal dose in this vulnerable population. In this paper, the evidence on this topic that has emerged since the report was compiled. Studies falling into three categories were included: studies reporting both free BPA and either conjugated BPA or Total BPA in biologic fluids from children age 0-5 years; 2) studies of neonatal toxicokinetics using animal models; and 3) physically based pharmacokinetic models of internal BPA dose in infants and young children. Human urinary biomarker data from infants published in

2008 demonstrated that infants are widely exposed to BPA and capable of efficient glucuronidation of BPA. However, contamination of urine samples with BPA from field and laboratory settings was a major limitation in human biomarker studies. In toxicokinetic studies of BPA conducted in animals, neonatal BPA glucuronidation was found to be limited in rodents, but not non-human primates. Studies in which internal dose of BPA was estimated using PBPK models were highly sensitive to intake rates and toxicokinetic parameters inputted in the model. Uncertainty in PBPK models stemmed from lack of biomarker-derived neonatal intake rates for neonates and the use of

toxicokinetic data based on studies of morphine, which may not be an appropriate model of neonatal BPA toxicokinetics.

Introduction

Bisphenol A (BPA) is a high production volume chemical used in the synthesis of polycarbonate plastics and epoxy resins, materials with widespread commercial applications that include drink and food containers, soda and food can liners, electronics and car parts, and dental sealants.12 BPA is also present as a thin coating on the surface of thermal printing papers where it serves as an ink developing agent, and as a result is widely found on cash register receipts and paper with a high recycled paper content.3

Widespread BPA exposure in the general population has been documented in several countries and regions of the world with detection in human urine exceeding 90% in representative samples of the population.4-7 Exposure is thought to occur mainly through the diet due to its presence in food packaging. Human health concerns arise from its estrogenic properties, which can potentially confer a range of health effects across the life span, including developmental and reproductive effects, diabetes and metabolic diseases, obesity, cancer of the reproductive organs, and heart disease.8-11

Early life exposures to endocrine disruptors such as BPA are of particular concern due to their potential to impact key stages of rapid growth and development that occur in utero, postnatally, and throughout childhood and adolescence.12 Although much research has focused on prenatal exposure to BPA, infancy and childhood may present exposure scenarios with implications for health that are unique to those life stages, such as disruption of estrogen- and testosterone-driven synaptogenesis which peaks in early childhood.13

Urinary biomarker data demonstrate that BPA body burden and is highest in children.14 Limited biomarker data exist in children under the age of 6 years, but

estimates of weight-normalized intake rates in young children based on source aggregation suggest that the highest levels of exposure may occur in infants,15 and diet is the main known contributor to exposure.16 BPA may be present in infant formula and baby food due to leaching from packaging or bottles, or can be indirectly passed by the mother through breast milk.

Upon ingestion of BPA, the biologically active parent molecule undergoes rapid first pass metabolism to BPA glucuronide, which is biologically inert and rapidly cleared in the urine.17 The formation of BPA glucuronide is driven by uridine diphosphate glucuronosyltransferases (UGTs) in the liver and gut. Enzyme activity for hepatic UGTs is limited at birth, a condition that manifests clinically as jaundice in the newborn. UGT enzyme activity increases with age until it reaches adult levels, which occurs between the ages of 3-6 months and 10 years, depending on the UGT isoform.18 Since glucuronidation is the main detoxication pathway for BPA in humans and other species, an understanding of the impact of low infant hepatic UGT activity on BPA metabolism is key to evaluating the potential health risks from early life exposure to BPA, especially in the neonatal period and infancy.

In this paper, we have reviewed studies of BPA metabolism in infants published in the peer-reviewed literature since the release of a comprehensive monograph by the

National Toxicology Program Center for the Evaluation of Risks to Human Reproduction

(NTP-CERHR).1 We present conclusions based on newly available evidence since the report and identify current research gaps.

Methods

Articles published between June 2008 and August 2013 that were catalogued in

PubMed were included. The NTP-CERHR monograph on Bisphenol A provides a thorough review of the literature through the middle of 2008 on human toxicokinetics of

BPA. To identify literature relevant to postnatal and early life toxicokinetics published since 2008, a search on the term “Bisphenol” was performed, and titles and abstracts were reviewed to identify publications of original research falling into three categories: 1) studies reporting both free BPA and either conjugated BPA or Total BPA in biologic fluids from children age 0-5 years; 2) studies of neonatal toxicokinetics using animal models; and 3) physically based pharmacokinetic (PBPK) models of internal BPA dose in infants and young children. Publications reporting measurements of BPA in human biological samples were included even if the studies were not designed specifically to investigate toxicokinetics. An upper limit of age five was chosen because enzyme activity develops rapidly and usually reaches adult levels by this age.19 In order to ensure adequate laboratory method specificity, only studies in which analytes were quantified by high (or ultra) performance liquid chromatography or gas chromatography coupled with mass spectrometry were considered; studies in which measurements made using enzyme- linked immunosorbent assay, or electrochemical and ultraviolet detection were excluded because they are not adequately specific and may overestimate concentrations of BPA in samples.20 In addition, human studies were limited to those in which BPA was measured in urine, since serum BPA concentrations in children in the general population are too low to be detected.21 Only studies relevant to early life toxicokinetics following birth were included; studies of cord blood, amniotic fluid, or fluids collected from mothers

during pregnancy, which better reflect prenatal exposure and metabolism, were not included in this review.

Human Biomarker Studies

Inadvertent contamination of biologic samples from sources of BPA in both field and laboratory settings significantly complicates the determination of metabolic profiles for BPA in biological fluids and tissues.22 BPA contamination from background sources is indistinguishable from free BPA eliminated in the urine following absorption, metabolism, and distribution in the body; background concentrations also tend to be highly variable, preventing the quantification of a mean background level with which to blank-correct samples. To circumvent the issue of sample contamination in a study of

BPA toxicokinetics in adults, volunteers were administered d16-BPA, a stable isotope that behaves like BPA the body, but is distinguishable from background BPA during laboratory analysis.17 A study of this design, which yields toxicokinetic parameters for use in modeling internal dose concentrations of the active parent molecule given various exposure scenarios, is not possible to conduct in children for obvious ethical reasons.

However, despite limitations, measurements of free BPA and BPA conjugates in urine of infants and young children who are exposed through everyday activities can nonetheless be useful in exploring changes in the BPA metabolic profile that occur at an early age.

We identified four studies of infant populations in which free BPA and Total BPA or

BPA glucuronide in urine were separately quantified; these are presented in Table 1.

The first infant urinary BPA concentrations were reported in 2009 by Calafat et al. for a population of 41 premature infants in a neonatal intensive care unit (NICU).23

Urine was collected from cotton balls and cotton diaper filling. Polyvinyl chloride bags and tubing used in NICU settings are a known source of BPA exposure,23 and urinary

Total BPA concentrations in the study population were higher than any of the concentrations since reported for infants in the general population. The median concentration of 28.6 ug/mL is an order of magnitude higher than the median in the U.S. general population age 6 years and greater.24 The maximum Total BPA concentration of

946 ug/mL exceeded urinary concentrations measured in an occupationally exposed population.25 The detection frequencies for both free BPA and Total BPA were 92% and

100%, respectively, but free BPA concentrations were much lower compared with Total

BPA concentrations (median = 1.7 ug/mL), demonstrating that the infants were capable of a significant level of BPA conjugation. The exact ages for infants in the study were not available, but investigators conducted a visual age assessment and only infants no more than 44 weeks corrected gestational age (gestational age + weeks since birth) were included in the study. Free BPA concentrations correlated closely with Total BPA concentrations (r=0.86). On the one hand, the correlation between free and Total BPA concentrations in the NICU study could indicate that a consistent percent of free BPA went unconjugated such that the infants with the highest Total BPA concentrations also had the highest free BPA concentrations. However, this correlation could be a result of

BPA contamination; in other words the Total BPA concentrations are high because of contamination with BPA.

Urinary free and Total or glucuronidated BPA concentrations have been measured in three different healthy infant populations.26-28 In a population of full-term infants age 1-

5 months in Germany, free BPA was detected in 9% of 91 samples from 47 infants.26

Total BPA was detected in 66% of the population demonstrating exposure was prevalent in the study population. Medians of both analytes were under the limit of quantification

(LOQ). Urine was collected using polyethylene urine collection bags, which would not be expected to contain BPA. This study provided the first urinary BPA measurements in a healthy infant population. Given the proximity of Total BPA concentrations to the method LOQ, unless glucuronidation were severely impaired, free BPA concentrations, being a fraction of the Total, would be expected to be below the LOQ. The authors noted that no free BPA was detected in five samples from one subject whose high Total BPA concentration was an outlier. Introduction of BPA to the samples during sample handling may have been responsible for the detection of free BPA in a small percent of the samples in this study.

Mendonca et al. measured free and Total BPA in the urine of 29 healthy infants age 2 – 15 months of age in Boston in a study investigating the contribution of breast milk to BPA exposure in infants.27 It was not reported whether infants in the study were full term at birth. The median Total BPA concentration was 1.8 ug/mL. The detection frequency for free BPA was notably high (28%), although the median free BPA concentration was below the limit of detection. The authors report that the small sample size and low free BPA detection frequency prevented the evaluation of free BPA as an indicator of metabolic capacity of the study population and that the wood pulp and cotton diapers used for sample collection are possible but unconfirmed sources of BPA contamination. Free BPA detected in urine from one infant with a high urinary Total

BPA concentration of 89 ug/mL could indicate incomplete glucuronidation of BPA in that subject, although this finding would be inconsistent with a report of no free BPA in

an infant with an unusually high urinary Total BPA concentration of 17.35 ug/mL in the study by Volkel et al.26

Nachman et al. measured free BPA and BPA glucuronide in urine collected from

12 healthy full-term infants age 1 – 6 weeks in Baltimore, Maryland.28 BPA glucuronide was detected in all samples (median = 0.66 ug/L), confirming exposure to BPA in all 12 infants. Free BPA was not detected in any sample, except for one, for which the replicate was a nondetect. The lack of detection of free BPA despite confirmation of BPA exposure in all infants suggests efficient conjugation of BPA in infants, specifically by the glucuronidation pathway, at a very young age. Given the small sample size, these results should be interpreted with caution as there may be variability in the capacity to conjugate BPA in this age group and the study population may have inadvertently included only those infants with higher glucuronidation enzyme activities. Unique features of the laboratory analysis method, direct measurement of BPA glucuronide and derivatization with dansyl chloride, may have contributed to the low incidence of sample contamination. This study did not report Total BPA, and thus any BPA sulfate that might have been present in the urine is not accounted for.

In summary, few studies have been published reporting urinary BPA concentrations in infants, and all have low sample sizes. Of the four that exist, all report concentrations of both free BPA and either Total BPA or BPA glucuronide and provide some insight into ability of humans to conjugate BPA in infancy. In all studies, a significant degree glucuronidation was found to occur in infants. High detection frequency of free BPA and correlation of free BPA concentrations with Total BPA concentrations in the urine of premature infants in a NICU demonstrated that though

BPA conjugation occurs in this sensitive subpopulation, it may be less efficient than it is in full term or older infants. The youngest infant of known age studied was 1 week old, thus the results of these studies may not apply in the first week of life, which could be a critical period in terms of both impaired metabolic capacity and developmental susceptibility. When Total BPA concentrations are close to the LOQ, free BPA may not be detectable even under conditions of impaired BPA glucuronidation. Detection of free

BPA among infants with higher than average exposures was inconsistent between studies.

Toxicokinetic Studies in Animals

Neonatal BPA toxicokinetics for oral and subcutaneous routes of exposure were studied by Doerge and colleagues through the administration of 100 ug/kg doses of d6-

BPA in a variety of animal models.29-31 (Table 2) The dose of 100 ug/kg was considered by the authors to be close enough to the range of human exposure to be relevant to the general population, and also high enough to be detectable in serum enabling the determination of toxicokinetic parameters. High performance liquid chromatography

(HPLC-MS/MS) with 13C-BPA as an internal standard was used to identify and quantify both free and Total d6-BPA. The use of d6 BPA for dosing solutions eliminated any possible influence of background contamination on the measurements.

Doerge et al. reported a significant effect of age on serum unconjugated BPA concentrations at the Cmax in neonatal CD-1 mice SD and rats, evidence that early life impairment of BPA conjugation impacts internal BPA dose in rodents.29, 30 In both mice and rats, unconjugated BPA concentrations decreased with age and were not statistically different from those observed in adult animals by age postnatal day (PND) 21. In

primates, considered to be a more appropriate model for human BPA toxicokinetics than rodents, age-related differences in BPA conjugation were not statistically significant when comparing PND 5, 35, 70, or adult.31 At all postnatal ages, free BPA concentrations in primates (based on the Cmax and area under the curve) accounted for less than 1% of

Total BPA concentrations in serum. Interestingly, the percent serum BPA concentrations at the Cmax in 4 adult female monkeys (29 ± 19% free BPA) were higher than they were in the younger ages groups, suggesting that interindividual variability in BPA glucuronidation may exist independent of age.

In the three studies described, reduced metabolic capacity in the neonatal period was observed to impact internal dose of BPA in rodents, but not primates. The 100 ug/kg dose in these experiments corresponds to about the same dose administered to adult human volunteers studied by Volkel et al.,17 five times the estimated 95th percentile daily intake for 6-11 year olds based on urinary concentrations in the United States (assuming a body weight of 40 kg),14 and twice the reference dose of 50 ug/kg for BPA.32 The null finding in neonatal rhesus monkeys at this dose suggests that contrary to previous thinking, differences between neonatal and adult metabolism may not significantly impact internal dosimetry of BPA following oral exposure. While data from rhesus monkeys would be expected to be the most relevant to humans of the three animal models explored, the possibility of species differences between humans and monkeys cannot be dismissed.

PBPK Models

PBPK models of internal dose in infants and children combine age-dependent toxicokinetic parameters, external exposure data, and allometric scaling of other physiologic features to calculate serum concentrations of the substance of interest. One influential input in a PBPK model estimating early life internal free BPA concentrations is the age-dependent rate at which BPA undergoes glucuronidation in the liver. Given estimates that the oral route accounts for as much as 99% of BPA exposure,16 first pass metabolism of BPA plays an important role in determining whether BPA enters the blood compartment as either the active (unconjugated) or inactive (conjugated) form.

Toxicokinetic parameters for glucuronidation vary greatly by substance and by age. By way of example, morphine and acetaminophen, the metabolism of which are well studied in infants and young children in clinical settings, both undergo glucuronidation, but because they have different affinities for different UGT isoforms, and those isoforms in turn have different developmental timelines, hepatic metabolism of morphine is estimated to occur at adult levels by age 1 year, compared with by age 10 years for acetaminophen.33

Two PBPK models have been published estimating serum free BPA concentrations at early ages. Mielke et al. reported estimated steady state as well as peak blood concentrations for free BPA for several age groups ranging from birth to age 4.5 years.34 (Table 3) Steady state concentration, a measure of internal dose, represents the average serum free BPA concentration over time. The peak concentration on the other hand is analogous to the Cmax in a controlled toxicokinetic study, and is an appropriate metric for BPA internal dose since most BPA exposure is thought to occur episodically

with food intake, resulting in blood concentrations that fluctuate with time. Intake rates from two European Union reports35 were applied as inputs in the model along with a number of age-appropriate physiologic and toxicokinetic parameters. Mielke et al. estimated that the highest peak serum concentration would be 0.34 ng/mL in newborns and infants up to 3 months of age who were fed from polycarbonate bottles. For infants fed from non-polycarbonate bottles, a lower estimated peak of 0.071 ng/mL was reported.

Steady state concentrations reported by Mielke et al. are presented in Table 3. This model is unique in its inclusion of both sulfation and glucuronidation as presystemic metabolic pathways.

In a second study, Edginton and Ritter estimated steady state serum concentrations of free BPA and BPA glucuronide with a model that reflects reduced glucuronidation capacity in the 0-3 months age group followed by rapid maturation of this metabolic pathway.36 When age-appropriate external exposure estimates were used as inputs in the model, the resulting steady state serum free BPA concentrations (presented only in graph form) were similar to those reported by Mielke et al. except for in the newborn age group, for which Edginton and Ritter assumed a lower external exposure compared with that chosen by Mielke et al. Peak serum concentrations of free or conjugated BPA were not estimated by Edginton and Ritter.

The PBPK models presented here produced quantitative data on serum free and conjugated BPA levels in humans under age 5 years that would not otherwise not be available for this age group. A number of age-specific human physiologic parameters were included in the models. Significant uncertainty in both models arises from the use of toxicokinetic parameters derived from clinically obtained data on early life metabolism of

morphine, which is primarily glucuronidated by the isoform UGT2B7. UGT2B15 is known to play a much larger role in glucuronidation of BPA than UGT2B7 does;37 however comparable age-specific toxicokinetic data for UGT2B15 are not available.

Discussion

Summary of the Findings

Although not a large body of literature, the publication of 4 human biomarker studies in infant populations, 3 animal studies in neonatal mice, rats and monkeys, and 2

PBPK models of age-dependent internal BPA dosimetry represent significant progress and an increase in data availability relating to neonatal BPA metabolism in the last 5 years. In 2008, only two studies of neonatal metabolism in rats, one in vitro38 and the other in vivo39 represented the entirety of literature on this topic. The NTP-CERHR specifically identified the need for more measurements of free and conjugated BPA in neonatal biological samples as well as the need for PBPK models of early life internal dose in their list of 10 Critical Data Needs for the evaluation of BPA toxicity.1

Three types of studies were considered in this review, and the findings differed by study type. BPA glucuronide accounted for most or all of the BPA measured in human urine from four different infant populations, which both confirmed exposure in this population and demonstrated substantial capacity to metabolize BPA at the levels at which the infants were exposed. However, detection of free BPA in 92% of urine samples in a population of premature infants in a NICU indicates that prematurity and high exposure may be factors that mediate the impact of neonatal toxicokinetics on internal

BPA dose. In contrast, no significant age-related differences in BPA glucuronidation

were found in a study of neonatal and adult monkeys suggesting that reduced neonatal conjugation capacity may not significantly alter BPA internal dose in human neonates.

PBPK models predicted that reduced glucuronidation capacity increases the bioavailability of free BPA in neonates and young infants under age 3 months. Each study type has a unique set of limitations, some of which can be addressed in future research.

Inconsistency among biomarker studies may be due to differences in exposure levels or developmental stages of the study populations. In the case of the study of premature infants in a NICU, urinary concentrations were in some cases 2-3 orders of magnitude higher than measurements reported in other infant studies, reflecting higher exposures in a NICU setting. It is possible that in a high exposure scenario, free BPA may be more likely to be detected and that it could reflect low conjugation capacity that is not detectable in lower exposure scenarios. Also, impaired glucuronidation capacity may be more pronounced or more widespread in premature infants compared with those carried to term. Inconsistency between the findings of Mendonca et al. compared with the other two studies in which free BPA was not detected, or detected infrequently is not easily explained. The low sample sizes and different population demographics may account for differences. However, inadvertent entry of BPA into samples during handling is always a consideration when interpreting free BPA concentrations in any biological samples. BPA glucuronide would not be expected to be found outside the body and thus contamination should not affect levels of BPA glucuronide in biological samples.

Studies of neonatal monkeys provide the strongest evidence to date that neonatal glucuronidation may be close to adult levels within a short time of birth. Serum free BPA

levels were <1% of Total BPA in neonatal monkeys at PND 5, even at peak levels

(Cmax) following an oral dose of 100 ug/mL. On the other hand, 29 ± 19% free BPA accounted for Total BPA concentrations in adult monkeys in the same study, raising the possibility of interindividual variability that was not captured by the small sample size.

The use of d6-BPA for dosing in all animals studies presented in this review adds considerable validity to these studies since it eliminates the possibility that BPA contamination during handling could be mistaken for BPA that has passed unmetabolized into the general blood compartment. However, as with any animal study, interspecies extrapolation is a limitation. Biological differences between monkeys and humans, such as differences in UGT isoform sequences or UGT ontogeny may exist such that BPA conjugation occurs at a lower rate or develops at a later age in humans than in other primates.

Consistency of the PBPK models with each other and with toxicokinetic data from an intentionally exposed group of adult volunteers supports the validity of the models. Given the same weight-normalized intake of BPA, steady state serum free BPA concentrations were predicted to be higher in infants age 0-3 months compared with adults by a factor of 3 in the model by Mielke el al., and a factor of 11 in the model by

Edginton and Ritter. However, it should be noted that both models relied on studies of

UGT2B7 ontogeny, not UGT2B15 for inputs of toxicokinetic parameters. While

UGT2B7 may contribute to BPA conjugation, UGT2B15 enzyme activity was about 3 times higher than UGT2B7 activity in an in vitro model testing several recombinant UGT isoforms.37 Also, a significant limitation of PBPK models is the scarcity of urine biomarker data in infants from which to estimate exposures. Exposure estimates in the

model were based on source aggregations which are much more uncertain that estimates from urine biomarkers.

The internal dose of BPA in neonates, infants, and young children, remains an important factor in estimates of health risk from BPA exposure early in life. The quality of internal dose estimates obtained via PBPK models depends on the quality of pharmacokinetic data available to be used as inputs in the model. Human biomarker studies and animal toxicokinetic studies provide context in which to consider the validity of PBPK models. We make the following recommendations to address data gaps and reduce uncertainty in estimates of internal dose in this age group.

Future Research Needs

1. More measurements of urinary free BPA and Total BPA are needed in this age

group and from larger and more diverse study populations. The contribution of

these measurements is twofold. First, back-calculation of intake rates from urinary

Total BPA concentrations would provide better estimates of exposure in this age

group. The contrast between the estimates for the 0-3 month age group in the two

PBPK studies reviewed here demonstrate the strong influence of external

exposure on modeled internal dose. Secondly, separate quantification of free BPA

in urine is an important indicator of presystemic metabolism in infant populations

under real life conditions. With proper quality control measures to ensure that

samples are not contaminated during collection, handling, and storage, these data

have the potential to demonstrate age-dependent variations in the capacity for

BPA conjugation.

2. Separate quantification of BPA glucuronide and BPA sulfate concentrations in

early life urine samples are needed to explore the potential role of sulfation, which

may contribute more to BPA metabolism at early ages than it does in adults. BPA

sulfate can undergo enzymatic hydrolysis in the presence of B-glucuornidase such

that typically measurements of Total BPA by enzymatic hydrolysis potentially

include a combination of free BPA and BPA glucuronide, as well as BPA sulfate.

3. More study of UGT isoforms responsible for BPA metabolism in human and

nonhuman primates is needed to better gauge the validity of extrapolation of BPA

toxicokinetic data from monkeys to humans.

4. In vitro studies that better characterize the ontogeny and interindividual variability

of glucuronidation by the UGT2B15 isoform are needed. These would improve

scaling of toxicokinetic parameters in PBPK models of internal dose in infants

and young children.

Conjugation of BPA via glucuronidation (and potentially sulfation) is an important detoxication pathway that may be impaired at birth and for a period of time in early human development. Understanding the timing and extent of impairment will improve assessments of human health risk in this vulnerable subpopulation. However, it should be noted that even given rapid and efficient first pass metabolism of BPA, biologically active BPA may still potentially reach target tissues either via bypassing presystemic conjugation as plasma-bound BPA or through deconjugation at target tissues. In addition, although BPA exposures have recently decreased in the general population,40 this trend is likely due in part to the introduction of replacements, some of

which are similar in structure to BPA and may act by a similar mode of action raising the possibilities of other environmental exposures of concern in the general population and/or possible cumulative effects of exposure to several related phenolic compounds including BPA.

This paper is a review of literature on a topic that has generated passionate debate among researchers in this field. The study of BPA toxicokinetics cannot on its own provide an answer regarding the health risks associated with BPA exposure, but it is a crucial component of any evaluation of the toxicity of this ubiquitous compound.

Table 1. Measurements of Urinary Free BPA, BPA Glucuronide, and Total BPA in Humans from Birth to One Year of Age

Authors and Country Population N LOD Free BPA* BPA- Total BPA* Detection Publication Age (ug/L) glucuronide* (ug/L) Frequency Year (ug/L) (free/Total) Volkel et al. Germany 1 – 5 months 47 0.15 ug/L** Data not Not meas Median < 9%/66% 201126 (Munich) presented LOD

Calafat et al. U.S. Premature, ≤ 41 0.4 ug/L GM: 1.8, Not meas GM: 30.3, 92%/100% 200923 (Boston) 44 weeks Median: 1.7 Median: 28.6 corrected gestational age Mendonca et U.S. 2 – 15 29 0.4 ug/L Mean: 0.5, Not meas Mean: 6.0, 28%/93% al. 201227 (Boston) months GM < LOD, GM: 2.3, Median < LOD Median: 1.8

Nachman et U.S. 1 – 6 weeks 12 0.02 ug/L† < LOD Mean: 0.87, Not meas 0%/100% al. 201328 (Baltimore) Median: 0.66

* Mean, geometric mean (GM) and median if reported. ** LOQ = 0.45ug/L † LOQ = 0.1 ug/L

Table 2. Studies of Postnatal Metabolism in Animal Models

Authors Animal N PND Dose Dosing Biological Laboratory Analytes % free and Model age at (ug/kg) Route fluids Analysis BPA* Publication dosing analyzed method Year Doerge et al. CD-1 72 pups (12 PND 100 Oral serum LC- Free and 23% (per 2011. 29 mouse per 3, 10, and SC EC/MS/MS Total Cmax), 3% age/route and 21 BPA (per AUC) group)

Doerge et al. SD rat 72 pups (12 PND 100 Oral serum LC- Free and 6.6% (per 2010 30 per 3, 10, and SC EC/MS/MS Total Cmax), age/route and 21 BPA 1.4% (per group) AUC)

Doerge et al. Rhesus 4 adult PND 100 Oral serum, LC- Free and <1% (per 2010 31 monkey female, 6 5,‡ 35, and IV urine, EC/MS/MS Total Cmax and neonatal (3 70 feces BPA AUC) female, 3 (oral); male) PND 77 (IV)

* Denominator: Total BPA. Data for lowest age group, orally dosed is presented

Table 3. PBPK Models of Infant Internal Dose of BPA

Authors and Published Sources of Age Groups Free BPA Steady Free BPA Peak Serum Bioavailability* Publication Model Inputs* Modeled State Serum Concentration (ng/mL) Year Concentration (ng/mL) Mielke and Intake Rates: Newborn 0.096 (pc bottle) 0.34 (pc bottle) Not reported X. 200934 EFSA 200635; EU 200841 3 months 0.0008 (breast), 0.020 0.0033 (breast), 0.071 (non-pc bottle) 0.096 (non-pc bottle) 0.34 (pc TK Parameters: (pc bottle) bottle) Kuester and Sipes 200742 6 months 0.030 0.17 6-12 months, 0.0085 0.049 TK Scaling factors: 1.5 years 0.010 0.059 Edginton and Ritter 1.5-4.5 years 0.012 0.069 200633 Adult 0.004 0.023

Edginton and Intake Rates: Newborn Presented in graph. Not modeled. 88% Ritter 200936 EFSA 200635; Ye et al. 3 months Exact values not 48% 200843 6 months reported. 32% 1.5 years 23% TK Parameters: adult 18% Volket et al. 200217

TK Scaling factors: Edginton and Ritter 200633

* TK Parameters=toxicokinetic parameters for adults, TK scaling factors=fraction of adult enzyme activity based on age of infant or child ** Fraction absorbed orally compared with IV absorption

CHAPTER 2:

Urinary Free Bisphenol A and Bisphenol A Glucuronide Concentrations in

Newborns

Rebecca M. Nachman, MPH1, Stephen D. Fox, BS2, W. Christopher Golden, MD3, Erica

Sibinga, MD3, Timothy D. Veenstra, PhD2, John D. Groopman, PhD1, Peter S. J. Lees,

PhD1

1Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of

Public Health, Baltimore, MD

2Laboratory of Proteomics, Advanced Technology Program, SAIC-Frederick, Inc.,

Frederick National Laboratory for Cancer Research, Frederick, MD

3Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, MD

Published April 2013 in Journal of Pediatrics. (J Pediatr 162(4):870-2.)

Corresponding Author/Reprint Request Author:

Rebecca Nachman, MPH

Department of Environmental Health Sciences

Johns Hopkins Bloomberg School of Public Health

615 N Wolfe Street, Room E7402

Baltimore, MD 21205 [email protected]

Tel: 443-803-6026 / Fax: 410-955-0617

Key words: Neonate, Biomarker, Endocrine Disruptor, Environment, Exposure

Short title: Free Bisphenol A, Bisphenol A Glucuronide in Newborns

Sources of Funding: This study was supported in part with a grant from The Johns

Hopkins Center for a Livable Future and by grants from the NIH (P01 ES006052, P30

ES003819, Contract N01-CO-12400, and training grant T32 ES007141). Neither sponsor had a role in the study design, collection, analysis, or interpretation of data, the writing of the report or the decision to submit the paper for publication. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States Government. The authors declare no conflicts of interest. No payment was received by anyone to produce the manuscript.

Acknowledgments

We would like to thank Pamela Donohue, ScD (Department of Pediatrics) Johns Hopkins

University School of Medicine for her contribution to study concept and design. We also thank the Pediatric Nurse Practitioners in the Full Term Nursery at the Johns Hopkins

Hospital for screening patients for eligibility.

Abstract

Infants are exposed to the endocrine disruptor bisphenol A (BPA) through breast milk and baby formula. The conjugation of BPA, a detoxication pathway, may be limited in infants. We demonstrate BPA exposure in 11 neonates and 1 young infant, but find no evidence of a low capacity for BPA conjugation.

Introduction

An estimated 5-6 billion pounds of the hormonally active compound Bisphenol A

(BPA) are produced globally per year.1 Human exposure is virtually universal, with over

90% of the United States population over age 5 exhibiting detectable urinary concentrations of BPA.2 Human health concerns arise from the estrogen mimetic properties of BPA that can confer a variety of health impacts across the lifespan. Prior analyses have demonstrated that BPA exposure can occur in neonates via maternal sources (including breast milk) as well as leaching from packaging of liquid formula or polycarbonate baby bottles.3, 4

BPA is detoxicated to its inactive form primarily via glucuronidation and rapidly

5 excreted in urine (t1/2 < 6 hours). Glucuronidation is limited in newborns relative to older children and adults, leading to concern that significant quantities of free BPA (the toxicologically active compound) may exist in the neonate and have deleterious long term developmental effects.6-8 Thus, the balance of free BPA and BPA-glucuronide is critical for determining exposure risk in neonates.

Unfortunately, data on specific BPA and BPA-glucuronide urinary concentrations in newborns and young children are sparse. Furthermore, progress in BPA quantification

32 has been impeded by sample contamination and the lack of specific and sensitive quantitative methods that can independently measure free BPA and BPA-glucuronide at environmental levels of exposure.9 We have examined the quantitative levels of urinary free BPA and BPA-glucuronide in a population of healthy newborns using a highly sensitive high performance liquid chromatography-tandem mass spectrometry (HPLC-

MS/MS) method that greatly reduces exogenous sources of contamination.10

Methods

In January and February 2012, mothers and healthy, full-term neonates (≥ 37 0/7 weeks gestation) were recruited from the Newborn Nursery at the Johns Hopkins

Hospital. Recruitment and follow up protocols were approved by the Johns Hopkins

Bloomberg School of Public Health Institutional Review Board, and participant informed consent was obtained. Newborns were excluded if they were large or small size for gestational age, were noted to have intrauterine growth restriction, had an APGAR score of less than 5 at 5 minutes of age, had delayed voiding or stooling (occurring greater than

24 hours after birth), had blood incompatibility (ABO or Rh) with their mother, were admitted to the Neonatal Intensive Care Unit (NICU) for management of hyperbilirubinemia, or had other risk factors for hyperbilirubinemia (i.e., cephalohematoma, polycythemia). Neonates born to mothers with documented tobacco use in pregnancy, a positive urine toxicology screen (for cocaine, marijuana, heroin, or methadone) at delivery, and/or anti-epileptic drug use in pregnancy were also excluded.

Infants with hyperbilirubinemia not requiring a NICU admission were included. We

33 further limited recruitment to infants who were to receive pediatric primary care at the

Johns Hopkins’ Harriet Lane Primary Care Clinic.

Urine samples were collected using BPA-free pediatric urine collection bags (U-

Bag, Hollister, Inc. Libertyville, IL) during the neonates’ regular well-child care visits

(occurring at or greater than one week of age). Immediately upon collection, each urine sample was transported to the lab on ice, transferred to a precleaned glass vial, and stored at -80 °C until analysis. Samples were analyzed for free BPA and BPA- glucuronide using HPLC-MS/MS according to a modified previously published method.10

Free BPA and BPA-glucuronide were derivatized with dansyl chloride and measured directly, with d6-BPA and d6-BPA-glucuronide as internal standards, eliminating the need for enzymatic hydrolysis and extraction steps prior to sample analysis. The limit of detection was 0.02 ng/mL and the limit of quantification was 0.1 ng/mL.

Results

A total of 12 mothers and their babies were enrolled into the study. The median age at the time of urine collection was 17 days (Table 1). The average concentration of

BPA-glucuronide, as measured in all of the duplicate urine samples, was 0.87 ± 0.51 ng/mL (median: 0.66 ng/mL). Free BPA was not found in any of the urine samples with the exception of one sample (subject 6) whose replicate sample was a non-detect. (Figure

Questionnaire data revealed that 10 of 12 newborns had some formula intake served in plastic bottles; 5 consumed formula made from powder and 4 drank liquid or

“ready to feed” formula, which requires no addition of water.

Discussion

A critical issue for the assessment of health risks from BPA exposure is the balance of free BPA and its glucuronide conjugate. The presumptive estrogenic effects of

BPA are only mediated by the availability of free BPA, with BPA-glucuronide viewed as a biologically inactive derivative. This study explored the potential problem of free BPA exposure in newborns under the age of 6 weeks. Decreased hepatic glucuronidation has been documented in full term and premature neonates relative to older children and adults6 and is manifested clinically by hyperbilirubinemia in the immediate newborn period.

In a previous study, free BPA was detected in 92% of urines from 42 premature infants in a NICU, suggesting a period of reduced BPA metabolic capacity at birth.11

Thus, we hypothesized that free BPA would be present in the urine of most, if not all infants in our study. Free BPA concentrations were also quantified in 3% of 45 full term infants in Germany, but since median exposure levels in the population were below the method limit of quantification (0.45 μg/L), it is not possible to determine whether infant physiology was responsible for the low detection frequency of free BPA.12 Besides the

German study, we know of no other published BPA biomarker data for full term healthy infants.

The detection of BPA-glucuronide in all infants demonstrates universal exposure to BPA in our study population. We were surprised to determine that BPA-glucuronide was the only detectable BPA compound in the urine of these newborns. These data fundamentally challenge our prior assumptions of the toxicology of this environmental

35 contaminant. Further research is needed to understand whether the absence of detectable concentrations of free BPA in our study population reflects early development of one or more enzyme isoforms responsible for the formation of BPA-glucuronide, or whether at very low levels of exposure to BPA, enzyme activity in neonatal tissues is sufficient to quickly inactivate BPA.

Table 1. Urinary Concentrations of Free BPA and BPA-glucuronide for 12 Infants

Free BPA BPA-Glucuronide (ng/mL) (ng/mL) Subject Age Sex Feeding Type1 Formula Replicate 1 Replicate 2 Replicate 1 Replicate 2 (days) Type 1 7 M Formula Liquid <0.1 <0.1 0.65 0.75 2 9 F Formula Liquid <0.1 <0.1 0.48 0.42 3 11 F Breast Milk - <0.1 <0.1 0.54 0.54 4 13 F Formula Liquid <0.1 <0.1 0.67 0.65 5 14 M Formula Liquid <0.1 <0.1 0.55 0.51 6 16 M Formula + Breast Milk Powder 0.85 <0.1 2.21 1.80 7 17 F Formula Powder <0.1 <0.1 1.39 1.27 8 18 M Formula Powder <0.1 <0.1 1.24 0.85 9 21 M Breast Milk - <0.1 <0.1 0.26 0.37 10 22 M Formula Powder <0.1 <0.1 0.93 1.16 11 25 F Formula2 + Breast Milk - <0.1 <0.1 0.35 0.39 12 44 M Formula + Breast Milk Powder <0.1 <0.1 1.63 1.23

1 On the day of sample collection. 2 Formula type not recorded

Figure 1: Representative HPLC-MS/MS Chromatogram of Neonatal Urine Sample

CHAPTER 3:

Urinary Free Bisphenol A and Bisphenol A Glucuronide Concentrations in a

Sample of Neonates in Baltimore, Maryland

Rebecca M. Nachman, MPH1, Stephen D. Fox, BS2, W. Christopher Golden, MD3, Erica

Sibinga, MD3, Timothy D. Veenstra, PhD2, John D. Groopman, PhD1, Peter S. J. Lees,

PhD1

1Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of

Public Health, Baltimore, MD

2Laboratory of Proteomics, Advanced Technology Program, SAIC-Frederick, Inc.,

Frederick National Laboratory for Cancer Research, Frederick, MD

3Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, MD

Address correspondence to:

Rebecca Nachman, MPH

Department of Environmental Health Sciences

Johns Hopkins Bloomberg School of Public Health

615 N Wolfe Street, Room E7402

Baltimore, MD 21205 [email protected]

Tel: 443-803-6026 / Fax: 410-955-0617

Abstract

Exposure to Bisphenol A (BPA) is widespread in the general population in countries around the world, raising concerns about potential adverse health effects stemming from its estrogenic properties. Upon ingestion, BPA undergoes presytemic metabolism via glucuronidation in the liver, a detoxicating pathway. Hepatic glucuronidation capacity is poor in humans at birth. As a result, serum concentrations of free BPA, the biologically active unchanged parent molecule, are hypothesized to be higher in neonates compared with those in older children or adults with the same weight-normalized intake of BPA.

Uncertainty in estimates of internal dose of BPA in neonates arise from a scarcity of BPA biomarker measurements in this age group and contamination of biological samples with

BPA from field and laboratory sources. In a previous study, we measured free BPA and

BPA glucuronide in a cohort of 11 neonates and 1 young infant over age 7 days using a method that reduces sample contamination. Urine samples were negative for the presence of free BPA even though detectable BPA glucuronide in all the samples indicated exposure in 100% of the study population. In this study of 44 neonates, two samples were collected per individual at ages 3-6 days and 7-28 days. No free BPA was detected in any sample. BPA glucuronide was detected in 71% of the samples (median: 0.27 ug/L). BPA intake rates were back-calculated and the median fell within the lower range of existing estimates obtained through source aggregation. Based on our results, we expect serum

BPA concentrations in newborns to be at or below serum estradiol concentrations.

Introduction

Bisphenol A (BPA) is a high production volume chemical used to synthesize polycarbonate plastic and epoxy resins, materials used in plastic food containers, food and drink can liners, dental sealants, medical equipment and electronics, creating opportunities for exposure in the general population.1 BPA is also a developing agent that coats thermal printing paper such as cash register receipts which has led to the introduction of BPA into the recycled paper supply.2

Pervasive BPA exposure in the general population is well-documented with detection frequencies in urine from representative populations exceeding 90% in countries around the world.3-6 Exposure occurs mainly through the diet.7, 8 Human health concerns arise from BPA’s estrogenic properties which may confer a variety of adverse health effect throughout the lifespan including developmental and reproductive effects, diabetes and metabolic diseases, obesity, cancer of the reproductive organs, and heart disease.9-12

Upon ingestion, the biologically active BPA undergoes first pass metabolism by

Phase II conjugation in the gut and liver. The resulting conjugates, BPA glucuronide and potentially BPA sulfate, are biologically inert and rapidly cleared in the urine (t1/2 < 6 hours).13 Glucuronidation is the main conjugation pathway, with sulfation possibly playing a minor secondary role.14

Poor glucuronidation capacity at birth combined with rapid developmental changes early in life has led to concerns about the unique health risks associated with

BPA exposure in infants.15 Infants’ reduced ability to form glucuronide conjugates was first discovered through investigation of gray baby syndrome, a condition caused by poor

41 clearance of the drug choramphenicol.16 Poor glucuronidation also manifests itself clinically as jaundice in the first weeks of life.17 Although glucuronidation is known to also be limited in utero, the fetus, unlike the newborn, is protected by the metabolism of

BPA by the mother before it crosses the placental barrier.

The impact of poor glucuronidation capacity on internal dose of BPA in infants has been investigated using animal and physically-based pharmacokinetic (PBPK) models.18-22 Sensitive and specific methods exist for the separate quantification of free

BPA and BPA conjugates in human fluids and tissues.23 However, study of BPA metabolism in humans has been hampered by contamination of human biological samples with BPA from laboratory and field sources, which is indistinguishable from free BPA that has passed unmetabolized into the blood or urine. BPA contamination tends to occur with high variability, making blank correction of sample concentrations by subtraction of background concentrations impossible. In addition, biomonitoring data are scarce for children under age 6 as most major biomonitoring studies include only children age 6 years and older.4, 24, 25 Some biomonitoring studies include children as young as 3 years,26 but collection of biological samples for monitoring of environmental contaminants of any kind is rare in children between birth and 1 year and for BPA is limited to a handful of small studies.27-30 To our knowledge, no biomarker measurements exist for full-term healthy infants under 1 week of age even though this is the period of time when postnatal glucuronidation capacity is at its lowest.16, 31 It is also a period of permanent organizational changes in the brain that are driven by estrogen.32

In this study, we utilized laboratory analysis and sample collection methods designed to limit opportunities BPA contamination of samples, to measure concentrations

42 of free BPA to BPA glucuronide in the urine of newborns. We previously reported no free BPA in the urine of 12 infants age 7-44 days with confirmed BPA exposure, suggesting efficient BPA metabolism in this young age group. In this phase of our research, we have measured free BPA and BPA glucuronide in newborns both before and after the first week of life. Given that serum bilirubin concentrations peak at age 3-5 days,33 we hypothesized that infants would be less efficient at metabolizing BPA in the first week of life than in subsequent weeks resulting in detectable free BPA in the urine of neonates in this age group.

Methods

Population and Study Design

Mothers and their healthy full-term neonates were recruited through the Newborn

Nursery at the Johns Hopkins Hospital during their postpartum hospital stay between

December 2012 and August 2013. Recruitment and follow up protocols were approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board, and participant informed consent was obtained. Newborns were excluded if they were less than 37 weeks gestational age at birth, were large or small size for gestational age, were noted to have intrauterine growth restriction, had an APGAR score of less than 5 at

5 minutes of age, had delayed voiding or stooling (occurring greater than 24 hours after birth), were admitted to the Neonatal Intensive Care Unit (NICU) for management of hyperbilirubinemia, or had other risk factors for hyperbilirubinemia (e.g., cephalohematoma, polycythemia). Infants born to mothers with documented tobacco use in pregnancy, a positive urine toxicology screen for cocaine, marijuana, heroin, or

43 methadone at delivery, and/or anti-epileptic drug use in pregnancy were also excluded.

Infants with hyperbilirubinemia not requiring a NICU admission were included.

Additionally, we recruited only infants who were to receive pediatric primary care at the

Johns Hopkins’ Harriet Lane Clinic (HLC). Further explanation of the exclusion criteria is found in Appendix I.

Two urine samples were collected from each infant on two different days, first at age 3-6 days and again at age 7-28 days during scheduled pediatric well-visit checkups at the HLC. The timing of these checkups followed the standard checkup schedule followed by all patients at the HLC and is in line with the recommendations of the American

Academy of Pediatrics.34

Research Visit 1: Participant Recruitment Protocol

Potential participants were screened for study eligibility by pediatric nurse practitioners in the full-term Newborn Nursery. The names of mothers whose newborns were determined to be eligible for the study and who had indicated to caretakers at the hospital their willingness to speak with study investigators and allow study investigators to have access to their medical information were provided in person to the study investigator at the start of the morning rounds each day. The investigator then approached the eligible participants in their hospital rooms. The purpose of the study was explained to the mother and also the father if he was present. Common questions from participants during recruitment included requests for more details about the urine collection methods, and questions about the purpose of the study. After signed consent was obtained, names and phone numbers for the mother and father were collected by the investigator to assist

44 in the communication of HLC appointment times. These data were kept only on a hard copy so that they could be destroyed at the conclusion of the study. The date and time of the first well-visit checkup were obtained from the pediatric nurse practitioners who scheduled the appointment. A phone call was made to participants to confirm the date of the appointment. The date of the second appointment was obtained from participants by phone.

Research Visits 2 and 3: Collection of Urine Samples and Administration of

Questionnaires

Upon arrival at the HLC, participants were greeted by the study investigator in the clinic waiting room. According to regular HLC procedures, participants were called first to a triage examination room where nurses collect physical measurement for patients before transferring them to an examination room where they will be seen by medical residents and attending physicians. The urine samples were collected using BPA-free pediatric urine collection bags (U-Bag, Hollister, Inc. Libertyville, IL) which were placed on the infant in the triage examination room when parents undressed the infant for weighing by the triage nurse. If the sample was not obtained during the triage examination, the bag was left in place and a diaper was placed on the infant over the bag.

The investigator accompanied infants and their families to the examination room where the questionnaire was administered. The investigator read the questions to the mother or caregiver present at the appointment and recorded answers on hard copies. In addition, the investigator completed a sample collection form on which notes were recorded regarding pertinent observations during sample collection (e.g., feces on the

45 bag, leaking of the bag, unusual urine color). Preprinted labels with randomly assigned sample ID were used on questionnaires, sample forms, and vials to ensure that samples and questionnaire data would be successfully matched and to reduce the use of personal identifiers (See Appendix II for copies of the questionnaire and sample forms.)

If the urine sample had not yet been successfully collected by the time the questionnaire was completed, the investigator remained in the examination room with the family until the arrival of the resident or attending physician. At that time, the investigator offered to leave the room and be summoned if the infant urinated. On some occasions, a urine sample was not yet obtained by the conclusion of the patient’s appointment. In that event, the participant was asked if they could stay for some extra time. In some cases, the decision was made to remove the bag without having collected a sample. Study participants received a gift card as compensation for the time and effort of meeting with the investigator regardless of whether a sample was successfully obtained.

Urine collection bags were immediately sealed and placed upright in a precleaned glass transfer vial that was placed in a cooler on ice and transported to the laboratory within 3 hours of sample collection. Upon arrival in the laboratory, samples were transferred using glass pipettes to certified pre-cleaned glass autosampler vials with polytetrafluoroethylene lined caps (National Scientific Company [Part of Thermo Fisher

Scientific], Rockwood, TN) in aliquots of 0.5 mL and stored at -80°C until they were transferred to the laboratory for analysis.

For further details on procedures for placing and removing the urine bags, and reducing opportunities for sample contamination during sample collection, handling and storage, see Appendix III and Appendix IV. A large supply of bags was purchased at the

46 beginning of the study in order to ensure that they would all be from the same lot. These bags were tested for the presence of BPA using a BPA-free synthetic urine mixture to generate QC samples. No free BPA or BPA glucuronide was detected in any QC sample

(Figure 1). The QC sample preparation and testing protocol is found in Appendix V.

Laboratory Analysis

Samples were analyzed using high performance liquid chromatography with tandem mass spectrometry (HPLC-MS/MS) following derivation with dansyl chloride according to a modified previously published method.35 This highly sensitive method eliminates the potential for BPA contamination during laboratory analysis, a major issue due to the ubiquity of BPA in laboratory equipment, by circumventing the use of plastic solid phase extraction columns and by derivatization of BPA during sample workup resulting in an analyte distinguishable from BPA originating from laboratory sources such as solvents, HPLC columns, or capillary tubing used in the analysis. Modifications to the previously published method included direct measurement of BPA glucuronide, which eliminated the need for enzymatic hydrolysis, a step that introduces potentially interfering compounds into the sample. Extraction was also no longer necessary. D6-BPA glucuronide as added as an internal standard. In addition, the HPLC column was changed from a 1mm ID C8 to a 2mm ID C18 which also required a flow increase from 80 to 250 ul/min. The limit of quantitation (LOQ) was 0.1 ug/L for both free BPA and BPA glucuronide (Figure 2).

Positive controls (1 ug/L and 10 ug/L) were prepared by dissolving 4 mg BPA in

4 mL methanol and performing serial dilutions then spiking pooled urine from Phase I,

47 which was known to contain no free BPA. The concentration of the spiking solution was confirmed by UV spectrometer (extinction coefficient: 6300 M-1 at 279 nm).36

Data Analysis

Statistical analysis was performed using STATA 10 (StataCorp, College Station,

TX) and Excel 2010 (Microsoft, Redmond, WA).

First exploratory data analysis including scatter plots, histograms, stem and leaf plots and box and whisker plots, and lowess curves was performed to identify trends in the BPA glucuronide concentrations with age and other factors. Descriptive statistics were calculated to describe the population and urinary BPA glucuronide concentrations.

Next BPA glucuronide concentrations were converted to two other units to make them comparable to other published literature. Presentation of concentrations in units of molar concentration facilitates comparison between substances with different molecular weights. In addition BPA glucuronide concentrations were converted to ug BPA/L in order to make them comparable to the more commonly reported free and Total BPA concentrations in the published literature. The quantification of Total BPA involves enzymatic hydrolysis, which cleaves the glucuronide group prior to quantification.

Concentrations of Total BPA represent BPA glucuronide concentrations if free BPA and other BPA conjugates such as BPA-sulfate are not present in the urine, thus it is useful to present Total BPA and BPA glucuronide in comparable units. For this unit conversion,

BPA glucuronide concentrations were adjusted to the molecular weight of BPA (228.29 g/mol) which is 56% of the weight of BPA glucuronide (404.41 g/mol). The adjusted

BPA glucuronide concentrations were calculated as follows (Equation 1):

BPAGadj = BPAG * (MWBPA/MWBPAG) (Equation 1)

where,

BPAGadj = adjusted concentration of BPA glucuronide

BPAG = unadjusted concentration of BPA glucuronide

MWBPA = molecular weight of BPA

MWBPAG = molecular weight of BPA glucuronide

BPA intake rates were back-calculated from BPA glucuronide concentrations expressed as ug BPA/L, using the method utilized by Lakind and Naiman to back- calculate BPA exposures from urinary Total BPA concentrations available from the

National Health and Examination Survey.37 The back-calculation equation is as follows:

Urinary BPA 24-hour glucuronide (ugBPA/L) * urine output (mL/d) ______= ug BPA/kgbw-d (Equation 2)

body weight (kg) * 1000 mL/L

For the BPA glucuronide and urinary output variables in Eq 2, several different inputs were used in order to obtain a range of estimates. For urinary BPA glucuronide concentrations, the inputs were mean, geometric mean, 5%-ile and 95%-ile based on data from the entire study population (all ages). In addition, intake rates were back-calculated using only Visit 2 data (ages 7-28 days) since Visit 2 urinary concentrations were less variable and less likely to artificially high as a result of low fluid intake.

Two different inputs were chosen for urinary output. The first, 300 mL/day is the reference value published by The International Commission on Radiological Protection,38 which was the source of the urinary output reference values used by Lakind and Naiman in their calculation of BPA intake in the U.S. population ≥ age 6 years.37 The second urinary output value, 150 mL/day, is based on published urine void volumes for neonates39 and the assumption of 8-10 diapers a day, which was the most commonly reported number of dirty diapers on questionnaires from the study population (Appendix

II).

In the final stage of the analysis of the data, simple and multiple linear regressions were performed in order to determine whether factors such as age and formula or breast milk intake were associated with urinary BPA glucuronide concentration, a biomarker of

BPA exposure. The data were converted to panel data using the command xtset with subject id as the group variable and visit as the time variable; this made it possible to account for any correlation between measurements from the same individual during the analysis. The command xtgee (for generalized estimating equation with correlated within-group residuals) was then used to perform a linear regression with the natural log of the BPA glucuronide concentrations as the dependent variable and visit and feeding status as the independent variables (Equations 3, 4 and 5). A uniform correlation matrix was specified since there was only one time lag. After the regression was performed, the command xtcorr was used to determine the correlation between measurements from the same individual.

Yij = β0 + β1visit ij + εij (Equation 3)

Yij = β0 + β1feed ij + εij (Equation 4)

Yij = β0 + β1visit1 ij + β2feed2 ij + εij (Equation 5)

Where,

Yij = natural log of the BPA glucuronide concentration for individual i at visit j

feed ij = Feeding status for individual i at visit j

visitij = Visit 1 or 2 (This functions as a surrogate for age in the model.)

εij = variability around the population mean urinary BPA glucuronide

concentration.

β0 = not interpretable in this model since categorical variables were assigned

values > 0.

β1 and β2 = the average difference in the BPA glucuronide concentration

comparing either visit or feeding status and holding the other variable

(feeding status or visit) constant. For example, β1 could represent the

difference in the mean concentration among breast fed and formula fed

infants at Visit 1. If the coefficient is statistically significant, then that

variable can be considered a factor that impacts BPA exposure.

The statistical significance of β1 and β2 indicates whether feeding status or age are determinants of BPA exposure. By including more than one factor in the model, the influence of each factor on BPA exposure can be evaluated while controlling for the other factor(s). For example, the influence of feeding status on BPA glucuronide concentration

51 can be assessed without confounding by visit (i.e., age). Also, this model takes into account the correlation between measurements from the same individual, which increases the power of the analysis, making it possible to detect associations between factors like feeding status and BPA exposure despite the small sample size.

Two separate one-way analyses of variance (ANOVAs) were also performed on the log transformed BPA glucuronide concentrations at Visit 1 and Visit 2 to compare infants with three different feeding statuses: a) fed exclusively breast milk, b) fed exclusively formula, and c) fed a combination of breast milk and formula.

Results

Study Population Characteristics and Attrition

Population characteristics are presented in Table 1 and Figures 3 & 4. Out of 66 eligible mothers and infants, 51 were enrolled into the study. At least one sample was collected from each of 44 participants. Both samples (at ages 2-6 days and age 7-28 days) were collected from 34 participants. For five participants, a sample was successfully collected at Visit 1, but not Visit 2, and for five others, a sample was collected at Visit 2, but not Visit 1.

At the time of sample collection, 51% of the neonates in the study were fed exclusively formula, 28% were exclusively breast fed, and 21% were fed a combination of breast milk and formula. The percent of infants exclusively breast fed were slightly higher at Visit 1. At Visit 2, the percent of neonates fed formula only or a combination of breast milk and formula rose slightly. (Table 2)

The most common reason for dropping out of the study was running early or late for a scheduled appointment. Communication between investigators and participants between research visits was particularly challenging, and in most cases several phone calls were required before the participant could be reached. When an appointment was missed, participants were given an opportunity to return to the clinic at an agreed upon date to meet with investigators. All mothers who were given this option agreed to it.

Exact numbers were not recorded, but it is estimated that approximately 5-10 samples were collected during these make-up visits

In two cases, infants were dropped from the study because they were admitted to the hospital for health reasons unrelated to the study. In one case, a participant voluntarily withdrew at the time of the first sample collection because the father had reservations about the use of the urine collection bag. Following this occurrence, a greater effort was made to involve the fathers during the recruitment process. Also, as a result of increased outreach to fathers, it was found that fathers could be helpful in maintaining contact with participants between appointments and could alleviate the burden of phone communication with investigators on the mother.

Urinary Free BPA and BPA Glucuronide Concentrations

Concentrations of free BPA and BPA glucuronide in the urine of the 44 neonates in the study are presented in Table 3 and Figure 5. BPA glucuronide concentrations are reported in three different units, ug/L, ug BPA/L and nM. Concentrations are reported as ug BPA/L to facilitate comparison with the more commonly reported Total BPA, for

53 which BPA conjugates are subjected to enzymatic hydrolysis before quantification of

BPA.

Free BPA concentrations were below the LOQ in 100% of 78 urine samples collected (Figure 6a). BPA glucuronide was detected in 71% of the samples (78% at Visit

2, 64% at Visit 2) (Figure 6b). BPA glucuronide concentrations were lognormally distributed with a geometric mean of 0.29 ug/L (95% Confidence Interval [CI]:0.22,

0.40). The geometric mean calculation calculated separately for Visit 1 and Visit 2 were

0.44 ug/L (95% CI: 0.28, 0.70) and 0.20 ug/L (95% CI: 0.14, 0.28). For the purposes of calculations and data analysis, samples with a concentration below the LOQ were assigned a value equal to the LOQ divided by the square root of 2, which approximates the mean of concentrations below the LOQ when data is lognormally distributed and the detection frequency is high.

The maximum BPA glucuronide concentration was 11.21 ug/L, (6.33 ug BPA/L).

The four highest BPA glucuronide concentrations, 3.67, 4.41, 8.49 and 11.21 ug/L

(Adjusted values: 2.07, 2.49, 4.79 and 6.32 ug BPA/L) were measured in samples collected from neonates at Visit 1 with known low intake of fluids in the 24 hour period prior to sample collection compared with other infants in the study.

BPA Intake Rates Back-Calculated from Urinary ug BPA/L in Urine based on BPA

Glucuronide Concentrations

BPA intake rates back-calculated from urinary concentrations of BPA glucuronide are presented in Tables 4 and 5. When calculated using data from both visits, the geometric mean intake was 0.078 ug/kg-bw/d or 0.16 ug/kg-bw/d, depending on the

54 estimated urine output in 24 hours. Intake rates back-calculated using only data from

Visit 2 were slightly lower than those calculated using all the data, with a geometric mean intake of 0.53 ug/kg-bw/d or 0.11 ug/kg-bw/d, depending on estimated urine output in 24 hours. The mean, 5%-ile and 95%-ile intake rates were also calculated, and the mean and

95%-ile were more influenced by the inclusion or exclusion of the Visit 1 data than the geometric mean or 5%-ile were. The most conservative estimate, 1.9 ug/kg-bw/day, was the 95%-ile calculated using all data, including several high outliers from Visit 1, and assuming a urinary output of 300 mL per day.

For context and comparison, published estimates of serum free BPA concentrations obtained using physically based pharmacokinetic models, and the intake rates from which the serum concentrations were modeled are presented in Table 6.

Results of Regression Analysis of Potential Predictors of BPA glucuronide

Concentrations

Evidence of the modest correlation between measurements from the same subject can be seen in Figure 7 in which lines connect BPA glucuronide concentrations for Visit

1 and Visit 2 from the same individual. The results of simple and multiple linear regressions are presented in Table 7. The coefficient for visit when not controlling for other factors was -0.81 (p-value: 0.001) indicating that BPA glucuronide concentrations were higher at Visit 1 compared with Visit 2 and the difference was statistically significant. The coefficient for feeding status was -0.45 (p -value: 0.037), comparing formula fed with neonates who drank a combination formula and breast milk and without controlling for other factors. Data from neonates who drank exclusively breast milk were

55 not included in the regression analysis because their urinary BPA glucuronide concentrations were highly variable at Visit 1. (Figure 8) This variability is likely attributable to dehydration; maternal lactation may be limited in the first week, and the highest urinary BPA glucuronide concentrations were observed in breast fed infants whose mothers reported a low number of wet diapers in the past 24 hours. The effect of feeding status on urinary BPA glucuronide concentration was remained significant and measurements of BPA glucuronide from the same individual were moderately correlated

(ρ=0.46) after controlling for the effect of visit number (Table 7). Geometric mean BPA concentrations by visit and feeding types are presented in Table 8. Other factors for which data were collected using the maternal questionnaire were not found to be significantly associated with urinary BPA glucuronide concentrations. These included sex, formula type (ready to feed versus powder), type of water mixed with formula, and number of ounces ingested per single feeding, number of ounces ingested in 24 hours. In contrast with the results of the regression analysis, the ANOVA resulted in a finding of no statistically significant difference between feeding status groups. For the ANOVA, all three feeding status groups were compared (breast milk only, formula only, and a combination of breast milk and formula); however, the analysis was limited to data from only Visit 2 since the BPA glucuronide concentrations for the exclusively breast fed group were highly variable at Visit 1 thereby violating Bartlett’s test of equal variances and raising concerns about the validity of urinary BPA glucuronide concentrations in this group as biomarkers of BPA exposure.

Discussion

Determination of the balance of free and conjugated BPA in biological fluids is a critical component of evaluations of health risk from BPA exposure in humans. Hepatic glucuronidation, the main detoxication pathway for BPA, is pre-systemic when exposure occurs by ingestion, but is known to be lower in neonates than it is in older infants, children, and adults, possibly resulting in a higher internal dose of BPA in infants. In a previous phase of our research, we reported BPA exposure in 100% of 12 neonates and young infants age 7-44 days, but no free BPA in the urine of these infants. In this phase of our research, we measured free BPA and BPA glucuronide in the urine of 44 neonates, collecting samples at two time points: age 3-6 days and age 7-28 days. We hypothesized that we would detect free BPA in the urine collected at the first time point. BPA glucuronide was detected in 71% of 78 samples, but no urinary free BPA was detected in samples from either age group. This study is the first of its kind to be conducted in healthy full-term neonates. Concentrations of free BPA and Total BPA in urine have been published for infants over the age of 1 month, but BPA biomarker data for this age group and younger are sparse. Together with our previously published research, these are the first two studies in which no free BPA was reported in biological samples from an infant population. Volkel et al. reported detectable free BPA 9% of 91 urine samples from 47 infants age 1-5 months compared with detection of Total BPA in 66% of the same samples.28 In a population of infants age 2-15 months, free BPA was detected in 28% and Total BPA was detected in 93% of 29 samples.29 Our finding suggests that free BPA detected in biological samples in older populations with similar exposures levels likely results from contamination of samples during sample collection, handling, storage or

57 analysis. Sample contamination is a recognized problem among researchers in this area of study.23, 40

This study utilized unique laboratory analysis methodology designed to reduce contamination; examples include use of glass storage containers and pipettes, direct measurement of BPA glucuronide, and derivatization of samples without prior treatment of samples with any reagent or filtering of sample using extraction columns. Further details on steps taken to avoid contamination in the field and laboratory are presented in

Appendix III.

Given that free BPA concentrations fell below the LOQ of 0.1 ug/L in all the samples collected for this study, we can infer that the fraction of free BPA in urine samples from our study population could not have exceeded the value of the LOQ divided by the measured BPA glucuronide concentration (adjusting for differences in the molecular weight of BPA versus BPA glucuronide). If we consider the samples with BPA glucuronide concentrations in the highest quartile, at least 80% of the BPA in the urine of those samples was in glucuronidated form. In other words, in in 20 of 78 samples (i.e.,

26% of the samples), in which BPA glucuronide concentrations exceeded 0.05 ug

BPA/L, free BPA could account for no more than 20% of all BPA in the urine sample;

80% or more of the BPA is in the form BPA glucuronide. In 9 of 78 samples (12% of samples), the BPA glucuronide concentration exceeded 0.1 ug BPA/L, thus at least 90% of the BPA in the sample was in the form BPA glucuronide. BPA-sulfate was not quantified in the samples and may have been present in urine from our study population.

If unquantified BPA sulfate was indeed present in our samples, then BPA conjugates

(BPA glucuronide + BPA sulfate) comprised an even a greater percent of the Total BPA

58 in urine from the study population than we have estimated from our measurements. The low percent of free BPA is indicative of efficient conjugation of BPA and suggests that the serum free BPA concentrations in our subjects were also very low.

The BPA intake rates reported in this study are the first estimates of intake for this age group to be calculated from biological samples. Previous intake estimates for neonates relied on source aggregation, a method which introduces uncertainty into the estimates and does not account for exposures from unidentified sources. The mean and

95%-ile intake rates calculated from BPA glucuronide concentrations in our study population fell within the range of existing conservative estimates of BPA intake in neonates which are based on source aggregation. The mean, median, and 95%-ile back- calculated intake rates using Visit 2 data were 0.11, 0.20, and 0.78 ug/kg-bw/d, respectively, compared with intake estimates from source aggregation that range from 0.2 ug/kg-bw/d to 11 ug/kg-bw/d depending on the exposure scenario.21, 22, 41 It was surprising that BPA-glucuronide concentrations were as high as they were in our study population since polycarbonate bottles have been off the market for several years and were unlikely to have been used by anyone in the study.42, 43 In addition, although BPA has been detected in liquid formula, it is not found at detectable levels in powder formula, and the use of BPA in formula packaging was outlawed earlier this year.44, 45 Thus the presence of BPA glucuronide in the urine of infants in our study suggests unknown and possibly non-dietary sources of exposure.

Although no free BPA was detected in any sample in this study, we know that free BPA could have been present in our urine samples at concentrations up to the LOQ of 0.1 ug/L (0.43 nM). We would expect serum concentrations of free BPA to be a

59 fraction of the concentrations in urine and thus lower than 0.1 ug/L (0.43 nM). By comparison, estrogen concentrations in serum of neonates are elevated at birth but fall to prepubertal levels < 0.020 ug/L (< 0.073 nM) within weeks.46 Serum estrogen concentrations at ages 1-15 days range from 0.01 to 0.5 ug/L (0.037-0.18 nM).47, 48 While this study does not provide any direct evidence, the possibility that free BPA concentrations in serum might exceed the serum concentration of endogenous estrogen is not inconceivable given the LOQ limitations, especially after serum estradiol levels fall to prepubertal levels. The health implications of these possible very low serum free BPA concentrations are beyond the scope of this study. Estimates of the relative potency of

BPA compared with endogenous estrogen are highly variable. BPA is estimated to be 20 to 27,000 times less potent than estradiol in terms of its affinity for the alpha or beta estrogen receptor.14 BPA may induce responses via receptors other than alpha and beta estrogen receptor.14

The mean intake rate which we back calculated from the concentrations in samples collected at Visit 2 is comparable to the intake rate from which Edginton and

Ritter estimated serum steady state free BPA and BPA glucuronide concentrations using a PBPK model.21 (Note: All serum and urine concentration values from the study by

Edginton and Ritter referenced here are approximate since they were presented in the publication as graphs only, not tables.) According to the sensitivity analysis, an influential parameter in Edginton and Ritter’s model was the percent of the relevant enzyme activity in neonates compared with adults, which was estimated to be 6.4% of adult enzyme activity. Based on this model, Edginton and Ritter estimated the steady

60 state serum free BPA concentration in neonates to be approximately 0.01 ug/L (0.043 nM), or 5% - 6% the estimated serum BPA glucuronide concentration.

In four of the urine samples from our study, BPA glucuronide concentrations exceeded 3.5 ug/L (2 ug BPA/L), a threshold at which we would have expected to detect free BPA in urine if the free concentration were 5% of the concentration of glucuronidated BPA, yet free BPA was not detected. However, it must be considered that the high exposures in these four infants are likely due to low intake of fluids, given their ages (three were 3 days old and one was 5 days old) and given that 3 of the four were exclusively breast fed, which could have resulted in dehydration of the infants and/or poor clearance of BPA from exposure in utero. It is worth noting that the BPA glucuronide concentration in a 3-day old formula-fed infant was 11.21 ug/L (6.33 ug/L); given the LOQ of 0.1 ug/L, free BPA in the urine of this subject could not have exceeded

2% of the Total BPA concentration. In an adult population, it would be standard procedure to quantify urinary dilution by measuring creatinine concentrations or specific gravity normalizing urinary BPA glucuronide concentrations. However, creatinine is not a valid measure of dilution in infants less than 2 weeks since it is still highly influenced by in utero exposure to maternal creatinine during this period.49 For the same reason, specific gravity was presumed to be an invalid measure of urinary dilution.

In this study, newborns who were fed exclusively formula had significantly higher urinary BPA glucuronide concentrations compared with newborns who were fed a combination of formula and breast milk according to results of a multiple linear regression that accounted for correlation between measurements from the same individual. This finding suggests that formula intake may be a risk factor for BPA

61 exposure. In addition to regression analysis, ANOVA was used to test for a difference between all feeding groups (i.e., formula only, breast milk only, and combination of both formula and breast milk) using only data from Visit 2; per the ANOVA, there was no difference between urinary BPA glucuronide concentrations for any of the three feeding groups.

Our results contrast with the results of a study of premature infants in a NICU

(corrected gestational age ≤ 44 weeks), in which free BPA was detected in 92% and Total

BPA was detected in 100% of the 41 samples.27 The NICU population was both more highly exposed and more immature developmentally than our full term study population since and were exposed via equipment in the NICU. Thus, the findings of our study may not apply to premature infants or infants with high exposures from non-dietary sources such as medical equipment. Free BPA concentrations were correlated with Total BPA concentrations in NICU study population. On the one hand, this correlation could indicate that a consistent percent of free BPA went unconjugated such that the infants with the highest Total BPA concentrations also had the highest free BPA concentrations.

However, this correlation could be a result of BPA contamination in that high Total BPA concentrations could have been high due to contamination with BPA.

Conclusion

We observed no evidence of reduced BPA metabolism in the 44 neonates in this study.

To our knowledge, the intake rates reported in this study are the first of their kind to be back-calculated from urinary biomarker measurements in a healthy infant population less than 1 month of age. Urine collection bags are an effective way to collect urine samples

62 in neonates and should be considered for use in future environmental biomontitoring studies of neonates and young infants. Measured concentrations of biomarkers in urine collected from neonates in first week are difficult to interpret because of variability in fluid intake in that age group which in turn influences both the dilution of the urine and clearance of endogenous and exogenous compounds passed on in utero. Formula intake may be a risk factor for BPA exposure.

Table 1. Population Characteristics

Visit 1 Visit 2 Total (Age 2-6 (Age 7-28 (Both Visits) days) days) 1 * Number of Subjects 44 39 39

Male** 25 (57%) 21 (54%) 21 (54%)

Female** 19 (43%) 18 (46%) 18 (46%)

† Age at sample collection -- 4.3 (1.0) 12.1 (3.8)

Gestational Age at Birth † 39.1 (1.12) -- -- (weeks)

Birth weight (grams)† 3117 (346.7) -- --

* Samples were missing for 5 subjects at visit 1 and for another 5 subjects at visit 2. ** N (%) † N(SD)

Table 2. Breast Milk and Formula Intake in the Study Population

Total Visit 1 Visit 2

(Both (Age 2-6 (Age 7-28 Visits) days) days) Number of 78 39 39 Observations

Breast milk only 19 (24%) 11 (28%) 8 (21%)

Formula only 41 (53%) 20 (51%) 21(54%)

Breast milk and 18 (23%) 8 (21%) 10 (25%) formula

Table 3. Concentrations of Free BPA and BPA Glucuronide in the Urine of 44 Neonates*

Age Mean N (%) Analyte GM 5% 25% 50% 75% 95% Min Max Group (SD)

* BPA=bisphenol A, BPAG=BPA glucuronide. The LOQ for both free BPA and BPA glucuronide was 0.1 ug/L (0.25 nM). The LOQ for BPA- glucuoride adjusted to the molecular weight of BPA is 0.06 ug BPA/L. ** N=78 refers to the number of samples. The number of participants was 44; two samples were collected from 34 of the 44 participants,one sample was collected at Visit 1 only for 5 participants and one sample was collected at Visit 2 only for 5 participants.

Table 4. Intake Rates in Newborns Back-Calculated from Urinary BPA glucuronide Concentrations for 44 Neonates (Both Age Groups Included)

Urinary BPA- Body Urine Glucuronide BPA Intake Weight Output Concentration (ug/kg-bw/d) (kg) (mL/d) Parameter Mean 0.42 Geometric Mean 0.16 3.117 300 5%-ile 0.37 95%-ile 1.9 Mean 0.21 Geometric Mean 0.078 3.117 150 5%-ile 0.018 95%-ile 0.97

Table 5. Intake Rates Back-Calculated from Urinary BPA glucuronide Concentrations for 39 Neonates (Age 7-27 days)

Urinary BPA- Body Urine Glucuronide BPA Intake Weight Output Concentration (ug/kg-bw/d) (kg (mL/d) Parameter Mean 0.20 Geometric Mean 0.11 3.117 300 5%-ile 0.037 95%-ile 0.78 Mean 0.098 Geometric Mean 0.053 3.117 150 5%-ile 0.018 95%-ile 0.39

Table 6. European Food Safety Authority (EFSA) BPA Intake Rates41 and Corresponding Modeled Serum or Plasma Concentrations of Free BPA21, 22

Model 121 Serum Model 222 Serum Model 222 Serum Intake, Age Exposure Source Steady State, Steady State, Peak, ug/kg-bw/d ug/L (nM)** ug/L (nM) ug/L (nM) 11* -- 0.096 (0.42) 0.34 (1.4) Newborn* Breast milk 0.25* ~0.01 (0.043) -- -- 3 month Breast milk 0.2 ~0.002 (0.008) 0.0008 (0.004) 0.0033 (0.014) (breast fed) 3 month Infant formula (non-pc bottle) 2.3 -- 0.02 (0.088) 0.071 (0.31) (glass bottle) 3 months 11 -- 0.096 (0.42) 0.34 (1.4) Infant formula (pc bottle) (pc bottle) 4.0 ~0.03 (0.14) -- -- Infant formula (pc bottle), 13 ~0.05 (0.22) 0.030 (0.13) 0.17 (0.74) 6 months commercial food/beverages 8.3 ~0.03 (0.14) -- -- 1.5 years commercial food/beverages 5.3 ~0.02 (0.088) 0.010 (0.044) 0.059 (0.26) Adult commercial food/beverages 1.5 ~0.006 (0.026) 0.0038 (0.017) 0.024 (0.11) Newborn*** n/a 1*** ~0.04 (0.018) -- -- Adult*** n/a 1*** ~0.004 (0.002) -- --

* Newborn intake estimates were generated by the PBPK modeling investigators, not by EFSA. ** Steady state serum concentrations were published in graph form only; values presented here are visual estimates from the graphs. *** Steady state serum concentrations for an intake rate of 1 ug/kgbw/d were modeled for comparison across age groups and are not meant to represent age-appropriate intake estimates.

Table 7. Association of Individual Factors with Urinary BPA Glucuronide Concentration

Dependent Independent Correlation Coefficient(s) P-value(s) Variable Variable(s) Coefficient† ln(BPAG)* Visit -0.81 0.001 0.28

ln(BPAG) Feeding Status** -0.72 0.037 0.39

Visit, -0.50, 0.027, ln(BPAG) 0.46 Feeding Status -0.72 0.032

* BPAG = urinary BPA glucuronide concentration

** Comparing neonates fed formula exclusively to neonates fed a combination of formula and breast milk

† Correlation between BPAG from the same individual

Table 8. Geometric Mean Urinary BPA Glucuronide Concentrations by Age Group and Feeding Status (ug/L)

Visit Breast Milk Formula Only Combination of Only Formula and Breast Milk 1 0.71 0.44 0.22 2 0.11 0.24 0.15 All data 0.32 0.32 0.18

Figure 1. Representative Chromatogram for a Synthetic Urine Rinsate Blank

RT: 5.0 - 24.0 SM: 3B NL: 5.91E2 m/z= 461.50-462.50 F: + p ESI SRM 100 ms2 [email protected] [461.750-462.250] MS 80 jhu40_130907102532 60 40 20 RT: 9.9 NL: 5.18E5 SN : 86908 m/z= 467.50-468.50 F: + p ESI SRM 100 ms2 [email protected] [467.750-468.250] MS ICIS 80 jhu40_130907102532 60 d -BPAG 40 6 20 Relative Relative Abundance NL: 1.16E4 14.6 m/z= 169.50-170.50+170.50-171.50 100 12.0 RT: 20.0 F: + p ESI SRM ms2 SN : 632 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS jhu40_130907102532 60 BPA 40 20

RT: 20.0 NL: 5.33E5 SN : 25609 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS jhu40_130907102532 60 40 d6-BPA 20

6 8 10 12 14 16 18 20 22 24 Time (min)

72 Figure 2. Chromatogram for Calibration Standard H, BPAG and BPA at 0.1 ng/mL (LOQ)

RT: 5.0 - 24.0 SM: 3B RT: 9.9 NL: 1.34E4 SN : 3668 m/z= 461.50-462.50 F: + p ESI SRM 100 ms2 [email protected] [461.750-462.250] MS ICIS 80 jhu08_130906180208 60 BPAG 40 20

RT: 9.9 NL: 4.91E5 SN : 179556 m/z= 467.50-468.50 F: + p ESI SRM 100 ms2 [email protected] [467.750-468.250] MS ICIS 80 jhu08_130906180208 60 d -BPAG 40 6 20 Relative Relative Abundance RT: 20.0 NL: 3.01E4 SN : 2602 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS jhu08_130906180208 60 40 BPA 20

RT: 20.0 NL: 5.31E5 SN : 27781 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS jhu08_130906180208 60 40 d6-BPA 20

6 8 10 12 14 16 18 20 22 24 Time (min)

73 Figure 3. Eligible Participants Screen, Enrolled, and Followed Up

66 15 Eligible/Declined Screened/Eligible

51 7 Enrolled/Dropped Enrolled/Consented

34 (Visit 1 and Visit 2)

44 5 (Visit 1 only) Followed Up

5 (Visit 2 only)

Figure 4. Age Distribution of Study Population at the Time of Sample Collection*

* Bimodal distribution reflects repeated sample collection design with sample collected at ages 2- 6 days and again at ages 7-28 days.

Figure 5. Urinary free BPA and BPA Glucuronide Concentrations in Neonates by Age

76 Figure 6a. Representative Chromatogram for Urine Sample with a BPA Glucuronide Concentration of 0.8 ug/L

RT: 5.0 - 24.0 SM: 3B RT: 9.9 NL: 3.11E4 SN: 2784 m/z= 461.50-462.50 F: + p ESI SRM 100 ms2 [email protected] [461.750-462.250] MS ICIS 80 jhu10_130904143506 60 BPAG 40 20

RT: 9.9 NL: 5.74E5 SN: 91460 m/z= 467.50-468.50 F: + p ESI SRM 100 ms2 [email protected] [467.750-468.250] MS ICIS 80 jhu10_130904143506 60 40 d6-BPAG 20 Relative Relative Abundance NL: 8.59E4 9.3 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS jhu10_130904143506 60 8.6 40 7.6 20

RT: 20.0 NL: 5.97E5 SN: 16334 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS 8.4 jhu10_130904143506 60 d -BPA 40 6 20

6 8 10 12 14 16 18 20 22 24 Time (min)

77 Figure 6b. Representative Chromatogram for Urine Sample with a BPA Glucuronide Concentration of 8.5 ug/L

RT: 5.0 - 24.0 SM: 3B RT: 9.9 NL: 3.53E5 SN: 17166 m/z= 461.50-462.50 F: + p ESI SRM 100 ms2 [email protected] [461.750-462.250] MS ICIS 80 jhu35_130905032320 60 BPAG 40 20

RT: 9.9 NL: 5.14E5 SN: 53738 m/z= 467.50-468.50 F: + p ESI SRM 100 ms2 [email protected] [467.750-468.250] MS ICIS 80 jhu35_130905032320 60 d -BPAG 40 6 20 Relative Relative Abundance NL: 2.54E5 8.6 9.3 m/z= 169.50-170.50+170.50-171.50 100 7.6 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS jhu35_130905032320 60 40 7.4 8.4 20

NL: 6.63E5 9.4 m/z= 169.50-170.50+170.50-171.50 100 F: + p ESI SRM ms2 [email protected] [169.750-170.250, 80 170.750-171.250] MS ICIS RT: 20.0 jhu35_130905032320 60 SN: 4237 40 d -BPA 5.7 8.4 6 20

6 8 10 12 14 16 18 20 22 24 Time (min)

78 Figure 7. Repeated Urinary BPA glucuronide Concentrations in Neonates at Visit 1 (Age 2-6 Days) and Visit 2 (Age 7-28 Days) with Line Connecting Measurements from the Same Individual

Figure 8. Urinary BPA Glucuronide Concentrations by Visit and Feeding Type

Chapter 4: CONCLUSIONS

Summary of the Findings

The estimation of internal dose of free BPA following ingestion is a critical component of the evaluation of health risks from exposure to BPA. First pass metabolism of BPA to BPA glucuronide, a detoxication pathway, is highly efficient in adults, but thought to be less so in neonates and young infants, who are more vulnerable to the potential health effects associated with BPA exposure.

In Chapter 1, we conducted a review of the literature on neonatal BPA metabolism. We included studies published from August 2008 to August 2013 that fell into three categories: 1) Studies reporting separate measurement of free BPA and BPA glucuronide or Total BPA in the urine of infants; 2) Studies of BPA toxicokinetics in animals; and 3) PBPK models of BPA internal dose in infants and young children.

Results were inconsistent between studies. Reported measurements of urinary free BPA and Total BPA in infants suggest that glucuronidation may be very efficient in older full- term infants.1, 2 In young premature infants, both free BPA and BPA glucuronide were detected in almost all the samples and the two measurements were correlated, suggesting glucuronidation of BPA occurred, but was limited in this population.3 In toxicokinetic animal studies, results varied depending on the animal model. Serum free BPA was significantly higher in neonatal rodents than in adult rodents with the same weight- adjusted BPA dose.4, 5 In primates, the impact of age on internal dose of free BPA was not statistically significant.6 In two studies, serum free BPA concentrations in neonates, infants, children and adults were estimated using PBPK models.7, 8 Limited

81 glucuronidation capacity and increased body burden of BPA were reflected in the inputs chosen for pharmacokinetic parameters and intake rates in both models. Both pharmacokinetic parameters and intake rates were highly uncertain due to incomplete data, but were the most influential inputs in the models, demonstrating the need for human biomarker measurements in neonates.

In Chapter 2 we reported free BPA and BPA glucuronide concentrations in urine from a cohort of 11 neonates and 1 young infant in Baltimore Maryland. BPA glucuronide was detected in all of the samples (median: 0.66 ug/L) confirming exposure to BPA in 100% of the study population. Free BPA was detected in only one sample, for which the replicate was negative. This was the first study to report urinary biomarkers of

BPA exposure and metabolism in healthy full-term infants under one month of age.

These unexpected findings challenged our hypothesis that given immature hepatic metabolism in neonatal levels of free BPA would be detectable in the urine of infants in this age range.

In Chapter 3 we reported free BPA and BPA glucuronide concentrations in urine from a cohort of neonates in Baltimore Maryland. We hypothesized that free BPA would be detectable in urine of neonates in the first week of life, but not in subsequent weeks due to poor glucuronidation capacity at birth and a rapid increase in enzyme activity in the neonatal period. Building upon the previous study, we expanded the size of the study population to 44 subjects, and collected two measurements from each subject at two time points, first at age 3-6 days and again at age 7-28 days. No free BPA was detected in any sample and BPA glucuronide was detected in 71% of the samples (median: 0.29 ug/L).

Given the detection of no free BPA at the method LOQ of 0.1 ug/L, we conclude that

BPA glucuronide is more efficient in the first week of life than we hypothesized.

Measurements of BPA-glucuronide concentrations in the study population were also used to back-calculate BPA intake rates. These intake rates are the first exposure estimates to be calculated from biomarkers for this age group and are an improvement upon previous estimates obtained by source aggregation.

Future Research and Implications

An important research gap identified during the review of the literature on neonatal BPA metabolism was the need for more human biomarker data in neonates.

Existing studies of healthy full-term infants focused on infants greater than or equal to one month of age. In addition, detectable free BPA concentrations, thought to be the result of sample contamination during sample collection, handling or analysis, impeded interpretation of the results. High quality data from animals exist, but as with any animal study, interspecies extrapolation remains an issue. PBPK models were found to be highly sensitive to parameters with substantial uncertainty such as intake rates calculated by source aggregation and toxicokinetic parameters based on studies of morphine. Morphine is metabolized by UGT2B7, one of the 16 known isoforms of uridine diphosphate glucuronosyltranferase (UGT) that catalyze glucuronidation reactions in the liver. BPA is metabolized primarily by UGT2B15, morphine may not be a good model of BPA metabolism.

This research resulted in the first measurements of urinary biomarkers of either

BPA exposure or BPA metabolism in healthy full term neonates. Since no urine sample from any infant in either of the two cohorts studied was found to contain a detectable

83 level of free BPA, the main conclusion of this study is that glucuronidation of BPA is more efficient in neonates than previously thought. Furthermore, this research demonstrated methods that eliminated the contamination of sample by background sources of BPA in the field and laboratory; contamination of samples has presented a particular challenge in this field and has impeded interpretation of urinary BPA biomarker measurements in other studies.9 The data generated by this research support the results of toxicokinetic studies conducted in primates and in addition provide better estimates of BPA intake than those that were previously available.

While this study addresses a critical data gap, further research is needed to improve evaluations of health risks from BPA exposure from postnatal exposure.

Addressing gaps in our knowledge of the isoform UGT 2B15, which is mainly responsible for BPA glucuronidation, would improve toxicokinetic inputs in models of

BPA internal dose resulting in more accurate and precise estimates of serum free BPA in neonates. Very little is known about the ontogeny of UGT 2B15 or whether an analogous isoform exists in primates or other animal models.

Although we did not measure free BPA in serum, we are able to conclude from our urinary measurements that free BPA concentrations in serum would not exceed our

LOQ of 0.1 ug/L and would likely be much lower. Serum free BPA concentrations one order of magnitude lower than our limit of detection are still in the range of the concentrations of endogenous estrogen in neonates after they fall from initially elevated birth levels. To understand the health significance of these concentrations of free BPA, more research is needed at the mechanistic level. Estimates of BPA’s potency relative to estradiol are highly variable, ranging from 20 to 27,000 times less potent.10 More research

84 is needed to understand whether it is biologically feasible for BPA to induce a biological effect given both low serum concentrations and low potency. Modes of action other than binding to the nuclear estrogen receptor (for example binding to the androgen receptor) should be explored.

Lastly, quantification of BPA sulfate in neonatal biological samples is also needed in order to evaluate whether sulfation of BPA is an additional BPA detoxication pathway in neonates.

Final Conclusions

We generated the first data on free BPA and BPA glucuronide concentrations in the urine of neonates as young as 3 days old in two cohorts from Baltimore, Maryland, thereby addressing a major research gap identified by the National Toxicology Program

Committee for the Evaluation of Risks to Human Reproduction.10 Given the method

LOQ of 0.1 ug/L, we concluded that BPA glucuronide accounted for at least 80% of the

BPA in urine of these infants, which demonstrates a high capacity to metabolize BPA by glucuronidation, the main detoxication pathway of BPA in humans. Intake rates back- calculated from the urinary BPA glucuronide concentrations in our study population are consistent with the lower range of estimates that have been used by other researchers to model serum BPA concentrations in this age group. Our data is consistent with serum free BPA concentrations that are in the range of or lower than endogenous estrogen concentrations in serum of neonates once they have fallen from initially elevated levels at birth (within the first 14 days). Further research is needed to elucidate a biologically

85 plausible mechanism of BPA toxicity given the very low internal dose of BPA following exposure.

REFERENCES

Introduction and Specific Aims

1. National Toxicology Program (Center for the Evaluation of Risks to Human Reproduction). NTP-CERHR monograph on the potential human reproductive and developmental effects of bisphenol A. 2008;NIH Publication No. 08 – 5994.

2. von Goetz N, Wormuth M, Scheringer M, Hungerbuhler K. Bisphenol a: how the most relevant exposure sources contribute to total consumer exposure. Risk Anal. 2010;30:473- 487.

3. Volkel W, Colnot T, Csanady GA, Filser JG, Dekant W. Metabolism and kinetics of bisphenol a in humans at low doses following oral administration. Chem Res Toxicol. 2002;15:1281-1287.

4. Fox SD, Falk RT, Veenstra TD, Issaq HJ. Quantitation of free and total bisphenol A in human urine using liquid chromatography-tandem mass spectrometry. J Sep Sci. 2011;34:1268-1274.

Chapter 1

2. European Commission Joint Research Center (Institute for Health and Consumer Protection). European union risk assessment report: bisphenol-A. Luxembourg: Office for Official Publications of the European Communities; 2003;EUR 20843 EN.

3. Liao C, Kannan K. Widespread occurrence of bisphenol A in paper and paper products: implications for human exposure. Environ Sci Technol. 2011;45:9372-9379.

4. Calafat AM, Ye X, Wong LY, Reidy JA, Needham LL. Exposure of the U.S. population to bisphenol A and 4-tertiary-octylphenol: 2003-2004. Environ Health Perspect. 2008;116:39-44.

5. Bushnik T, Haines D, Levallois P, Levesque J, Van Oostdam J, Viau C. Lead and bisphenol A concentrations in the Canadian population. Health Rep. 2010;21:7-18.

6. Koch HM, Kolossa-Gehring M, Schroter-Kermani C, Angerer J, Bruning T. Bisphenol A in 24 h urine and plasma samples of the German Environmental Specimen Bank from 1995 to 2009: a retrospective exposure evaluation. J Expo Sci Environ Epidemiol. 2012;22:610-616.

7. Zhang Z, Alomirah H, Cho HS, et al. Urinary bisphenol A concentrations and their implications for human exposure in several Asian countries. Environ Sci Technol. 2011;45:7044-7050.

8. Lang IA, Galloway TS, Scarlett A, et al. Association of urinary bisphenol A concentration with medical disorders and laboratory abnormalities in adults. JAMA. 2008;300:1303-1310.

9. Braun JM, Kalkbrenner AE, Calafat AM, et al. Impact of early-life bisphenol A exposure on behavior and executive function in children. Pediatrics. 2011;128:873-882.

10. Trasande L, Attina TM, Blustein J. Association between urinary bisphenol A concentration and obesity prevalence in children and adolescents. JAMA. 2012;308:1113- 1121.

11. Acevedo N, Davis B, Schaeberle CM, Sonnenschein C, Soto AM. Perinatally Administered Bisphenol A Acts as a Mammary Gland Carcinogen in Rats. Environ Health Perspect. 2013.

12. Selevan SG, Kimmel CA, Mendola P. Identifying critical windows of exposure for children's health. Environ Health Perspect. 2000;108 Suppl 3:451-455.

13. Hajszan T, Leranth C. Bisphenol A interferes with synaptic remodeling. Front Neuroendocrinol. 2010;31:519-530.

14. Lakind JS, Naiman DQ. Bisphenol A (BPA) daily intakes in the United States: estimates from the 2003-2004 NHANES urinary BPA data. J Expo Sci Environ Epidemiol. 2008;18:608-615.

15. European Commission Joint Research Center (Institute for Health and Consumer Protection). European Union risk assessment report: 4-4' isoproylidenediphenol (Bisphenol A), Part 2 Human Health. Luxembourg: Publications Office of the European Union; 2008;EINECS No. 201-245-8.

16. Wilson NK, Chuang JC, Morgan MK, Lordo RA, Sheldon LS. An observational study of the potential exposures of preschool children to pentachlorophenol, bisphenol-A, and nonylphenol at home and daycare. Environ Res. 2007;103:9-20.

17. Volkel W, Colnot T, Csanady GA, Filser JG, Dekant W. Metabolism and kinetics of bisphenol a in humans at low doses following oral administration. Chem Res Toxicol. 2002;15:1281-1287.

18. McCarver DG, Hines RN. The ontogeny of human drug-metabolizing enzymes: phase II conjugation enzymes and regulatory mechanisms. J Pharmacol Exp Ther. 2002;300:361-366.

19. Alcorn J, McNamara PJ. Pharmacokinetics in the newborn. Adv Drug Deliv Rev. 2003;55:667-686.

20. Dekant W, Volkel W. Human exposure to bisphenol A by biomonitoring: methods, results and assessment of environmental exposures. Toxicol Appl Pharmacol. 2008;228:114-134.

21. Ye X, Zhou X, Wong LY, Calafat AM. Concentrations of bisphenol A and seven other phenols in pooled sera from 3-11 year old children: 2001-2002 National Health and Nutrition Examination Survey. Environ Sci Technol. 2012;46:12664-12671.

22. Ye X, Zhou X, Hennings R, Kramer J, Calafat AM. Potential external contamination with bisphenol A and other ubiquitous organic environmental chemicals during biomonitoring analysis: an elusive laboratory challenge. Environ Health Perspect. 2013;121:283-286.

23. Calafat AM, Weuve J, Ye X, et al. Exposure to bisphenol A and other phenols in neonatal intensive care unit premature infants. Environ Health Perspect. 2009;117:639- 644.

24. Centers for Disease Control and Prevention. Fourth National Report on Human Exposure to Environmental Chemicals, updated tables, February 2012. 2012.

25. Li D, Zhou Z, Qing D, et al. Occupational exposure to bisphenol-A (BPA) and the risk of self-reported male sexual dysfunction. Hum Reprod. 2010;25:519-527.

26. Volkel W, Kiranoglu M, Fromme H. Determination of free and total bisphenol A in urine of infants. Environ Res. 2011;111:143-148.

27. Mendonca K, Hauser R, Calafat AM, Arbuckle TE, Duty SM. Bisphenol A concentrations in maternal breast milk and infant urine. Int Arch Occup Environ Health. 2012.

28. Nachman RM, Fox SD, Golden WC, et al. Urinary free bisphenol A and bisphenol A- glucuronide concentrations in newborns. J Pediatr. 2013;162:870-872.

29. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult CD-1 mice: inter-species comparisons with Sprague- Dawley rats and rhesus monkeys. Toxicol Lett. 2011;207:298-305.

30. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult Sprague-Dawley rats. Toxicol Appl Pharmacol. 2010;247:158-165.

31. Doerge DR, Twaddle NC, Woodling KA, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult rhesus monkeys. Toxicol Appl Pharmacol. 2010;248:1-11.

32. U.S. Environmental Protection Agency. IRIS Database.

33. Edginton AN, Schmitt W, Voith B, Willmann S. A mechanistic approach for the scaling of clearance in children. Clin Pharmacokinet. 2006;45:683-704.

34. Mielke H, Gundert-Remy U. Bisphenol A levels in blood depend on age and exposure. Toxicol Lett. 2009;190:32-40.

35. European Food Safety Authority. Opinion of the scientific panel on food additives, flavourings, food processing aids, and materials in contact with food (AFC). The EFSA journal. 2006;759:1-10.

36. Edginton AN, Ritter L. Predicting plasma concentrations of bisphenol A in children younger than 2 years of age after typical feeding schedules, using a physiologically based toxicokinetic model. Environ Health Perspect. 2009;117:645-652.

37. Hanioka N, Naito T, Narimatsu S. Human UDP-glucuronosyltransferase isoforms involved in bisphenol A glucuronidation. Chemosphere. 2008;74:33-36.

38. Matsumoto J, Yokota H, Yuasa A. Developmental increases in rat hepatic microsomal UDP-glucuronosyltransferase activities toward xenoestrogens and decreases during pregnancy. Environ Health Perspect. 2002;110:193-196.

39. Domoradzki JY, Thornton CM, Pottenger LH, et al. Age and dose dependency of the pharmacokinetics and metabolism of bisphenol A in neonatal sprague-dawley rats following oral administration. Toxicol Sci. 2004;77:230-242.

40. Wells J, Koontz. Decline in urinary bisphenol A concentrations in the U.S. Epidemiology. 2013.

41. European Food Safety Authority. Toxicokinetics of bisphenol A: Scientific opinion of the panel on food additives, flavourings, food processing aids, and materials in contact with food (AFC). The EFSA journal. 2008;759:1-10.

42. Kuester RK, Sipes IG. Prediction of metabolic clearance of bisphenol A (4,4 '- dihydroxy-2,2-diphenylpropane) using cryopreserved human hepatocytes. Drug Metab Dispos. 2007;35:1910-1915.

43. Ye X, Bishop AM, Needham LL, Calafat AM. Automated on-line column-switching HPLC-MS/MS method with peak focusing for measuring parabens, triclosan, and other environmental phenols in human milk. Anal Chim Acta. 2008;622:150-156.

Chapter 2

1. U.S. Centers for Disease Control and Prevention. Fourth National Report on Human Exposure to Environmental Chemicals. Department of Health and Human Services; 2009. Available from: http://www.cdc.gov/exposurereport/pdf/FourthReport.pdf. Accessed July 31, 2012.

2. Calafat AM, Ye X, Wong LY, Reidy JA, Needham LL. Exposure of the U.S. population to bisphenol A and 4-tertiary-octylphenol: 2003-2004. Environ Health Perspect. 2008;116:39-44.

3. National Toxicology Program (Center for the Evaluation of Risks to Human Reproduction). NTP-CERHR monograph on the potential human reproductive and developmental effects of bisphenol A. 2008;NIH Publication No. 08 – 5994.

4. Ackerman LK, Noonan GO, Heiserman WM, Roach JA, Limm W, Begley TH. Determination of bisphenol A in U.S. infant formulas: updated methods and concentrations. J Agric Food Chem. 2010;58:2307-2313.

5. Volkel W, Colnot T, Csanady GA, Filser JG, Dekant W. Metabolism and kinetics of bisphenol a in humans at low doses following oral administration. Chem Res Toxicol. 2002;15:1281-1287.

6. Alcorn J, McNamara PJ. Pharmacokinetics in the newborn. Adv Drug Deliv Rev. 2003;55:667-686.

7. Edginton AN, Ritter L. Predicting plasma concentrations of bisphenol A in children younger than 2 years of age after typical feeding schedules, using a physiologically based toxicokinetic model. Environ Health Perspect. 2009;117:645-652.

8. Braun JM, Kalkbrenner AE, Calafat AM, et al. Impact of early-life bisphenol A exposure on behavior and executive function in children. Pediatrics. 2011;128:873-882.

9. Dekant W, Volkel W. Human exposure to bisphenol A by biomonitoring: methods, results and assessment of environmental exposures. Toxicol Appl Pharmacol. 2008;228:114-134.

10. Fox SD, Falk RT, Veenstra TD, Issaq HJ. Quantitation of free and total bisphenol A in human urine using liquid chromatography-tandem mass spectrometry. J Sep Sci. 2011;34:1268-1274.

11. Calafat AM, Weuve J, Ye X, et al. Exposure to bisphenol A and other phenols in neonatal intensive care unit premature infants. Environ Health Perspect. 2009;117:639- 644.

12. Volkel W, Kiranoglu M, Fromme H. Determination of free and total bisphenol A in urine of infants. Environ Res. 2011;111:143-148.

Chapter 3

1. European Commission Joint Research Center (Institute for Health and Consumer Protection). European union risk assessment report: bisphenol-A. Luxembourg: Office for Official Publications of the European Communities; 2003;EUR 20843 EN.

2. Liao C, Kannan K. Widespread occurrence of bisphenol A in paper and paper products: implications for human exposure. Environ Sci Technol. 2011;45:9372-9379.

3. Calafat AM, Ye X, Wong LY, Reidy JA, Needham LL. Exposure of the U.S. population to bisphenol A and 4-tertiary-octylphenol: 2003-2004. Environ Health Perspect. 2008;116:39-44.

4. Bushnik T, Haines D, Levallois P, Levesque J, Van Oostdam J, Viau C. Lead and bisphenol A concentrations in the Canadian population. Health Rep. 2010;21:7-18.

5. Koch HM, Kolossa-Gehring M, Schroter-Kermani C, Angerer J, Bruning T. Bisphenol A in 24 h urine and plasma samples of the German Environmental Specimen Bank from

1995 to 2009: a retrospective exposure evaluation. J Expo Sci Environ Epidemiol. 2012;22:610-616.

6. Zhang Z, Alomirah H, Cho HS, et al. Urinary bisphenol A concentrations and their implications for human exposure in several Asian countries. Environ Sci Technol. 2011;45:7044-7050.

7. Wilson NK, Chuang JC, Morgan MK, Lordo RA, Sheldon LS. An observational study of the potential exposures of preschool children to pentachlorophenol, bisphenol-A, and nonylphenol at home and daycare. Environ Res. 2007;103:9-20.

8. von Goetz N, Wormuth M, Scheringer M, Hungerbuhler K. Bisphenol A: how the most relevant exposure sources contribute to total consumer exposure. Risk Anal. 2010;30:473-487.

9. Lang IA, Galloway TS, Scarlett A, et al. Association of urinary bisphenol A concentration with medical disorders and laboratory abnormalities in adults. JAMA. 2008;300:1303-1310.

10. Braun JM, Kalkbrenner AE, Calafat AM, et al. Impact of early-life bisphenol A exposure on behavior and executive function in children. Pediatrics. 2011;128:873-882.

11. Acevedo N, Davis B, Schaeberle CM, Sonnenschein C, Soto AM. Perinatally Administered Bisphenol A Acts as a Mammary Gland Carcinogen in Rats. Environ Health Perspect. 2013.

12. Trasande L, Attina TM, Trachtman H. Bisphenol A exposure is associated with low- grade urinary albumin excretion in children of the United States. Kidney Int. 2013;83:741-748.

13. Volkel W, Colnot T, Csanady GA, Filser JG, Dekant W. Metabolism and kinetics of bisphenol A in humans at low doses following oral administration. Chem Res Toxicol. 2002;15:1281-1287.

14. National Toxicology Program (Center for the Evaluation of Risks to Human Reproduction). NTP-CERHR monograph on the potential human reproductive and developmental effects of bisphenol A. 2008;NIH Publication No. 08 – 5994.

15. Selevan SG, Kimmel CA, Mendola P. Identifying critical windows of exposure for children's health. Environ Health Perspect. 2000;108 Suppl 3:451-455.

16. McCarver DG, Hines RN. The ontogeny of human drug-metabolizing enzymes: phase II conjugation enzymes and regulatory mechanisms. J Pharmacol Exp Ther. 2002;300:361-366.

17. Watchko JF, Lin Z, Clark RH, et al. Complex multifactorial nature of significant hyperbilirubinemia in neonates. Pediatrics. 2009;124:e868-77.

18. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult CD-1 mice: inter-species comparisons with Sprague- Dawley rats and rhesus monkeys. Toxicol Lett. 2011;207:298-305.

19. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult Sprague-Dawley rats. Toxicol Appl Pharmacol. 2010;247:158-165.

20. Doerge DR, Twaddle NC, Woodling KA, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult rhesus monkeys. Toxicol Appl Pharmacol. 2010;248:1-11.

21. Edginton AN, Ritter L. Predicting plasma concentrations of bisphenol A in children younger than 2 years of age after typical feeding schedules, using a physiologically based toxicokinetic model. Environ Health Perspect. 2009;117:645-652.

22. Mielke H, Gundert-Remy U. Bisphenol A levels in blood depend on age and exposure. Toxicol Lett. 2009;190:32-40.

23. Dekant W, Volkel W. Human exposure to bisphenol A by biomonitoring: methods, results and assessment of environmental exposures. Toxicol Appl Pharmacol. 2008;228:114-134.

24. Calafat AM, Kuklenyik Z, Reidy JA, Caudill SP, Ekong J, Needham LL. Urinary concentrations of bisphenol A and 4-nonylphenol in a human reference population. Environ Health Perspect. 2005;113:391-395.

25. Centers for Disease Control and Prevention. Fourth National Report on Human Exposure to Environmental Chemicals, updated tables, February 2012. 2012.

26. Health Canada. Second Report on Human Biomonitoring of Environmental Chemicals in Canada. 2013.

27. Calafat AM, Weuve J, Ye X, et al. Exposure to bisphenol A and other phenols in neonatal intensive care unit premature infants. Environ Health Perspect. 2009;117:639- 644.

28. Volkel W, Kiranoglu M, Fromme H. Determination of free and total bisphenol A in urine of infants. Environ Res. 2011;111:143-148.

29. Mendonca K, Hauser R, Calafat AM, Arbuckle TE, Duty SM. Bisphenol A concentrations in maternal breast milk and infant urine. Int Arch Occup Environ Health. 2012.

30. Nachman RM, Fox SD, Golden WC, et al. Urinary free bisphenol A and bisphenol A- glucuronide concentrations in newborns. J Pediatr. 2013;162:870-872.

31. Alcorn J, McNamara PJ. Pharmacokinetics in the newborn. Adv Drug Deliv Rev. 2003;55:667-686.

32. Hajszan T, Leranth C. Bisphenol A interferes with synaptic remodeling. Front Neuroendocrinol. 2010;31:519-530.

33. National Institute for Health and Care Excellence (NICE), National Health Service (NHS) (United Kingdom). Neonatal Jaundice. 2010.

34. American Academy of Pediatrics. Committee on Practice and Ambulatory Medicine and Bright Futures Steering Committee. Recommendations for preventive pediatric health care. Pediatrics. 2007;120:1376.

35. Fox SD, Falk RT, Veenstra TD, Issaq HJ. Quantitation of free and total bisphenol A in human urine using liquid chromatography-tandem mass spectrometry. J Sep Sci. 2011;34:1268-1274.

36. Grasselli J, ed. Atlas of Spectral Data and Physical Constants for Organic Compounds. Cleveland: CRC Press, Inc.; 1975.

37. Lakind JS, Naiman DQ. Bisphenol A (BPA) daily intakes in the United States: estimates from the 2003-2004 NHANES urinary BPA data. J Expo Sci Environ Epidemiol. 2008;18:608-615.

38. Valentin J. Basic anatomical and physiological data for use in radiological protection: reference values: ICRP Publication 89. Annals of the ICRP. 2002;32:1-277.

39. Olsen LH, Grothe I, Rawashdeh YF, Jorgensen TM. Urinary flow patterns in first year of life. J Urol. 2010;183:694-698.

40. Ye X, Zhou X, Hennings R, Kramer J, Calafat AM. Potential external contamination with bisphenol A and other ubiquitous organic environmental chemicals during

95 biomonitoring analysis: an elusive laboratory challenge. Environ Health Perspect. 2013;121:283-286.

41. European Food Safety Authority. Opinion of the scientific panel on food additives, flavourings, food processing aids, and materials in contact with food (AFC). The EFSA journal. 2006;759:1-10.

42. Layton L. No BPA for baby bottles in US: 6 makers announce decision on chemical. Washington Post. March 6, 2009 2009;http://www.washingtonpost.com/wp- dyn/content/article/2009/03/05/AR2009030503285.html.

43. 77 FR 41899 (July 17, 2012) (Abandonment of BPA use in baby bottles).

44. Ackerman LK, Noonan GO, Heiserman WM, Roach JA, Limm W, Begley TH. Determination of bisphenol A in U.S. infant formulas: updated methods and concentrations. J Agric Food Chem. 2010;58:2307-2313.

45. 78 FR 41840 (July 12, 2013) (Abandonment of BPA in formula packaging).

46. Mayo Clinic Medical Laboratories. Test ID: EEST (Estradiol, Serum). Available at: http://www.mayomedicallaboratories.com/test- catalog/setup.php?unit_code=81816&format=pdf. Accessed 10/01, 2013.

47. Elmlinger MW, Kuhnel W, Ranke MB. Reference ranges for serum concentrations of lutropin (LH), follitropin (FSH), estradiol (E2), prolactin, progesterone, sex hormone- binding globulin (SHBG), dehydroepiandrosterone sulfate (DHEAS), cortisol and ferritin in neonates, children and young adults. Clin Chem Lab Med. 2002;40:1151-1160.

48. Gassler N, Peuschel T, Pankau R. Pediatric reference values of estradiol, testosterone, lutropin, follitropin and prolactin. Clin Lab. 2000;46:553-560.

49. Guignard JP, Drukker A. Why do newborn infants have a high plasma creatinine? Pediatrics. 1999;103:e49.lactin. Clin Lab. 2000;46:553-560.

Conclusions

1. Volkel W, Kiranoglu M, Fromme H. Determination of free and total bisphenol A in urine of infants. Environ Res. 2011;111:143-148.

2. Mendonca K, Hauser R, Calafat AM, Arbuckle TE, Duty SM. Bisphenol A concentrations in maternal breast milk and infant urine. Int Arch Occup Environ Health. 2012.

3. Calafat AM, Weuve J, Ye X, et al. Exposure to bisphenol A and other phenols in neonatal intensive care unit premature infants. Environ Health Perspect. 2009;117:639- 644.

4. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult CD-1 mice: inter-species comparisons with Sprague- Dawley rats and rhesus monkeys. Toxicol Lett. 2011;207:298-305.

5. Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult Sprague-Dawley rats. Toxicol Appl Pharmacol. 2010;247:158-165.

6. Doerge DR, Twaddle NC, Woodling KA, Fisher JW. Pharmacokinetics of bisphenol A in neonatal and adult rhesus monkeys. Toxicol Appl Pharmacol. 2010;248:1-11.

8. Mielke H, Gundert-Remy U. Bisphenol A levels in blood depend on age and exposure. Toxicol Lett. 2009;190:32-40.

9. Ye X, Zhou X, Hennings R, Kramer J, Calafat AM. Potential external contamination with bisphenol A and other ubiquitous organic environmental chemicals during biomonitoring analysis: an elusive laboratory challenge. Environ Health Perspect. 2013;121:283-286.

10. National Toxicology Program (Center for the Evaluation of Risks to Human Reproduction). NTP-CERHR monograph on the potential human reproductive and developmental effects of bisphenol A. 2008;NIH Publication No. 08 – 5994.

APPENDIX I: Exclusion Criteria

Exclusion Criteria Justification Premature infants (less than 37 weeks and Physiologic differences between 0/7 days gestational age) premature and full term infants may impact BPA metabolism Infants who are small for gestational The function of the liver in infants who age/intrauterine growth retarded are SGA, IUGR or LGA may differ from (SGA/IUGR) infants or large for gestational liver function in an infant of average size. age (LGA) Infants with delayed voiding or stooling (> This condition may lead to high levels of 24 hours of age) bilirubin in the blood, which may impact BPA metabolism. Infants with prenatally or postnatally These conditions may impact BPA diagnosed renal or hepatic abnormalities metabolism. Infants with high bilirubin due to High bilirubin may impact BPA  blood group incompatibility (ABO or metabolism given similarities between the Rh) (Phase 1 only) metabolic pathways of the two  high rbc load (infants of diabetic substances. mothers )  red blood cell extravasation (cephalohematoma, bruising) Infants at high risk for developing jaundice These conditions may lead to high levels in first year of life due to of bilirubin in the blood, which may  Crigler-Najaar syndrome types I and impact BPA metabolism. II  Lucey Driscoll syndrome  congenital hypothyroidism  pyloric stenosis Infants with APGAR scores less that 5 at 5 Indicate infant may be at risk for minutes) ischemia. This condition may lead to high levels of bilirubin in the blood, which may impact BPA metabolism. Mother is a smoker. Nicotine is a known inducer of UGT1A8, and UGT1A10. The role of these two UGT isoforms in BPA metabolism is unknown. If they were to play a role in BPA metabolism, exposure to infants to nicotine in utero could impact infant metabolism of BPA in the newborn.

Infants of mothers on barbiturates or other Barbiturates are known inducers of antiepileptic/anticonvulsant medication. UGT2B15, the UGT isoform mainly responsible for the metabolism of BPA. Antiepileptic and anticonvulsant medications are known inducers of other UGT isoforms that may play a role in BPA metabolism. In utero exposure and exposure through breast milk to these medications may impact infant metabolism of BPA. Infants of mothers with a positive drug THC, heroin, methadone, and psilocybin screening test. (mushrooms) undergo glucuronidation. In utero exposure and exposure to these substances through breast milk may impact infant metabolism of BPA. Infants who will not receive pediatric care at Dr. Erica Sibinga, pediatrician at the the Harriet Lane Pediatric Clinic. HLPC, is a collaborator and co- investigator on the study. She will facilitate communication between JHSPH investigators and the pediatric residents, faculty and nurse practitioners at HLC who care for participating patients.

APPENDIX II: Sample Collection Form and Questionnaires

100

JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH

Infant Urine Sample Collection Form

Study Title: Development of BPA Metabolizing Enzymes in Infants Principal Investigator: Peter S. J. Lees Student Investigator: Rebecca M. Nachman IRB No.: IRB00003838 PI Version Number/Date: Version 1 / September 6, 2011

Date: ______Technician name: ______

Participant ID: ______

Age of infant today: ______days

Comments: ______

______

101

JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH

Research Visit Interview Questionnaire (to be completed at Visit 2)

Interviewer name: ______Age of infant today: ______days Participant ID: ______Sex of child: ______

Q1a. What did your baby drink today? (check all that apply) _ breast milk _ formula _ other specify: ______

Q1b. If the baby drank formula, it was (check all that apply) _ powder _ “ready to feed” (no need to add water)

Q1c. The baby drank milk or formula from a cup or bottle made of (check all that apply) _ glass _ plastic _ the baby drank directly from the breast

Q2a. What food did your baby eat today? _ infant cereal (dry cereal with liquid added just before eating) _ Baby food sold in a store (in jar or other container, not home-made) _ Other: (List other foods not in the above categories): ______

Q2b. If the baby ate cereal, it was mixed with (check all that apply) _ breast milk _ formula _ water

Q3. Today the baby did these things _ Crawling or playtime on the floor _ Mouthing of toys and objects _ Fed him- or herself using own hands 102

Q4. What kind of diaper was/is your baby wearing while the urine collection the bag was/is on him/her? _ cloth diaper with no disposable portion _ cloth diaper with a disposable insert _ disposable diapers

Q5. Has your doctor told you your baby is yellow? _ yes _ no

Q6. Is your child currently taking or has your child taken any medicine since the last interview? _ yes (Please list the medications and the dates they were taken below.) _ no

Name of Medication Dates Taken*

* Provide a date to the best of your memory even if it may not be the exact date.

103

JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH

Research Visit Interview Questionnaire (to be completed at Visit 2)

Interviewer name: ______Age of infant today: ______days Participant ID: ______Sex of child: ______

Q1-. What did your baby drink today? (check all that apply) _ breast milk _ formula _ other specify: ______

Q2. If the baby drank formula, it was (check all that apply) o powder o “ready to feed” (no need to add water)

Q3. If the baby drank powder formula, it was mixed with: _ Tap water _ Bottled water Brand: ______Other Specify: ______

Q4. The baby drank milk or formula from a cup or bottle made of (check all that apply) _ glass _ plastic Brand: ______Year of purchase: ______the baby drank directly from the breast

Q5. What kind of diaper was/is your baby wearing while the urine collection the bag was/is on him/her? _ cloth diaper with no disposable portion _ cloth diaper with a disposable insert _ disposable diapers

Q6. Has your doctor told you your baby is yellow? _ yes _ no 104

Q7: Which of these best describes your child’s race? _ American Indian or Alaska Native _ Asian _ Black or African American _ Hispanic or Latino _ Native Hawaiian or Other Pacific Islander _ White _ Other _ Unknown

Q8. Is your child currently taking or has your child taken any medicine since the last interview? _ yes (Please list the medications and the dates they were taken below.) _ no Name of Medication Dates Taken* (mm/dd/yyyy – mm/dd/yyyy)

* If you don’t remember the exact date, provide a date to the best of your memory even if it may not be the exact date.

105

JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH

Research Visit Interview Questionnaire (to be completed at Visit 2)

Interviewer name: ______Age of infant today: ______days Participant ID: ______Sex of child: ______

Q1-. What did your baby drink today? (check all that apply) _ breast milk _ formula _ other specify: ______

Q2. If the baby drank formula, it was (check all that apply) _ powder _ “ready to feed” (no need to add water)

Q3. If the baby drank powder formula, it was mixed with: _ Tap water _ Bottled water Brand: ______Other Specify: ______

Q4. The baby drank milk or formula from a cup or bottle made of (check all that apply) _ glass _ plastic Brand: ______Year of purchase: ______the baby drank directly from the breast

Q4a. How many ounces of formula and/or breast milk does your baby typically drink in one feeding? ______

Q4b. How many ounces of formula and/or breast milk does your baby typically drink in 24 hours? ______

106

Q4c. How many wet diapers does your baby typically have in 24 hours? ______

Q6. Has your doctor told you your baby is yellow? _ yes _ no

Q8. Is your child currently taking or has your child taken any medicine since the last interview? _ yes (Please list the medications and the dates they were taken below.) _ no

Name of Medication Dates Taken* (mm/dd/yyyy – mm/dd/yyyy)

* Provide a date to the best of your memory even if it may not be the exact date.

107

APPENDIX III: Reducing Opportunities for Sample Contamination during Sample

Handling, Collection and Storage

Contamination of samples with BPA from field and laboratory sources is a common problem an impediment to the interpretation of BPA biomarker data. The standard protocol for dealing with background levels of an analyte of interest, that is, the analysis of field blanks to determine a mean background level and blank-correction of sample measurements, is not possible in the context of BPA biomonitoring because the contamination occurs sporadically at levels well within the range of concentrations found in human biological samples. Thus it is imperative that measures be taken to reduce and if possible eliminate the incidence of sample contamination in the field and laboratory.

This Appendix contains details on the procedures that were followed in the field and laboratory in order to avoid contamination of samples.

Suspected Sources of BPA Contamination

The following were considered to be potential sources of BPA contamination.

Each is discussed in more detail throughout this appendix.

 Dust and fibers

 Paper

 Plastic

 Reagents and chemicals

 Tap water

 Personal care products

 Glass (prior to cleaning with hexane)

108

Dust and Fibers

Measures were taken to keep the bench as dust free as possible by regularly wiping down surfaces with deionized water and Kimwipes (Kimberly-Clark Professional,

Roswell, GA). Ethanol and other organic solvents were not used to wipe down surfaces just prior to handling of samples since they may cause BPA to leach from surfaces that contain BPA. The bench surface, pipetters, pipette stands, pipette bulbs, electrical cords, bottles of chemicals, and the outer surfaces of packages of supplies are examples of surfaces that were regularly wiped down with Kimwipes dampened with deionized water.

Blue absorbent pads were not used in order to keep fibers from the pads away from the work space. Instead, work surfaces were lined with aluminum foil to catch small drips and spills. During sample collection and handling, care was taken not to pass items over the tops of open vials. Glass vials and pipettes were kept covered at all times except for brief moments when in use.

Paper

To avoid BPA from paper, receipts were handled as little as possible by investigators and were kept away from the laboratory and field workspaces. Paper towels were used sparingly and Kimwipes were used to wipe down surfaces. Gloves were pulled slowly from their cardboard packaging to avoid release of cardboard particles into the air.

Items arriving in cardboard boxes were wiped down with damp Kimwipes.

Polypropylene sample storage boxes (Fisherbrand, Leicestershire, England) were used to store vials in the -80°C freezer instead of the paper boxes or boxes with polycarbonate lids.

109

Plastic

All plastic equipment was considered to be a potential source of BPA, even if was made of polyethylene or polypropylene, which are not made from BPA. Polycarbonate was not used at any time. Equipment made from polyethylene or polypropylene was washed with water only. Solvents were kept out of contact with plastic. If plastic items were notably dusty, they were wiped with a Kimwipe dampened with deionized water or blown clean with a stream of nitrogen gas. The urine bag manufacturer was not able to provide information about the material from which the bags were made, but they were similar in appearance to bags made from polyethylene.

Reagents and Chemicals

Solvents used to clean glassware and water used to make synthetic urine were

HPLC grade. The purest available form of each chemical was used to in the making of the synthetic urine. Powder chemicals were weighed in glass scintillation vials that had been previously cleaned with hexane.

Tap Water

Generally, tap water was not used in the work space. Items that were hand- washed such as reusable pipettes, large glassware, and polypropylene storage boxes were washed with Micro 90 cleaning solution (Cole-Parmer, Vernon Hills, IL) in tap water, but were then rinsed thoroughly with deionized water. For glass, this procedure was also followed up with hexane washing according to the instructions below.

110

Personal Care Products

Gloves were always used when handling clean glassware to avoid contamination by personal care products or hand soaps.

Glass

This research involved the abundant use of glassware including Pasteur pipettes, reusable serological pipettes, autosampler vials, large wide mouthed vials (i.e. the transfer vials used to hold full urine collection bags), volumetric flasks and solvent bottles with glass stoppers, beakers, funnels, conical tubes and vials, and scintillation vials. All glassware was first cleaned by routine cleaning procedures (either in the sink with Micro 90 soap or in a large automatic laboratory glass washer). Glassware was then rinsed thoroughly with HPLC-grade hexane and dried with a stream of nitrogen gas under a ventilation hood.

Glass cleaning with hexane was done in large batches to conserve materials. Used hexane was stored in a solvent bottle covered in foil for reuse over a period of a few days.

Two sets purple nitrile gloves were worn during hexane cleaning, and gloves were changed if they became wet. Metal tongs were used when necessary to avoid getting hexane on the gloves. To keep the hexane as BPA-free as possible, plastic wash bottles typically used in the lab were not used for hexane washing. Instead, the cleaning was accomplished by pouring hexane from one glass vessel to the next under a ventilation hood. A layer of foil (dull side up) was used to catch inadvertent spills.

111

Several beakers and graduated cylinders of various sizes were always washed first. Beakers were stored upside-down in the hood and used for a month at a time. The wide-mouthed 4-ounce glass vials (Supelco, Bellefonte, Pennsylvania) used to transfer full urine collection bags from the clinic to the lab were purchased pre-cleaned. However, the outer surfaces were dusty upon arrival, thus the outside of each vial was first wiped down with a Kimwipe dampened with deionized water. The caps were removed and set aside while the vials were cleaned with hexane and blown dry with nitrogen gas. The caps were blown free of dust using a stream of nitrogen gas. The cleaned capped vials were individually stored in a zip lock polyethylene bag until use in the field.

Glass serological and Pasteur pipettes were washed by placing them into a graduated cylinder filled with hexane. Each pipette was then blown dry with a stream of nitrogen gas. Glass pipettes were stored in a glass graduated cylinder with an overturned glass beaker on top to keep dust out.

The certified pre-cleaned autosampler vials (National Scientific Company,

Rockwood, TN) that were used to store the samples until analysis were not hexane- cleaned. They were stored in their original packaging until use, then blown out with nitrogen and stored under an overturned beaker. Caps were also blown out with nitrogen gas.

112

APPENDIX IV: Use of Pediatric Urine Collection Bags for the Collection of Urine

Samples from Neonates and Young Infants

Pediatric urine collection bags (U-Bag, Hollister, Inc. Libertyville, IL) were used for the collection of samples from subjects who participated in our research on BPA exposure and metabolism. (Figure 1) Discussion of the testing of these bags for the presence of BPA is found in Appendix V. This section describes how to use the bag and tips for successfully and quickly obtaining urine samples from newborns using the bags.

When using urine collection bags for biomonitoring, ideally the investigator will place the bag on the infant and will remove it immediately after the baby voids. Handling of the bag by the investigator and immediate removal of the bag after a void reduces the chance for leakage of the sample, sample contamination or degradation of the analyte(s), all of which could occur if a parent puts the bag on the infant in advance of meeting with the investigator.

The bags come in two sizes: newborn and pediatric. The newborn bags were recently redesigned and the newly designed newborn bag is larger and fits better than the old newborn bag has two layers, which helps prevent leaks. More than one bag should be stocked for each sample as the adhesive on the bag often becomes soiled from sweat or ointment making it necessary to remove and replace the bag one or more times.

Requiring that the investigator be present when the bag is placed on the infant as well as when the infant voids has the potential to be time-consuming for the investigator and participants. It was observed during research visits that infants often urinated upon waking and being undressed. Thus, parent were instructed to allow the infant to sleep until taken to an examination room to be undressed and weighed. By adhering to this

113 protocol, it was possible in many cases to obtain a sample within minutes of placing the bag on the infant. To reduce leakage, the sides of the bag were eased apart before placing the bag on the infant. Cleaning the skin with a wipe also helped to reduce leaks in some cases, but only if there was a long wait for the baby to void; taking time to wipe the skin before placing the bag can lead to a lost sample.

After the infant voids, it is important that the bag be removed immediately and placed in a cooler to prevent leakages and/or degradation of compounds in the urine.

Once the bag is removed, the tabs can be folded upon themselves to form a seal. This seal is not perfect, so it is recommended that the bag then be placed upright in a tall container with a lid. A four-ounce wide-mouthed glass vial with a Teflon coated lid was used for this purpose in our study. The glass vial was placed in a zip-locked plastic bag to further prevent sample contamination during transfer of the sample to the lab. When a sample was not obtained early in the appointment, parents were encouraged to feed the baby to keep the baby hydrated. A wait of up to 1 hour was not uncommon if the sample was not obtained in the first few minutes of the appointment. In some cases, the wait was as long as 3 hours.

114

Figure 1. Pediatric Urine Collection Bag

115

APPENDIX V:

Quality Control Sample Preparation Procedure for the Measurement of BPA and

BPA Conjugates in Infant Urine

1. Test Principle

Measurements of Total Bisphenol A (BPA) and free BPA in human urine reflect recent exposure to and metabolism of BPA. Measurement of Total and free BPA in humans between birth and 12 months is desirable because of a critical window when infants may be more susceptible to developmental effects associated with BPA exposure. Collection of urine samples from infants requires specialty equipment that differs from the equipment used to collect adult samples. This set of procedures outlines steps for the preparation of quality control (QC) blanks to test for possible BPA contamination from infant urine sampling equipment.

2. Data Management

Samples will be assigned a Sample ID at random. Investigators performing the lab analysis will be blind to the contents of the QC samples until after the lab analysis has been completed.

3. Preparation of Reagents and Other Materials

A. Synthetic Urine (Source: 2005 CDC Laboratory Procedure Manual: Bisphenol A)

Prepare synthetic urine in a 1L glass volumetric flask (previously cleaned with

hexane per section 4) with a glass stopper.

116

1. 500 mL HPLC water

2. 3.8 g Potassium Chloride

3. 8.5 g Sodium Chloride

4. 24.5 g Urea

5. 1.03 g Citric Acid

6. 0.34 g Ascorbic Acid

7. 1.18 g Potassium Phosphate

8. 1.4 g Creatinine

9. 0.64 g Sodium Hydroxide (add slowly)

10. 0.47 g Sodium Bicarbonate

11. 0.28 mL Sulfuric Acid

12. Fill up to 1 L with HPLC water

Solution is good for 1 month when refrigerated at 4C. Check that urine pH is in physiologic range and adjust if necessary. Do not dip anything into the master mixture of synthetic urine. To test the pH, our a few drops out of the flask, onto a piece of pH paper.

B. Other Materials

1. Mass Spec Certified Glass Autosampler Vial Kit, clear, 2 mL, no crimping,

PTFE coated blue cap (National Scientific, Rockwood, TN)

117

2. Ten 2 mL glass reusable serological pipettes (KIMAX-51. 2mL in 1/100, No.

37034)

3. Forty “U-Bag” Single Specimen Urine Collection Bags (Fig 1) (Item # 7511

and 7515) (Mabis Healthcare, Waukegan, IL)

4. Socorex Calibra Digital 832 Macropipette (0.2 – 2mL) with pastuer pipette

adaptor

5. Standard 150 mm glass pastuer pipettes

6. Purple Nitrile gloves, #G4162 (Denville Scientific, Metuchen, NJ)

7. A 1 liter volumetric flask with glass stopper

8. Ten 50 mL beakers (for testing baby wipes)

9. Beakers for holding urine collection bags upright after filled with synthetic

urine

10. 500 mL beakers to hold the master mixture while pipetting

11. Various cleaned beakers and graduated cylinders to hold clean glassware and

pipettes and protect from dust contamination

12. Reagent bottle large enough for 100 mL or more.

4. Glassware Cleaning

Start with clean, dry glassware. Wash all glassware (except autosampler vials and transfer vial caps) with HPLC grade hexane and dry in oven at 50C or with a stream of

Nitrogen gas. Never solvent-rinse plastics at it will increase leaching of BPA from the plastic. Blow out the autosampler vials and caps with Nitrogen gas. Blow out the transfer vial caps with hexane (blow out with Nitrogen gas).

118

5. Preparation of QC samples

Note: In an initial round of testing, the serological pipettes, the synthetic urine mixture, and baby wipes were found to be BPA free (Figure 1). A second round testing was done to test bags from the lot used to collect samples for Phase II. Also a newly purchased

Socorex pipette was tested in the second round of testing (Figure 2).

1. Clean all glassware with hexane as specified above.

2. Set aside and refrigerate 100 mL of the synthetic urine in a glass reagent bottle to

be used to make positive controls.

3. Heat synthetic urine to 37 oC in a volumetric flask with a stopper using the

incubator shaker (Innova 4200, New Brunswick Scientific). This will take several

hours. To test the temperature, fill a 2nd 1L volumetric flask with water and heat it

alongside the synthetic urine. A thermometer can be inserted into the flask filled

with water rather than the synthetic urine to avoid contamination of the test urine.

4. Rinsate blanks will be prepared to test the following potential sources of BPA

contamination during collection of urine samples from infants:

 Synthetic urine mixture

 Urine collection bags, newborn size

 Urine collection bags, pediatric size

 Reusable glass serological pipettes

 Baby wipes

 Socorex pipetter/pastuer pipettes

119

Instructions by type of blank:

a. QC blanks for synthetic urine mixture.

Pour enough synthetic urine (previously warmed to 37C) from the volumetric

flask into a hexane-cleaned beaker to create the desired number of blanks (see

Figures 2 and 3). Recap the volumetric flask and place a larger inverted clean

beaker over the top of the beaker of synthetic urine when not pipetting in

order to keep dust from getting into mixture. Place the master mixture back in

the incubator to maintain its temperature while it is not is use. Using a clean

reusable borosilicate glass serological pipette, transfer 0.5 mL mixture into

each of 10 autosampler vials. Label the vials with randomly assigned sample

IDs. Also, create a code sheet for the sample IDs and track the order in which

the vials were filled (in case the mixture becomes contaminated during

pipetting).

b. QC blank newborn collection bags.

Prepare another beaker of the warmed synthetic urine from which to pipette

the mixture into the urine collection bags. Place the master mixture back in the

incubator to maintain its temperature when it is not in use. Using clean

reusable borosilicate glass serological pipette, transfer 5-10 mL synthetic

urine (37 oC) into each urine collection bag. Close and seal bag by shaking the

urine to the bottom of the bag and folding the adhesive in half onto itself.

Place the bag upright in a cleaned beaker. After 15 mins, hold the bag upright

and shake until urine collects at the bottom of the bag. Open the bag valve on

120

bottom corner by removing the blue sticker to uncover a hole in the bag.

Pipette 0.5 – 1 mL liquid into each of 4 autosample vials and label. Two will

be analyzed and two are backup samples to be stored. Test a total of 10

newborn sized bags and 10 pediatric sized bags, resulting in 4 samples from

each bag (a set a duplicates and 2 backup samples). As with the master

mixture blanks, label the vials with sample ID and add the sample IDs to the

code sheet, tracking the order in which the vials were filled. c. QC blank new pipette.

Pour enough synthetic urine from the volumetric flask into a hexane-cleaned

beaker to create the desired number of blanks Label the vials with sample ID

and add the sample IDs to the code sheet, tracking the order in which the vials

were filled. d. QC blank for baby wipe

Baby wipes tested previous using a slightly different protocol included here

for reference. Ten cleaned beakers were filled with 50 mL synthetic urine

each. In each beaker, a single baby wipe was submerged. The beaker was

gently swirled and allowed to sit for 5 minutes. After 5 minutes,

approximately 1.5 mL was poured directly from each beaker into a 1.5 mL

sample vial. e. Store all QC samples at -80 oC until analysis.

121

Figure 1. Pediatric Urine Collection Bag.

122

Figure 2. Potential Sources of BPA Contamination during Urine Collection

Bag N

Skin Washed Transfer Sample Urine Pipette with Wipe Vial* Vial

Bag P

* The urine collection bags were placed upright in a 4 oz precleaned glass wide mouthed vial (Supelco, Bellefonte, Pennsylvania) with a Teflon coated cap during transfer back to the lab.

123

Figure 3. Previous test of serological pipettes and baby wipes

Sample Vial Master Mixture Serological Beaker Pipette Sample Vial

Beaker with Sample Wipe Vial

124

Figure 4. Preparation of 3 Different Types of QC Blanks for Further Testing of Sample Collection Equipment

Autosampler Vials (N =5 for analysis, N=5 backup)

Beaker Reusable 10 New- Autosampler Vial (500 mL) Glass born (N= 10 X 2 for analysis,

Serological Bags N=10 x 2 for backup) Cover Pipette Master With Mixture Autosampler Vials Larger (N =2 for analysis, N=2 backup) 1 L beaker

10 New- Autosampler Vial born (N= 10 X 2 for analysis, Bags N=10 x 2 for backup)

Autosampler Vials

(N =2 for analysis, N=2 backup)

New Pipette Autosampler Vial (N= 10 X 2 for analysis, N=10 x 2 for backup)

125

CURRICULUM VITAE

REBECCA M. NACHMAN

Department of Environmental Health Sciences

615 North Wolfe Street, Room E7402

Baltimore, MD 21205 [email protected]

EDUCATION

Johns Hopkins Bloomberg School of Public Health, Baltimore, MD

Doctor of Philosophy (PhD) Candidate, Environmental Health Sciences, October 2013

Thesis Title: Biomarkers of Bisphenol A Exposure and Metabolism in a Population of

Neonates in Baltimore, Maryland

Advisor: Peter S.J. Lees, PhD

Johns Hopkins Bloomberg School of Public Health, Baltimore, MD

Master of Public Health (MPH), May 2007

Thesis Title: Association between Personal and Ambient Exposures to Fine Particulate

Matter in Congestive Heart Failure Patients in Baltimore, MD

Advisor: Alison Geyh, PhD

State University of New York (SUNY), Geneseo, Geneseo, NY

Bachelor of Arts (BA), English, May 1998

126

GRANTS AND AWARDS

NIEHS Training Program in Environmental Health -- Pre-doctoral Trainee

2010 – 2012

Johns Hopkins Bloomberg School of Public Health

Reach the Decision Makers Training Program – Fellow

2011

University of California, San Francisco, Program on Reproductive Health and the

Environment

Center for a Livable Future – Pre-Doctoral Fellow

2010 – 2013

Johns Hopkins Bloomberg School of Public Health

Wendy Klag Memorial Fund Award

2012 – 2013

Johns Hopkins Bloomberg School of Public Health

The Dr. C. W. Krusé Memorial Fund

2012 – 2013

Johns Hopkins Bloomberg School of Public Health

127

PROFESSIONAL EXPERIENCE

Internship, Office of Research and Development

Summer 2007

U.S. Environmental Protection Agency, Washington , DC

Research Assistant, Risk Sciences and Public Policy Institute

April 2003 – May 2007

Johns Hopkins Bloomberg School of Public Health, Baltimore, MD

TEACHING EXPERIENCE

Teaching Assistant, Industrial Hygiene Laboratory (5 credits)

Spring 2012

Johns Hopkins Bloomberg School of Public Health

Course Instructors: Dr. Ana Rule and Dr. Peter S.J. Lees

Teaching Assistant, Environmental Health (5 credits)

Summer 2008

Johns Hopkins Bloomberg School of Public Health

Course Instructor: Dr. John Links

128

Teaching Assistant, Public Health Practice (4 credits)

Fall 2006

Johns Hopkins Bloomberg School of Public Health

Course Instructor: Dr. Lynn Goldman

SERVICE

Academic Ethics Committee

2006-2007

Johns Hopkins Bloomberg School of Public Health

CONFERENCE PRESENTATIONS AND POSTER SESSIONS

Nachman RM, Cuevas AJ, Canvasser J, Nelson B, Schwingl P. 2012. Advocating for reproductive health protection by chemical regulation under the Toxic Substances

Control Act (TSCA). Oral presentation at the American Public Health Association

Annual Meeting, Oct 27 – 31, 2012.

Ramos-Bonilla JP, White RH, Nachman RM, Samet JS, Breysse PN. 2009. Use of personal monitoring equipment in the assessment of ambient air pollution exposure.

Poster presented at the International Society of Exposure Science Conference, Nov 1 – 5,

2009.

129

Nachman RM and Geyh AS. Use of ambient concentrations of PM2.5 as surrogates for personal exposure in Baltimore, MD. Poster presented at Delta Omega Poster

Competition, February 2009.

PEER-REVIEWED PUBLICATIONS

Nachman RM, Fox SD, Golden WC, Sibinga E, Veenstra TD, Groopman JD, Lees PS.

2013. Urinary Free Bisphenol A and Bisphenol A-Glucuronide Concentrations in

Newborns. J Pediatr 162(4):870-2.

130