Centre for Drug research Division of Pharmaceutical Chemistry

Faculty of Pharmacy University of Helsinki

Finland

UDP-GLUCURONOSYLTRANSFERASES IN SICKNESS AND HEALTH

STUDIES ON BILIRUBIN AND STEROID GLUCURONIDATION

Nina Sneitz

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Pharmacy of the University of Helsinki, for public examination in Auditorium 1,

Korona Information Centre, on 25 October 2013, at 12 noon.

Helsinki 2013

Supervisor: Dr. Moshe Finel Centre for Drug Research (CDR) Faculty of Pharmacy University of Helsinki Finland Reviewers: Professor Peter Mackenzie Department of Clinical Pharmacology School of Medicine Flinders University Australia

Professor Hannu Raunio Faculty of Health Sciences School of Pharmacy University of Eastern Finland Finland Opponent: Professor Philip Lazarus Department of Pharmaceutical Sciences College of Pharmacy Washington State University College of Pharmacy USA © Nina Sneitz 2013 ISBN 978-952-10-9292-3 (Paperback) ISBN 978-952-10-9293-0 (PDF) ISSN 1799-7372 http://ethesis.helsinki.fi/

Helsinki University Printing House Helsinki 2013

3

CONTENTS

Abstract 5 Acknowledgements 7 List of original publications 8 Abbreviations 9 1 Introduction 10 2 Review of the Literature 12 2.1 UDP-glucuronosyltransferases in brief 12 2.2 Bilirubin metabolism and Familial hyperbilirubinemias 15 2.2.1 Bilirubin: properties and metabolism 15 2.2.2 Gilbert syndrome 17 2.2.3 Crigler-Najjar syndromes 18 2.2.3.1 Current therapy for Crigler-Najjar syndrome 20 2.2.3.2 Gene therapy in treatment of Crigler-Najjar syndrome 20 2.3 Steroid hormones and their metabolism 22 2.3.1 Glucuronidation of steroid hormones 24 2.3.2 Steroids and health 26 2.4 Steroid glucuronidation and UGT structure 28 3 Aims of the study 30 4 Materials and Methods 31 4.1 Materials 31 4.1.1. Chemicals 31 4.1.2. Microsomes and RNA sources 33 4.2 Patients 34 4.3. Animal experiments 35 4.4 Methods 36 4.4.1 Molecular biology 36 4.4.1.1 Cloning 36 4.4.1.2. UGT expression in Sf9 cells 36 4.4.1.3. UGT quantification and visualisation 36 4.4.2 Glucuronidation assays 37 4.4.3 Analytical methods and data analysis 37

4.4.4 Molecular modeling 39 5 Results and discussion 40 5.1. UGT1A1 and Crigler-Najjar syndrome 40 5.1.1 UGT1A1 mutations in CN patients in the Netherlands 40 5.1.2 Bilirubin glucuronidating activity of ten UGT1A1 missense mutants 42

4

5.1.3 Activity of UGT1A1 missense mutants toward other substrates 43 5.1.4 Biomarker for efficacy of gene therapy for Crigler-Najjar syndrome 44 5.2 Expression and characterisation of UGT2A2 47 5.3 Glucuronidation of selected steroids 50 5.3.1. Glucuronidation of estradiol, ethinylestradiol and estriol by UGT2A1 and UGT2A2 50 5.3.2. Estrogen glucuronidation and UGT1A10 substrate binding site 50 5.3.3. Glucuronidation of estriol and estradiol stereoisomers and enantiomeric steroids 52 6 Summary and Conclusions 59 References 61 Appendix: Original publications I-VI 75

5

ABSTRACT

UDP-glucuronosyltransferase (UGT) enzymes are responsible for the metabolism of many xenobiotics, as well as several endogenous compounds like bilirubin and certain steroids. This thesis involves three major themes: UGT1A1 and Crigler-Najjar syndrome; expression and characterisation of a novel UGT, UGT2A2; and the glucuronidation of various endogenous and synthetic steroids. Fourteen different UGT1A1 mutations were identified in 19 CN patients, four of which were novel. The mutant proteins were expressed in baculovirus infected Sf9 insect cells and their activity towards bilirubin and other substrates was determined. Residual glucuronidation activity and the severity of hyperbilirubinemia were found to correlate well in most cases. In addition, a major advancement to aid the development of gene therapy for CN was achieved by in discovering that ezetimibe glucuronidation correlates with reduction of serum bilirubin in gene therapy treated Gunn rats. This compound might be useful in examining the efficacy of gene therapy in humans in the future. When the novel UGT2A2 was cloned, expressed, and studied alongside UGTs 2A1 and 2A3, results indicated that UGT2A2 is a functional enzyme with broad substrate specificity. The previously suggested exon sharing between UGT2A1 and UGT2A2 was also confirmed, and UGT2A2 was found to be expressed mainly in nasal mucosa, similar to UGT2A1. In addition, UGT2A1 and UGT2A2 were found to differ largely in their regioselectivity towards estrogens, and this encouraged us to further investigate the glucuronidation of steroids and the differences in how they are glucuronidated by the different human UGT enzymes. The glucuronidation of estriol, 16-epiestriol, 17-epiestriol, 13-epiestradiol, ent-estradiol, ent-androsterone, and ent-etiocholanolone by different UGT enzymes was studied and analytical methods for analysing these compounds were developed. UGTs 1A10 and 2B7 were found to be the most active UGTs in the glucuronidation of almost every steroid studied. Several patterns in the regio- and stereoselectivity of steroid glucuronidation were observed for the first time. The effects of phenylalanines 93 and 90 on human UGT1A10 on glucuronidation were also studied, exploiting site directed mutagenesis. The results revealed that F93 is involved in the interactions of the enzyme with estrogens, but the interactions with smaller molecules were found to be more complex and less drastic. A homology model of UGT1A10 was created based on the results. Knowledge of glucuronidation of endogenous compounds is important, since it may be directly or indirectly linked to clinical conditions and also because certain drugs that inhibit UGTs may inhibit the metabolic pathways of endogenous substrates that are metabolised by UGTs. Understanding the glucuronidation of steroids could be used to

6

make better predictions on the role of glucuronidation in the metabolism of newly designed drugs that resemble steroids. Detailed information on steroid glucuronidation may also be used in determining UGT structure via function and the results obtained during this PhD work may be used in the development of more detailed homology models of other UGT enzymes.

7

ACKNOWLEDGEMENTS

This study was carried out in the years 2008-2013 at the Faculty of Pharmacy, University of Helsinki. Besides the Faculty of Pharmacy, financial support from Magnus Ehrnrooth foundation, Finnish Chemical Society, Graduate School of Pharmaceutical Research, Orion-Farmos Research Foundation and Finnish Cultural Foundation are gratefully acknowledged.

My deepest thanks go to my supervisor, docent Moshe Finel who has always given me help and encouragement when I needed it. I am and I will always be inspired by your enthusiasm and dedication to science. I am also deeply grateful to Dr. Piter Bosma and Docent Liisa Laakkonen and all my other co-authors for their help and contribution. I am also sincerely grateful for the pre-examiners, Professor Peter Mackenzie and Professor Hannu Raunio, for all the effort they made in reviewing and improving this thesis.

My past and present colleagues at pharmaceutical chemistry deserve my sincerest thanks for the unreserved and friendly atmosphere of the department: Anna, Gusse, Hongbo, Inkku, Johanna T., Katriina, Linda, Mika, Mikko, Nenad, Nina, Niina, Paula, Päivi, Raisa, Sanna S., Taina, Tiina S., Titti, and all others – Thank you! Very special thanks go to my roommate and great friend Johanna M. I have enjoyed working with this PhD project almost every moment of it and much of that I owe to you. You made it fun to come to work even on the rainy days.

All my wonderful friends deserve my thanks as well! The most fun I have ever spent has been with you guys. Especially I want to thank Ilkka, Jussi, Janne, Eevi, Jere, Pate, Stefu, Salla, Teemu, Tomppa, Tuomas, and Valtteri: Thank you so much for everything! Special thanks go to Sini, my oldest and dearest friend, for being there no matter what. You are the best! My dear friend Sakari also deserves special thanks: your sense of humour has always been a great relief from weekday routines.

My dear parents Ronald and Kaisu have my very special thanks. You brought me up believing in my skills and always supported and encouraged me to study. Thank you so much for being such great parents. And, although I know this is an utter waste, I would like to thank my cat Sissi for always being happy to see me and forcing me to stay off the computer from time to time.

Finally, I want to thank my loving husband Antti. You have been by my side during this whole project and I can not imagine a better person for that. I´m eagerly waiting for the next big “project” in our lives when our soon-to-be-born child comes into the world.

Helsinki, July 2013

Nina Sneitz

8

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following publications: I Sneitz N, Court MH, Zhang X, Laajanen K, Yee KK, Dalton P, Ding X,

Finel M (2009) Human UDP-glucuronosyltransferase UGT2A2: cDNA construction, expression, and functional characterization in comparison with UGT2A1 and UGT2A3. Pharmacogenet Genomics 19:923-934.

II Sneitz N, Bakker CT, de Knegt RJ, Halley DJ, Finel M, Bosma PJ (2010)

Crigler-Najjar syndrome in The Netherlands: identification of four novel UGT1A1 alleles, genotype-phenotype correlation, and functional analysis of 10 missense mutants. Hum Mutat 31:52-9.

III Höglund C1, Sneitz N1, Radominska-Pandya A, Laakkonen L, Finel M

(2011) Phenylalanine 93 of the human UGT1A10 plays a major role in the interactions of the enzyme with estrogens. Steroids 76:1465-73.

IV Sneitz N, Krishnan K, Covey DF, Finel M (2011) Glucuronidation of the

steroid enantiomers ent-17 -estradiol, ent-androsterone and ent-etiocholanolone by the human UDP-glucuronosyltransferases. J Steroid Biochem Mol Biol 127:282-8.

V Montenegro-Miranda1 PS, Sneitz N1, de Waart DR, Ten Bloemendaal L,

Duijst S, de Knegt RJ, Beuers U, Finel M, Bosma PJ (2012) Ezetimibe: A biomarker for efficacy of liver directed UGT1A1 gene therapy for inherited hyperbilirubinemia. Biochim Biophys Acta 1822:1223-9.

VI Sneitz N, Vahermo M, Mosorin J, Laakkonen L, Poirier D and Finel M

(2013) Regio- and stereo-specificity of the human UDP-glucuronosyltransferases in the glucuronidation of estriol, 16-epiestriol, 17-epiestriol and 13-epiestradiol. Drug Metab Dispos 41:582-91.

The publications are referred to in the text by their roman numerals.

1 Authors contributed equally to these studies

9

ABBREVIATIONS

AAV Adeno-associated virus CN-I Crigler-Najjar syndrome type I CN-II Crigler-Najjar syndrome type II COMT Catechol-O-methyltransferase CYP Cytochrome P450 DMSO Dimethyl sulfoxide E2 Estradiol E3 Estriol EED Ethinyl estradiol Ent Enantiomeric FDA US Food and Drug Administration GC Genome copy GS Gilbert syndrome HIM Human intestinal microsomes HLM Human liver microsomes HPLC High performance liquid chromatography ITR Inverted terminal repeat MeOH Methanol MS Mass spectrometry PCR Polymerase chain reaction pDNA Plasmid DNA qRT-PCR Quantitative real-time polymerase chain reaction RT-PCR Reverse transcription polymerase chain reaction scAAV Self-complementary adeno-associated virus SN2 Bimolecular nucleophilic substitution SULT Sulfotransferase UDP Uridine 5’-diphosphate UDPGA Uridine 5’-diphospho glucuronic acid UGT UDP-glucuronosyltransferase UGT Gene encoding UDP-glucuronosyltransferase

10

1. INTRODUCTION

Metabolism, or biotransformation, is the most important elimination pathway for drugs and most of the drug metabolism takes place in the liver. Drug metabolism is usually divided into two phases: phase I includes functionalization reactions and phase II conjugation reactions (Fig. 1) (Gonzales et al., 2011; Wienkers and Heath, 2005; Evans and Relling, 1999). Metabolic reactions produce metabolites that can be active, even toxic, or inactive. Therefore the FDA encourages characterising metabolites at an early stage of drug discovery (www.fda.gov). Approximately 10% of the top 200 prescribed drugs that were recorded in the United States in 2002 are metabolised, at least partly, by uridine 5’-diphospho glucuronosyltransferases (UDP-glucuronosyltransferases, UGTs). UGTs and cytochrome P450 enzymes (CYPs) are together responsible for the metabolism of most hepatically cleared drugs (Williams et al., 2004, www.rxlist.com).

Routes of drug elimination. A-D represent different elimination routes without metabolism (A) or Figure 1. through one or two metabolising phases (B-D). CYP: Cytochrome P450 enzymes; FMO: Flavin- containing monooxygenase; EH: Epoxide hydrolase; UGT: UDP-glucuronosyltransferase; SULT: Sulfotransferase; GST: Glutathione S-transferase.

While UGTs are important for the metabolism of drugs and other compounds that are foreign to humans, i.e. xenobiotics, they also have numerous endogenous substrates including bilirubin, bile acids and certain hormones, like steroids and thyroid hormones. There are many reasons why it is important to understand the glucuronidation of endogenous substrates. Firstly, endogenous glucuronidation may be linked directly to clinical conditions, like familial hyperbilirubinemias (see section 2.2 and publications II and V) and this knowledge may be utilized in genetic counseling of the patients and in

Phase II

CYPs FMOs EHs Etc.

UGTs SULTs GSTs Etc.

Phase I

A

Excretion

B C D

11

developing new treatments. In addition, certain drugs inhibit UGTs and, therefore, may also inhibit the metabolic pathways of endogenous substrates that are primarily metabolised by UGTs. This is the case, for example, with atazanavir and increased risk of hyperbilirubinemia (Bentue-Ferrer et al., 2009), also making endogenous glucuronidation important for drug development. In cases other than bilirubin, like with steroid glucuronidation, the correlation between clinical conditions and glucuronidation is less direct and more research is needed to make relevant predictions, for example on the interplay between steroid glucuronidation and development of cancer (see section 2.3.). Understanding steroid glucuronidation may also help us in the future to make better predictions on the role of glucuronidation in the metabolism of synthetic estrogens that may be designed for cancer research and therapy (VI). While characterizing a novel UGT, UGT2A2 (I), numerous differences were encountered in regioselectivity of estrogen glucuronidation by UGTs, that were otherwise closely homologous. As relatively large and rigid molecules, steroids may serve as good candidates in determining UGTs structure via function and this approach has been exploited on several occasions (III, IV, VI) using both natural and synthetic steroids. In the following review of the literature, first, UGTs will be briefly discussed followed by a more detailed description of the selected endogenous UGT substrates. Bilirubin and familial hyperbilirubinemias will be discussed, as well as the emerging gene therapy treatments for Crigler-Najjar syndrome. This review will then cover the glucuronidation of steroids and the current knowledge on the correlation between steroid glucuronidation and medical conditions; followed, finally, by the use of steroids in predicting UGT structure.

12

2. REVIEW OF THE LITERATURE

2.1 UDP-GLUCURONOSYLTRANSFERASES IN BRIEF

Uridine 5'-diphospho-glucuronosyltransferases (UGTs, EC 2.4.1.17) are approximately 56 kDa sized integral membrane proteins of the endoplasmic reticulum (ER). The human UGT1 and UGT2 subfamilies contain 19 different enzymes which share close sequence similarities, especially among enzymes of the same subfamily (Mackenzie et al., 2005). UGT1As are encoded by a single large gene, UGT1. Each UGT1A protein is the product of a unique first exon, that encodes the N-terminal half of the protein, and four exons, 2-5, that encode the C-terminal half of the protein and are shared among all UGT1As. UGT2A1 and UGT2A2 are also encoded by unique first exons and five shared exons, exons 2-6. UGT2A3 and UGT2Bs are comprised of six exons, 1-6, that are not shared and are unique for each UGT, even if they are highly homologous, particularly exons 2-6 that encode the C-terminal half of each enzyme (Mackenzie et al., 2005). Additionally, an alternative splicing at the exon 5 of the UGT1 locus leads to nine shorter UGT1A proteins. These alternative UGT isoforms lack enzymatic activity, but they may have a regulatory role in UGT function (Girard et al., 2007; Bellemare et al., 2010; Levesque et al., 2007). UGTs are expressed in various different tissues (Table 1). Liver carries the most abundant expression, but significant extrahepatic UGT expression is also evident, particularly in the kidney and gastrointestinal tract. Moreover, some UGTs, like UGT1A10, are exclusively or nearly exclusively extrahepatic. It is also noteworthy that many UGTs are expressed in steroid target tissues, like the testes and prostate. This will be discussed further in section 2.3.1. While individual UGT enzymes are not discussed in detail in this section, UGT2As deserve some additional remarks since they are the subject of publication I. In 2008, when this PhD project was started, UGTs from subfamilies 1A and 2B had already been studied extensively and their expression patterns were well characterised. Much less was known about the three members of subfamily 2A: UGT2A1, UGT2A2, and UGT2A3. One major work on UGT2A1 had been published earlier showing, among other things, that this enzyme is mainly expressed in the nasal epithelium and that it has considerably broad substrate specificity (Jedlitschky et al., 1999). UGT2A1 was suggested to be involved in odorant signal termination (Jedlitschky et al., 1999; Lazard et al., 1991) and glucuronidated odorant molecules have indeed been shown to be unable to induce olfactory response (Thiebaud et al., 2013). In 2008, a comprehensive work about UGT2A3 was published (Court et al., 2008). The expression pattern and activity of UGT2A3 were found to be very different compared to UGT2A1. The expression tissues included liver, adipose tissue, and parts of the GI tract and only few bile acids were found to be substrates for this enzyme. Subsequently, the characterisation of the remaining UGT2A enzyme, UGT2A2, was addressed (I, section 5.2).

13

Table 1. UGT tissue expression. The table highlights the tissues that contain the highest mRNA levels of a given UGT enzyme. Other expression tissues besides the ones listed also exist. (Court et al., 2008; Court et al., 2012; Ohno and Nakajin, 2009)

UGT Tissue UGT1A1 Liver, small intestine UGT1A3 Liver, small intestine, colon UGT1A4 Liver, small intestine, colon trachea, kidney UGT1A5 Liver, small intestine, colon, trachea UGT1A6 Liver, stomach, small intestine, colon, trachea, kidney, nasal

epithelium UGT1A7 Liver, small intestine, colon, esophagus, kidney UGT1A8 Small intestine, colon, adrenal gland, nasal epithelium UGT1A9 Liver, small intestine, colon, kidney UGT1A10 Small intestine, colon, adipose tissue, nasal epithelium UGT2A1 Nasal epithelium UGT2A2 Nasal epithelium UGT2A3 Liver, small intestine, colon, adipose tissue UGT2B4 Liver, colon UGT2B7 Liver, small intestine, pancreas, kidney, uterus UGT2B10 Liver, testes UGT2B11 Liver, kidney, breast, prostate UGT2B15 Liver, stomach, breast, prostate UGT2B17 Liver, stomach, small intestine, colon, pancreas, adipose tissue,

testes, nasal epithelium UGT2B28 Liver

Contrary to CYPs, the complete crystal structure of any UGT is yet to be solved and only a partial crystal structure of the UDPGA binding domain of UGT2B7 is available (Miley et al., 2007). In addition, several computational models of UGTs have also been made based on the known 3D structure of related bacterial and plant sugar transferases (Fujiwara et al., 2009; Laakkonen and Finel, 2010; Li and Wu, 2007). The lack of high resolution 3D structure of any full size UGT hinders the utilization of computational approaches in glucuronidation studies, for example at early stages of drug development. This highlights the importance of in vitro assays in predicting UGT catalysed metabolism. It has been postulated that UGTs function as dimers, based on different biochemical techniques, like co-immunoprecipitation, rescue of the activity of mutants, inhibition of activity by the alternative spliced variants, fluorescence energy transfer and others. There

14

is evidence of both homo- and heterodimerisation; also with the alternative splice variants. (Bellemare et al., 2010; Finel and Kurkela, 2008; Fujiwara et al., 2010; Kurkela et al., 2003; Meech and Mackenzie, 1997; Operana and Tukey, 2007). The proposed membrane topology of dimeric UGT is presented schematically in Figure 2. Most of the polypeptide, including the enzymes active site, resides in the ER lumen.

Schematic representation of dimeric UGT topology at ER membrane. Most of the polypeptide Figure 2. resides in the ER lumen. (Adapted from Finel and Kurkela, 2008)

UGTs catalyse the transfer of glucuronic acid from uridine 5 -diphosphoglucuronic acid (UDPGA) to various exo- and endogenous compounds. These substrate molecules are commonly regarded as aglycones (nonsugars) in comparison to glucuronides that contain the transferred glucuronic acid. Glucuronidation is evidently an SN2 type bimolecular nucleophilic substitution and the conjugation site is typically a hydroxyl-, carboxyl- or amino-group of the aglycone substrate. (Miley et al., 2007; Meech and Mackenzie, 1998; Patana et al., 2008). The proposed reaction mechanism is shown in Figure 3. The formed glucuronides are usually, but not always, biologically inactive and more hydrophilic than the corresponding aglycones.

CYTOSOL

ER LUMEN

COOH COOH

NH2 NH2

15

OOHOH

OH

HOOC

O PO

UMP

O

OOH

OHOH

HOOCO-R

O H

R

NHNH

R

NHNH

R O

O

R

UGT+ UDP

+

His Asp

His

SN2 type reaction mechanism of glucuronidation. The aglycone (R-OH) is deprotonated by the Figure 3. catalytic dyad of aspartic acid and histidine followed by the nucleophilic attack on C1 carbon of UDPGA by the aglycone ion leading to the detachment of UDP. Protonated histidine is stabilized by the neighboring aspartic acid. (Miley et al., 2007; Yin et al., 1994)

2.2 BILIRUBIN METABOLISM AND FAMILIAL HYPERBILIRUBINEMIAS

In adults, approximately 250-350 mg of bilirubin is produced daily, mainly through the catabolism of heme proteins. Normal plasma concentration of bilirubin in human adults is 5-15 µM. Neonates, however, can have significantly higher concentrations (up to 170 µM) of bilirubin during their first few weeks of life due to immature hepatic conjugation of bilirubin and a limited ability to transport it (Dennery et al., 2001; Kirkby and Adin, 2006). The physiological condition associated with an excess amount of bilirubin is referred to as hyperbilirubinemia, and it is most visibly characterized by jaundice caused by the yellow pigmented colour of bilirubin. A small number of hyperbilirubinemias are due to a rare familial condition called Crigler-Najjar syndrome (CN). In the following chapters, the metabolism of bilirubin will be discussed, in addition to the mechanisms triggering Crigler-Najjar syndrome and the more common benign condition called Gilbert syndrome (GS). The pharmacological implications of hyperbilirubinemias and different therapeutical approaches to the treatment of CN will be taken into consideration.

2.2.1 Bilirubin: properties and metabolism The main source of bilirubin is the breakdown of haemoglobin from red blood cells in the spleen. Heme is degraded by heme oxygenase, resulting in the release of iron and the formation of carbon monoxide and biliverdin (Tenhunen et al., 1969). Biliverdin is then further reduced to bilirubin by biliverdin reductase (Tenhunen et al., 1970).

16

Bilirubin is a linear tetrapyrrole featuring Z,Z configuration (Fig. 4). At physiological pH it has six intermolecular hydrogen bonds which stabilize the Z,Z configuration of bilirubin and prevent its interactions with polar groups in aqueous milieu. Therefore bilirubin is very lipophilic with a logP value of approximately 5 (U.S. National Library of Medicine).

NH

O

NH

O

O

HN

HN

O

H

O

OH

HNNH

NH

O

HN

O

OH

O

HO

O

Bilirubin Structure. Bilirubin in its linear Z,Z conformation (A) and ridge tile conformation (B). In Figure 4. ridge tile conformation each propionic acid group has a hydrogen bond with the opposing dipyrrinone lactam and pyrrole (Boiadjiev and Lightner, 2001).

Hydrophobicity is the main reason behind the toxic effects of bilirubin. Due to its high hydrophobicity, bilirubin can passively diffuse across biological membranes, including the blood-brain barrier, if it is not bound to albumin or if it is unconjugated, namely without any glucuronic acid (Zucker et al., 1999). In the central nervous system, bilirubin exhibits considerable toxicity to neurons, including the inhibition of protein kinase C, an essential enzyme for synaptic signalling (Grojean et al., 2000). In vivo and in vitro experiments have also indicated that bilirubin induces apoptosis (Grojean et al., 2000; Grojean et al., 2001; McDonald et al., 1998). The mode of apoptosis is apparently mitochondria mediated: bilirubin diffuses across mitochondrial membranes, causing mitochondrial swelling and the release of cytochrome c to the cytosol, triggering apoptosis (Rodrigues et al., 2000). It has also been shown that even a short time exposure of neurons to bilirubin can inhibit their function (Zhang et al., 2003). Due to this hydrophobicity, bilirubin cannot be excreted as it is and must be modified enzymatically first (Kirkby and Adin, 2006; Boiadjiev and Lightner, 2001). Bilirubin has high affinity to albumin and over 99% of unconjugated bilirubin circulates in the plasma

A B

17

bound to albumin (Ostrow et al., 2003). Unconjugated and albumin-bound bilirubin is transported to the liver, where it dissociates from albumin and moves into hepatocytes either by diffusion or active uptake (Tiribelli and Ostrow, 1996). Inside hepatocytes, UGT1A1 catalyses the transfer of glucuronic acid from UDPGA to the carboxylic groups of bilirubin to form mono- and di-glucuronides. Of the known 19 UGT isoforms, only UGT1A1 is capable of glucuronidating bilirubin (Bosma et al., 1994). Bilirubin glucuronides are excreted into bile via the efflux transporter ABCC2 (MRP2) and, eventually, into the intestine (Paulusma et al., 1997).

2.2.2 Gilbert syndrome Gilbert syndrome, a common and benign familial hyperbilirubinemia, was first described over a hundred years ago (Gilbert and Lereboulet, 1901). Estimations about the prevalence of GS in the Caucasian population vary between 6 and 16% (Monaghan et al., 1996; Owens and Evans, 1975). Patients with GS have approximately 30% of UGT1A1 activity compared to normal UGT1A1 activity and this causes slightly elevated, usually 15-50 M, serum bilirubin levels and, in some cases, mild jaundice (Monaghan et al., 1996, Burchell and Hume, 1999). The condition, however, is benign and requires no treatment. Today there are 19 mutations reported in the literature causing GS (see Canu et al., 2013 for a recent review). In the Caucasian and African populations the most common mutation is the elongation of a TATA-box in the promoter region of UGT1A1 by an additional TA repeat. This is also known as UGT1A1*28 polymorphism or A(TA)7TAA and it causes underexpression of UGT1A1 (Bosma et al., 1995). This polymorphism has approximately 40% frequency in these populations, but only 16% in Asian populations (Beutler et al., 1998). In Asian populations the missense mutation G71R (UGT1A1*6) is also a significant cause of GS (Akaba et al., 1998). Although GS is benign, it may have clinical significance. While bilirubin is traditionally considered to be a metabolic waste product, it also has antioxidant capacity and it has been shown to prevent the oxidation of lipids and lipoproteins (Stocker et al., 1987; Sedlak et al., 2009). The antioxidant properties of bilirubin are directly related to the total serum antioxidant capacity and several studies have associated either low bilirubin levels with higher risk, and/or high bilirubin levels with lower risk, of cardiovascular diseases (Horsfall et al., 2012; Kimm et al., 2009; Perlstein et al., 2008; Rantner et al., 2008; Vitek et al., 2002). The actual mechanism behind this association is not entirely clear and not all studies have associated the UGT1A1*28 mutation with decreased risk of cardiovascular diseases (Rantner et al., 2008; Hsieh et al., 2008). GS has also been associated with increased risk of certain cancers. The UGT1A1*6 genotype has been associated with the risk of developing colorectal cancer in a case control study with Chinese patients (Tong et al., 2007). The UGT1A1*28 genotype, on the other hand, has been linked to both decreased (Jiraskova et al., 2012) and increased (Bajro et al., 2012) risk of colorectal cancer in males. The case control studies of UGT1A1*28

18

and the risk of breast cancer are also conflicting. Some analyses have associated UGT1A1*28 genotype with increased risk of breast cancer (Shatalova et al., 2006; Adegoke et al., 2004), while others failed to find a significant association (Guillemette et al., 2001). UGT1A1 has numerous other endobiotic and xenobiotic substrates besides bilirubin, including steroid hormones and drugs. Reduced UGT1A1 function may potentially affect the metabolism of these compounds, too. The best known example is irinotecan, a pro-drug that is used in the treatment of metastatic colorectal cancer. Irinotecan is converted by endogenous carboxyesterase enzymes to the active metabolite, SN-38, that is a topoisomerase I inhibitor. SN-38 is glucuronidated by UGT1A1 to an inactive glucuronide and failure to do this may cause severe drug toxicity (Iyer et al., 1998). Patients with GS have been shown to have an increased risk of SN-38 mediated toxicity, especially at high irinotecan doses. Due to this, in 2005 the US Food and Drug Administration (FDA) recommended UGT1A1*28 testing on the irinotecan drug label (Marques and Ikediobi, 2010).

2.2.3 Crigler-Najjar syndromes Cricler-Najjar syndrome is a familial hyperbilirubinemia that was first described by John Fielding Crigler and Victor Assad Najjar in 1952 (Crigler and Najjar, 1952). Crigler and Najjar, who also gave their name to the syndrome, described six children born to three different families that suffered from severe unconjugated hyperbilirubinemia. Five of these children died before reaching 15 months of age and one lived up to be 15 years old. It was concluded that the children had died of brain damage caused by kernicterus. Later it was found that there is variation in the severity of hyperbilirubinemia between CN patients and the syndrome was divided into two clinically distinct forms: CN type I (CN-I) and CN type II (CN-II) (Arias et al., 1969). CN-I patients do not have any bilirubin conjugation activity and their serum bilirubin levels vary between 340-685 µM (normal range 5-15 M). CN-II patients, on the other hand, have some bilirubin glucuronidation activity left and their serum bilirubin levels range from 100 to 340 µM (Costa, 2006). The distinction between GS, CN-II, and CN-I is not sharp and there is a continuous spectrum in the severity of hyperbilirubinemia (Fig. 5).

19

Graphic representation of UGT1A1 activity and its correlation to familial hyperbilirubinemias Figure 5. Gilbert syndrome (GS) and Crigler-Najjar syndromes I and II (CN-I and CN-II, respectively). (Adapted from Strassburg, 2010)

The residual bilirubin glucuronidation activity can be further enhanced with phenobarbital treatment. Response to phenobarbital treatment can also be used to distinguish between the two syndromes: While CN-I patients do not response to the treatment, CN-II patients response to the treatment and their serum bilirubin levels are reduced due to phenobarbital induced glucuronidation (Arias et al., 1969). Type II CN can, however, also lead to kernicterus and irreversible neurological damage, for example post surgically, or if the syndrome is severe (Poddar et al., 2002; Thapa et al., 1987; Chalasani et al., 1997) Today CN syndrome is known to be caused by deficiency in UGT1A1 function (Bosma et al., 1992) and there are at least 102 mutations causing CN syndrome reported in the literature. Mutations have been found in all five exons of UGT1A1 and also in the splice sites and they include point mutations, deletions, and insertions (see Canu et al., 2013 for a recent review). CN syndromes are inherited recessively and are estimated to affect 1/106 new-borns (Bosma, 2003). In some cases, however, the inheritance seems dominant with incomplete penetrance. This might be the case, for example, if the patient is heterozygous for CN-I type mutation and homozygous for GS type promoter mutation (Chalasani et al., 1997). Insufficient bilirubin conjugation in CN patients leads to accumulation of unconjugated bilirubin in the serum and hyperbilirubinemia. In concentrations over 200-300 µM, serum albumin binding becomes saturated and concentrations of free, unbound, bilirubin rise drastically. Toxic levels of bilirubin are therefore reached at these concentrations (Wennberg, 2000). There is a natural mutant rat model, the Gunn rat, for CN syndrome (Gunn, 1944). Gunn rats have a deletion at nucleotide 1242 of bilirubin glucuronidating UGT, which results in a frameshift mutation and a premature stop codon (Sato et al., 1991). This rat model has proven to be useful in various studies of CN syndrome, including demonstration that CN syndrome is caused by deficient bilirubin glucuronidation (Axelrod et al., 1957). A mouse model for CN has also been introduced recently (Nguyen et al., 2008).

UGT1A1 Activity

Hyperbilirubinemia

100 % 0 %

Phenotype: GS CN-II CN-I

20

2.2.3.1 Current therapy for Crigler-Najjar syndrome CN-II patients do retain some UGT1A1 activity and their condition is usually more easily treatable compared to CN-I patients. The treatment of CN-II patients typically consists of UGT1A1 inducer phenobarbital (Arias et al., 1969). Some CN-II patients, however, also require phototherapy or even a liver transplant (II, V). The absence of bilirubin glucuronidation activity is fatal for humans and before the development of phototherapy, CN-I was a lethal disease (Crigler and Najjar, 1952). Nowadays CN syndromes are treatable and the patients can reach adulthood without irreversible brain damage (van der Veere et al., 1996). The discovery of phototherapy greatly improved the prognosis for CN-I and it remains the most prevalent treatment for CN syndromes (Bosma, 2003; van der Veere et al., 1996; Cremer et al., 1958). The effectiveness of phototherapy is based on the ability of blue light to convert bilirubin from its 4Z,15Z configuration to different photoisomers, e.g. to 4E,15Z-cyclobilirubin. These photoisomers are not able to compose as many intramolecular hydrogen bonds as physiological bilirubin and are therefore more hydrophilic and easily excretable (Lightner et al., 1979). Phototherapy is, however, cumbersome and it affects the life quality of CN-I patients, considering that some of them may need phototherapy up to 14 hours per day. Another drawback of phototherapy is that it usually becomes less effective as patients age due to the thickening of the skin, and the risk of bilirubin encephalopathy therefore increases with age (van der Veere et al., 1996). The only permanent therapy for CN-I patients in the long term is liver transplantation and usually CN-I patients will need the liver transplant at some point in their lives, when phototherapy is no longer effective enough to maintain safe bilirubin levels (van der Veere et al., 1996). Since only a small amount of liver is needed for a sufficient amount of glucuronidation, auxiliary liver transplantation has also proven to be effective (Rela et al., 1999). Since there are difficulties with both phototherapy and liver transplantation, namely the lack of organ donors and side effects of lifelong immune suppression, there is a substantive need for alternative treatments for CN-I. Hepatocyte transplantation has been one of the more recent approaches that have been carried out in CN patients (Fox et al., 1998). Nonetheless, hepatocyte transplantation has thus far yielded only a partial lowering of bilirubin levels and the results have persisted for less than one year in most cases (Lysy et al., 2008).

2.2.3.2 Gene therapy in treatment of Crigler-Najjar syndrome Gene therapy is another new approach for CN treatment. The syndrome has many characteristics that make it an ideal disease for gene therapy treatment. It is caused by a single gene and for therapeutic effect it is sufficient if approximately 5% of UGT1A1 activity is regained (Fox et al., 1998). Moreover, there is an animal model, the Gunn rat, available for preclinical trials. The efficacy of the treatment in human patients, however, cannot be monitored by serum bilirubin quantification, since CN-I patients need daily phototherapy. Therefore it is impossible to determine whether the serum bilirubin is

21

lowered because of phototherapy or as a consequence of gene therapy. This issue is addressed in publication V. Several groups have succeeded in lowering bilirubin levels in vivo in Gunn rats by hepatic gene transfer of the UGT1A1. Various different vectors have been used, including plasmid DNA (pDNA) (Jia and Danko, 2005a; Jia and Danko, 2005b), retrovirus based viral vectors (Bellodi-Privato et al., 2005), lentiviral vectors (Nguyen et al., 2007; Seppen et al., 2006; Schmitt et al., 2010), recombinant simian virus 40 (SV40) vectors (Sauter et al., 2000), adenovirus vectors (Askari et al., 1996; Dimmock et al., 2011; Toietta et al., 2005), and adeno-associated virus vectors (AAV) (Seppen et al., 2006; Pastore et al., 2012; Montenegro-Miranda et al., 2011). AAV vectors have been shown to be effective also in the treatment of CN in a mouse model (Bortolussi et al., 2012). In many cases, a single dose of a vector may provide a lifelong correction of bilirubin levels (Nguyen et al., 2007; Seppen et al., 2006; Toietta et al., 2005; Bortolussi et al., 2012). AAV vectors (Fig. 6) are among the most promising vector systems for gene therapy. They are non-pathogenic (lack all viral genes), have a broad host spectrum, are stable, and maintain a high level of expression for a long time. Recombinant AAV vectors that are used in gene transfer contain the desired gene and suitable promoter or enhancer element, flanked by two inverted terminal repeats (ITR). ITRs facilitate hairpin-type intrastrand base pairing which enables self-priming in DNA replication. A recombinant AAV also needs rep and cap genes and a helper virus for its replication and integration to the genome. These elements are usually provided in trans. (Grieger and Samulski, 2012) AAV transduction is a multistep process including attachment to cell membrane, virion internalization, intracellular trafficking, and transport to the nucleus (Buning et al., 2008). Before transduction to the genome, AAV vectors need to be converted from single to double stranded, which is the rate limiting step of AAV mediated transduction (Ferrari et al., 1996). Self-complementary AAVs (scAAV) vectors circumvent the problem by having a dimeric inverted repeat sequence. A mutated ITR at the middle of the sequence forms a closed hairpin structure at one end of the self-complementary sequence while wild-type ITRs form two hairpins at the open end. With the rate limiting step bypassed, scAAV vectors have been shown to have higher transduction efficiencies than conventional AAV vectors (McCarty et al., 2001).

22

Schematic figure of wild type and recombinant AAV genomes. Wild type AAV (A) contains rep Figure 6. and cap regions flanked by inverted terminal repeats (ITRs). Recombinant AAV (B) contains gene of interest and suitable promoter or enhancer element in place of rep and cap regions. Recombinant AAVs however need trans helper system containing rep and cap genes. scAAV vectors (C) contain mutated ITR in the middle of the sequence, which forms a closed hairpin structure at one end of the dimeric inverted sequence. Wild type ITRs form two hairpins at the other end.

2.3 STEROID HORMONES AND THEIR METABOLISM

Steroids are a large class of generally lipophilic compounds, with a core structure of four rings of 17 carbons (Fig. 7). They are formed from cholesterol in a process called steroidogenesis and they differ from each other by functional groups attached to the carbon rings and by the oxidation state of the rings.

12

3

45

67

8

910

1112

13

14 15

16

17

A B

C D

1

2

3

4 5

10 9

11

1213

6 7

8 14 15

16

17

Carbon atom and ring numbering of steroids. Gonane, the simplest possible steroid, is presented in Figure 7. 2D and 3D structure.

3’ITR

Wild type AAV

Rep Cap

Promoter Transgene

Helper system (trans)

Rep Cap

AAV shuttle

A

B

C scAAV vector

Promoter Transgene Mutated ITR Transgene Promoter

3’ITR

3’ITR

5’ITR

5’ITR

5’ITR

23

Steroidogenesis takes place mainly in the adrenal cortex, testes, ovaries, and in the placenta during pregnancy. Numerous different enzymes are involved in the production of steroids, many of them CYP enzymes (Hanukoglu, 1992). Metabolism of endogenous estrogens is presented in Figure 8.

Endogenous metabolism of estrogens. Estrone and estradiol are generated from androstenedione Figure 8. and testosterone, respectively. Estriol is mainly generated from estradiol and estrone, but it can also arise from 16-hydroxyandrostenedione (16-OH AD), or from 16-hydroxydehydroepiandrosterone (16-OH DHEA). Estriol stereoisomers, 17-epiestriol and 16-epiestriol, are formed from 16 - hydroxyestrone or estriol in subsequent reduction and oxidation reactions (Eliassen et al., 2009). The natural estrogens studied in publications I-III and VI are highlighted.

Enantiomeric steroids (ent-steroids) have opposite configurations at all chiral centres in comparison to their corresponding natural steroid, resulting in a mirror image of the molecule (Fig. 9). Enantiomers have identical chemical and physical properties, except for their ability to rotate plane-polarized light. Their three-dimensional structures, on the other hand, are different and this may result in different types of binding in different types of binding pockets. Enantiomeric structures are interesting starting points for drug discovery

O

OH

OH

OH

OH

O

OH

OH

OH

OH

OH

OH

OH

OH

OH

OH

OH

O

OH

Estrone 17 -Estradiol

16 -Hydroxyestrone

Estriol

16-Ketoestradiol

Catechol estrogens 16-OH AD and 16-OH DHEA

17-epiestriol

16-epiestriol

Androstenedione Testosterone

I, II, III, VI

I, III, VI

Publications

VI

24

because the stereochemical changes of the parent molecule can drastically affect the pharmacological properties of the molecules.

OH

OH

H H

H

OH

H H

H

OH

17 -estradiol (A) with its enantiomer, ent-estradiol (B). Figure 9.

The clinical potential of ent-steroids has been studied for several years (reviewed by Covey, 2009). Some ent-steroids, like ent-estradiol and ent-progesterone, have been shown to have neuroprotective effects (Green et al., 2001; VanLandingham et al., 2006). Ent-androsterone and ent-etiocholanolone have also been shown to be positive allosteric modulators of -aminobutyric acid (GABA) type A receptors and to be substantially more active on them than their natural counterparts (Katona et al., 2008; Krishnan et al., 2012).

2.3.1 Glucuronidation of steroid hormones It is commonly believed that one of the major functions of UGTs is the glucuronidation of steroids. UGTs can conjugate the naturally occurring steroid hydroxyl groups or the hydroxyl groups gained though catabolism. Conjugated steroids have a significantly reduced or non-existent ability to bind to steroid receptors, and to generate biological response (You, 2004). Glucuronidated steroids are also more readily removed from the tissue and more easily excretable in urine than their corresponding aglycones. Many steroid glucuronides are indeed detectable in urine (Smith and Kellie, 1967; Kirdani et al., 1977). Alongside sulfotransferases (SULTs) and CYPs, UGTs are one of the major enzyme families involved in steroid catabolism (You, 2004). As ‘high affinity, low activity’ enzymes, SULTs are effective in conjugating steroids at low concentrations. The importance of UGTs is emphasized at high steroid concentrations, as the capacity of SULTs is saturated and the ‘low affinity, high capacity’ UGTs step in. Estradiol (E2) glucuronides were discovered as early as 1936 when Cohen and Marrian reported estrogen glucuronides from human pregnancy urine (Cohen and Marrian, 1936). Since then, several steroids have been found to be UGT substrates and steroid glucuronidation has also been studied extensively in vitro. (Table 2)

A B

25

Table 2. Selected UGT steroid hormone substrates and their major glucuronidating UGT enzymes.

Trivial name Structure

Major

glucuronidating

enzymes

Reference

Estrone (E1) H

H

H

O

OH

1A1, 1A8, 1A10

(Smith and Kellie, 1967; Lepine et al., 2004)

Estradiol (E2) H

H

H

OH

OH

UGT1A1, UGT1A10, UGT2B7, UGT2B17

(Smith and Kellie, 1967; Itaaho et al., 2008)

Estriol (E3) H

H

H

OH

OH

OH

UGT1A10, UGT2B7

(Sandberg and Slaunwhite, 1965, VI)

Testosterone H

H

H

OH

O

UGT2A1, UGT2B15, UGT2B17

(Fishman and Sie, 1956; Sten et al., 2009a)

Dihydro-testosterone

(DHT) O

H

H

OH

HH

UGT2B15, UGT2B17

(Kirdani et al., 1977; Murai et al., 2006)

Androsterone H

H

H

O

OHH

UGT2B7, UGT2B17

(Coffman et al., 1998; Sten et al., 2009b)

26

Several UGT enzymes are capable of steroid glucuronidation. Generally, UGT1A enzymes are known to glucuronidate estrogens E1, E2, E3 and catechol estrogens (Lepine et al., 2004; Itaaho et al., 2008; Starlard-Davenport et al., 2007, VI). The extrahepatic UGT1A10 seems to be the major UGT isoform in conjugation of estrogens at the 3-OH position and it has also been shown to be up-regulated by E2 in breast cancer cells (Starlard-Davenport et al., 2008b). UGT2Bs, on the other hand, generally catalyse estrogen glucuronidation at D-ring, as well as glucuronidation of androgens that lack the phenolic 3-OH (Sten et al., 2009b; Sten et al., 2009a; Chouinard et al., 2008). Especially UGT2B7 and UGT2B17 have high capacities for androgen conjugation. UGT2B7 seems to be able to glucuronidate most steroids, including estrogens (Itaaho et al., 2008; Radominska-Pandya et al., 2001). Interestingly, UGT1As and 2Bs exhibit stereoselectivity in estrogen glucuronidation. UGT1As, with the exception of 1A3 and 1A4, generally catalyse the conjugation of 3-hydroxyl of estrogens, while UGT2Bs, with the exception of 2B15, glucuronidate D-ring hydroxyls (Itaaho et al., 2008, VI). The nasal epithelium enzymes UGT2A1 and UGT2A2 have exceptionally broad substrate specificity, even for UGTs, and their substrates include many estrogens and androgens (Jedlitschky et al., 1999; Sten et al., 2009b; Sten et al., 2009a, I). UGT2A1 and UGT2A2 also exhibit less stereoselectivity and they are often capable of glucuronidating both A and D ring hydroxyls of steroids, but at different rates (Itaaho et al., 2008, VI). Steroids have been shown to be glucuronidated in steroid target tissues (Chung and Coffey, 1978; Adams et al., 1989) and steroid glucuronides have also been found in several of these tissues, including prostate, breast cyst fluid, and ovary follicular fluid (Belanger et al., 1991). Expression of UGTs, determined by mRNA levels, has also been documented in these tissues and especially many UGT2B enzymes are expressed in steroidogenic tissues (Table 1). Based on the abundance of UGTs in steroidogenic tissues and the evidence of their glucuronidation in situ, it is proposed that UGTs are involved in maintaining steroid homeostasis via glucuronidation in steroid target tissues. Human prostate adenocarcinoma LNCaP cells expressing UGT2B17 have been found to have decreased response to androgens, which suggests that glucuronidation plays a role in regulating the availability of steroids at their target sites (Belanger et al., 1998). The metabolic pathways of enantiomeric steroids are poorly characterised and in publication IV, how these molecules are glucuronidated compared to their natural counterparts was investigated for the first time.

2.3.2 Steroids and health Steroids are known to be involved in normal physiology and several medical conditions, especially in cancers that are located in steroid target tissues. For example, increased steroid concentrations have been associated with higher risk of breast cancer in postmenopausal women, (Key et al., 2002) and in some studies also among premenopausal women (Walker et al., 2011). In estrogen dependent cancers, the activation

27

of estrogen receptors induces the growth of cancer cells. These cancers can be treated by blocking estrogen receptors with antiestrogens, such as tamoxifen, in the case of breast cancer (Cuzick et al., 2013). Enantiomeric steroids were initially considered potential antiestrogens, but later found to bind only weakly to estrogen receptors and also to lack estrogenic effects in murine tissues (Payne and Katzenellenbogen, 1979). The 4-hydroxylated estrogens have also been associated with breast cancer (Liehr and Ricci, 1996). In normal breast tissue, 2-hydroxyestradiol is the dominant catechol estrogen, but in breast cancer cells 4-hydroxyestradiol predominates. Both of these hydroxylated estradiols can form reactive quinones that can bind to DNA and lead to mutations and eventually induce cancer (Cavalieri et al., 2002; Stack et al., 1996). The 2-hydroxyestradiol is apparently less carcinogenic, since the 2,3-quinone formed from 2-hydroxyestradiol is less reactive than the corresponding 3,4-quinone formed from 4-hydroxyestradiol. The 2-hydroxyestradiol also forms smaller amount of these reactive quinones since it is methylated by catechol-O-methyltransferase (COMT) at higher rates than 4-hydroxyestradiol (Emons et al., 1987; Zahid et al., 2006). An elevated ratio of 4-/2-hydroxyestradiol formation is therefore suggested to be a biomarker of benign or malignant breast tumours (Liehr and Ricci, 1996). Besides COMT, UGTs are also capable of inactivating these catechol estrogens, and thus may protect cells from the carcinogenic effects of them (Raftogianis et al., 2000). Some studies have also associated higher testosterone levels with increased risk for prostate cancer (Hyde et al., 2012). Taken together, it is postulated that the glucuronidation of steroids could prevent the possibly carcinogenic effects of altered steroid concentrations. Although this is an exciting concept, there is little evidence of this actually being the case. This may be partly due to the fact that steroids have other catabolic routes besides glucuronidation and also because UGTs show extensive overlapping in substrate specificity. That is, if a given UGT enzyme has impaired activity, other UGTs can compensate for it in many cases. Although correlation between steroid glucuronidation and diseases is less clear than with UGT1A1 and hyperbilirubinemias, there is some clinical evidence available, especially concerning cancers. A whole gene deletion of UGT2B17 (UGT2B17*2) has been reported in humans and the frequency of this polymorphism is about 11% in Caucasians and 2% in African Americans (Wilson et al., 2004). Individuals homozygous for this deletion have been associated with increased prostate cancer risk in the Caucasian, but not in the African American population (Park et al., 2006). Several opposite results, with no such correlation, have also been reported (Gallagher et al., 2007; Olsson et al., 2008; Setlur et al., 2010), so this correlation remains controversial. UGT2B15 is also polymorphic, and the allele UGT2B15(Y85) (UGT2B15*2) has higher Vmax values for dihydrotestosterone glucuronidation than the wild type UGT2B15(D85) (Levesque et al., 1997). Several studies have associated the lower activity allele D85 with

28

increased risk of prostate cancer (MacLeod et al., 2000; Park et al., 2004; Okugi et al., 2006; Grant et al., 2013), while others have failed to find any correlation (Cunningham et al., 2007; Gsur et al., 2002). UGT1A enzymes may also be involved in abnormal estrogen glucuronidation among breast cancer patients. The glucuronidation of 4-hydroxyestrone has been shown to be significantly reduced in cancer breast tissue, with a concomitant reduced expression of UGT1A10 (African American women) and UGT2B7 (Caucasian women) (Starlard-Davenport et al., 2008a). Association between endometrial cancers and certain UGT1A1 and UGT2B7 polymorphisms has also been investigated. The results of this study indicated that UGT polymorphisms do not contribute to the risk of endometrial cancer (McGrath et al., 2009). As discussed in section 2.2.2, the role of UGT1A1*28 polymorphism in the risk of breast cancer and colorectal cancer remains controversial.

2.4 STEROID GLUCURONIDATION AND UGT STRUCTURE As mentioned before, the high-resolution crystal structure of any UGT is yet to be solved. Therefore, the substrate specificity of a given UGT is difficult, or impossible to predict, although this kind of information would often be very useful in drug development. Instead many other approaches have been made in trying to determine the critical amino acids and the structure of UGTs and their binding sites. These approaches include molecular modelling, site directed mutagenesis, chimerisation, pharmacophore models, and structure-activity studies (Laakkonen and Finel, 2010; Xiong et al., 2006; Itaaho et al., 2010; Smith et al., 2003). As relatively large and rigid molecules, steroids are prospective candidates in determining UGT structure via function. Examining the different configurations of the hydroxyl groups on rings A and D provides insight into the binding of steroids to the UGT active site and the possible binding to critical amino acid residues. This approach is used in publication VI, where in addition to novel information on substrate specificity, some basic ideas of differences in the active sites between individual UGT enzymes could be drawn. Combining this strategy with site directed mutagenesis adds value to this approach and even enables the making of molecular models that are based on the obtained information. This approach was employed in publication III where a molecular model of the UGT1A10 binding site was built based on the information gained from estrogen glucuronidation studies with two UGT1A10 mutants. Some of the differences observed in the glucuronidation of different steroids may be due to their different physiochemical properties rather than three-dimensional structure. Enantiomeric steroids, on the other hand, are interesting in this area since they have the

29

same physicochemical properties as the corresponding endogenous steroids, and therefore any differences between the glucuronidation of enantiomers and their natural counterparts are solely due to the differences in three-dimensional shape. Enantiomeric steroids were used as study tools in publication IV.

30

3. AIMS OF THE STUDY

a) To investigate the UGT1A1 missense mutant enzymes found in Dutch Crigler-Najjar patients and to work with the development of gene therapy treatment for the syndrome (II, V). A functional analysis of the missense mutant UGT1A1 enzymes was pursued by expressing 10 missense mutants and wild type UGT1A1 in baculovirus-infected insect cells and microsomal preparations were made of them. The glucuronidation activity of the mutants with bilirubin was then studied in order to estimate the correlation of genotype and phenotype (the severity of hyperbilirubinemia). Furthermore, other substrates were tested with the mutants to see whether the mutations would lead to any unexpected activity, which would be interesting for determining the structure of UGT active site. Promising results for CN-I gene therapy in Gunn rats, the animal model for the disease, have been reported by our collaborator Dr. Piter Bosma and his team (Liver Center of the Academical Medical Center of the University of Amsterdam). As part of this collaboration a nontoxic and specific bio-marker for this treatment was sought (V). It was anticipated that the glucuronide concentration of this biomarker would correlate with the recovery of UGT1A1 activity thereby confirming the efficacy of gene therapy treatment even under conditions of on-going phototherapy. b) To express and characterize the novel UGT enzyme UGT2A2 (I). The UGT2A2 gene had previously been found in the human genome and some mRNA expression reported in segments of the intestinal tract. A complete and thorough characterization was sought, including substrate specificity and a detailed tissue expression pattern. c) To study the glucuronidation of steroid hormones (III, IV, VI). The main goals were to extend and improve the picture of steroid glucuronidation and to determine UGT specific stereospecificities of the glucuronidation in order to help determine UGT structure via function. Site specific mutations (III) and enantiomeric steroids (IV) were also used to gain more detailed information.

31

4. MATERIALS AND METHODS

4.1 MATERIALS

4.1.1. Chemicals The chemicals that were used in publications I-VI are listed in Table 3. The chemicals were of analytical grade and the high performance liquid chromatography (HPLC) eluents were of HPLC grade or mass spectrometry (MS) grade, when LC-MS was used. The ent-androsterone, ent-etiocholanolone, ent-17 -estradiol, and 13-epiestradiol were synthesized in the laboratory of Prof. Douglas Covey according to methods that are described elsewhere (Green et al., 2001; Katona et al., 2008; Ayan et al., 2011). 1-hydroxypyrene -D-glucuronide was synthesized in our laboratory as described earlier (Luukkanen et al., 2001).

Table 3. Chemicals used in this study.

Chemical Supplier Use Publi-cation

Acetic acid J.T. Baker, Deventer, Netherlands HPLC I-III Acetonitrile Fluka, Germany HPLC I-VI Alamethicin Sigma-Aldrich, St. Louis, MO Activity assay V, VI 4-Aminobiphenyl Sigma-Aldrich, St. Louis, MO Substrate I Ammonium acetate Riedel-de Haën, Seelze, Germany HPLC I, IV ent-Androsterone Prof. DF Covey, Washington

University in St. Louis, School of Medicine, St. Louis, MO

Substrate IV

Bilirubin Sigma-Aldrich, St. Louis, MO Substrate II, V Deuterated methanol Sigma-Aldrich, St. Louis, MO NMR VI Dimethylsulfoxide Fluka, Germany Activity assay I-VI Disodium hydrogen phosphate dihydrate

Fluka, Germany Activity assay I-VI

Entacapone Orion Pharma, Espoo, Finland Substrate III 13-Epiestradiol (18-epiestradiol)

Prof. DF Covey, Washington University in St. Louis, School of Medicine, St. Louis, MO

Substrate VI

17 -Estradiol Sigma-Aldrich, St. Louis, MO Substrate I-V ent-17 -Estradiol Prof. DF Covey, Washington

University in St. Louis, School of Medicine, St. Louis, MO

Substrate IV

-Estradiol 3- -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I-V

-Estradiol 17- -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I, II, IV

32

Table 3. Continued.

16 , 17 -Estriol Sigma-Aldrich, St. Louis, MO Substrate VI 16 , 17 -Estriol Sigma-Aldrich, St. Louis, MO Substrate I-III, VI 16 , 17 -Estriol Steraloids, Newport, RI Substrate VI Estriol 3- -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I-III, VI

Estriol 16 - -D-glucuronide,

Sigma-Aldrich, St. Louis, MO Standard I, VI

Estriol 17 - -D-glucuronide,

Sigma-Aldrich, St. Louis, MO Standard VI

17 -Ethinylestradiol Sigma-Aldrich, St. Louis, MO Substrate I-III, V ent-Etiocholanolone Prof. DF Covey, Washington

University in St. Louis, School of Medicine, St. Louis, MO

Substrate IV

Ezetimibe Kemprotec Ltd, Middlesbrough, UK/MSD,Hertfordshire, UK

Substrate V

Formic acid 98-100% Sigma-Aldrich, St. Louis, MO HPLC V, VI 7-Hydroxycoumarin

-D-glucuronide Ultrafine Synergy-House, Manchester, UK

Standard I-III

1-Hydroxypyrene Sigma-Aldrich, St. Louis, MO Substrate I, II 1-Hydroxypyrene -D-glucuronide

Faculty of Pharmacy, University of Helsinki, Finland

Standard I, II

Hyodeoxycholic acid Sigma-Aldrich, St. Louis, MO Substrate I Methanol Fluka, Germany/Merck, Darmstadt,

Germany Activity assay, HPLC

I-VI

4-Methylumbelliferone

Sigma-Aldrich, St. Louis, MO Substrate I-III

4-Methylumbelliferyl -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I-III

MgCl2 Riedel-de Haën, Seelze, Germany Activity assay I-VI

-Naphthol Sigma-Aldrich, St. Louis, MO Substrate I-III -Naphthyl -D-

glucuronide Sigma-Aldrich, St. Louis, MO Standard I-III

4-Nitrophenol Sigma-Aldrich, St. Louis, MO Substrate I-III 4-Nitrophenyl -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I-III

Perchloric acid 70-72%

Merck, Darmstadt, Germany Activity assay I-VI

2-phenylphenol Fluka, Germany Substrate I 3-phenylphenol Chiron Trondheim, Norway Substrate I 4-phenylphenol Fluka, Germany Substrate I D-Saccharic acid 1, 4-lactone

Sigma-Aldrich, St. Louis, MO Activity assay I-III

Scopoletin Sigma-Aldrich, St. Louis, MO Substrate I-III

33

Table 3. Continued.

Sodium phosphate monobasic dihydrate

Sigma-Aldrich, St. Louis, MO Activity assay, HPLC

I-VI

Testosterone-glucuronide

National Measurement Institute, Pymble, Australia

Standard IV

[14C] UDP glucuronic acid

PerkinElmer Life and Analytical Sciences, Boston, MA

Co-substrate I-III

UDP-glucuronic acid triammonium salt

Sigma-Aldrich, St. Louis, MO Co-substrate I-VI

UDP-glucuronic acid trisodium salt

Sigma-Aldrich, St. Louis, MO Co-substrate VI

Umbelliferone Sigma-Aldrich, St. Louis, MO Substrate I, II Umbelliferyl -D-glucuronide

Sigma-Aldrich, St. Louis, MO Standard I, II

Water Millipore, Molsheim, France Activity assay, HPLC

I-VI

ZnSO4 Sigma-Aldrich, St. Louis, MO Activity assay I, II

4.1.2. Microsomes and RNA sources Enzyme and RNA sources other than in-house expressed UGTs (see 4.4.1.2.) are listed in Table 4.

Table 4. Enzyme and RNA sources used in this study.

Enzyme/RNA source Supplier Publi-cation

Gunn rat microsomes Tytgat Institute for Intestinal and Liver Research, Academic Medical Center of the University of Amsterdam, The Netherlands

V

Human adult liver Department of Pharmacology and Experimental Therapeutics, Tufts University School of Medicine, Boston, MA, USA

II

Human adult lung Department of Pharmacology and Experimental Therapeutics, Tufts University School of Medicine, Boston, MA, USA

II

Human adult nasal mucosa

Monell Chemical Senses Center,Philadelphia, PA, USA

II

Human fetal lung University of Washington Birth Defects Research Laboratory, USA

II

Human fetal nasal mucosa

University of Washington Birth Defects Research Laboratory, USA

II

Human lung RNA BD-Clontech, USA II

34

Table 4. Continued.

Human Pooled fetal liver total RNA

BD-Clontech, USA II

Human Pooled small intestine mucosa RNA

Ambion, USA II

Human UGT2B15 Supersomes®

BD Gentest Woburn, MA, USA IV, VI

Pooled human intestinal microsomes (HIM)

BD Gentest Woburn, MA, USA VI

Pooled human liver microsomes (HLM)

BD Gentest Woburn, MA, USA VI

4.2 PATIENTS

Publications II and V included investigations on Dutch Crigler-Najjar patients. Details of the patients, including the type of Crigler-Najjar syndrome, gender, mutations, presence of the UGT1A1*28 promoter mutation (7 TA repeats compared to the wild type 6 TA repeats), and the treatments of individual patients are listed in Table 5. The information about the treatments of the patients is as of 2009 and the information that was published first in publication II of this thesis, provided by Dr. Piter Bosma, was obtained under the approval of the Ethical committee of the University of Amsterdam.

Table 5. Crigler-Najjar patient data. OLT: orthotopic liver transplant, PT: phototherapy (hours per day), Phe: patient treated with phenobarbital.

Patient

no. Gender Mutant proteins *28 Treatment Reference

Crigler-Najjar Type I

1 M Q49X/del exon 4 6/6 OLT (Gantla et al., 1998)

2 F K407X/Y74X 6/6 PT 9 (Zmetakova et al., 2007;

Kadakol et al., 2000)

3 F K407X/K407X 6/6 PT 7 (Zmetakova et al., 2007)

4 M K407X/K407X 6 /6 PT 10 (Zmetakova et al., 2007)

5 F K407X/K407X 6/6 PT 10 (Zmetakova et al., 2007)

6 M K407X/K407X 6/6 PT 10-14 (Zmetakova et al., 2007)

7 F K407X/K407X 6/6 PT 10 (Zmetakova et al., 2007)

8 F K402T/K402T 6/6 PT 10 II

9 M L443P/L443P 6/6 ? (D'Apolito et al., 2007)

35

Table 5. Continued.

Crigler-Najjar Type II

10 F L15R/L15R 7/7 PT 2 (Seppen et al., 1996)

11 M L15R/K407X 7/6 PT 2, Phe (Seppen et al., 1996)

12 F L15R/S191F 7/7 PT 2 II

13 M G71R/K407X 6/6 Phe (Zmetakova et al., 2007;

Aono et al., 1993)

14 M R209W/R209W 6/6 (Bosma et al., 1993)

15 F R209W/R209W 6/6 (Bosma et al., 1993)

16 M R336W/A498X 7/6 PT 12,

Phe

(Ciotti et al., 1998, II)

17 M R336W/A498X 7/6 PT 12,

Phe

(Ciotti et al., 1998,

II)

18 M L15R/L15R 7/7 Phe (Seppen et al., 1996)

19 F P387H/wt 6/7 Phe II

4.3. ANIMAL EXPERIMENTS

Publication V included animal experiments that were conducted in the Tytgat Institute for Intestinal and Liver Research, Netherlands. The experiments were performed in accordance with the animal ethics committee guidelines of the Academic Medical Center of Amsterdam. The Gunn rats were 8–10 weeks old, weighing 180-200 g (males) or 120-150 g (females). The rats received an intra-portal venous injection of scAAV2/8-LP1-UGT1A1 that provides the hepatic expression of human UGT1A1 (Montenegro-Miranda et al., 2011). The doses of scAAV were either high (1 × 1012 genome copies [GC]/kg), medium (3 × 1011 GC/kg), low (1 × 1011 GC/kg), or none (negative control). Rat liver microsomes were isolated as previously described (Seppen et al., 2006). The rats received ezetimibe (2 mg/kg) i.p. Blood samples were collected using heparin tubes, before the injection and at different time intervals following it. Heparin tubes were centrifuged for 1 min at 10,000 × g to obtain plasma.

36

4.4 METHODS

4.4.1 Molecular biology

4.4.1.1 Cloning Recombinant human UGTs were either cloned in our lab (Kurkela et al., 2003; Kuuranne et al., 2003) or obtained from the collaborating laboratories of Professor Peter Mackenzie (UGTs 1A5, 2B10, 2B11, 2B17, and 2B28) or Professor Michael Court (UGT2A3). UGT2A1 was cloned in the laboratory of Dr. Ding (I). UGT2A2 was constructed by amplifying exon 1 of UGT2A2 from genomic DNA and ligating it to exons 2-6 of UGT2A1 after some DNA manipulation. UGT2A1 cDNA was provided by Dr. Xinxin Ding and Dr. Xiuling Zhang from Wadsworth Center, New York State Department of Health. UGT1A1 and UGT1A10 mutations (II, III) were created using the QuikChange system (Stratagene, La Jolla, CA, USA) or by a procedure described by Deng and Nickoloff, (1992). The amplified genes were subcloned into pFB-XHA or pFB-XHC vectors (Kurkela et al., 2003). The cloned genes were sequenced entirely to confirm the sequence and to exclude additional alterations.

4.4.1.2. UGT expression in Sf9 cells All the recombinant UGTs and the mutants, except UGT2B15, were expressed in-house as His-tagged proteins. The expression was done in baculovirus infected Sf9 insect cells according to the Bac-to-Bac® Baculovirus Expression System (Invitrogen). Briefly, recombinant baculoviruses were generated by transposition in the E.coli bacterial strain DH10Bac, after which bacmid DNA was isolated. Sf9 cells were transfected with the bacmid using the Cellfectin reagent and viruses were harvested from the transfected insect cell medium according to the manufacturer's instructions. The optimal infection size was determined by infecting parallel cell suspensions of constant cell concentration with different amounts of recombinant baculovirus, and by assaying enzyme activity in the collected cells 48 h post infection. For protein production, the cells (2x106 cells/ml) were cultured in suspension with constant stirring at 27 C in HyClone SFX-Insect medium (Thermo Scientific, Waltham, MA) supplemented with 10% heat-inactivated fetal bovine serum. The cells were harvested approximately 48 hours after infection. The membranes (microsomes) were isolated from the harvested cells as described before (Kurkela et al., 2003).

4.4.1.3. UGT quantification and visualisation The total protein concentration of the isolated membranes was measured using the bicinchonic acid (BCA) assay kit according to the manufacturer's instructions (Pierce, Rockford, IL). The relative expression levels of individual UGTs, carrying C-terminal His-Tag, were determined using the monoclonal tetra-His antibodies (Qiagen, Hilden, Germany). The relative expression levels reflect the ratio of UGT protein versus total protein and they were used to normalize the actual rates obtained from UGT activity

37

assays. In publication II, Western-blotting was performed to confirm proper expression of all UGT1A1 mutant proteins, i.e. that none of them contained a pre-mature stop codon. Expression of UGT2A1, UGT2A2, and UGT2A3 in human tissues was screened using RT-PCR and subsequently quantified by quantitative real-time PCR (qRT-PCR). The expression of UGT2A2 mRNA was confirmed from fetal nasal mucosa by PCR amplification and sequencing (I). This work was done in a collaborative laboratory at the Tufts University, Boston, Massachusetts.

4.4.2 Glucuronidation assays The glucuronidation activity assay samples contained 50 mM phosphate buffer pH 7.4, 10 mM MgCl2, 0.01-20 mg/mL of enzyme source (total protein), 1-5 mM UDPGA and 1-4000 µM of aglycone substrate, dissolved in up to 100% dimethyl sulfoxide (DMSO). The final DMSO concentration in the assays was max 10% and the total sample volume was 100 l. This amount of DMSO does not significantly reduce UGT activity (Zhang et al., 2011). When the enzyme source was HLM, HIM, or rat liver microsomes, 5% alamethicin was also added to the mixture to activate the latent UGT activity in these microsomes (V,VI). The reactions were initiated by the addition of UDPGA, then carried on for 15-150 min at 37 °C and terminated by the addition of either 10 l ice-cold 4 M perchloric acid, 100 l methanol (MeOH), or 20 µl of cold ZnSo4 and 200 µl MeOH depending on the substrate used. The samples were thereafter transferred to ice for 10 min and the proteins were finally sedimented by centrifugation at 16100 g (room temperature). The supernatants from these centrifugations were transferred in to vials and analysed immediately.

4.4.3 Analytical methods and data analysis Supernatants were analysed by HPLC, using either a Shimadzu LC-10 model (Shimadzu Corporation, Kyoto, Japan), an Agilent 1100 (Agilent Technologies, Palo Alto, CA, USA), or an ultra-performance liquid chromatography quadrupole time of flight mass spectrometry (UPLC-QTOF) (Waters/Micromass, Manchester, UK). The analytical methods used for separation of the glucuronides and aglycone substrates were developed by the author or described in previous publications from our laboratory. Details of the methods are provided in publications I-VI. Detection of glucuronides was performed in most cases with a fluorescence detector, but if the glucuronides did not exhibit fluorescence, an ultraviolet (UV) detector or MS was used. In most cases the glucuronides were identified and quantified using authentic glucuronides as external standards. If authentic glucuronides were not available, the quantification was done with external standard curve generated with the unconjugated substrate, as in the case of bilirubin, or by a combination of radioactive and fluorescence detection, as for 4-phenylphenol. In the latter case, the reactions contained 50 µmol/l unlabelled UDPGA, and 11.1 µmol/l radiolabeled [14C] UDPGA. The samples were

38

analysed with HPLC combined with a radiochemical detector (Model 9701, Reeve Analytical, Glasgow, UK). The glucuronides were quantified by the peak areas of the [14C] UDPGA-containing metabolites. In the case of estriol stereoisomers, where three different glucuronides were possible and authentic glucuronides were not available for determining the retention time of each glucuronide, we used NMR for the identification of distinct estriol glucuronides (VI). Large amounts of individual glucuronides were biosynthesised with recombinant enzymes, fractionated using an Agilent 1100 HPLC, and the specific fractions were collected. The glucuronide fractions were concentrated using a rotary evaporator and lyophilized. The lyophilized glucuronides were redissolved in deuterated methanol and subjected to NMR spectroscopy on a Varian 300 MHz MercuryPlus spectrometer (Varian, Inc., Palo Alto, CA). The enzyme kinetic constants were determined by fitting the experimental data using GraphPad Prism 4.02 or 5.04 for Windows (GraphPad Software, San Diego, CA, USA) (I, III, IV, VI). The data was fitted to either the Michaelis–Menten, Michaelis–Menten with substrate inhibition, or biphasic kinetic models. The equations used were as follow: Michaelis-Menten equation:

1. = Vmax × [S]Km + [S]

Michaelis-Menten with substrate inhibition:

2. = Vmax × [S]

Km + [S] × (1+[S]Ki

)

Biphasic kinetics:

3. = Vmax1 × [S] Km1+[S]

+ Vmax2 × [S] Km2+[S]

(Initial) reaction rate Vmax Maximum enzyme velocity [S] Substrate concentration Km Michaelis-Menten constant, substrate concentration at ½ Vmax

Ki Dissociation constant for enzyme-inhibitor complex

39

4.4.4 Molecular modelling A homology model for UGT1A10 (III) was constructed using Modeler 9v6 similarly to the previous work with UGT1A1 (Laakkonen and Finel, 2010). Briefly, human UGT2B7 was used as a template for the enzyme C-terminal domain and three templates were used for the N-terminal domain: UGT72B1 from Arabidopsis thaliana, GtfB from Amycolatopsis orientalis and macrolide glycosyltransferase from Streptomyces antibioticus. The substrate structures were built and optimized in MOPAC2009 (http://openmopac.net) and docked into the active site based on the interactions with catalytic histidine (His37) and UDPGA. (III, Laakkonen and Finel, 2010)

40

5. RESULTS AND DISCUSSION

5.1. UGT1A1 AND CRIGLER-NAJJAR SYNDROME

5.1.1 UGT1A1 mutations in CN patients in the Netherlands The UGT1A1 missense mutant enzymes found in Dutch Crigler-Najjar patients were investigated in publication II. In total, 14 different UGT1A1 mutations were found in the set of CN-patients, of which 4 were novel (Table 5). Eleven patients had at least one of the two mutations frequently detected in the Netherlands, K407X and L15R. Many of them are nonrelated, which suggests that both are founder mutations. Of the identified mutations, the missense mutants G71R, S191F, R209W, R336W, P387H, K402T, and L443P identified from patients 8, 9, 12-17 and 19 were studied more closely (Table 5, Fig. 10). Two mutations, L15R and Y486D, were not included because both had been studied before (Seppen et al., 1996; Kurkela et al., 2007) and also because L15R mutation is located in UGT1A signal sequence and probably affects the expression level of the protein more than the activity of the produced protein. Nonsense mutations, i.e. point mutations such as Q49X that result in a premature stop codon, were also not included since they are most likely degraded before or after translation. Instead, three previously reported mutations were included for comparison: L175P, Q331R, and G395V (Seppen et al., 1994; D'Apolito et al., 2007; Moghrabi et al., 1993). As seen in Figure 10, mutations G71R, L175Q, S191F, and R209W lie in the putative aglycone binding area and may affect the binding of aglycone substrate. L175 may be one of the hydrophobic residues forming a hydrophobic pocket for aglycone binding. Mutations P387H, G395V, and K402T lie in the highly conserved UGT signature sequence. This sequence contains the UDPGA binding site and the conserved amino acids are thought to be important or even essential for UDPGA binding and enzyme activity (Miley et al., 2007). Mutations Q331R and R336W are also located near the UDPGA binding site and may affect the affinity of UDPGA to the enzyme. L443P lies further away from the UDPGA binding site, in the middle of -helix C 7 (Miley et al., 2007). Proline is known to disrupt or break -helixes because its amine group is bonded to the side chain and thus cannot form a hydrogen bond. Therefore it is possible that a mutation from leucine to proline might end the -helix and disrupt the three dimensional structure of the enzyme, affecting its catalytic activity.

41

Schematic representation of UGT1A1 primary sequence and the locations of the missense Figure 10. mutations examined.

It is noteworthy that mutations in exons 2-5, region common to all UGT1A enzymes, are present in all nine UGT1A enzymes. That is, patients with mutations Q331R, R336W, P387H, G395W, K402T, and L443P are prone to reduced activity in the glucuronidation of many other endobiotic and xenobiotic UGT substrates besides bilirubin. For example, propofol is an anesthetic that is almost exclusively metabolised by UGT1A9 (Court, 2005), hence mutations in UGT1 exons 2-5 are likely to affect the metabolism of this drug. Western-blot analyses were performed to confirm proper expression of all missense mutants, revealing an increased Mr for mutant K402T (Fig. 11). Closer examination of the sequence of this mutant revealed that this is most likely due to an additional N-glycosylation site generated by the mutation. The mutant sequence 400NATR403 is strongly predicted to contain a glycosylation site while the wild type 400NAKR403 sequence is not (NetNGlyc 1.0 Server).

Western-Blot wild type UGT1A1and UGT1A1K402T. K402T has increased molecular mass due to Figure 11. the new N-glycosylation site. (II)

UGT1A1 UGT1A1K402T

42

5.1.2 Bilirubin glucuronidation activity of ten UGT1A1 missense mutants One of the main aims of this study was to evaluate the effect of the missense mutations on the bilirubin glucuronidation activity and to examine the correlation of in vitro results with the severity of hyperbilirubinemia. The correlation between residual UGT1A1 activity at the protein level and patient phenotype has relevance in predicting the severity of the syndrome in newly identified patients. For this, bilirubin glucuronidation assays were performed, with all the mutant proteins as well as wild type UGT1A1 (Fig. 12).

1A1

G71R

L175

QS19

1F

R209W

Q331R

R336W

P387H

G395V

K402T

L443

P0

50

100

500

1000

1500

2000

0123456

20406080100

Bilir

ubin

glu

curo

nida

tion

rate

pmol

/min

/mg

% o

f wt 1

A1 a

ctivi

ty

Bilirubin glucuronidating activity of wild type UGT1A1 and ten UGT1A1 mutants. The results Figure 12. are normalised for differences in enzyme expression as determined by dot-blot analysis. (II)

Mutants G71R, L175Q, S191F, R209W, and R336W showed some residual bilirubin glucuronidating activity and they are all indeed found in patients diagnosed with CN-II. G71R is a relatively common GS-causing mutation in Asian population. Patient 13 exhibiting this mutation, however, presents a CN-II phenotype, but this is explained by compound heterozygous genotype G71R/Y486D. Mutants Q331R, P387H, G395V, K402T, and L443P did not show any bilirubin glucuronidation activity, but only K402T and L443P were found in patients with CN-I. Mutations P387H and G395V were identified in heterozygous CN-II patients and the other allele probably had enough residual bilirubin glucuronidating activity which could explain the phenotype. Mutation Q331R was identified in a homozygous patient with CN-II. Assuming that this patient was indeed homozygous for the Q331R, there is an apparent

43

contradiction in this case between the bilirubin glucuronidation activity in vitro and the severity of the disease in vivo. One explanation for this might be that the in vivo activity is derived from heterodimer formation that rescues part of the enzymatic activity (Kurkela et al., 2007). It might also be possible that this mutation causes folding defects and these folding defects are more severe in the insect cell expression system because of the different composition of chaperone proteins compared to humans. Further research is needed to understand this observation in full.

5.1.3 Activity of UGT1A1 missense mutants toward other substrates UGT1A1 has many other substrates besides bilirubin and we also studied the effect of the missense mutations on the glucuronidating activity of eight UGT1A1 substrates (Table 6). The results show that the effects of the mutations on bilirubin glucuronidation generally correlate well with the activity toward other substrates.

Table 6. Activity of 10 UGT1A1 missense mutants towards tested substrates. The symbols represent the mutant proteins glucuronidation activity compared to wild type UGT1A1. -: no activity detected; (+): 0.1-2.5% activity; +: 2.5-10% activity; ++: 10-25% activity; +++: 25-50% activity and ++++: over 50% activity.

Substrate G71R L175Q S191F R209W Q331R R336W P387H G395V K402T L443P

Bilirubin ++ + + + - (+) - - - -

17 -estradiol +++ (+) ++ (+) - - - - - -

17 -ethinyl-estradiol

++++ (+) +++ (+) - - - (+) (+) -

1-hydroxy-pyrene

++ - ++ - - - - (+) - -

4-methyl-umbelliferone

++++ (+) +++ (+) - - - (+) - -

-naphthol +++ (+) + (+) (+) - - - - -

4-nitrophenol +++ (+) ++ + (+) - - - - -

Scopoletin ++++ ++ ++ ++ (+) (+) - (+) - -

Umbelliferone ++++ (+) ++ + - - - (+) - (+)

44

Some mutant enzymes (Q331R, G395V, K402T, and L443P) exhibited activity for one or more substrates, although they did not glucuronidate bilirubin. It is possible that these mutations are targeted to an amino acid residue that is specific for bilirubin binding. A more likely explanation is, however, that since the sensitivity of the bilirubin method was lower than with most other methods, some of these mutant proteins did in fact exhibit minor bilirubin glucuronidation activity, but the yield of glucuronides remained below the detection limit of the method.

5.1.4 Biomarker for efficacy of gene therapy for Crigler-Najjar syndrome As noted in section 2.2.3.2, the efficacy of gene therapy for CN cannot be monitored by serum bilirubin quantification since daily phototherapy of CN-I patients cannot be halted. The possibilities to determine the efficacy of scAAV2/8-LP1-UGT1A1 treatment were investigated in publication V. One option to estimate the efficacy of gene therapy treatment would naturally appear to be the assessment of bile glucuronides. Closer inspection, however, revealed that not only does bile analysis require an invasive endoscopy, the amount of bilirubin glucuronides in bile correlated poorly with the clinical status of CN patients (see publication V for detailed results). This indicates that the amount of bilirubin glucuronides in bile cannot be used in evaluating whether or not it is safe to cease phototherapy. Therefore we turned to exogenous UGT1A1 substrates in search of nontoxic and specific marker that would confirm the activity level of UGT1A1 and, thereby, the efficacy of the gene therapy treatment. Three compounds were tested for this purpose: estradiol, ethinylestradiol and ezetimibe (Fig. 13). These compounds were selected based on previous knowledge of their nontoxicity and selectivity for (human) UGT1A1. In the case of estradiol, selectivity applies when giving this compound intravenously to bypass the high activity of UGT1A10 in the intestine. (Lepine et al., 2004; Itaaho et al., 2008; Ghosal et al., 2004; Ebner et al., 1993).

HO

H

OH

H H

HO

OH

H H

H F

OH

N

O

HO FA B C

Structures of estradiol (A), ethinylestradiol (B), and ezetimibe (C). Figure 13.

The glucuronidation of estradiol, ethinylestradiol, and ezetimibe was first tested in vitro, using liver microsomes from Gunn rats that received a single injection of scAAV2/8-LP1-UGT1A1, encoding human UGT1A1 (Fig. 14). Assays for bilirubin glucuronidation were also included to ensure UGT1A1 activity and to monitor the therapeutic effect of the virus

45

dose. The results revealed a good correlation between the different levels of the gene therapy treatment and bilirubin glucuronidation rates. The highest bilirubin glucuronidation rates were observed with microsomes from rats that were treated with a high dose of scAAV, while microsomes from non-treaded control rats showed no bilirubin glucuronidation activity. Even with the lowest virus dose the reduction of serum bilirubin in rats was in the magnitude that would be therapeutic in CN-I patients (results not shown).

Male Gunn rats

High (n=4) Medium (n=6) Ctrl (n=4)0

250

500

750

1000

1250

Virus dose

pmol

/min

/mg

Female Gunn rats

Medium (n=4) Ctrl (n=4)

EzetimibeEthinylestradiol

BilirubinEstradiol

Glucuronidation activities measured from Gunn rat liver microsomes. The doses of AAV carrying Figure 14. UGT1A1 were high (1×1012 GC/kg), medium (3×1011 GC/kg), or none (Ctrl). Untreated male control Gunn rats showed no activity for bilirubin or ezetimibe glucuronidation.

Estradiol and ethinylestradiol glucuronidation at the 3-OH proved to be unsuitable for the biomarker role, because the glucuronidation activity in control microsomes was in the same range as in microsomes from rats that were treated with the medium and high amounts of scAAV. This indicates that one or more hepatic rat UGT2B can catalyse these reactions at relatively high rates. Ezetimibe appeared more suitable, since its glucuronidation rate in untreated control male Gunn rat microsomes was below the detection limit and clearly higher in microsomes from treated rats. In female rats, however, background ezetimibe glucuronidation activity was observed. One explanation for this could be gender differences in the expression of UGT2Bs in rats. One or more UGT2B enzyme, which catalyses the glucuronidation of ezetimibe, might be expressed at higher levels in female rats compared to male rats and this could explain the increased ezetimibe glucuronidation activity observed in liver microsomes from female Gunn rats. Gender differences in UGT activity have been reported in rats, for example female Wistar rats have been found to conjugate bilirubin at two-fold higher rates compared to males (Muraca and Fevery, 1984). In humans, UGT2B15 and UGT2B7 have been also reported to glucuronidate ezetimibe, and UGT2B15 is known to be up regulated by estrogen (Ghosal et al., 2004; Harrington et al., 2006). We could speculate that a rat equivalent of UGT2B15 could be responsible for the increased ezetimibe glucuronidation activity in

46

female Gunn rats, but since rat UGTs are not thoroughly characterized, this remains a hypothesis. Ezetimibe conjugation was subsequently studied in Gunn rats in vivo in a collaborating laboratory. Serum concentrations of ezetimibe and its glucuronide were measured and the correlation between ezetimibe glucuronidation and bilirubin reduction was determined. A clear correlation between the reduction of unconjugated bilirubin concentration and increased ezetimibe glucuronidation was observed in both male and female rats that were treated with UGT1A1-carrying scAAV (Fig. 15).

Correlation between ezetimibe glucuronidation and the degree of serum bilirubin correction (i.e. Figure 15. reduction of bilirubin) 20 min after injection in rats that were treated with UGT1A1-carrying scAAV. (V)

The reduction was seen in both sexes in vivo, and although the results from the in vitro assays demonstrated a large difference between control male and female samples, the in vivo assays revealed that there is a clear difference in the ezetimibe glucuronidation rate between treated and untreated rats, also in female rats (results not shown). While these results suggest that ezetimibe is a suitable biomarker for UGT1A1 activity in vivo, it should be highlighted that since some UGT1A3 activity in ezetimibe glucuronidation has been reported (Ghosal et al., 2004), ezetimibe might be suitable as a biomarker mainly for CN patients with a mutation in exons 2–5 and not in exon 1. In future clinical studies, the suitability of ezetimibe as a biomarker for gene therapy should be evaluated separately for patients who have the inactivating mutation in the first exon. Furthermore, although ezetimibe appears to be a good biomarker for UGT1A1 activity in rats, the same does not necessarily apply in humans due to interspecies differences in glucuronidation and further studies are therefore needed.

R2 = 0,886

0

20

40

60

80

0 0,2 0,4 0,6 0,8

eze-gluc / eze

perc

enta

ge c

orre

ctio

n of

ser

um b

iliru

bin

R2 = 0,7231

0

20

40

60

80

0,00 0,20 0,40 0,60 0,80

eze-gluc / eze

perc

enta

ge c

orre

ctio

n of

ser

um b

iliru

binFemale rats Male rats

47

5.2 EXPRESSION AND CHARACTERISATION OF UGT2A2 The novel UGT2A2 enzyme was characterised in publication I. The expression of UGT2A1, UGT2A2, and UGT2A3, relative to 18S rRNA expression, was determined by qRT-PCR from human fetal nasal mucosa, adult nasal mucosa, fetal lung, adult lung, fetal liver, adult liver, and adult small intestine mucosa (Fig 16).

Expression of UGT2A1, UGT2A2, and UGT2A3 isoform mRNA in human tissues. Results are Figure 16. normalized to 18S rRNA content. # represents individual donors and in pooled samples the number of donors is represented in brackets. (I)

UGT2A2, as well as UGT2A1, mRNA was mainly detected in human fetal and adult nasal mucosa. UGT2A2 exon 1 expression was previously reported in duodenum and jejunum in a review article with limited experimental details (Tukey and Strassburg, 2001) and we too found UGT2A2 in the small intestine, although only in trace amounts. With UGT2A1 and 2A3, our data was in agreement with the previously reported results (Jedlitschky et al., 1999; Court et al., 2008; Somers et al., 2007; Thum et al., 2006). According to our results,

1316

1678

315

0 0 0 0 0 0 0 0

422

683

143

0 0 3 0 0 1 0 40 0 11 4 2 0 10 15 19

152

243

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

UG

T m

RN

A co

nten

t(c

opie

s / 1

0^9

copi

es 1

8S rR

NA)

UGT2A1

UGT2A2

UGT2A3

Tissue

48

UGT2A2’s expression pattern is closely similar to the expression pattern of UGT2A1. An interesting finding was also that the mRNA levels were significantly higher in fetal nasal samples compared to adult ones. The prenatal expression might be an evolutionary relic, but it is also possible that these enzymes are involved in glucuronidating compounds present in the amniotic fluid around the developing fetus and thus limiting the intake of possibly harmful chemicals. For example, several drugs are able to cross the placental wall and reach the fetus. If the drug is subsequently excreted renally by the fetus, it enters the amniotic fluid and is subject to recirculation via amniotic aspiration, which occurs nasally as well as via mouth (Morgan, 1997). There are also indications of prenatal olfactory function in humans (Faas et al., 2000; Schaal et al., 2000; Schaal et al., 1998), so nasal UGTs might be responsible for olfactory signal termination also in the fetus. Keeping in mind the exon sharing between UGT2A1 and UGT2A2 (see section 2.1), UGT2A2 cDNA was constructed by first amplifying exon 1 from genomic DNA and then ligating it to the exons 2-6 of UGT2A1 that was cloned in a collaborating laboratory. The cloned UGT2A2, as well as the cloned UGT2A1 and UGT2A3, were expressed in Sf9 cells and their activities were tested using various substrates (Table 7).

Table 7. Glucuronidation rates of UGT2As with selected aglycone substrates. The rates are expressed as mean ±SD (pmol/min/mg) and the concentration of the aglycone substrates was 200 µM. Nd, no glucuronides detected; Det, glucuronides detected below quantification limit.

Substrate 2A1 2A2 2A3 4-aminobiphenyl Nd Nd Nd Estradiol (3-glucuronide) 0.70 ± 0.03 1.50 ± 0.03 Nd Estradiol (17-glucuronide) 2.2 ± 0.1 Det Nd Estriol (3-glucuronide) Nd Nd Nd Estriol (16 or 17-glucuronide) 14.1 ± 0.5 0.10 ± 0.03 Nd Ethinylestradiol Det 9.5 ± 0.5 Nd 1-hydroxypyrene 198 ± 2 5.3 ± 0.1 Nd Hyodeoxycholic acid 2.1 ± 0.1 8.1 ± 0.5 4.9 ± 0.4 4-methylumbelliferone 548 ± 40 2.9 ± 0.1 Nd 1-naphthol 314 ± 24 4.1 ± 0.1 Nd 4-nitrophenol 60.1 ± 4 3.6 ± 0.1 Nd 2-phenylphenol 15.7 ± 0.6 1.3 ± 0.1 Nd 3-phenylphenol 339 ± 6 62 ± 1 Nd 4-phenylphenol 249 ± 22 74 ± 2 Nd Scopoletin 283 ± 12 7.7 ± 0.1 Nd Umbelliferone 233 ± 10 0.60 ± 0.03 Nd

Generally, our results confirmed the broad substrate specificity of UGT2A1 (Jedlitschky et al., 1999) and the narrow selectivity of UGT2A3 (Court et al., 2008). UGT2A1 glucuronidated all the compounds that were recognized as UGT2A1 substrates before, and

49

in our set of substrates, UGT2A3 exhibited glucuronidating activity only towards the bile acid hyodeoxycholic acid was found to have similar substrate specificity as UGT2A1, but lower glucuronidation rates. Exceptions to this rule were found with hyodeoxycholic acid conjugation, in which UGT2A2 exhibited a higher glucuronidation rate, and in the glucuronidation of estrogens, where there were differences in both regiospecificity and in glucuronidation rates (discussed in more detail in section 5.3.1). The difference in the glucuronidation rate of hyodeoxycholic acid by UGT2A1 and UGT2A2 was recently confirmed by another group, which also demonstrated that these enzymes, particularly UGT2A2, glucuronidate other bile acids at high rates (Perreault et al., 2013). Why these enzymes glucuronidate bile acids is unclear, since they are not expressed extensively in tissues that are exposed to them (Court et al., 2012). To assess whether the generally lower glucuronidation activity of UGT2A2 was due to weaker substrate affinity or lower turnover rates, we performed enzyme kinetic assays with selected substrates, namely 4-nitrophenol, 4-phenylphenol, 4-methylumbelliferone, and the co-substrate UDPGA. With UGT2A1, 4-nitrophenol followed Michaelis–Menten kinetics, 4-methylumbelliferone Michaelis–Menten kinetics with substrate inhibition and 4-phenylphenol biphasic kinetics. With UGT2A2, all substrates followed Michaelis-Menten kinetics. Kinetic constants are presented in Table 8.

Table 8. Kinetic constants for the glucuronidation of 4-nitrophenol, 4-methylumbelliferone, and 4- phenylphenol by UGT2A1 and UGT2A2 and for a UDPGA with either 500 µM 4-methyl- umbelliferone (UGT2A1), or 750 µM 4-phenylphenol (UGT2A2). Values are represented as mean ± SD from triplicate incubations in µM (Km values) or pmol/min/mg (Vmax values). Ki value (in µM) is presented in the case of Michaelis–Menten kinetics with substrate inhibition.

UDPGA 4-nitrophenol 4-methyl-umbelliferone 4-phenylphenol

Km Vmax Km Vmax Km Vmax Km Vmax

UGT2A1 39.5 ± 8.9

292 ± 12

303 ± 35 141 ± 5 32.3 ± 2.2 537 ± 13

0.9 ± 0.3 (Km1) 116 ± 10 (Vmax1)

Ki, 2135 ± 240

614 ± 221 (Km2) 325 ± 49 (Vmax2)

UGT2A2 55.1 ± 7.7

91.4 ± 3

4469 ± 295

69 ± 3 2998 ± 123 39 ± 0.9

245.0 ± 12 124 ± 2

UGT2A2 was found to have significantly higher Km values for all the aglycone substrates tested while both enzymes have approximately equal affinity to UDPGA. The Km values for UDPGA were slightly lower compared to previous studies with other UGTs (see, for example, Patana et al., 2007; Luukkanen et al., 2005). This might be a trait specific for UGT2A1 and UGT2A2, or arise from the aglycone compound used. The Vmax values for UGT2A2 were also generally lower compared to UGT2A1, but the differences were not as striking as with the Km values. It is, however, worth noting that since the affinity of UGT2A2 to aglycone substrates is rather poor, the substrate concentrations needed to nearly reach these Vmax values would be almost impossible to achieve in a physiological environment.

50

The very clear biphasic kinetics of UGT2A1 in 4-phenylphenol glucuronidation, confirmed by computer analyses, was an interesting finding. The enzyme seems to have two binding sites for this substrate, a low-affinity site and a high affinity one. The high affinity binding site possessed a Km value that is exceptionally low for UGTs, which are generally considered to be ´low affinity, high capacity´ enzymes. If UGT2A1 forms dimers, it may be argued that the dimerisation plays a role in the biphasic kinetics. In our studies we used recombinant UGT2A1, meaning that homodimers could form, but not heterodimers with other UGTs. It is thus likely that both binding sites lay in UGT2A1, either in one or more subunits. In any case, additional studies are needed to identify the specific binding sites of the substrates. In summary, we expressed, and characterized the novel UGT2A2, which was found to be a functional enzyme that is expressed in the nasal mucosa. Exon sharing between UGT2A1 and UGT2A2 was also confirmed and the substrate specificities of these enzymes were found to be closely similar, but not identical.

5.3 GLUCURONIDATION OF STEROIDS

5.3.1. Glucuronidation of estradiol, ethinylestradiol, and estriol by UGT2A1 and UGT2A2 Several differences in the regioselectivity of estradiol, ethinylestradiol, and estriol (see Figures 8 and 13 for molecular structures) glucuronidation were found with UGT2A1 and UGT2A2. The identification of the conjugation site was, however, limited in the case of estriol, since the HPLC method for estriol glucuronidation had a limited resolution and the 16- and 17-glucuronides could not be distinguished from each other in this study. UGT2A1 exhibited a clear preference for the 17-OH (or 16-OH in the case of estriol) of the tested estrogens (Table 7) and with ethinylestradiol, that has a bulky ethinyl group attached to C17, hindering the conjugation at the 17-OH, UGT2A1 was practically inactive. On the other hand, for UGT2A2, ethinylestradiol was one of the best substrates among these three steroids while UGT2A2 was nearly inactive in estriol glucuronidation.

5.3.2. Estrogen glucuronidation and UGT1A10 substrate binding site Amino acids F90 and F93 of UGT1A10 were previously suggested to be important in estrogen glucuronidation (Starlard-Davenport et al., 2007) and in publication III we created several mutations in these residues of UGT1A10 to further study the effect of these amino acids on estrogen conjugation. The activity of the mutants was screened using single concentrations (indicated in brackets) of estradiol (100 M), estriol (500 M), and ethinylestradiol (100 M). Results with other substrates and kinetic assays are not discussed here. The results with the three estrogens are presented in Figure 17.

51

Glucuronidation of estradiol (A), estriol (B) and ethinylestradiol (C) with UGT1A10 and its 12 Figure 17. mutants. Only the formation of 3-glucuronide is presented.

The results suggest that wild type UGT1A10 is able to glucuronidate estradiol at a relatively high rate, but as the D-ring substituents increase in size, the glucuronidation rates are reduced (estriol) or even abolished (ethinylestradiol). Interestingly, mutating phenylalanine 93 to an amino acid with a smaller side chain, like alanine or glycine, cancels out this phenomenon. Therefore we suggest that the D ring substituents of these estrogens clash with F93 at the UGT1A10 active site, creating steric hindrance and inhibiting glucuronidation. A UGT1A10 homology model was created based on recent work in our group (Laakkonen and Finel, 2010) and it shows both F90 and F93 located in a surface helix at the far end of the substrate binding site, but within reach of large substrates like estrogens (Fig. 18).

Active site of UGT1A10 with the N-terminal domain (in white) up, and the C-terminal domain (in Figure 18. dark gray) down. Phenylalanines 90 and 93 catalytic histidine H37 are shown. Estradiol (white), 4- nitrophenol (gray) and UDPGA (dark gray) are docked to the active site. (III)

In summary, the results suggest that F93 of the human UGT1A10 is involved in the interactions of the enzyme with ring D of estrogens, but the interactions with smaller molecules were found to be less drastic. Most of the F90 mutants were also inactive in estrogen glucuronidation, suggesting that this residue is also involved in estrogen binding.

OH

OH

OH

OH

OH

OH

OHB C A

52

1A1

1A3

1A4

1A61A

71A8

1A9

1A10 2A

12A2

2A3

2B4

2B7

2B102B

112B

15*2B

172B

281

10

100

1000

10000

16-glucuronide3-glucuronide

17-glucuronide

A. Estriol

pmol

/min

/mg

1A1

1A3

1A4

1A61A

71A8

1A9

1A10 2A

12A

22A

32B

42B7

2B102B

112B

15*

2B17

2B28

1

10

100

1000

10000

16-glucuronide3-glucuronide

17-glucuronide

B. 16-epiestriol

pmol

/min

/mg

1A1

1A3

1A4

1A61A

71A8

1A9

1A10 2A

12A2

2A3

2B4

2B7

2B102B

112B

15*2B

172B

281

10

100

1000

10000

17-glucuronide

3-glucuronide16-glucuronide

C. 17-epiestriol

pmol

/min

/mg

OH

OH

OH

OH

OH

OH

OH

OH

OH

5.3.3. Glucuronidation of estriol and estradiol stereoisomers and enantiomeric steroids The differences found between UGT2A1 and UGT2A2 regioselectivity towards estrogens (I) prompted further study of the glucuronidation of steroids and the differences in the stereo- and regiospecificity between different UGT enzymes. In publication VI a more refined HPLC method was developed for the separation of estriol 16- and 17-glucuronides in addition to methods to study 16- and 17-epiestriol and 13-epiestradiol. The glucuronidation of enantiomeric steroids was also of interest, thus this was studied in publication IV. The glucuronidation of estriol, 16-epiestriol and 17-epiestriol was first screened by single point assays with all the human UGTs (Fig 19).

Glucuronidation rates of the estriol (A), 16-epiestriol (B), and 17-epiestriol (C) at 200 µM substrate Figure 19. concentration. The glucuronidation rates are corrected according to their relative expression level, except for UGT2B15 (indicated with *). Logarithmic scale is used to make low rates visible. (VI)

53

UGTs 1A10 and 2B7 were the most active enzymes with all the substrates. UGT1A10 was the most active UGT in 3-OH glucuronidation, while UGT2B7 was the most active in the glucuronidation of D-ring hydroxyls. Regio- and stereochemical differences in the glucuronidation patterns were found especially among UGTs 2B4, 2B7, and 2B17. UGT2B17 appeared to glucuronidate D-ring hydroxyls exclusively when they were in the

-configuration. With UGTs 2B4 and 2B7, the configuration of C17 seemed to dominate the glucuronidation pattern: the two enzymes glucuronidated 16-OH when C17 was in -configuration and 17-OH when C17 was in -configuration. Enzyme kinetic assays were performed with the most active UGT enzymes, UGTs 1A10 and 2B7, as well as with HIM and HLM (results not shown). We had previously noted a stimulation of estrogen glucuronidation in the UGT1A10 mutant F93G (see chapter 5.3.2. and III) and thus enzyme kinetic assays were performed with this mutant too, to examine the effect of the mutation on the conjugation of 16- and 17-epiestriol. Enzyme kinetic parameters are presented in Table 9.

Table 9. Kinetic constants of estriol, 16-epiestriol and 17-epiestriol glucuronidation by UGT1A10, UGT2B7 and UGT1A10 mutant F93G. All assays except those with UGT1A10 and UGT2B7 with 16-epiestriol followed Michaelis-Menten kinetics. With UGT1A10 and UGT2B7 and 16- epiestriol curves fit best to the Michaelis-Menten with substrate inhibition model.

UGT1A10 1A10F93G UGT2B7 Estriol \ glucuronide formed (3-g) (3-g) (16-g) Vmax (pmol/min/mg) 1200 ± 28.1 7535 ± 618 1509 ± 27.7 Km (µM) 68.4 ± 4.7 230 ± 29.7 10.3 ± 0.8 16-epiestriol \ glucuronide formed (3-g) (3-g) (16-g) Vmax (pmol/min/mg) 7700 ± 740 2249 ± 201 3866 ± 788 Km (µM) 59.8 ± 8.0 186 ± 28.0 38.4 ± 11.4 Ki (µM) 98.1 ± 16.6 162 ± 86.4 17-epiestriol \ glucuronide formed (3-g) (3-g) (17-g) Vmax (pmol/min/mg) 958 ± 75.0 5712 ± 216 820 ± 18.3 Km (µM) 337 ± 42.1 51.1 ± 5.4 0.6 ± 0.1

Kinetic assays revealed that in most cases UGT1A10 and UGT2B7 exhibited relatively low affinity and high turnover rates for the different estriols. The exception was UGT2B7 which had exceptionally low Km value, 0.6 µM, for 17-epiestriol glucuronidation. We had previously noted a high affinity of UGT2B7 to epiestradiol (epiestradiol also has the C17 in -configuration; Itaaho et al., 2008) and these results support our current findings.

54

The effect of the F93G mutation on enzyme kinetics was also interesting, when compared to wild type UGT1A10. When estriol or 16-epiestriol was the substrate, the mutant exhibited approximately threefold higher Km value. With 17-epiestriol, the situation was opposite and the mutant exhibited higher Vmax value and lower Km value in comparison with the wild type UGT1A10. Since it seemed obvious that the configuration of carbons 16 and 17 at the D ring affects the glucuronidation of estrogens, we wanted to see whether the spatial relations between the 17-OH and the C13 methyl group also affected the results. 13-epistradiol was the only currently available estrogen with C13 in the -configuration so this substrate was used to screen UGT activity and compare the results with those gained with natural estradiol (Fig. 20).

Glucuronidation rates and structures of the estradiol (A) and 13-epiestradiol (B) at 200 µM Figure 20. substrate concentration. The glucuronidation rates are corrected according to their relative expression level, except for UGT2B15 (indicated with *). Logarithmic scale is used to make low rates visible. (VI)

The results indicated that configurational change at C13 lowered the 3-OH glucuronidation activity of most UGT1As, including the F93G mutant of UGT1A10. On the other hand, with some enzymes that did not glucuronidate estradiol at the 3-OH, like UGT1A9 and 2B17, 3-glucuronides were detectable when 13-epiestradiol was the substrate. Conjugation at the 17-OH seemed to be stimulated in many cases, possibly due to less steric hindrance at the D-ring. As with 3-OH glucuronidation, some enzymes,

1A1

1A3

1A4

1A6

1A7

1A8

1A9

1A10 2A

12A

22A

32B

42B

72B

102B

112B

15*

2B17

2B28

1A10

-F93G

1

10

100

1000

100003-glucuronide17-glucuronide

A. Estradiol

pmol

/min

/mg

1A1

1A3

1A4

1A6

1A7

1A8

1A9

1A10 2A

12A

22A

32B

42B

72B

102B

112B

15*

2B17

2B28

1A10

-F93G

1

10

100

1000

100003-glucuronide

B. 13-epiestradiol

17-glucuronide

pmol

/min

/mg

OH

OH

OH

OH

55

namely UGT2A2, UGT2B4 and UGT2B15, showed newly gained activity at the 17-OH when 13-epiestradiol was the substrate. In publication IV, we studied the glucuronidation of the enantiomeric estradiol, androsterone, and etiocholanolone. Like with estriols, the glucuronidation rates were first screened at single substrate concentration, 200 µM for ent-estradiol and 50 µM for ent-androsterone and ent-etiocholanolone (Fig. 21).

Glucuronidation rates of the ent-estradiol (A), ent-androsterone, and ent-etiocholanolone (B).The Figure 21. glucuronidation rates are corrected according to their relative expression level, except for UGT2B15 (indicated with *).

For ent-estradiol, the results resembled the glucuronidation pattern of natural estradiol. UGT1As that were active for this substrate conjugated it at the 3-OH. The exception was UGT1A4 which conjugated ent-estradiol at the 17-OH. With natural estradiol, UGT1A4 is also selective for this hydroxyl. With natural estradiol and epiestradiol (17 -estradiol), UGT2A1 and UGT2A2 glucuronidate both hydroxyls, but with ent-estradiol, these enzymes were selective only for 17-OH.

A. Ent-estradiol

B. Ent-androsterone

H

HH

OH

OH

H

H

OH

O

H

H

H

H

OH

O

H

H

and ent-etiocholanolone

56

For UGT2Bs, the results with ent-estradiol, resembled in many cases the glucuronidation pattern of natural estradiol; 2Bs, with the exception of UGT2B15, conjugate it at the 17-OH. Results with UGT2B17 supported our previous finding that this enzyme appears to glucuronidate the D-ring hydroxyls exclusively when they are in the -configuration. UGT2B17 is active toward the natural estradiol but nearly inactive in epiestradiol (Itaaho et al., 2008) and ent-17 -estradiol glucuronidation. In this case, the ent-estradiol glucuronidation resembles the glucuronidation of epiestradiol rather than natural estradiol. Enzyme kinetic analyses were subsequently performed for UGTs 1A10, 2A1, and 2B7, the most active enzymes in ent-estradiol glucuronidation (results not shown). UGT1A10 was confirmed to be the most active enzyme in ent-estradiol glucuronidation, and also to have the lowest Km value for the substrate. Ent-androsterone and ent-etiocholanolone have a single glucuronidation site (3-OH) that is in the -configuration. It is noteworthy that this 3-OH is different from the 3 OH-in estrogens, since in estrogens the A-ring is aromatic, making the 3-OH a phenolic hydroxyl, while in the androgens the A ring is not aromatic and the 3-OH can be in either

or configuration. Most UGTs of subfamily 1A that have good activity toward the 3-OH of ent-estradiol (like UGT1A10), do not glucuronidate ent-androsterone and ent-etiocholanolone at the 3-OH. Instead, the glucuronidation of these hydroxyls resembles the 17-OH glucuronidation of different estrogens. Therefore, we suggest that androsterone, etiocholanolone, ent-androsterone, and ent-etiocholanolone are bound to UGT1A10 and several other UGT1As active sites in the opposite orientation to that of estrogens when they are bound to the same site, so that the 3-OH of these androgens occupy the position that, with bound estrogens, is occupied by the 17-OH (Fig. 22). This assumption is however purely hypothetical since the structure of UGT active site is unknown.

Schematic representation of the orientation of ent-estradiol (A) and ent-etiocholanolone (B) in Figure 22. UGT active site.

In summary, with ent-estradiol the UGTs of subfamily 1A and 2A exhibit a higher degree of regioselectivity than enantioselectivity but with UGT2Bs, the glucuronidation resembles more epiestradiol glucuronidation than estradiol glucuronidation. The

H

HH

OHOHH

H

OH

O

H

H

A B

57

glucuronidation of ent-etiocholanolone and ent-androsterone was catalysed mainly by UGT2A and 2B enzymes and it resembled the glucuronidation of steroid C17 hydroxyls. The results of publications IV and VI are summarised in Table 10. The intensity of the blue colour indicates the probability of a given UGT to glucuronidate a hydroxyl in a certain carbon position and configuration. The intensity of the colour is drawn according to the percentage of substrates glucuronidated. For example, UGT1A10 glucuronidated phenolic C3 hydroxyls in 6/6 situations.

Table 10. Heat map presentation of steroid glucuronidation by different UGTs. Intensity of blue colour correlates positively with the probability of a given UGT to glucuronidate a hydroxyl at certain position and stereochemical configuration. White colour: no activity.

A-ring D-ring Carbon number C3

C16

C17

Configuration - Steroids i ii iii iv v vi

1A1 1A3

1A4

1A5

1A6 1A7

1A8 1A9 1A10

2A1 2A2

2A3 2B4

2B7

2B10 2B11 2B15

2B17

2B28

i) Estradiol, ent-estradiol, estriol, 16-epiestriol, 17-epiestriol, 13-epiestradiol ii) Ent-androsterone, ent-etiocholanolone iii) Estriol, 17-epiestriol iv) 16-epiestriol v) Ent-estradiol, 17-epiestriol vi) Estradiol, estriol, 16-epiestriol

58

The preference of UGT1As (with the exception of UGTs 1A3 and 1A4) and UGT2B15 to phenolic C3 hydroxyls is clearly visible. The UGTs 1A3 and 1A4 seem more flexible with their regio- and stereoselectivity, although UGT1A3 seems to be active towards D-ring hydroxyls when they are in the -configuration and UGT1A4 seems to be completely unable to glucuronidate phenolic C3 hydroxyls. The broad substrate specificity of UGTs 2A1 and 2A2 is also obvious and, especially, UGT2A1 seems to be very flexible for the hydroxyl group orientation. UGTs 2B4 and 2B7 seem to prefer D-ring hydroxyls, except for C17 in the -configuration, and C3 hydroxyls in the -configuration. As noted before, UGT2B17 seems to be able to glucuronidate D–ring hydroxyls only in the -configuration.

59

6. SUMMARY AND CONCLUSIONS

This thesis mainly explored three topics: UGT1A1 and familial hyperbilirubinemias (II, V), the glucuronidation of different endogenous and synthetic steroids (I, III, IV, VI), as well as characterisation of a novel UGT, UGT2A2 (I). Fourteen different UGT1A1 mutations were identified in 19 CN patients, four of which were novel. The results not only broadened the mutational spectrum of CN-syndromes, but also hinted at a founder effect of the disease in the Netherlands, which may be useful information for the genetic counseling of Dutch CN patients. In vitro studies were performed for the first time with seven of these missense mutants as well as three mutations that were reported previously. Residual UGT1A1 activity and the severity of hyperbilirubinemia were found to correlate relatively well with each other in most cases. The information gained should help in predicting the severity of the syndrome in newly identified patients, especially those who are compound heterozygotes. Additionally, one (small) step forward in developing gene therapy for CN-I was obtained in finding that ezetimibe glucuronidation correlates with reduction of serum bilirubin in gene therapy treated Gunn rats. Ezetimibe thus appears to be a useful biomarker for UGT1A1 activity in vivo, and, using it, the problem of estimating the efficacy of gene therapy treatment in patients receiving phototherapy could be overcome. This PhD project also involved the expression and characterization of a novel UGT enzyme, UGT2A2, and additional studies with UGT2A1 and UGT2A3. UGT2A2 was found to be a functional enzyme with broad substrate specificity. The previously suggested exon sharing between UGT2A1 and UGT2A2 was confirmed and UGT2A2 was found to be expressed mainly in the nasal mucosa. Although otherwise very similar, UGT2A1 and UGT2A2 were also found to have large differences in their regiosselectivity towards estrogens and this encouraged us to further investigate the conjugation of steroids and the differences in the stereo- and regiospecificity in the glucuronidation between different human UGTs. The glucuronidation of estriol, 16-epiestriol, 17-epiestriol, 13-epiestradiol, ent-estradiol, ent-androsterone, and ent-etiocholanolone by different UGT enzymes was studied, most of them for the first time, and analytical methods for analysing these compounds were developed. UGTs 1A10 and 2B7 were found to be the most active UGT enzymes in the glucuronidation of almost every steroid studied. While UGT1A10 is mainly expressed in the intestine, the most important UGT in in vivo glucuronidation of steroids is probably the hepatic UGT2B7. Other UGTs, particularly UGT2A1, also exhibited high turnover rates in some cases. Several patterns in the regio- and stereoselectivity of steroid conjugation were observed for the first time, like the activity of UGT2B17 toward D –ring hydroxyls only in the -configuration.

60

The findings with steroid glucuronidation are expected to assist in predicting the glucuronidation of drugs that resemble estrogens, as some future anti-cancer drugs are expected to. The results may also be useful in the future when the interplay of steroid glucuronidation and cancer, like the connection of risk of certain cancers and UGT polymorphisms, becomes clearer. The results on steroid glucuronidation are also useful in exploring the three dimensional structure of the UGT active site and a model of the UGT1A10 binding site was constructed based on information gained from estrogen glucuronidation studies of UGT1A10 mutants.

61

REFERENCES

Adams JB, Phillips NS, Young CE. 1989. Formation of glucuronides of estradiol-17 beta by human mammary cancer cells. J Steroid Biochem 33:1023-1025.

Adegoke OJ, Shu XO, Gao YT, Cai Q, Breyer J, Smith J, Zheng W. 2004. Genetic polymorphisms in uridine diphospho-glucuronosyltransferase 1A1 (UGT1A1) and risk of breast cancer. Breast Cancer Res Treat 85:239-245.

Akaba K, Kimura T, Sasaki A, Tanabe S, Ikegami T, Hashimoto M, Umeda H, Yoshida H, Umetsu K, Chiba H, Yuasa I, Hayasaka K. 1998. Neonatal hyperbilirubinemia and mutation of the bilirubin uridine diphosphate-glucuronosyltransferase gene: A common missense mutation among japanese, koreans and chinese. Biochem Mol Biol Int 46:21-26.

Aono S, Yamada Y, Keino H, Hanada N, Nakagawa T, Sasaoka Y, Yazawa T, Sato H, Koiwai O. 1993. Identification of defect in the genes for bilirubin UDP-glucuronosyl-transferase in a patient with crigler-najjar syndrome type II. Biochem Biophys Res Commun 197:1239-1244.

Arias IM, Gartner LM, Cohen M, Ezzer JB, Levi AJ. 1969. Chronic nonhemolytic unconjugated hyperbilirubinemia with glucuronyl transferase deficiency. clinical, biochemical, pharmacologic and genetic evidence for heterogeneity. Am J Med 47:395-409.

Askari FK, Hitomi Y, Mao M, Wilson JM. 1996. Complete correction of hyperbilirubinemia in the gunn rat model of crigler-najjar syndrome type I following transient in vivo adenovirus-mediated expression of human bilirubin UDP-glucuronosyltransferase. Gene Ther 3:381-388.

Axelrod J, Schmid R, Hammaker L. 1957. A biochemical lesion in congenital, nonobstructive, non-haemolytic jaundice. Nature 180:1426-1427.

Ayan D, Roy J, Maltais R, Poirier D. 2011. Impact of estradiol structural modifications (18-methyl and/or 17-hydroxy inversion of configuration) on the in vitro and in vivo estrogenic activity. J Steroid Biochem Mol Biol 127:324-330.

Bajro MH, Josifovski T, Panovski M, Jankulovski N, Nestorovska AK, Matevska N, Petrusevska N, Dimovski AJ. 2012. Promoter length polymorphism in UGT1A1 and the risk of sporadic colorectal cancer. Cancer Genet 205:163-167.

Belanger A, Brochu M, Lacoste D, Noel C, Labrie F, Dupont A, Cusan L, Caron S, Couture J. 1991. Steroid glucuronides: Human circulatory levels and formation by LNCaP cells. J Steroid Biochem Mol Biol 40:593-598.

Belanger A, Hum DW, Beaulieu M, Levesque E, Guillemette C, Tchernof A, Belanger G, Turgeon D, Dubois S. 1998. Characterization and regulation of UDP-glucuronosyltransferases in steroid target tissues. J Steroid Biochem Mol Biol 65:301-310.

Bellemare J, Rouleau M, Girard H, Harvey M, Guillemette C. 2010. Alternatively spliced products of the UGT1A gene interact with the enzymatically active proteins to inhibit glucuronosyltransferase activity in vitro. Drug Metab Dispos 38:1785-1789.

62

Bellodi-Privato M, Aubert D, Pichard V, Myara A, Trivin F, Ferry N. 2005. Successful gene therapy of the gunn rat by in vivo neonatal hepatic gene transfer using murine oncoretroviral vectors. Hepatology 42:431-438.

Bentue-Ferrer D, Arvieux C, Tribut O, Ruffault A, Bellissant E. 2009. Clinical pharmacology, efficacy and safety of atazanavir: A review. Expert Opin Drug Metab Toxicol 5:1455-1468.

Beutler E, Gelbart T, Demina A. 1998. Racial variability in the UDP-glucuronosyltransferase 1 (UGT1A1) promoter: A balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci U S A 95:8170-8174.

Boiadjiev SE, Lightner DA. 2001. An enantiomerically pure bilirubin. absolute configuration of R, R)-dimethylmesobilirubin-XIII . Tetrahedron: Asymmetry 12:2551-2564.

Bortolussi G, Zentilin L, Baj G, Giraudi P, Bellarosa C, Giacca M, Tiribelli C, Muro AF. 2012. Rescue of bilirubin-induced neonatal lethality in a mouse model of crigler-najjar syndrome type I by AAV9-mediated gene transfer. FASEB J 26:1052-1063.

Bosma PJ. 2003. Inherited disorders of bilirubin metabolism. J Hepatol 38:107-117.

Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, Oostra BA, Lindhout D, Tytgat GN, Jansen PL, Oude Elferink RP. 1995. The genetic basis of the reduced expression of bilirubin UDP-glucuronosyltransferase 1 in gilbert's syndrome. N Engl J Med 333:1171-1175.

Bosma PJ, Chowdhury NR, Goldhoorn BG, Hofker MH, Oude Elferink RP, Jansen PL, Chowdhury JR. 1992. Sequence of exons and the flanking regions of human bilirubin-UDP-glucuronosyltransferase gene complex and identification of a genetic mutation in a patient with crigler-najjar syndrome, type I. Hepatology 15:941-947.

Bosma PJ, Goldhoorn B, Oude Elferink RP, Sinaasappel M, Oostra BA, Jansen PL. 1993. A mutation in bilirubin uridine 5'-diphosphate-glucuronosyltransferase isoform 1 causing crigler-najjar syndrome type II. Gastroenterology 105:216-220.

Bosma PJ, Seppen J, Goldhoorn B, Bakker C, Oude Elferink RP, Chowdhury JR, Chowdhury NR, Jansen PL. 1994. Bilirubin UDP-glucuronosyltransferase 1 is the only relevant bilirubin glucuronidating isoform in man. J Biol Chem 269:17960-17964.

Buning H, Perabo L, Coutelle O, Quadt-Humme S, Hallek M. 2008. Recent developments in adeno-associated virus vector technology. J Gene Med 10:717-733.

Burchell B, Hume R. 1999. Molecular genetic basis of Gilbert's syndrome. J Gastroenterol Hepatol 14:960-966.

Canu G, Minucci A, Zuppi C, Capoluongo E. 2013. Gilbert and crigler najjar syndromes: An update of the UDP-glucuronosyltransferase 1A1 (UGT1A1) gene mutation database. Blood Cells Mol Dis 50:273-280.

Cavalieri EL, Li KM, Balu N, Saeed M, Devanesan P, Higginbotham S, Zhao J, Gross ML, Rogan EG. 2002. Catechol ortho-quinones: the electrophilic compounds that form depurinating DNA adducts and could initiate cancer and other diseases. Carcinogenesis 23: 1071-1077.

63

Chalasani N, Chowdhury NR, Chowdhury JR, Boyer TD. 1997. Kernicterus in an adult who is heterozygous for crigler-najjar syndrome and homozygous for gilbert-type genetic defect. Gastroenterology 112:2099-2103.

Chouinard S, Yueh MF, Tukey RH, Giton F, Fiet J, Pelletier G, Barbier O, Belanger A. 2008. Inactivation by UDP-glucuronosyltransferase enzymes: The end of androgen signaling. J Steroid Biochem Mol Biol 109:247-253.

Chung LW, Coffey DS. 1978. Androgen glucuronide. II. DIfferences in its formation by human normal and benign hyperplastic prostates. Invest Urol 15:385-388.

Ciotti M, Chen F, Rubaltelli FF, Owens IS. 1998. Coding defect and a TATA box mutation at the bilirubin UDP-glucuronosyltransferase gene cause crigler-najjar type I disease. Biochim Biophys Acta 1407:40-50.

Coffman BL, King CD, Rios GR, Tephly TR. 1998. The glucuronidation of opioids, other xenobiotics, and androgens by human UGT2B7Y(268) and UGT2B7H(268). Drug Metab Dispos 26:73-77.

Cohen SL, Marrian GF. 1936. The isolation and identification of a combined form of oestriol in human pregnancy urine. Biochem J 30:57-65.

Costa E. 2006. Hematologically important mutations: Bilirubin UDP-glucuronosyltransferase gene mutations in gilbert and crigler-najjar syndromes. Blood Cells Mol Dis 36:77-80.

Court MH. 2005. Isoform-selective probe substrates for in vitro studies of human UDP-glucuronosyltransferases. Methods Enzymol 400:104-116.

Court MH, Hazarika S, Krishnaswamy S, Finel M, Williams JA. 2008. Novel polymorphic human UDP-glucuronosyltransferase 2A3: Cloning, functional characterization of enzyme variants, comparative tissue expression, and gene induction. Mol Pharmacol 74:744-754.

Court MH, Zhang X, Ding X, Yee KK, Hesse LM, Finel M. 2012. Quantitative distribution of mRNAs encoding the 19 human UDP-glucuronosyltransferase enzymes in 26 adult and 3 fetal tissues. Xenobiotica 42:266-277.

Covey DF. 2009. ent-Steroids: Novel tools for studies of signaling pathways. Steroids 74:577-585.

Cremer RJ, Perryman PW, Richards DH. 1958. Influence of light on the hyperbilirubinaemia of infants. Lancet 1:1094-1097.

Crigler JF,Jr, Najjar VA. 1952. Congenital familial nonhemolytic jaundice with kernicterus. Pediatrics 10:169-180.

Cunningham JM, Hebbring SJ, McDonnell SK, Cicek MS, Christensen GB, Wang L, Jacobsen SJ, Cerhan JR, Blute ML, Schaid DJ, Thibodeau SN. 2007. Evaluation of genetic variations in the androgen and estrogen metabolic pathways as risk factors for sporadic and familial prostate cancer. Cancer Epidemiol Biomarkers Prev 16:969-978.

Cuzick J, Sestak I, Bonanni B, Costantino JP, Cummings S, DeCensi A, Dowsett M, Forbes JF, Ford L, LaCroix AZ, Mershon J, Mitlak BH, Powles T, Veronesi U, Vogel V, Wickerham DL, SERM Chemoprevention of Breast Cancer Overview Group. 2013. Selective oestrogen receptor

64

modulators in prevention of breast cancer: An updated meta-analysis of individual participant data. Lancet 381:1827-1834.

D'Apolito M, Marrone A, Servedio V, Vajro P, De Falco L, Iolascon A. 2007. Seven novel mutations of the UGT1A1 gene in patients with unconjugated hyperbilirubinemia. Haematologica 92:133-134.

Deng WP, Nickoloff JA. 1992. Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal Biochem 200:81-88.

Dennery PA, Seidman DS, Stevenson DK. 2001. Neonatal hyperbilirubinemia. N Engl J Med 344:581-590.

Dimmock D, Brunetti-Pierri N, Palmer DJ, Beaudet AL, Ng P. 2011. Correction of hyperbilirubinemia in gunn rats using clinically relevant low doses of helper-dependent adenoviral vectors. Hum Gene Ther 22:483-488.

Ebner T, Remmel RP, Burchell B. 1993. Human bilirubin UDP-glucuronosyltransferase catalyzes the glucuronidation of ethinylestradiol. Mol Pharmacol 43:649-654.

Eliassen AH, Ziegler RG, Rosner B, Veenstra TD, Roman JM, Xu X, Hankinson SE. 2009. Reproducibility of fifteen urinary estrogens and estrogen metabolites over a 2- to 3-year period in premenopausal women. Cancer Epidemiol Biomarkers Prev 18:2860-2868.

Emons G, Merriam G, Pfeiffer D, Loriaux D, Ball P, Knuppen R. 1987. Metabolism of exogenous 4-and 2-hydroxyestradiol in the human male. J Steroid Biochem 28: 499-504.

Evans WE, Relling MV. 1999. Pharmacogenomics: Translating functional genomics into rational therapeutics. Science 286:487-491.

Faas AE, Sponton ED, Moya PR, Molina JC. 2000. Differential responsiveness to alcohol odor in human neonates: Effects of maternal consumption during gestation. Alcohol 22:7-17.

Ferrari FK, Samulski T, Shenk T, Samulski RJ. 1996. Second-strand synthesis is a rate-limiting step for efficient transduction by recombinant adeno-associated virus vectors. J Virol 70:3227-3234.

Finel M, Kurkela M. 2008. The UDP-glucuronosyltransferases as oligomeric enzymes. Curr Drug Metab 9:70-76.

Fox IJ, Chowdhury JR, Kaufman SS, Goertzen TC, Chowdhury NR, Warkentin PI, Dorko K, Sauter BV, Strom SC. 1998. Treatment of the crigler-najjar syndrome type I with hepatocyte transplantation. N Engl J Med 338:1422-1426.

Fujiwara R, Nakajima M, Oda S, Yamanaka H, Ikushiro S-, Sakaki T, Yokoi T. 2010. Interactions between human UDP-glucuronosyltransferase (UGT) 2B7 and UGT1A enzymes. J Pharm Sci 99:442-454.

Fujiwara R, Nakajima M, Yamamoto T, Nagao H, Yokoi T. 2009. In silico and in vitro approaches to elucidate the thermal stability of human UDP-glucuronosyltransferase (UGT) 1A9. Drug Metab Pharmacokinet 24:235-244.

65

Gallagher CJ, Kadlubar FF, Muscat JE, Ambrosone CB, Lang NP, Lazarus P. 2007. The UGT2B17 gene deletion polymorphism and risk of prostate cancer. A case-control study in caucasians. Cancer Detect Prev 31:310-315.

Gantla S, Bakker CT, Deocharan B, Thummala NR, Zweiner J, Sinaasappel M, Roy Chowdhury J, Bosma PJ, Roy Chowdhury N. 1998. Splice-site mutations: A novel genetic mechanism of crigler-najjar syndrome type 1. Am J Hum Genet 62:585-592.

Ghosal A, Hapangama N, Yuan Y, Achanfuo-Yeboah J, Iannucci R, Chowdhury S, Alton K, Patrick JE, Zbaida S. 2004. Identification of human UDP-glucuronosyltransferase enzyme(s) responsible for the glucuronidation of ezetimibe (zetia). Drug Metab Dispos 32:314-320.

Gilbert A, Lereboulet P. 1901. La cholemie simple familiale. Sem Med 21:241-243.

Girard H, Levesque E, Bellemare J, Journault K, Caillier B, Guillemette C. 2007. Genetic diversity at the UGT1 locus is amplified by a novel 3' alternative splicing mechanism leading to nine additional UGT1A proteins that act as regulators of glucuronidation activity. Pharmacogenet Genomics 17:1077-1089.

Gonzalez FJ, Coughtrie M, Tukey RH. Drug metabolism. In: Goodman & Gilmans’s The Pharmacological Basis of Therapeutics. Brunton LL. Chabner B, Knollman b, Editors. 12th Edition, pp. 123-143. McGraw-Hill, New York 2011.

Grant DJ, Hoyo C, Oliver SD, Gerber L, Shuler K, Calloway E, Gaines AR, McPhail M, Livingston JN, Richardson RM, Schildkraut JM, Freedland SJ. 2013. Association of uridine diphosphate-glucuronosyltransferase 2B gene variants with serum glucuronide levels and prostate cancer risk. Genet Test Mol Biomarkers 17:3-9.

Green PS, Yang SH, Nilsson KR, Kumar AS, Covey DF, Simpkins JW. 2001. The nonfeminizing enantiomer of 17beta-estradiol exerts protective effects in neuronal cultures and a rat model of cerebral ischemia. Endocrinology 142:400-406.

Grieger JC, Samulski RJ. 2012. Adeno-associated virus vectorology, manufacturing, and clinical applications. Methods Enzymol 507:229-254.

Grojean S, Koziel V, Vert P, Daval JL. 2000. Bilirubin induces apoptosis via activation of NMDA receptors in developing rat brain neurons. Exp Neurol 166:334-341.

Grojean S, Lievre V, Koziel V, Vert P, Daval JL. 2001. Bilirubin exerts additional toxic effects in hypoxic cultured neurons from the developing rat brain by the recruitment of glutamate neurotoxicity. Pediatr Res 49:507-513.

Gsur A, Preyer M, Haidinger G, Schatzl G, Madersbacher S, Marberger M, Vutuc C, Micksche M. 2002. A polymorphism in the UDP-Glucuronosyltransferase 2B15 gene (D85Y) is not associated with prostate cancer risk. Cancer Epidemiol Biomarkers Prev 11:497-498.

Guillemette C, De Vivo I, Hankinson SE, Haiman CA, Spiegelman D, Housman DE, Hunter DJ. 2001. Association of genetic polymorphisms in UGT1A1 with breast cancer and plasma hormone levels. Cancer Epidemiol Biomarkers Prev 10:711-714.

Gunn CK. 1944. Hereditary acholuric jaundice in the rat. Can Med Assoc J 50:230-237.

66

Hanukoglu I. 1992. Steroidogenic enzymes: Structure, function, and role in regulation of steroid hormone biosynthesis. J Steroid Biochem Mol Biol 43:779-804.

Harrington WR, Sengupta S, Katzenellenbogen BS. 2006. Estrogen regulation of the glucuronidation enzyme UGT2B15 in estrogen receptor-positive breast cancer cells. Endocrinology 147:3843-3850.

Horsfall LJ, Nazareth I, Petersen I. 2012. Cardiovascular events as a function of serum bilirubin levels in a large, statin-treated cohort. Circulation 126:2556-2564.

Hsieh CJ, Chen MJ, Liao YL, Liao TN. 2008. Polymorphisms of the uridine-diphosphoglucuronosyltransferase 1A1 gene and coronary artery disease. Cell Mol Biol Lett 13:1-10.

Hyde Z, Flicker L, McCaul KA, Almeida OP, Hankey GJ, Chubb SA, Yeap BB. 2012. Associations between testosterone levels and incident prostate, lung, and colorectal cancer. A population-based study. Cancer Epidemiol Biomarkers Prev 21:1319-1329.

Itaaho K, Laakkonen L, Finel M. 2010. How many and which amino acids are responsible for the large activity differences between the highly homologous UDP-glucuronosyltransferases (UGT) 1A9 and UGT1A10? Drug Metab Dispos 38:687-696.

Itaaho K, Mackenzie PI, Ikushiro S, Miners JO, Finel M. 2008. The configuration of the 17-hydroxy group variably influences the glucuronidation of beta-estradiol and epiestradiol by human UDP-glucuronosyltransferases. Drug Metab Dispos 36:2307-2315.

Iyer L, King CD, Whitington PF, Green MD, Roy SK, Tephly TR, Coffman BL, Ratain MJ. 1998. Genetic predisposition to the metabolism of irinotecan (CPT-11). role of uridine diphosphate glucuronosyltransferase isoform 1A1 in the glucuronidation of its active metabolite (SN-38) in human liver microsomes. J Clin Invest 101:847-854.

Jedlitschky G, Cassidy AJ, Sales M, Pratt N, Burchell B. 1999. Cloning and characterization of a novel human olfactory UDP-glucuronosyltransferase. Biochem J 340 ( Pt 3):837-843.

Jia Z, Danko I. 2005a. Long-term correction of hyperbilirubinemia in the gunn rat by repeated intravenous delivery of naked plasmid DNA into muscle. Mol Ther 12:860-866.

Jia Z, Danko I. 2005b. Single hepatic venous injection of liver-specific naked plasmid vector expressing human UGT1A1 leads to long-term correction of hyperbilirubinemia and prevention of chronic bilirubin toxicity in gunn rats. Hum Gene Ther 16:985-995.

Jiraskova A, Novotny J, Novotny L, Vodicka P, Pardini B, Naccarati A, Schwertner HA, Hubacek JA, Puncocharova L, Smerhovsky Z, Vitek L. 2012. Association of serum bilirubin and promoter variations in HMOX1 and UGT1A1 genes with sporadic colorectal cancer. Int J Cancer 131:1549-1555.

Kadakol A, Ghosh SS, Sappal BS, Sharma G, Chowdhury JR, Chowdhury NR. 2000. Genetic lesions of bilirubin uridine-diphosphoglucuronate glucuronosyltransferase (UGT1A1) causing crigler-najjar and gilbert syndromes: Correlation of genotype to phenotype. Hum Mutat 16:297-306.

67

Katona BW, Krishnan K, Cai ZY, Manion BD, Benz A, Taylor A, Evers AS, Zorumski CF, Mennerick S, Covey DF. 2008. Neurosteroid analogues. 12. potent enhancement of GABA-mediated chloride currents at GABAA receptors by ent-androgens. Eur J Med Chem 43:107-113.

Key T, Appleby P, Barnes I, Reeves G, Endogenous Hormones and Breast Cancer Collaborative Group. 2002. Endogenous sex hormones and breast cancer in postmenopausal women: Reanalysis of nine prospective studies. J Natl Cancer Inst 94:606-616.

Kimm H, Yun JE, Jo J, Jee SH. 2009. Low serum bilirubin level as an independent predictor of stroke incidence: A prospective study in korean men and women. Stroke 40:3422-3427.

Kirdani R, Barua NR, Sandberg AA. 1977. Studies on phenolic steriods in human subjects. XXI. renal metabolism, conjugation and excretion of four androgens: A comparison with estriol. J Clin Endocrinol Metab 44:1121-1129.

Kirkby KA, Adin CA. 2006. Products of heme oxygenase and their potential therapeutic applications. Am J Physiol Renal Physiol 290:F563-71.

Krishnan K, Manion BD, Taylor A, Bracamontes J, Steinbach JH, Reichert DE, Evers AS, Zorumski CF, Mennerick S, Covey DF. 2012. Neurosteroid analogues. 17. inverted binding orientations of androsterone enantiomers at the steroid potentiation site on gamma-aminobutyric acid type A receptors. J Med Chem 55:1334-1345.

Kurkela M, Garcia-Horsman JA, Luukkanen L, Morsky S, Taskinen J, Baumann M, Kostiainen R, Hirvonen J, Finel M. 2003. Expression and characterization of recombinant human UDP-glucuronosyltransferases (UGTs). UGT1A9 is more resistant to detergent inhibition than other UGTs and was purified as an active dimeric enzyme. J Biol Chem 278:3536-3544.

Kurkela M, Patana AS, Mackenzie PI, Court MH, Tate CG, Hirvonen J, Goldman A, Finel M. 2007. Interactions with other human UDP-glucuronosyltransferases attenuate the consequences of the Y485D mutation on the activity and substrate affinity of UGT1A6. Pharmacogenet Genomics 17:115-126.

Kuuranne T, Kurkela M, Thevis M, Schanzer W, Finel M, Kostiainen R. 2003. Glucuronidation of anabolic androgenic steroids by recombinant human UDP-glucuronosyltransferases. Drug Metab Dispos 31:1117-1124.

Laakkonen L, Finel M. 2010. A molecular model of the human UDP-glucuronosyltransferase 1A1, its membrane orientation, and the interactions between different parts of the enzyme. Mol Pharmacol 77:931-939.

Lazard D, Zupko K, Poria Y, Nef P, Lazarovits J, Horn S, Khen M, Lancet D. 1991. Odorant signal termination by olfactory UDP glucuronosyl transferase. Nature 349:790-793.

Lepine J, Bernard O, Plante M, Tetu B, Pelletier G, Labrie F, Belanger A, Guillemette C. 2004. Specificity and regioselectivity of the conjugation of estradiol, estrone, and their catecholestrogen and methoxyestrogen metabolites by human uridine diphospho-glucuronosyltransferases expressed in endometrium. J Clin Endocrinol Metab 89:5222-5232.

Levesque E, Beaulieu M, Green MD, Tephly TR, Belanger A, Hum DW. 1997. Isolation and characterization of UGT2B15(Y85): A UDP-glucuronosyltransferase encoded by a polymorphic gene. Pharmacogenetics 7:317-325.

68

Levesque E, Girard H, Journault K, Lepine J, Guillemette C. 2007. Regulation of the UGT1A1 bilirubin-conjugating pathway: Role of a new splicing event at the UGT1A locus. Hepatology 45:128-138.

Li C, Wu Q. 2007. Adaptive evolution of multiple-variable exons and structural diversity of drug-metabolizing enzymes. BMC Evol Biol 7:69.

Liehr JG, Ricci MJ. 1996. 4-hydroxylation of estrogens as marker of human mammary tumors. Proc Natl Acad Sci U S A 93:3294-3296.

Lightner DA, Wooldridge TA, McDonagh AF. 1979. Configurational isomerization of bilirubin and the mechanism of jaundice phototherapy. Biochem Biophys Res Commun 86:235-243.

Luukkanen L, Mikkola J, Forsman T, Taavitsainen P, Taskinen J, Elovaara E. 2001. Glucuronidation of 1-hydroxypyrene by human liver microsomes and human UDP-glucuronosyltransferases UGT1A6, UGT1A7, and UGT1A9: Development of a high-sensitivity glucuronidation assay for human tissue. Drug Metab Dispos 29:1096-1101.

Luukkanen L, Taskinen J, Kurkela M, Kostiainen R, Hirvonen J, Finel M. 2005. Kinetic characterization of the 1A subfamily of recombinant human UDP-glucuronosyltransferases. Drug Metab Dispos 33:1017-1026.

Lysy PA, Najimi M, Stephenne X, Bourgois A, Smets F, Sokal EM. 2008. Liver cell transplantation for crigler-najjar syndrome type I: Update and perspectives. World J Gastroenterol 14:3464-3470.

Mackenzie PI, Bock KW, Burchell B, Guillemette C, Ikushiro S, Iyanagi T, Miners JO, Owens IS, Nebert DW. 2005. Nomenclature update for the mammalian UDP glycosyltransferase (UGT) gene superfamily. Pharmacogenet Genomics 15:677-685.

MacLeod SL, Nowell S, Plaxco J, Lang NP. 2000. An allele-specific polymerase chain reaction method for the determination of the D85Y polymorphism in the human UDP-glucuronosyltransferase 2B15 gene in a case-control study of prostate cancer. Ann Surg Oncol 7:777-782.

Marques SC, Ikediobi ON. 2010. The clinical application of UGT1A1 pharmacogenetic testing: Gene-environment interactions. Hum Genomics 4:238-249.

McCarty DM, Monahan PE, Samulski RJ. 2001. Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis. Gene Ther 8:1248-1254.

McDonald JW, Shapiro SM, Silverstein FS, Johnston MV. 1998. Role of glutamate receptor-mediated excitotoxicity in bilirubin-induced brain injury in the gunn rat model. Exp Neurol 150:21-29.

McGrath M, Lepine J, Lee IM, Villeneuve L, Buring J, Guillemette C, De Vivo I. 2009. Genetic variations in UGT1A1 and UGT2B7 and endometrial cancer risk. Pharmacogenet Genomics 19:239-243.

Meech R, Mackenzie PI. 1997. UDP-glucuronosyltransferase, the role of the amino terminus in dimerization. J Biol Chem 272:26913-26917.

69

Meech R, Mackenzie PI. 1998. Determinants of UDP glucuronosyltransferase membrane association and residency in the endoplasmic reticulum. Arch Biochem Biophys 356:77-85.

Miley MJ, Zielinska AK, Keenan JE, Bratton SM, Radominska-Pandya A, Redinbo MR. 2007. Crystal structure of the cofactor-binding domain of the human phase II drug-metabolism enzyme UDP-glucuronosyltransferase 2B7. J Mol Biol 369:498-511.

Moghrabi N, Clarke DJ, Boxer M, Burchell B. 1993. Identification of an A-to-G missense mutation in exon 2 of the UGT1 gene complex that causes crigler-najjar syndrome type 2. Genomics 18:171-173.

Monaghan G, Ryan M, Seddon R, Hume R, Burchell B. 1996. Genetic variation in bilirubin UPD-glucuronosyltransferase gene promoter and gilbert's syndrome. Lancet 347:578-581.

Montenegro-Miranda PS, ten Bloemendaal L, Kunne C, de Waart DR, Bosma PJ. 2011. Mycophenolate mofetil impairs transduction of single-stranded adeno-associated viral vectors. Hum Gene Ther 22:605-612.

Morgan DJ. 1997. Drug disposition in mother and foetus. Clin Exp Pharmacol Physiol 24:869-873.

Muraca M, Fevery J. 1984. Influence of sex and sex steroids on bilirubin uridine diphosphate-glucuronosyltransferase activity of rat liver. Gastroenterology 87:308-313.

Murai T, Samata N, Iwabuchi H, Ikeda T. 2006. Human UDP-glucuronosyltransferase, UGT1A8, glucuronidates dihydrotestosterone to a monoglucuronide and further to a structurally novel diglucuronide. Drug Metab Dispos 34:1102-1108.

Nguyen N, Bonzo JA, Chen S, Chouinard S, Kelner MJ, Hardiman G, Belanger A, Tukey RH. 2008. Disruption of the ugt1 locus in mice resembles human crigler-najjar type I disease. J Biol Chem 283:7901-7911.

Nguyen TH, Aubert D, Bellodi-Privato M, Flageul M, Pichard V, Jaidane-Abdelghani Z, Myara A, Ferry N. 2007. Critical assessment of lifelong phenotype correction in hyperbilirubinemic gunn rats after retroviral mediated gene transfer. Gene Ther 14:1270-1277.

Ohno S, Nakajin S. 2009. Determination of mRNA expression of human UDP-glucuronosyltransferases and application for localization in various human tissues by real-time reverse transcriptase-polymerase chain reaction. Drug Metab Dispos 37:32-40.

Okugi H, Nakazato H, Matsui H, Ohtake N, Nakata S, Suzuki K. 2006. Association of the polymorphisms of genes involved in androgen metabolism and signaling pathways with familial prostate cancer risk in a japanese population. Cancer Detect Prev 30:262-268.

Olsson M, Lindstrom S, Haggkvist B, Adami HO, Balter K, Stattin P, Ask B, Rane A, Ekstrom L, Gronberg H. 2008. The UGT2B17 gene deletion is not associated with prostate cancer risk. Prostate 68:571-575.

Operana TN, Tukey RH. 2007. Oligomerization of the UDP-glucuronosyltransferase 1A proteins: Homo- and heterodimerization analysis by fluorescence resonance energy transfer and co-immunoprecipitation. J Biol Chem 282:4821-4829.

70

Ostrow JD, Pascolo L, Shapiro SM, Tiribelli C. 2003. New concepts in bilirubin encephalopathy. Eur J Clin Invest 33:988-997.

Owens D, Evans J. 1975. Population studies on gilbert's syndrome. J Med Genet 12:152-156.

Park J, Chen L, Ratnashinge L, Sellers TA, Tanner JP, Lee JH, Dossett N, Lang N, Kadlubar FF, Ambrosone CB, Zachariah B, Heysek RV, Patterson S, Pow-Sang J. 2006. Deletion polymorphism of UDP-glucuronosyltransferase 2B17 and risk of prostate cancer in african american and caucasian men. Cancer Epidemiol Biomarkers Prev 15:1473-1478.

Park J, Chen L, Shade K, Lazarus P, Seigne J, Patterson S, Helal M, Pow-Sang J. 2004. Asp85tyr polymorphism in the udp-glucuronosyltransferase (UGT) 2B15 gene and the risk of prostate cancer. J Urol 171:2484-2488.

Pastore N, Nusco E, Vanikova J, Sepe RM, Vetrini F, McDonagh A, Auricchio A, Vitek L, Brunetti-Pierri N. 2012. Sustained reduction of hyperbilirubinemia in gunn rats after adeno-associated virus-mediated gene transfer of bilirubin UDP-glucuronosyltransferase isozyme 1A1 to skeletal muscle. Hum Gene Ther 23:1082-1089.

Patana AS, Kurkela M, Finel M, Goldman A. 2008. Mutation analysis in UGT1A9 suggests a relationship between substrate and catalytic residues in UDP-glucuronosyltransferases. Protein Eng Des Sel 21:537-543.

Patana AS, Kurkela M, Goldman A, Finel M. 2007. The human UDP-glucuronosyltransferase: identification of key residues within the nucleotide-sugar binding site. Mol Pharmacol 72:604-611.

Paulusma CC, Kool M, Bosma PJ, Scheffer GL, ter Borg F, Scheper RJ, Tytgat GN, Borst P, Baas F, Oude Elferink RP. 1997. A mutation in the human canalicular multispecific organic anion transporter gene causes the dubin-johnson syndrome. Hepatology 25:1539-1542.

Payne DW, Katzenellenbogen JA. 1979. Binding specificity of rat alpha-fetoprotein for a series of estrogen derivatives: Studies using equilibrium and nonequilibrium binding techniques. Endocrinology 105:745-753.

Perlstein TS, Pande RL, Beckman JA, Creager MA. 2008. Serum total bilirubin level and prevalent lower-extremity peripheral arterial disease: National health and nutrition examination survey (NHANES) 1999 to 2004. Arterioscler Thromb Vasc Biol 28:166-172.

Perreault M, Gauthier-Landry L, Trottier J, Verreault M, Caron P, Finel M, Barbier O. 2013. The human UDP-glucuronosyltransferase UGT2A1 and UGT2A2 enzymes are highly active in bile acid glucuronidation. Drug Metab Dispos (epub ahead of print).

Poddar B, Bharti B, Goraya J, Parmar VR. 2002. Kernicterus in a child with crigler-najjar syndrome type II. Trop Gastroenterol 23:33-34.

Radominska-Pandya A, Little JM, Czernik PJ. 2001. Human UDP-glucuronosyltransferase 2B7. Curr Drug Metab 2:283-298.

Raftogianis R, Creveling C, Weinshilboum R, Weisz J. 2000. Estrogen metabolism by conjugation. J Natl Cancer Inst Monogr 2000:113-124.

71

Rantner B, Kollerits B, Anderwald-Stadler M, Klein-Weigel P, Gruber I, Gehringer A, Haak M, Schnapka-Kopf M, Fraedrich G, Kronenberg F. 2008. Association between the UGT1A1 TA-repeat polymorphism and bilirubin concentration in patients with intermittent claudication: Results from the CAVASIC study. Clin Chem 54:851-857.

Rela M, Muiesan P, Vilca-Melendez H, Dhawan A, Baker A, Mieli-Vergani G, Heaton ND. 1999. Auxiliary partial orthotopic liver transplantation for crigler-najjar syndrome type I. Ann Surg 229:565-569.

Rodrigues CM, Sola S, Silva R, Brites D. 2000. Bilirubin and amyloid-beta peptide induce cytochrome c release through mitochondrial membrane permeabilization. Mol Med 6:936-946.

Sandberg AA, Slaunwhite WR,Jr. 1965. Studies on phenolic steroids in human subjects. vii. metabolic fate of estriol and its glucuronide. J Clin Invest 44:694-702.

Sato H, Aono S, Kashiwamata S, Koiwai O. 1991. Genetic defect of bilirubin UDP-glucuronosyltransferase in the hyperbilirubinemic gunn rat. Biochem Biophys Res Commun 177:1161-1164.

Sauter BV, Parashar B, Chowdhury NR, Kadakol A, Ilan Y, Singh H, Milano J, Strayer DS, Chowdhury JR. 2000. A replication-deficient rSV40 mediates liver-directed gene transfer and a long-term amelioration of jaundice in gunn rats. Gastroenterology 119:1348-1357.

Schaal B, Marlier L, Soussignan R. 1998. Olfactory function in the human fetus: Evidence from selective neonatal responsiveness to the odor of amniotic fluid. Behav Neurosci 112:1438-1449.

Schaal B, Marlier L, Soussignan R. 2000. Human foetuses learn odours from their pregnant mother's diet. Chem Senses 25:729-737.

Schmitt F, Remy S, Dariel A, Flageul M, Pichard V, Boni S, Usal C, Myara A, Laplanche S, Anegon I, Labrune P, Podevin G, Ferry N, Nguyen TH. 2010. Lentiviral vectors that express UGT1A1 in liver and contain miR-142 target sequences normalize hyperbilirubinemia in gunn rats. Gastroenterology 139:999-1007, 1007.e1-2.

Sedlak TW, Saleh M, Higginson DS, Paul BD, Juluri KR, Snyder SH. 2009. Bilirubin and glutathione have complementary antioxidant and cytoprotective roles. Proc Natl Acad Sci U S A 106:5171-5176.

Seppen J, Bakker C, de Jong B, Kunne C, van den Oever K, Vandenberghe K, de Waart R, Twisk J, Bosma P. 2006. Adeno-associated virus vector serotypes mediate sustained correction of bilirubin UDP glucuronosyltransferase deficiency in rats. Mol Ther 13:1085-1092.

Seppen J, Bosma PJ, Goldhoorn BG, Bakker CT, Chowdhury JR, Chowdhury NR, Jansen PL, Oude Elferink RP. 1994. Discrimination between crigler-najjar type I and II by expression of mutant bilirubin uridine diphosphate-glucuronosyltransferase. J Clin Invest 94:2385-2391.

Seppen J, Steenken E, Lindhout D, Bosma PJ, Elferink RP. 1996. A mutation which disrupts the hydrophobic core of the signal peptide of bilirubin UDP-glucuronosyltransferase, an endoplasmic reticulum membrane protein, causes crigler-najjar type II. FEBS Lett 390:294-298.

Setlur SR, Chen CX, Hossain RR, Ha JS, Van Doren VE, Stenzel B, Steiner E, Oldridge D, Kitabayashi N, Banerjee S, Chen JY, Schafer G, Horninger W, Lee C, Rubin MA, Klocker H,

72

Demichelis F. 2010. Genetic variation of genes involved in dihydrotestosterone metabolism and the risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 19:229-239.

Shatalova EG, Loginov VI, Braga EA, Kazubskaia TP, Sudomoina MA, Blanchard RL, Favorova OO. 2006. Association of polymorphisms in SULT1A1 and UGT1A1 genes with breast cancer risk and phenotypes in russian women. Mol Biol (Mosk) 40:263-270.

Smith ER, Kellie AE. 1967. Oestrogen conjugates of human late-pregnancy urine. Biochem J 104:83-94.

Smith PA, Sorich MJ, McKinnon RA, Miners JO. 2003. Pharmacophore and quantitative structure-activity relationship modeling: Complementary approaches for the rationalization and prediction of UDP-glucuronosyltransferase 1A4 substrate selectivity. J Med Chem 46:1617-1626.

Somers GI, Lindsay N, Lowdon BM, Jones AE, Freathy C, Ho S, Woodrooffe AJ, Bayliss MK, Manchee GR. 2007. A comparison of the expression and metabolizing activities of phase I and II enzymes in freshly isolated human lung parenchymal cells and cryopreserved human hepatocytes. Drug Metab Dispos 35:1797-1805.

Stack DE, Byun J, Gross ML, Rogan EG, Cavalieri EL. 1996. Molecular characteristics of catechol estrogen quinones in reactions with deoxyribonucleosides. Chem Res Toxicol 9: 851-859.

Starlard-Davenport A, Lyn-Cook B, Radominska-Pandya A. 2008a. Identification of UDP-glucuronosyltransferase 1A10 in non-malignant and malignant human breast tissues. Steroids 73:611-620.

Starlard-Davenport A, Lyn-Cook B, Radominska-Pandya A. 2008b. Novel identification of UDP-glucuronosyltransferase 1A10 as an estrogen-regulated target gene. Steroids 73:139-147.

Starlard-Davenport A, Xiong Y, Bratton S, Gallus-Zawada A, Finel M, Radominska-Pandya A. 2007. Phenylalanine(90) and phenylalanine(93) are crucial amino acids within the estrogen binding site of the human UDP-glucuronosyltransferase 1A10. Steroids 72:85-94.

Sten T, Bichlmaier I, Kuuranne T, Leinonen A, Yli-Kauhaluoma J, Finel M. 2009a. UDP-glucuronosyltransferases (UGTs) 2B7 and UGT2B17 display converse specificity in testosterone and epitestosterone glucuronidation, whereas UGT2A1 conjugates both androgens similarly. Drug Metab Dispos 37:417-423.

Sten T, Kurkela M, Kuuranne T, Leinonen A, Finel M. 2009b. UDP-glucuronosyltransferases in conjugation of 5alpha- and 5beta-androstane steroids. Drug Metab Dispos 37:2221-2227.

Stocker R, Yamamoto Y, McDonagh AF, Glazer AN, Ames BN. 1987. Bilirubin is an antioxidant of possible physiological importance. Science 235:1043-1046.

Strassburg CP. 2010. Hyperbilirubinemia syndromes (gilbert-meulengracht, crigler-najjar, dubin-johnson, and rotor syndrome). Best Pract Res Clin Gastroenterol 24:555-571.

Tenhunen R, Marver HS, Schmid R. 1969. Microsomal heme oxygenase. characterization of the enzyme. J Biol Chem 244:6388-6394.

73

Tenhunen R, Ross ME, Marver HS, Schmid R. 1970. Reduced nicotinamide-adenine dinucleotide phosphate dependent biliverdin reductase: Partial purification and characterization. Biochemistry 9:298-303.

Thapa BR, Yachha SK, Mehta S. 1987. Crigler-najjar syndrome type II with kernicterus. Indian Pediatr 24:680-683.

Thiebaud N, Veloso Da Silva S, Jakob I, Sicard G, Chevalier J, Menetrier F, Berdeaux O, Artur Y, Heydel JM, Le Bon AM. 2013. Odorant metabolism catalyzed by olfactory mucosal enzymes influences peripheral olfactory responses in rats. PLoS One 8:e59547.

Thum T, Erpenbeck VJ, Moeller J, Hohlfeld JM, Krug N, Borlak J. 2006. Expression of xenobiotic metabolizing enzymes in different lung compartments of smokers and nonsmokers. Environ Health Perspect 114:1655-1661.

Tiribelli C, Ostrow JD. 1996. New concepts in bilirubin and jaundice: Report of the third international bilirubin workshop, april 6-8, 1995, trieste, italy. Hepatology 24:1296-1311.

Toietta G, Mane VP, Norona WS, Finegold MJ, Ng P, McDonagh AF, Beaudet AL, Lee B. 2005. Lifelong elimination of hyperbilirubinemia in the gunn rat with a single injection of helper-dependent adenoviral vector. Proc Natl Acad Sci U S A 102:3930-3935.

Tong Z, Li H, Goljer I, McConnell O, Chandrasekaran A. 2007. In vitro glucuronidation of thyroxine and triiodothyronine by liver microsomes and recombinant human UDP-glucuronosyltransferases. Drug Metab Dispos 35:2203-2210.

Tukey RH, Strassburg CP. 2001. Genetic multiplicity of the human UDP-glucuronosyltransferases and regulation in the gastrointestinal tract. Mol Pharmacol 59:405-414.

van der Veere CN, Sinaasappel M, McDonagh AF, Rosenthal P, Labrune P, Odievre M, Fevery J, Otte JB, McClean P, Burk G, Masakowski V, Sperl W, Mowat AP, Vergani GM, Heller K, Wilson JP, Shepherd R, Jansen PL. 1996. Current therapy for crigler-najjar syndrome type 1: Report of a world registry. Hepatology 24:311-315.

VanLandingham JW, Cutler SM, Virmani S, Hoffman SW, Covey DF, Krishnan K, Hammes SR, Jamnongjit M, Stein DG. 2006. The enantiomer of progesterone acts as a molecular neuroprotectant after traumatic brain injury. Neuropharmacology 51:1078-1085.

Vitek L, Jirsa M, Brodanova M, Kalab M, Marecek Z, Danzig V, Novotny L, Kotal P. 2002. Gilbert syndrome and ischemic heart disease: A protective effect of elevated bilirubin levels. Atherosclerosis 160:449-456.

Walker K, Bratton DJ, Frost C. 2011. Premenopausal endogenous oestrogen levels and breast cancer risk: A meta-analysis. Br J Cancer 105:1451-1457.

Wennberg RP. 2000. The blood-brain barrier and bilirubin encephalopathy. Cell Mol Neurobiol 20:97-109.

Wienkers LC, Heath TG. 2005. Predicting in vivo drug interactions from in vitro drug discovery data. Nat Rev Drug Discov 4:825-833.

74

Williams JA, Hyland R, Jones BC, Smith DA, Hurst S, Goosen TC, Peterkin V, Koup JR, Ball SE. 2004. Drug-drug interactions for UDP-glucuronosyltransferase substrates: A pharmacokinetic explanation for typically observed low exposure (AUCi/AUC) ratios. Drug Metab Dispos 32:1201-1208.

Wilson W,3rd, Pardo-Manuel de Villena F, Lyn-Cook BD, Chatterjee PK, Bell TA, Detwiler DA, Gilmore RC, Valladeras IC, Wright CC, Threadgill DW, Grant DJ. 2004. Characterization of a common deletion polymorphism of the UGT2B17 gene linked to UGT2B15. Genomics 84:707-714.

Xiong Y, Bernardi D, Bratton S, Ward MD, Battaglia E, Finel M, Drake RR, Radominska-Pandya A. 2006. Phenylalanine 90 and 93 are localized within the phenol binding site of human UDP-glucuronosyltransferase 1A10 as determined by photoaffinity labeling, mass spectrometry, and site-directed mutagenesis. Biochemistry 45:2322-2332.

Yin H, Bennett G, Jones JP. 1994. Mechanistic studies of uridine diphosphate glucuronosyltransferase. Chem Biol Interact 90:47-58.

You L. 2004. Steroid hormone biotransformation and xenobiotic induction of hepatic steroid metabolizing enzymes. Chem Biol Interact 147:233-246.

Zahid M, Kohli E, Saeed M, Rogan E, Cavalieri E. 2006. The greater reactivity of estradiol-3, 4-quinone vs estradiol-2, 3-quinone with DNA in the formation of depurinating adducts: implications for tumor-initiating activity. Chem Res Toxicol 19: 164-172.

Zhang H, Tolonen A, Rousu T, Hirvonen J, Finel M. 2011. Effects of cell differentiation and assay conditions on the UDP-glucuronosyltransferase activity in caco-2 cells. Drug Metab Dispos 39:456-464.

Zhang L, Liu W, Tanswell AK, Luo X. 2003. The effects of bilirubin on evoked potentials and long-term potentiation in rat hippocampus in vivo. Pediatr Res 53:939-944.

Zmetakova I, Ferak V, Minarik G, Ficek A, Polakova H, Ferakova E, Kadasi L. 2007. Identification of the deletions in the UGT1A1 gene of the patients with crigler-najjar syndrome type I from slovakia. Gen Physiol Biophys 26:306-310.

Zucker SD, Goessling W, Hoppin AG. 1999. Unconjugated bilirubin exhibits spontaneous diffusion through model lipid bilayers and native hepatocyte membranes. J Biol Chem 274:10852-10862.

INTERNET SOURCES: FDA Guidance for Industry: Safety Testing of Drug Metabolites http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm079266.pdf (last accessed on 13 August 2013) U.S. National Library of Medicine http://chem.sis.nlm.nih.gov/chemidplus (last accessed on 24 March 2013) NetNGlyc 1.0 Server http://www.cbs.dtu.dk/services/NetNGlyc/ (last accessed on 17 May 2013) RxList - The Internet Drug Index www.rxlist.com (last accessed on 2 April 2013)