(Title of the Thesis)*

CLONING AND FUNCTIONAL CHARACTERIZATION OF THE RETINOIC ACID-CATABOLIZING ENZYME CYP26B1 IN MOUSE DEVELOPMENT

Glenn Alexander MacLean

A thesis submitted to the Department of Pathology and Molecular Medicine

In conformity with the requirements for

the degree of Doctor of Philosophy

Queen‟s University

Kingston, Ontario, Canada

(September 2007)

Retinoic acid (RA) is an active metabolite of vitamin A that is essential for embryonic development, and homeostasis of adult tissues. RA is a ligand for the nuclear retinoic acid receptor, and RA-mediated signaling is critical for regulation of cell proliferation, differentiation and apoptosis. There is a spatio-temporal distribution of RA in the developing embryo such that some tissues are rich in RA, while others are devoid. This patterned distribution of RA is tightly controlled through the coordinated expression of

RA-synthesizing (retinaldehyde dehydrogenase) and RA-catabolizing (CYP26) enzymes.

In this thesis, I describe the cloning of a mouse gene encoding one of the CYP26 proteins, Cyp26b1. Cyp26b1 was shown to be highly expressed in the embryo, with transcripts localized to the hindbrain, limb buds and branchial arches. We also used homologous recombination to generate a line of transgenic mice with a loss-of-function deletion in Cyp26b1. These mice die shortly after birth with severe malformations affecting the limbs, craniofacial structures and epidermis; phenotypes that are all reminiscent of RA teratogenesis. We present an extensive characterization of the craniofacial and epidermal abnormalities in Cyp26b1-/- animals, and examine several molecular pathways that may be deregulated. CYP26B1 null embryos exhibit a truncated mandible, lack numerous facial bones, and show reduced ossification of the calvaria.

Molecular analysis of Cyp26b1-/- embryos indicates hindbrain and branchial arch patterning is largely unaffected in early to mid-gestational mutants. However, there appear to be some subtle abnormalities in neural crest cell migration, which may contribute to the development of some of the observed phenotypes. CYP26B1 null

i mutants also lack hair follicles, which appears to be due to a downregulation of -catenin mediated signaling. Thus, in addition to cloning and characterizing the expression of murine Cyp26b1, we have demonstrated in vivo, that regulation of RA distribution by

CYP26B1 is essential for morphogenesis of the epidermis and craniofacial structures.

ii Co-Authorship

Chapter 3 is a peer-reviewed manuscript, as is the appendix, and contributions by my colleagues are noted below. Chapters 4 and 5 were prepared by myself, and will be submitted for publication.

Chapter 31 Suzan Abu-Abed, my co-author performed the in situ hybridization experiments with the probe for Cyp26a1 Pascal Dollé prepared Figure 6, and also performed the staining on sections in Figure 5 F and Figure 7 F, G

Chapter 4 Daniel Metzger assisted in generating the mutant embryonic stem cells and mouse lines.

Chapter 5 Victor Tron assisted me in analyzing sections and immunohistochemistry for desmin and muscle-specific actin (Figure 5.2) was performed at the Pathology lab at Kingston General Hospital.

Appendix2 Hui Li, my co-author, had a primary role in analyzing the testes of mutant mice

1-MacLean, G., Abu-Abed, S, Dollé, P., Tahayato, A., Chambon, P., Petkovich, M. 2001. Cloning of a novel retinoic-acid metabolizing cytochrome P450, Cyp26B1, and comparative expression analysis with Cyp26A1 during early murine development. Mechanisms of Development, 107:195-201.

2-MacLean, G., Li, H., Metzger, D,. Chambon, P., Petkovich, M. 2007. Apoptotic extinction of germ cells in testes of Cyp26b1 knockout mice. Endocrinology, 148: 4560-4567.

iii Acknowledgements

First of all, I would like to thank my supervisor, mentor and friend, Dr. Martin Petkovich.

Marty, over the years, I have grown both as a scientist and a person, and you have provided an environment to allow that to happen. You have given me a great deal of freedom, but whenever I have needed guidance, you have been there for me. I think much differently than when I entered the lab, and you are the primary reason for that. I also can not emphasize enough what your friendship has meant to me over the years. You have been one of the most influencial people in my life, and for that I will always be grateful.

To Naomi Dore, Don Cameron and Tracie Pennimpede. I am very fortunate because I think that I am one of the few people that can say that i) they have a complete passion for their work, and ii) they work with their best friends. You are the ones who made the lab special to me.

You all know how much I love the lab, but in all honesty, I will miss each of you more than my beady-eyed minions. It has been a pleasure to work with each of you, without your friendship and support, I‟m not sure how I could have gotten through the rough patches. We have had a lot of great times both in and out of the lab, and I treasure those memories.

To the rest of my friends in Kingston, especially Megan, Doug, Ashley, Carolyn,

Erin (x2), Ben, Mark, Dave and Tara. I really can‟t express how much you all mean to me. You have all been tremendous friends, and whether you knew it or not, you‟ve all helped me through some very difficult times, and we have shared more than a few laughs along the way. I sincerely hope that we keep in touch wherever life‟s paths take us.

I would also like to thank Dr. Ali Tahayato and Dr. Olivier Loudig for keeping an eye on the Rock, when he was just a little pebble. You were both instrumental in my early development as a scientist, and I am grateful for that.

iv Thanks to Larry the geranium for brightening room 357 every day.

Thank you also to the Department of Pathology and Molecular Medicine, and especially

Barb Saunders for her patience and assistance over the years.

I would like to also thank Lloyd, Darrell, Kate, Bernie, Kir and Sam. You have always been supportive of me, and helped remind me of what is truly important in life.

Finally, I dedicate this thesis to Mom, the one person to whom it meant the most.

v Table of Contents

Abstract ...... i Co-Authorship ...... iii Acknowledgements ...... iv Table of Contents ...... vi List of Figures ...... x Abreviations……………………………………………………………………………………..xii Chapter 1 General Introduction ...... 1 Chapter 2 Literature Review ...... 5 2.1 Historical Perspective………..…………………………………………...... 6 2.2 Overview of Vitamin A Metabolism………………………..….……………9 2.3 Retinoic Acid is a Transcriptional Regulator……………………….…….13 2.3.1 Nuclear retinoic acid receptors…………………………………………13 2.3.2 Ligands for retinoic acid receptors……………………………………..15 2.4 Regulation of Retinoic Acid Signaling………………………….…………17 2.4.1 Retinoic acid receptor expression and function………………………...17 2.4.2 Retinoid binding proteins……………………………………………….19 2.5 Retinoic Acid Localization...... 21 2.6 Regulation of Retinoic Acid Synthesis………………………………….…22 2.6.1 Regulation of RA Synthesis by Alcohol Dehydrogenases…………….22 2.6.1 Regulation of RA Synthesis by Retinaldehyde Dehydrogenases……....24 2.7 Regulation of Retinoic Acid Catabolism……………………………….....26 2.7.1 Hydroxylation of retinoic acid…………………………………………26 2.7.2 Identification of CYP26A1 as retinoic acid hydroxylase……………...27 2.7.3 Retinoic acid is the biologically active form of vitamin A…………….29 2.7.4 Identification of CYP26B1 and CYP26C1…………………………….30 2.8 Thesis Objectives…………………………………………………………...32 2.8.1 Cloning and characterization of murine Cyp26b1 (Chapter 3)…….…..32 2.8.2 Generation of CYP26B1 null animals (Chapters 4 and 5)……………..33

Chapter 3: Cloning of a Novel Retinoic-Acid Metabolizing Cytochrome P450, CYP26B1, and Comparative Expression Analysis with CYP26A1 during Early Murine Development ...... 35 3.1 Summary…………………………………………………………………….36 3.2 Introduction…………………………………………..……………………..37 3.3 Materials and Methods……………………………………………………..39 3.3.1 RNA extraction, RT-PCR and Northern blotting………………………39 3.3.2 Whole mount in situ hybridization……………………………………..39 3.4 Results…………………………………………………………………….....40 3.4.1 Identification of mouse Cyp26b1……………………………………….40 3.4.2 Expression at E7.0-E7.5………………………………………………..40 3.4.3 Expression at E8.0-E8.5……………………………...……………… 45 3.4.5 Expression at E9.5-E10.5………………………………………………50 3.4.6 Expression at E11.5…………………………………………………….56 3.5 Discussion………………………………………………………………...…61

Chapter 4: CYP26B1, a Retinoic Acid Catabolizing Enzyme is Dispensable for

Hindbrain Patterning, but Essential for Craniofacial Development...... 63

4.1 Summary……………………………………………………………………64

4.2 Introduction………………………………..…………………………….....65

4.3 Materials and Methods………………………………………………….....69

4.3.1 Construction of the targeting vector and generation of Cyp26b1

deficient mice………………………………………………………………...69

4.3.2 Embryo Collection……………………………………………………..70

4.3.3 Alizarin Red and Alcian Blue staining…………………………………71

4.3.4 Whole-mount in situ hybridization…………………………………...... 71

vii 4.4 Results………………………………………………………………….....72

4.4.1 Generation of a Cyp26b1 targeting construct………………………..72

4.4.2 Cyp26b1-/- embryos die perinatally and exhibit numerous

malformations………………………………………..…………………….78

4.4.3 CYP26B1 is required for normal craniofacial ossification………….81

4.4.4 Hindbrain patterning is unaffected in CYP26B1 null embryos……..88

4.4.5 Subtle abnormalities in branchial arch patterning are observed

in Cyp26b1-/- embryos……………………………………………………..91

4.4.6 Neural crest cell migration in Cyp26b1 null embryos……………….97

4.4.7 Cyp26a1 and Cyp26c1 expression in the absence of CYP26B1……100

4.5 Discussion…………………………………………………………………103

4.5.1 Cyp26b1 is dispensable for molecular patterning of the hindbrain….103

4.5.2 Cyp26b1 and neural crest cell migration…………………………….106

4.5.3 Branchial arch patterning in the absence of CYP26B1……………107

4.5.4 Craniofacial malformations in Cyp26b1-/- embryos are due to

disruptions in multiple signaling pathways…………………………………109

Chapter 5 The Retinoic Acid Metabolizing Enzyme CYP26B1 is Required for -Catenin Mediated Induction of Embryonic Hair Placodes………………………111 5.1 Summary…………………………………………………………………112 5.2 Introduction……………………………………………………………...... 113 5.3 Materials and Methods……………………………………………………117 5.3.1 Collection of Cyp26b1-/- embryos…………………………………...... 117 5.3.2 Embryo fixation and histology………………………………………117 5.3.3 In situ hybridization…………………………………………………..118 5.3.4 Skin Permeability Assay……………………………………………...118 viii 5.4 Results……………………………………………………………………...119 5.4.1 Epidermal abnormalities in Cyp26b1-/- embryos…………………...... 119 5.4.2 CYP26B1 is required for proper muscle development……………….123 5.4.3 Lack of hair follicles in Cyp26b1-/- embryos………………….………126 5.4.4 Molecular defects in vibrissal follicles of Cyp26b1-/- mice………...... 129 5.5 Discussion………………………………………………………………….132 5.5.1 Cyp26b1 is required for induction of hair follicles……………………132 5.5.2 CYP26B1 is required for -catenin mediated initiation of vibrissal placode formation………………………………………………….134 5.5.3 Epidermal abnormalities in the absence of CYP26B1……………...... 139

Chapter 6-General Discussion………………………………………………………..141 6.1 Expression of Cyp26 genes defines RA distribution in vivo…………….144 6.2 RA-catabolizing activity is essential for embryonic development…….. 145 6.3 CYP26-mediated catabolism of RA is required for normal cell fate...... 147 6.4 Future Research Directions…………………………………………...... 149 6.4.1 Generation of compound Cyp26 mutants…………………………...... 150 6.4.2 CYP26B1 in adult brain…………………………………………….....152 6.5 Concluding Remarks……………………………………………………...153

References……………………………………………………………………………...155 Appendix-Apoptotic Extinction of Germ Cells in Testes of Cyp26b1 Knockout Mice…………………………………………………………………………173 Curriculum Vitae…………………………………………………………………….....212

ix List of Figures Chapter 2 Figure 2.1 Metabolic Pathway of Retinoic Acid..………………………..12

Chapter 3 Figure 3.1 Amino acid alignment of CYP26 members……………………42 Figure 3.2 Size of Cyp26b1 mRNA transcripts…………………………... 44 Figure 3.3 Cyp26a1 and Cyp26b1 expression in post-implantation mouse embryos………………………………………………………………….…47 Figure 3.4 Details of Cyp26a1 expression at E7.5………………………...49 Figure 3.5 Cyp26a1 and Cyp26b1 expression in E9.5 and E10.5 embryos………………………………………………………………….....53 Figure 3.6 Details of Cyp26b1 rhombomeric expression in dissected hindbrains…………………………………………………………………..55 Figure 3.7 Dynamic Cyp26b1 expression during early limb bud development………………………………….…………………………….58 Figure 3.8 Cyp26a1 and Cyp26b1 expression in E11.5 embryos…………60

Chapter 4 Figure 4.1 Homologous recombination is used to integrate LoxP sites into the Cyp26b1 locus……………………………………………………..75 Figure 4.2 Cre-mediated excision of Cyp26b1…………………………….77 Figure 4.3 Gross malformations in Cyp26b1 null embryos………….….…80 Figure 4.4 Severe craniofacial abnormalities in Cyp26b1 null mice………85 Figure 4.5 Malformations of Meckel‟s cartilage and the mandible in Cyp26b1-/- embryos…………………………………………………………87 Figure 4.6 Cyp26b1 is largely dispensable for hindbrain patterning….……90 Figure 4.7 Subtle changes in branchial arch markers in Cyp26b1 null embryos………………………………………………..…………………….94 Figure 4.8 Dlx expression is unaffected in Cyp26b1-/- embryos……………96

Figure 4.9 Differences in neural crest cell patterning in Cyp26b1 null mutant embryos……………………………………………………………99 Figure 4.10 Cyp26a1 and Cyp26c1 are not upregulated in the hindbrain or neural crest cells of Cyp26b1-/- embryos……………………………….102

Chapter 5 Figure 5.1 Epidermal and stromal abnormalities in Cyp26b1-/- embryos…………………………………………………………………...122 Figure 5.2 Cyp26b1-/- embryos have deficiencies in muscle formation.....125 Figure 5.3 Lack of vibrissal and pelage follicle formation in Cyp26b1-/- embryos…………………………………………………………………...128 Figure 5.4 Molecular abnormalities in follicles of Cyp26b1-/- embryos…………………………………………………………………...131 Figure 5.5 Molecular pathways in vibrissal hair placode formation……..138

List of Abbreviations

4-OH-RA…………………………….4-hydroxy-retinoic acid ADH…………………………………Alcohol Dehydrogenase AER………………………………….Apical Ectodermal Ridge ALDH………………………………..Aldehyde Dehydrogenase BA……………………………………Branchial Arch Bmp…………………………………..Bone morphogenetic protein CRABP……………………………….Cellular retinoic acid binding protein E………………………………………Embryonic day Fgf…………………………………….Fibroblast growth factor HAT…………………………………..Histone Acetylase IHC…………………………………....Immunohistochemistry NCC…………………………………..Neural Crest Cell Pr……………………………………...Presumptive rhombomere R………………………………………Rhombomere RA…………………………………….Retinoic acid RAR…………………………………..Retinoic acid receptor RARE…………………………………Retinoic acid response element RBP…………………………………...Retinol binding protein RXR………………………………...... Retinoid X receptor Shh……………………………………Sonic hedgehog VAD………………………………….Vitamin A deficiency

xii

Chapter 1

General Introduction

How cells self-organize into functional structures is one of the most intriguing questions in biology with implications for the future of stem cell based medicine. Tissue formation, or morphogenesis is the result of coordinated proliferation, differentiation and apoptosis of cell populations. During embryonic development, fields of cells respond to localized molecular signals that dictate cell fate, and thus shape early tissues both structurally and functionally (Lander 2007). These signals and their respective transducing pathways must be tightly regulated to allow for control of cellular processes.

Understanding mechanisms underlying embryonic pattern formation have significance therapeutically for diseases such as cancer, where deregulation of signaling pathways results in perturbations in the control of cell fate, and may lead to new approaches to facilitate wound healing and tissue regeneration. The studies described in this thesis focus on retinoic acid (RA), one of the key early signaling molecules involved in embryonic patterning (Pennimpede et al. 2006). How RA is generated and distributed in embryonic tissues plays a significant role in determining the structure and function of essentially all tissues.

The signals that govern morphogenesis are tissue-specific, but there are often common elements that govern development of very diverse structures. For example, the vitamin A metabolite, RA has been identified as a regulator of cell proliferation, differentiation and apoptosis; and thus is critical for tissue morphogenesis and homeostasis (Ross et al. 2000; Pennimpede et al. 2006). RA is essential for embryonic patterning of structures including the limb, brain, neural tube and skin; hence

2 understanding mechanisms regulating RA metabolism and function are widely applicable.

RA modulates cellular processes through transcriptional regulation of RA- responsive genes (Chambon 1996). Essentially all cells have the potential to respond to

RA, but RA-mediated signaling only occurs in a specific population of cells at any given time. The question of how the RA signal is regulated spatially and temporally is fundamental to the understanding of embryonic development and disease progression.

Retinoid metabolism is regulated at multiple steps between vitamin A ingestion and gene signaling, as will be discussed in Chapter 2. It is my belief that the key to controlling RA signaling during embryonic patterning is through the generation of a finely tuned spatiotemporal distribution of RA. The tissue distribution of RA is largely controlled through coordinated expression of genes encoding RA-synthetic

(retinaldehyde dehydrogenase-ALDH) and RA-catabolic (CYP26) proteins, thus defining regions where RA will be available for signaling (Reijntjes et al. 2005; Pennimpede et al.

2006). In this thesis, I describe the identification and characterization of the mouse homolog of one of these genes, Cyp26b1. Furthermore, I present a murine genetic model lacking Cyp26b1 and demonstrate the importance of this enzyme for embryonic development, and discuss the role of CYP26B1 in regulating morphogen signaling.

To assist the reader, the literature review chapter provides background information on i) early nutritional studies of vitamin A and RA, ii) a brief overview of dietary vitamin A metabolism, iii) transcriptional regulation by RA and its receptors, the nuclear retinoic acid receptors, iv) different mechanisms to regulate RA signaling, v) the

3 distribution of RA in developing embryos, vi) regulation of RA synthesis and vii) regulation of RA catabolism and the significance of this activity in establishing and maintaining morphogenetic fields.

Chapter 2

Literature Review

Vitamin A and its biological metabolites (retinoids) have numerous physiological functions, and are required for processes including vision, embryonic development and tissue homeostasis. There is no de novo synthesis of retinoids, they must be obtained from dietary sources. After ingestion, retinoids must be efficiently absorbed and stored until needed. When required, retinoids are mobilized, and converted to biologically active forms, which are utilized by the cell (Blomhoff and Blomhoff 2006). Finally, when retinoids are no longer required in a given tissue, they are catabolized and then excreted. Each of these steps in retinoid metabolism is tightly regulated to ensure a timely and proper dispersion of RA. In this chapter, I will discuss in detail each of these steps in retinoid metabolism, and the significance of regulation at each stage for embryonic patterning.

2.1 Historical Perspective

Decades of important nutritional studies have revealed the significance of vitamin A for biology (McCollum and Davis 1913; Mason 1935; Wilson et al. 1953; Thompson et al. 1964). These early studies were important for defining the role of vitamin A, and for making the observation that retinoids were required for i) vision and ii) embryonic development; and there are mechanistic differences in how retinoids are utilized in these processes.

One of the first processes where the role of vitamin A was appreciated was the visual cycle. Night blindness is a symptom of vitamin A deficiency, and for centuries had been treated with topical administration of liver extracts (Wolf 1978). In the early

1900s, vitamin A was identified as a fat-soluble compound in liver that was likely responsible for the effectiveness of these treatments (McCollum and Davis 1913).

Vitamin A (also known as retinol), has since been shown to be required for vision as retinoids such as 11-cis-retinal serve as the chromophore of various visual pigments

(Dowling and Wald 1960).

Unrelated to the role of retinol in vision; other biological functions were discovered as it was observed that pregnant rodents maintained in a vitamin A deficient (VAD) state exhibited elevated rates of reproductive failure and abnormalities in epithelial tissues

(Mason 1935; Wilson et al. 1953). It was demonstrated that maternal VAD resulted in abnormal morphogenesis of numerous fetal tissues including the eyes, genito-urinary tract, lungs, heart and limb; malformations that could be largely prevented through the administration of retinol depending on the time and dose of treatment (Wilson et al. 1953;

Thompson et al. 1964). It was also observed that vitamin A acid, known as retinoic acid

(RA) was sufficient to reverse some phenotypes of VAD in adult rodents, and was able to support embryonic development in VAD rats (Dowling and Wald 1960; Thompson et al.

1964; Takahashi et al. 1975). These early experiments established that i) there are developmental time points when the embryo must have minimum levels of retinol to ensure normal morphogenesis, ii) most tissues require vitamin A, and iii) RA was able to rescue some of the abnormalities associated with VAD, and thus may be the active form of vitamin A. Conversely, it has also been demonstrated that vitamin A, and more specifically RA can act as a teratogen, as maternal exposure to exogenous RA has been shown to result in a spectrum of birth defects in numerous animal models as well as

7 humans (Kalter and Warkany 1961; Kochhar 1973; Lammer et al. 1985; Collins and Mao

1999; McCaffery et al. 2003). RA induced abnormalities are observed in several tissues including the limb, neural tube, heart, gonad and craniofacial structures (Kochhar 1973;

Taylor et al. 1980; Collins and Mao 1999; McCaffery et al. 2003). Similar to the rescue of phenotypes in VAD models, tissues affected by exogenous RA depend on the time and level of RA exposure. These early nutritional studies provided evidence that retinoid status had to be tightly controlled at specific times and in certain tissues to allow for normal development.

2.2 Overview of Vitamin A Metabolism

Nutritional studies showed that developing tissues are sensitive to maternal retinoid status, thus indicating that there is likely tight regulation of endogenous retinoid metabolism to ensure proper distribution, particularly during development. Vertebrates are unable to synthesize vitamin A de novo, hence it must be obtained from dietary sources, most commonly derived from plant carotenoids, or as retinyl esters from animal sources or in fortified milk and grain products (Blomhoff and Blomhoff 2006). For efficient absorption and storage, ingested carotenoids such as -carotene, -carotene and

-cryptoxanthin are converted to retinal or other retinoids. For example, -carotene in the small intestine is cleaved by , -carotene-15,15‟-monooxygenase to yield two retinal molecules which are then reduced to form retinol to facilitate absorption by enterocytes

(Goodman et al. 1965; Olson and Hayaishi 1965). After absorption, retinol is re- esterified and packaged into chylomicrons, which are secreted into the lymphatic system, where the majority of retinyl esters are transported to the liver (Ross and Zolfaghari

2004; Blomhoff and Blomhoff 2006). In the liver, retinyl esters are either hydrolysed for immediate transport, or stored in stellate cells, which ensures that adequate retinoids will be available for mobilization when dietary retinoid content is low.

When required in other tissues, retinyl esters are hydrolysed, liberating retinol, which forms a complex with a specific retinol-binding protein (RBP) and transthyretin.

The RBP:retinol complex is secreted into plasma and travels in the bloodstream to target tissues where it has been suggested that a cell surface receptor mediates uptake.

Alternatively, retinyl esters may be delivered by the circulatory system to tissues by 9 lipoproteins such as very low-density lipoprotein, and low-density lipoprotein (Blomhoff and Blomhoff 2006).

Upon entry into the cell, retinol must be converted to the more active compound,

RA. Initially, retinol is converted to the aldehyde retinal, via a reversible oxidation reaction (Figure 2.1). Several enzymes have been identified that are capable of catalyzing this reaction, including cytosolic medium-chain alcohol dehydrogenases

(ADH) and membrane-bound short-chain dehydrogenase/reductase (SDR) enzymes

(Duester et al. 2003). Subsequently, retinal acts as a substrate for retinaldehyde dehydrogenases (ALDH) that catalyze the irreversible formation of RA (Niederreither and Dolle 2006). RA is the most biologically active form of RA; this metabolite is able to modulate gene expressing by binding to nuclear receptor proteins, as will be discussed.

Within cells, RA can be bound by the cellular retinoic acid binding proteins

(CRABP), CRABPI and CRABPII, although the biological function of these proteins is not well defined (Wei et al. 1987; Giguere et al. 1990). RA is further oxidized by the

Cyp26 proteins (CYP26A1, CYP26B1 and CYP26C1), which are cytochrome P450 enzymes that catalyze the conversion of RA to more polar metabolites including 4-oxo-

RA, 4-OH-RA and 18-OH-RA (White et al. 1996; White et al. 2000a; Tahayato et al.

2003). These oxidation reactions are thought to inactivate RA, thus limiting expression of RA-responsive genes in tissues where the CYP26 proteins are expressed (see below).

The oxidation of RA is also believed to enhance its aqueous solubility thereby facilitating its elimination via excretion.

Figure 2.1 Metabolic Pathway of Retinoic Acid. Vitamin A (Retinol) is converted to retinal via a reversible oxidation reaction catalyzed by alcohol dehydrogenases. Retinal in turn is converted to retinoic acid in a reaction facilitated by retinaldehyde dehydrogenases. Retinoic acid is the biologically active form of vitamin A, and is converted to 4-OH-retinoic acid, 18-OH-retinoic acid and 4-oxo-retinoic acid by the

CYP26 to facilitate excretion.

2.3 Retinoic Acid is a Transcriptional Regulator

2.3.1 Nuclear retinoic acid receptors Although the biological effects of RA had been appreciated for decades, questions remained as to how RA exerted its effects. Particularly during embryonic development, it was unknown how RA modulated cellular processes. However, in 1987, a nuclear receptor, the retinoic acid receptor (RAR ), was cloned for which RA acted as a ligand

(Giguere et al. 1987; Petkovich et al. 1987). This seminal discovery defined a novel family of nuclear receptors; and the identification of RA as a transcriptional regulator provided a mechanistic basis for the observation that RA could regulate cell proliferation and differentiation. Subsequently, other RARs were identified, and the family is now comprised of three separate genes (RAR , RAR , RAR ), each with several isoforms as a result of differential splicing and the use of alternate promoters (Brand et al. 1988;

Krust et al. 1989; Chambon 1996; Germain et al. 2006). RARs form functional heterodimers with another class of nuclear receptors, the retinoid X receptors (RXR).

RXRs are also functional binding partners for other receptor families including the vitamin D receptor, thyroid receptor, peroxisome proliferator activated receptor and liver-

X receptor among others. RAR/RXR heterodimers bind to specific DNA sequences termed retinoic acid response elements (RAREs). RAREs typically consist of a direct hexamer repeat motif, PuGG/T TCA, separated by 5 base pairs (DR5), although RAREs can also contain either 1 or 2 base pair (DR1, DR2) spacers (Chambon 1996; Germain et al. 2006).

According to a model of transcriptional regulation proposed by Dilworth and

Chambon (Dilworth and Chambon 2001), in the absence of ligand, RAR and RXR bind to RAREs and recruit the transcriptional repressor proteins, nuclear receptor corepressor

(NCoR) and silencing mediator of retinoid and thyroid hormone receptors (SMRT) (Chen and Evans 1995; Zamir et al. 1996). NCoR and SMRT in turn recruit a protein corepressor complex with histone deacetylase activity (HDAC) that facilitates chromatin compaction and gene silencing (Alland et al. 1997; Heinzel et al. 1997; Nagy et al. 1997).

Upon ligand binding, the receptors undergo a conformational change that increases

RAR/RXR affinity for DNA, and also results in a shift in the receptors that shields repressor-binding sites and exposes activator-binding sites within the ligand-binding motif (Valcarcel et al. 1994). Several activators such as p300/CBP and members of the

SRC/p160 family including have been shown to interact with RAR/RXR (Wei 2003).

CBP and p/CIP exhibit histone acetylase activity (HAT) that acetylates specific lysine residues on histones that changes local chromatin conformation (Torchia et al. 1997;

Schiltz et al. 1999). Furthermore, other activator protein complexes with HAT activity or the SWI/SNF chromatin remodeling complex are recruited to further reduce chromatin compaction thus facilitating gene expression (Fukuda et al. 2006; Flajollet et al. 2007).

The identification of the RARs, and elucidation of their role in regulating gene expression provided an explanation for RA-mediated effects at the cellular level. Thus far, several hundred genes have been identified as being potentially transcriptionally regulated by RA, either directly (containing a RARE) or indirectly (downstream target of a directly regulated gene) (Balmer and Blomhoff 2002). These genes are involved in

14 numerous pathways that affect many biological processes including regulation of cell growth, differentiation and apoptosis, and are thus key components of gene regulatory cascades leading to morphogenetic events.

2.3.2 Ligands for retinoic acid receptors

There are multiple metabolites of retinol in vivo, and identification of the endogenous ligands for the RARs was a critical step in understanding retinoid-regulated gene expression. All-trans-RA (at-RA) has been identified as the physiological ligand for the RARs, although the receptors can also bind the at-RA isomer 9-cis-RA with high affinity in vitro (Allenby et al. 1993; Allenby et al. 1994). It was initially believed that 9- cis-RA was the natural ligand for RXR as ligand trapping studies showed that 9-cis-RA, but not at-RA bound RXR in vitro (Heyman et al. 1992; Levin et al. 1992). Several groups have corroborated these results, but there is still a lack of convincing evidence that 9-cis-RA is the physiological ligand for RXR. Several studies have been unable to directly detect 9-cis-RA in vivo using highly sensitive chromatographic techniques, thus raising the possibility that RXR either remains unliganded, or has other ligands in vivo

(Ulven et al. 2001; Schmidt et al. 2003; Kane et al. 2005). It has been proposed that fatty acids such as linoleic, linolenic and docosahexaenoic acid may be active ligands, as they can be readily detected in tissues, and are able to bind RXR in vitro, albeit with a lower affinity than 9-cis-RA (de Urquiza et al. 2000; Goldstein et al. 2003). Recent genetic studies have shown that 9-cis-RA is not the ligand for RXR in developing mouse skin,

15 although no conclusions regarding signaling in other tissues can be drawn (Calleja et al.

2006).

Several other retinoids including retinol, 4-oxo-retinol, 4-OH-retinol, 4-oxo-RA and 4-OH-RA have been shown to bind RAR in vitro, but it is controversial whether or not these metabolites have any endogenous biological function (Repa et al. 1993; Achkar et al. 1996). Genetic studies have been conducted that provide compelling evidence that at-RA is the most biologically active retinoid, and that these other compounds have a limited role in RA-mediated transcriptional regulation as will be discussed later

(Niederreither et al. 2002a).

2.4 Regulation of Retinoic Acid Signaling

2.4.1 Retinoic acid receptor expression and function

The vast number of retinoid responsive genes involved in many diverse pathways indicates that RA-signaling must be tightly regulated to ensure that signaling cascades are only initiated in appropriate tissues at the correct development stage. One obvious means of controlling RA-signaling would be through regulated expression of the transducers of the signal, the RARs. RARs exhibit gene and isoform specific expression patterns as will be briefly described. Throughout mouse development, RAR is ubiquitous with slightly lower expression levels in some regions of the central nervous system (Dolle et al. 1990;

Ruberte et al. 1993). In contrast, RAR is expressed in specific regions, including craniofacial structures and epithelia of tissues such as the trachea, stomach and digestive tract (Mollard et al. 2000). RAR is also present in the proximal mesenchyme, and later in development is detected in interdigital regions of the limb as well as neural structures

(Dolle et al. 1989; Dolle et al. 1993; Ruberte et al. 1993). From embryonic day (E) 8.5-

10.5, RAR is detected in the tail bud, frontonasal mass, and mesenchyme of the branchial arches and limb buds (Ruberte et al. 1990; Mollard et al. 2000). At E13.5, RAR is expressed in pre-cartilagenous elements, and this expression decreases as ossification is initiated (Ruberte et al. 1990).

Considering that RARs are the mediators of RA-signaling, tissue specific receptor expression could potentially be an effective means of regulating morphogen signaling. Interestingly, gene targeting studies in mice have shown that there is

17 significant functional redundancy amongst the receptors. Generally, when the dominant isoform of any given RAR is deleted, homozygous mice are viable, fertile and morphologically similar to wild-type littermates. However, when all the isoforms of a receptor are targeted, mice develop normally, but show some postnatal effects reminiscent of VAD. For example, RAR 1 null animals show no morphological abnormalities, while mice lacking the entire RAR gene, ablating isoforms 1 and 2, show increased post-natal lethality and the majority of mutant homozygotes die within 1-

2 months (Li et al. 1993; Lufkin et al. 1993). In addition to high post-natal lethality,

RAR null mutants display testicular abnormalities, and some minor skeletal malformations with low penetrance (Lufkin et al. 1993; Ghyselinck et al. 1997).

Similarly, RAR 2-/- mice are phenotypically normal, while RAR 2 null mice show elevated post-natal mortality. RAR 1/ 2 null animals exhibit numerous abnormalities in the axial skeleton, which are generally minor vertebral transformations that are variable, but highly penetrant (Lohnes et al. 1993; Ghyselinck et al. 1997). RAR null mice are the least affected of any of the receptor knockout animals, as RAR 2-/- mutants are phenotypically normal, and RAR 1-4 null mutants display only homeotic vertebral transformations and abnormalities in the retrolenticular membrane (Mendelsohn et al.

1994b; Ghyselinck et al. 1997).

In contrast to these genetic models, when knockout mice were developed that lacked at least two of the receptors, observed phenotypes were much more severe

(Lohnes et al. 1994; Mendelsohn et al. 1994a; Kastner et al. 1995; Lohnes et al. 1995).

Mutants often exhibited embryonic lethality, and almost all of the characteristics of VAD 18 were recapitulated by various combinations of compound mutant mice. These compound mutant animals provided strong evidence that i) the RARs are responsible for mediating

RA signaling and ii) there is significant functional redundancy amongst members of this nuclear receptor family.

In terms of pattern formation, it is obvious that removal of multiple RAR genes has severe consequences for fetal development. However, any single RAR is largely dispensable for murine embryonic development, as even when there is high post-natal lethality, embryos develop to E18.5 in an expected Mendelian ratio. This indicates that

RA signaling during embryonic patterning is not tightly regulated at the level of receptor expression, as multiple genes must be disrupted to have a deleterious effect on the embryo.

2.4.2 Retinoid binding proteins

As previously mentioned, there are several proteins that participate in the absorption, storage and mobilization of vitamin A during metabolism. Early studies postulated that these proteins played critical roles in regulating the distribution and activity of RA. For example, RBP binds retinol and transports it to target tissues via the bloodstream. RBP is unable to cross the placenta, but is expressed in the embryonic yolk sac indicating that it may be important for retinol transfer to the embryo (Sapin et al.

1997; Quadro et al. 2004b). However, RBP-/- mice are viable and fertile with reduced levels of circulating retinol, but have offspring with no apparent abnormalities (Quadro et

19 al. 1999; Quadro et al. 2004a). Therefore, when mice are maintained on a vitamin A sufficient diet, RBP is dispensable for embryonic development, indicating that there may be alternative means of retinol delivery from mother to fetus.

CRABPI and CRABPII are RA binding proteins that are expressed in both embryonic and adult tissues, and are largely conserved across species. Thus, it was speculated that these proteins were important for either sequestering RA in the cell, or aiding in cellular RA metabolism or delivery to receptors. However, mice lacking either, or both of CRABPI and CRABPII exhibit no discernable phenotypes except for an extra forelimb digit in CRABPII null mice (Fawcett et al. 1995; Lampron et al. 1995; Schmidt et al. 2003). This indicates that with a diet sufficient in vitamin A, the CRABPs are not essential for embryogenesis. Although RBP, CRABPI and CRABPII are dispensable for development under laboratory conditions, these proteins may play significant roles in specific physiological states, or dietary conditions when vitamin A is not abundant.

However, it is clear that these proteins do not have a prominent role in regulating RA distribution or signaling during embryonic patterning.

2.5 Retinoic Acid Localization

It has been demonstrated by different means that RA is not ubiquitous in the embryo; there are RA-rich and RA-poor tissues (Mendelsohn et al. 1991; Rossant et al.

1991; Balkan et al. 1992; Colbert et al. 1993). Even within tissues there is often a localized distribution of RA. For example, the anterior hindbrain is devoid of RA in the mouse, chick and frog, while all three species have higher RA levels in the posterior hindbrain and anterior somites (Chen et al. 1994; Maden et al. 1998; Sirbu et al. 2005).

A line of transgenic mice (RARE-LacZ) containing 3 copies of a RARE upstream of the promoter of the hsp68 (heat-shock protein) gene fused to a LacZ reporter gene are a widely used model to visualize RA-mediated signaling in vivo (Rossant et al. 1991). RA distribution is dynamic and changes dramatically as the embryo develops, and tissues undergo morphogenesis. In E8.5 mouse embryos, RA is detected in the trunk of the embryo, and is absent from the tailbud and anterior embryo, with the exception of a region surrounding the presumptive eye. As the limb buds develop and initiate outgrowth, they appear to lack RA, until mesenchymal digit condensations are visible

(Rossant et al. 1991). This distribution is largely maintained until E11.5; in older embryos, the pattern of RA-signaling becomes more complex, as some tissues become divided into regions that are either RA-rich, or devoid of RA. Many of these tissues have been shown to be sensitive to perturbations in endogenous retinoid status, therefore it is believed that the spatio-temporal distribution of RA in these regions is critical for morphogenesis.

As has been discussed, the RARs, RBP and CRABPs do not appear to play major roles in the spatial regulation of RA signaling when retinol is sufficient in the diet.

However, given that there is a definite distribution of RA, it is possible that the coordinated expressions of RA-synthetic and RA-catabolic enzymes are critical in defining where RA is present, and consequentially where and when RA-mediated gene expression occurs.

2.6 Regulation of Retinoic Acid Synthesis

2.6.1 Regulation of RA Synthesis by Alcohol Dehydrogenases

The RA precursor retinol, is the most abundant retinoid in plasma, and can be reversibly oxidized to the aldehyde retinal. This reaction is catalyzed in vitro by several enzyme families including alcohol dehydrogenases (ADH) and short-chain dehydrogenase/reductases (SDR) (Duester et al. 2003). Retinal then undergoes another oxidation reaction to form RA, as catalyzed by retinaldehyde dehydrogenases (ALDH- see below). ADH1, ADH3 and ADH4 have been identified as alcohol dehydrogenases with high specificity for retinol in mice (Holmes 1977; Zgombic-Knight et al. 1995).

However, with the exception of ADH3, (which when knocked out results in slight growth retardation), gene targeting studies have shown that ablation of any of these genes has no effect on growth or survival when mice are maintained on a retinol-sufficient diet

(Deltour et al. 1999b; Deltour et al. 1999a; Molotkov et al. 2002b). ADH3 and ADH4 null mutant mice show significant increases in mortality when maintained on a VAD diet

22 indicating that these proteins may be important in regulating retinol oxidation when it is limiting (Molotkov et al. 2002a; Molotkov et al. 2002b). ADH1 null mice show increased sensitivity to vitamin A toxicity, thus ADH1 may play a role in oxidizing exogenous retinol to retinal to facilitate excretion (Molotkov et al. 2002a). The lack of embryonic patterning phenotypes in these mouse models when retinol was sufficient indicated that retinol in plasma may be available to all tissues, and the expression of

ALDH enzymes is the main regulatory step in determining the spatio-temporal distribution of RA (Duester et al. 2003).

However, a recent mouse model, trex, describes mice with a mutation in the SDR family member RDH10 that significantly reduces protein stability. RDH10 is expressed in somites, lateral plate mesoderm, otic vesicles, nasal pit and eye cup; regions that all have comparatively elevated levels of RA (Sandell et al. 2007). Homozygous trex mice die in utero by E13.0 and exhibit severe forelimb and craniofacial abnormalities, lack corneas and the ventral retina, and exhibit hypoplastic organ development (Sandell et al.

2007). Trex/trex embryos also have dramatically reduced levels of RA in the nasal process and trunk indicating that RDH10 activity is required for fetal RA synthesis in these tissues, and this activity not compensated for by other ADH or RDH enzymes. This mouse model is the first example of a retinol oxidizing enzyme that is essential for establishing and maintaining the spatio-temporal distribution of RA in the embryo.

2.6.2 Regulation of RA Synthesis by Retinaldehyde Dehydrogenases

Similar to RDH10, the retinaldehyde dehydrogenase (ALDH) enzymes have been shown to be important for embryonic patterning. ALDHs, which are also referred to as

RALDHs, are cytosolic aldehyde dehydrogenases with high catalytic specificity for retinal. Thus far, four ALDH (1-4) proteins have been identified; ALDH 1-3 catalyze the irreversible oxidation all-trans-retinal to form at-RA, while ALDH4 can generate 9-cis-

RA in vitro (Rongnoparut and Weaver 1991; Zhao et al. 1996; Li et al. 2000; Mic et al.

2000; Lin et al. 2003). Aldh1a1-3 exhibit dynamic expression patterns during mouse embryogenesis that overlap with regions that are rich in RA. Aldh1a2 is the most widely expressed ALDH in the mouse embryo, with expression beginning at E7.0 in mesoderm lateral to the primitive streak (Niederreither et al. 1997). During early to mid-gestation

(E8.5-E11.5), Aldh1a2 is detected in somites, mesoderm of the trunk and regions of the eye. As the limb buds develop, Aldh1a2 is expressed in interdigital regions

(Niederreither et al. 1997). Later in development, Aldh1a2 is restricted to several regions including the nasal cavities, tooth buds, eye, metanephros, mesenephros, and mesenchymal layers of the esophagus, stomach and intestine (Niederreither et al. 2002b).

In contrast, Aldh1a1 expression is much more restricted beginning at E9.5 in the dorsal retina and ventral mesencephalon (Haselbeck et al. 1999). During organogenesis,

Aldh1a1 detected in the epithelium of the respiratory tract and trachea, the olfactory mesenchyme and in regions of the lung, liver, intestine and kidney (Niederreither et al.

2002b). Aldh1a3 expression begins at E8.5 in the dorsal retina and olfactory epithelium, persists in the nasal process throughout much of development, and transcripts are also

24 detected in the kidney and seminal vesicles (Li et al. 2000; Mic et al. 2000; Niederreither et al. 2002b). Aldh1a4 is absent in the embryo, and is restricted to the kidney and liver in adult mice (Lin et al. 2003; Niederreither and Dolle 2006). Embryonic tissues that express the various Aldh genes generally correspond with regions that have been shown to contain RA. Moreover, many of these tissues are sensitive to the effects of vitamin A deprivation as shown by nutritional studies. However, the strongest evidence that the

ALDH proteins are responsible for participating in the regulation of the physiological distribution of RA come from studies with transgenic mice lacking one or more of the retinaldehyde dehydrogenases.

Aldh1a2 null mice die in utero at approximately E9.5 with severe malformations, as embryos do not undergo axial rotation, lack limb buds and posterior branchial arches, have a shortened frontonasal region and a truncated posterior axis (Niederreither et al.

1999). These malformations are caused by a deficiency of RA as reduced reporter activation is detected in Aldh1a2-/- embryos generated on a background of RARE-LacZ mice (Niederreither et al. 1999). Maternal supplementation with RA is able to partially rescue the Aldh1a2-/- phenotype allowing embryos to develop further in gestation, and this model has been used to show the importance of Aldh1a2 in the morphogenesis of structures including the hindbrain, heart and limb (Niederreither et al. 1999; Niederreither et al. 2000; Niederreither et al. 2001; Niederreither et al. 2002d).

In contrast to the striking phenotype of Aldh1a2 null mice, Aldh1a1 and

Aldh1a3-/- mice exhibit much subtler abnormalities. Aldh1a1 null mutants are viable and fertile with no discernable developmental malformations (Fan et al. 2003). Aldh1a3-/-

25 mice die within hours of birth due to respiratory distress as a result of choanal atresia

(Dupe et al. 2003). Other than nasal abnormalities and a shortened ventral retina,

Aldh1a3-/- mice appear morphologically normal. Taken together, these genetic models show that ALDH1A2 has an essential role in regulating embryonic RA distribution, and consequently patterning, while ALDH1 is dispensable and ALDH3 is only required for nasal morphogenesis.

2.7 Regulation of Retinoic Acid Catabolism

2.7.1 Hydroxylation of retinoic acid

RA is a small, lipophilic molecule that is diffusible in vivo. Conceptually, a biological system that has regulated expression of retinaldehyde dehydrogenases can define where RA will be present. However, regulated synthesis alone would be insufficient to control the duration that tissues would be exposed to RA, and could not prevent RA from diffusing to tissues that are sensitive to RA. In line with Lewis

Wolpert‟s theoretical work predicting the generation of biological patterns in response to a morphogen gradient (Wolpert 1969), it was postulated that RA distribution was controlled through the coordinated regulation of RA-synthetic and catabolic enzymes.

The identification of polar metabolites of RA including 4-OH-RA and 4-oxo-RA indicated that RA was further oxidized both in vitro and in vivo (Frolik et al. 1979;

Roberts et al. 1979). Treatment of mice with oral doses of at-RA resulted in the accumulation of polar metabolites of RA in feces and urine, providing evidence that

26 oxidation of RA may facilitate its excretion (Kalin et al. 1981). This oxidation was shown to be catalyzed by cytochrome P-450 enzymes as treatment with broad spectrum cytochrome P450 inhibitors such as ketoconazole, clotrimazole, miconazole and metyrapone abrogated the oxidation of RA (Williams and Napoli 1987; Van Wauwe et al.

1988). These early experiments let to the proposal that the oxidation of RA by specific cytochrome P-450 enzymes was part of an inactivation or detoxification pathway for RA.

2.7.2 Identification of CYP26A1 as retinoic acid hydroxylase

In 1996, a cytochrome P450 enzyme capable of converting RA to 4-OH-RA and

4-oxo-RA was identified in the regenerating tail fin of zebrafish (White et al. 1996). This enzyme, P450RAI (later renamed CYP26A1) defined a novel family of cytochrome P450 enzymes capable of specific RA catabolism. Subsequently, orthologs of Cyp26a1 were cloned in human, mouse, chick and frog (Fujii et al. 1997; Ray et al. 1997; White et al.

1997; Hollemann et al. 1998; Swindell et al. 1999). Cyp26a1 is widely expressed in embryonic and adult tissues that are devoid of RA, strengthening the argument that

CYP26A1 degrades RA in vivo. Furthermore, Cyp26a1 is inducible by RA in some tissues and cell lines, which validated earlier studies showing that generation of 4-OH-

RA was inducible by RA (Duell et al. 1996). Two RAREs have been characterized in the promoter region of Cyp26a1, and both elements appear to function synergistically to provide a highly sensitive transcriptional response to low levels of RA (Loudig et al.

2000; Loudig et al. 2005). This indicates that Cyp26a1 expression is part of a positive feedback loop to regulate levels of RA in specific tissues in response to influxes of RA.

Cyp26a1 expression during embryonic development begins at E7.0-7.5 with expression in the extra-embryonic and embryonic endoderm and posterior mesoderm

(Fujii et al. 1997; de Roos et al. 1999; MacLean et al. 2001). During early murine development (E8.0-E11.5), Cyp26a1 is strongly expressed in the tailbud and caudal neuroepithelium, with transient expression in rhombomere 2 (at E8.5), the first branchial arch (E9.0-10) and cervical mesenchyme (E9.0-E10.5) (Fujii et al. 1997; MacLean et al.

2001). Later, during organogenesis, Cyp26a1 is detected in pre-cartilagenous elements, the retina, toothbud, stomach and the diaphragm (Abu-Abed et al. 2001; Abu-Abed et al.

2002). The importance of Cyp26a1 in regulating the distribution of RA in vivo has been demonstrated by the generation of Cyp26a1 null animals which die during gestation exhibiting variable phenotypes including spina bifida, exencephaly, posterior truncation, homeotic transformations of cervical vertebrae and abnormalities in hindbrain patterning; phenotypes that are all reminiscent of RA teratogenesis (Abu-Abed et al. 2001; Sakai et al. 2001). When Cyp26a1-/- mice are bred onto a mouse background containing the

RARE-LacZ transgene, there is increased reporter activation in regions such as the tailbud where Cyp26a1 is normally expressed, indicating that the in vivo function of

CYP26A1 is to define spatio-temporal boundaries for RA in the embryo (Sakai et al.

2001).

2.7.3 Retinoic acid is the biologically active form of vitamin A

CYP26A1 generates 4-OH-RA and 4-oxo-RA, and it has been argued that these oxidized metabolites are physiologically active forms of vitamin A (Pijnappel et al. 1993;

Reijntjes et al. 2007). Xenopus embryos treated with either RA or 4-oxo-RA exhibit a loss of anterior structures, and 4-oxo-RA can also act as a teratogen in mice (Satre et al.

1989; Pijnappel et al. 1993). Furthermore, studies in the vitamin A deficient quail model have shown that 4-OH-RA and 4-oxo-RA are able to rescue embryo development, and it has also been suggested that these metabolites are important for hindbrain development in zebrafish (Reijntjes et al. 2005; Reijntjes et al. 2007). However, genetic studies in mice have provided strong evidence that polar metabolites of RA are not the most biologically active retinoids. When the levels of RA in Cyp26a1-/- embryos are lowered by also knocking out one allele of Aldh1a2, Cyp26a1-/- Aldh1a2+/- embryos develop normally and are viable (Niederreither et al. 2002a). In Cyp26a1-/- mice, there is no compensation by other Cyp26 genes (see below) in the posterior part of the embryo, thus there will be no polar metabolites of RA generated. Therefore, the Cyp26a1-/-

Aldh1a2+/- mutants provide compelling evidence that RA is the active metabolite of vitamin A.

The expression and knockout studies of Cyp26a1 definitively showed that regulation of RA catabolism was critical for morphogenesis. However, these studies also indicated that other RA-catabolizing enzymes may exist and have a role in pattern formation. For example, Cyp26a1 is not expressed in the developing limb bud at E10.5, but there is an absence of RA in this region as shown by RARE-LacZ mice (Rossant et al.

1991). Considering that RA is diffusible, and detected in high levels in mesenchyme adjacent to the limb buds, it appeared that other enzymes must be responsible for preventing RA from entering these structures. Similarly, there is a lack of Cyp26a1 expression in the hindbrain and craniofacial structures; regions that are devoid of RA

(Rossant et al. 1991).

2.7.4 Identification of CYP26B1 and CYP26C1

In 2000, a second Cyp26 enzyme, CYP26B1 was cloned from human and shown to have comparable efficiency and substrate specificity to CYP26A1 in vitro (White et al.

2000a). Subsequently, the mouse ortholog of Cyp26b1 was cloned and characterized, as outlined in Chapter 2 (MacLean et al. 2001). A third Cyp26 member, CYP26C1 has also been identified, although CYP26C1 appears to have slightly different enzymatic properties (Tahayato et al. 2003; Taimi et al. 2004). Analysis of CYP26C1 activity in vitro has shown that in addition to metabolizing at-RA, CYP26C1 hydroxylates 9-cis-RA with a much greater efficiency than either CYP26A1 or CYP26B1 (Taimi et al. 2004).

The Cyp26 genes exhibit largely distinct expression patterns indicating that they may be responsible for protecting different tissues from inappropriate RA signaling. As previously mentioned, murine Cyp26a1 is largely restricted to the tailbud, and cervical mesenchyme (Fujii et al. 1997; MacLean et al. 2001). Cyp26c1 is expressed in the hindbrain between E8.0 and E9.5, and also overlaps in some regions with Cyp26a1 with transcripts detected in the tooth buds and mandibular-maxillary cleft of the first branchial arch (Fujii et al. 1997; MacLean et al. 2001; Tahayato et al. 2003). Cyp26b1 in contrast,

30 is more widely expressed in the embryo than either Cyp26a1 or Cyp26c1, as is described in detail in Chapter 2. Briefly, Cyp26b1 is expressed in the hindbrain from E8.0 to

E10.5, the developing limb buds, the branchial arches and craniofacial structures and various organs in older embryos (MacLean et al. 2001; Abu-Abed et al. 2002). Many of the tissues that express Cyp26 genes have been shown to be sensitive to the effects of

RA; common phenotypes of RA teratogenesis include spina bifida, posterior truncation, limb truncations, cleft palate, hindbrain posteriorization, micrognanthia and other craniofacial abnormalities. Of these phenotypes, spina bifida, posterior truncation and subtle changes in hindbrain patterning are observed in Cyp26a1 null mice; and these tissues all express Cyp26a1 during gastrulation (Abu-Abed et al. 2001; Sakai et al. 2001).

Considering the expression pattern of Cyp26b1, it is reasonable to hypothesize that

Cyp26b1 null animals would manifest some of the other phenotypes of RA teratogenicity.

We and others have generated Cyp26b1-/- animals, and observed numerous developmental abnormalities due to aberrant retinoid signaling (Yashiro et al. 2004;

MacLean et al. 2007). These studies categorically show that CYP26B1 is critical for embryonic development and morphogenesis of tissues including the limb, testes and neural crest derived structures. The molecular mechanisms contributing to these phenotypes have been identified in some detail, but future studies will look at further delineating affected pathways and examining the role of CYP26B1 mediated RA- catabolism in tissue homeostasis of adult mice.

2.8 Thesis Objectives:

The goal of this thesis was to clone and characterize the expression and function of murine Cyp26b1. The following sections will briefly explain the experimental rationale, hypotheses and techniques utilized in successfully accomplishing these goals.

2.8.1 Cloning and characterization of murine Cyp26b1 (Chapter 3)

Hypothesis: Regions of the mouse embryo which appear devoid or RA activity such as limb and hindbrain exhibit limited expression of Cyp26a1, hence other Cyp26 genes complement the activity of CYP26A1 and help to block RA activity in these tissues.

When I first joined the Petkovich lab, human CYP26B1 had been first cloned by

Dr. Jay White (White et al. 2000c). The first goal of my project was to clone the murine ortholog of CYP26B1, which I initially attempted by screening 3 mouse cDNA libraries using both a human cDNA probe, and later a mouse EST that was highly homologous to human CYP26B1. Concurrently, I used the mouse EST to generate an anti-sense RNA probe to characterize the embryonic expression of Cyp26b1 by whole mount in situ hybridization. After observing that Cyp26b1 is expressed in the limb bud beginning at outgrowth, I collected E11.5 limb buds, extracted RNA from them and using primers based on mouse EST and human cDNA sequence amplified a full-length Cyp26b1 transcript. These results were subsequently published and are presented in Chapter 3.

2.8.2 Generation of CYP26B1 null animals (Chapters 4 and 5)

Hypothesis: CYP26B1 may be important for limiting RA exposure in embryonic regions including the limb, hindbrain and branchial arches, and disruption of Cyp26b1 expression result in inappropriate RA exposure in these tissues and hence develop phenotypes reminiscent of RA teratogenicity.

In order to study the function of CYP26B1 in vivo, we proceeded to generate a transgenic mouse line in which homologous recombination was used to eliminate the endogenous Cyp26b1 locus. Many genes involved in RA metabolism had already been deleted in mouse; thus, there was an existing framework into which observations from

Cyp26b1 null animals could be applied. Furthermore, generating transgenic mice would allow us to later make compound knockout mice lacking multiple Cyp26 genes, or other components of RA signaling. Finally, given the embryonic expression of Cyp26b1 and the severe phenotype of Cyp26a1-/- mice, we hypothesized that Cyp26b1 null animals would exhibit an embryonic lethal phenotype (Abu-Abed et al. 2001; MacLean et al.

2001; Sakai et al. 2001). Therefore, I designed a targeting strategy that would allow for the generation of the conditional knockout of Cyp26b1 in specific tissues or time points through the use of the Cre-LoxP system. I designed a construct containing 8.9 kb of the endogenous murine Cyp26b1 locus that encompassed exons 3-6. This region includes the heme-binding site, which is critical for cytochrome P450 activity, so I proceeded to generate a loss-of-function allele. LoxP sites were inserted in intron 2 and after exon 6 so that in the presence of Cre, exons 3-6 would be excised. A neomycin resistance marker

(neo) was added after exon 6 to allow for the selection of transfected embryonic stem

(ES) cells into which the construct had integrated into genomic DNA. The neo was flanked by LoxP sites, so that it could later be removed to prevent any deleterious effects by the neo product. Homologous recombination was confirmed by both southern blot and PCR as summarized in Figures 4.1 and 4.2. Null animals were generated and exhibit severe embryonic malformations. The characterization and analysis of the craniofacial abnormalities in Cyp26b1-/- mice are presented in Chapter 4. Cyp26b1-/- also display abnormalities in skin and hair follicle development, and these results are discussed in

Chapter 5.

Chapter 3

Cloning of a Novel Retinoic-Acid Metabolizing Cytochrome P450, CYP26B1, and

Comparative Expression Analysis with CYP26A1 during Early Murine

Development

3.1 Summary

Tight regulation of retinoic acid (RA) distribution in the embryo is critical for normal morphogenesis. The RA-metabolizing enzymes CYP26A1 and CYP26B1 are believed to play important roles in protecting certain embryonic tissues from inappropriate RA signaling. We have cloned the murine Cyp26b1 cDNA and compared its expression pattern to that of Cyp26a1 from embryonic day (E) 7 to 11.5 using in situ hybridization. Northern blot analysis shows the presence of two Cyp26b1 transcripts of approximately 2.3 and 3.5 kb in embryonic limb bud. Whereas Cyp26a1 is expressed in gastrulating embryos by E7, Cyp26b1 is first expressed at E8.0 in prospective rhombomeres 3 and 5. Cyp26b1 expression expands to specific dorso-ventral locations in rhombomeres 2 to 6 between E8.5-E9.5, whereas Cyp26a1 hindbrain expression is limited to rhombomere 2 at E8.5. No (or very weak) Cyp26b1 expression is observed in the tail bud, a major site of Cyp26a1 expression. Differential expression is seen in branchial arches, with Cyp26a1 being mainly expressed in neural crest-derived mesenchyme, and Cyp26b1 in specific ectodermal and endodermal areas. Cyp26b1 is markedly expressed in the ectoderm and distal mesoderm of the limb buds from the beginning of their outgrowth. Cyp26a1 transcripts are seen later and at lower levels in limb ectoderm, and both transcripts are excluded from the apical ectodermal ridge.

3.2 Introduction

Retinoic acid (RA), an active metabolite of vitamin A, is a critical signaling molecule during vertebrate development (Maden 1998). Maternal vitamin A deficiency

(Wilson et al. 1953; Maden 1998; White et al. 2000b), or teratogenic RA excess (Maden

1998; Mulder et al. 2000) results in a spectrum of congenital abnormalities. Thus, embryonic RA levels must be strictly controlled.

In vivo, RA is generated in a tissue-specific manner due to differential expression of three retinaldehyde dehydrogenases, RALDH-1, RALDH-2 and RALDH-3 (Zhao et al. 1996; Niederreither et al. 1997; Haselbeck et al. 1999; Li et al. 2000; Mic et al. 2000) and references therein). Regions of the embryo expressing Raldh genes coincide with those which have been shown to contain RA (Mendelsohn et al. 1991; Rossant et al.

1991); (Maden et al. 1998), and embryos lacking RALDH-2 suffer early morphogenetic defects (Niederreither et al. 1999).

Certain embryonic regions that appear devoid of RA (Mendelsohn et al. 1991;

Rossant et al. 1991) express the RA-degrading enzyme CYP26A1 (Iulianella et al. 1999;

McCaffery et al. 1999), a novel cytochrome p450 that metabolizes RA into 4-hydroxy and 4-oxo metabolites (White et al. 1996; Fujii et al. 1997; White et al. 1997; Abu-Abed et al. 1998; Hollemann et al. 1998; Swindell et al. 1999). Cyp26a1 knockout mice die during mid-late gestation and exhibit caudal regression, spina bifida, sirenomelia, as well as hindbrain patterning defects and vertebral transformations (Abu-Abed et al. 2001;

Sakai et al. 2001).

CYP26B1 was recently characterized in human as a CYP26A1 homologue of comparable enzymatic activity (White et al. 2000a). We have now cloned its murine counterpart and analyzed its mRNA distribution. We show that the two genes are expressed in essentially distinct domains during early embryogenesis (E7-E11.5). Thus, the complex spatio-temporal expression patterns of RA-synthesizing and degrading enzymes may provide a combinatorial mechanism for generating morphogenetic cues.

3.3 Materials and Methods

3.3.1 RNA extraction, RT-PCR and northern blotting

Total RNA from E11.5 fore- and hindlimb buds (30 embryos) was extracted with

TRIzol (Gibco, BRL). Two mouse ESTs (Genbank Accesion #: AW488377,

AW047279) highly homologous to human Cyp26b1 were used to design PCR primers

(sense 5'ATGCTGTTTGAAGGCTTGGAGTTG, antisense 5'CTACACCGTAGCACTC

AACATGGC). RT-PCR was performed using the ThermoScriptTM RT-PCR System

(Gibco) and TakaraExTaqTM Polymerase (TaKaRa Biomedicals).

Poly (A)+ RNA was prepared from limb bud RNA using an Oligotex kit (Qiagen), and 2 g was elecrophoresed on an agarose/formaldehyde gel, transferred to Hybond-N+

(Amersham Pharmacia Biotech) and hybridized as described (Abu-Abed et al. 1998).

The Cyp26b1 probe used spanned putative exons 2-5.

3.3.2 Whole-mount in situ hybridization

An RNA labelling kit (Roche) was used to synthesize Cyp26a1 and Cyp26b1 antisense riboprobes from plasmids containing full length cDNA and exons 2-5, respectively. In situ hybridization was performed on CD-1 mouse embryos (Charles

River) as described (Decimo et al. 1995; Niederreither et al. 1997). Some of the hybridized embryos were postfixed in 2.5 % glutaraldehyde and embedded using the JB-

4 embedding kit (Polysciences) for sectioning.

3.4 Results

3.4.1 Identification of mouse CYP26b1

Full length murine Cyp26b1 was amplified by RT-PCR, using E11.5 limb bud

RNA as template (see Materials and Methods). The 1539 base pair murine cDNA shows

93% amino acid identity with human CYP26B1, and 43% with both human and mouse

CYP26A1 (Figure 3.1).

Northern blot analysis of poly(A+) E11.5 limb bud RNA, hybridized with a murine Cyp26b1 cDNA probe, revealed a doublet with transcript sizes of approximately

2.3 and 3.5 kb (Figure 3.2).

3.4.2 Expression at E7.0-E7.5

Cyp26a1 expression appears by E6.5-7.0 in the extra-embryonic and embryonic endoderm, and then in primitive streak and posterior mesoderm (Fujii et al. 1997; de

Roos et al. 1999). Transient expression is seen in anterior headfold tissues by E7.0-7.5

(Figure 3.3 A). We also observed Cyp26a1 expression at E7.5 within the cardiac crescent and lateral vascular embryonic network, as well as in the allantoic network (Figure 3.4

A,B). In contrast, Cyp26b1 transcripts were not detected in E7.0-7.5 embryos (Figure 3.3

B).

Figure 3.1: Amino acid alignment of Cyp26 members. Amino acid alignment of murine (m) CYP26B1 with human (h) CYP26B1, along with human and mouse

CYP26A1, is shown. Conserved residues with respect to the mouse CYP26B1 sequence are shaded.

Figure 3.2: Size of Cyp26b1 mRNA transcripts. Northern blot analysis was performed with poly(A)+enriched RNA isolated from E11.5 mouse limb buds. Two RNA species are observed with sizes of approximately 3.5 and 2.3 kb (arrowheads).

3.4.4 Expression at E8.0-E8.5

Cyp26a1 expression in tailbud mesoderm and posterior neural plate has been documented (Fujii et al. 1997), Figure 3.3 C,E). We found that Cyp26a1 expression is spatially restricted in the E8.0 heart tube endocardium, both rostrally in the outflow tract and caudally in the atria and sinus venosus, from which labelling extends in the lateral vascular networks (Figure 3.4 C). Expression is still seen in the endocardium, but not in other vascular elements, by E8.5-E9 (Figures 3.3 E and 3.5 A,G, and data not shown;

(Moss et al. 1998).

Cyp26b1 expression is initiated at E8.0 in prospective rhombomere (r) 3, and at higher levels in pre-r5 (Figure 3.3 D). Rhombomeric expression expands as soon as the embryos have turned (Figure 3.3 F; see below). When compared to Cyp26a1, which is transiently expressed in r2 by E8.5 (Fujii et al., 1997), Cyp26b1 hindbrain expression is more intense (compare Figures 3.3 E and F). In contrast to Cyp26a1, Cyp26b1 transcripts are not detected in the tail bud region (compare Figures 3.3 C,E and D,F).

Extended staining reveals Cyp26b1 expression between the developing somites and at the level of caudal branchial arches (Figure 3.3 F), whereas Cyp26a1 expression is observed in the rostral portion of the mandibular arch (Figure 3.3 E).

Figure 3.3: Cyp26a1 and Cyp26b1 expression in post-implantation mouse embryos.

(A ,B) E7.5, lateral views. Cyp26a1 is expressed in the anterior region of the embryo

(developing headfolds) and in the extraembryonic endoderm. No expression of Cyp26b1 is seen at this stage. (C, D) E8.0, lateral views. Cyp26a1 expression is observed in the posteriormost mesoderm and neuroepithelium. Cyp26b1 is detected in developing rhombomeres 3 and 5. (E, F) E8.5, lateral views. Strong Cyp26a1 expression is seen in posterior mesoderm and neuroepithelium. Weak expression in rhombomere 2, mandibular arch and heart outflow tract is also evident (arrowheads). Cyp26b1 expression has appeared in rhombomere 6, as well as in the caudal pharyngeal arch region (arrowhead) and intersomitic regions (see higher magnification in inset). A: anterior, hf: headfold; md: mandibular arch; ot: outflow tract; P: posterior, r2-r6: rhombomeres 2 to 6; ve: visceral endoderm.

Figure 3.4: Details of Cyp26a1 expression at E7.5 (A,B: cranial and caudal views, respectively) and E8.0 (C) embryos. Specific expression is seen along the cardiac crescent and developing embryonic and allantoic vascular networks at E7.5, and in the outflow tract and atria/sinus venosus regions of the heart at E8.0. Note that signal is comparatively much higher in the posterior region of the embryo. a: atrium; al: allantois; cc: cardiac crescent; hf: headfold; lv: lateral vascular network; ot: outflow tract; pm: posterior mesoderm; ps: primitive streak; sv: sinus venosus; v: ventricle.

3.4.5 Expression at E9.5-E10.5 Cyp26b1 hindbrain expression becomes more complex at E9.5. Highest expression is seen in r5 and 6, while expression appears weaker in r3, and non-uniform in r2 and 4

(Figure 3.5 D). Hindbrain flatmounts show expression throughout r5-6 and in the ventral portion (basal plate) of r2-4, the floor plate being unlabelled (Figure 3.6 A). Three longitudinal columns are labeled in r4. The same columnar arrangement becomes apparent in r5-6 by E10.5 (Figure 3.6 B).

Cyp26a1 expression in cervical and branchial arch mesenchyme has been reported

(Fujii et al. 1997; de Roos et al. 1999); Figure 3.5 C,E,G). In contrast, Cyp26b1 branchial expression is observed in both ectodermal and endodermal portions of the caudal branchial arches (Figure 3.5 F). Only Cyp26a1 is expressed along the maxilla- mandibular cleft (compare Figure 3.5 E and F). Cyp26b1 expression was not seen in the growing tail tip, where Cyp26a1 expression is maintained (compare Figure 3.5 A,G and

B,H). Prolonged staining revealed weak Cyp26b1 signal in tail bud mesoderm, indicating co-expression with Cyp26a1, albeit at much lower levels (not shown).

Cyp26b1 expression is detected in limb buds at outgrowth (Figure 3.5 B, H).

Expression first appears along the ventral aspect of the buds (Figure 3.7 A,B). Shortly thereafter, expression expands dorsally, but remains weak in the central region where the apical ectodermal ridge (AER) will develop (Figure 3.7 C,E). On sections, expression appears restricted to the ectoderm and/or superficial mesenchyme (Figure 3.7 F).

Cyp26b1 transcripts are distally restricted during further limb outgrowth (Figure 3.7 D).

By E10.5, expression occurs throughout the distal mesenchyme or progress zone, but

50 remains highest along the dorsal and ventral aspects (Figure 3.7 G). Cyp26a1 transcripts were not detected in limbs at these stages (see Figure 3.5 A,G).

Figure 3.5: Cyp26a1 and Cyp26b1 expression in E9.5 and E10.5 embryos. (A,B)

E9.5, lateral views. Cyp26a1 is expressed in cervical mesenchyme, tailbud mesoderm and neuroepithelium (arrowhead). Cyp26b1 is expressed in rhombomeres 2-6, forelimb bud and developing hindlimb bud (indicated by arrowhead). (C,D) E9.5, dorsal views of the hindbrain region. Cyp26a1 expression is no longer seen in the hindbrain, but is present in the adjacent cervical mesenchyme, laterally and caudally to the otic vesicle

(arrowheads). Cyp26b1 is differentially expressed in rhombomeres 2 to 6. (E,F) E9.5, details of the branchial arch labelling patterns. Cyp26a1 expression in cervical mesenchyme encompasses the second arch and is also seen in the caudal region of the first arch. Restricted signal is also seen along the maxillo-mandibular cleft. In contrast,

Cyp26b1 expression is observed in defined areas of the caudal branchial arches

(arrowheads). Histological analysis (inset) shows labeling along the inner (endodermal, left) and outer (ectodermal, right) edges of the branchial arches. (G,H) E10.5, lateral views. Cyp26a1 expression persists in tailbud (arrowhead) and cervical mesenchyme.

Cyp26b1 expression is restricted to the hindbrain (note that transcripts now extend towards the met-mesencephalic boundary, arrowhead) and limb buds. b2: second branchial arch; cm: cervical mesenchyme; fl: forelimb; hl: hindlimb; md: mandibular process of the first branchial arch; r2-r6: rhombomeres; ov: otic vesicle.

Figure 3.6: Details of Cyp26b1 rhombomeric expression in dissected hindbrains. Hindbrains were examined at E9.5 (A) and E10.5 (B) as flatmounts after removal of the underlying mesenchyme and sectioning of the dorsal midline.

Thus, the ventral midline (floor plate) appears as central in these preparations. D: dorsal; FP: floor plate, r2-r6: rhombomeres; V: ventral.

3.4.6 Expression at E11.5 Cyp26a1 expression has been described in the maxillo-mandibular cleft region

(Figure 3.8 A) and the neural retina (results not shown; (Fujii et al. 1997; McCaffery et al. 1999). Cyp26b1 expression is not detected in neural retina (not shown). Cyp26b1 expression has decreased in the caudal hindbrain (myelencephalon), whereas expression in the met-mesencephalic region (first seen in E10.5 embryos) has increased (Figures 3.5

H and 3.8 B, arrowheads). Cyp26b1 expression has also expanded in maxillary and mandibular components of the first branchial arch, as well as in the second arch (Figure

3.8 B). Furthermore, expression has appeared within the developing nasal process (Figure

3.8 B). In the limbs, Cyp26b1 expression remains strong along the distal margin of the footplates (Figure 3.8 B,D), but is excluded from the AER (Figure 3.8 F). Cyp26a1 expression is detected at lower levels in two stripes flanking the AER (Figure 3.8 C,E, arrowheads).

Figure 3.7: Dynamic Cyp26b1 expression during early limb bud development.

(A,C,E) show three sequential stages of forelimb outgrowth (E9-E9.5 embryos, profile views), illustrating the expansion of Cyp26b1 transcripts from a ventral location to both dorsal and ventral locations, and no or very little transcripts in the central, prospective apical ectodermal ridge territory. Black arrows in (A) point to remnants of the intersomitic expression domains. (B,D) show the same embryos as in (A,C), such that forelimb buds are seen from their caudal aspect. Note the distalization of Cyp26b1 transcripts during limb outgrowth. (F,G) Histological sections at E9.0 (early limb budding) and E10.5, respectively. Orientation of the limb axes is indicated. AER: apical ectodermal ridge; D: dorsal; Di: distal; Pr: proximal; V: ventral.

Figure 3.8: Cyp26a1 and Cyp26b1 expression in E11.5 embryos. (A,B) Lateral views.

At this stage, Cyp26a1 expression in the branchial region is more restricted than that of

Cyp26b1, which has expanded up to the first arch, and has also appeared in the nasal process. Cyp26b1 expression at the level of the met-mesencephalic boundary (arrowhead) has become stronger than within the hindbrain. Cyp26b1 expression remains prominent in the developing limbs. (C-F) High magnification views of the forelimb extremity

(footplate). Cyp26b1 is highly expressed along the distal margin of the developing footplate, except for the apical ectodermal ridge (D,F). Although Cyp26a1 limb expression is markedly weaker (C,E), two labeled stripes can be distinguished along each side of the ridge (arrowheads). AER: apical ectodermal ridge; fl: forelimb; hl: hindlimb; md: mandibular process; mx: maxillary process; np: nasal process.

3.5 Discussion

The importance of Cyp26 enzymes in regulating morphogenesis through modulating RA distribution has recently been demonstrated in Cyp26a1-/- mice (Abu-

Abed et al. 2001; Sakai et al. 2001). CYP26A1 null embryos die during gestation and exhibit phenotypes characteristic of RA teratogenesis such as spina bifida and posterior truncation. However, the limbs and craniofacial structures, which are common targets of exogenous RA, develop normally in these mice, at least with respect to gross morphology. A second RA-catabolizing enzyme, CYP26B1 was recently cloned in human and shown to have substrate specificity and catalytic activity similar to those for

CYP26A1 (White et al. 2000a). In this work, we have cloned the murine ortholog of

Cyp26b1, and analyzed the expression of this gene during early embryogenesis.

Cyp26a1 and Cyp26b1 have comparable enzymatic activity in vitro, but are expressed in different regions throughout development (White et al. 2000a). Cyp26a1 is generally restricted to the tailbud, cervical mesenchyme and ectoderm of the first branchial arch. In contrast, Cyp26b1 is highly expressed in the hindbrain, limbs and nasal process. The distribution of RA in vivo has been characterized using transgenic reporter mouse lines; the pattern of reporter activation is generally complementary to that of Cyp26a1 and Cyp26b1 expression (Mendelsohn et al. 1991; Rossant et al. 1991).

Cyp26a1 is predominately expressed in the caudal embryo, and severe malformations in this region are observed when Cyp26a1 is deleted (Abu-Abed et al. 2001; Sakai et al.

2001). Coupled with in vitro metabolic assays, these results indicate that Cyp26 enzymes

61 have a critical role in establishing tissue boundaries where RA-mediated signaling will not occur.

The dynamic expression pattern of Cyp26b1 in the developing hindbrain, suggests that this enzyme is important for hindbrain patterning, and possibly the development of branchial arch derived structures. Retinoid status is important for hindbrain patterning, as low levels of RA are required for specification of anterior rhombomeres, while higher levels of RA are needed for development of posterior rhombomeres (Gavalas 2002).

Furthermore, hindbrain patterning has been demonstrated to be particularly sensitive to excess RA exposure. Considering the dynamic expression pattern of Cyp26b1 in the hindbrain, this enzyme may be important for defining the anterior boundary of RA in the neural tube. Neural crest cells (NCC) migrating from r2, 4 and 6 populate the first, second and third branchial arches, and many of the branchial arch derived tissues are sensitive to exogenous RA. Therefore, Cyp26b1 expression in the hindbrain may also be important for NCC specification and migration. Functional analysis of Cyp26b1 in vivo by gene targeting will hopefully demonstrate the role of CYP26B1 in regulating retinoid signaling across cell fields in the hindbrain and limb bud.

Chapter 4

CYP26B1, a Retinoic Acid Catabolizing Enzyme is Dispensable for Hindbrain

Patterning, but Essential for Craniofacial Development

4.1 Summary

Retinoic acid (RA) is a potent morphogen that regulates cellular growth and differentiation of several tissues in the developing embryo. Levels of RA must be tightly controlled during embryogenesis as either an excess or deficiency of RA can have severe consequences for morphogenesis. There is a spatio-temporal distribution of RA in vivo, such that some tissues are rich in RA, while others are devoid. This patterned distribution is established and maintained by the coordinated expression of RA- synthesizing (ALDH) and RA-degrading (CYP26) enzymes. CYP26B1 is a RA- metabolizing cytochrome P450 that is expressed in the embryonic limb, hindbrain and craniofacial elements that we have cloned and characterized. We have also generated transgenic mice lacking CYP26B1 and observed numerous developmental abnormalities.

Cyp26b1-/- mice die shortly after birth and exhibit shortened limbs, open eyes at birth and a lack of male germ cells. Mutant mice also have a cleft palate, show reduced calvarial ossification and have numerous malformed craniofacial structures. Molecular analysis of the hindbrain indicates that despite the high expression levels, Cyp26b1 is dispensible for rhombomere identity and patterning. However, there do appear to be subtle differences in neural crest cell identity and migration, which may be responsible for some of the craniofacial phenotypes observed.

4.2 Introduction

Craniofacial malformations are among the most common human birth defects accounting for roughly half of all birth defects (Christianson et al., 2006). A critical process in craniofacial development is the controlled formation, proliferation and migration of neural crest cells (NCC) (Tapadia et al. 2005). NCCs are multipotent progenitor cells capable of developing into bone, cartilage, muscle or neural structures that contribute to the formation of the cranium and facial structures (Kaufman and Bard J.

B. 1999; Chai and Maxson 2006). In mammals, NCCs are derived from the neural folds, along the dorsal neural tube at the junction of the surface ectoderm and the neural plate.

As the neural folds fuse, NCCs migrate throughout the embryo and give rise to numerous structures. Depending on the location along the anterior-posterior axis of the neural tube,

NCCs form distinct populations, and are classified as either cranial, cardiac, vagal or trunk. Cranial NCCs (CNCC) originate in either the forebrain, midbrain or hindbrain and contribute to the formation of the frontal and interparietal skull plates, cranial ganglia, the inner ear, hyoid, as well as bones of the jaw and nasal process (Chai and Maxson 2006).

Genetic and pharmacologic studies have shown that disruption of NCC patterning, migration or survival can result in a spectrum of craniofacial abnormalities both in experimental models, and in humans (Sulik et al. 1988).

Retinoic acid (RA) is a morphogen that regulates cell differentiation, proliferation and survival . RA has been identified as a teratogen and has the potential to affect many developing tissues, depending on the embryonic age at the time of administration

(Kochhar 1973). Manifestations of RA teratogenicity include cleft palate, inner ear

65 malformations, thymic agenesis, craniosynostosis and hypoplasia of the mandible and maxilla (Yip et al. 1980; Frenz et al. 1996; Mulder et al. 1998; Mulder et al. 2000;

Collins et al. 2006). CNCCs contribute to the development of all of these structures, and it has been proposed that some of the teratogenic effects of RA may be due to abnormalities in NCC migration or differentiation (Uehara et al. 2007). NCC derived tissues are also sensitive to a depletion of RA, inhibition of RA synthesis or signaling can result in fetal abnormalities affecting the branchial arches, forebrain, skull and face

(Schneider et al. 2001; Halilagic et al. 2003; Mark et al. 2004).

RA is able to modulate gene expression by acting as a ligand for the nuclear retinoic acid receptors (RAR) (Chambon 1996). RARs form functional heterodimers with the retinoid X receptors (RXR), that bind to specific DNA sequences known as retinoic acid response elements (RAREs). Upon ligand binding, RAR and RXR undergo a conformational shift that facilitates the binding of transcriptional activator proteins.

These proteins have histone acetylase activity (HAT) that mediate chromatin relaxation and thus promote gene expression (Dilworth and Chambon 2001).

There are three distinct RAR genes ( , each with several isoforms that exhibit tissue specific expression patterns (Dolle et al. 1990; Mollard et al. 2000).

However, there is significant functional redundancy amongst the RARs. Gene targeting experiments in mice have shown that deletion of a single receptor typically results in mild phenotypes, while compound mutants lacking two or more RARs exhibit more severe phenotypes resembling those observed in cases of maternal VAD (Kastner et al. 1995;

Lohnes et al. 1995). Therefore, receptor expression may not be a key step in defining tissues where RA-mediated signaling occurs.

RA synthesis is catalyzed by a family of retinaldehyde dehydrogenases (ALDH1-

3), which irreversibly oxidize retinal to form RA (Niederreither and Dolle 2006).

Conversely, Cyp26 enzymes, a family of cytochrome P450 proteins, catabolize RA by generating more polar metabolites that have limited activity in vivo. The cyp26 and aldh genes exhibit tightly regulated expression patterns that are responsible for defining the tissue specific distribution of RA (Pennimpede et al. 2006).

Aldh1a2 is the predominant retinaldehyde dehydrogenase in the embryo, with expression beginning at E7.5 in the embryonic mesoderm, and later spreading to the somites, flank mesoderm and posterior hindbrain mesenchyme (Niederreither et al. 1997;

Sirbu et al. 2005). Cyp26 genes exhibit a complimentary expression pattern, particularly in the hindbrain. Cyp26a1 is initially expressed at E7.75 in the anterior headfolds with a posterior boundary of presumptive rhombomere (pr) 2/3. Shortly after, Cyp26c1 is expressed in pr4/5 at embryonic day (E) 8.0 to establish an anterior boundary to RA

(MacLean et al. 2001; Sirbu et al. 2005). Cyp26b1 is the final cyp26 to be detected in hindbrain with expression in rhombomeres (r) 2 and 4 at E8.25; however, once initiated,

Cyp26b1 hindbrain expression is maintained throughout the hindbrain until E10.5

(MacLean et al. 2001). Cyp26b1 also appears to be the major cyp26 gene expressed in the branchial arches with strong expression at E11.5, while Cyp26a1 and Cyp26c1 are only transiently detected at the maxillo-mandibular cleft of the first arch at E9.5

(MacLean et al. 2001; Tahayato et al. 2003).

Studies have shown that maintenance of RA levels by these proteins is essential for hindbrain and early NCC patterning. Aldh1a2-/- mice have an abnormal hindbrain with an enlarged anterior aspect, and lack the second and third branchial arches

(Niederreither et al. 1999; Niederreither et al. 2000). Furthermore, mice with hypomorphic Aldh1a2 alleles die perinatally with malformations of branchial arch- derived structures characteristic of the human birth disorder, DiGeorge Syndrome

(Vermot et al. 2003). Cyp26a1 null mutant mice exhibit a mild defect in hindbrain patterning, and Cyp26a1/Cyp26c1 double knockout animals have an enlarged posterior hindbrain at the expense of anterior structures (Abu-Abed et al. 2001; Uehara et al.

2007). These genetic models support observations in zebrafish, chick and frog that RA may act as a posteriorizing signal in the hindbrain, and correct RA-mediated signaling is required for NCC development (Glover et al. 2006) .

Cyp26b1 has previously been shown to be critical for modulating RA levels thus allowing for proper development of the limb and testes (Yashiro et al. 2004; Bowles et al.

2006; MacLean et al. 2007). It was also observed that Cyp26b1 null animals exhibit severe craniofacial abnormalities; although neither these malformations nor underlying molecular mechanisms have been described in any detail. Here we present a detailed analysis of Cyp26b1-/- embryos, and provide evidence that although hindbrain patterning is largely unaffected in these animals, there are subtle abnormalities in NCC and branchial arch development possibly contributing to the observed phenotypes.

4.3 Materials and Methods

4.3.1 Construction of the targeting vector and generation of Cyp26b1-deficienct mice

A bacterial artificial chromosome (BAC) clone containing the murine Cyp26b1 locus was obtained from a SV129 genomic library screened by the Centre for Applied

Genomics (TCAG, Toronto, ON). Restriction digest of the clone with BamHI generated an 8.9 kb fragment spanning exons 3-6 that was gel-purified and cloned into the vector pBS (Stratagene, La Jolla, CA). SpeI and NotI restriction sites were added by site- directed mutatgenesis into intron 2 and the 3‟-UTR respectively. An oligonucleotide containing a LoxP site as well as DraIII, NaeI and PvuII restriction sites was ligated into the introduced SpeI site. A neomycin resistance marker flanked by LoxP sites was cloned into the NotI site downstream of the stop codon (Metzger and Feil 1999). The targeting construct was linearized and then electroporated into embryonic stem (ES) cells as previously described (Abu-Abed et al. 2001). Homologous recombination of the 3‟ end of the construct in G418 resistant ES cells was established by Southern blot analysis, and integration of both 5‟ and 3‟ ends of the construct was verified by PCR.

ES cells with one targeted Cyp26b1 allele (L3/+) were injected into blastocysts as described elsewhere (Abu-Abed et al. 2001). Agouti mice with germ-line transmission of the L3 allele were bred with CMV-Cre mice which act as a deletor strain for floxed alleles (Dupe et al. 1997), and resultant offspring were genotyped by PCR for excision of exons 3-6 corresponding to the Cyp26b1 allele. Cyp26b1+/− mice were then crossed to generate Cyp26b1−/− animals.

PCR was used to genotype Cyp26b1 mice using DNA isolated from either tail clips or yolk sacs. A standard PCR was performed using Taq polymerase (Sigma-

Aldrich, Oakville, ON, Canada), and 3 primers, P1: 5‟-CAGTAGATGTTTGAGT

GACACAGCC, P2: 5‟-GAGGAAGTGTCAGGAGAAGTGG and P3:5‟-GGGCCAC

CAAGGAAGATGCTGAGG. P1 and P2 amplify a product of 223 bp from the wild-type allele, or a product of 284 bp from a targeted (L3) allele, but will not amplify a null allele. P1 and P3 will amplify a 364 bp product from only the excised allele. All reactions included 1.5 mM MgCl2, 0.2 mM dNTPs, 0.4 mM of P1, and 0.2 mM of P2 and

P3. Each PCR consisted of 30 cycles of 94 ˚C for 20 sec, 58 ˚C for 45 sec and 72 ˚C for

50 sec. Resultant products were visualized by ethidium bromide staining after electrophoresis in a 2% agarose gel.

4.3.2 Embryo Collection

Timed matings were established by housing female mice with males overnight.

Females were checked daily for vaginal plugs, and noon of the day that a plug appeared was denoted as E0.5. On the appropriate day, pregnant females were humanely euthanized by cervical dislocation, and embryos were collected in ice-cold phosphate- buffered saline (PBS).

All animal experimentation was reviewed and approved by the University‟s

Animal Care and Use Committee.

4.3.3 Alizarin Red and Alcian Blue staining

For alizarin red/alcian blue staining of E18.5 skeletons, embryos were collected, skinned, eviscerated and fixed overnight in ethanol. Post-fixation, embryos were stained for 48 hours with 0.015% alcian blue (Sigma) in 20% glacial acetic acid/76% ethanol.

After dehydrating overnight in 95% ethanol, embryos were placed in 1% KOH until ribs became clearly visible (usually 8-10 hours). Bones were then stained overnight with a solution of 0.015% alizarin red (Sigma) in 1% KOH, and skeletons were subsequently cleared with 20% glycerol/1 % KOH until ready to be photographed.

E14.5 embryos were processed by fixing for 48 hours in 95% ethanol, then transferred to acetone overnight. Embryos were stained in solution of 0.03% alcian blue

(Sigma) in 20% glacial acetic acid/76% ethanol. After two rinses in 95% ethanol, embryos were placed in 1% KOH for 2 hours and then left to clear in 80% glycerol/

1% KOH.

4.3.4 Whole-mount in situ hybridization

In situ hybridization was performed on embryos fixed overnight in 4% paraformaldehyde as previously described (Decimo et al. 1995; Niederreither et al. 1997)

All antisense riboprobes were synthesized using an RNA digoxgenin labeling kit (Roche,

Laval, PQ, Canada) and the appropriate RNA polymerase (Roche). For each probe, embryos of each genotype (+/+, +/-, -/-) were processed concurrently, and the colour reaction was started at the same time, thus the duration of staining for each genetic group was the same for any given probe. Each probe presented was analyzed on at least 3

Cyp26b1-/- embryos.

4.4 Results

4.4.1 Generation of a Cyp26b1 targeting construct

In order to generate a targeting construct to ablate Cyp26b1 expression, a mouse expressed sequence tag (Genebank Accesion #: AW047279) was used as a probe to screen a mouse SV129 genomic library (performed at The Centre for Applied Genomics,

University of Toronto). From positive bacterial artificial chromosome (BAC) clones

(confirmed by PCR and Southern blot, results not shown), a BamHI digest generated an

8.9 kb fragment that spanned exons 3-6 of Cyp26b1. Sequence alignment with Cyp26a1, and other characterized cytochrome-P450 genes indicated that the putative heme binding site of Cyp26b1 is encoded by exon 6. This motif is critical for cytochrome-P450 function, hence deletion of exons 3-6 should yield a nonfunctional peptide, even if exons

1 and 2 are expressed and translated.

The 8.9 kb fragment was subcloned into a modified pBS vector, and site directed mutagenesis was performed to add a SpeI site into intron 2, and a NotI site after exon 6.

An oligonucleotide containing a LoxP site was ligated into the SpeI site, while a neomycin resistance marker (neo) flanked by LoxP sites was ligated into the NotI site

(Figure 4.1 A). The resultant vector was linearized by a KpnI digest and electroporated in SV129 embryonic stem (ES) cells.

ES cells were grown in media supplemented with neomycin to select for clones into which the construct had integrated. In order to determine whether homologous recombination had occurred, cell lysates were collected and screened by Southern blot.

Integration of the targeting construct into the endogenous Cyp26b1 allele will add an

72 additional PvuII restriction site. Therefore, when a DNA probe external to the BamHI fragment is used to screen PvuII digested genomic DNA by Southern blot, the wild-type allele will give a band at 5.2 kb, while a targeted allele will yield a band at 3.8 kb (Figure

4.1 B). Cells positive for inclusion of the 3‟ end of the targeting construct were screened by PCR for integration of the 5‟ LoxP site by a nested PCR strategy (results not shown).

384 cell lysates were screened and three clones were positive for homologous recombination. These clones were expanded, and two were observed to have abnormal karyotypes, so the remaining clone was selected for blastocyst microinjection. After microinjection, and blastocyst implantation into pseduopregnant female mice, resultant offspring were screened for germ line transmission of a targeted allele. Eventually, four founder mice, two males and two females were obtained and crossed with CMV-Cre mice, which express Cre recombinase ubiquitously and act as a general deletor strain for

„floxed‟ (Flanked by LoxP sites) loci (Dupe et al. 1997). DNA extracted from tail clips was used as template to genotype offspring by PCR. When primers P1, P2 and P3 are added together in PCR mix with limited extension time, P1 and P2 will amplify a product of 223 bp from a wild-type allele, and 284 bp from an L2 or L3 allele. Primers P1 and P4 will amplify a product of 364 bp from a null allele only (Figure 4.2 A, B).

Figure 4.1: Homologous recombination is used to integrate LoxP sites into the

Cyp26b1 locus. (A) A targeting construct containing 8.9 kb of sequence homologous to mCyp26b1 is used to integrate three LoxP sites and a neomycin resistance marker (neo) into the endogenous locus. LoxP sites are represented by the small triangles, and the probe used for Southern blotting is shown. (B) Southern blot analysis performed on

DNA isolated from ES cell lysates. When digested with PvuII and hybridized with the indicated probe, a wild-type allele will show a band at 5.2 kb, while a targeted allele will show a band at 3.8 kb.

Figure 4.2: Cre-mediated excision of Cyp26b1. (A) After exposure to Cre recombinase, three possible products are possible, containing 1 LoxP site (null allele), 2 LoxP sites

(L2-conditional allele) or 3 LoxP sites (L3-Targeted allele). Cre excision products can be differentiated by PCR using primers P1, P2 and P3 as indicated. (B) Sample mouse genotype by PCR using P1, P2 and P3.

4.4.2 Cyp26b1-/- embryos die perinatally and exhibit numerous malformations

Cyp26b1+/- mice are healthy and fertile, with no obvious phenotypes. When

Cyp26b1 heterozygous mice are mated, Cyp26b1-/- embryos collected at various developmental stages are obtained at the expected Mendelian ratio (results not shown).

Although born at the expected frequency, Cyp26b1-/- mutants die shortly after birth due to apparent respiratory distress. Cyp26b1 null mutants exhibit numerous gross abnormalities including truncated fore and hind limbs, open eyes at birth, a misshapen head, small malformed external ears, a shortened nasal process, loose skin and a distended abdomen (Figure 4.3 A, B). Cyp26b1-/- null mutants also develop a cleft palate, which has been confirmed by histological analysis (results not shown). E18.5 alizarin red/alcian blue stained skeletons show that Cyp26b1-/- animals lack mineralization of numerous structures of the skull (see below), and also exhibit irregular curvature of the spine (Figure 4.3 C, D). As has been previously reported, Cyp26b1-/- mice have truncated limbs due to reductions in zeugopod and stylopod development, and also display oligodactyly with two or three digits per appendage (Yashiro et al. 2004).

Interestingly, it was observed that while the scapula is largely unaffected, the ilium is consistently malformed (Figure 4.3 C, D), which has not been noted before (Yashiro et al.

2004). Alcian blue stained E14.5 embryos clearly show that Cyp26b1-/- mice lack clavicles, which may result in difficulties breathing, and contribute to mortality (Figure

4.3 E, F).

Figure 4.3: Gross malformations in Cyp26b1 null embryos. (A, B) Lateral views of

E18.5 (A) Wild-type and (B) Cyp26b1-/- embryos showing truncated limbs, a misshapen head and distended abdomen. (C, D) Alizarin red/alcian blue stained skeletons of E18.5

(C) Wild-type and (D) Cyp26b1-/- embryos exhibiting numerous deficiencies in ossification and an irregularly curved spine. (E, F) E14.5 alcian blue stained embryos.

Comparison of wild-type (E) and Cyp26b1-/- (F) embryos shows that Cyp26b1 null mutants lack clavicles as indicated by arrows. (Cl) clavicle.

4.4.3 CYP26B1 is required for normal craniofacial ossification

When compared to wild-type littermates, alizarin red/alcian blue stained

Cyp26b1-/- skeletons exhibit severe deficits in formation and mineralization of structures of the skull and face (Figure 4.4 A-F). Lateral views show that Cyp26b1 -/- skulls have a shortened mandible and are missing most of the ossified maxilla and premaxilla (Figure

4.4 A, D). The frontal plate of the skull is less prominent in Cyp26b1 -/- embryos, and the parietal plate is also consistently smaller. In addition, the atlas and the axis are fused in

Cyp26b1 null animals (Figure 4.4 D, asterisk), which appears to be the one abnormality that is shared with Cyp26a1-/- mice (Abu-Abed et al. 2003).

The superior view of wild-type skulls shows the parietal, frontal and interparietal plates are highly ossified with very little space between the plates (Figure 4.4 B). In contrast, skulls from Cyp26b1-/- mice have a striking reduction in the size of the frontal and parietal plates such that there is an unossified region on the top of the skull, and the palate can be clearly visualized below (Figure 4.4 E).

When the mandible is removed, the inferior view of skull shows that there are numerous malformations (Figure 4.4 C, F). Cyp26b1-/- mutants lack the jugal and ptgerygoid bones as well as the zygomatic process of the maxilla. Mutants also show reduced ossification of the maxilla and premaxilla and an abnormal palatine. Finally, in wild-type embryos, there is a band of cartilage separating the exoccipital and basioccipital bones; in null mutants, this cartilaginous region is missing and the bones appear to be fused or overgrown (Figure 4.4 C, F, arrowheads).

Cyp26b1-/- embryos also exhibit abnormal ear development, as the external pinnae are generally smaller than those of wild-type littermates (Figure 4.3 A, B). There are also defects in inner ear formation, as the ectotympanic process is reduced or missing in

Cyp26b1-/- animals (Figure 4.4 G, H). Cyp26b1 null mutants also lack the squamosal bone (Figure 4.4 G, H).

Isolation of the trachea and larynx clearly shows that the hyoid bone is absent in

Cyp26b1-/- embryos (Figure 4.4 I, J) and that the cartilage of the larynx is abnormally formed. It is unclear if the hyoid cartilage is fused to the larynx, or if the hyoid cartilage is missing, and the larynx cartilage is also malformed.

Alcian blue stained embryos were examined at E14.5 to observe potential irregularities in precursor elements of craniofacial bones. Most strikingly, when viewed laterally, Meckel‟s cartilage is severely truncated in mutant mice (Figure 4.5 A, D).

When viewed from below, it is apparent that in addition to truncation, the distal aspect of

Meckel‟s cartilage does not fuse (Figure 4.5 B, D). At this stage it is also obvious that there are several regions of cartilage deposition in the presumptive frontal plate and maxilla (Figure 4.5, A, D yellow arrowheads) that are absent in Cyp26b1-/- embryos; this may represent either a delay in formation, or a complete absence of these structures.

The mandibles of Cyp26b1-/- embryos are obviously shortened in comparison with wild-type littermates (Figure 4.4 A, D, Figure 4.5 C, F), but closer inspection reveals many structural deficiencies. The condylar process is significantly smaller, and the angular process is absent in Cyp26b1 null mutants (Figure 4.5 C, F). Furthermore, the

82 dentary is reduced in both width and length, appears more porous, and is no longer continuous with the coronoid process (Figure 4.5 F).

Figure 4.4: Severe craniofacial abnormalities in Cyp26b1 null mice. (A, D) Alizarin red/alcian blue stained skeletons from E18.5 embryos. Lateral views of wild-type (A) and

Cyp26b1-/- (D) skulls showing that mutants have smaller frontal and parietal plates, reduced ossification of the maxilla and premaxilla and a fusion of the atlas and axis as indicated by the asterisk (*). (B, E) Superior views of wild type (B) and mutant (E) skulls demonstrating the lack of calvarial ossification in the absence of CYP26B1. (C, F)

Inferior views of skulls with the mandible removed from wild-type (C) and Cyp26b1 null

(F) embryos. Cyp26b1-/- mice exhibit a bilateral loss of the zygomatic process of the maxilla and jugal bone, and display abnormal pterygoid and palatine bones. There is also a fusion of the basioccipital and exocciptial bones whereas in wild-type mice there is a band of cartilage separating these structures (arrowheads). (G, H) Lateral views of the ear structures of wild-type (G) and Cyp26b1-/- (H) embryos. Mutant embryos posses a shortened ectotympanic process and lack the squamosal bone. (I, J) Isolated tracheas from wild-type (I) and Cyp26b1-/- (J) embryos demonstrating the lack of a hyoid bone and cartilage in null embryos. (At) Atlas; (Ax) Axis; (Bo) Basioccipital; (Eo)

Exocciptial; (Et) Ectotympanic process; (H) Hyoid; (Jg) Jugal; (FP) Frontal Plate; (IP)

Interparietal Plate; (Md) Mandible; (Mx) Maxilla; (P) Parietal Plate; (Pl) Palatine; (Pmx)

Premaxilla; (Ptg) Pterygoid; (Sq) Squamosal; (Zmx) Zygomatic process of maxilla.

Figure 4.5: Malformations of Meckel’s cartilage and the mandible in Cyp26b1-/- embryos. (A, B, D, E) Alcian blue stained E14.5 embryos. Lateral views of wild-type

(A) and mutant (D) embryos demonstrating the truncation of Meckel‟s cartilage as well as reduced cartilage formation in the maxilla and presumptive frontal plate in Cyp26b1-/- mice (yellow arrowheads). Inferior views of wild-type (B) and Cyp26b1-/- (E) embryos showing the lack of fusion of the distal aspect of Meckel‟s cartilage (red arrow) in

Cyp26b1 null mutants. (C, F) Alizarin red stained mandibles isolated from E18.5 littermates. (C) Superior view, mandibles from Cyp26b1-/- embryos are much smaller than wild-type counterparts and have a truncated condylar process. (F) Lateral view, mandibles from Cyp26b1 null mutants lack the angular process, have a significantly reduced condylar process and the dentary appears more porous and is discontinuous with the coronoid process. (Agp) Angular process; (Cdp) condylar process; (Crp) Coronoid process; (D) Dentary; (Mc) Meckel‟s cartilage; (Mx) Maxilla.

4.4.4 Hindbrain patterning is unaffected in CYP26B1 null embryos

In order to examine the integrity of the hindbrain, in situ hybridization with various molecular markers was performed. Meis2 is expressed in rhombomeres (r) 2 and

3, and no differences in expression were observed between Cyp26b1-/- and wild-type embryos at E9.5 (results not shown). HoxB1 encodes a homeobox protein that is normally expressed in r4 and the tailbud. No change in rhombomeric expression is evident at either E8.5 (results not shown) or E9.5 (Figure 4.6 A-D) when Cyp26b1-/- and

Cyp26b1+/- or +/+ embryos are compared. R3 and 5 were analyzed by looking at Krox-20 expression; at E8.5, neither rhombomere appears abnormal in Cyp26b1-/- embryos

(Figure 4.6 E-H). However, there is a novel stream of expression that runs posterior and ventral to r5 (yellow arrows, Figure 4.6 E, G). Kriesler is usually confined to r5 and r6, but in Cyp26b1-/- embryos, there is a posterior expansion in expression in the ventral neural tube (Figure 4.6 I-L, red arrows). The posterior boundary of the hindbrain was analyzed by examining RAR expression. RAR is usually detected in the neural tube with an anterior boundary of expression at r6/r7 (Serpente et al. 2005), and in Cyp26b1 -/- mutants, this boundary is maintained (Figure 4.6 L-O, red arrow).

Figure 4.6: Cyp26b1 is largely dispensable for hindbrain patterning. Whole-mount in situ hybridization with probes for hindbrain markers. (A-D) At E9.5, HoxB1 expression in r4 is unchanged in Cyp26b1+/- (A, B) or Cyp26b1-/- (C, D) embryos when viewed laterally or dorsally. (E-H) Krox-20 expression at E8.5 indicates that r3 and r5 form normally in Cyp26b1+/- (E, F) and Cyp26b1-/- (G, H) embryos. However, Cyp26b1-/- embryos exhibit a novel stream of expression posterior and ventral to r5 (yellow arrow).

(I-L) Kreisler is detected in r5 and r6 of E9.5 Cyp26b1+/- (I, J) and Cyp26b1-/- (K, L) embryos. Similar to Krox-20, there appears to be a posterior and ventral expansion of expression in Cyp26b1 null mutants (red arrow). (M-P) RAR expression is identical in

Cyp26b1+/- (M, N) and Cyp26b1-/- (O, P) embryos, with the same anterior boundary of expression (red arrow). (BA) branchial arch; (ov) otic vesicle; (r) rhombomere.

4.4.5 Subtle abnormalities in branchial arch patterning are observed in Cyp26b1-/- embryos

Many of the malformed structures in E18.5 Cyp26b1-/- mutants are branchial arch derived tissues. Despite the severe phenotypes observed, there are no obvious morphological abnormalities in branchial arch development during mid-gestation (E8.5-

E11.5). In situ hybridization was performed to analyze the early molecular patterning of the branchial arches. In E9.5 wild-type embryos, Fgf8 is expressed in the outgrowing limb bud, nasal process, second, third and fourth branchial pouches, as well as the maxillo-mandibular cleft of the first branchial arch (Figure 4.7 A). At this stage in

Cyp26b1-/- embryos, there appears to be enhanced Fgf8 expression in the branchial pouches, although pouch morphology appears normal (Figure 4.7 B, p2, p3, p4). In order to further characterize branchial pouch development, Eya1 expression was examined at

E10.5. Eya1 is normally detected in maxillary and mandibular components of the first branchial arch, as well as branchial pouches 2-4 at E10.5. No obvious changes in Eya1 expression were observed in Cyp26b1-/- embryos at this stage of development (Figure 4.7

C, D). Similarly, Pax9, a marker of branchial pouches 1-4 was analyzed and no differences were observed between Cyp26b1-/- and wild-type embryos (results not shown).

In the first branchial arch, Fgf8 expression is similar to that of wild-type littermates (Figure 4.7 A, B). In order to further define signaling pathways in the first branchial arch, expression of Bmp4 and Tbx1 was examined. Bmp4 is normally expressed in the anterior and distal ectoderm of the mandibular process, as well as the posterior

91 ectoderm of the maxillary process (Figure 4.7 E). However, in Cyp26b1 null mutants,

Bmp4 expression in this region is reduced at E9.5 (Figure 4.7 E, F arrowheads).

Expression of Bmp4 was also examined in E10.5 embryos, and no abnormalities were detected (results not shown), indicating that the downregulation at E9.5 may be transient.

Tbx1 encodes a T-box transcription factor that is expressed in the first three branchial arches as they develop. In wild-type embryos at E9.5, Tbx1 is readily detected in the first and second branchial arches, but this expression is absent in Cyp26b1-/- embryos (Figure

4.7 G, H).

The Dlx genes encode homeodomain transcription factors that are involved in branchial arch patterning and craniofacial development (Depew et al. 2005). There are six Dlx genes that are arranged as tightly linked pairs in the mouse. Dlx1 and Dlx2 are expressed in the mandibular and maxillary components of the first branchial arch, as well as the second branchial arch. The expression of both of these genes was examined in

E10.5 embryos, and no differences between Cyp26b1-/- and wild-type embryos were observed (Figure 4.8 A, B; results not shown). Dlx5 and Dlx6 are also expressed in the first two branchial arches, but display more distal localization than Dlx1 or Dlx2.

Expression of Dlx5 and Dlx6 at E10.5 was also the same in Cyp26b1-/- and wild-type embryos (Figure 4.8 C, D; results not shown). dHAND is a downstream target of Dlx6 that is expressed in the first and second branchial arches (Charite et al. 2001). As with the Dlx genes, no differences in dHAND expression were observed in Cyp26b1 null mutants, even when embryos were left to overstain (Figure 4.8, E, F).

Figure 4.7: Subtle changes in branchial arch markers in Cyp26b1 null embryos.

Whole-mount in situ hybridization on E9.5 embryos with various markers of branchial arch development. Fgf8 is expressed in the ectoderm of the nasal process, the maxillo- mandibular arch, second, third and fourth branchial pouches of wild-type mice (A). In

Cyp26b1-/- mice, there is a slight expansion of expression in the first branchial arch and induction of Fgf8 in the branchial pouches. Similarly, Bmp4 is expressed at the cleft of the first branchial arch in wild-type mice (C) but not Cyp26b1 littermates (D, arrow).

Tbx1 is detected in the first and second branchial arches wild-type mice (E), but is absent in Cyp26b1-/- mice (F). (Ba) Branchial arch; (p) Branchial pouch.

Figure 4.8: Dlx expression is unaffected in Cyp26b1-/- embryos. Whole-mount in situ hybridization on E10.5 embryos. Dlx2 is expressed in the first and second branchial arches in both wild-type (A) and Cyp26b1-/- (B) embryos. Similarly, Dlx6 expression in the branchial arches is the same in Cyp26b1+/+ (C) and Cyp26b1-/- (D) embryos. dHAND is detected in the distal ectoderm of the first and second branchial arches in wild-type (E) and Cyp26b1-/- (F) overstained embryos. (BA) Branchial Arch.

4.4.6 Neural crest cell migration in Cyp26b1 null embryos

Although hindbrain patterning appeared largely unaffected in Cyp26b1-/- embryos, several markers were analyzed to examine migration and development of NCCs. Snail

(Snai1) encodes a neural-crest cell associated transcription factor that is expressed in the presomitic mesoderm and dorsal headfolds. When wild-type and Cyp26b1-/- embryos are compared at E8.0, there is a discernable downregulation of snail in the headfolds (Figure

4.9 A). Sox9 also appears downregulated at E9.0 in the absence of CYP26B1, as the domain of expression does not extend as far anteriorly in mutants (Figure 4.9 B). CrabpI was examined as a marker of migrating neural crest cells, and there was a general upregulation throughout the neural tube in Cyp26b1-/- embryos. However, in the mesenchyme between the second and third branchial arches (Figure 4.9 C, E, red arrowheads) there is reduced CrabpI expression. Conversely, there is an increase in signal in the cervical mesenchyme between the first and second branchial arches in

Cyp26b1-/- embryos (Figure 4.9 C, E, dashed circles).

Figure 4.9: Differences in neural crest cell patterning in Cyp26b1 null mutant embryos. Whole-mount in situ hybridization on mid-gestational embryos with neural crest cell markers. (A) In E8.0 embryos, Snail is downregulated in Cyp26b1-/- embryos in comparison with age-matched littermates. (B) The anterior boundary of Sox9 expression is shifted posteriorly in Cyp26b1-/- mutants as indicated by the square brackets. (C-F)

CrabpI expression in E9.5 embryos. In general, there is an upregulation of CrabpI in

Cyp26b1-/- embryos (E, F) in comparison with heterozygous littermates (C, D). There are also apparent differences in expression dorsal to the first branchial arch (C, E) white hashed circle) and in the mesenchyme between the second and third branchial arches.

4.4.7 Cyp26a1 and Cyp26c1 expression in the absence of CYP26B1

Cyp26a1 has been shown to be inducible by RA in vivo, so the expression of both

Cyp26a1 and Cyp26c1 was examined to determine if there is any compensation in the absence of CYP26B1 ((Fujii et al. 1997); G MacLean and M Petkovich, unpublished results). At E9.5, Cyp26a1 is induced in the outgrowing limb bud in Cyp26b1-/- embryos as has been reported (Figure 4.10 A, B, arrows, (Yashiro et al. 2004). No change in

Cyp26a1 expression was observed in the hindbrain or branchial arches at this stage

(Figure 4.10 A, B). Similarly, Cyp26c1 expression is unchanged at E9.5, as transcripts are detected in the hindbrain and maxillo-mandibular cleft of the first branchial arch in both Cyp26b1-/- and Cyp26b1+/+ embryos (Figure 4.10 C, D).

100

Figure 4.10: Cyp26a1 and Cyp26c1 are not upregulated in the hindbrain or neural crest cells of Cyp26b1-/- embryos. Whole-mount in situ hybridization on E9.5 embryos for Cyp26a1

(A, B) and Cyp26c1 (C, D). Cyp26a1 is detected in the outgrowing limb buds of Cyp26b1 null embryos (B, arrow), but not wild-type embryos (A). Cyp26c1 expression is unchanged between wild-type (C) and Cyp26b1-/- embryos (D).

101

102

4.5 Discussion:

RA distribution is established and maintained through the coordinated expression of ALDH and Cyp26 enzymes. This distribution is critical for morphogenesis, as numerous studies have demonstrated that disruption of either family of genes can result in severe embryopathy (Niederreither et al. 1999; Abu-Abed et al. 2001; Sakai et al.

2001; Yashiro et al. 2004). CYP26B1 has previously been shown to be essential in limiting the exposure of the developing limb and testes to RA, thus allowing for proper development (Yashiro et al. 2004; Bowles et al. 2006; MacLean et al. 2007). In this work, we have provided the first detailed analysis of the extensive craniofacial abnormalities associated with Cyp26b1 ablation. Cyp26b1 null animals present many phenotypes reminiscent of RA teratogenicity including cleft palate, micrognanthia, and defects in structures of the inner ear and calvaria; thus CYP26B1 mediated regulation of

RA levels is critical for craniofacial morphogenesis.

4.5.1 Cyp26b1 is dispensable for molecular patterning of the hindbrain

Early studies in chick and frog proposed that RA acted as a posteriorizing factor in the developing hindbrain, and that the anterior hindbrain had to remain in an RA- depleted state (Glover et al. 2006). RA is detected in the presumptive mouse hindbrain with an anterior limit at pr2/3 at E7.6 (Sirbu et al. 2005). As the embryo develops, the anterior boundary of RA retreats posteriorly, until E9.25, when RA extends up to r5 (Mic et al. 2002). Cyp26b1 is the last of the Cyp26 genes to be expressed in the hindbrain, but from E9.0 onwards, it is the predominant Cyp26 member in this region with expression

103 spanning r2-r6 (MacLean et al. 2001). Despite the high levels of Cyp26b1 expression, through analysis of hindbrain markers, we have shown that r2-r6 are essentially normal in

Cyp26b1-/- embryos (Figure 4.7). The only abnormalities detected were in the expression of Krox-20 and Kriesler. Both markers indicated that r5 and r6 were intact and the correct size, but there appeared to be novel expression of both genes posterior and ventral to the last rhombomere. NCCs from r6 normally migrate into the third branchial arch and contribute to the parathyroid gland and thymus, tissues that appear normal in Cyp26b1-/- embryos (Kaufman and Bard J. B. 1999). However, the murine hindbrain develops into the cerebellum, pons and medulla oblongata (Kaufman and Bard J. B. 1999). These structures have not been analyzed in older embryos, hence the subtle differences observed in Krox-20 and Kriesler expression may contribute to phenotypes evident later in brain development.

The lack of hindbrain patterning abnormalities in Cyp26b1-/- mice is in contrast to

Cyp26a1-/- embryos, which display minor changes in hindbrain formation, characterized by an enlarged r4 and a partial transformation of r2 into r4 (Abu-Abed et al. 2001).

Cyp26c1-/- mice show no hindbrain defects, but Cyp26a1/Cyp26c1 compound knockout mice exhibit phenotypes more severe than Cyp26a1-/- embryos, as r4 is markedly larger, and there are patterning abnormalities in r1, r2 and r3 (Uehara et al. 2007). Cyp26a1 and

Cyp26c1 are both expressed prior to Cyp26b1 in the presumptive hindbrain, with

Cyp26a1 being detected in the hindbrain with a posterior boundary of presumptive pr 2 at

E7.75 (Sirbu et al. 2005). At this time, RA has an anterior boundary of pr2, until E8.0 when RA is cleared from pr4, presumably due to the activity of CYP26C1. Cyp26b1 is

104 first expressed at E8.0 in r3 and r5, and at this stage, RA has an anterior boundary of r5

(MacLean et al. 2001; Sirbu et al. 2005). Shortly afterwards, Cyp26b1 becomes the predominant Cyp26 gene detected in the hindbrain. Based on observations from Cyp26a1 and Cyp26a1/Cyp26c1 knockout models, it appears that the activity of CYP26A1 (and to a lesser degree CYP26C1) is important for early hindbrain patterning, while our results indicate that CYP26B1 is dispensable for hindbrain development (Abu-Abed et al. 2001;

Uehara et al. 2007). Although there does not appear to be any compensation by Cyp26a1 or Cyp26c1 in the hindbrain of Cyp26b1-/- embryos at E9.5 (Figure 4.10), there may be

Cyp26 induction at an earlier time point. It is also possible that CYP26C1 is adequate in establishing a boundary for RA in the hindbrain, as there is some overlap in expression between Cyp26c1 and Cyp26b1 (Tahayato et al. 2003). Furthermore, by the time

Cyp26b1 is expressed, Aldh1a2 expression, and consequentially RA may have retreated far enough posteriorly that RA is unable to diffuse into the anterior hindbrain of

Cyp26b1-/- embryos.

The lack of a striking hindbrain phenotype in Cyp26b1 null mutants is surprising, considering the expansive expression of Cyp26b1 which is largely conserved across species including mouse, chick, and zebrafish (Reijntjes et al. 2003; Reijntjes et al. 2005;

Hernandez et al. 2007). The hindbrain is recognized as being particularly sensitive to high levels of exogenous RA; therefore CYP26B1 may be required to protect the developing hindbrain from maternal influxes of retinol or RA. This postulate can be tested by treating pregnant females with a sub-teratogenic dose of RA to observe if

105

Cyp26b1-/- embryos are more sensitive to hindbrain malformations induced by ectopic

RA exposure.

4.5.2 Cyp26b1 and Neural Crest Cell Migration

Some NCCs migrate out of rhombomeres 2, 4 and 6 and populate the branchial arches (BA), and other regions in the skull . Despite the lack of any striking defects in hindbrain patterning, markers for NCCs and BAs were analyzed, as many of the affected structures in Cyp26b1-/- embryos are NCC-derived tissues. Expression of the early migrating NCC markers Snail and Sox9 appear to be downregulated in the absence of

CYP26B1 (Figure 4.9). Both Snail-/- and Sox9 -/- mutants die early during gestation thus preventing detailed analysis of NCC derived structures (Bi et al. 2001; Carver et al.

2001). However, conditional knockout models utilizing a Wnt1-Cre line that selectively excises floxed alleles in migrating NCCs resulted in mice exhibiting some common phenotypes with Cyp26b1 null embryos. Sox9 conditional knockout animals have a cleft secondary palate, a shortened mandible, lack Meckel‟s cartilage and other endochondral bones of the skull (Mori-Akiyama et al. 2003). Snail conditional knockout mutants are phenotypically normal, but when derived on a genetic background of Slug (Snai2)-/- mice, offspring have a cleft palate, a shortened mandible and lack calvarial fusion similar to

Cyp26b1-/- embryos (Murray et al. 2007). Although the Sox9 and Snail conditional mutants exhibit some phenotypic similarities with Cyp26b1 knockout animals, there are only very subtle changes in the expression of these genes in the absence of CYP26B1.

106

Therefore, the phenotypes in Cyp26b1-/- embryos may be due to changes in expression of downstream targets of either Sox9 or Snail.

4.5.3 Branchial arch patterning in the absence of CYP26B1

The branchial arches are sensitive to perturbations in endogenous retinoid status, as maternal RA deficiency can result in defects in formation of posterior arches (BA3-6), while exogenous RA can result in abnormal development of BA1 and 2 (Mark et al.

2004). BA1 and 2 contribute to numerous structures including the maxilla, mandible, hyoid, Meckel‟s cartilage and the bones of the inner ear; elements that all compromised in Cyp26b1-/- embryos. Many of these phenotypes are similar to those observed in the human microdeletion disorder, 22q11.2 deletion syndrome (22qllDS) (Scambler 2000).

DiGeorge syndrome is associated with 22q11DS, and the T-box transcription factor Tbx1 has been identified as a candidate gene for pathogenesis of the syndrome. Tbx1-/- mice are unable to inflate their lungs and die shortly after birth. Null mutants present with a cleft palate, reduced hyoid, abnormal ears, a short mandible and heart defects (Jerome and Papaioannou 2001). Tbx1 has been proposed as a regulator of Cyp26 expression, as

Cyp26a1, Cyp26b1 and Cyp26c1 are all downregulated in Tbx1 null embryos, and abnormal RA-mediated signaling may contribute to the observed phenotypes (Roberts et al. 2006). However, Tbx1 is downregulated by exogenous RA in chick and zebrafish; which corroborates our observation of Tbx1 loss in BA1 of Cyp26b1-/- embryos (Figure

4.7) (Roberts et al. 2006; Zhang et al. 2006). Furthermore, in vitamin-A deficient quail, or mice with a hypomorphic allele for Aldh1a2, there is a posterior expansion of Tbx1

107 expression, indicating TBX1 may regulate its own expression by limiting RA levels through induction of Cyp26 expression (Vermot et al. 2003; Roberts et al. 2005).

Although Tbx1 was downregulated in Cyp26b1-/- embryos, no significant abnormalities were observed in expression of other BA1 markers that were examined.

Fgf8 expression was unchanged in E9.5 and E10.5 embryos, while Bmp4 appeared downregulated at E9.5, but not E10.5 (Figure 4.7, E, F; results not shown).

The Dlx genes encode homeobox transcription factors that are involved in branchial arch development (Depew et al. 2005). In mice, there are 6 Dlx genes arranged as convergently transcribed bigene pairs, with Dlx1/2 and Dlx5/6 are expressed in the first branchial arch. It has been observed that various Dlx knockout models share some craniofacial phenotypes with Cyp26b1-/- embryos suggesting a mechanistic link between retinoid status and Dlx expression (Depew et al. 2005). Furthermore, it has also been shown in RA-treated embryos that Dlx expression can be downregulated in BA1 by RA, and it was subsequently proposed that altered signaling from branchial epithelial was responsible for retinoid-induced craniofacial malformations (Vieux-Rochas et al. 2007).

Considering these results, it was unexpected that expression of Dlx1, 2, 5 and 6 was unaltered in Cyp26b1-/- embryos (Figure 4.8; results not shown). A possible explanation for these discrepancies is that in the pharmacological study, RA was administered as a single dose, and would therefore only be exposed to tissues for a brief time prior to clearance. In Cyp26b1 null mutants, elevated RA levels would be sustained within a tissue over an extended time frame, especially since there does not appear to be any compensation by either Cyp26a1 or Cyp26c1 (Figure 4.10). It is well documented that 108 the duration and timing of exposure to ectopic RA has significant effects fetal malformations (Kochhar 1973). Furthermore, the RA administered is likely in excess of endogenous RA levels, even in the absence of CYP26B1; therefore, Dlx genes may only be affected when the developing embryo is exposed to an acute, teratogenic dose of RA.

However, the similarity of phenotypes between Dlx compound mutant mice and

Cyp26b1-/- embryos is intriguing, and may indicate that expression of downstream targets of DLX proteins are affected by elevated levels of RA.

4.5.4 Craniofacial malformations in Cyp26b1-/- embryos are due to disruptions in multiple signaling pathways

Cyp26b1-/- mutant mice have numerous craniofacial abnormalities, many of which have been observed in models of RA teratogenicity. Considering the high expression of

Cyp26b1 in the hindbrain, and the importance of this tissue in NCC formation, we originally hypothesized that the observed craniofacial malformations would be due to irregular NCC development or migration. A comprehensive analysis of the hindbrain, branchial arches and branchial pouches has revealed these structures to appear morphologically normal with only subtle changes in tissue patterning in the early embryo. Furthermore, NCCs appear to develop normally, but there are some anomalies in migration as shown by changes in CrabpI and Sox9 expression (Figure 4.9). The presented results indicate the craniofacial malformations in Cyp26b1 null mice are likely due to disruptions in multiple signaling pathways that are independently affected by RA.

109

The lack of molecular defects in many of the markers analyzed indicates some of the observed malformations may be due to changes in local RA distribution later during development. One tissue where this could occur the nasal cavity, which is deformed in

Cyp26b1-/- mutants, and also a site of Aldh1a1, Aldh1a2 and Aldh1a3 expression, while

Cyp26b1 is detected in the nasal process (Niederreither et al. 2002b). CYP26B1 activity in this region may regulate the local distribution of RA, allowing for nasal morphogenesis after branchial arches have extended, and begun to differentiate into specific tissues. Further studies with Cyp26b1-/- mice will examine molecular markers in older embryos to determine the role of retinoid signaling in late forming morphogenetic fields.

110

Chapter 5

The Retinoic Acid Metabolizing Enzyme CYP26B1 is required for -Catenin mediated induction of embryonic hair placodes

111

5.1 Summary Retinoic acid (RA) is a signaling molecule that regulates cell proliferation, differentiation and apoptosis, and along with other retinoid-based molecules is used therapeutically as a treatment for numerous skin conditions. The developing embryo exhibits a tightly regulated distribution of RA, and disruption of endogenous RA levels by genetic, dietary or pharmacological means can result in severe embryopathy. The coordinated expression of RA-synthesizing (retinaldehyde dehydrogenase) and RA- catabolizing (CYP26) enzyme defines the spatio-temporal specific distribution of RA in vivo, and is a key regulatory mechanism in control of RA-mediated signaling. We have generated a line of transgenic mice lacking the RA-degrading enzyme CYP26B1, and observed numerous developmental deficiencies in mutant embryos. In addition to limb, craniofacial and gonad abnormalities, we show here that Cyp26b1-/- embryos exhibit a striking reduction in vibrissal and pelage follicles. Most follicles appear to terminate prior to placode formation, and any follicles that do form are abnormally spaced.

Molecular analysis of vibrissal rudiments in embryonic day (E) 12.5 embryos reveals a downregulation of -catenin, and its downstream targets Shh and Bmp4. In addition to follicular defects, Cyp26b1-/- embryos display epidermal stromal abnormalities, as cells in this region are disordered and not arranged in ordered layers as in wild-type embryos; this may contribute to the poorly anchored skin observed in CYP26B1 null mutants.

112

5.2 Introduction The epidermis and associated appendages; hair follicles, sweat glands, sebaceous glands and nails all form a protective barrier that is essential for survival. The skin develops from a sheet of neuroectoderm that will differentiate appropriately in response to morphogenetic cues (Fuchs 2007). Of particular interest is the formation of hair follicles as they contain a population of stem cells contained within a well defined niche

(Moore and Lemischka 2006). Study of the molecular signaling pathways underlying hair follicle stem cell survival and proliferation is important for understanding embryonic patterning, and holds potential for stem cell based therapeutics.

The epidermis does not form as a homogenous cell population; rather cells are multipotent and can form either appendages or epidermis. A delicate balance between appendage stimulating and appendage inhibiting signaling pathways determines epidermal cell fate (Fuchs 2007). The interplay between these two pathways results in the spacing of hair follicles within the epidermis. Mammalian hair follicle development begins with formation of a hair placode consisting of enlarged keratinocytes that bunch together. Inductive molecular signals drive proliferation within the placode, resulting in downwards growth and formation of the hair germ. As downwards growth continues, keratinocytes differentiate forming the inner root sheath, which will be surrounded by layers of outer root sheath cells (Schmidt-Ullrich and Paus 2005). Once the inner and outer root sheaths have formed, the follicle is prepared to grow hair shafts. Mice develop two major types of hair follicles, the vibrissal (whisker) follicles which develop first at embryonic day (E) 12, and the pelage follicles which begin to form at E15 (Schmidt-

113

Ullrich and Paus 2005). In adult mice, hair follicles undergo continuous cycles of renewal consisting of four main stages: growth (anagen), regression (catagen), rest

(telogen) and release (exogen).

A balance of inductive and repressive pathways dictates hair follicle formation.

In mice, Wnt mediated stabilization of -catenin is necessary for follicle induction.

When -catenin is conditionally knocked out in skin, there is a lack of placode formation, and hair cycling also terminates in adult animals (Huelsken et al. 2001). Similarly, when the Wnt inhibitor Dickkopf 1 (Dkk-1) is ectopically expressed in skin there is a downregulation of Wnt10b and -catenin, and a loss of placodes (Andl et al. 2002). Dkk-

1 is normally expressed in the mesenchyme surrounding hair follicles and is believed to limit follicle formation mediated by Wnt/ -catenin signaling (Andl et al. 2002; Schmidt-

Ullrich and Paus 2005).

The association between retinoids and skin homeostasis is well characterized.

Vitamin A deficiency can result in abnormal keratinization of the epidermis, and retinoid- based therapeutics are used to treat skin conditions such as acne, psoriasis and photo- damage (Wolbach and Howe 1978; Lee and Koo 2005; Singh and Griffiths 2006).

Retinoic acid (RA), the most biologically active retinoid, is a ligand for the nuclear retinoic acid receptor (RAR) (Chambon 1996). RAR forms functional heterodimers with the retinoid X receptor (RXR), which then bind to specific sites in DNA known as retinoic acid response elements (RAREs). Upon ligand binding, RAR/RXR recruit activator proteins that promote transcription of retinoid responsive genes (Chambon

1996; Dilworth and Chambon 2001).

114

RA is synthesized by retinaldehyde dehydrogenase (ALDH) enzymes, and catabolized by Cyp26 enzymes, a family of cytochrome P450 proteins that specifically oxidize RA as part of an inactivation pathway (Niederreither and Dolle 2006;

Pennimpede et al. 2006). RA can regulate cell proliferation, apoptosis and differentiation; hence the proper distribution of RA is critical for morphogenesis. The coordinated expression of Aldh and Cyp26 genes defines the spatiotemporal distribution of RA in tissues, helping to define proliferating and differentiating regions in the embryo.

Disruption of the endogenous distribution of RA has significant effects on the developing embryo, as Aldh1a2, Cyp26a1 or Cyp26b1 null mutant mice all die during gestation, or just after birth with severe malformations (Niederreither et al. 1999; Abu-Abed et al.

2001; Sakai et al. 2001).

The expression of RA metabolizing enzymes is dynamic during morphogenesis, as gene expression must ultimately meet tissue requirements for cells to be in a either a proliferating or differentiating state. An example of this is the adult mouse hair follicle where Aldh1a1, Aldh1a2 and Aldh1a3 expression changes depending on the stage in the hair cycle indicating RA may regulate the progression from one stage in the hair cycle to the next (Everts et al. 2004; Everts et al. 2007). The localization of Cyp26 genes has not been studied in mature mouse skin, although one of the genes, Cyp26b1 has been detected in the dermis surrounding developing hair follicles in the embryo (Abu-Abed,

2002). Aldh1a1-3 are not detected in embryonic skin (Niederreither et al. 2002b), but two of the nuclear receptors, RAR and RAR are expressed in vibrissal hair placodes

(Viallet and Dhouailly 1994). Conditional ablation of RAR and RAR in epidermis had

115 no effect on keratinocyte renewal, raising questions regarding the function of these receptors in skin with endogenous levels of RA (Metzger et al. 2003).

We have developed a line of transgenic mice that lack Cyp26b1. These embryos die shortly after birth with severe malformations affecting the limb, testes and craniofacial structures (Yashiro et al. 2004; Bowles et al. 2006; MacLean et al. 2007). It was also observed that Cyp26b1 null embryos appear to lack essentially all whiskers and have improperly anchored skin (Figure 4.3). We have further characterized these observations and noted abnormalities in epidermal and hair follicle formation in these animals.

116

5.3 Materials and Methods 5.3.1 Collection of Cyp26b1-/- embryos Cyp26b1-/- embryos were generated and genotyped as previously described

(MacLean et al. 2007), Chapter 4). Timed matings were established by housing female mice with males overnight. Females were checked for vaginal plugs, and noon of the day that a plug appeared was denoted as E0.5. On the appropriate day, pregnant females were humanely euthanized by cervical dislocation and embryos collected in ice-cold phosphate-buffered saline.

All animal experiments were reviewed and approved by the University‟s Animal

Care and Use Committee.

5.3.2 Embryo fixation and histology

Embryos were collected and fixed overnight in either Bouin‟s solution (Figure

5.3) or 10% buffered formalin (Figures 5.1, 5.2), before being transferred to 70% ethanol.

Embryos were then embedded in paraffin in the appropriate orientation and cut into 5 M sections with a microtome. For general morphology, sections were stained by a standard protocol with Harris‟ Hematoxylin and 0.5% Eosin.

Immunohistochemistry for Desmin and Muscle Specific Actin (MSA) was performed at the core pathology laboratory, Queen‟s University.

117

5.3.3 Whole-mount in situ hybridization

Embryos were fixed overnight in 4% paraformaldehyde at 4 C, and then whole- mount in situ hybridization was performed with digoxigenin (DIG) labeled anti-sense riboprobes as previously described (Decimo et al. 1995; Niederreither et al. 1997). All riboprobes were synthesized using a DIG labeling kit, and the appropriate RNA polymerase (Roche). For each probe, embryos of each genotype (+/+, +/-, -/-) were processed concurrently, and the colour reaction was started at the same time, thus the duration of staining for each genetic group was the same for any given probe.

5.3.4 Skin Permeability Assay

Fresh, unfixed E18.5 embryos were collected and washed briefly in phosphate buffered saline. Embryos were then placed separately in 10 mL of X-gal staining buffer consisting of 100mM NaPO4, 1.3 mM MgCl2, 3 mM K3Fe(CN)6, 3 mM K4Fe(CN)6 and

1 mg/mL X-gal at pH 4.5. Samples were left overnight in the dark at room temperature.

118

5.4 Results 5.4.1 Epidermal abnormalities in Cyp26b1 -/- embryos Cyp26b1-/- mice die perinatally and exhibit malformations of the limb, testes, skull and face as has been previously described. Mutant embryos collected at E18.5 also have improperly anchored skin, a distended abdomen and lack whiskers (Figure 4.3).

Histological analysis was performed to further characterize any epidermal or stromal abnormalities in the absence of CYP26B1. Transverse sections of E14.5 embryos stained with Harris Hematoxylin (H) and Eosin (E), show that Cyp26b1-/- embryos have a number of epidermal malformations. The trunk epidermis at this stage is characterized by periderm, which overlays an ordered stratum intermedium (SI). The basal layer of the

SI consists of a continuous layer of cells forming an ordered boundary from the stratum basale (SB) (Figure 5.1 A, C). In Cyp26b1-/- embryos it is difficult to differentiate the SB from the SI, as the stroma is not organized into discrete layers (Figure 5.1 B, D).

Furthermore, the stromal cells appear spaced further apart in CYP26B1 null embryos, possibly indicating an increase in ground substance. Sections were stained with Alcian

Blue to determine if there was increased deposition of glycosaminoglycans (results not shown), but results were inconclusive, possibly due to the fixative (10% buffered formalin) used in tissue preparation.

The periderm normally forms as a discrete ordered layer, but this is not always observed in Cyp26b1-/- mutants (Figure 5.1 E, F). A representative abnormal region is shown, where the periderm is disorganized and there does not appear to be a distinct basal layer separating the SI from the SB. This phenotype is not consistent throughout

119 the entire trunk epidermis of Cyp26b1-/- embryos, but was not observed in wild-type littermates.

An assay was performed to determine if the disorganized periderm affected skin permeability in Cyp26b1-/- embryos. Skin contains endogenous -galactosidases which are active at low pH. By E18.5, embryos have formed an epidermis that is impermeable to X-gal, therefore when embryos are placed in the appropriate solution, no staining will be observed unless X-gal is able to permeate the skin (Turksen and Troy 2002). No differences in X-gal staining was observed between Cyp26b1+/+ and Cyp26b1-/- embryos, although as a positive control, staining was observed in the umbilical cord (results not shown).

120

Figure 5.1 Epidermal and Stromal abnormalities in Cyp26b1-/- embryos. H and E stained transverse sections of E14.5 embryos. Wild-type embryos (A, C) have several stromal layers, while Cyp26b1-/- embryos (B, D) possess an unorganized stroma (red arrows). (E, F) High magnification views of the periderm showing that the basal layer of the stratum intermedium forms a distinct boundary separating the stratum basale in wild- type embryos (E, yellow arrows) but not Cyp26b1-/- embryos (F). (H) Heart; (L) Lung;

(P) Periderm; (SB) Stratum Basale; (SC) Spinal Cord; (SI) Stratum Intermedium. Scale bars=100 M

121

122

5.4.2 CYP26B1 is required for proper muscle formation

In order to determine if the distended abdomen in Cyp26b1 null mutants was related to a lack of muscle formation, immunohistochemistry (IHC) on E14.5 transverse sections was performed with antibodies to Desmin and Muscle-Specific Actin (MSA).

The Desmin stained sections show that in Cyp26b1-/- embryos there is an accumulation of material in between layers of muscle (Figure 5.2 A, B). There also appear to be reductions in dorsal muscle flanking the spinal cord, and lateral muscle of the body wall.

Similar results were observed on sections stained for MSA (Figure 5.2 C-F). The MSA stained sections also clearly show that panniculus carnosus, which normally forms as an ordered layer of cells, is disorganized in Cyp26b1-/- embryos, possibly contributing to the distended appearance of mutants.

It should also be noted that although the plane of these sections is the same through the heart and lungs, the liver is clearly visible in Cyp26b1 null mice, prior to the stage at which it is visible in wild-type littermates (Figure 5.2 C, D). This phenotype is currently being analyzed for significance.

123

Figure 5.2 Cyp26b1-/- embryos have deficiencies in muscle formation. Immuno- histochemistry for Desmin (A, B) and Muscle Specific Actin (C-F) shows that there is reduced dorsal and lateral muscle formation in Cyp26b1-/- embryos (B, D). High magnification images also show that the panniculus carnosus forms irregularly in

Cyp26b1-/- mice (F) in comparison with wild-type littermates. (H) Heart; (L) Lung; (Li)

Liver; (PC) Panniculus Carnosus; (SC) Spinal Cord. Scale bars=100 M.

124

125

5.4.3 Lack of hair follicles in Cyp26b1-/- embryos

Gross analysis of E18.5 embryos revealed that Cyp26b1-/- mice lack whisker follicles readily evident in wild-type littermates (results not shown). This phenotype was investigated further with sagittal sections of E14.5 embryos. By E14.5 in wild-type embryos, whisker follicles have clearly formed in distinct rows that will later give rise to mystacial and submental vibrissae (Figure 5.3 A)(Yamakado and Yohro 1979). In contrast, Cyp26b1-/- embryos lack an ordered array and develop only a few irregularly spaced follicles (Figure 5.3 B).

Pelage follicles develop later than vibrissal follicles, but are clearly visible as hair placodes in E14.5 wild-type mice (Figure 5.3 C, E). However, in Cyp26b1-/- embryos, well defined placodes are not evident in the trunk epidermis, although there are some regions where the epidermis is slightly thickened (Figure 5.3 D, F).

126

Figure 5.3 Lack of vibrissal and pelage follicle formation in Cyp26b1-/- embryos. H and E stained sagittal sections of E14.5 embryos. (A) Wild-type embryos possess several distinct rows of vibrissal follicles, while Cyp26b1-/- embryos (B) only have a few irregularly spaced follicles. In the body wall, wild-type mice (C, E, arrowheads) have numerous developing hair placodes, while similar structures are not observed in

Cyp26b1-/- littermates (D, F). Scale bars=100 M.

127

128

5.4.4 Molecular defects in vibrissal follicles of Cyp26b1-/- mice

Whisker follicles begin to develop at E11.5, and sinus hair rudiments can be detected at E12.5. In situ hybridization was used to examine molecular pathways in initiation of vibrissal follicle formation. In E12.5 wild-type embryos, Sonic Hedgehog

(Shh) is expressed in five rows that will form the mystacial follicles (Figure 5.4 A).

However, in Cyp26b1-/- embryos, Shh expression is reduced to a few irregularly spaced follicles (Figure 5.4 B). Similarly, -catenin normally expressed in developing follicles of E12.5 embryos, is essentially absent in Cyp26b1-/- embryos (Figure 5.4 C, D). Bone morphogenetic protein 4 (BMP4), another marker of vibrissal follicles is clearly detected in wild-type embryos, but follicle expression is reduced in Cyp26b1-/- littermates (Figure

5.4 E, F).

129

Figure 5.4 Molecular abnormalities in follicles of Cyp26b1 -/- embryos. Whole mount in situ hybridization performed on E12.5 embryos (A-D). In wild-type mice (A), shh is detected in vibrissal hair rudiments, while similar expression is not observed in

Cyp26b1 -/- embryos (B). -catenin is also normally expressed in presumptive whisker follicles (C), but not in Cyp26b1-/- littermates (D). Bmp4 expression is also reduced in

Cyp26b1-/- mice in comparison with wild-type embryos (E, F).

130

131

5.5 Discussion

RA is a potent signaling molecule able to regulate the cellular processes of proliferation, differentiation and apoptosis. In the developing embryo, the endogenous distribution of RA is such that some regions are rich in RA, while others are devoid of

RA. Gene targeting studies have shown that maintenance of this distribution is critical for morphogenesis as removal of RA-synthesizing (ALDH) or RA-catabolizing (Cyp26) enzymes can result in embryopathy reminiscent of either maternal vitamin A deficiency or RA teratogenesis (Niederreither et al. 1999; Abu-Abed et al. 2001; Sakai et al. 2001;

Yashiro et al. 2004). These genetic studies have shown the function of Cyp26 enzymes is to define boundaries for RA as a means of regulating expression of retinoid-responsive genes. It has already been shown that Cyp26a1 is required to prevent caudal abnormalities, while Cyp26b1 is essential for limb, male gonad and craniofacial development (Abu-Abed et al. 2001; Sakai et al. 2001; Yashiro et al. 2004; Bowles et al.

2006; MacLean et al. 2007). The current study expands our knowledge of retinoid signaling by demonstrating a role for CYP26B1 in hair follicle development, and epidermal maturation.

5.5.1 Cyp26b1 is required for induction of hair follicles

RA signaling has been implicated in formation of hair, as the distribution of the

RARs and Aldhs changes in the hair follicle depending on the stage of hair cycle (Everts et al. 2004; Everts et al. 2007). However, little is known regarding the role of retinoid signaling during embryonic follicle formation. The only Cyp26 gene detected in

132 embryonic skin is Cyp26b1, with expression in the dermis surrounding hair follicles

(Abu-Abed et al. 2002). Aldh1a1-3, are not detected in embryonic skin, although RAR and RAR are expressed in hair placodes (Viallet and Dhouailly 1994; Niederreither et al.

2002b). It is unclear if RA is present in embryonic epidermis. A transgenic mouse line containing multiple RAREs upstream of a LacZ gene (RAREhsplacZ), is commonly used as a reporter model for detection of RA in vivo (Rossant et al. 1991). In these animals, at

E14.5 there is reporter activation in the whisker pad, and diffuse staining throughout the body (Rossant et al. 1991). However, there does not appear to be discrete localization of

RA either within or surrounding hair follicles. It is still possible for RA to be in the hair follicles but at a level below the detection limit of the model, or analysis of embryos at this age could reflect a limitation of these mice, as the RARElacZ transgene quickly loses activity during development (P.Dollé, personal communication). Therefore, there may be low levels of epidermal RA that are required for RAR and RAR signaling, or unliganded RARs may be required for follicle morphogenesis, as has been demonstrated in other experimental models (Koide et al. 2001; Weston et al. 2003). Regardless,

Cyp26b1 expression surrounding hair follicles is quite strong, suggesting a function in either keeping RA within the follicle, or preventing it from diffusing into the follicle

(Abu-Abed et al. 2002).

When Cyp26b1 is ablated, there is a striking loss of both vibrissal and pelage follicles. Histological analysis of E14.5 embryos showed that Cyp26b1-/- mutants had only a few irregularly spaced vibrissal follicles, in contrast to wild-type littermates, which developed several ordered rows of follicles (Figure 5.3 A, B). Similarly, pelage follicles,

133 which are visible as hair placodes at E14.5 were greatly reduced in E14.5 Cyp26b1-/- mice (Figure 5.3 C-F). Visual analysis of E18.5 embryos revealed that whiskers are absent in Cyp26b1-/- animals, hence we believe that vibrissal follicle formation terminates at the hair placode stage, rather than there being a delay in the process. Histological analysis of younger embryos is being performed to determine at what stage vibrissal follicle formation terminates, and older embryos are also being analyzed to ensure that development of pelage follicles is arrested, and not delayed.

5.5.2 CYP26B1 is required for -catenin mediated initiation of vibrissal placode formation

In order to elucidate the molecular mechanisms underlying hair follicle loss in

Cyp26b1-/- embryos, we analyzed several molecular markers of hair follicle formation.

Sonic hedgehog (Shh) is a secreted signaling peptide that has been implicated in numerous signaling pathways. In hair follicles, Shh is not required for placode induction, but is necessary for placode progression, as follicles in Shh null embryos arrest shortly after induction (St-Jacques et al. 1998). At E12.5, Shh is normally detected in vibrissal follicles, which are visualized as distinct rows of follicle rudiments. However, in

Cyp26b1-/- animals, Shh in the nasal process is only detected in a few irregularly spaced patches (Figure 5.4 A, B). Shh expression can be affected by retinoids depending on the experimental context, so other markers were analyzed to determine if the observations were specific to Shh expression, or indicative of a problem with structure formation.

134

Bmp4 has been proposed as a repressive signaling factor that is involved in proper follicle spacing, as conditional deletion of the BMP receptor (Bmpr1a) in epidermis results in an increased density of embryonic hair follicles (Andl et al. 2004). However, it does not appear that induction of ectopic Bmp4 is involved in development of the phenotype in Cyp26b1-/- embryos, as Bmp4 expression was downregulated in vibrissal follicles (Figure 5.4 E, F). It was also noted that Bmp4 expression was unaffected in nasal process in mutant embryos, which indicates that the defect in signaling is upstream of Bmp4.

-catenin is a component of the Wnt signaling pathway that is expressed in the embryonic and adult hair follicles (Ridanpaa et al. 2001). Conditional knockout of - catenin in epidermis results in a loss of hair follicles, and downregulation of Shh and

Bmp4 in follicles, demonstrating the -catenin signaling is upstream of Shh and Bmp4

(Huelsken et al. 2001). Similar to results with Shh, at E12.5 there is a striking loss of - catenin expression in vibrissal rudiments in Cyp26b1-/- embryos. -catenin is required for early placode formation, therefore it appears that in Cyp26b1 null embryos there is a complete failure of vibrissal placode formation.

Wnt signaling is essential for follicle formation, as ectopic expression of the Wnt inhibitor dickkopf (Dkk-1) completely inhibits placode formation, possibly by blocking upregulation of Wnt10b in placodes (Andl et al. 2002). In the epidermis, Dkk-1 is normally expressed in mesenchyme surrounding follicles, and thus may define follicle spacing by limiting where Wnt signaling occurs. Wnt10b (and possibly other Wnt members) are expressed in the hair follicle and stabilize -catenin, which initiates a 135 signaling cascade favouring placode formation (Figure 5.5), and eventual follicle formation. Links between RA and Wnt genes have been demonstrated in the tailbud where analysis of Cyp26a1-/- RAR -/- led to the proposal that exogenous RA downregulates Wnt3a, possibly contributing to caudal malformations observed in

Cyp26a1-/- embryos (Abu-Abed et al. 2003). Furthermore, it has been demonstrated in vitro that RA can increase Dkk-1 expression in cultured cells (Verani et al. 2007). One explanation for the lack of follicles in Cyp26b1-/- embryos is that in the absence of

CYP26B1, there is an increase in Dkk-1 expression. This results in downregulation of

Wnt10b and consequentially lower -catenin levels, which culminates in an inhibition of placode formation. Immunohistochemistry with antibodies to Dkk-1 have been performed on Cyp26b1-/- embryos with inconclusive results.

Another signaling molecule that may be deregulated in Cyp26b1-/- hair follicles is

Fgf10. Fgf10 null mice have reduced numbers of whiskers, along with open eyes at birth and a cleft palate; phenotypes that are all observed in Cyp26b1-/- embryos (Ohuchi et al.

2003; Alappat et al. 2005; Tao et al. 2006). Analysis of Fgf10-/- embryos determined that

Fgf10 is upstream of Shh, as it is not required for induction of Shh expression, but is required for correct localization of Shh (Ohuchi et al. 2003). It has been suggested that

Wnt signaling regulates FGF10 in other tissues, but this effect has not been examined in the hair follicle (Li et al. 2005; Cohen et al. 2007). Future experiments in Cyp26b1 null embryos will clarify the relationship between RA, Fgf10, Dkk-1 and Wnt10b (Figure 5.5).

136

Figure 5.5 Molecular Pathways in vibrissal hair placode formation. (A) In wild-type embryos, Wnt signaling results in -catenin stabilization, thus facilitating expression of Shh and

Bmp4. Shh is required for hair placode formation, while Bmp4 limits follicle growth. Fgf10 is also required for follicle formation, but it is unclear if FGF10 acts up or downstream of -

Catenin. (B) In Cyp26b1-/- embryos, there is downregulation of Bmp4 and Shh as a result of a lack of -catenin. We postulate that in the absence of CYP26B1, levels of RA increase in the hair follicle and either induce Dkk-1 or downregulate Wnt10b directly, thus inhibiting hair placode formation.

137

138

5.5.3 Epidermal abnormalities in the absence of CYP26B1

In addition to the lack of hair follicles, there are other irregularities in epidermal development in Cyp26b1-/- embryos. In E14.5 embryos, there are normally distinct layers of tightly packed cells that can be visualized in the stroma (Figure 5.1 A, C). However, in Cyp26b1-/- mutants there is a lack of any discernable layers, and there appear to be gaps between cells (Figure 5.1 B, D). Analysis of H and E stained sections indicates this could be due to increased ground substance as a result of mucin accumulation in

Cyp26b1-/- embryos (V. Tron, personal communication). To confirm this, sections were stained with 2.5% alcian blue, which is often used to visualize mucin; however, results were inconclusive. Samples were fixed in 10% buffered formalin, which can wash away peptidoglycans in some tissues (R. Kisilvesky, personal communication), so these analyses are being repeated with tissues fixed in ethanol/acetic acid.

Close examination of the epidermis of Cyp26b1-/- embryos revealed that there are some regions where the epidermal basal layer does not develop properly (Figure 5.1 E,

F). A skin permeability assay was conducted on E18.5 embryos, and no differences were observed between Cyp26b1-/- embryos and wild-type littermates. It is possible that

Cyp26b1-/- embryos may exhibit permeability deficiencies is post-natal life, but conditional knockout mice will have to be generated to answer this question.

IHC performed with antibodies to desmin and muscle specific actin revealed that

Cyp26b1-/- embryos have deficiencies in trunk muscle development that may contribute to the distended abdomen that is observed in older mutant embryos (Figure 4.3). In mutant embryos, the panniculus carnosus is disordered, and there also appears to be

139 reduced formation of dorsal and flank muscle (Figure 5.2). Skeletal muscle arises from somites, as the dermomyotome forms the body wall musculature. Muscle progenitor cells, known as myoblasts follow a differentiation pathway that is regulated by transcription factors including Myf5, MyoD and Myogenin (Brent and Tabin 2002).

Cyp26b1 is expressed intersomitically at E8.5 (Figure 3.3), therefore we will investigate various markers of myoblast differentiation to observe is this process is affected in the absence of CYP26B1.

These experiments clearly show that Cyp26b1 plays an important role in epidermal development, and particularly in hair follicle formation. It will be interesting to study the function of Cyp26b1 in adult hair follicles. There are regions in the follicle that presumably lack RA, hence Cyp26b1 expression in these regions may be involved in maintaining cells in a specific state of differentiation depending on the stage of the hair cycle (Everts et al. 2007). By waxing mice to induce anagen, and then treating at various stages with Cyp26 inhibitors, we may be able to determine the role of Cyp26b1 in hair cycling. Considering that Cyp26b1 is also detected hippocampus, which is a site of neurogenesis in adult rodents, these experiments may help define the role of retinoid metabolism in stem cell survival.

140

Chapter 6

General Discussion

141

Embryonic patterning is a complex process in which a multitude of signaling molecules, or morphogens dictate tissue fate through regulation of cell proliferation, differentiation, migration and apoptosis. Many of the signaling pathways relevant for embryonic development are also involved in disease progression, hence research directed towards understanding the molecular physiology driving these process will likely yield strategies for developing novel therapeutics.

Several morphogen signaling pathways have been characterized, and many common features have been identified, namely: i) a molecule (morphogen) capable of modulating gene expression and consequently cell fate, ii) a mechanism through which cells respond to different levels of signaling molecule, by changes in either concentration or time of exposure, and iii) a process of either regulated transport, or synthesis and degradation allowing for the generation of a differential distribution of signaling compound across a field of cells (Wolpert 1969). While many morphogens (eg: Fgf,

Bmp, Shh) are peptide signaling molecules, the vitamin A metabolite, RA is also a morphogen, as it satisfies the above requirements. A small, lipophilic molecule, RA is diffusible in vivo, and therefore may not require regulated transport to act in a paracrine manner. It is well documented that RA is not ubiquitous in the embryo, and the expression of RA-synthetic (ALDH) and RA-catabolic enzymes (Cyp26) are critical to defining and maintaining the tissue-specific distribution of RA in vivo (Pennimpede et al.

2006). Early theoretical work in developmental biology proposed that morphogen gradients were generated by synthetic („source‟) and catabolic („sink‟) activities (Wolpert

142

1969). Between the source and sink, a morphogen gradient would be established across a field of cells, which would respond specifically to different levels of morphogen. In terms of RA-signaling, ALDH (source) and Cyp26 (sink) fit cohesively into this model; however, we would like to propose that instead of participating in the formation of a concentration gradient of RA, Cyp26 may function to provide an exclusion barrier to RA.

Essentially all cells have the potential to respond to RA, as RAR , a transducer of

RA signaling is ubiquitous (Dolle et al. 1990; Mollard et al. 2000). However, RA- mediated signaling only occurs in a very specific population of cells at any given time. If this signaling is disrupted or occurs in inappropriate tissues, the effects can be severe for the embryo, and so RA signaling must be tightly regulated to ensure normal embryonic patterning. Throughout the retinoid metabolic pathway, there are several potential steps for regulation; during absorption, storage, mobilization, activation, utilization and catabolism. The significance of regulation at some of these steps depends on maternal retinoid status, but with a retinoid sufficient diet, I believe that the synthesis and degradation of RA by ALDH and Cyp26 enzymes are the key regulatory steps in controlling RA-mediated signaling during embryonic development.

In this thesis, I have detailed the cloning and characterization of a gene encoding one of the RA-catabolic enzymes, Cyp26b1. We have generated a line of transgenic mice lacking Cyp26b1, and characterized for the first time the craniofacial and epidermal malformations manifested in these animals. We have also generated a line of transgenic mice that can be used to conditionally delete Cyp26b1 in a temporal and spatial specific manner to observe CYP26B1 function in adult homeostasis. In this chapter I have

143 attempted to put the results of this thesis in context of our understanding of retinoid metabolism for embryonic development and outline possible research directions for this model.

6.1 Expression of Cyp26 genes defines RA distribution in vivo

As described in Chapter 3, Cyp26b1 was the second Cyp26 gene cloned in mouse.

Cyp26b1 was observed to have a distinct expression pattern that differed from Cyp26a1

(MacLean et al. 2001). While Cyp26a1 is detected in the tailbud, cervical mesenchyme and cleft of the first branchial arch, Cyp26b1 is expressed in the hindbrain, limb buds and nasal process in older embryos. These results suggested CYP26A1 and CYP26B1 were responsible for protecting different tissues from RA. Subsequently, another Cyp26 gene,

Cyp26c1 was identified in mouse, human, zebrafish and chick (Tahayato et al. 2003;

Reijntjes et al. 2004; Taimi et al. 2004; Kawakami et al. 2005). Cyp26c1 expression overlaps with Cyp26a1 in the first branchial arch, and is also detected in the hindbrain, although with a different expression profile than Cyp26b1 (Tahayato et al. 2003). All three proteins efficiently catabolize at-RA, while CYP26C1 can also hydroxylate 9-cis-

RA in vitro (Taimi et al. 2004).

Currently, the most relevant model for visualizing RA distribution in vivo is the

RARE-LacZ mouse (Rossant et al. 1991). At mid-gestation, these animals indicate RA is present in the trunk with an anterior boundary at the posterior hindbrain. RA is also detected in the region surrounding the presumptive eye, but is absent from the limb and tail buds. When observations from these mice are compared with the expression of

144

Cyp26 and Aldh genes, it is apparent that Cyp26 genes are expressed in regions where RA is absent, while the Aldhs are detected in regions where RA is present ((Rossant et al.

1991; Niederreither et al. 1997; MacLean et al. 2001; Niederreither et al. 2002b; Reijntjes et al. 2005). While the expression data was compelling, gene targeting studies had to be performed to definitively establish the function of ALDHs and CYP26s in vivo.

6.2 RA-catabolizing activity is essential for embryonic development

It has been established for decades that certain tissues in the embryo were sensitive to maternal influxes of exogenous RA (Kochhar 1973; Collins and Mao 1999).

Many of these tissues derive from regions where Cyp26 genes are expressed, strengthening the argument that CYP26 activity protected these areas from endogenous retinoids. Transgenic mice lacking various Cyp26 genes have validated this hypothesis.

When Cyp26a1 is knocked out, mutant embryos die during gestation and exhibit variable phenotypes including sirenomelia, spina bifida and exencephaly; observations that are all associated with RA teratogenesis (Abu-Abed et al. 2001; Sakai et al. 2001).

Cyp26a1-/- mice have been generated on a background of RARE-LacZ mice, and show increased levels of RA in the tailbud, where Cyp26a1 is normally expressed (Sakai et al.

2001). When endogenous levels of RA are lowered by generating Cyp26a1-/- Aldh1a2+/- mice, the animals survive despite lacking CYP26A1 (Niederreither et al. 2002a).

Similarly, Cyp26a1-/- RAR -/- mice are viable, and molecular markers of the tailbud that are abnormal in Cyp26a1-/- embryos are restored in double mutants (Abu-Abed et al.

2003). These studies led to the conclusions that i) RA is the active metabolite of vitamin

145

A and ii) CYP26A1 is responsible for restricting gene expression mediated by RAR in the tailbud.

Other common manifestations of RA teratogenicity are cleft palate, limb truncation and defects in branchial arch derived tissues; none of these malformations were observed in Cyp26a1-/- embryos (Kochhar 1973). Considering the expression pattern of Cyp26b1, it was postulated that CYP26B1 was protecting these tissues from inappropriate RA-signaling. This hypothesis was validated when Cyp26b1-/- embryos were generated and observed to have truncated limbs and severe craniofacial abnormalities. When crossed with RARE-LacZ mice, Cyp26b1-/- mutants were reported to have elevated levels of RA in limb buds (other tissues were not commented upon)

(Yashiro et al. 2004). Furthermore, in vitro cell based assays with null embryos indicate

CYP26B1 limits RA in the embryonic male gonad (Bowles et al. 2006; MacLean et al.

2007). Therefore, it is clear that the malformations in Cyp26b1-/- embryos are due to RA- signaling occurring in tissues where it is normally prevented by CYP26B1 mediated catabolism. The genetic disruption of Cyp26a1 and Cyp26b1 has clearly demonstrated the importance of these enzymes in regulating RA distribution in vivo.

Conversely, when the gene encoding the major RA-synthetic enzyme in the embryo, Aldh1a2 is deleted, embryos abort by E9.5 with malformations resembling those associated with maternal vitamin A deficiency (Niederreither et al. 1999). Therefore, the coordinated activity of Cyp26 and ALDH enzymes is critical for establishing a distribution of RA that defines tissues where RA-mediated signaling can occur which is required for morphogenesis.

146

6.3 CYP26-mediated catabolism of RA is required for normal cell fate

RA is a regulator of cell differentiation, proliferation and apoptosis with the exact effects of RA dependent on the cellular context. The Cyp26 knockout models are valuable tools in establishing the function of RA in embryonic development at the cellular level. In Cyp26b1-/- animals, there is increased apoptosis in the limb buds, which likely contributes to truncations in these animals (Yashiro et al. 2004). The embryonic testes of Cyp26b1-/- embryos also show a striking loss of germ cells due to apoptosis, although it is unclear if this is a direct effect of RA or due to release of a Sertoli cell derived factor (MacLean et al. 2007). However, not all phenotypes with a loss of tissue are due to apoptosis in Cyp26 mutants. Cyp26a1-/-/Cyp26c1-/- embryos have a severe reduction in the number of head mesenchymal cells, but rates of apoptosis and proliferation are unaffected (Uehara et al. 2007). Instead, these mutants show reduced production of migratory cranial neural crest cells. Cyp26a1-/- mice exhibit severe posterior malformations but there is no detected increase in caudal apoptosis; therefore, the effects of RA in causing a phenotype are specific to the affected tissue (Abu-Abed et al. 2001).

Cellular abnormalities underlying the craniofacial and epidermal phenotypes of

Cyp26b1-/- embryos have not been determined. Many of the abnormalities observed in

Cyp26b1 null animals such as the shortened mandible, missing eyelids, and reduced calvaria indicate deficiencies in tissue formation, which could be indicative of increased apoptosis in some regions. However, in early embryos, there are no obvious size differences in the branchial arches or head, hence apoptosis may either occur later in

147 development, or be localized to specific cell populations. It is also possible that

CYP26B1 activity maintains some tissues in a proliferative state during development.

For example, Cyp26b1 is expressed at the leading edge of the embryonic eyelid, a region rich in proliferating cells. This region also expresses Fgf10, which is required for cell proliferation in the eyelid, as Fgf10 knockout animals lack eyelids (Tao et al. 2005).

Therefore, CYP26B1 activity in the eyelid may deplete RA, thus allowing for Fgf10 expression which is necessary to maintain cell proliferation. The relationship between

Cyp26b1 and Fgf10 has not been examined, but would be a logical direction for continuing studies.

It is essential to identify cellular processes that are deregulated in Cyp26 mutants, as these results may be extrapolated to various disease models in which there are changes in apoptosis or proliferation.

148

6.4 Future Research Directions

The Cyp26b1 knockout mouse model develops with many diverse phenotypes, which present numerous exciting research avenues. Chapters 4 and 5 describe the craniofacial and epidermal malformations in these animals, along with some of the molecular pathways that are affected possibly contributing to defects, yet, there is still much to be done to delineate the role of CYP26B1 in these tissues. For the craniofacial defects, it must be determined if subtle changes in neural crest cell or branchial arch specification affect signaling cascades culminating in the observed phenotypes.

Additionally, it may be possible that changes in local RA distribution in older embryos are responsible for some of the malformations.

In terms of the epidermis, it must be clarified whether follicle induction is being terminated as a result of downregulation of Wnt signaling, and if the same pathway is responsible for the lack of vibrissal and pelage follicles. It will also be interesting to determine the function, if any, of CYP26B1 in the hair cycle of adult mice.

In addition to continuing to analyze Cyp26b1 null embryos, the transgenic mouse lines that have been developed are a valuable tool that can be utilized for i) the production of compound mutant mice lacking multiple Cyp26 genes, or other RA- metabolic genes, and ii) the generation of conditional knockouts of Cyp26b1 to analyze function in adult animals.

149

6.4.1 Generation of compound Cyp26 mutants

Cyp26a1 and Cyp26b1 are obviously essential for embryonic development, but in some tissues, the phenotypes of single knockout animals may be masked by compensation by other Cyp26 proteins. This has been demonstrated to be the case with

Cyp26a1 and Cyp26c1 in early head development (Uehara et al. 2007). Cyp26c1 null embryos have no discernable abnormalities, and Cyp26a1 null mutants exhibit anterior neural tube abnormalities with low penetrance (Abu-Abed et al. 2001; Uehara et al.

2007). However, Cyp26a1-/- Cyp26c1-/- compound mutant mice fail to develop past E11.0 exhibiting severe abnormalities including a reduced head and failure of neural tube closure. Analyses of these double mutant animals have established that the combined activities of CYP26A1 and CYP26C1 is required for normal migration of cranial neural crest cells (Uehara et al. 2007).

Considering the expression pattern of Cyp26b1 in the hindbrain, it is intriguing that rhombomere segmentation and patterning is essentially normal in Cyp26b1-/- embryos. Presumably, CYP26C1 activity in this tissue prevents RA from diffusing anteriorly to result in more significant malformations in Cyp26b1-/- embryos.

Interestingly, when Cyp26b1 and Cyp26c1 are targeted by antisense morpholinos in zebrafish, there are only very subtle hindbrain defects unless Cyp26a1 is removed as well, resulting in a severely posteriorized hindbrain (Hernandez et al. 2007). However, in the mouse, Cyp26a1 expression is limited to early stages of hindbrain development, so

Cyp26b1/Cyp26c1 double knockout animals may show severe abnormalities in hindbrain development.

150

Cyp26a1 and Cyp26b1 do not show significant overlap in expression, but it will be interesting to study this double knockout model, as there may be coordinated activities of these genes resulting in significant phenotypic alterations. For example, in Cyp26b1-/- embryos, we along with other groups have observed an upregulation of Cyp26a1 in the limb buds beginning at outgrowth (Yashiro et al. 2004). Although this induction of

Cyp26a1 is not sufficient to rescue limb development in the absence of CYP26B1, it may explain why there is still limb outgrowth with truncations not evident until E11.5/E12.5.

Analysis of Cyp26a1/Cyp26b1 double mutant embryos may yield information on signaling pathways required for limb bud initiation and outgrowth of the prospective limb field.

We have already generated Cyp26a1-/-/Cyp26b1-/- embryos, although no extensive analysis has been performed (G MacLean and M Petkovich, unpublished results). Thus far, no double mutant embryos have been obtained beyond E9.5, indicating that there may be an early embryonic lethal phenotype, possibly due to defects in vascular development.

151

6.4.2 CYP26B1 in adult brain

Retinoids have been associated with learning and memory in several animal models, and have also been linked to neurological diseases including Alzheimers and motor neuron disease (Chiang et al. 1998; Krezel et al. 1998; Misner et al. 2001;

Corcoran et al. 2002; Goodman and Pardee 2003). Surprisingly, there has not been a great deal of research regarding the role of RA metabolism in adult brain function.

The only Cyp26 that has been convincingly localized in brain is Cyp26b1, which is expressed in the hippocampus, subiculum and amygdaloid of adult mice (Abu-Abed et al. 2002). RAR and RAR are also detected in the hippocampus, while Aldh1a2 is expressed in the meninges adjacent to the hippocampus (Krezel et al. 1999; McCaffery et al. 2006). This juxtaposition of Aldh1a2 and Cyp26b1 expression in brain is reminiscent of observations during embryogenesis. The hippocampus is a site of neurogenesis, raising the possibility that CYP26B1 may maintain hippocampal neurons in a proliferative state. Alternatively, CYP26B1 may prevent RA-mediated apoptosis in the brain, a function possibly lost in some neurological diseases.

Generation of Cyp26b1 conditional knockout animals would be an ideal model with which to address these questions. Cyp26b1 L2/L2 (Figure 4.2) animals could be crossed with a line of mice that express Cre recombinase in hippocampus (Sahly et al.

2007). Behavioral testing along with histological and molecular analyses could be performed to delineate the role of retinoid metabolism in brain function possibly leading to novel therapies for neurological disorders.

152

6.5 Concluding Remarks

In this thesis, I have described the cloning and functional characterization of the

RA-catabolizing enzyme, CYP26B1. Experiments performed by my colleagues and myself have made significant contributions to the fields of retinoids and developmental biology. Our characterization of Cyp26b1 expression in the embryo demonstrated that

CYP26 activity could be responsible for limiting RA-signaling in tissues including the limb, hindbrain and branchial arches. Generating a line of transgenic mice lacking

CYP26B1 validated this hypothesis. Cyp26b1-/- mice die just after birth and exhibit severe limb and craniofacial malformations, indicating that the RA-catabolizing activity of CYP26B1 is essential for embryonic development. We have begun to characterize some of the phenotypes of Cyp26b1-/- embryos, with some surprising observations.

Cyp26b1-/- embryos develop hypoplastic testes, which we have shown is due to apoptotic extinction of testicular germ cells (MacLean et al. 2007). While other groups had postulated that CYP26B1 activity prevented inappropriate meiosis of male embryonic germ cells, we were the first to categorically demonstrate that meiosis precedes apoptosis in germ cells of Cyp26b1-/- males (Bowles et al. 2006; Koubova et al. 2006; MacLean et al. 2007). We plan to study the significance of CYP26B1 activity in spermatogenesis by generating a testes-specific conditional knockout of Cyp26b1. Cyp26b1-/- embryos display severe craniofacial defects, and yet thus far do not appear to exhibit any striking molecular abnormalities in hindbrain, branchial arch or neural crest cell patterning. The lack of early molecular defects was surprising, and future experiments will focus on

153 tissue-specific signaling pathways later in development. Finally, we also noted deficiencies in hair follicle formation in Cyp26b1-/- embryos, apparently due to deficiencies in -Catenin mediated signaling. We will continue to define the molecular signals responsible for the lack of follicles, we will also study the role of CYP26B1 in adult hair cycling by genetic and pharmacologic means. In addition to further characterizing these phenotypes, the Cyp26b1-/- and Cyp26b1L2/L2 mice have great potential for investigating the significance of retinoid signaling in other tissues in embryonic and adult tissues.

154

References

Abu-Abed, S., P. Dolle, D. Metzger, B. Beckett, P. Chambon, and M. Petkovich. 2001. The retinoic acid-metabolizing enzyme, CYP26A1, is essential for normal hindbrain patterning, vertebral identity, and development of posterior structures. Genes Dev 15: 226-40.

Abu-Abed, S., P. Dolle, D. Metzger, C. Wood, G. MacLean, P. Chambon, and M. Petkovich. 2003. Developing with lethal RA levels: genetic ablation of Rarg can restore the viability of mice lacking Cyp26a1. Development 130: 1449-59.

Abu-Abed, S., G. MacLean, V. Fraulob, P. Chambon, M. Petkovich, and P. Dolle. 2002. Differential expression of the retinoic acid-metabolizing enzymes CYP26A1 and CYP26B1 during murine organogenesis. Mech Dev 110: 173-7.

Abu-Abed, S.S., B.R. Beckett, H. Chiba, J.V. Chithalen, G. Jones, D. Metzger, P. Chambon, and M. Petkovich. 1998. Mouse P450RAI (CYP26) expression and retinoic acid-inducible retinoic acid metabolism in F9 cells are regulated by retinoic acid receptor gamma and retinoid X receptor alpha. J Biol Chem 273: 2409-15.

Achkar, C.C., F. Derguini, B. Blumberg, A. Langston, A.A. Levin, J. Speck, R.M. Evans, J. Bolado, Jr., K. Nakanishi, J. Buck, and et al. 1996. 4-Oxoretinol, a new natural ligand and transactivator of the retinoic acid receptors. Proc Natl Acad Sci U S A 93: 4879-84.

Alappat, S.R., Z. Zhang, K. Suzuki, X. Zhang, H. Liu, R. Jiang, G. Yamada, and Y. Chen. 2005. The cellular and molecular etiology of the cleft secondary palate in Fgf10 mutant mice. Dev Biol 277: 102-13.

Alland, L., R. Muhle, H. Hou, Jr., J. Potes, L. Chin, N. Schreiber-Agus, and R.A. DePinho. 1997. Role for N-CoR and histone deacetylase in Sin3-mediated transcriptional repression. Nature 387: 49-55.

Allenby, G., M.T. Bocquel, M. Saunders, S. Kazmer, J. Speck, M. Rosenberger, A. Lovey, P. Kastner, J.F. Grippo, P. Chambon, and A.A. Levin. 1993. Retinoic acid receptors and retinoid X receptors: interactions with endogenous retinoic acids. Proc Natl Acad Sci U S A 90: 30-4.

Allenby, G., R. Janocha, S. Kazmer, J. Speck, J.F. Grippo, and A.A. Levin. 1994. Binding of 9- cis-retinoic acid and all-trans-retinoic acid to retinoic acid receptors alpha, beta, and gamma. Retinoic acid receptor gamma binds all-trans-retinoic acid preferentially over 9- cis-retinoic acid. J Biol Chem 269: 16689-95.

Andl, T., K. Ahn, A. Kairo, E.Y. Chu, L. Wine-Lee, S.T. Reddy, N.J. Croft, J.A. Cebra-Thomas, D. Metzger, P. Chambon, K.M. Lyons, Y. Mishina, J.T. Seykora, E.B. Crenshaw, 3rd, and S.E. Millar. 2004. Epithelial Bmpr1a regulates differentiation and proliferation in postnatal hair follicles and is essential for tooth development. Development 131: 2257-68.

155

Andl, T., S.T. Reddy, T. Gaddapara, and S.E. Millar. 2002. WNT signals are required for the initiation of hair follicle development. Dev Cell 2: 643-53.

Balkan, W., M. Colbert, C. Bock, and E. Linney. 1992. Transgenic indicator mice for studying activated retinoic acid receptors during development. Proc Natl Acad Sci U S A 89: 3347- 51.

Balmer, J.E. and R. Blomhoff. 2002. Gene expression regulation by retinoic acid. J Lipid Res 43: 1773-808.

Bi, W., W. Huang, D.J. Whitworth, J.M. Deng, Z. Zhang, R.R. Behringer, and B. de Crombrugghe. 2001. Haploinsufficiency of Sox9 results in defective cartilage primordia and premature skeletal mineralization. Proc Natl Acad Sci U S A. 98: 6698-703. Epub 2001 May 22.

Blomhoff, R. and H.K. Blomhoff. 2006. Overview of retinoid metabolism and function. J Neurobiol 66: 606-30.

Bowles, J., D. Knight, C. Smith, D. Wilhelm, J. Richman, S. Mamiya, K. Yashiro, K. Chawengsaksophak, M.J. Wilson, J. Rossant, H. Hamada, and P. Koopman. 2006. Retinoid signaling determines germ cell fate in mice. Science 312: 596-600.

Brand, N., M. Petkovich, A. Krust, P. Chambon, H. de The, A. Marchio, P. Tiollais, and A. Dejean. 1988. Identification of a second human retinoic acid receptor. Nature 332: 850-3.

Brent, A.E. and C.J. Tabin. 2002. Developmental regulation of somite derivatives: muscle, cartilage and tendon. Curr Opin Genet Dev 12: 548-57.

Calleja, C., N. Messaddeq, B. Chapellier, H. Yang, W. Krezel, M. Li, D. Metzger, B. Mascrez, K. Ohta, H. Kagechika, Y. Endo, M. Mark, N.B. Ghyselinck, and P. Chambon. 2006. Genetic and pharmacological evidence that a retinoic acid cannot be the RXR-activating ligand in mouse epidermis keratinocytes. Genes Dev 20: 1525-38.

Capel, B. 2000. The battle of the sexes. Mech Dev 92: 89-103.

Carver, E.A., R. Jiang, Y. Lan, K.F. Oram, and T. Gridley. 2001. The mouse snail gene encodes a key regulator of the epithelial-mesenchymal transition. Mol Cell Biol 21: 8184-8.

Chai, Y. and R.E. Maxson, Jr. 2006. Recent advances in craniofacial morphogenesis. Dev Dyn 235: 2353-75.

Chambon, P. 1996. A decade of molecular biology of retinoic acid receptors. Faseb J 10: 940-54.

Charite, J., D.G. McFadden, G. Merlo, G. Levi, D.E. Clouthier, M. Yanagisawa, J.A. Richardson, and E.N. Olson. 2001. Role of Dlx6 in regulation of an endothelin-1-dependent, dHAND branchial arch enhancer. Genes Dev 15: 3039-49.

156

Chen, J.D. and R.M. Evans. 1995. A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature 377: 454-7.

Chen, Y., L. Huang, and M. Solursh. 1994. A concentration gradient of retinoids in the early Xenopus laevis embryo. Dev Biol 161: 70-6.

Chiang, M.Y., D. Misner, G. Kempermann, T. Schikorski, V. Giguere, H.M. Sucov, F.H. Gage, C.F. Stevens, and R.M. Evans. 1998. An essential role for retinoid receptors RARbeta and RXRgamma in long- term potentiation and depression. Neuron 21: 1353-61.

Cohen, E.D., Z. Wang, J.J. Lepore, M.M. Lu, M.M. Taketo, D.J. Epstein, and E.E. Morrisey. 2007. Wnt/beta-catenin signaling promotes expansion of Isl-1-positive cardiac progenitor cells through regulation of FGF signaling. J Clin Invest 117: 1794-1804.

Colbert, M.C., E. Linney, and A.S. LaMantia. 1993. Local sources of retinoic acid coincide with retinoid-mediated transgene activity during embryonic development. Proc Natl Acad Sci U S A 90: 6572-6.

Collins, M.D., C. Eckhoff, R. Weiss, E. Resnick, H. Nau, and W.J. Scott, Jr. 2006. Differential teratogenesis of all-trans-retinoic acid administered on gestational day 9.5 to SWV and C57BL/6N mice: emphasis on limb dysmorphology. Birth Defects Res A Clin Mol Teratol 76: 96-106.

Collins, M.D. and G.E. Mao. 1999. Teratology of retinoids. Annu Rev Pharmacol Toxicol 39: 399-430.

Corcoran, J., P.L. So, and M. Maden. 2002. Absence of retinoids can induce motoneuron disease in the adult rat and a retinoid defect is present in motoneuron disease patients. J Cell Sci 115: 4735-41. de Roos, K., E. Sonneveld, B. Compaan, D. ten Berge, A.J. Durston, and P.T. van der Saag. 1999. Expression of retinoic acid 4-hydroxylase (CYP26) during mouse and Xenopus laevis embryogenesis. Mech Dev 82: 205-11. de Urquiza, A.M., S. Liu, M. Sjoberg, R.H. Zetterstrom, W. Griffiths, J. Sjovall, and T. Perlmann. 2000. Docosahexaenoic acid, a ligand for the retinoid X receptor in mouse brain. Science 290: 2140-4.

Decimo, D., E. Georges-Labouesse, and P. Dolle. 1995. In situ hybridization to cellular RNA. In Gene Probes 2, a Practical Approach (ed. B.D. James and S.J. Higgins). Oxford University Press, New York.

Deltour, L., M.H. Foglio, and G. Duester. 1999a. Impaired retinol utilization in Adh4 alcohol dehydrogenase mutant mice. Dev Genet 25: 1-10.

Deltour, L., M. H. Foglio, and G. Duester. 1999b. Metabolic deficiencies in alcohol dehydrogenase Adh1, Adh3, and Adh4 null mutant mice. Overlapping roles of Adh1 and

157

Adh4 in ethanol clearance and metabolism of retinol to retinoic acid. J Biol Chem 274: 16796-801.

Depew, M.J., C.A. Simpson, M. Morasso, and J.L. Rubenstein. 2005. Reassessing the Dlx code: the genetic regulation of branchial arch skeletal pattern and development. J Anat 207: 501-61.

Dilworth, F.J. and P. Chambon. 2001. Nuclear receptors coordinate the activities of chromatin remodeling complexes and coactivators to facilitate initiation of transcription. Oncogene 20: 3047-54.

Dolle, P., T. Lufkin, R. Krumlauf, M. Mark, D. Duboule, and P. Chambon. 1993. Local alterations of Krox-20 and Hox gene expression in the hindbrain suggest lack of rhombomeres 4 and 5 in homozygote null Hoxa-1 (Hox-1.6) mutant embryos. Proc Natl Acad Sci U S A 90: 7666-70.

Dolle, P., E. Ruberte, P. Kastner, M. Petkovich, C.M. Stoner, L.J. Gudas, and P. Chambon. 1989. Differential expression of genes encoding alpha, beta and gamma retinoic acid receptors and CRABP in the developing limbs of the mouse. Nature 342: 702-5.

Dolle, P., E. Ruberte, P. Leroy, G. Morriss-Kay, and P. Chambon. 1990. Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differential pattern of transcription during mouse organogenesis. Development 110: 1133-51.

Dowling, J.E. and G. Wald. 1960. The Biological Function of Vitamin a Acid. Proc Natl Acad Sci U S A 46: 587-608.

Duell, E.A., S. Kang, and J.J. Voorhees. 1996. Retinoic acid isomers applied to human skin in vivo each induce a 4- hydroxylase that inactivates only trans retinoic acid. J Invest Dermatol 106: 316-20.

Duester, G., F.A. Mic, and A. Molotkov. 2003. Cytosolic retinoid dehydrogenases govern ubiquitous metabolism of retinol to retinaldehyde followed by tissue-specific metabolism to retinoic acid. Chem Biol Interact 143-144: 201-10.

Dupe, V., M. Davenne, J. Brocard, P. Dolle, M. Mark, A. Dierich, P. Chambon, and F.M. Rijli. 1997. In vivo functional analysis of the Hoxa-1 3' retinoic acid response element (3'RARE). Development 124: 399-410.

Dupe, V., N. Matt, J.M. Garnier, P. Chambon, M. Mark, and N.B. Ghyselinck. 2003. A newborn lethal defect due to inactivation of retinaldehyde dehydrogenase type 3 is prevented by maternal retinoic acid treatment. Proc Natl Acad Sci U S A 100: 14036-41.

Everts, H.B., L.E. King, Jr., J.P. Sundberg, and D.E. Ong. 2004. Hair cycle-specific immunolocalization of retinoic acid synthesizing enzymes Aldh1a2 and Aldh1a3 indicate complex regulation. J Invest Dermatol 123: 258-63.

158

Everts, H.B., J.P. Sundberg, L.E. King, Jr., and D.E. Ong. 2007. Immunolocalization of enzymes, binding proteins, and receptors sufficient for retinoic acid synthesis and signaling during the hair cycle. J Invest Dermatol 127: 1593-604.

Fan, X., A. Molotkov, S. Manabe, C.M. Donmoyer, L. Deltour, M.H. Foglio, A.E. Cuenca, W.S. Blaner, S.A. Lipton, and G. Duester. 2003. Targeted disruption of Aldh1a1 (Raldh1) provides evidence for a complex mechanism of retinoic acid synthesis in the developing retina. Mol Cell Biol 23: 4637-48.

Fawcett, D., P. Pasceri, R. Fraser, M. Colbert, J. Rossant, and V. Giguere. 1995. Postaxial polydactyly in forelimbs of CRABP-II mutant mice. Development 121: 671-9.

Flajollet, S., B. Lefebvre, C. Cudejko, B. Staels, and P. Lefebvre. 2007. The core component of the mammalian SWI/SNF complex SMARCD3/BAF60c is a coactivator for the nuclear retinoic acid receptor. Mol Cell Endocrinol 270: 23-32.

Frenz, D.A., W. Liu, V. Galinovic-Schwartz, and T.R. Van De Water. 1996. Retinoic acid- induced embryopathy of the mouse inner ear. Teratology 53: 292-303.

Frolik, C.A., A.B. Roberts, T.E. Tavela, P.P. Roller, D.L. Newton, and M.B. Sporn. 1979. Isolation and identification of 4-hydroxy- and 4-oxoretinoic acid. In vitro metabolites of all-trans-retinoic acid in hamster trachea and liver. Biochemistry 18: 2092-7.

Fuchs, E. 2007. Scratching the surface of skin development. Nature 445: 834-42.

Fujii, H., T. Sato, S. Kaneko, O. Gotoh, Y. Fujii-Kuriyama, K. Osawa, S. Kato, and H. Hamada. 1997. Metabolic inactivation of retinoic acid by a novel P450 differentially expressed in developing mouse embryos. Embo J 16: 4163-73.

Fukuda, H., N. Sano, S. Muto, and M. Horikoshi. 2006. Simple histone acetylation plays a complex role in the regulation of gene expression. Brief Funct Genomic Proteomic 5: 190-208.

Gavalas, A. 2002. ArRAnging the hindbrain. Trends Neurosci 25: 61-4.

Germain, P., P. Chambon, G. Eichele, R.M. Evans, M.A. Lazar, M. Leid, A.R. De Lera, R. Lotan, D.J. Mangelsdorf, and H. Gronemeyer. 2006. International Union of Pharmacology. LX. Retinoic acid receptors. Pharmacol Rev 58: 712-25.

Ghyselinck, N.B., V. Dupe, A. Dierich, N. Messaddeq, J.M. Garnier, C. Rochette-Egly, P. Chambon, and M. Mark. 1997. Role of the retinoic acid receptor beta (RARbeta) during mouse development. Int J Dev Biol 41: 425-47.

Giguere, V., S. Lyn, P. Yip, C.H. Siu, and S. Amin. 1990. Molecular cloning of cDNA encoding a second cellular retinoic acid-binding protein. Proc Natl Acad Sci U S A 87: 6233-7.

Giguere, V., E.S. Ong, P. Segui, and R.M. Evans. 1987. Identification of a receptor for the morphogen retinoic acid. Nature 330: 624-9.

159

Glover, J.C., J.S. Renaud, and F.M. Rijli. 2006. Retinoic acid and hindbrain patterning. J Neurobiol 66: 705-25.

Goldstein, J.T., A. Dobrzyn, M. Clagett-Dame, J.W. Pike, and H.F. DeLuca. 2003. Isolation and characterization of unsaturated fatty acids as natural ligands for the retinoid-X receptor. Arch Biochem Biophys 420: 185-93.

Goodman, A.B. and A.B. Pardee. 2003. Evidence for defective retinoid transport and function in late onset Alzheimer's disease. Proc Natl Acad Sci U S A 100: 2901-5.

Goodman, D.W., H.S. Huang, and T. Shiratori. 1965. Tissue Distribution and Metabolism of Newly Absorbed Vitamin a in the Rat. J Lipid Res 6: 390-6.

Greco, T.L. and A.H. Payne. 1994. Ontogeny of expression of the genes for steroidogenic enzymes P450 side-chain cleavage, 3 beta-hydroxysteroid dehydrogenase, P450 17 alpha-hydroxylase/C17-20 lyase, and P450 aromatase in fetal mouse gonads. Endocrinology 135: 262-8.

Halilagic, A., M.H. Zile, and M. Studer. 2003. A novel role for retinoids in patterning the avian forebrain during presomite stages. Development 130: 2039-50.

Haselbeck, R.J., I. Hoffmann, and G. Duester. 1999. Distinct functions for Aldh1 and Raldh2 in the control of ligand production for embryonic retinoid signaling pathways. Dev Genet 25: 353-64.

Heinzel, T., R.M. Lavinsky, T.M. Mullen, M. Soderstrom, C.D. Laherty, J. Torchia, W.M. Yang, G. Brard, S.D. Ngo, J.R. Davie, E. Seto, R.N. Eisenman, D.W. Rose, C.K. Glass, and M.G. Rosenfeld. 1997. A complex containing N-CoR, mSin3 and histone deacetylase mediates transcriptional repression. Nature 387: 43-8.

Hernandez, R.E., A.P. Putzke, J.P. Myers, L. Margaretha, and C.B. Moens. 2007. Cyp26 enzymes generate the retinoic acid response pattern necessary for hindbrain development. Development. 134: 177-87.

Heyman, R.A., D.J. Mangelsdorf, J.A. Dyck, R.B. Stein, G. Eichele, R.M. Evans, and C. Thaller. 1992. 9-cis retinoic acid is a high affinity ligand for the retinoid X receptor. Cell 68: 397- 406.

Hollemann, T., Y. Chen, H. Grunz, and T. Pieler. 1998. Regionalized metabolic activity establishes boundaries of retinoic acid signalling. Embo J 17: 7361-72.

Holmes, R.S. 1977. The genetics of alpha-hydroxyacid oxidase and alcohol dehydrogenase in the mouse: evidence for multiple gene loci and linkage between Hao-2 and Adh-3. Genetics 87: 709-16.

Huelsken, J., R. Vogel, B. Erdmann, G. Cotsarelis, and W. Birchmeier. 2001. beta-Catenin controls hair follicle morphogenesis and stem cell differentiation in the skin. Cell 105: 533-45.

160

Iulianella, A., B. Beckett, M. Petkovich, and D. Lohnes. 1999. A molecular basis for retinoic acid-induced axial truncation. Dev Biol 205: 33-48.

Jerome, L.A. and V.E. Papaioannou. 2001. DiGeorge syndrome phenotype in mice mutant for the T-box gene, Tbx1. Nat Genet 27: 286-91.

Kalin, J.R., M.E. Starling, and D.L. Hill. 1981. Disposition of all-trans-retinoic acid in mice following oral doses. Drug Metab Dispos 9: 196-201.

Kalter, H. and J. Warkany. 1961. Experimental production of congenital malformations in strains of inbred mice by maternal treatment with hypervitaminosis A. Am J Pathol 38: 1-21.

Kane, M.A., N. Chen, S. Sparks, and J.L. Napoli. 2005. Quantification of endogenous retinoic acid in limited biological samples by LC/MS/MS. Biochem J 388: 363-9.

Kastner, P., M. Mark, and P. Chambon. 1995. Nonsteroid nuclear receptors: what are genetic studies telling us about their role in real life? Cell 83: 859-69.

Kaufman, M.H. and Bard J. B. 1999. The Anatomical Basis of Mouse Development. Harcourt Brace and Company, Toronto, ON.

Kawakami, Y., A. Raya, R.M. Raya, C. Rodriguez-Esteban, and J.C. Belmonte. 2005. Retinoic acid signalling links left-right asymmetric patterning and bilaterally symmetric somitogenesis in the zebrafish embryo. Nature 435: 165-71.

Kochhar, D.M. 1973. Limb development in mouse embryos. I. Analysis of teratogenic effects of retinoic acid. Teratology 7: 289-98.

Koide, T., M. Downes, R.A. Chandraratna, B. Blumberg, and K. Umesono. 2001. Active repression of RAR signaling is required for head formation. Genes Dev 15: 2111-21.

Koubova, J., D.B. Menke, Q. Zhou, B. Capel, M.D. Griswold, and D.C. Page. 2006. Retinoic acid regulates sex-specific timing of meiotic initiation in mice. Proc Natl Acad Sci U S A 103: 2474-9.

Krezel, W., N. Ghyselinck, T.A. Samad, V. Dupe, P. Kastner, E. Borrelli, and P. Chambon. 1998. Impaired locomotion and dopamine signaling in retinoid receptor mutant mice. Science 279: 863-7.

Krezel, W., P. Kastner, and P. Chambon. 1999. Differential expression of retinoid receptors in the adult mouse central nervous system. Neuroscience 89: 1291-300.

Krust, A., P. Kastner, M. Petkovich, A. Zelent, and P. Chambon. 1989. A third human retinoic acid receptor, hRAR-gamma. Proc Natl Acad Sci U S A 86: 5310-4.

Lammer, E.J., D.T. Chen, R.M. Hoar, N.D. Agnish, P.J. Benke, J.T. Braun, C.J. Curry, P.M. Fernhoff, A.W. Grix, Jr., I.T. Lott, and et al. 1985. Retinoic acid embryopathy. N Engl J Med 313: 837-41.

161

Lammers, J.H., H.H. Offenberg, M. van Aalderen, A.C. Vink, A.J. Dietrich, and C. Heyting. 1994. The gene encoding a major component of the lateral elements of synaptonemal complexes of the rat is related to X-linked lymphocyte-regulated genes. Mol Cell Biol 14: 1137-46.

Lampron, C., C. Rochette-Egly, P. Gorry, P. Dolle, M. Mark, T. Lufkin, M. LeMeur, and P. Chambon. 1995. Mice deficient in cellular retinoic acid binding protein II (CRABPII) or in both CRABPI and CRABPII are essentially normal. Development 121: 539-48.

Lander, A.D. 2007. Morpheus unbound: reimagining the morphogen gradient. Cell 128: 245-56.

Lee, C.S. and J. Koo. 2005. A review of acitretin, a systemic retinoid for the treatment of psoriasis. Expert Opin Pharmacother 6: 1725-34.

Levin, A.A., L.J. Sturzenbecker, S. Kazmer, T. Bosakowski, C. Huselton, G. Allenby, J. Speck, C. Kratzeisen, M. Rosenberger, A. Lovey, and J.F. Grippo. 1992. 9-cis retinoic acid stereoisomer binds and activates the nuclear receptor RXR alpha. Nature 355: 359-61.

Li, C., L. Hu, J. Xiao, H. Chen, J.T. Li, S. Bellusci, S. Delanghe, and P. Minoo. 2005. Wnt5a regulates Shh and Fgf10 signaling during lung development. Dev Biol 287: 86-97.

Li, E., H.M. Sucov, K.F. Lee, R.M. Evans, and R. Jaenisch. 1993. Normal development and growth of mice carrying a targeted disruption of the alpha 1 retinoic acid receptor gene. Proc Natl Acad Sci U S A 90: 1590-4.

Li, H. and K.H. Kim. 2004. Retinoic acid inhibits rat XY gonad development by blocking mesonephric cell migration and decreasing the number of gonocytes. Biol Reprod 70: 687-93.

Li, H., E. Wagner, P. McCaffery, D. Smith, A. Andreadis, and U.C. Drager. 2000. A retinoic acid synthesizing enzyme in ventral retina and telencephalon of the embryonic mouse. Mech Dev 95: 283-9.

Lin, M., M. Zhang, M. Abraham, S.M. Smith, and J.L. Napoli. 2003. Mouse retinal dehydrogenase 4 (RALDH4), molecular cloning, cellular expression, and activity in 9- cis-retinoic acid biosynthesis in intact cells. J Biol Chem 278: 9856-61.

Lohnes, D., P. Kastner, A. Dierich, M. Mark, M. LeMeur, and P. Chambon. 1993. Function of retinoic acid receptor gamma in the mouse. Cell 73: 643-58.

Lohnes, D., M. Mark, C. Mendelsohn, P. Dolle, D. Decimo, M. LeMeur, A. Dierich, P. Gorry, and P. Chambon. 1995. Developmental roles of the retinoic acid receptors. J Steroid Biochem Mol Biol 53: 475-86.

Lohnes, D., M. Mark, C. Mendelsohn, P. Dolle, A. Dierich, P. Gorry, A. Gansmuller, and P. Chambon. 1994. Function of the retinoic acid receptors (RARs) during development (I). Craniofacial and skeletal abnormalities in RAR double mutants. Development 120: 2723- 48.

162

Loudig, O., C. Babichuk, J. White, S. Abu-Abed, C. Mueller, and M. Petkovich. 2000. Cytochrome P450RAI(CYP26) promoter: a distinct composite retinoic acid response element underlies the complex regulation of retinoic acid metabolism. Mol Endocrinol 14: 1483-97.

Loudig, O., G.A. Maclean, N.L. Dore, L. Luu, and M. Petkovich. 2005. Transcriptional co- operativity between distant retinoic acid response elements in regulation of Cyp26A1 inducibility. Biochem J 392: 241-8.

Lufkin, T., D. Lohnes, M. Mark, A. Dierich, P. Gorry, M.P. Gaub, M. LeMeur, and P. Chambon. 1993. High postnatal lethality and testis degeneration in retinoic acid receptor alpha mutant mice. Proc Natl Acad Sci U S A 90: 7225-9.

MacLean, G., S. Abu-Abed, P. Dolle, A. Tahayato, P. Chambon, and M. Petkovich. 2001. Cloning of a novel retinoic-acid metabolizing cytochrome P450, Cyp26B1, and comparative expression analysis with Cyp26A1 during early murine development. Mech Dev 107: 195-201.

MacLean, G., H. Li, D. Metzger, P. Chambon, and M. Petkovich. 2007. Apoptotic extinction of germ cells in testes of Cyp26b1 knockout mice. Endocrinology.

Maden, M. 1998. The role of retinoids in developmental mechanisms in embryos. Subcell Biochem 30: 81-111.

Maden, M., E. Sonneveld, P.T. van der Saag, and E. Gale. 1998. The distribution of endogenous retinoic acid in the chick embryo: implications for developmental mechanisms. Development 125: 4133-44.

Mark, M., N.B. Ghyselinck, and P. Chambon. 2004. Retinoic acid signalling in the development of branchial arches. Curr Opin Genet Dev 14: 591-8.

Mark, M., T. Lufkin, J.L. Vonesch, E. Ruberte, J.C. Olivo, P. Dolle, P. Gorry, A. Lumsden, and P. Chambon. 1993. Two rhombomeres are altered in Hoxa-1 mutant mice. Development 119: 319-38.

Mason, K.E. 1935. Foetal death, prolonged gestation, and difficult parturition in the rat as a result of vitamin A-deficiency. Am J Anat 57: 303-349.

McCaffery, P., E. Wagner, J. O'Neil, M. Petkovich, and U.C. Drager. 1999. Dorsal and ventral rentinoic territories defined by retinoic acid synthesis, break-down and nuclear receptor expression. Mech Dev 85: 203-14.

McCaffery, P., J. Zhang, and J.E. Crandall. 2006. Retinoic acid signaling and function in the adult hippocampus. J Neurobiol 66: 780-91.

McCaffery, P.J., J. Adams, M. Maden, and E. Rosa-Molinar. 2003. Too much of a good thing: retinoic acid as an endogenous regulator of neural differentiation and exogenous teratogen. Eur J Neurosci 18: 457-72.

163

McClellan, K.A., R. Gosden, and T. Taketo. 2003. Continuous loss of oocytes throughout meiotic prophase in the normal mouse ovary. Dev Biol 258: 334-48.

McCollum, E.V. and M. Davis. 1913. The Necessity of certin Lipins in the Diet During Growth. J Biol Chem 15: 167-175.

McLaren, A. 1984. Meiosis and differentiation of mouse germ cells. Symp Soc Exp Biol 38: 7-23.

Mendelsohn, C., D. Lohnes, D. Decimo, T. Lufkin, M. LeMeur, P. Chambon, and M. Mark. 1994a. Function of the retinoic acid receptors (RARs) during development (II). Multiple abnormalities at various stages of organogenesis in RAR double mutants. Development 120: 2749-71.

Mendelsohn, C., M. Mark, P. Dolle, A. Dierich, M.P. Gaub, A. Krust, C. Lampron, and P. Chambon. 1994b. Retinoic acid receptor beta 2 (RAR beta 2) null mutant mice appear normal. Dev Biol 166: 246-58.

Mendelsohn, C., E. Ruberte, M. LeMeur, G. Morriss-Kay, and P. Chambon. 1991. Developmental analysis of the retinoic acid-inducible RAR-beta 2 promoter in transgenic animals. Development 113: 723-34.

Metzger, D. and R. Feil. 1999. Engineering the mouse genome by site-specific recombination. Curr Opin Biotechnol 10: 470-6.

Metzger, D., A.K. Indra, M. Li, B. Chapellier, C. Calleja, N.B. Ghyselinck, and P. Chambon. 2003. Targeted conditional somatic mutagenesis in the mouse: temporally-controlled knock out of retinoid receptors in epidermal keratinocytes. Methods Enzymol 364: 379- 408.

Metzger, D., E. Scheer, A. Soldatov, and L. Tora. 1999. Mammalian TAF(II)30 is required for cell cycle progression and specific cellular differentiation programmes. Embo J 18: 4823- 34.

Mic, F.A., R.J. Haselbeck, A.E. Cuenca, and G. Duester. 2002. Novel retinoic acid generating activities in the neural tube and heart identified by conditional rescue of Raldh2 null mutant mice. Development 129: 2271-82.

Mic, F.A., A. Molotkov, X. Fan, A.E. Cuenca, and G. Duester. 2000. RALDH3, a retinaldehyde dehydrogenase that generates retinoic acid, is expressed in the ventral retina, otic vesicle and olfactory pit during mouse development. Mech Dev 97: 227-30.

Misner, D.L., S. Jacobs, Y. Shimizu, A.M. de Urquiza, L. Solomin, T. Perlmann, L.M. De Luca, C.F. Stevens, and R.M. Evans. 2001. Vitamin A deprivation results in reversible loss of hippocampal long-term synaptic plasticity. Proc Natl Acad Sci U S A 98: 11714-9.

Mollard, R., S. Viville, S.J. Ward, D. Decimo, P. Chambon, and P. Dolle. 2000. Tissue-specific expression of retinoic acid receptor isoform transcripts in the mouse embryo. Mech Dev 94: 223-32.

164

Molotkov, A., L. Deltour, M.H. Foglio, A.E. Cuenca, and G. Duester. 2002a. Distinct retinoid metabolic functions for alcohol dehydrogenase genes Adh1 and Adh4 in protection against vitamin A toxicity or deficiency revealed in double null mutant mice. J Biol Chem 277: 13804-11.

Molotkov, A., X. Fan, L. Deltour, M.H. Foglio, S. Martras, J. Farres, X. Pares, and G. Duester. 2002b. Stimulation of retinoic acid production and growth by ubiquitously expressed alcohol dehydrogenase Adh3. Proc Natl Acad Sci U S A 99: 5337-42.

Moore, K.A. and I.R. Lemischka. 2006. Stem cells and their niches. Science 311: 1880-5.

Mori-Akiyama, Y., H. Akiyama, D.H. Rowitch, and B. de Crombrugghe. 2003. Sox9 is required for determination of the chondrogenic cell lineage in the cranial neural crest. Proc Natl Acad Sci U S A 100: 9360-5.

Moss, J.B., J. Xavier-Neto, M.D. Shapiro, S.M. Nayeem, P. McCaffery, U.C. Drager, and N. Rosenthal. 1998. Dynamic patterns of retinoic acid synthesis and response in the developing mammalian heart. Dev Biol 199: 55-71.

Mulder, G.B., N. Manley, J. Grant, K. Schmidt, W. Zeng, C. Eckhoff, and L. Maggio-Price. 2000. Effects of excess vitamin A on development of cranial neural crest-derived structures: a neonatal and embryologic study. Teratology 62: 214-26.

Mulder, G.B., N. Manley, and L. Maggio-Price. 1998. Retinoic acid-induced thymic abnormalities in the mouse are associated with altered pharyngeal morphology, thymocyte maturation defects, and altered expression of Hoxa3 and Pax1. Teratology 58: 263-75.

Munsterberg, A. and R. Lovell-Badge. 1991. Expression of the mouse anti-mullerian hormone gene suggests a role in both male and female sexual differentiation. Development 113: 613-24.

Murray, S.A., K.F. Oram, and T. Gridley. 2007. Multiple functions of Snail family genes during palate development in mice. Development 134: 1789-97.

Nagy, L., H.Y. Kao, D. Chakravarti, R.J. Lin, C.A. Hassig, D.E. Ayer, S.L. Schreiber, and R.M. Evans. 1997. Nuclear receptor repression mediated by a complex containing SMRT, mSin3A, and histone deacetylase. Cell 89: 373-80.

Niederreither, K., S. Abu-Abed, B. Schuhbaur, M. Petkovich, P. Chambon, and P. Dolle. 2002a. Genetic evidence that oxidative derivatives of retinoic acid are not involved in retinoid signaling during mouse development. Nat Genet 31: 84-8.

Niederreither, K. and P. Dolle. 2006. Molecular mediators of retinoic acid signaling during development. Adv Dev Biol 16: 107-143.

165

Niederreither, K., V. Fraulob, J.M. Garnier, P. Chambon, and P. Dolle. 2002b. Differential expression of retinoic acid-synthesizing (RALDH) enzymes during fetal development and organ differentiation in the mouse. Mech Dev 110: 165-71.

Niederreither, K., P. McCaffery, U.C. Drager, P. Chambon, and P. Dolle. 1997. Restricted expression and retinoic acid-induced downregulation of the retinaldehyde dehydrogenase type 2 (RALDH-2) gene during mouse development. Mech Dev 62: 67-78.

Niederreither, K., V. Subbarayan, P. Dolle, and P. Chambon. 1999. Embryonic retinoic acid synthesis is essential for early mouse post- implantation development [see comments]. Nat Genet 21: 444-8.

Niederreither, K., J. Vermot, V. Fraulob, P. Chambon, and P. Dolle. 2002c. Retinaldehyde dehydrogenase 2 (RALDH2)- independent patterns of retinoic acid synthesis in the mouse embryo. Proc Natl Acad Sci U S A 99: 16111-6.

Niederreither, K., J. Vermot, N. Messaddeq, B. Schuhbaur, P. Chambon, and P. Dolle. 2001. Embryonic retinoic acid synthesis is essential for heart morphogenesis in the mouse. Development 128: 1019-31.

Niederreither, K., J. Vermot, B. Schuhbaur, P. Chambon, and P. Dolle. 2000. Retinoic acid synthesis and hindbrain patterning in the mouse embryo. Development 127: 75-85.

Niederreither, K., J. Vermot, B. Schuhbaur, P Chambon, and P. Dolle. 2002d. Embryonic retinoic acid synthesis is required for forelimb growth and anteroposterior patterning in the mouse. Development 129: 3563-74.

Ohuchi, H., H. Tao, K. Ohata, N. Itoh, S. Kato, S. Noji, and K. Ono. 2003. Fibroblast growth factor 10 is required for proper development of the mouse whiskers. Biochem Biophys Res Commun 302: 562-7.

Olson, J.A. and O. Hayaishi. 1965. The enzymatic cleavage of beta-carotene into vitamin A by soluble enzymes of rat liver and intestine. Proc Natl Acad Sci U S A 54: 1364-70.

Pennimpede, T., D. Cameron, and M. Petkovich. 2006. Regulation of murine embryonic patterning and morphogenesis by retinoic acid signaling. Adv Dev Biol 16: 66-104.

Peters, A.H., A.W. Plug, M.J. van Vugt, and P. de Boer. 1997. A drying-down technique for the spreading of mammalian meiocytes from the male and female germline. Chromosome Res 5: 66-8.

Petkovich, M., N.J. Brand, A. Krust, and P. Chambon. 1987. A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 330: 444-50.

Pijnappel, W.W., H.F. Hendriks, G.E. Folkers, C.E. van den Brink, E.J. Dekker, C. Edelenbosch, P.T. van der Saag, and A.J. Durston. 1993. The retinoid ligand 4-oxo-retinoic acid is a highly active modulator of positional specification. Nature 366: 340-4.

166

Quadro, L., W.S. Blaner, L. Hamberger, P.M. Novikoff, S. Vogel, R. Piantedosi, M.E. Gottesman, and V. Colantuoni. 2004a. The role of extrahepatic retinol binding protein in the mobilization of retinoid stores. J Lipid Res 45: 1975-82.

Quadro, L., W.S. Blaner, D.J. Salchow, S. Vogel, R. Piantedosi, P. Gouras, S. Freeman, M.P. Cosma, V. Colantuoni, and M.E. Gottesman. 1999. Impaired retinal function and vitamin A availability in mice lacking retinol-binding protein. Embo J 18: 4633-44.

Quadro, L., L. Hamberger, M.E. Gottesman, V. Colantuoni, R. Ramakrishnan, and W.S. Blaner. 2004b. Transplacental delivery of retinoid: the role of retinol-binding protein and lipoprotein retinyl ester. Am J Physiol Endocrinol Metab 286: E844-51.

Ray, W.J., G. Bain, M. Yao, and D.I. Gottlieb. 1997. CYP26, a novel mammalian cytochrome P450, is induced by retinoic acid and defines a new family. J Biol Chem 272: 18702-8.

Reijntjes, S., A. Blentic, E. Gale, and M. Maden. 2005. The control of morphogen signalling: regulation of the synthesis and catabolism of retinoic acid in the developing embryo. Dev Biol 285: 224-37.

Reijntjes, S., E. Gale, and M. Maden. 2003. Expression of the retinoic acid catabolising enzyme CYP26B1 in the chick embryo and its regulation by retinoic acid. Gene Expr Patterns 3: 621-7.

Reijntjes, S., E. Gale, and M. Maden. 2004. Generating gradients of retinoic acid in the chick embryo: Cyp26C1 expression and a comparative analysis of the Cyp26 enzymes. Dev Dyn 230: 509-17.

Reijntjes, S., A. Rodaway, and M. Maden. 2007. The retinoic acid metabolising gene, CYP26B1, patterns the cartilaginous cranial neural crest in zebrafish. Int J Dev Biol 51: 351-60.

Repa, J.J., K.K. Hanson, and M. Clagett-Dame. 1993. All-trans-retinol is a ligand for the retinoic acid receptors. Proc Natl Acad Sci U S A 90: 7293-7.

Ridanpaa, M., R. Fodde, and M. Kielman. 2001. Dynamic expression and nuclear accumulation of beta-catenin during the development of hair follicle-derived structures. Mech Dev 109: 173-81.

Roberts, A.B., C.A. Frolik, M.D. Nichols, and M.B. Sporn. 1979. Retinoid-dependent induction of the in vivo and in vitro metabolism of retinoic acid in tissues of the vitamin A- deficient hamster. J Biol Chem 254: 6303-9.

Roberts, C., S. Ivins, A.C. Cook, A. Baldini, and P.J. Scambler. 2006. Cyp26 genes a1, b1 and c1 are down-regulated in Tbx1 null mice and inhibition of Cyp26 enzyme function produces a phenocopy of DiGeorge Syndrome in the chick. Hum Mol Genet. 15: 3394-410. Epub 2006 Oct 17.

Roberts, C., S.M. Ivins, C.T. James, and P.J. Scambler. 2005. Retinoic acid down-regulates Tbx1 expression in vivo and in vitro. Dev Dyn 232: 928-38.

167

Rongnoparut, P. and S. Weaver. 1991. Isolation and characterization of a cytosolic aldehyde dehydrogenase-encoding cDNA from mouse liver. Gene 101: 261-5.

Ross, A.C. and R. Zolfaghari. 2004. Regulation of hepatic retinol metabolism: perspectives from studies on vitamin A status. J Nutr 134: 269S-275S.

Ross, S.A., P.J. McCaffery, U.C. Drager, and L.M. De Luca. 2000. Retinoids in embryonal development. Physiol Rev 80: 1021-54.

Rossant, J., R. Zirngibl, D. Cado, M. Shago, and V. Giguere. 1991. Expression of a retinoic acid response element-hsplacZ transgene defines specific domains of transcriptional activity during mouse embryogenesis. Genes Dev 5: 1333-44.

Ruberte, E., P. Dolle, A. Krust, A. Zelent, G. Morriss-Kay, and P. Chambon. 1990. Specific spatial and temporal distribution of retinoic acid receptor gamma transcripts during mouse embryogenesis. Development 108: 213-22.

Ruberte, E., V. Friederich, P. Chambon, and G. Morriss-Kay. 1993. Retinoic acid receptors and cellular retinoid binding proteins. III. Their differential transcript distribution during mouse nervous system development. Development 118: 267-82.

Sahly, I., V. Fabre, S. Vyas, A. Milet, J.D. Rouzeau, M. Hamon, M. Lazar, and F. Tronche. 2007. 5-HT1A-iCre, a new transgenic mouse line for genetic analyses of the serotonergic pathway. Mol Cell Neurosci.

Sakai, Y., C. Meno, H. Fujii, J. Nishino, H. Shiratori, Y. Saijoh, J. Rossant, and H. Hamada. 2001. The retinoic acid-inactivating enzyme CYP26 is essential for establishing an uneven distribution of retinoic acid along the anterio- posterior axis within the mouse embryo. Genes Dev 15: 213-25.

Sandell, L.L., B.W. Sanderson, G. Moiseyev, T. Johnson, A. Mushegian, K. Young, J.P. Rey, J.X. Ma, K. Staehling-Hampton, and P.A. Trainor. 2007. RDH10 is essential for synthesis of embryonic retinoic acid and is required for limb, craniofacial, and organ development. Genes Dev 21: 1113-24.

Sapin, V., S.J. Ward, S. Bronner, P. Chambon, and P. Dolle. 1997. Differential expression of transcripts encoding retinoid binding proteins and retinoic acid receptors during placentation of the mouse. Dev Dyn 208: 199-210.

Satre, M.A., J.D. Penner, and D.M. Kochhar. 1989. Pharmacokinetic assessment of teratologically effective concentrations of an endogenous retinoic acid metabolite. Teratology 39: 341-8.

Scambler, P.J. 2000. The 22q11 deletion syndromes. Hum Mol Genet 9: 2421-6.

Schiltz, R.L., C.A. Mizzen, A. Vassilev, R.G. Cook, C.D. Allis, and Y. Nakatani. 1999. Overlapping but distinct patterns of histone acetylation by the human coactivators p300 and PCAF within nucleosomal substrates. J Biol Chem 274: 1189-92.

168

Schmidt, C.K., A. Brouwer, and H. Nau. 2003. Chromatographic analysis of endogenous retinoids in tissues and serum. Anal Biochem 315: 36-48.

Schmidt-Ullrich, R. and R. Paus. 2005. Molecular principles of hair follicle induction and morphogenesis. Bioessays 27: 247-61.

Schneider, R.A., D. Hu, J.L. Rubenstein, M. Maden, and J.A. Helms. 2001. Local retinoid signaling coordinates forebrain and facial morphogenesis by maintaining FGF8 and SHH. Development 128: 2755-67.

Serpente, P., S. Tumpel, N.B. Ghyselinck, K. Niederreither, L.M. Wiedemann, P. Dolle, P. Chambon, R. Krumlauf, and A.P. Gould. 2005. Direct crossregulation between retinoic acid receptor {beta} and Hox genes during hindbrain segmentation. Development 132: 503-13.

Singh, M. and C.E. Griffiths. 2006. The use of retinoids in the treatment of photoaging. Dermatol Ther 19: 297-305.

Sirbu, I.O., L. Gresh, J. Barra, and G. Duester. 2005. Shifting boundaries of retinoic acid activity control hindbrain segmental gene expression. Development 132: 2611-22.

St-Jacques, B., H.R. Dassule, I. Karavanova, V.A. Botchkarev, J. Li, P.S. Danielian, J.A. McMahon, P.M. Lewis, R. Paus, and A.P. McMahon. 1998. Sonic hedgehog signaling is essential for hair development. Curr Biol 8: 1058-68.

Sulik, K.K., C.S. Cook, and W.S. Webster. 1988. Teratogens and craniofacial malformations: relationships to cell death. Development 103 Suppl: 213-31.

Swindell, E.C., C. Thaller, S. Sockanathan, M. Petkovich, T.M. Jessell, and G. Eichele. 1999. Complementary domains of retinoic acid production and degradation in the early chick embryo. Dev Biol 216: 282-96.

Tahayato, A., P. Dolle, and M. Petkovich. 2003. Cyp26C1 encodes a novel retinoic acid- metabolizing enzyme expressed in the hindbrain, inner ear, first branchial arch and tooth buds during murine development. Gene Expr Patterns 3: 449-54.

Taimi, M., C. Helvig, J. Wisniewski, H. Ramshaw, J. White, M. Amad, B. Korczak, and M. Petkovich. 2004. A novel human cytochrome P450, CYP26C1, involved in metabolism of 9-cis and all-trans isomers of retinoic acid. J Biol Chem 279: 77-85.

Takahashi, Y.I., J.E. Smith, M. Winick, and D.S. Goodman. 1975. Vitam A deficiency and fetal growth and development in the rat. J Nutr 105: 1299-310.

Tao, H., K. Ono, H. Kurose, S. Noji, and H. Ohuchi. 2006. Exogenous FGF10 can rescue an eye- open at birth phenotype of Fgf10-null mice by activating activin and TGFalpha-EGFR signaling. Dev Growth Differ 48: 339-46.

169

Tao, H., M. Shimizu, R. Kusumoto, K. Ono, S. Noji, and H. Ohuchi. 2005. A dual role of FGF10 in proliferation and coordinated migration of epithelial leading edge cells during mouse eyelid development. Development 132: 3217-30.

Tapadia, M.D., D.R. Cordero, and J.A. Helms. 2005. It's all in your head: new insights into craniofacial development and deformation. J Anat 207: 461-77.

Taylor, I.M., M.J. Wiley, and A. Agur. 1980. Retinoic acid-induced heart malformations in the hamster. Teratology 21: 193-7.

Thompson, J.N., J.M. Howell, and G.A. Pitt. 1964. Vitamin a and Reproduction in Rats. Proc R Soc Lond B Biol Sci 159: 510-35.

Torchia, J., D.W. Rose, J. Inostroza, Y. Kamei, S. Westin, C.K. Glass, and M.G. Rosenfeld. 1997. The transcriptional co-activator p/CIP binds CBP and mediates nuclear-receptor function. Nature 387: 677-84.

Toyooka, Y., N. Tsunekawa, Y. Takahashi, Y. Matsui, M. Satoh, and T. Noce. 2000. Expression and intracellular localization of mouse Vasa-homologue protein during germ cell development. Mech Dev 93: 139-49.

Turksen, K. and T.C. Troy. 2002. Permeability barrier dysfunction in transgenic mice overexpressing claudin 6. Development 129: 1775-84.

Uehara, M., K. Yashiro, S. Mamiya, J. Nishino, P. Chambon, P. Dolle, and Y. Sakai. 2007. CYP26A1 and CYP26C1 cooperatively regulate anterior-posterior patterning of the developing brain and the production of migratory cranial neural crest cells in the mouse. Dev Biol 302: 399-411.

Ulven, S.M., T.E. Gundersen, A.K. Sakhi, J.C. Glover, and R. Blomhoff. 2001. Quantitative axial profiles of retinoic acid in the embryonic mouse spinal cord: 9-cis retinoic acid only detected after all-trans-retinoic acid levels are super-elevated experimentally. Dev Dyn 222: 341-53.

Valcarcel, R., H. Holz, C.G. Jimenez, D. Barettino, and H.G. Stunnenberg. 1994. Retinoid- dependent in vitro transcription mediated by the RXR/RAR heterodimer. Genes Dev 8: 3068-79.

Van Wauwe, J.P., M.C. Coene, J. Goossens, G. Van Nijen, W. Cools, and W. Lauwers. 1988. Ketoconazole inhibits the in vitro and in vivo metabolism of all-trans-retinoic acid. J Pharmacol Exp Ther 245: 718-22.

Verani, R., I. Cappuccio, P. Spinsanti, R. Gradini, A. Caruso, M.C. Magnotti, M. Motolese, F. Nicoletti, and D. Melchiorri. 2007. Expression of the Wnt inhibitor Dickkopf-1 is required for the induction of neural markers in mouse embryonic stem cells differentiating in response to retinoic acid. J Neurochem 100: 242-50.

170

Vermot, J., K. Niederreither, J.M. Garnier, P. Chambon, and P. Dolle. 2003. Decreased embryonic retinoic acid synthesis results in a DiGeorge syndrome phenotype in newborn mice. Proc Natl Acad Sci U S A 100: 1763-8.

Viallet, J.P. and D. Dhouailly. 1994. Retinoic acid and mouse skin morphogenesis. I. Expression pattern of retinoic acid receptor genes during hair vibrissa follicle, plantar, and nasal gland development. J Invest Dermatol 103: 116-21.

Vieux-Rochas, M., L. Coen, T. Sato, Y. Kurihara, Y. Gitton, O. Barbieri, K.L. Blay, G. Merlo, M. Ekker, H. Kurihara, P. Janvier, and G. Levi. 2007. Molecular dynamics of retinoic Acid-induced craniofacial malformations: implications for the origin of gnathostome jaws. PLoS ONE 2: e510.

Wei, L.N. 2003. Retinoid receptors and their coregulators. Annu Rev Pharmacol Toxicol 43: 47- 72.

Wei, L.N., J.R. Mertz, D.S. Goodman, and M.C. Nguyen-Huu. 1987. Cellular retinoic acid- and cellular retinol-binding proteins: complementary deoxyribonucleic acid cloning, chromosomal assignment, and tissue specific expression. Mol Endocrinol 1: 526-34.

Weston, A.D., B. Blumberg, and T.M. Underhill. 2003. Active repression by unliganded retinoid receptors in development: less is sometimes more. J Cell Biol 161: 223-8.

White, J.A., B. Beckett-Jones, Y.D. Guo, F.J. Dilworth, J. Bonasoro, G. Jones, and M. Petkovich. 1997. cDNA cloning of human retinoic acid-metabolizing enzyme (hP450RAI) identifies a novel family of cytochromes P450. J Biol Chem 272: 18538-41.

White, J.A., Y.D. Guo, K. Baetz, B. Beckett-Jones, J. Bonasoro, K.E. Hsu, F.J. Dilworth, G. Jones, and M. Petkovich. 1996. Identification of the retinoic acid-inducible all-trans- retinoic acid 4- hydroxylase. J Biol Chem 271: 29922-7.

White, J.A., H. Ramshaw, M. Taimi, W. Stangle, A. Zhang, S. Everingham, S. Creighton, S.P. Tam, G. Jones, and M. Petkovich. 2000a. Identification of the human cytochrome P450, P450RAI-2, which is predominantly expressed in the adult cerebellum and is responsible for all-trans-retinoic acid metabolism. Proc Natl Acad Sci U S A 97: 6403-8.

White, J.C., M. Highland, and M. Clagett-Dame. 2000b. Abnormal development of the sinuatrial venous valve and posterior hindbrain may contribute to late fetal resorption of vitamin A- deficient rat embryos. Teratology 62: 374-84.

White, J.C., M. Highland, M. Kaiser, and M. Clagett-Dame. 2000c. Vitamin A deficiency results in the dose-dependent acquisition of anterior character and shortening of the caudal hindbrain of the rat embryo. Dev Biol 220: 263-84.

Williams, J.B. and J.L. Napoli. 1987. Inhibition of retinoic acid metabolism by imidazole antimycotics in F9 embryonal carcinoma cells. Biochem Pharmacol 36: 1386-8.

171

Wilson, J.G., C.B. Roth, and J. Warkany. 1953. An analysis of the syndrome of malformations induced by maternal vitamin A deficiency. Effects of restoration of vitamin A at various times during gestation. Am J Anat 92: 189-217.

Wolbach, S.B. and P.R. Howe. 1978. Nutrition Classics. The Journal of Experimental Medicine 42: 753-77, 1925. Tissue changes following deprivation of fat-soluble A vitamin. S. Burt Wolbach and Percy R. Howe. Nutr Rev 36: 16-9.

Wolf, G. 1978. A historical note on the mode of administration of vitamin A for the cure of night blindness. Am J Clin Nutr 31: 290-2.

Wolpert, L. 1969. Positional information and the spatial pattern of cellular differentiation. J Theor Biol 25: 1-47.

Yamakado, M. and T. Yohro. 1979. Subdivision of mouse vibrissae on an embryological basis, with descriptions of variations in the number and arrangement of sinus hairs and cortical barrels in BALB/c (nu/+; nude, nu/nu) and hairless (hr/hr) strains. Am J Anat 155: 153- 73.

Yashiro, K., X. Zhao, M. Uehara, K. Yamashita, M. Nishijima, J. Nishino, Y. Saijoh, Y. Sakai, and H. Hamada. 2004. Regulation of retinoic acid distribution is required for proximodistal patterning and outgrowth of the developing mouse limb. Dev Cell 6: 411- 22.

Yip, J.E., V.G. Kokich, and T.H. Shepard. 1980. The effect of high doses of retinoic acid on prenatal craniofacial development in Macaca nemestrina. Teratology 21: 29-38.

Zamboni, L. and S. Upadhyay. 1983. Germ cell differentiation in mouse adrenal glands. J Exp Zool 228: 173-93.

Zamir, I., H.P. Harding, G.B. Atkins, A. Horlein, C.K. Glass, M.G. Rosenfeld, and M.A. Lazar. 1996. A nuclear hormone receptor corepressor mediates transcriptional silencing by receptors with distinct repression domains. Mol Cell Biol 16: 5458-65.

Zgombic-Knight, M., H.L. Ang, M.H. Foglio, and G. Duester. 1995. Cloning of the mouse class IV alcohol dehydrogenase (retinol dehydrogenase) cDNA and tissue-specific expression patterns of the murine ADH gene family. J Biol Chem 270: 10868-77.

Zhang, L., T. Zhong, Y. Wang, Q. Jiang, H. Song, and Y. Gui. 2006. TBX1, a DiGeorge syndrome candidate gene, is inhibited by retinoic acid. Int J Dev Biol 50: 55-61.

Zhao, D., P. McCaffery, K.J. Ivins, R.L. Neve, P. Hogan, W.W. Chin, and U.C. Drager. 1996. Molecular identification of a major retinoic-acid-synthesizing enzyme, a retinaldehyde- specific dehydrogenase. Eur J Biochem 240: 15-22.

172

Appendix

Apoptotic Extinction of Germ Cells in Testes of Cyp26b1 Knockout Mice

173

A.1 Summary

Cyp26b1 encodes a retinoic acid (RA) metabolizing cytochrome P450 enzyme that is expressed in embryonic tissues undergoing morphogenesis including the testes. We have generated transgenic mice lacking Cyp26b1, and have observed increased RA levels in embryonic testes. Cyp26b1−/− germ cells prematurely enter meiosis at embryonic day

(E)13.5, and appear to arrest at pachytene stage. Furthermore, after E13.5, a rapid increase in apoptosis is observed in male germ cells derived from Cyp26b1−/− embryos; germ cells are essentially absent in mutant male neonates. In contrast, testicular somatic cells appear to develop normally in the absence of Cyp26b1. Moreover, ovarian germ and somatic cells appear unaffected by lack of CYP26B1. We also show that the synthetic retinoid Am580, which is resistant to CYP26 metabolism, induces meiosis of male germ cells in cultured gonads, suggesting that abnormal development of germ cells in the Cyp26b1−/− testes results from excess RA rather than the absence of CYP26B1- generated metabolites of RA. These results provide evidence that CYP26B1 maintains low levels of RA in the developing testes which blocks entry into meiosis and acts as a survival factor to prevent apoptosis of male germ cells.

174

A.2 Introduction

Retinoic acid (RA) is a potent signaling molecule that regulates cell proliferation, differentiation, and apoptosis during embryonic development. The morphogenesis of embryonic tissues depends on spatio-temporally controlled RA signaling that is achieved through regulation of RA synthesis and catabolism (Pennimpede et al. 2006). RA dispersal must be tightly regulated for the normal development of numerous tissues, including the limb, hindbrain, neural tube, and male gonad.

RA signaling is mediated through two families of nuclear receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs) which form functional heterodimers. Each family consists of three separate genes , and with several isoforms that differ in tissue distribution (Chambon 1996). In the absence of RA,

RAR/RXR heterodimers bind to RA response elements (RAREs) and are complexed with transcriptional corepressors associated with histone deacetylase activity, resulting in chromatin condensation, and transcriptional silencing. In the presence of RA, corepressors are released permitting liganded RAR/RXR to recruit coactivator complexes possessing histone transacetylase activity which induces chromatin decondensation and subsequent activation of target gene transcription.

RA is synthesized by a family of retinaldehyde dehydrogenases (ALDH1A1,

ALDH1A2 and ALDH1A3) that irreversibly oxidize retinal to form RA (Niederreither and Dolle 2006). Conversely, RA is catabolized by a family of cytochrome P450 (CYP) enzymes, CYP26A1, CYP26B1 and CYP26C1, which convert RA to more polar metabolites for excretion (White et al. 1996; White et al. 2000a; Tahayato et al. 2003).

175

Regions where Cyp26 genes are expressed are devoid of RA, and the complementary expression of CYP26 and ALDH is responsible for defining tissue distribution of RA.

Perturbations of this endogenous distribution has severe consequences for the embryo, as mice lacking either Cyp26a1 or Cyp26b1 die in utero, or shortly after birth, and exhibit abnormalities consistent with those seen in RA teratogenesis (Abu-Abed et al. 2001;

Sakai et al. 2001; Yashiro et al. 2004).

While the role of CYP26B1 in regulating limb development has been documented

(Yashiro et al. 2004), the function of this enzyme in other organ systems including the embryonic testes has not been studied extensively. Cyp26b1 is expressed in somatic cells of the embryonic testes (Abu-Abed et al. 2002; Bowles et al. 2006), while Aldh1a2 and

Aldh1a3 are detected in the mesonephros (Mic et al. 2002; Niederreither et al. 2002c).

Organ culture experiments have shown that embryonic testes grown in the presence of

RA express Stra8, which normally is expressed in embryonic germ cells of ovaries

(Koubova et al. 2006). Furthermore, expression of Stra8 and Scp3 is upregulated in E13.5 testes in Cyp26b1 null embryos, raising the possibility that Cyp26b1 expression prevents entry of male germ cells into meiosis (Bowles et al. 2006). We have generated a transgenic mouse line lacking Cyp26b1, and observed that in addition to preventing meiotic entry of male germ cells, CYP26B1 is a critical survival factor for germ cells in the embryonic testis. Analysis of Cyp26b1−/− mice revealed increased levels of RA in the embryonic testes, resulting in initiation of meiosis and increased apoptosis of male germ cells as early as E13.5. Testes of neonatal Cyp26b1−/− animals are largely devoid of germ cells, while somatic (Sertoli and Leydig) cell differentiation appears unaffected.

176

A.3 Materials and Methods A.3.1 Construction of the targeting vector and generation of Cyp26b1-deficient mice

A bacterial artificial clone containing the murine Cyp26b1 locus was obtained from a SV129 genomic library screened by the Centre for Applied Genomics (TCAG,

Toronto, ON). Restriction digest of the clone with BamHI generated an 8.9 kb fragment spanning exons 3-6 that was gel-purified and cloned into the vector pBS (Stratagene, La

Jolla, CA). SpeI and NotI restriction sites were added by site-directed mutagenesis into intron 2 and the 3‟-UTR respectively. An oligonucleotide containing a LoxP site as well as DraIII, NaeI and PvuII, restriction sites was ligated into the introduced SpeI site. A neomycin resistance marker flanked by LoxP sites was cloned into the NotI site downstream of the stop codon (Metzger et al. 1999). The targeting construct was linearized and then electroporated into embryonic stem (ES) cells as previously described

(Abu-Abed et al. 2001). Homologous recombination of the 3‟ end of the construct in

G418 resistant ES cells was established by Southern blot analysis, and integration of both

5‟ and 3‟ ends of the construct was verified by PCR.

ES cells with one targeted Cyp26b1 allele (L3/+) were injected into blastocysts as described elsewhere (Abu-Abed et al. 2001). Agouti mice with germ-line transmission of the L3 allele were bred with CMV-Cre mice, which act as a deletor strain for floxed alleles (Dupe et al. 1997). Resultant offspring were genotyped by PCR for excision of exons 3-6 corresponding to the Cyp26b1 allele. Cyp26b1+/− mice were then crossed to generate Cyp26b1−/− animals.

177

PCR was used to genotype Cyp26b1 mice using DNA isolated from either tail clips or yolk sacs. A standard PCR was performed using Taq polymerase (Sigma-

Aldrich, Oakville, ON, Canada), and 3 primers, P1: 5‟-CAGTAGATGTTTGAGTGACA

CAGCC, P2:5‟-GAGGAAGTGTCAGGAGAAGTGG and P3:5‟-GGGCCACCAAGGA

AGATGCTGAGG. P1 and P2 amplify a product of 223 bp from the wild-type allele, or a product of 284 bp from a targeted (L3) allele, but will not amplify a null allele. P1 and P3 will amplify a 364 bp product from only the excised allele. All reactions included 1.5 mM MgCl2, 0.2 mM dNTPs, 0.4 mM of P1 and 0.2 mM of P2 and P3. Each PCR consisted of 30 cycles of 94 ˚C for 20 sec, 58 ˚C for 45 sec and 72 ˚C for 50 sec.

Resultant products were visualized by ethidium bromide staining after electrophoresis in a 2% agarose gel.

A.3.2 Embryo collection and histology

Timed matings were established by housing female mice with males overnight.

Females were checked for vaginal plugs, and noon of the day that a plug appeared was denoted as E0.5. Freshly dissected gonads from wild-type and mutant mice were fixed for

1 h in Bouin‟s solution and paraffin embedded. Sections (5 m) were cut, dewaxed, and stained with hematoxylin and eosin as previously described (Mark et al. 1993). At least four gonads from embryos of each genotype were examined at every developmental stage.

All animal experimentation was reviewed and approved by the University‟s

Animal Care and Use Committee. 178

A.3.3 Detection of RA activity

For RA detection, a previously described retinoic acid reporter cell line was used

(Loudig et al. 2005). The reporter cell line was derived from stable transfection of mouse

P19 teratocarcinoma cells with a transgene consisting of the murine Cyp26a1 promoter fused to a firefly luciferase gene. Eight XY gonads from E12.5 Cyp26b1+/+, Cyp26b1+/− or Cyp26b1−/− embryos were minced finely in minimal essential medium (Gibco

Invitrogen Corporation, Burlington, ON), subjected to freeze/thaw, and centrifuged at

12,000g for 10 min at 4oC. The supernatant was added to confluent P19 reporter cells growing in a 24-well plate wrapped in aluminum foil and incubated for 24 h at 37˚C, 5%

CO2. Then, media was aspirated and cells were washed twice with PBS. Cells were lysed by adding 50 μL of Passive Lysis Buffer (Promega, San Luis Obispo, CA) and cell lysates were transferred to a 96-well assay plate. Readings were taken using a Berthold

MicroLumat Plus LB 96V luminometer in the presence of Firefly Luciferase Assay

Substrate (Promega). Each sample pool was assayed in triplicate.

A.3.4 Immunohistochemistry and Terminal Deoxynucleotidyl Transferase dUTP Nick end Labeling (TUNEL) analysis

For immunohistochemistry, sections were blocked in PBS with 10% goat or rabbit serum and 0.1% Triton X-100 for 30 min at room temperature before staining with primary antibodies. Sections were incubated overnight at 4oC with antibodies directed against Müllerian-inhibiting substance (MIS, 1:500; Santa Cruz Biotechnology, Santa

Cruz, CA), mouse vasa homologue protein (MVH, 1:5000, kindly provided by Dr.

179

Toshiaki Noce), 3β-hydroxysteroid dehydrogenase (3βHSD, 1:5000, a kind gift from Dr.

Anita Payne) or activated caspase-3 (1:200, Cell Signaling Technology, Inc., Danvers,

MA). After three washes in PBS, the sections were incubated for 30 min at room temperature with biotinylated anti-rabbit or anti-goat secondary antibody at a dilution of

1:500 (Vector Laboratories, Burlingame, CA). Sections were washed three times in PBS for 5 min each and then incubated with peroxidase-conjugated streptavidin (Zymed

Laboratories, South San Francisco, CA) for 10 min at room temperature. Antibody localization was determined by application of diaminobenzidine (DAB, Zymed

Laboratories). Negative controls, omitting the primary antisera, were included in each experiment.

For the TUNEL assay, sections were dewaxed and boiled for 10 min in 10 mM citrate buffer. TUNEL was performed using the DeadEnd Colorimetric TUNEL System

(Promega) and sections were counterstained with hematoxylin. The TUNEL-positive cells in the entire cross section were counted using a light microscope. At least six cross sections were analyzed for each developmental stage.

A.3.5 Chromosome spreading, preparation and immunofluorescence

For analysis of male meiotic chromosomes, embryonic testes were removed between E12.5 and E16.5, and chromosome spreads were prepared as described previously (Peters et al. 1997). Slides containing chromosome spreads were subjected to immunofluorescent staining as described (Lammers et al. 1994). Polyclonal rabbit anti-

SCP3 antiserum (1:500, a kind gift from Dr. Christa Heyting) was applied overnight at

180 room temperature. Then slides were washed three times in PBS for 5 min each and incubated with goat anti-rabbit Alexa 488 (Molecular Probes, Eugene, OR) diluted 1:500 for 60 min at 37oC. After a further three washes in PBS, slides were mounted with

Fluorescence Mounting Medium (Dako, Carpinteria, CA) and analyzed by confocal microscopy with a Leica TCS MP2 confocal microscope.

A.3.6 Organ Culture of Embryonic Gonads

E12.5 genital ridges were collected from wild-type embryos and cultured as described previously (Li and Kim 2004). Briefly, paired genital ridges were dissected and cultured on Millicell CM filters (Millipore, Bedford, MA) floating on the surface of

Dulbecco's modified Eagle medium (Invitrogen, Burlington, ON, Canada) supplemented with penicillin/streptomycin (25 units/mL) and incubated at 37˚C with 5% CO2. One genital ridge was cultured for 48 hours in medium containing 1 nM of the synthetic retinoid Am580 (Biomol Research Lab, Plymouth Meeting, PA). The other genital ridge from the same embryo was cultured in medium containing an equivalent volume of dimethyl sulfoxide (solvent for Am580) as control. Two independent organ culture experiments were conducted, with each experiment consisting of two pairs of genital ridges.

181

A. 4 Results A.4.1 Generation of CYP26B1-deficient mice

Mice lacking CYP26B1 were generated by deleting exons 3-6 of the endogenous locus (Figure A.1). Cyp26b1+/− mice were fertile and phenotypically indistinguishable from wild-type littermates. When Cyp26b1+/− mice were intercrossed, Cyp26b1−/− offspring were born at the expected Mendelian ratio, but died within hours of birth due most likely to respiratory distress. Cyp26b1−/− embryos exhibit truncated limbs, and severe craniofacial abnormalities as have been previously reported (Yashiro et al. 2004).

A.4.2 Loss of CYP26B1 function elevates RA levels in embryonic testes

The effect of CYP26B1 deficiency on RA levels in embryonic testes was assessed using a RA reporter cell line, which has been shown to be highly sensitive to RA (Loudig et al. 2005). Eight E12.5 testes from Cyp26b1+/+, Cyp26b1+/− or Cyp26b1−/− mice were collected and pooled. Extracts from each sample set were incubated with the reporter cells. Luciferase activity in response to the presence of RA was assayed as a means of determining the relative RA content of samples. Testes from Cyp26b1−/− mice showed approximately a 3-fold increase in RA levels in comparison to samples from Cyp26b1+/− or Cyp26b1+/+ testes (Figure A.2).

182

Figure A.1. Generation of null alleles of Cyp26b1. (A) The endogenous murine

Cyp26b1 locus is shown, consisting of six exons spanning 13.6 kb. The targeting vector consists of a subcloned 8.9 kb BamHI fragment with three LoxP sites (arrowheads) and a neomycin resistance marker inserted. After chimeric mice harboring a targeted allele were generated, they were crossed with mice constitutively expressing Cre recombinase and resultant offspring were screened by PCR for possible alleles (null, L3).

Heterozygous mice were crossed to generate null mutant mice. (B) Results of a genotyping PCR differentiating wild-type (+/+), heterozygous (+/−) and homozygous null (−/−) mice. PCR mix containing primers P1, P2 and P3 amplifies a product of 223 bp from the wild-type allele and 364 bp from a null allele.

183

184

Figure A.2 Elevated RA levels in E12.5 testes from Cyp26b1−/− mice. Extracts prepared from eight Cyp26b1+/+, Cyp26b1+/− or Cyp26b1−/− testes were incubated with RA reporter cells. Relative luciferase activity produced in response to RA was assayed as a means to determine relative RA levels of the samples. The results are plotted as relative levels of

RA in Cyp26b1+/− or Cyp26b1−/− samples versus Cyp26b1+/+ littermates. Each sample was assayed in triplicate and error bar represents SEM.

185

186

A.4.3 Loss of germ cells in Cyp26b1−/− testes

Testes of Cyp26b1−/− neonates were significantly smaller than wild-type littermates (results not shown). Immunohistochemistry was performed on testis sections using an antibody directed against mouse vasa homologue protein (MVH), a cytoplasmic antigen present in germ cells (Toyooka et al. 2000).

In contrast to Cyp26b1+/+ testes where many germ cells were observed (Figure

A.3, A and C), virtually no germ cells were found in Cyp26b1 −/− testes (Figure A.3, B and D). Neonatal ovaries were also examined, and there were comparable numbers of germ cells in Cyp26b1−/−and wild-type littermates (results not shown).

Male and female gonads have also been examined in Cyp26a1−/− mutant neonates, and no abnormalities in cell number or morphology were noted (Li and Petkovich, unpublished results).

187

Figure A.3 Germ cell loss in newborn Cyp26b1−/− testes. Newborn Cyp26b1+/+ and

Cyp26b1−/− testes stained with hematoxylin and eosin (A and B) or with antibody to

MVH (C and D). Bar in (D), 200 μm (A-D).

188

189

A.4.4 Histological abnormalities of germ cells in embryonic Cyp26b1-/- testes

As virtually no germ cells were found in Cyp26b1−/− neonatal testes, embryonic testes were examined at various stages. The morphology of the testis cords in Cyp26b1−/− mice was normal at all stages examined, while there were significant germ cell abnormalities. At E12.5, Cyp26b1−/− testes were indistinguishable from wild-type testes

(data not shown). In mutants at E13.5, some germ cells exhibited dark, condensed nuclei

(Figure A.4B), typical of apoptotic cells, while such cells were rarely observed in the testes of wild-type littermates (Figure A.4A). By E14.5, some germ cells in Cyp26b1−/− testes appeared to be apoptotic cells, while others appeared to be in zygotene stage of meiosis (Figure A.4D). At E16.5, there were fewer germ cells in Cyp26b1−/− mice (Figure

A.4F) compared to control mice (Figure A.4E). Cells that were present appeared to be either apoptotic or in zygotene/pachytene stage of meiosis.

A.4.5 Increased apoptosis in Cyp26b1-/- testes

TUNEL analysis was performed on E13.5, E14.5 and E16.5 Cyp26b1−/− testes to determine if germ cell loss in Cyp26b1−/− testes was due to apoptosis. Wild-type testes had very few TUNEL positive cells (Figure A.5, A, C and E), while Cyp26b1−/− testes had significantly more TUNEL-positive cells at all stages examined (Figure A.5, B, D, F and G). Most TUNEL-positive cells were found within the testis cords, indicating that apoptotic cells were germ cells. Similar results were obtained using a marker of apoptotic death, the active form of caspase-3 (Figure A.5, insets).

190

Figure A.4 Histology of embryonic Cyp26b1+/+ and Cyp26b1−/− testes. (A and B)

E13.5; (C and D) E14.5; (E and F), E16.5. Insets show individual germ cells. Some germ cells exhibited apoptotic morphology (arrowheads) whereas others appeared to be meiotic

(arrows). Bar in (F), 25 μm (A-F).

191

192

Figure A.5 Increased number of TUNEL-positive cells in Cyp26b1−/− mice. (A and B)

E13.5; (C and D) E14.5; (E and F), E16.5. Insets show immunohistochemical staining of activated caspase-3. Bar in (F), 100 μm (A-F). (G) Bar graph indicating the number of

TUNEL-positive cells per section. Values are mean ± SEM. *, significant at P < 0.05; **, significant at P < 0.01 compared with the controls (Student‟s t-test).

193

194

A.4.6 Germ cells enter meiosis in embryonic Cyp26b1−/− testes

Chromosome spread preparations from E12.5 to E16.5 were stained with an antibody to SCP3, which is an indicator of leptotene stage of meiosis, to differentiate between the germ cells at different meiotic stages (McClellan et al. 2003). In wild-type mice, there was no SCP3 staining in germ cells collected from E12.5-E16.5 testes (results not shown). In Cyp26b1−/− testes at E12.5, germ cells showed no SCP3 staining (Figure

A.6A), whereas at E13.5, there was diffuse SCP3 labeling indicative of leptotene stage

(Figure A.6B). At E14.5, in addition to cells in leptotene, samples with thickened threads characteristic of zygotene stage were detected (Figure A.6C). Finally, by E16.5, some germ cells were in zygotene while others had progressed to pachytene stage with distinct threads representative of paired chromosomes (Figure A.6D).

195

Figure A.6 Meiotic staging in Cyp26b1−/− germ cells. SCP3 immunostaining of germ cells shows the stage of meiotic prophase in Cyp26b1−/− testes at E12.5 (A-no staining),

E13.5 (B-leptotene), E14.5 (C-zygotene), and E16.5 (D-pachytene). Bar in (D), 5 μm (A-

D).

196

197

A.4.7 Sertoli and Leydig cells develop normally in Cyp26b1−/− testes

Immunohistochemistry was performed for Müllerian-inhibiting substance (MIS) and 3β-hydroxysteroid dehydrogenase (3βHSD), Sertoli and Leydig cell-specific proteins respectively (Munsterberg and Lovell-Badge 1991; Greco and Payne 1994), to determine if somatic cells develop normally in the absence of CYP26B1. MIS and 3βHSD immunoreactivities were similar in Cyp26b1+/+ and Cyp26b1−/− testes at both E14.5 and

E16.5 (Figure A.7), suggesting that Sertoli and Leydig cells develop normally in

Cyp26b1−/− testes.

A.4.8 Exposure of embryonic testes in culture to a CYP26B1-resistant retinoid increases apoptosis of germ cells

Loss of CYP26B1 function could result in morphological changes due to either increased RA levels or an absence of CYP26B1-generated metabolites. To distinguish between these possibilities, E12.5 urogenital ridges from wild-type mice were cultured in the presence of the synthetic retinoid Am580, which is a RAR-selective agonist that is resistant to CYP26B1 metabolism (our unpublished data). Treatment with 1 nM Am580 mimicked the germ cell abnormalities in Cyp26b1−/− testes, including initiation of meiosis and increased apoptosis (Figure A.8, B, D and F), while such changes were rarely seen in control testes cultured with vehicle (dimethyl sulforxide) only (Figure A.8, A, C and E).

This suggests that ectopic RA, rather than the lack of specific CYP26B1-generated metabolites are responsible for phenotypes observed in Cyp26b1−/− testes.

198

Figure A.7 Somatic cell differentiation is unaffected in Cyp26b1−/− testes. Embryonic

Cyp26b1+/+ (A, B, E, F) and Cyp26b1−/− (C, D, G, H) testes stained with antibodies to

MIS or 3 HSD. (A-D) E14.5; (E-H) E16.5. Bars, 25 μm.

199

200

Figure A.8 The synthetic retinoid Am580 induces meiosis and apoptosis of germ cells in embryonic testes. E12.5 testes were cultured for 48 hours in the presence of

DMSO or 1 nM Am580. (A and B) Testicular sections stained with hematoxylin and eosin; (C and D) TUNEL analysis; (E and F) Testicular sections stained with antibody to

SCP3. Some germ cells exhibited apoptotic morphology (arrowheads) whereas others appeared to be meiotic (arrow). Bar in (D), 25 μm (A-D); bar in (F), 5 μm (E and F).

201

202

A.5 Discussion

Loss of CYP26B1 function in embryonic testes leads to increased ectopic RA exposure that precociously triggers meiosis and apoptosis in germ cells. This suggests that CYP26B1 is responsible for establishing a barrier to protect embryonic testes from inappropriate exposure to RA, thus maintaining germ cells in mitotic quiescence until meiosis is required for spermatogenesis, and maintaining the viability of male germ cells during embryonic development.

In Cyp26b1−/− testes at E13.5, meiotic germ cells were observed, which is the same time that germ cells in ovaries enter meiosis (McLaren 1984). It has been proposed that embryonic germ cells are intrinsically programmed to enter meiosis (McLaren 1984).

Germ cells enter meiosis around E13.5 not only in ovaries, but also in regions outside the gonad such as the intervening mesonephric region and adrenal primordia of male embryos (Zamboni and Upadhyay 1983; McLaren 1984). Recently, it was speculated that

Cyp26B1 prevents premature meiosis in embryonic male germ cells, based on the observation that expression of Stra8 and Scp3 were upregulated in Cyp26b1−/− testes

(Bowles et al. 2006). In the present study, we have categorically demonstrated that male germ cells progressed through meiotic prophase in embryonic Cyp26b1−/− testes. These results provide additional evidence that CYP26B1 acts as a meiotic inhibitor in embryonic testes.

Furthermore, we have observed the striking loss of testicular germ cells due to apoptosis in the absence of Cyp26b1. Increased apoptosis is observed at the same time

(E13.5), when Cyp26b1−/− male germ cells enter meiosis. As a result, male germ cells

203 were essentially absent in newborn Cyp26b1−/− testes. This finding is significant, as it shows that in addition to preventing meiosis in germ cells, CYP26B1 activity is required for germ cell survival in the embryonic testes. We also show that RA is present in

Cyp26b1−/− testes at E12.5, and yet no morphological abnormalities are detected in germ cells until E13.5. This may reflect an indirect effect of RA or a delay between RA exposure and triggering of meiosis and apoptosis.

CYP26B1 metabolizes RA into polar derivatives, which appear to be inactive as shown by genetic, and biochemical analysis (Niederreither et al. 2002a). However, it has been suggested that 4-oxo-RA may have some specific biological role, particularly in lower vertebrates (Pijnappel et al. 1993; Achkar et al. 1996). Thus, one possible explanation for our observations is that metabolites generated by CYP26B1 may prevent the apoptosis and entry of male germ cells into meiosis. However, this is unlikely because the synthetic retinoid Am580 is not metabolized by CYP26B1, and yet can still induce meiosis and apoptosis in cultured embryonic testes. This suggests that RA is responsible for the abnormal testicular phenotypes in Cyp26b1−/− embryos.

Finally, we observed that testis cord formation and somatic cell differentiation appear essentially normal in Cyp26b1−/− embryonic testes. The male pathway of gonad development, which is determined by the expression of Sry in somatic cells, occurs within a narrow window (between E11.5 and E12.5). Although Sry was discovered a decade ago, no direct target genes of SRY have been identified (Capel 2000). Cyp26b1 is first detected in embryonic gonads of both sexes by E11.5, and becomes male-specific by

E12.5, shortly after the onset of Sry transcription, indicating that Cyp26b1 could act

204 downstream of Sry. While Sry is essential for testis cord formation, Sertoli and Leydig cell differentiation, these male-specific events occur normally in Cyp26b1−/− mice. These observations suggest that Sry and Cyp26b1 act along different pathways.

RA is synthesized in the mesonephros by ALDH1A2 and ALDH1A3, and degraded by CYP26B1, which is expressed in somatic cells of the male gonad. Under these conditions, germ cells develop normally and are maintained in a mitotic state (Fig.

9) until E13.5 when they arrest in quiescence. However, in the absence of Cyp26b1, RA diffuses from the mesonephros, across the basal lamina, where it either directly or indirectly induces meiosis and apoptosis in male germ cells. Although it appears that meiosis precedes apoptosis in germ cells, we can not exclude the possibility that some germ cells undergo apoptosis without entering meiosis. Two potential mechanisms for these observations are presented in Figure A.9; RA can either act on germ cells directly

(Figure A.9B), or induce or repress expression of a secreted factor(s) from Sertoli cells that in turn modulates germ cell meiosis and apoptosis (Figure A.9C). We presently favor the former explanation, as Cyp26b1 expression in Sertoli cells enclosing germ cells could provide a barrier to prevent RA from reaching the interior of the seminiferous cord.

Furthermore, RARs have been detected in male germ cells (Bowles et al. 2006), and female germ cells appear to enter meiosis in response to RA in vivo. However, it is also possible that meiotic germ cells undergo apoptosis in response to a factor secreted by male somatic cells. Future research will focus on discriminating between these two models, and identifying genes responsible for initiating germ cell entry into meiosis and subsequent extinction in the presence of RA.

205

Figure A.9 Models for the role of CYP26B1 in embryonic development of male germ cells. (A) In wild-type male gonads, Cyp26b1 is expressed in Sertoli and interstitial cells and degrades RA before it can cross the basal lamina. (B) In Model 1, in the absence of

CYP26B1, RA crosses the basal lamina into the seminiferous cord where RA acts directly on germ cells inducing or repressing the expression genes that result in meiosis and apoptosis. (C) In Model 2, in the absence of CYP26B1, RA diffuses into the seminiferous cord where it induces an unidentified factor(s) in Sertoli cells that is secreted and initiates meiosis and apoptosis of germ cells.

206

207

References

Capel, B. 2000. The battle of the sexes. Mech Dev 92: 89-103.

Chambon, P. 1996. A decade of molecular biology of retinoic acid receptors. Faseb J 10: 940-54.

Koubova, J., D.B. Menke, Q. Zhou, B. Capel, M.D. Griswold, and D.C. Page. 2006. Retinoic acid regulates sex-specific timing of meiotic initiation in mice. Proc Natl Acad Sci U S A 103: 2474-9.

208

Li, H. and K.H. Kim. 2004. Retinoic acid inhibits rat XY gonad development by blocking mesonephric cell migration and decreasing the number of gonocytes. Biol Reprod 70: 687-93.

Mark, M., T. Lufkin, J.L. Vonesch, E. Ruberte, J.C. Olivo, P. Dolle, P. Gorry, A. Lumsden, and P. Chambon. 1993. Two rhombomeres are altered in Hoxa-1 mutant mice. Development 119: 319-38.

McClellan, K.A., R. Gosden, and T. Taketo. 2003. Continuous loss of oocytes throughout meiotic prophase in the normal mouse ovary. Dev Biol 258: 334-48.

McLaren, A. 1984. Meiosis and differentiation of mouse germ cells. Symp Soc Exp Biol 38: 7-23.

Metzger, D., E. Scheer, A. Soldatov, and L. Tora. 1999. Mammalian TAF(II)30 is required for cell cycle progression and specific cellular differentiation programmes. Embo J 18: 4823- 34.

Munsterberg, A. and R. Lovell-Badge. 1991. Expression of the mouse anti-mullerian hormone gene suggests a role in both male and female sexual differentiation. Development 113: 613-24.

209

Niederreither, K. and P. Dolle. 2006. Molecular mediators of retinoic acid signaling during development. Adv Dev Biol 16: 107-143.

Niederreither, K., J. Vermot, V. Fraulob, P. Chambon, and P. Dolle. 2002b. Retinaldehyde dehydrogenase 2 (RALDH2)- independent patterns of retinoic acid synthesis in the mouse embryo. Proc Natl Acad Sci U S A 99: 16111-6.

Pennimpede, T., D. Cameron, and M. Petkovich. 2006. Regulation of murine embryonic patterning and morphogenesis by retinoic acid signaling. Adv Dev Biol 16: 66-104.

Peters, A.H., A.W. Plug, M.J. van Vugt, and P. de Boer. 1997. A drying-down technique for the spreading of mammalian meiocytes from the male and female germline. Chromosome Res 5: 66-8.

210

White, J.A., H. Ramshaw, M. Taimi, W. Stangle, A. Zhang, S. Everingham, S. Creighton, S.P. Tam, G. Jones, and M. Petkovich. 2000. Identification of the human cytochrome P450, P450RAI-2, which is predominantly expressed in the adult cerebellum and is responsible for all-trans-retinoic acid metabolism. Proc Natl Acad Sci U S A 97: 6403-8.

Zamboni, L. and S. Upadhyay. 1983. Germ cell differentiation in mouse adrenal glands. J Exp Zool 228: 173-93.

211

Glenn MacLean 61 Stanley St., Apt#4 Kingston, ON K7K 1Y2 [email protected]

Education: 09/1999-Present Ph.D Candidate, Department of Pathology, Queen's University Supervisor: Martin Petkovich

09/1995-05/1999 B. Sc (Honours) with distinction, Department of Biochemistry, McMaster University

Research Experience: 09/1999-Present Ph.D. Thesis Project, "Cloning, Embryonic Expression and Targeted Disruption of murine Cyp26B1"; Supervisor: Dr. Martin Petkovich

09/1998-04/1999 Honours Thesis Project, "Structure of Rat Deoxyuridine Triphosphatase Gene"; Supervisor: Dr. John Capone

05/1997-09/1999 Research Assistant for Dr. Michael Buchanan, Department of Pathology and Molecular Medicine, McMaster University

Scholarships and Academic Achievements: 09/2005 Ontario Graduate Scholarship 09/2005 Queen‟s Graduate Award 09/2004 R. Samuel McLaughlin Fellowship 09/2004 Queen‟s Graduate Award 09/2002- NSERC PSG B Scholarship (2 year award) 2004 09/2001 Ontario Graduate Scholarship 09/2001 Queen's Graduate Award 09/2000 R. Samuel McLaughlin Fellowship 09/2000 Queen's Graduate Award 09/1999 Ontario Graduate Scholarship, Science and Technology 1995-1999 Dean's Honour Society, McMaster University

Publications:

MacLean, G., Li, H., Metzger, D., Chambon, P., Petkovich, M., (2007). Apoptotic extinction of germ cells in testes of Cyp26b1 knockout mice. Endocrinology, 148, 4560-4567.

Loudig, O., MacLean, G., Dore, N., Luu, L., Petkovich, M., (2005). Transcriptional co- operativity between distant Retinoic acid response elements in regulation of Cyp26A1 inducibility. Biochem. J., 392, 241-248.

212

Abu-Abed, S., Dollé, P., Metzger, D., Wood, C., MacLean, G., Chambon, P., Petkovich, M., (2003). Developing with lethal RA levels: genetic ablation Rarg can restore the viability of mice lacking Cyp26a1. Development, 130:1449-1459.

Abu-Abed, S., MacLean, G., Fraulob, V., Chambon, P., Petkovich, M., Dollé, P., (2002) Differential expression of the retinoic acid-metabolizing enzymes CYP26A1 and CYP26B1 during murine organogenesis. Mech. Dev., 110:173-177.

MacLean, G., Abu-Abed, S., Dollé, P., Tahayato, A., Chambon, P., Petkovich, M., (2001) Cloning of a novel retinoic-acid metabolizing cytochrome P450, Cyp26B1, and comparative expression analysis with Cyp26A1 during early murine development. Mech. Dev., 107:195-201.

Buchanan, M., MacLean, G., Brister, (2001) Selective and sustained inhibition of surface-bound thrombin activity by intimatan/heparin cofactor II and its relevance to assessing systemic anticoagulation in vivo, ex vivo and in vitro. Thromb Haemost 86(3):909-13.