UNIVERSITY OF CINCINNATI

Date:______

I, ______, hereby submit this work as part of the requirements for the degree of: in:

It is entitled:

This work and its defense approved by:

Chair: ______

Expression Microarray Analysis of Renal Development and Human Renal

Disease

A dissertation submitted to the Division of Graduate Studies

and Research of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in the Graduate Program in Molecular and Developmental Biology

of the College of Medicine

2006

by

Kristopher Robert Schwab

B.A., Blackburn College, 2001

Committee Chair: S. Steven Potter, Ph.D.

Tom Doetschman, Ph.D.

Chia-Yi Kuan, M.D., Ph.D.

Dan Wiginton, Ph.D.

James Wells, Ph.D.

Abstract

Renal morphogenesis involves the reciprocal inductive interactions between the ureteric bud and metanephric mesenchyme forming the collecting ducts and nephrons within adult kidney. We applied microarray technology to the study of renal morphogenesis in order to better understand the molecular mechanisms underlying development. Additionally, the techniques employed in the expression analysis of the embryonic kidney were extended to the study of renal disease.

Embryonic kidneys representing different stages of renal development were analyzed using expression microarrays. Renal developmental analysis revealed many novel genes and genetic pathways involved in renal development.

In addition, the normal renal development data provides a baseline for the analysis of gene targeted mice possessing disruptions in renal morphogenesis.

Microarray analysis was also performed on the Hoxa11/Hoxd11 compound null renal defect throughout renal development. In conclusion, these microarray studies greatly advance our knowledge of gene expression within the normal renal morphogenesis and identify possible downstream candidate genes regulated by the Hox11 genes.

Wnt signaling is crucial for normal renal morphogenesis. In Drosophila, the pygopus gene encodes a transcriptional co-activator required for canonical

Wnt signaling. The targeted deletion of the mammalian orthologs of pygopus,

Pygo1 and Pygo2, in mice was investigated in renal development. A disruption

ii in ureteric number tip and morphology was identified in Pygo1/Pygo2 compound null kidneys. Additionally, canonical Wnt signaling as measure by the Bat-gal transgene is reduced within the ureteric compartment in Pygo1/Pygo2 null kidneys. Overall, these experiments suggest that Pygo function is required for activation of canonical Wnt signaling in the ureteric compartment of the developing kidney.

Focal segmental glomerulosclerosis (FSGS) is characterized by the segmental scarring of the glomerulus, ultimately resulting loss of nephron function. To understand the molecular pathogenesis of this disease, gene expression analysis was performed on FSGS patient kidney biopsies and compared to normal kidney tissue. Hundreds of genes were identified significantly changed with in the FSGS patient groups. Furthermore, gene expression changes were identified in subsets of patients possessing different clinical manifestations of FSGS. In conclusion, the molecular analysis of gene expression in the FSGS kidney provides a better understanding of expression changes during renal disease.

iii iv Acknowledgements

First and foremost, I would like to thank my advisor S. Steven Potter. I greatly appreciate his enthusiasm for developmental biology and excellent guidance he has given me over the last few years. Also, these studies were greatly aided by the technical expertise of both Larry Patterson and Heather

Hartman. I would also like to thank the members of my committee, Tom

Doetschman, Alex Kuan, Dan Wiginton, and James Wells, for their assistance and guidance.

Additionally, I would like to thank the past and present members of the

Potter, Patterson, and Michael Bates laboratories for their friendship and help given to me over the years.

Many thanks to the faculty, students, technicians, and staff of the

Molecular and Developmental Biology program for constructing such an excellent environment for the study for developmental biology.

Finally, a very special thanks goes to my wife, Jennifer, and the rest of our families for their love during this significant endeavor. I am grateful for your support.

v Table of contents

Abstract ii

Acknowledgements v

Table of contents 1

Chapter 1. General Introduction

Overview 5

Mammalian kidney development 7

Role of Homeoxbox (Hox) genes in kidney development 15

Wnt signaling in renal development 21

Fsgs (Focal segmental glomerulosclerosis) 24

Significance of microarray analysis in renal development and 26

disease

References 30

Figures 44

Chapter 2. A catalogue of gene expression in the developing kidney

Abstract 51

Introduction 53

Materials and methods 57

Results 60

Discussion 77

Acknowledgements 79

References 80

Figures and tables 91

1 Chapter 3. Comprehensive microarray analysis of Hoxa11/Hoxd11 mutant kidney development

Abstract 106

Introduction 107

Materials and methods 110

Results 114

Discussion 124

Appendix 134

Acknowledgements 135

References 136

Figures and tables 148

Chapter 4. Pygo1 and Pygo2 roles in Wnt signaling in mammalian kidney development

Abstract 159

Introduction 160

Materials and methods 163

Results 169

Discussion 179

Acknowledgements 185

References 186

Figures and tables 193

2 Chapter 5. Microarray analysis of focal segmental glomerulosclerosis (FSGS)

Abstract 210

Introduction 211

Materials and methods 213

Results 216

Discussion 229

Appendix 231

Acknowledgements 232

References 233

Figures and tables 239

Chapter 6. General Discussion

Microarray expression analysis of kidney development 247

Microarray expression analysis of Hoxa11/Hoxd11 null mutants 249

Analysis of Pygo1/Pygo2 null mutants 252

Focal segmental glomerulosclerosis 254

References 257

3 Chapter 1

General Introduction

4 Overview

How a single fertilized egg differentiates into the many diverse tissues and organ systems of an organism is a fundamental question of developmental biology. Embryogenesis relies on many different processes, such as cell proliferation, apoptosis, polarization, migration, and differentiation to occur within a specific spatio-temporal manner generating the basic body plan, limbs, and organs of the organism. A common theme in organogenesis is the requirement of specific instructive interactions between different tissue types; one example is the interaction of the epithelial tube with the surrounding unorganized mesenchyme in the developing metanephric kidney. The ablation of either the mesenchyme or epithelial component in this system results in a remarkable loss of developmental potential. Likewise, the disruption of specific molecular signals and cell-cell interactions between these tissues result in abnormal development.

The formation of the developing kidney relies on many complex genetic and signaling pathways between the ureteric bud, an epithelial outgrowth from nephric duct which extends throughout the intermediate mesoderm, and the surrounding metanephric mesenchyme (Dressler, 2006). The ureteric bud must undergo many rounds of branching while inducing the surrounding metanephric mesenchyme to undergo a mesenchymal to epithelial transformation forming renal vesicles. These vesicles undergo nephrogenesis, a complex process of elongation and differentiation forming the nephron, the functional unit of the kidney. During these processes, the metanephric mesenchyme is also actively

5 regulating ureteric branching, while the ureteric bud tips regulate nephron induction. These reciprocal inductive interactions generate the 800,000 to

1,200,000 nephrons present in the adult human kidney functioning in the removal of metabolic wastes from the bloodstream. These developmental interactions of the metanephric kidney can easily be studied in metanephric kidney organ culture, recapitulating in vivo development and providing an excellent experimental system for further analysis (Grobstein, 1955; Grobstein, 1956).

After development has completed successfully, the organ must fulfill its physiological role within the organism while properly responding to insults and stresses from the environment or intrinsic to organism, such as genetic and autoimmune diseases. The adult kidney filters nearly 200 liters of fluid a day from the bloodstream, concentrating waste products into a significantly smaller volume while reclaiming electrolytes, metabolites, and water. Many human renal diseases result from stresses, insults, or genetic defects within the glomeruli, the site of blood filtration. The glomerulus consists of podocytes with complex cellular processes surrounding a specialized endothelium allowing the selective passage of small molecules. Focal segmental glomerulosclerosis (FSGS) is an example of such a devastating condition, resulting in the ablation of podocytes and segmental scarring of the glomeruli causing loss of nephron function which eventually leads to renal failure. Although the etiology of a small percentage of

FSGS cases is due to alleles containing mutations in podocyte-specific genes, the majority of FSGS cases are idiopathic. Due to the heterogeneity of FSGS cases, many clinicians have suggested that FSGS is a pathological diagnosis of

6 different, uncharacterized glomerular diseases resulting in similar lesions.

Consequently, many difficulties arise in clinical treatment of this disease due the ineffectiveness of common therapies in many FSGS patients.

These studies aim to further our understanding of both kidney

development and kidney disease using expression microarray technology. The

recent development of microarray technology has offered the ability to measure

the expression of thousands of genes in a given cell population providing a

wealth of gene expression information. The ability to measure gene expression

changes in small cell populations has given the developmental biologist the

ability to ascertain the complete developmental gene expression profile of

developing organs and the clinician the ability to study the molecular

pathogenesis of many diseases. The development of the microarray has greatly

enhanced our resolution of the molecular mechanisms occurring in both

development and disease.

Mammalian kidney development

The human adult kidney contains around one million nephrons, the

functional waste filtering units of the kidney (reviewed in Vize et al., 2002). A

nephron consists of epithelial tube with distinct segments (the proximal

convoluted tubule, the loop of Henle, and the distal convoluted tubule) each with specific transport functions and a glomerulus attached to the proximal convoluted tubule, the site of the vasculature integration containing many specialized cells.

The distal convoluted tubule connects the nephron to the collecting duct system

7 derived from the ureteric bud which converges into the renal pelvis. The development of the metanephric kidney proceeds through a series of evolutionary conserved steps in mammals.

Prior to the development of the metanephric kidney, two primitive kidneys are formed within the anterior intermediate mesoderm of the developing embryo: the pronephros and the mesonephros (Bouchard, 2004). These primitive kidneys precede the development of the adult metanephric kidney and form in an anterior to posterior fashion within the urogenital ridge. The pronephros is the first primitive kidney to develop consisting of the anterior nephric duct and pronephric tubules. Although pronephric tubules and pronephric duct form functioning kidneys in fish and amphibian larvae, the pronephros is believed to be rudimentary and inactive in the developing mammalian embyro. The nephric duct then extends caudally within the urogenital ridge giving rise to the second primitive kidney, the mesonephros, also containing attached mesonephric tubules. In some mammals, the mesonephric kidney functions in filtration in adult fish and amphibians, but is thought to be inactive within both humans and rodents. Both of these structures degenerate as kidney development proceeds except for portions of the mesonephros which incorporated into the male reproductive system. Studies of targeted and transgenic strains of mice have revealed genes critical to the formation of the pronephros and the mesonephros.

Embryos with compound null alleles of Pax2 and Pax8, both paired box transcription factors, possess a disruption in pronephric and mesonephric development due loss of intermediate mesoderm undergoing epithelial

8 transformation forming the nephric duct (Bouchard et al., 2002). Additionally, the targeted knockout of the Wnt ligand, Wnt9b, results in the complete loss of pronephric and mesonephric tubules demonstrating a critical role for Wnt signaling in mesenchymal to epithelial transformation of the intermediate mesoderm (Carroll et al., 2005). Although these primitive kidneys are not thought to possess critical physiological functions in mammalian development, the progression through these two primitive kidney stages has been preserved during mammalian urogenital morphogenesis indicating developmental constraint of the metanephric kidney formation in the evolution of higher vertebrates.

At E10.5, the metanephric kidney can first be identified as an outgrowth of the caudal nephric duct, the ureteric bud, surrounded by a condensing mesenchyme within the intermediate mesoderm, the metanephric mesenchyme

(Fig. 1). Early organ culture experiments have demonstrated the reciprocal inductive interactions between these two tissues early in development

(Grobstein, 1956; Saxen and Sariola, 1987). When isolated ureteric bud is cultured separately from the metanephric mesenchyme, the ability to undergo branching morphogenesis is completely lost. Similarly, metanephric mesenchyme cultured separately from the ureteric bud does not initiate nephrogenesis and later undergoes apoptosis due to loss of survival and nephrogenic signals from the ureteric bud. Mouse genetics has greatly increased our understating of gene function in the initiation of metanephric kidney development. Gdnf, a TGF-beta family member signaling molecule expressed by the metanephric mesenchyme binds the receptor tyrosine kinase, Ret, expressed

9 within the ureteric bud, driving ureteric bud outgrowth and subsequent branching

(Pichel et al., 1996; Schuchardt et al., 1996). Loss of either of these genes results in bilateral renal agenesis demonstrating their essential role in metanephric kidney initiation.

In addition, mouse genetic studies have elucidated many transcription factors and co-factors that are required for the initial expression of Gdnf in the metanephric kidney (Fig. 2). The transcription factors Sall1, Six1, Pax2, Foxc1/2,

Hox11 paralogous genes, and Eya1 are essential for the initial expression of

Gdnf within the nascent metanephric mesenchyme. Pax2 is required for metanephric kidney initiation in addition to its role in patterning the nephric duct.

Deletion of Pax2 results in loss of Gdnf expression, a direct Pax2 transcriptional target (Brophy et al., 2001; Torres et al., 1995). Eya1 null embryos also possess a similar renal phenotype with the complete loss of Gdnf within the mesenchyme even though Pax2 expression in not affected (Xu et al., 2003). Additionally, Sall1, a zinc finger transcription factor, is required for maintaining proper levels of Gdnf expression in the metanephric mesenchyme. Loss of Sall1 during development results in severe ureteric branching defects due to inadequate Gdnf expression

(Nishinakamura et al., 2001). In contrast to the previously described genes activating transcription of Gdnf, the forkhead-domain containing transcription factors, Foxc1 and Foxc2, restrict Gdnf expression at caudal aspect of the intermediate mesoderm. Compound null Foxc1/Foxc2 embryos possess multiple, ectopic ureteric buds along the extent of the nephric duct resulting from the anterior expansion of Gdnf from the posterior metanephric mesenchyme into

10 anterior mesenchyme of the intermediate mesoderm (Kume et al., 2000). These studies demonstrate the transcription factors regulating the repression and activation of Gdnf within the metanephric mesenchyme driving the initial ureteric bud outgrowth forming metanephros.

Later at E11.5, the ureteric bud has undergone one round of branching morphogenesis and can be easily identified as a “T” shaped structure surrounded by condensing mesenchyme (Fig. 1). By E13.5, the ureteric bud has undergone many rounds of branching driven by the Gdnf/Ret pathway and has induced the metanephric mesenchyme surrounding the tips to form nephrons (Fig. 1). The conversion of metanephric mesenchyme into a nephron is complex multi-step process in which the condensing mesenchyme forms a compact epithelial sphere, known as renal vesicle, and then progressively elongates forming early tubules, described as “comma” shaped and “S” shaped bodies. During this process, the segments of the nephron are patterned and the specialized cells of the glomerulus are undergoing differentiation. Recently, many studies of targeted mice have revealed critical roles for Fgf8, Lhx1, and Wnt9b in governing the mesenchymal to epithelial tranformation of the metanephric mesenchyme (Fig.

2). The LIM-class homeobox gene, Lhx1, is expressed during the initial formation of the renal vesicle and throughout nephrogenesis and within the nephric duct and branching ureteric bud tips (Karavanov et al., 1998). The conditional removal of Lhx1 within the metanephric mesenchymes results in a remarkable halt in nephrogenesis at the renal vesicle stage resulting in the complete loss of nephrons and glomeruli (Kobayashi et al., 2005). The fibroblast growth factor

11 ligand, Fgf8, is also expressed within condensing metanephric mesenchyme and early renal vesicles (Grieshammer et al., 2005). Inactivation of Fgf8 within the metanephric mesenchyme results in the loss of Lhx1 expression and increased cell death in the nascent nephron ultimately producing a kidney completely devoid of nephrons at the final stages of development (Grieshammer et al., 2005;

Perantoni et al., 2005). In a significant paper by Carroll et al. (2005) Wnt9b, a

Wnt ligand expressed within the nephric duct and ureteric bud tips, was identified as the critical ureteric signal initiating the process of the mesenchymal to

epithelial transformation of the condensed metanephric mesenchyme. The

Wnt9b target null kidney is rudimentary and devoid of nephrons due to the loss of

Wnt4 and Fgf8 activation within the condensing metanephric mesenchyme.

Although these studies have identified some of the major molecules underlying

the induction of nephrogenesis, further research is needed to fully understand

nephron formation from the early renal vesicle stage to the mature nephron

containing specialized tubule segments and a glomerulus.

At later stages of development (E14.5 to after birth) branching

morphogenesis and nephrogenesis are proceeding at the cortex of the

developing kidney, while collecting ducts, nephrons, and glomeruli are maturing

below (Fig. 1). Wnt11 expression within the ureteric bud tips is required to

maintain Gdnf expression metanephric mesenchyme suggesting a positive

feedback loop regulating branching morphogenesis (Majumdar et al., 2003).

During the later stages of development, the uncondensed mesenchyme, or

stromal mesenchyme, surrounding the induced metanephric mesenchyme plays

12 an important role in maintaining branching morphogenesis of the ureteric tips

(Fig. 2). Loss of Foxd1, a forkhead box-containing transcription factor and stromal cell marker, results in hypoplastic kidneys possessing significantly fewer ureteric bud branch points and large, undifferentiated mesenchyme condensates

(Hatini et al., 1996). Additionally, the Foxd1 null embryonic kidney displays an aberrant expansion of Ret expression from the ureteric bud tip into the ureteric trunk, demonstrating stromal regulation of ureteric tree gene expression

(Levinson et al., 2005). The retinoic acid receptors Rara and Rarb2, also specifically expressed within the stroma, are required for maintenance of Ret expression within the ureteric bud tips suggesting a role for retinoid signaling in regulation of kidney development (Batourina et al., 2001). These experimental findings suggest that an unknown molecular signal emanating from the cortically located stroma, directly or indirectly, regulates ureteric bud branching and expression of Ret after initiation of the developing kidney.

The period of kidney development from E14.5 to after birth is also characterized by the functional maturation of the collecting ducts, nephrons, and glomeruli (Fig. 1). The mature glomerulus contains capillary loops, consisting of a specialized fenestrated endothelium surrounded by the glomerular basement membrane, encapsulated by cellular processes called foot processes extending from podocytes. The glomerular basement membrane and surrounding podocytes allow the selective passage of small molecules from the bloodstream while preventing large molecules from entering the glomerular space. The podocyte-specific bHLH transcription factor Pod1 is necessary for normal

13 glomerular development demonstrated by the knockout mouse possessing non- functional, immature glomeruli with undifferentiated podocytes (Quaggin et al.,

1999). The filtrate from the glomerulus flows through the tubules of the nephron: the proximal convoluted tubule, loop of Henle, and the distal convoluted tubule.

Each of these nephron segments possess different functions in the re-absorption of water, glucose, amino acids, electrolytes, and other small molecules from the glomerular filtrate back into the bloodstream. The Pou domain transcription factor

Pou3f3 (Brn-1) is required for proper nephron segment identity. Loss of Pou3f3 in the developing kidney arrests loop of Henle formation at an early stage resulting in the dramatic reduction of segment-specific electrolyte transporters (Nakai et al., 2003). Finally, the filtrate is further concentrated as it passes through the collecting duct derived from the ureteric bud. The maturation of the collecting duct system is characterized by the expression of the aquaporin water channel

Apq2 after birth which is the major water channel responsible for the concentration of urine (Nielsen et al., 2002). Also, a POU and homeodomain- containing transcription factor, Hnf1b (Tcf2), has been implicated in the transcription regulation of Pkhd1 expression, a membrane protein localized to the cilia of collecting duct cells (Hiesberger et al., 2005; Ward et al., 2003).

Mutations of Pkhd1 result in the formation of renal collecting duct-derived cysts characterized by autosomal-recessive polycystic kidney disorder during development (Igarashi and Somlo, 2002). As the embryo reaches the later stages of development, new molecular signals emerge in the developing kidney controlling, not only the continual branching morphogenesis of the ureteric bud

14 and mesenchymal induction of new nephrons, but also the differentiation and maturation the collecting ducts and nephrons.

Role of the Homeobox (Hox) genes in kidney development

The Homeobox (Hox) genes are transcription factors conserved

throughout evolution and are essential for proper patterning and segmentation of

the developing embryo. Hox genes are also characterized by a highly conserved

DNA-binding domain, the homeodomain, possessing an helix-turn-helix motif

(Kissinger et al., 1990). Mutations affecting the expression and function of Hox

genes in Drosophila result in striking changes of the segment identity, termed

“homeotic” transformations. The antennapedia phenotype undoubtedly

demonstrates the power these genes possess in the regulation of Drosophila segment identity by transforming the antenna imaginal disc into leg structures

(McGinnis et al., 1984). This homeotic transformation results from a chromosomal rearrangement in which antennapedia expression is expanded into the head (Frischer et al., 1986). In addition, the mammalian Hox gene, Hoxb6, can functionally substitute for the Drosophila antennapedia gene demonstrating the conserved function of these genes throughout evolution (Malicki et al., 1990).

Although the incredible phenotypes due to Hox gene mutations in

Drosophila are not found in vertebrates, these transcription factors have critical roles in governing segment identity in development. The four mammalian Hox clusters (termed A, B, C, and D) are thought to arise from the duplication of a common ancestral Homeobox cluster on a single chromosome (Krumlauf, 1992).

15 Each Hox cluster can further be divided into thirteen paralogous groups based on position and homology giving a total of 39 Hox genes (Fig. 3A). For example, the

Hox11 paralogs present on Hox clusters A, B, and C possesses the greatest homology and functional overlap in development of the mammalian body plan.

Developmental expression of the Hox genes within the embryo is also unique. The linear arrangement of the Hox paralogs parallels the expression of

Hox genes within the developing embryo, such that the 3' Hox genes (such as the Hox1 paralogs) are expressed first within the most anterior aspects of the neural tube, mesoderm, and endoderm while the 5' Hox genes (such as the Hox8 paralogs) are expressed within more posterior regions (Graham et al., 1989)

(Fig3A). This temporal and spatial colinearity of Hox gene expression is required for normal anterior/posterior patterning of the axial skeleton (Deschamps and van

Nes, 2005). In addition, each developing segment of the embryo expresses more than one Hox gene suggesting the combinatorial regulation governing segmental identity described as the “Hox code” (Kessel and Gruss, 1991). The combinatorial action of the Hox genes is also supported by experimental evidence demonstrating the ability of different Hox homeodomains binding similar

DNA sequences (Desplan et al., 1988; Hoey and Levine, 1988). Recent data argues that the divergent homeodomains of 3' and 5' Hox genes are not functionally equivalent in developing structures and organs outside of the axial skeleton, such as the kidney and reproductive tracts suggesting that each Hox paralog's DNA binding domain is functionally distinct (Zhao and Potter, 2001;

Zhao and Potter, 2002).

16 Recent expression and knockout analysis has suggested crucial roles for

Hox genes during metanephric kidney development. Deletion of the three Hox11 paralogs (Hoxa11, Hoxc11, and Hoxd11) has demonstrated a crucial role in the normal morphogenesis of the metanephric kidney (Patterson and Potter, 2003;

Wellik et al., 2002). Prior to metanephric kidney initiation, expression of these genes is found within the intermediate mesoderm before ureteric bud outgrowth, while later in development expression is limited to the condensing metanephric mesenchyme surrounding the branching ureteric bud (Patterson et al., 2001).

Single targeted mutations of either Hoxa11 or Hoxd11 results in normal renal morphogenesis in mice (Davis and Capecchi, 1994; Small and Potter, 1993).

However, Hoxa11 and Hoxd11 (Hoxa11/Hoxd11) compound null embryos exhibit both severe renal defects and severely reduced zeugopod (Davis et al., 1995).

Developmental analysis of Hoxa11/Hoxd11 compound knockout embryos

revealed disruption in the interactions between the ureteric bud and metanephric

mesenchyme resulting in phenotypes ranging from small, dysplastic kidneys to

complete bilateral kidney agenesis (Patterson et al., 2001). Additionally, the mid-

ventral domain of these compound knockout embryos is severely affected

consisting of mesenchyme lacking both ureteric branching and nephrogenesis.

Several kidney markers were also found to be significantly down regulated within

the Hoxa11/Hoxd11 null kidney. Expression of both Wt1 and Pax2, markers of

the nephrogenic mesenchyme, was reduced in these mutant kidneys, even within

the poles of the kidney which are undergoing somewhat normal branching

morphogenesis and nephrogenesis. Indeed, the expression of Gdnf, the growth

17 factor regulating ureteric outgrowth, was also significantly decreased within these mutants as expected from the loss of Wt1 and Pax2. Also, the marker of the stromal mesenchyme, Foxd1, was completely lost within the mid-ventral domain of the mutant kidney identifying a role for the Hox11 paralogs in patterning the stromal mesenchyme. Although the ureteric tree within these mutants is dysmorphic with few ureteric bud tips, normal expression of the ureteric bud tip markers, Wnt11 and Ret, is identified within the mutant kidneys. Thus, the loss of the previous markers suggests the metanephric mesenchyme in these compound mutants is unable to properly interact with the ureteric bud resulting in defects in ureteric branching and nephrogenesis.

The metanephric mesenchyme of the Hox11 null embryo fails to express

Gdnf resulting in the loss of ureteric bud outgrowth and subsequent bilateral renal agenesis phenotype (Wellik et al., 2002). This severe kidney phenotype generated by the loss of all Hox11 paralogs identifies the Hoxa11/Hoxd11 phenotype as a hypomorphic phenotype representing the partial loss of function of the Hox11 paralogs, and furthermore, suggesting that these genes possess a conserved, functionally redundant role during development. Surprisingly, the transcription factors Wt1, Pax2, and Eya1 continue to be expressed within the uninduced E11 metanephric mesenchyme of the triple Hox11 knockout. Since these critical transcription factors remain expressed within the metanephric mesenchyme prior to apoptosis due to loss of the ureteric bud signal, the Hox11 paralogs may participate downstream of these genes in the regulation of Gdnf expression. In addition, Six2, an ortholog of the Drosophila transcription factor

18 sine oculis (so), is completely lost from the metanephric mesenchyme at the time of kidney initiation. In Drosophila, the Pax ortholog, eyeless (ey) operates upstream of eyes absent (eya) and so in the development of the eye (Halder et

al., 1998; Pignoni et al., 1997). This data suggests that the Hox11 paralogs

interacts with the vertebrate pax-eya-six pathway activating Gdnf expression in

mammalian kidney development, rather than influencing the patterning of the

urogenital tract resulting in homeotic transformations.

Recently, the developmental expression analysis of all 39 Hox genes in

metanephric kidney has been completed expanding our knowledge of the

potential roles of important genes in development (Patterson and Potter, 2004)

(Fig. 3A). This study identified the spatial colinearity of the Hox genes in renal

development which is characterized by the expression of 3’ Hox genes are expressed in more anterior regions of developing structures compared to the expression of 5’ Hox genes. The expression of many 3' Hox genes is restricted to the anterior intermediate mesoderm derived structures, the ureteric tree and ureteric bud tips. In contrast, the 5' Hox genes (Hox9 thru Hox12 paralogs) displayed expression restricted to the metanephric mesenchyme and mesenchymal-derived structures.

Also, different segments of the forming nephron displayed overlapping domains of Hox gene expression, while Hox gene expression within ureteric- derived duct system displayed no segmental expression (Fig. 3B). Amazingly, paralogous Hox genes were found to possess diverse expression patterns. For

19 example, while all Hox9 paralogs possess expression within the condensing metanephric mesenchyme, Hoxd9 also displays expression from the ureteric tree into the connecting segment of nephron. Additionally, Hox genes on the same

Hox cluster possessed similar expression patterns, as in the case of Hoxa9,

Hoxa10, and Hoxa11 which all demonstrated overlapping expression patterns within the condensing mesenchyme and the junction of the nascent nephron and ureteric bud tip suggesting that these flanking Hox genes are under the control of the same enhancer elements (Fig. 3B). Indeed, this phenomenon has been demonstrated within the Hox B cluster in which a retinoic acid-responsive enhancer drives expression of Hoxb5, Hoxb6, and Hoxb8 in development

(Oosterveen et al., 2003). Overall, the expression data identifies a possible additional Hox function within the kidney in the regulation of nephron segment patterning and identity.

In conclusion, Hox genes play critical roles in kidney development as demonstrated by the loss of all three Hox11 paralogs and subsequent failure of

Gdnf expression in the metanephric mesenchyme. Additionally, expression analysis has provided evidence suggesting the regulation of nephron segment identity by Hox genes (Fig. 3B). However, little is known about the Hox downstream target genes that may be regulating the process of nephrogenesis.

More research is needed to fully understand the functions of Hox genes in kidney development which is complicated by both functional redundancy of Hox paralogs and combinatorial function of many Hox genes in a given tissue.

Wnt signaling in renal development

20 Wnt ligands can elicit two divergent downstream signaling pathways within a cell, the canonical Wnt pathway involving β-catenin and the noncanonical Wnt pathway which operates independent of β-catenin. Canonical Wnt signaling is dependent on the stabilization of cytosolic β-catenin resulting in the nuclear localization of a transcriptional complex consisting of β-catenin, Legless,

Pygopus, and Tcf/Lef transcription factors. In the absence of Wnt ligand, β- catenin is continually phosphorylated by the GSK3 complex, ubiquinated, and degraded within the cytosol. The noncanonical Wnt pathway involves eliciting cellular responses through other mechanisms including calcium signaling, JNK signaling, and heterotrimeric G proteins (reviewed in Veeman et al., 2003).

Although significant advances have been made in describing the developmental function of Wnt signaling, the responses elicited by Wnt signaling in the developing kidney remain poorly defined. Initially, the inductive capabilities of certain Wnt ligands were demonstrated by co-culture of uninduced metanephric mesenchyme with cells expressing Wnt proteins (Herzlinger et al.,

1994; Kispert et al., 1998). These in vitro studies identified the Wnt ligands

Wnt1, Wnt3a, Wnt4, Wnt7a, and Wnt7b as metanephric mesenchyme inducers

of nephrogenesis, while Wnt5a and Wnt11 possess no inductive activity.

Additionally, the treatment of uninduced metanephric mesenchyme with LiCl, an

inhibitor of GSK3, results in the robust induction of tubules mimicking the

inductive capabilities of Wnt1 (Davies and Garrod, 1995). Although LiCl

treatment of cells results in β-catenin stabilization and subsequent activation of

21 Wnt targets, lithium can also activate other signaling pathways, such as the JNK signaling pathway (Hedgepeth et al., 1997).

Genetic studies of Wnt4, Wnt9b, and Wnt11 demonstrate the unique functions each gene possesses in renal morphogenesis. During renal morphogenesis, ureteric tip expression of Wnt11 is required to maintain normal levels of Gdnf and Ret expression in the kidney driving normal branching morphogenesis. The loss of Wnt11 causes the loss of branching potential resulting in a significant decrease in the number of nephrons present in the kidney. Previous studies in other developing systems suggest Wnt11 signals

specifically through the noncanonical Wnt pathway (Du et al., 1995). The

specific downstream pathway in which Wnt11 regulates Gdnf/Ret expression

during kidney development remains to be elucidated. Both Wnt4 and Wnt9b

induce epithelial transformation of the metanephric mesenchyme driving

nephrogenesis. Although both Wnt ligands possess inductive capabilities, Wnt9b

is thought to operate through the canonical Wnt pathway, while Wnt4 functions in

activating the noncanonical Wnt pathway. The loss of nephric induction in Wnt9b

knockout mice can be rescued by the introduction of a transgene expressing

Wnt1 specifically within the ureteric bud (Carroll et al., 2005). Additionally,

Wnt9b expression correlates with the activation of a Tcf/Lef reporter transgene in

craniofacial morphogenesis (Lan et al., 2006). Recently, Wnt4 has been

implicated in the beta-catenin-independent activation of the JNK pathway in the

developing vertebrate eye (Cai et al., 2002; Maurus et al., 2005). Contrary to this

finding, the inhibition of Wnt4 activity within induced metanephric mesenchyme

22 cultures results in decreased Tcf/Lef complex formation and subsequent inhibition of nephrogenesis (Yoshino et al., 2001). The expression of both canonical and noncanonical Wnts in the developing kidney suggests that these two pathways may be acting in parallel during morphogenesis. Clearly, more information is needed to dissect the function of canonical and noncanonical Wnt signaling in kidney development.

Recent genetic studies in Drosophila have identified a specific requirement of pygopus, a PHD-containing transcriptional coactivator interacting with β-catenin, in canonical Wnt signaling (Belenkaya et al., 2002; Kramps et al.,

2002; Parker et al., 2002). Pygopus is a component of the β-catenin transcriptional complex and is thought to be involved in both nuclear localization and transcriptional activation (Hoffmans et al., 2005; Townsley et al., 2004).

Additionally, experimental evidence suggests that transcriptional activation of

Wnt target relies on the formation of a “chain of adaptors” consisting of Lgs/Bcl9,

β-catenin, and Tcf/Lef transcription factors directing the Pygopus to activate gene expression (Stadeli and Basler, 2005). In mammals, two pygopus orthologs,

Pygo1 and Pygo2, have been identified containing conserved N-terminal and C- terminal domains (Li et al., 2004). Despite Pygo function in Wnt signaling during

Drosophila development, it is not yet known if the mammalian orthologs of Pygo mediates Wnt signaling in the developing kidney.

FSGS (Focal Segmental glomerulosclerosis)

23 Of the renal pathologies, FSGS remains one of the least understood glomerular diseases in pediatric nephrology today (Collins et al., 2005). FSGS results in the progression of proteinuria, renal insufficiency, and renal failure requiring dialysis or transplant to compensate for loss of renal function. This disease can be classified as primary (idiopathic) with an unknown etiology, or as secondary resulting from other conditions or diseases, such as the human immunodeficiency virus (HIV). Different subtypes of lesions can be identified within different FSGS patients, but each subtype is identified as segmental, only a portion of a single glomerulus is scarred, and focal, only a portion of the total glomeruli population is affected (Meyrier, 2003). The prognosis is poor due to rapid progression to renal failure, variable response to corticosteroids, and the recurrence of the disease in about 40% of FSGS patients receiving renal transplant suggesting that the cause is not intrinsic to the kidney in all cases

(Daskalakis and Winn, 2006). Therefore, primary FSGS is viewed as a heterogeneous condition with no identifiable single disease entity found in all affected individuals (Devarajan and Spitzer, 2002). Due to this heterogeneity, others have stated that FSGS is, in fact, not a disease, but the pathological classification of the same histological manifestation caused by different insults to the glomerulus (Meyrier, 2003).

Despite the heterogeneous nature of FSGS, all forms of this condition affect the podocytes, the specialized glomerular cells surrounding capillaries responsible for the selective filtration of small molecules from the bloodstream

(Kwoh et al., 2006). Podocyte injury is thought to result in lesions causing a

24 cascade of events eventually leading to the scarring of the glomerulus and the loss of nephron function. Within the FSGS glomerulus, the diffuse loss of podocyte foot processes, as well as the glomerular basement membrane surrounding the specialized endothelial cells, results in plasma proteins crossing freely from the bloodstream into the urinary space. The influx of plamsa proteins, such as albumin, not only results in podocyte damage and foot process effacement, but also has severe consequences for the downstream tubules causing both inflammation and fibrosis (Davies et al., 1985; Remuzzi et al.,

2004).

Recent molecular studies of FSGS in humans and mouse models have identified genes crucial in maintaining normal podocyte morphology and function during injury. Mutations in the NPHS1 gene encoding Nephrin cause the Finnish- type congenital nephrotic syndrome (Kestila et al., 1998). Nephrin specifically localizes to foot processes of podocytes forming slit diaphragms acting as molecular sieves (Ruotsalainen et al., 1999). When NPHS1 is deleted in mice, neonates develop severe nephrotic syndrome causing death (Putaala et al.,

2001). Mutations in podocin (NPHS2) are also associated with FSGS at a frequency of around 10% in children (Caridi et al., 2003). Further analysis suggest that the most common FSGS-inducing mutation of NPHS2 result in a misfolded protein causing mislocalization of Nephrin, suggesting a role of

Podocin in maintaining the slit diaphragm (Huber et al., 2003b; Nishibori et al.,

2004). Additionally, mice with one null allele of NPHS2 develop proteinuria by six months of age, while mice possessing both null alleles of NPHS2 possess

25 decreased Nephrin and have absent slit diaphragms resulting is death by three weeks of age (Roselli et al., 2004). Loss of the Cd2ap gene in mice presents a similar renal phenotype to the NPHS2 knockout glomerular pathology and is dependent on gene dosage (Shih et al., 1999). Cd2ap may play multiple roles in the podocyte including anchoring Nephrin to the actin cytoskeleton, trafficking of podocyte-specific proteins, and regulation of cell survival through the phosphoinositide 3 (PI3) kinase pathway (Huber et al., 2003a; Schiffer et al.,

2004; Shih et al., 1999). Additionally, variants of the Cd2ap gene result in variability of FSGS progression (Kim et al., 2003). A mutant form of alpha-actinin-

4 (Actn4) has also been identified in an autosomal-dominant form of FSGS due to in abnormalities in filament assembly (Kaplan et al., 2000). Although genes identified in these studies do not account for a significant portion of FSGS cases, these genes do demonstrate that the primary lesion of FSGS occurs by podocyte loss or susceptibility to injury.

Significance of microarray analysis in renal development and disease

Microarray technology has allowed the unprecedented analysis of thousands of genes expressed in a single sample offering the molecular biologist a new, powerful tool to investigate the transcriptional state of cells, tissues, and organs. Initially, microarray technology required large amounts of total RNA in the range of micrograms from a given sample, prohibiting the analysis of small cell populations or tissues. Given the extremely small size of developing structures in a mammalian embryo, microarray analysis was extremely difficult

26 requiring the isolation of hundreds of developing organs in order to isolate enough material for a single hybridization. In order to circumvent the large amount of material needed for a single microarray, some researchers created immortalized cell lines representing the developmental potential of specific

tissues in vivo (Levashova et al., 2003; Valerius et al., 2002b). Although such a

methodology may be thought artificial, the immortalized mL3 cell line derived

from the metanephric mesenchyme expresses Gdnf and drives ureteric

branching morphogenesis when co-cultured with isolated ureteric buds

demonstrating the preservation of some developmental characteristics

(Levashova et al., 2003; Valerius et al., 2002b). These developmental cell lines

were further exploited to identify possible downstream targets of Hoxa11 using

microarray analysis of both HoxA11-transfected and immortalized Hoxa11

knockout metanephric mesenchyme cell lines (Valerius et al., 2002a). Although

these experiments have been useful in analysis of the initial transcriptional state

of early developing kidney, mK3 cells do not fully recapitulate the complex

molecular signals and interactions occurring during renal development in vivo.

One example of the loss of development potential is the failure of mK3 cells to

undergo epithelial transformation. In a similar study, an immortalized

metanephric mesenchymal cell line was generated possessing the ability to form

epithelial bodies representing renal vesicles when exposed to soluble growth

factors previously demonstrated to be necessary for nephrogenesis in vitro

(Levashova et al. 2003). Therefore, cell line based experiments do not

completely represent the molecular changes occurring throughout renal

27 development within the ureteric bud and metanephric mesenchyme compartments. In order to fully understand gene expression changes throughout kidney development, the transcriptional states of individual developing structures of embryonic kidneys must be analyzed.

Additionally, microarray expression analysis of renal disease was similarly hindered due to the small amounts of tissue obtained from procedures, such as

biopsies. The application of expression microarray technology is especially

important for clinical practice, not only in the analysis of possible candidate

genes involved in pathogenesis, but also as a possible diagnostic tool in

predicting the prognosis and treatment of the disease. Initially, microarray

expression studies were performed on human diseases, such as cancer, in which

adequate amount of material could be obtained from a highly proliferative tissue.

Indeed, earliest microarray expression studies of renal disease were performed

on two common renal malignancies: Wilms’ tumor disease, a pediatric

malignancy in which the mesenchyme fails to undergoes differentiation due to

mutation of the Wt1 gene and renal cell carcinoma, a epithelial malignancy of the

tubules within the nephron (Li et al., 2002; Moch et al., 1999). As in renal

development, the quantity of material obtained from a single kidney biopsy

previously prevented the successful expression analysis of human renal disease

using microarray technology.

Technological advances in RNA amplification protocols have allowed

small amounts of total RNA in the picogram to nanogram range to be

28 successfully amplified into quantities needed for microarray hybridization (Baugh et al., 2001). The ability to amplify small quantities of RNA allows the microarray analysis of individual developing kidney segments, such a single E11.5 condensing metanephric mesenchyme, obtained through laser capture micro- dissection (LCM). Additionally, the expression study of renal diseases using microarrays can now be easily performed using small amounts of tissue obtained from biopsy specimens previously used for histological analysis.

The following microarray expression studies in renal development greatly enhance our knowledge, providing not only the identification of new candidate genes involved in morphogenesis, but also possible downstream target genes of the Hox11 paralogs and Pygo orthologs. These developmental studies incorporate both the latest microarray technology and novel techniques expanding the initial microarray study of Stuart et al. (2001). Additionally, FSGS patient biopsies were analyzed using microarray technology providing a more complete portrait of gene expression changes during renal disease.

In conclusion, the following thesis research seeks to understand the gene expression changes involved in both normal and abnormal kidney morphogenesis, as well as kidney disease. Additionally, the developmental functions of the Hox genes and Wnt signaling were further investigated in kidney development. Overall, these studies provide a significant expansion of our knowledge of renal developmental and disease.

29 References

Batourina, E., Gim, S., Bello, N., Shy, M., Clagett-Dame, M., Srinivas, S.,

Costantini, F., and Mendelsohn, C. (2001). Vitamin A controls

epithelial/mesenchymal interactions through Ret expression. Nat Genet

27, 74-8.

Baugh, L. R., Hill, A. A., Brown, E. L., and Hunter, C. P. (2001). Quantitative

analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res

29, E29.

Belenkaya, T. Y., Han, C., Standley, H. J., Lin, X., Houston, D. W., and

Heasman, J. (2002). pygopus Encodes a nuclear protein essential for

wingless/Wnt signaling. Development 129, 4089-101.

Bouchard, M. (2004). Transcriptional control of kidney development.

Differentiation 72, 295-306.

Bouchard, M., Souabni, A., Mandler, M., Neubuser, A., and Busslinger, M.

(2002). Nephric lineage specification by Pax2 and Pax8. Genes Dev 16,

2958-70.

Brophy, P. D., Ostrom, L., Lang, K. M., and Dressler, G. R. (2001). Regulation of

ureteric bud outgrowth by Pax2-dependent activation of the glial derived

neurotrophic factor gene. Development 128, 4747-56.

Cai, Y., Lechner, M. S., Nihalani, D., Prindle, M. J., Holzman, L. B., and Dressler,

G. R. (2002). Phosphorylation of Pax2 by the c-Jun N-terminal kinase and

enhanced Pax2-dependent transcription activation. J Biol Chem 277,

1217-22.

30 Caridi, G., Bertelli, R., Di Duca, M., Dagnino, M., Emma, F., Onetti Muda, A.,

Scolari, F., Miglietti, N., Mazzucco, G., Murer, L., Carrea, A., Massella, L.,

Rizzoni, G., Perfumo, F., and Ghiggeri, G. M. (2003). Broadening the

spectrum of diseases related to podocin mutations. J Am Soc Nephrol 14,

1278-86.

Carroll, T. J., Park, J. S., Hayashi, S., Majumdar, A., and McMahon, A. P. (2005).

Wnt9b plays a central role in the regulation of mesenchymal to epithelial

transitions underlying organogenesis of the mammalian urogenital system.

Dev Cell 9, 283-92.

Collins, A. J., Kasiske, B., Herzog, C., Chavers, B., Foley, R., Gilbertson, D.,

Grimm, R., Liu, J., Louis, T., Manning, W., Matas, A., McBean, M., Murray,

A., St Peter, W., Xue, J., Fan, Q., Guo, H., Li, S., Roberts, T., Snyder, J.,

Solid, C., Wang, C., Weinhandl, E., Arko, C., Chen, S. C., Dalleska, F.,

Daniels, F., Dunning, S., Ebben, J., Frazier, E., Johnson, R., Sheets, D.,

Forrest, B., Berrini, D., Constantini, E., Everson, S., Frederick, P., Eggers,

P., and Agodoa, L. (2005). Excerpts from the United States Renal Data

System 2004 annual data report: atlas of end-stage renal disease in the

United States. Am J Kidney Dis 45, A5-7.

Daskalakis, N., and Winn, M. P. (2006). Focal and segmental glomerulosclerosis:

varying biologic mechanisms underlie a final histopathologic end point.

Semin Nephrol 26, 89-94.

31 Davies, D. J., Messina, A., Thumwood, C. M., and Ryan, G. B. (1985).

Glomerular podocytic injury in protein overload proteinuria. Pathology 17,

412-9.

Davies, J. A., and Garrod, D. R. (1995). Induction of early stages of kidney tubule

differentiation by lithium ions. Dev Biol 167, 50-60.

Davis, A. P., and Capecchi, M. R. (1994). Axial homeosis and appendicular

skeleton defects in mice with a targeted disruption of hoxd-11.

Development 120, 2187-98.

Davis, A. P., Witte, D. P., Hsieh-Li, H. M., Potter, S. S., and Capecchi, M. R.

(1995). Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11.

Nature 375, 791-5.

Deschamps, J., and van Nes, J. (2005). Developmental regulation of the Hox

genes during axial morphogenesis in the mouse. Development 132, 2931-

42.

Desplan, C., Theis, J., and O'Farrell, P. H. (1988). The sequence specificity of

homeodomain-DNA interaction. Cell 54, 1081-90.

Devarajan, P., and Spitzer, A. (2002). Towards a biological characterization of

focal segmental glomerulosclerosis. Am J Kidney Dis 39, 625-36.

Dressler, G. R. (2006). The Cellular Basis of Kidney Development. Annu Rev

Cell Dev Biol.

Du, S. J., Purcell, S. M., Christian, J. L., McGrew, L. L., and Moon, R. T. (1995).

Identification of distinct classes and functional domains of Wnts through

32 expression of wild-type and chimeric proteins in Xenopus embryos. Mol

Cell Biol 15, 2625-34.

Frischer, L. E., Hagen, F. S., and Garber, R. L. (1986). An inversion that disrupts

the Antennapedia gene causes abnormal structure and localization of

RNAs. Cell 47, 1017-23.

Graham, A., Papalopulu, N., and Krumlauf, R. (1989). The murine and

Drosophila homeobox gene complexes have common features of

organization and expression. Cell 57, 367-78.

Grieshammer, U., Cebrian, C., Ilagan, R., Meyers, E., Herzlinger, D., and Martin,

G. R. (2005). FGF8 is required for cell survival at distinct stages of

nephrogenesis and for regulation of gene expression in nascent nephrons.

Development 132, 3847-57.

Grobstein, C. (1955). Inductive interaction in the development of the mouse

metanephros. J. Exp. Zool. 130, 319-340.

Grobstein, C. (1956). Inductive tissue interaction in development. Adv Cancer

Res 4, 187-236.

Halder, G., Callaerts, P., Flister, S., Walldorf, U., Kloter, U., and Gehring, W. J.

(1998). Eyeless initiates the expression of both sine oculis and eyes

absent during Drosophila compound eye development. Development 125,

2181-91.

Hatini, V., Huh, S. O., Herzlinger, D., Soares, V. C., and Lai, E. (1996). Essential

role of stromal mesenchyme in kidney morphogenesis revealed by

33 targeted disruption of Winged Helix transcription factor BF-2. Genes Dev

10, 1467-78.

Hedgepeth, C. M., Conrad, L. J., Zhang, J., Huang, H. C., Lee, V. M., and Klein,

P. S. (1997). Activation of the Wnt signaling pathway: a molecular

mechanism for lithium action. Dev Biol 185, 82-91.

Herzlinger, D., Qiao, J., Cohen, D., Ramakrishna, N., and Brown, A. M. (1994).

Induction of kidney epithelial morphogenesis by cells expressing Wnt-1.

Dev Biol 166, 815-8.

Hiesberger, T., Shao, X., Gourley, E., Reimann, A., Pontoglio, M., and Igarashi,

P. (2005). Role of the hepatocyte nuclear factor-1beta (HNF-1beta) C-

terminal domain in Pkhd1 (ARPKD) gene transcription and renal

cystogenesis. J Biol Chem 280, 10578-86.

Hoey, T., and Levine, M. (1988). Divergent homeo box proteins recognize similar

DNA sequences in Drosophila. Nature 332, 858-61.

Hoffmans, R., Stadeli, R., and Basler, K. (2005). Pygopus and legless provide

essential transcriptional coactivator functions to armadillo/beta-catenin.

Curr Biol 15, 1207-11.

Huber, T. B., Hartleben, B., Kim, J., Schmidts, M., Schermer, B., Keil, A., Egger,

L., Lecha, R. L., Borner, C., Pavenstadt, H., Shaw, A. S., Walz, G., and

Benzing, T. (2003a). Nephrin and CD2AP associate with phosphoinositide

3-OH kinase and stimulate AKT-dependent signaling. Mol Cell Biol 23,

4917-28.

34 Huber, T. B., Simons, M., Hartleben, B., Sernetz, L., Schmidts, M., Gundlach, E.,

Saleem, M. A., Walz, G., and Benzing, T. (2003b). Molecular basis of the

functional podocin-nephrin complex: mutations in the NPHS2 gene disrupt

nephrin targeting to lipid raft microdomains. Hum Mol Genet 12, 3397-405.

Igarashi, P., and Somlo, S. (2002). Genetics and pathogenesis of polycystic

kidney disease. J Am Soc Nephrol 13, 2384-98.

Kaplan, J. M., Kim, S. H., North, K. N., Rennke, H., Correia, L. A., Tong, H. Q.,

Mathis, B. J., Rodriguez-Perez, J. C., Allen, P. G., Beggs, A. H., and

Pollak, M. R. (2000). Mutations in ACTN4, encoding alpha-actinin-4,

cause familial focal segmental glomerulosclerosis. Nat Genet 24, 251-6.

Karavanov, A. A., Karavanova, I., Perantoni, A., and Dawid, I. B. (1998).

Expression pattern of the rat Lim-1 homeobox gene suggests a dual role

during kidney development. Int J Dev Biol 42, 61-6.

Kessel, M., and Gruss, P. (1991). Homeotic transformations of murine vertebrae

and concomitant alteration of Hox codes induced by retinoic acid. Cell 67,

89-104.

Kestila, M., Lenkkeri, U., Mannikko, M., Lamerdin, J., McCready, P., Putaala, H.,

Ruotsalainen, V., Morita, T., Nissinen, M., Herva, R., Kashtan, C. E.,

Peltonen, L., Holmberg, C., Olsen, A., and Tryggvason, K. (1998).

Positionally cloned gene for a novel glomerular protein--nephrin--is

mutated in congenital nephrotic syndrome. Mol Cell 1, 575-82.

35 Kim, J. M., Wu, H., Green, G., Winkler, C. A., Kopp, J. B., Miner, J. H., Unanue,

E. R., and Shaw, A. S. (2003). CD2-associated protein haploinsufficiency

is linked to glomerular disease susceptibility. Science 300, 1298-300.

Kispert, A., Vainio, S., and McMahon, A. P. (1998). Wnt-4 is a mesenchymal

signal for epithelial transformation of metanephric mesenchyme in the

developing kidney. Development 125, 4225-34.

Kissinger, C. R., Liu, B. S., Martin-Blanco, E., Kornberg, T. B., and Pabo, C. O.

(1990). Crystal structure of an engrailed homeodomain-DNA complex at

2.8 A resolution: a framework for understanding homeodomain-DNA

interactions. Cell 63, 579-90.

Kobayashi, A., Kwan, K. M., Carroll, T. J., McMahon, A. P., Mendelsohn, C. L.,

and Behringer, R. R. (2005). Distinct and sequential tissue-specific

activities of the LIM-class homeobox gene Lim1 for tubular morphogenesis

during kidney development. Development 132, 2809-23.

Kramps, T., Peter, O., Brunner, E., Nellen, D., Froesch, B., Chatterjee, S.,

Murone, M., Zullig, S., and Basler, K. (2002). Wnt/wingless signaling

requires BCL9/legless-mediated recruitment of pygopus to the nuclear

beta-catenin-TCF complex. Cell 109, 47-60.

Krumlauf, R. (1992). Evolution of the vertebrate Hox homeobox genes.

Bioessays 14, 245-52.

Kume, T., Deng, K., and Hogan, B. L. (2000). Murine forkhead/winged helix

genes Foxc1 (Mf1) and Foxc2 (Mfh1) are required for the early

organogenesis of the kidney and urinary tract. Development 127, 1387-95.

36 Kwoh, C., Shannon, M. B., Miner, J. H., and Shaw, A. (2006). PATHOGENESIS

OF NONIMMUNE GLOMERULOPATHIES. Annual Review of Pathology:

Mechanisms of Disease 1, 349-374.

Lan, Y., Ryan, R. C., Zhang, Z., Bullard, S. A., Bush, J. O., Maltby, K. M., Lidral,

A. C., and Jiang, R. (2006). Expression of Wnt9b and activation of

canonical Wnt signaling during midfacial morphogenesis in mice. Dev Dyn

235, 1448-54.

Levashova, Z. B., Plisov, S. Y., and Perantoni, A. O. (2003). Conditionally

immortalized cell line of inducible metanephric mesenchyme. Kidney Int

63, 2075-87.

Levinson, R. S., Batourina, E., Choi, C., Vorontchikhina, M., Kitajewski, J., and

Mendelsohn, C. L. (2005). Foxd1-dependent signals control cellularity in

the renal capsule, a structure required for normal renal development.

Development 132, 529-39.

Li, B., Mackay, D. R., Ma, J., and Dai, X. (2004). Cloning and developmental

expression of mouse pygopus 2, a putative Wnt signaling component.

Genomics 84, 398-405.

Li, C. M., Guo, M., Borczuk, A., Powell, C. A., Wei, M., Thaker, H. M., Friedman,

R., Klein, U., and Tycko, B. (2002). Gene expression in Wilms' tumor

mimics the earliest committed stage in the metanephric mesenchymal-

epithelial transition. Am J Pathol 160, 2181-90.

37 Majumdar, A., Vainio, S., Kispert, A., McMahon, J., and McMahon, A. P. (2003).

Wnt11 and Ret/Gdnf pathways cooperate in regulating ureteric branching

during metanephric kidney development. Development 130, 3175-85.

Malicki, J., Schughart, K., and McGinnis, W. (1990). Mouse Hox-2.2 specifies

thoracic segmental identity in Drosophila embryos and larvae. Cell 63,

961-7.

Maurus, D., Heligon, C., Burger-Schwarzler, A., Brandli, A. W., and Kuhl, M.

(2005). Noncanonical Wnt-4 signaling and EAF2 are required for eye

development in Xenopus laevis. Embo J 24, 1181-91.

McGinnis, W., Levine, M. S., Hafen, E., Kuroiwa, A., and Gehring, W. J. (1984).

A conserved DNA sequence in homoeotic genes of the Drosophila

Antennapedia and bithorax complexes. Nature 308, 428-33.

Meyrier, A. (2003). E pluribus unum: The riddle of focal segmental

glomerulosclerosis. Semin Nephrol 23, 135-40.

Moch, H., Schraml, P., Bubendorf, L., Mirlacher, M., Kononen, J., Gasser, T.,

Mihatsch, M. J., Kallioniemi, O. P., and Sauter, G. (1999). High-throughput

tissue microarray analysis to evaluate genes uncovered by cDNA

microarray screening in renal cell carcinoma. Am J Pathol 154, 981-6.

Nakai, S., Sugitani, Y., Sato, H., Ito, S., Miura, Y., Ogawa, M., Nishi, M., Jishage,

K., Minowa, O., and Noda, T. (2003). Crucial roles of Brn1 in distal tubule

formation and function in mouse kidney. Development 130, 4751-9.

38 Nielsen, S., Frokiaer, J., Marples, D., Kwon, T. H., Agre, P., and Knepper, M. A.

(2002). Aquaporins in the kidney: from molecules to medicine. Physiol Rev

82, 205-44.

Nishibori, Y., Liu, L., Hosoyamada, M., Endou, H., Kudo, A., Takenaka, H.,

Higashihara, E., Bessho, F., Takahashi, S., Kershaw, D., Ruotsalainen,

V., Tryggvason, K., Khoshnoodi, J., and Yan, K. (2004). Disease-causing

missense mutations in NPHS2 gene alter normal nephrin trafficking to the

plasma membrane. Kidney Int 66, 1755-65.

Nishinakamura, R., Matsumoto, Y., Nakao, K., Nakamura, K., Sato, A.,

Copeland, N. G., Gilbert, D. J., Jenkins, N. A., Scully, S., Lacey, D. L.,

Katsuki, M., Asashima, M., and Yokota, T. (2001). Murine homolog of

SALL1 is essential for ureteric bud invasion in kidney development.

Development 128, 3105-15.

Oosterveen, T., Niederreither, K., Dolle, P., Chambon, P., Meijlink, F., and

Deschamps, J. (2003). Retinoids regulate the anterior expression

boundaries of 5' Hoxb genes in posterior hindbrain. Embo J 22, 262-9.

Parker, D. S., Jemison, J., and Cadigan, K. M. (2002). Pygopus, a nuclear PHD-

finger protein required for Wingless signaling in Drosophila. Development

129, 2565-76.

Patterson, L. T., Pembaur, M., and Potter, S. S. (2001). Hoxa11 and Hoxd11

regulate branching morphogenesis of the ureteric bud in the developing

kidney. Development 128, 2153-61.

39 Patterson, L. T., and Potter, S. S. (2003). Hox genes and kidney patterning. Curr

Opin Nephrol Hypertens 12, 19-23.

Patterson, L. T., and Potter, S. S. (2004). Atlas of Hox gene expression in the

developing kidney. Dev Dyn 229, 771-9.

Perantoni, A. O., Timofeeva, O., Naillat, F., Richman, C., Pajni-Underwood, S.,

Wilson, C., Vainio, S., Dove, L. F., and Lewandoski, M. (2005).

Inactivation of FGF8 in early mesoderm reveals an essential role in kidney

development. Development 132, 3859-71.

Pichel, J. G., Shen, L., Sheng, H. Z., Granholm, A. C., Drago, J., Grinberg, A.,

Lee, E. J., Huang, S. P., Saarma, M., Hoffer, B. J., Sariola, H., and

Westphal, H. (1996). Defects in enteric innervation and kidney

development in mice lacking GDNF. Nature 382, 73-6.

Pignoni, F., Hu, B., Zavitz, K. H., Xiao, J., Garrity, P. A., and Zipursky, S. L.

(1997). The eye-specification proteins So and Eya form a complex and

regulate multiple steps in Drosophila eye development. Cell 91, 881-91.

Putaala, H., Soininen, R., Kilpelainen, P., Wartiovaara, J., and Tryggvason, K.

(2001). The murine nephrin gene is specifically expressed in kidney, brain

and pancreas: inactivation of the gene leads to massive proteinuria and

neonatal death. Hum Mol Genet 10, 1-8.

Quaggin, S. E., Schwartz, L., Cui, S., Igarashi, P., Deimling, J., Post, M., and

Rossant, J. (1999). The basic-helix-loop-helix protein pod1 is critically

important for kidney and lung organogenesis. Development 126, 5771-83.

40 Remuzzi, G., Chiurchiu, C., and Ruggenenti, P. (2004). Proteinuria predicting

outcome in renal disease: nondiabetic nephropathies (REIN). Kidney Int

Suppl, S90-6.

Roselli, S., Heidet, L., Sich, M., Henger, A., Kretzler, M., Gubler, M. C., and

Antignac, C. (2004). Early glomerular filtration defect and severe renal

disease in podocin-deficient mice. Mol Cell Biol 24, 550-60.

Ruotsalainen, V., Ljungberg, P., Wartiovaara, J., Lenkkeri, U., Kestila, M.,

Jalanko, H., Holmberg, C., and Tryggvason, K. (1999). Nephrin is

specifically located at the slit diaphragm of glomerular podocytes. Proc

Natl Acad Sci U S A 96, 7962-7.

Saxen, L., and Sariola, H. (1987). Early organogenesis of the kidney. Pediatr

Nephrol 1, 385-92.

Schiffer, M., Mundel, P., Shaw, A. S., and Bottinger, E. P. (2004). A novel role for

the adaptor molecule CD2-associated protein in transforming growth

factor-beta-induced apoptosis. J Biol Chem 279, 37004-12.

Schuchardt, A., D'Agati, V., Pachnis, V., and Costantini, F. (1996). Renal

agenesis and hypodysplasia in ret-k- mutant mice result from defects in

ureteric bud development. Development 122, 1919-29.

Shih, N. Y., Li, J., Karpitskii, V., Nguyen, A., Dustin, M. L., Kanagawa, O., Miner,

J. H., and Shaw, A. S. (1999). Congenital nephrotic syndrome in mice

lacking CD2-associated protein. Science 286, 312-5.

Small, K. M., and Potter, S. S. (1993). Homeotic transformations and limb defects

in Hox A11 mutant mice. Genes Dev 7, 2318-28.

41 Stadeli, R., and Basler, K. (2005). Dissecting nuclear Wingless signalling:

recruitment of the transcriptional co-activator Pygopus by a chain of

adaptor proteins. Mech Dev 122, 1171-82.

Stuart, R. O., Bush, K. T., and Nigam, S. K. (2001). Changes in global gene

expression patterns during development and maturation of the rat kidney.

Proc Natl Acad Sci U S A 98, 5649-54.

Torres, M., Gomez-Pardo, E., Dressler, G. R., and Gruss, P. (1995). Pax-2

controls multiple steps of urogenital development. Development 121,

4057-65.

Townsley, F. M., Cliffe, A., and Bienz, M. (2004). Pygopus and Legless target

Armadillo/beta-catenin to the nucleus to enable its transcriptional co-

activator function. Nat Cell Biol 6, 626-33.

Valerius, M. T., Patterson, L. T., Feng, Y., and Potter, S. S. (2002a). Hoxa 11 is

upstream of Integrin alpha8 expression in the developing kidney. Proc Natl

Acad Sci U S A 99, 8090-5.

Valerius, M. T., Patterson, L. T., Witte, D. P., and Potter, S. S. (2002b).

Microarray analysis of novel cell lines representing two stages of

metanephric mesenchyme differentiation. Mech Dev 110, 151-64.

Veeman, M. T., Axelrod, J. D., and Moon, R. T. (2003). A second canon.

Functions and mechanisms of beta-catenin-independent Wnt signaling.

Dev Cell 5, 367-77.

42 Vize, P. D., Woolf, A. S., and Bard, J. B. L. (2002). "The kidney : from normal

development to congenital diseases." Academic Press, Amsterdam ;

Boston.

Ward, C. J., Yuan, D., Masyuk, T. V., Wang, X., Punyashthiti, R., Whelan, S.,

Bacallao, R., Torra, R., LaRusso, N. F., Torres, V. E., and Harris, P. C.

(2003). Cellular and subcellular localization of the ARPKD protein;

fibrocystin is expressed on primary cilia. Hum Mol Genet 12, 2703-10.

Wellik, D. M., Hawkes, P. J., and Capecchi, M. R. (2002). Hox11 paralogous

genes are essential for metanephric kidney induction. Genes Dev 16,

1423-32.

Xu, P. X., Zheng, W., Huang, L., Maire, P., Laclef, C., and Silvius, D. (2003). Six1

is required for the early organogenesis of mammalian kidney.

Development 130, 3085-94.

Yoshino, K., Rubin, J. S., Higinbotham, K. G., Uren, A., Anest, V., Plisov, S. Y.,

and Perantoni, A. O. (2001). Secreted Frizzled-related proteins can

regulate metanephric development. Mech Dev 102, 45-55.

Zhao, Y., and Potter, S. S. (2001). Functional specificity of the Hoxa13

homeobox. Development 128, 3197-207.

Zhao, Y., and Potter, S. S. (2002). Functional comparison of the Hoxa 4, Hoxa

10, and Hoxa 11 homeoboxes. Dev Biol 244, 21-36.

43 Figures

Figure 1. Development of the metanephric kidney. Metanephric kidney

(metanephros) development is initiated at E10.5. The condensing metanephric mesenchyme (compact red cells) has induced the ureteric bud to evaginate from the posterior nephric duct. The mesonephric tubules (green) of the mesonephros can be identified at the anterior aspect of the urogenital ridge, while the mesenchyme (sparse red cells) of the intermediate mesoderm separates the mesonephros from the early metanephric field. At E11.5, the ureteric bud has branched once forming a “T” shaped structure surrounded by condensed metanephric (nephrogenic) mesenchyme (red). From E12.5 to E13.5, the ureteric branching morphogenesis continues and the metanephric mesenchyme

44 Figure 1 continued. has become two distinct populations: the condensed nephrogenic mesenchyme

(red), surrounding the ureteric tips, and the stromal mesenchyme (blue), an important regulator of branching morphogenesis. Additionally, the ureteric tips have induced the surrounding nephrogenic mesenchyme to undergo mesenchymal to epithelia transformation forming renal vesicles (box). The renal vesicles elongate to form an early nephron structure, the “S” shaped body (red).

As development continues from E14.5 to after birth, the nascent nephrons undergo further elongation and differentiation forming the mature nephron consisting of the glomerulus and tubules (red) which are derived from the metanephric mesenchyme, while the collecting duct (white) is derived from the ureteric bud.

45

Figure 2. Signaling and transcriptional regulation of ureteric branching morphogenesis and metanephric mesenchyme nephrogenesis during kidney development. Branching morphogenesis of the ureteric bud (yellow) is primarily driven by nephrogenic mesenchyme (NM, red) expression of Gdnf. The

Gdnf receptor, Ret, and Wnt11 are expressed by the ureteric tips of the ureteric

tree. Many transcriptions factors within the NM are critical for Gdnf activation in

the mesenchyme, while Foxc1/c2 represses ectopic mesenchymal Gdnf

expression. The stromal mesenchyme (SM, blue) is characterized by expression

of the retinoic acid receptors, Rara and Rarb2, and Foxd1. These transcription

factors influence Ret expression within the ureteric bud and tree through an

unknown mechanism which may or may not involve signaling through the NM.

46 Figure 2 continued.

The mesenchymal to epithelial transformation of the nephrogenic mesenchyme into a nephron (nephrogenesis) requires ureteric expression of Wnt9b. Wnt9b induces mesenchymal expression of Wnt4 and Fgf8 which are required for normal nephrogenesis. Fgf8 expression in the mesenchyme then activates expression of Lhx1 allowing nephrogenesis to advance from the renal vesicle stage.

47

Figure 3. Hox gene expression in the metanephric kidney and the “S” shaped developing nephron. (A) The diverse expression of Hox genes within the ureteric bud and metanephric mesenchyme compartments of the developing kidney. Hox genes are categorized depending on the ureteric-specific expression, metanephric mesenchyme-specific expression, or expression in both

48 Figure 3 continued. compartments. Notice the similar expression patterns of many different Hox genes suggesting functional redundancy in kidney development. The 3’ Hox genes are generally expressed in the anterior embryo during development, while

5’ Hox genes are generally expressed in the posterior embryo. (B) Hox expression in the collecting duct and developing nephron segments: the connecting segment (cs) and distal segment (ds), intermediate segment (is), and proximal segment (ps). For example, Hoxa9, Hoxa10, Hoxb8, Hoxd9, Hoxd10, and Hoxa11 are all expressed in the cs and ps suggesting a role in the patterning of these nephron segments. The unique Hox expression patterns suggest these important transcription factors may have significant roles in collecting duct and nephron patterning. Adapted from Patterson and Potter (2004).

49 Chapter 2

A catalogue of gene expression in the developing kidney.*

Kristopher Schwab2, Larry T. Patterson3, Bruce J. Aronow1, Ruth Luckas3,

Hung-Chi Liang2, and S. Steven Potter2

Division of Bioinformatics1, Developmental Biology2 and Nephrology3

Children's Hospital Medical Center

3333 Burnet Avenue

Cincinnati, OH 45229

Published in Kidney Int. 2003 Nov;64(5):1588-604.

50 Abstract

Background. Although many genes with important function in kidney morphogenesis have been described, it is clear that many more remain to be discovered. Microarrays allow a more global analysis of the genetic basis of kidney organogenesis.

Methods. In this study, Affymetrix U74Av2 microarrays, with over 12,000 genes represented, were used in conjunction with robust target microamplification techniques to define the gene expression profiles of the developing mouse kidney.

Results. Microdissected murine ureteric bud and metanephric mesenchyme as well as total kidneys at embryonic day E11.5, E12.5, E13.5,

E16.5, and adult were examined. This work identified, for example, 3847 genes expressed in the E12.5 kidney. Stringent comparison of the E12.5 versus adult recognized 428 genes with significantly elevated expression in the embryonic

kidney. These genes fell into several functional categories, including transcription

factor, growth factor, signal transduction, cell cycle, and others. In contrast,

surprisingly few differences were found in the gene expression profiles of the

ureteric bud and metanephric mesenchyme, with many of the differences clearly

associated with the more epithelial character of the bud. In situ hybridizations

were used to confirm and extend microarray-predicted expression patterns in the

developing kidney. For three genes, Cdrap, Tgfbi, and Col15a1, we observed

strikingly similar expression in the developing kidneys and lungs, which both

undergo branching morphogenesis.

51 Conclusion. The results provide a gene discovery function, identifying large numbers of genes not previously associated with kidney development. This study extends developing kidney microarray analysis to the powerful genetic system of the mouse and establishes a baseline for future examination of the many available mutants. This work creates a catalogue of the gene expression states of the developing mouse kidney and its microdissected subcomponents.

52 Introduction

The kidney provides a powerful model system for study of the principles of organogenesis. The developing kidney employs many developmental mechanisms, including budding, reciprocal inductive tissue interactions, stem cell growth and differentiation, cell polarization, mesenchyme to epithelia transformation, branching morphogenesis, angiogenesis, apoptosis, fusion

(nephrons to collecting ducts), proximal-distal segmentation (along the length of the nephron), and the differentiation of several interesting cell types.

Furthermore, nephrogenesis proceeds readily in culture [1], providing an important experimental advantage.

Significant advances have been made in understanding the genetic basis of kidney organogenesis. The un-induced metanephric mesenchyme produces glial-derived neurotrophic factor (GDNF), which interacts with the ret receptor on

the ureteric bud to promote outgrowth. Mice with mutation of either the GDNF or

ret, or its coreceptor the GFRa1 gene, have severe failure of kidney development

[2, 3]. There is also strong evidence implicating pleiotrophin as a mesenchyme-

synthesized inducer of the ureteric bud and branching morphogenesis [4]. The

initial signaling from the bud to the metanephric mesenchyme is less well

understood, but Wnt6, synthesized by the bud and able to induce tubulogenesis

in vitro, is a candidate [5].

Kidney organ culture and cell line studies have implicated a number of

genes in kidney development, including midkine [6], phosphatidylinositol 3 kinase

(PI3K) [7], transforming growth factor-β1 (TGF-β1) [8], galectin-3 [9], bone

53 morphogenetic protein-4 (BMP-4) [10], vascular endothelial growth factor (VEGF)

[11], hepatocyte growth factor (HGF ) [12], and others.

Gene-targeting studies have shown that Lim1 [13], Hoxa 11/Hoxd 11 [14–16],

Eya1 [17], WT-1 [18], Sall1 [19], a3 b1 integrin [20], a8 b1 integrin [21], Emx-2

[22], FOXd1 (BF-2) [23], RARa, and RARb [24] are all essential for kidney

development. BMP-7 is essential for continued kidney growth [25, 26], and Wnt4,

synthesized in the mesenchyme, is required for development past the aggregate

stage of nephrogenesis [27]. Cadherin-6 promotes mesenchyme to epithelia

conversion and nephron formation [28]. N-Myc promotes cell proliferation [29].

LIF appears an important signal from the later ureteric bud to the differentiating

nephrons [30]. Despite thisimpressive progress, it is clear that we have only

begun to understand the genetic basis of kidney formation.

Microarrays offer the opportunity to determine global definitions of the

gene expression states of developing organs. As microarrays approach

comprehensive coverage of the genome, it is becoming possible to assay

expression levels of every gene. As target amplification procedures become

more powerful, it is becoming possible to perform microarray analysis with ever-

decreasing amounts of starting RNA. The ultimate goal is to define the complete

gene expression patterns of individual cell types as they progress through kidney

organogenesis. Microarrays have been used to generate gene expression

profiles of cell lines representing specific stages of kidney development [31]. The

mK3 and mK4 cells correspond to early metanephric mesenchyme and later

metanephric mesenchyme, undergoing epithelial transformation. Microarrays

54 identified thousands of genes expressed in these two cell types and thereby implicated them in early kidney development. Comparison of the mK3 and mK4 gene expression profiles found a large number of genes differently expressed, likely reflecting changes in gene expression during epithelial transformation [31].

In another study, microarrays were used to examine kidney development in the rat [32]. Cluster analysis identified five groups of genes with interesting expression patterns. The entire set of microarray data, for 8741 genes, has been made available (http://organogenesis.ucsd.edu/). This work represents an important step in the use of microarrays to perform a global analysis of gene expression states in the developing kidney. In this report,we extend these previous studies by performing a microarray analysis of early kidney development in the mouse. Affymetrix murine U74Av2 genechip probe arrays with over 12,000 genes represented were used. Robust microamplification techniques were used to allow study of extremely small samples. Gene expression profiles were determined for microdissected embryonic day E11.5 ureteric buds and metanephric mesenchyme as well as E11.5, E12.5, E13.5,

E16.5, and adult total kidneys. The results identified extensive sets of genes, of multiple functional categories, expressed in developmental timing and compartment specific patterns. In situ hybridizations were used to corroborate microarray expression data and to further define the expression patterns of selected genes during kidney development. This work creates a catalogue of the gene expression states of the total developing mouse kidney and selected

55 subcomponents. This provides a baseline for the analysis of the many mouse mutants available with altered kidney development.

56 Materials and methods

Dissection of early metanephric and adult tissues and RNA isolation

E11.5 and E12.5 kidneys were dissected from wild-type CD-1 mice, pooled, and frozen at -80ºC. Ureteric buds and metanephric mesenchyme were microdissected from E11.5 metanephric kidneys following mild trypsinization, pooled separately, and then frozen at 80 C. E13.5, E16.5, and adult kidneys were isolated and frozen at -80ºC. All dissections were performed in phosphate- buffered saline (PBS).

Total RNA was prepared from ureteric bud, metanephric mesenchyme, and early kidney samples using Stratagene Absolutely RNA Nanoprep kit (La

Jolla, CA, USA) for small samples. Adult kidney and P1 whole mouse total RNA was prepared using RNAzol (Tel-Test, Friendwood, TX, USA).

RNA amplification and target RNA isolation

Total RNA was linearly amplified by using a previously described procedure [33] with 30 ng (ureteric bud, metanephric mesenchyme, and two adult kidney replicates) or 100 ng (remaining samples, including two adult kidney replicates). Briefly, total RNA was reverse transcribed into cDNA using a T7 promoter-dT primer [5’-GGCCAGTGAATTGTAATACGACTCACTATA-

GGGAGGCGG-(T)24], amplified through an in vitro transcription reaction using

T7 RNA polymerase, and the products then reverse transcribed into cDNA again, using random hexamer primers. A final in vitro transcription reaction using the

Bioarray High Yield RNA transcript labeling kit (Enzo Life Sciences, Farmingdale,

57 NY, USA) was performed producing biotinylated cRNA for microarray hybridization.

Microarray analysis of each developmental stage/tissue of the developing kidney and adult kidney was performed in biologic duplicate by obtaining tissues from different mice. In addition, technical replicates of the same adult kidney total

RNA sample were performed using both 30 ng and 100 ng amounts of total RNA.

Technical duplicates were also performed from the same P1 whole mouse total

RNA preparation. The same amplification procedure was used for all samples.

Gene expression profile analysis

Amplified, biotinylated cRNA samples were hybridized to Affymetrix murine U74AV2 microarrays, according to standard procedures as described by

Affymetrix (Santa Clara, CA, USA). Both Microarrary Suite 5.0 (Affymetrix) and

Gene-Spring 4.2.1 and 5.1 (Silicon Genetics, Inc., Redwood City, CA, USA) software were used for data analysis. For cluster analysis, data were normalized using per chip scaling in RMA [Affymetrix package in R that has been developed by the Bioconductor Consortium (http://bio-conductor.org)] followed by transformation of log2 signal back to linear signal that was then normalized to the day 1 whole mouse reference. All group to group comparisons (metanephric mesenchyme:ureteric bud, E11.5:E12.5, E11.5:E13.5, E11.5:E16.5,

E12.5:E13.5, E12.5:E16.5, E13.5:E16.5, and adult kidney:whole mouse for

genes with relative expression more than twofold higher in adult kidney) were

performed using Welch t test analysis of variance (ANOVA) with P < 0.05 without

multiple testing rate correction because most samples were collected and

58 analyzed with two replicates. Genes that passed the ANOVA were then ranked by average fold difference with cutoffs used that were appropriate for each comparison (>~ twofold for each group), pooled, and converted back to log expression ratios, which were then subjected to the hierarchical tree clustering algorithm with the Pearson distance metric as implemented in GeneSpring 5.1

(Silicon Genetics, Inc.). All Affyme-trix.cel files, RMA results, gene lists, and gene trees are available at the following website: http://genet.chmcc.org (login is

Nephrome, password is reviewer, click login, not login as guest, click continue on

U74Av2_mouse).

RNA tissue section in situ hybridization

E12.5, E13.5, E15.5, and E17.5 embryos were obtained from CD-1 mice.

Embryos and tissues underwent fixation and embedding as previously described

[15]. Four micron sections, including the developing kidney, were made. In situ

33 hybridizations to RNA in tissue sections were performed using a P-uridine triphosphate (UTP)-labeled riboprobe generated from cDNA clones. The following probes were made from the corresponding cDNA clone: Birc5, 373242;

Capn5, 733856; Cdrap, 427314; Col15a1, 406344; Mns1, 480118; Nr2f1,

420535; Rab6kifl, 764732; Smoh, 439010; and Tgfbi, 734101. All cDNA clones were obtained from the Incyte GEM1 mouse library (University of Cincinnati

Genomics and Microarray Laboratory). Slides were then dipped in Kodak NTB-2 nuclear track emulsion and exposed for 4 weeks at 4°C. After exposure, slides were developed using Kodak D-19 developer at 15°C. Slides were stained with

4’,6’-diamidino-2-phenylindole hydrochloride (DAPI).

59 Results

The goal of this work was to establish a reference data set of gene expression states during normal early mouse kidney development. This data set could facilitate the identification of new genes and pathways involved in kidney development and provide a standard for microarray studies of the many mouse mutants available with abnormal kidney development.

Reproducibility of the target amplification procedure

Powerful amplification techniques were necessary to analyze the small samples provided by the early embryonic mouse kidneys and their microdissected components. Several protocols were tried and the one described by Baugh et al [33], using two rounds of in vitro transcription, gave excellent results. Technical replicates were performed to measure reproducibility of the resulting microarray data. Total RNA was prepared from a single adult kidney and multiple small aliquots were removed and used in parallel for target amplification and hybridization to Affymetrix murine U74Av2 microarrays. Figure

1A shows a scattergram comparing the raw hybridization signals from two samples of 30 ng of total RNA. Each point represents a single gene, with the hybridization levels from the two microarrays shown on the two axes. A single line at 45° would designate perfect reproducibility. The two lines flanking the center line indicate points of twofold difference in signal intensity in the two microarray hybridizations. Figure 1A only shows genes with sufficient hybridization signal to be called expressed (P) by Microarray Suite 5.0 (MAS5).

We observed a high level of technical reproducibility for these genes with well

60 below 1% (43/12488) called more than a two-fold different in expression by

MAS5. Genes with very low levels of expression, near noise, and called absent or not-expressed by MAS5, gave more variation, but in performing analyses, we used expression level cutoffs to exclude these genes.

Biologic duplicates

In this study all kidney samples, embryonic and adult, were examined in biologic duplicate, using RNA from separate biologic samples. Figure 1B illustrates the reproducibility of biologic duplicates for E12.5 kidneys. In this case,

the biologic noise was added to the technical noise, and as expected, more

variation was observed than for technical replicates. Nevertheless, in biologic

replicates, there were still well below 1% of genes showing more than a twofold

difference in gene expression.

This modest noise, of less than 1%, is almost entirely eliminated by

performing experiments in biologic duplicate. The 1% of the approximately

12,000 genes on a microarray is 120 genes, a significant number. But the noise

is largely random, so only 1% of these 120 genes, or approximately one gene,

would again show noise-related variation on the duplicate.

The use of biologic duplicates provides a minimum of two times two, or

four, comparisons between two sample types. By screening out genes with low

expression levels and genes with less than threefold change in expression, we

have previously observed over 95% of genes identified as differently expressed

in such comparisons are validated by Northern blot, reverse transcription-

polymerase chain reaction (RT-PCR), and/or Western blot [31].

61 Gene expression state of the E12.5 kidney

The set of genes expressed in the embryonic kidney defines its potential for growth and development. Microarrays, by providing a more comprehensive and unbiased view, can promote the discovery of new genes not previously implicated in developmental processes. For example, 5542 genes were called expressed in both biologic duplicates of the E12.5 kidney by MAS5, with the

detection P value set at the default 0.04. Errors in absence-presence calls occur

most often for genes expressed at a low level, near noise. Excluding genes with

expression signal below 100 [glyceraldehyde-3-phosphate dehydrogenase

(GAPDH) signal, for example, was 5237 and 5149 for the replicates], MAS5

called 3914 genes expressed. Further screening of these genes by selection with

a detection P value of less than or equal to 0.03 found 3847 genes expressed in

the E12.5 developing kidney. This list of genes active during early kidney

development provides an important foundation for further study of kidney

morphogenesis. This gene collection is too extensive for detailed discussion, but

the data set is fully provided at a supplementary Web site. All Affymetrix. cel files,

RMA results, gene lists, and gene trees are available at http://genet.chmcc.org

(login is Nephrome, password is reviewer, click login, not login as guest, click

continue on U74Av2_mouse).

Genes with elevated expression levels in the E12.5 kidney

Among the genes expressed in the developing kidney are housekeeping

genes and others of relatively little developmental consequence, many of which

will be expressed in both the embryonic and adult kidney. Of particular interest

62 from a development perspective are the genes comparatively more active in the early developing kidney, which might play important roles in the many processes of kidney formation. A scattergram presentation of the many differences in the gene expression patterns of the E12.5 and adult kidneys is shown in Figure 1C.

We performed a stringent screen of the microarray data to identify genes with elevated expression levels in the E12.5 embryonic kidney when compared to adult. The E12.5 microarray analysis was performed in biologic duplicate, and the adult kidney analysis was performed in biologic duplicate, with one sample analyzed in technical quadruplicate, giving a total of five adult microarrays, and allowing a total of ten crosswise comparisons, between the two embryonic and

five adult microarray hybridizations. The microarray analysis software MAS5 found 1858 genes with consistently increased expression levels in the embryonic kidneys versus adult in all ten comparisons. Further screening of these genes required that all ten comparisons show greater than threefold (signal log ratio of

1.7) higher expression in embryonic kidney, and that the expression signal in embryonic kidney be greater than 150, to further reduce possible artifacts due to low signal to noise ratio.

This rigorous screen identified 428 genes with significantly elevated expression in the embryonic kidney. Approximately 10% of these genes are involved in the regulation of transcription (Table 1). Some of these (Taf1a,

Gtf2e2, Gtf2h1, and Taf 9) can be grouped as general transcription factors.

Others play an important role in regulating gene expression through their effects on chromatin configuration. These include the histone deacetylases genes

63 Hdac1 and Hdac2, the SWI/SNF-related Smarce1, BRG/brm-associated factor

53A, the HMG box genes Hmgb2 and Hmga2, the chromobox homolog Cbx2, the histone methylase gene silencer Setdb1, and the Polycomb group genes eed and EZH2, which form a complex together [34], and inactivate Hox genes through histone methylation [35].

Another group of genes were found that likely encode transcription factors based on motif, protein interactions, or strong homologies to known transcription factors. These include the zinc finger encoding genes Zfp105 and Zfp61, the zinc

finger homeobox gene Zfhx1a, the HMG gene Hmgn2, and others such as

Trp53bp1, Cbfb8, Mcmd4, Cnbp6, and Trim27.

The E12.5 kidney also expressed 19 better-characterized transcription factor genes at elevated levels. Several of these have been previously implicated in kidney development. The only Hox genes identified by these strict screening criteria were Hoxa 11 and Hoxd 11, which give absent or rudimentary kidneys when both are mutated [14, 15]. Pbx3, encoding a homeodomain protein that interacts with Hox proteins was also identified. Sox11 was previously associated with mesenchyme-epithelia transformation in kidney development by a differential display study [36]. The Six2 gene was previously shown to be down- regulated in compound Hoxa 11, Hoxd 11, Hoxc 11 triple mutant kidneys [16].

The Six, Eya, and Dach genes form a regulatory network in several developing systems, and EYA1 mutant mice have absent kidneys [17]. Tcf12 encodes a basic helix loop helix transcription factor. The forkhead FOXd1 (BF-2) gene, with important function in stromal cells [23], was expressed. The FOXc2 gene, also

64 encoding a forkhead transcription factor, was also expressed at elevated levels at E12.5. FOXc2 has been implicated in the repression of GDNF signaling, with

FOXc1 mutants showing urinary tract duplications and the FOXc1 and FOXc2 genes showing considerable functional redundancy in both cardiovascular and kidney development [37]. N-Myc was also identified, and mutation of N-Myc gives poorly developed kidneys [29]. The mesenchyme homeobox gene Meox2 is of interest. The Meox genes, also called Mox, have been shown to interact with

PAX proteins [38], which are known to play a major role in kidney development.

In addition, the Meox genes have been implicated in mesenchyme-epithelia interactions [39], again important in kidney formation. Other transcription factor genes identified include Zipro, a zinc finger proliferation gene, Pttg1, a pituitary tumor transforming gene, Maged1, Tea Domain gene TEAD2, the zinc finger

Snail2, the interleukin binding factor Ilf3, the MORF-related Mrgx, and ERH

(enhancer of rudimentary homolog).

It is clear from the microarray data that a major function of the early kidney is

simply to grow. Thirty five cell cycle-associated genes were found elevated in

expression in the E12.5 kidney, consistent with a high rate of cell division. An

additional 90 genes with high expression in the developing kidney were involved

in intermediary metabolism, ribosome biogenesis, DNA synthesis, and other

processes connected with rapidly dividing cells. A similar conclusion was drawn

in a previous microarray study of rat kidney development [32].

Twenty one genes with roles in signaling were expressed at elevated

levels in the mouse embryonic kidney. Several of these are involved in

65 guanosine triphosphatase (GTPase)-mediated signaling, including RAN binding protein 1 (Ranbp1), Rac GTPase activating protein Racgap1, the oncogene ect2, and the ras homologs Arhu and Ras-like family 2 locus 9 (Rasl2-9). Other genes of interest were 2310076D10, involved in mitogen-acti-vated protein kinase

(MAPK) inactivation, and the Wsb1 gene, with a SOCS (suppressor of cytokine

signaling) box.

Growth factor–related genes detected by this stringent screen in the early

kidney included GDNF, midkine, secreted frizzled-related protein 2 (Sfrp2),

follistatin-like protein (FSTL), and hepatoma-derived growth factor– related

protein2 (Hdgfrp2). Although kidney development functions for GDNF and

midkine have been extensively investigated [2, 6], the possible roles of Sfrp2,

Hdgfrp2 and follistatin-like protein in kidney formation are largely unexplored.

Several interesting extracellular matrix (ECM) -encoding genes were

expressed at elevated levels in the developing kidney. Chondroitin sulfate

proteoglycan 6 (also called Bamacan) has been previously associated with

mesangial cells and tubulogenesis. Fibulin 1 and laminin alpha 4 were also

expressed, as was collagen type V alpha 2. Glypican 3 mutations have been

linked to kidney malformations in both mice and humans [40, 41], likely due to

altered growth factor signaling [42]. Other ECM genes of interest include

microfibrillar-associated protein 2 and Emu2. The Emu2 gene is often expressed

in mesenchymal cells undergoing mesenchymal-epithelial interactions [43].

Many other genes of interest were found expressed at elevated levels in the

early embryonic kidney. These included the protease inhibitors Serpinf1 and

66 Serpinh1. In addition, six receptor encoding genes were found, including the nuclear receptor Nr2f1 (COUP-TF1), the ephrin receptor Ephb4, the frizzled receptor Fzd2, the laminin receptor 1, the netrin receptor Unc5h3, and the cytokine receptor-like factor Crlf1. Several of these receptors can be linked to known processes in kidney development. For example, it is known that the ureteric bud cell line up-regulates the expression of the membrane-bound ephrins when undergoing branching morphogenesis in culture [44]. There were also 125 expressed sequence tags (ESTs) of currently unknown function that showed high expression levels in the developing kidney.

It is important to emphasize that this list of 428 genes does not represent a complete compilation of developmentally important genes expressed in the early developing kidney. It excludes, for example, many Hox genes expressed in both the developing and adult kidney. The high stringency of the screen also eliminated some genes of known importance in kidney development. Other genes were not represented on the microarrays. This study does, however, implicate a large number of interesting new genes in the process of kidney formation.

Genes differentially expressed between the E11.5 metanephric mesenchyme and ureteric bud

Gene expression profiles were determined for isolated ureteric bud and metanephric mesenchyme of the early developing kidney. These two elements were separated by microdissection from E11.5 kidneys following mild trypsinization. The similarity of the gene expression patterns of these two

67 compartments that give rise to the kidney is shown in Figure 1D. Using a parametric t test (P < 0.05) we identified only 78 genes that were reproducibly differentially expressed between the E11.5 metanephric mesenchyme and ureteric bud (Table 2). This relatively short list of consistent differences indicates that the gene expression states of the metanephric mesenchyme and ureteric bud are surprisingly similar, perhaps reflecting their common intermediate mesoderm origin. We identified 53 genes with increased expression in the ureteric bud, and 25 genes with increased expression in the metanephric mesenchyme.

Ureteric bud–enriched genes

The list of ureteric bud-enriched genes includes many previously shown to be important in metanephric kidney development, or specifically associated with epithelial structures. Ten of the 53 genes more highly expressed in the ureteric bud function in forming cell-cell adhesions or ECM (Table 2). Claudin genes 3, 4,

7, and 8 were enriched in the ureteric bud. Claudins are integral membrane proteins that function in forming tight junctions between epithelial cells. Cartilage- derived retinoic acid-sensitive protein is an ECM protein containing a SH3 domain that is repressed in the presence of retinoic acid and functions in cell culture as a tumor growth inhibitor [45, 46]. Cadherin 16 was first identified as a kidney-specific cadherin and is expressed in the epithelia of the developing kidney [47, 48]. Laminin gamma 2 is a subunit of the laminin-5 complex and is found within the basal lamina of the ureteric bud [49]. In vitro studies have shown

68 that ureteric bud branching is decreased when laminin-5 is blocked through use of antibodies to itself and its receptors, the α3 and α6 integrins [50].

Genes with elevated expression in the bud also include growth factors and signal transducers. Two identified growth factors, TGF-α and WNT6, were previously implicated in kidney development. TGF-α is expressed in the ureteric bud tips and ducts and is able to induce metanephric kidney growth in vitro, while growth is inhibited by TGF-α antibodies [51, 52]. WNT6 is expressed in ureteric bud and can induce nephrogenesis in vitro [5].

Notable signaling transducers detected in the ureteric bud include GFRA1,

FGFr3, and ROS1. GFRA1 is co-receptor for RET, which is also expressed in the ureteric bud. Mutation of Gfra1 gives a similar phenotype to the GDNF and Ret knockout mice, in which there is an absent kidney due to loss of ureteric bud outgrowth [53]. FGFr3, a fibroblast growth factor (FGF) receptor, may function in early signaling within the metanephric kidney, as expression of a secreted dominant-negative FGF receptor in transgenic mice results in kidney agenesis

[54].

Several genes involved in calcium-mediated signaling showed elevated expression in the ureteric bud. This is consistent with the observed response of

Madin-Darby canine kidney (MDCK) cells, which quickly become polarized and form tight junctions and desmosomes following calcium level elevation [55].

VSN11 is a calcium-dependent regulator of cellular signaling [56]. The calcitonin

receptor is known to regulate Ca++ excretion in the kidney and might also play a

role in organogenesis [57]. Calbindin-28K, a calcium-binding protein, was

69 previously shown to be expressed in the ureteric bud [58], and TACSTD2, a tumor-associated calcium signal transducer, also showed elevated expression the ureteric bud.

Only four ureteric bud elevated genes, including Tcf2, Pea3, and En-2, encoded known transcription factors. Tcf2 is expressed in the forming collecting

ducts in the developing kidney as well as in the liver [59]. Tissue-specific

targeting in the liver shows that this gene is essential for bile duct morphogenesis

[60], suggesting a possible role in branching morphogenesis in the kidney.

Recently, Pea3 has been implicated in muscle innervation. Of interest, Pea3 is

downstream of GDNF in motor neurons [61]. En-2, an engrailed homolog, has

been found to regulate boundaries in the developing brain and has not been

previously implicated in kidney development [62].

Other genes identified encoded proteases and protein modifiers such as

CAPN5, hepsin, and transmembrane protease 2. Also, a cell cycle regulator,

Cdkn1a, was identified. Other genes with elevated expression in the ureteric bud

were involved in metabolism, the cytoskeleton, or of unknown function, including

ESTs.

Metanephric mesenchyme–enriched genes

We identified 25 genes more highly expressed in the metanephric

mesenchyme than in the ureteric bud. Six of these encoded the transcription

factors WT-1, FOXc1, SOX2, SOX18, HOXA 10, and C1D. The WT-1 gene has

previously been shown to be expressed in the metanephric mesenchyme and is

required for early kidney development [18]. FOXc1, a forkhead/winged helix

70 transcription factor, as discussed previously, is an apparent regulator of GDNF signaling. Sox2 and Sox18 encode transcription factors containing HMG boxes.

For Hoxa 10, we have confirmed restricted expression in the metanephric mesenchyme (Patterson, unpublished observations, 2003). C1D is a DNA high- affinity binding protein that serves as a nuclear receptor corepressor [63].

Genes encoding four signal transduction proteins, TEK, TIE1, SGK, and FES,

were enriched in the metanephric mesenchyme. TEK and TIE1, receptor tyrosine

kinases, function in vascular endothelium development and mutations in these

genes cause endothelial defects [64, 65]. Of particular interest, TIE-deficient cells

were unable to contribute to adult kidney epithelium in chimeras [65]. Sgk1

encodes a serum/glucocorticoid-regulated kinase. Targeted mutation of sgk1

gives an impaired sodium retention phenotype [66]. The protein kinase FES has

been reported expressed in epithelia cells, hematopoietic cells, and vascular

endothelial cells [67]. Transgenic studies indicate a role for FES in angiogenesis

[68]. Targeted mutation of the fes gene revealed function in cardiovascular

development and myeloid cell proliferation [69]. It is interesting that the E11.5

mesenchyme already expressed so many genes (Sox18, tek, tie, and fes)

associated with vasculogenesis.

Other genes of interest more highly expressed in the metanephric

mesenchyme included Vamp5, which encodes an integral membrane protein that

is enriched during in vitro myogenesis and is highly expressed in heart and

skeletal muscle, but with lower levels also detected in the adult kidney and other

organs by Northern blot [70]. Adult hemoglobin alpha and beta chains, lysozyme

71 (Lyzs), an antimicrobial enzyme produced by macrophages, and aminolevulinic acid synthase 2 (Alas2), an erythroid-specific enzyme functioning in heme production, are enriched within the metanephric mesenchyme. The elevated expression of these blood-related genes could simply reflect a greater blood content for the metanephric mesenchyme compared to the ureteric bud.

Alternatively, it could suggest a possible hematopoietic function for the metanephros, similar to that previously described for the mesonephros. Five EST genes, with unknown functions, were also enriched in the metanephric mesenchyme.

Genes with elevated expression at E16.5

The E16.5 kidney showed elevated expression of many genes when compared to E12.5 (Table 3). Of particular note, there were 32 aquaporin, ion

channel, solute carrier, and other transport molecules actively expressed at

E16.5, consistent with ongoing tubule maturation. A number of other genes with

previously assigned kidney function were also identified, including nephronectin

(integrin ligand), polyductin (polycystic kidney disease), neuropilin (semaphorin

receptor), uromodulin, glomerular epithelial protein 1, and renin. In addition, a

large number of genes not previously associated with kidney development, from

a variety of functional categories, were found (Table 3).

Cluster analysis of gene expression during kidney development

A more comprehensive view of the changes in gene expression that take

place during kidney development can be provided by hierarchical cluster analysis

(Fig. 2). Rows represent the different tissues examined, which included E11.5

72 metanephric mesenchyme, E11.5 ureteric bud, E11.5, E12.5, E13.5, E16.5, and adult total kidney, as well as P1 total mouse. Genes with similar expression patterns were clustered using hierarchical tree algorithm applied to the log relative gene expression values using Pearson correlation. Red represents high expression and blue represents low expression levels. The genes on the right showed higher expression in the adult kidney, and genes on the left had higher expression in the developing kidney. Figure 2 also provides graphic illustration of the similarities in gene expression patterns of the embryonic kidney at E11.5,

E12.5, and E13.5. Compared with the striking embryo-adult differences, the changes in gene expression at these early developmental time points are subtle.

The E16.5 kidney, in contrast, shows a distinctive gene expression profile when

compared with either the embryonic or adult kidney, consistent with its

intermediate state of differentiation

In situ hybridization confirmation of microarray results

To confirm and extend the microarray results, we performed in situ hybridizations. The microarray data identified a large number of interesting genes, many of which had been previously implicated in kidney development, providing significant validation of the approach. For other genes there had been only indirect, or in many cases no previous association with kidney formation. To corroborate predictions of the microarray data and to better define the expression

73 patterns of selected genes in the developing kidney, we performed in situ hybridizations.

The Smoothened (Smoh) gene was predicted by the microarray results to

be expressed in the early developing kidney. In situ hybridizations showed a

graded Smoh expression pattern in the E12.5 metanephric mesenchyme, with

higher levels in prestromal regions more distant from the ureteric bud (Fig. 3 A to

C). This is consistent with a recent study showing that SHH expression in the

ureteric bud is an important regulator of proliferation and differentiation of the

metanephric mesenchyme in the developing kidney [71].

Nr2f1 (COUP-TFI) expression was observed in the mesenchyme of the

E13.5 kidney (Fig.3Dto F). Similar to Smoh, expression was excluded from the

bud and its flanking mesenchyme. At E15.5, Nr2f1 expression remained off in the

developing collecting ducts and was maintained in mesenchyme and forming

tubules (Fig. 3Gto I). Nr2f1 is an orphan member of the nuclear receptor family

with no identified ligand [72]. It was previously shown to be highly expressed in

developing neural tissue of the brain [73] and when knocked out in the mouse

defects are observed in development of the peripheral nervous system ganglia

[74].

At E17.5, Birc5 (survivin) expression was observed in the developing

tubules and glomeruli (Fig. 3 J to L). This gene functions to enhance proliferation

and survival of cells.

Capn5 expression was limited to the epithelia of the collecting ducts at

E15.5 (Fig.3M to O). This gene is a member of the calpain family, which encodes

74 calcium-dependent intracellular proteases but lacks a calmodulin-like domain [75] and may act within signaling cascades [76].

The Mns1 gene (meiosis-specific nuclear structural protein 1) [77] showed expression in the early nephrogenic epithelia in the cortex (Fig. 3 P to R). The

Rab6kifl gene showed restricted expression in the mesenchyme surrounding the ureter in the E12.5 kidney (Fig. 3 S to U). The Rab6kifl encoded protein is localized to the golgi apparatus and includes a kinesin motor domain as well as a

Rab6 GTP interacting domain [78].

Although the focus of this study is on the developing kidney, the in situ hybridizations also defined expression domains in the other regions of the

embryo. For some genes the expression patterns in the developing lung, which

also undergoes extensive branching morphogenesis, were particularly striking.

The Cdrap (or MIA) gene was originally cloned as a secreted protein from

human melanoma cell lines. It was reported expressed in cartilage primordia [46],

as well as transiently in embryonic mammary buds, suggesting a possible role in

embryonic events involving invasive growth [79]. Cdrap is a secreted ECM

protein that adopts an SH3 domain-like fold in solution. The gene was recently

knocked out. Malformations were detected in mutant mice in collagen fiber

density, diameter and arrangement [80]. The mice are viable and have normal

mammary gland development, but other tissues were not extensively examined.

The microarray results suggested that Cdrap is expressed in the ureteric bud of

the developing kidney. As predicted, Cdrap showed high expression in the distal

ends of the branching ureteric bud of the E17.5 embryo (Fig.4 A to C). Of

75 interest, we also observed Cdrap expression in the distal tips of the branching bronchi of the lung (Fig.4 D to F).

The Tgfbi gene was isolated as a gene that is induced by TGF-β1, and prevents cell adhesion [81]. It is mutated in corneal dystrophies [82]. The protein is a secreted, RGD-containing collagen-associated protein, and is thought to interact with the ECM. Consistent with the microarray results, this gene was expressed in the mesenchyme of the developing kidney at both E12.5 and E15.5

(Fig. 4 G to L). Similar expression was found in the developing lung at E15.5

(Fig. 4 M to O). These results support and expand earlier expression analysis of this gene [83].

There is evidence from cell culture studies that the collagen Col15a1 gene may be modulated by cytokines [84]. The gene has been previously targeted causing muscular disease and defects detected in the cardiovascular system in mice [85]. In situ hybridizations confirmed the microarray predicted expression in the developing kidney, with distinct patches of expression observed in the

mesenchyme of both the kidney and lung at E13.5 and E15.5 (Fig. 4 P to A’).

This data support and expand a previous Col15a1 expression analysis [86].

76 Discussion

By coupling microarrays and robust target amplification techniques it is possible to perform global studies of the gene expression patterns of early

developing organs and microdissected subcomponents. In this report, we present

a catalogue of the gene expression states of the early developing mouse total

kidney as well as the separated ureteric bud and metanephric mesenchyme

compartments.

The 20 Affymetrix U74A microarrays used for this study generated over

240,000 gene expression level data points, which were analyzed by multiple

crosswise comparisons as well as hierarchical cluster analysis. Hundreds of

genes expressed at much higher levels in the developing kidney than the adult

were found, while far fewer differences in gene expression patterns were found

between the multiple early embryonic kidney samples. A surprisingly high fraction

of the known genes to emerge from this study had been formerly implicated in

kidney development or processes related to branching morphogenesis. This

serves to authenticate the screen. In some cases, however, previously

characterized genes had not been reported expressed in the developing kidney.

For selected genes, we performed in situ hybridizations to more precisely define

spatiotemporal expression patterns.

The in situ hybridization results confirmed the microarray-predicted kidney

development expression. Interesting similarities in expression patterns in the

kidney and lung, which both undergo branching morphogenesis, were also

observed for three genes, Cdrap, Tgfbi, and Col15a1. It is interesting to note,

77 however, that for some of these genes, targeted mutant mice have been made and no kidney phenotype observed. This could reflect expression without kidney development function, functional redundancy, or the presence of a previously undetected kidney phenotype in the mutant mice. Even severely malformed kidneys can often provide sufficient function for survival. It would be interesting to begin to distinguish these possibilities by reexamination of these mutants.

Not all of the expected genes are actually identified in the comparisons.

For example, over 30 Hox genes are expressed in the developing kidney, but most do not appear on the list of transcription factors expressed at elevated levels during development. There are several possible reasons for these absences. Some genes are simply not on the U74Av2 microarrays. Other genes are expressed at low, near noise, levels and are deliberately excluded to avoid artifacts. Additional genes are expressed at similar levels in both samples being compared, and are therefore not called different. Furthermore, the high stringency of the screening process eliminated some genes.

This work provides an extensive definition of gene expression states during mouse kidney development. The results suggest kidney development function for a large number of genes not previously implicated in this process.

The complete data set is available (http://genet.chmcc.org Login, Nephrome with password “reviewer”). Since this work was performed with the standard commercial Affymetrix microarray platform, the resulting gene expression profiles can provide baselines for multiple future studies of kidney development in mouse mutants.

78 Acknowledgements

We thank Heather Hartman for excellent technical assistance. We thank

Cathy Ebert for providing whole mouse P1 total RNA. This work was supported by NIH grants DK61916-01 (S.S.P.) and DK02702 (L.T.P.).

79 References

1. Saxen L, Lehtonen E: Embryonic kidney in organ culture. Differentiation

36:2–11, 1987

2. Moore MW, Klein RD, Farinas I, et al: Renal and neuronal abnormalities in mice lacking GDNF. Nature 382:76–79, 1996

3. Schuchardt A, D’Agati V, Larsson-Blomberg L, et al: Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret.

Nature 367:380–383, 1994

4. Sakurai H, Bush KT, Nigam SK: Identification of pleiotrophin as a mesenchymal factor involved in ureteric bud branching morphogenesis.

Development 128:3283–3293, 2001

5. Itaranta P, Lin Y, Perasaari J, et al: Wnt-6 is expressed in the ureter bud

and induces kidney tubule development in vitro. Genesis 32:259–268, 2002

6. Vilar J, Lalou C, Duong VH, et al: Midkine is involved in kidney

development and in its regulation by retinoids. J Am Soc Nephrol 13:668–676,

2002

7. Tang MJ, Cai Y, Tsai SJ, et al: Ureteric bud outgrowth in response to RET

activation is mediated by phosphatidylinositol 3-kinase. Dev Biol 243:128–136,

2002

8. Clark AT, Young RJ, Bertram JF: In vitro studies on the roles of

transforming growth factor-beta 1 in rat metanephric development. Kidney Int

59:1641–1653, 2001

80 9. Bullock SL, Johnson TM, Bao Q, et al: Galectin-3 modulates ureteric bud branching in organ culture of the developing mouse kidney. J Am Soc Nephrol

12:515–523, 2001

10. Raatikainen-Ahokas A, Hytonen M, Tenhunen A, et al: BMP-4 affects the differentiation of metanephric mesenchyme and reveals an early anterior- posterior axis of the embryonic kidney. Dev Dyn 217:146–158, 2000

11. Tufro A, Norwood VF, Carey RM, Gomez RA: Vascular endothelial growth factor induces nephrogenesis and vasculogenesis. J Am Soc Nephrol 10:2125–

2134, 1999

12. Santos OF, Barros EJ, Yang XM, et al: Involvement of hepatocyte growth factor in kidney development. Dev Biol 163:525–529, 1994

13. Shawlot W, Behringer RR: Requirement for Lim1 in head-organizer function. Nature 374:425–430, 1995

14. Davis AP, Witte DP, Hsieh-Li HM, et al: Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11. Nature 375:791–795, 1995

15. Patterson LT, Pembaur M, Potter SS: Hoxa11 and hoxd11 regulate branching morphogenesis of the ureteric bud in the developing kidney.

Development 128:2153–2161, 2001

16. Wellik DM, Hawkes PJ, Capecchi MR: Hox11 paralogous genes are essential for metanephric kidney induction. Genes Dev 16:1423–1432, 2002

17. Xu PX, Adams J, Peters H, et al: Eya1-deficient mice lack ears and kidneys and show abnormal apoptosis of organ primordia. Nat Genet 23:113–

117, 1999

81 18. Kreidberg JA, Sariola H, Loring JM, et al: WT-1 is required for early kidney development. Cell 74:679–691, 1993

19. Nishinakamura R, Matsumoto Y, Nakao K, et al: Murine homolog of

SALL1 is essential for ureteric bud invasion in kidney development. Development

128:3105–3115, 2001

20. Kreidberg JA, Donovan MJ, Goldstein SL, et al: Alpha 3 beta 1 integrin has a crucial role in kidney and lung organogenesis. Development 122:3537–

3547, 1996

21. Muller U, Wang D, Denda S, et al: Integrin alpha 8 beta 1 is critically important for epithelial-mesenchymal interactions during kidney morphogenesis.

Cell 88:603–613, 1997

22. Miyamoto N, Yoshida M, Kuratani S, et al: Defects of urogenital development in mice lacking Emx2. Development 124:1653–1664, 1997

23. Hatini V, Huh SO, Herzlinger D, et al: Essential role of stromal mesenchyme in kidney morphogenesis revealed by targeted disruption of winged helix transcription factor BF-2. Genes Dev 10:1467– 1478, 1996

24. Mendelsohn C, Batourina E, Fung S, et al: Stromal cells mediate retinoid- dependent functions essential for renal development. Development 126:1139–

1148, 1999

25. Dudley AT, Lyons KM, Robertson EJ: A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye.

Genes Dev 9:2795–2807, 1995

82 26. Luo G, Hofmann C, Bronckers AL, et al: BMP-7 is an inducer of nephrogenesis, and is also required for eye development and skeletal patterning.

Genes Dev 9:2808–2820, 1995

27. Stark K, Vainio S, Vassileva G, McMahon AP: Epithelial transformation of metanephric mesenchyme in the developing kidneyregulated by Wnt-4. Nature

372:679–683, 1994

28. Mah SP, Saueressig H, Goulding M, et al: Kidney development in cadherin-6 mutants: delayed mesenchyme-to-epithelial conversion and loss of nephrons. Dev Biol 223:38–53, 2000

29. Bates CM, Kharzai S, Erwin T, et al: Role of N-myc in thedeveloping mouse kidney. Dev Biol 222:317–325, 2000

30. Barasch J, Yang J, Ware CB, et al: Mesenchymal to epithelial conversion in rat metanephros is induced by LIF. Cell 99:377–386,1999

31. Valerius MT, Patterson LT, Witte DP, Potter SS: Microarrayanalysis of novel cell lines representing two stages of metanephric mesenchyme differentiation. Mech Dev 112:219–232, 2002

32. Stuart RO, Bush KT, Nigam SK: Changes in global gene expression patterns during development and maturation of the rat kidney. Proc Natl Acad

Sci USA 98:5649–5654, 2001

33. Baugh LR, Hill AA, Brown EL, Hunter CP: Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic AcidsRes 29:E2938, 2001

34. Sewalt RG, van der Vlag J, Gunster MJ, et al: Characterizationof interactions between the mammalian polycomb-group proteinsEnx1/EZH2 and

83 EED suggests the existence of different mammalian polycomb-group protein complexes. Mol Cell Biol 18:3586–3595, 1998

35. Cao R, Wang L, Wang H, et al: Role of histone H3 lysine 27methylation in polycomb-group silencing. Science 298:1039–1043, 2002

36. Plisov SY, Ivanov SV, Yoshino K, et al: Mesenchymal-epithelial transition in the developing metanephric kidney: Gene expression study by differential display. Genesis J Genet Dev 27:22–31, 2000

37. Kume T, Deng K, Hogan BL: Murine forkhead/winged helix genes Foxc1

(Mf1) and Foxc2 (Mfh1) are required for the early organogenesis of the kidney and urinary tract. Development 127:1387–1395, 2000

38. Stamataki D, Kastrinaki M, Mankoo BS, et al: Homeodomain proteins

Mox1 and Mox2 associate with Pax1 and Pax3 transcription factors. FEBS Lett

499:274–278, 2001

39. Quinn LM, Latham SE, Kalionis B: The homeobox genes MSX2 and

MOX2 are candidates for regulating epithelial-mesenchymalcell interactions in the human placenta. Placenta 21 (Suppl A):S50–S54, 2000

40. Cano-Gauci DF, Song HH, Yang H, et al: Glypican-3-deficient mice exhibit developmental overgrowth and some of the abnormalities typical of Simpson-

Golabi-Behmel syndrome. J Cell Biol 146:255–264, 1999

41. Grisaru S, Rosenblum ND: Glypicans and the biology of renal malformations. Pediatr Nephrol 16:302–306, 2001

84 42. Grisaru S, Cano-Gauci D, Tee J, et al: Glypican-3 modulates BMP-and

FGF-mediated effects during renal branching morphogenesis. Dev Biol 231:31–

46, 2001

43. Leimeister C, Steidl C, Schumacher N, et al: Developmental expression and biochemical characterization of Emu family members. Dev Biol 249:204–

218, 2002

44. Pavlova A, Stuart RO, Pohl M, et al: Evolution of gene expression patterns in a model of branching morphogenesis. Am J Physiol 277:F650–F663, 1999

45. Blesch A, Bosserhoff AK, Apfel R, et al: Cloning of a novelmalignant melanoma-derived growth-regulatory protein, MIA. Cancer Res 54:5695–5701,

1994

46. Dietz UH, Sandell LJ: Cloning of a retinoic acid-sensitive mRNA expressed in cartilage and during chondrogenesis. J Biol Chem 271:3311–3316,

1996

47. Thomson RB, Igarashi P, Biemesderfer D, et al: Isolation and cDNA cloning of Ksp-cadherin, a novel kidney-specific member of the cadherin multigene family. J Biol Chem 270:17594–17601,1995

48. Koschorz B, Wertz K, McLaughlin KJ, Kispert A: Kidney-specific cadherin

(cdh16) is expressed in embryonic kidney, lung,and sex ducts. Nature 402:141–

146, 1999

49. Lin Y, Olsson M, Ekblom P, et al: Differential expression ofmouse laminin

gamma2 and gamma2* chain transcripts. Cell Tissue Res 300:129–137, 2000

85 50. Bush KT, Pohl ML, Quaranta V, et al: Involvement of laminin binding integrins and laminin-5 in branching morphogenesis of the ureteric bud during kidney development. Dev Biol 238:289–302, 2001

51. Rogers SA, Ryan G, Hammerman MR: Metanephric transforming growth factor-alpha is required for renal organogenesis in vitro. Am J Physiol 262:F533–

F539, 1992

52. Bernardini N, Mattii L, Bianchi F, et al: TGF-alpha mRNA expression in renal organogenesis: A study in rat and human embryos. Exp Nephrol 9:90–98,

2001

53. Enomoto H, Araki T, Jackman A, et al: GFR alpha 1-deficient mice have

deficits in the enteric nervous system and kidneys. Neuron 21:317–324, 1998

54. Celli G, LaRochelle WJ, Mackem S, et al: Soluble dominant-negative

receptor uncovers essential roles for fibroblast growth factors in multi-organ

induction and patterning. EMBO J 17:1642–1655, 1998

55. Nigam SK, Rodriguez-Boulan E, Silver RB: Changes in intracellular

calcium during the development of epithelial polarity and junctions. Proc Natl

Acad Sci USA 89:6162–6166, 1992

56. Spilker C, Gundelfinger ED, Braunewell KH: Evidence for different

functional properties of the neuronal calcium sensor proteins VILIP-1 and VILIP-

3: from subcellular localization to cellular function. Biochim Biophys Acta

1600:118–127, 2002

57. Pondel M: Calcitonin and calcitonin receptors: bone and beyond. Int J Exp

Pathol 81:405–422, 2000

86 58. Liu L, Dunn ST, Christakos S, et al: Calbindin-D28k gene expression in the developing mouse kidney. Kidney Int 44:322–330, 1993

59. Lazzaro D, De Simone V, De Magistris L, et al: LFB1 and LFB3 homeoproteins are sequentially expressed during kidney development.

Development 114:469–479, 1992

60. Coffinier C, Gresh L, Fiette L, et al: Bile system morphogenesis defects and liver dysfunction upon targeted deletion of HNF1beta. Development

129:1829–1838, 2002

61. Haase G, Dessaud E, Garces A, et al: GDNF acts through PEA3 to regulate cell body positioning and muscle innervation of specific motor neuron pools. Neuron 35:893–905, 2002

62. Araki I, Nakamura H: Engrailed defines the position of dorsal di- mesencephalic boundary by repressing diencephalic fate. Development

126:5127–5135, 1999

63. Zamir I, Dawson J, Lavinsky RM, et al: Cloning and characterization of a corepressor and potential component of the nuclear hormone receptor repression complex. Proc Natl Acad Sci USA 94:14400–14405, 1997

64. Dumont DJ, Gradwohl G, Fong GH, et al: Dominant-negative and targeted null mutations in the endothelial receptor tyrosine kinase, tek, reveal a critical role in vasculogenesis of the embryo. Genes Dev 8:1897–1909, 1994

65. Puri MC, Rossant J, Alitalo K, et al: The receptor tyrosine kinase TIE is required for integrity and survival of vascular endothelial cells. EMBO J 14:5884–

5891, 1995

87 66. Wulff P, Vallon V, Huang DY, et al: Impaired renal sodium retention in the sgk1-knockout mouse. J Clin Invest 110:1263–1268,2002

67. Haigh J, McVeigh J, Greer P: The fps/fes tyrosine kinase is expressed in myeloid, vascular endothelial, epithelial, and neuronal cells and is localized in the

trans-golgi network. Cell Growth Differ 7:931–944, 1996

68. Greer P, Haigh J, Mbamalu G, et al: The Fps/Fes protein-tyrosine kinase

promotes angiogenesis in transgenic mice. Mol Cell Biol 14:6755–6763, 1994

69. Hackenmiller R, Simon MC: Truncation of c-fes via gene targeting results

in embryonic lethality and hyperproliferation of hematopoietic cells. Dev Biol

245:255–269, 2002

70. Zeng Q, Subramaniam VN, Wong SH, et al: A novel synaptobrevin/VAMP homologous protein (VAMP5) is increased during in vitro myogenesis and present in the plasma membrane. Mol Biol Cell 9:2423–2437, 1998

71. Yu J, Carroll TJ, McMahon AP: Sonic hedgehog regulates proliferation and differentiation of mesenchymal cells in the mouse metanephric kidney.

Development 129:5301–5312, 2002

72. Zhou C, Tsai SY, Tsai M: From apoptosis to angiogenesis: New insights into the roles of nuclear orphan receptors, chicken ovalbumin upstream promoter-transcription factors, during development. Biochim Biophys Acta

1470:M63–M68, 2000

73. Pereira FA, Qiu Y, Tsai MJ, Tsai SY: Chicken ovalbumin upstream promoter transcription factor (COUP-TF): Expression during mouse embryogenesis. J Steroid Biochem Mol Biol 53:503–508, 1995

88 74. Qiu Y, Pereira FA, DeMayo FJ, et al: Null mutation of mCOUP-TFI results in defects in morphogenesis of the glossopharyngeal ganglion, axonal projection, and arborization. Genes Dev 11:1925–1937, 1997

75. Matena K, Boehm T, Dear N: Genomic organization of mouseCapn5 and

Capn6 genes confirms that they are a distinct calpain subfamily. Genomics

48:117–120, 1998

76. Sato K, Kawashima S: Calpain function in the modulation of signal transduction molecules. Biol Chem 382:743–751, 2001

77. Furukawa K, Inagaki H, Naruge T, et al: cDNA cloning and functional characterization of a meiosis-specific protein (MNS1) with apparent nuclear association. Chromosome Res 2:99–113, 1994

78. Echard A, Jollivet F, Martinez O, et al: Interaction of a Golgi-associated kinesin-like protein with Rab6. Science 279:580–585, 1998

79. Xie WF, Kondo S, Sandell LJ: Regulation of the mouse cartilage-derived retinoic acid-sensitive protein gene by the transcription factor AP-2. J Biol Chem

273:5026–5032, 1998

80. Moser M, Bosserhoff AK, Hunziker EB, et al: Ultrastructural cartilage abnormalities in MIA/CD-RAP-deficient mice. Mol Cell Biol 22:1438–1445, 2002

81. Skonier J, Neubauer M, Madisen L, et al: cDNA cloning and sequence analysis of beta ig-h3, a novel gene induced in a human adenocarcinoma cell line after treatment with transforming growth factor-beta. DNA Cell Biol 11:511–

522, 1992

89 82. Mashima Y, Nakamura Y, Noda K, et al: A novel mutation at codon 124

(R124L) in the BIGH3 gene is associated with a superficial variant of granular corneal dystrophy. Arch Ophthalmol 117:90–93, 1999

83. Schorderet DF, Menasche M, Morand S, et al: Genomic characterization and embryonic expression of the mouse Bigh3 (Tgfbi) gene. Biochem Biophys

Res Commun 274:267–274, 2000

84. Kivirikko S, Mauviel A, Pihlajaniemi T, Uitto J: Cytokine modulation of type

XV collagen gene expression in human dermal fibroblast cultures. Exp Dermatol

8:407–412, 1999

85. Eklund L, Piuhola J, Komulainen J, et al: Lack of type XV collagen causes a skeletal myopathy and cardiovascular defects in mice. Proc Natl Acad Sci USA

98:1194–1199, 2001

86. Muona A, Eklund L, Vaisanen T, Pihlajaniemi T: Developmentally regulated expression of type XV collagen correlates with abnormalities in

Col15a1(-/-) mice. Matrix Biol 21:89–102, 2002

90 Figures and tables

91 Table 1 continued.

92

93

94 Table 3 continued

95

Figure1. Scattergram comparisons of microarray hybridizations. (A)

Technical replicate. Two 30 ng aliquots from a single adult kidney RNA sample were processed in parallel and hybridized to two U74Av2 microarrays. The resulting scattergram shows less than 1% of genes with more than a two-fold difference in hybridization signal. (B) Biologic replicate. RNA samples from two separately dissected E12.5 kidneys were processed in parallel. The resulting scattergram shows more variation than for the technical replicate, but again with well below 1% of genes giving over a twofold difference in hybridization signal.

96 (C) Comparison of E12.5 and adult kidney microarray hybridization patterns. A high level of biologic variation is observed. (D) Comparison of E11.5 ureteric bud

Figure 1 continued. and metanephric mesenchyme. Scattergrams show comparisons of hybridization signals. Only genes with sufficient signal to be called present by MAS5 for at least one of the two samples being compared are shown. The scattergrams show comparisons of raw hybridization signals of each probe set. For (A), for example, this results in 101 probe sets called over two-fold different in expression level on the two microarrays. The MAS5 comparison, however, performs a more sophisticated analysis of the probe set hybridization data and identified only 43 genes with over a twofold difference. For both methods twofold outliers are below

1% of total.

97

Figure 2. Hierarchical cluster analysis of genes expressed in the early developing and adult kidney. (A) Cluster analysis used several analysis of variance (ANOVA)-based statistical criteria, which identified genes strongly expressed in developing or adult kidney. Genes with similar expression profiles were clustered using hierarchical tree algorithm applied to the log relative gene expression values using Pearson correlation. Rows represent the different tissues examined. Abbreviations are: MM, metanephric mesenchyme; UB, ureteric bud; W, whole mouse; all others, total kidney. Red represents high expression, blue represents low expression, and yellow is intermediate,

98 normalized to P1 whole mouse. A more complete set of comparisons is available at http://genet.chmcc.org (login is Nephrome, password is reviewer, click login, not login as guest, click continue on U74Av2_mouse).

99

Figure 3. In situ hybridization analysis of gene expression patterns in the developing kidney. (A to C) At E12.5, Smoh expression was found in the metanephric mesenchyme with the highest expression localized to the

100 Figure 3 continued.

prestromal mesenchyme (arrows). Also, notice the absence of signal within the

branching ureteric bud (arrowhead). (D to F) Nr2f1 expression was diffuse

throughout the metanephric mesenchyme in the E13.5 developing kidney, while

expression was not found within the branching ureteric bud (arrowheads). (G to I)

Later in the developing kidney, at E15.5, Nr2f1 expression was maintained in the

mesenchyme, and was found in the smaller tubular structures (arrows), but was

absent in the mature ureteric stalks (arrowheads). (J to L) The Birc5 gene was

highly expressed in the tubular structures of the E17.5 kidney (arrowheads) and

in developing glomeruli (arrows). (M to O) Capn5 expression in the E15.5 kidney.

High expression was restricted to epithelia of the large diameter ureteric stalks

(arrows), while decreased expression is identified in the smaller ureteric tips

(arrowheads) located at the periphery of the developing kidney. (P to R) Mns1

expression in the E15.5 kidney was mainly localized to the early nephrogenic

epithelia (arrows) within the cortex of the developing kidney and an S-shaped

body (arrowhead). (S to U) Rab6kifl expression in the developing urogenital

system at E12.5 was localized to mesenchyme surrounding the ureter (arrow)

while absent from the mesenchyme surrounding the ureteric bud (arrowheads).

(M to O) taken at 100X; all other images taken at 200X. Serial sagittal sections.

(A, D, G, J, M, P, and S) show DAPI fluorescence. Dark-field imaging was used

to view the hybridization signal (B, E, H, K, N, Q, and T). (C, F, I, L, O, R, and U)

show overlays of dark-field image over the DAPI fluorescent image.

101

Figure 4. Similar expression patterns within the developing lung and

kidney suggesting roles in branching morphogenesis. (A to C) Cdrap showed high expression in the E17.5 kidney localized to the distal ends of the ureteric bud (arrowheads) at the periphery of the kidney. (D to F) A similar

102 Figure 4 continued. expression pattern for Cdrap was seen in the E15.5 lung with expression localized at the distal tips of the branching distal bronchi (arrows), while the proximal bronchi (arrowhead) lack Cdrap expression. (G to I) Tgfbi expression in the metanephric mesenchyme at E12.5 was diffuse with no detectable signal in the ureteric bud (arrowheads). There was high Tgfbi expression in the mesenchyme surrounding the Wolffian (mesonephric) duct (arrow). (J to L) The mesenchymal-specific expression pattern of Tgfbi was maintained in the E15.5

kidney within the nephrogenic mesenchyme surrounding the ureteric-derived

duct (arrowheads), but decreased or absent expression was identified in

mesenchymal-derived structures, including an S-shaped body (arrow) and

immature glomeruli (asterisks). (M to O) A similar expression pattern was found

in the lung at E15.5, with signal localized to the mesenchyme surrounding the

bronchi (arrows). (P to R) Early in kidney development, at E13.5, Col15a1 was

highly expressed within patches of mesenchyme surrounding the ureteric

epithelia (arrowheads). (S to U) Later, in the E15.5 metanephric kidney, Col15a1

was highly expressed within the mesenchyme surrounding the ureteric bud-

derived ducts (arrows), but absent in S-shaped bodies (arrowheads). (V to X) A

similar expression pattern for Col15a1 was found in the E13.5 lung, with high signal in patches of mesenchyme surrounding the bronchi (arrows). (Y to A’)

Later in development, the E15.5 lung showed a diffuse mesenchymal Col15a1 expression pattern surrounding the proximal bronchi (arrows) and distal branching bronchi (arrowhead). Sections were stained with DAPI (A, D, G, J, M,

103 Figure 4 continued.

P, S, V, and Y). Dark-field imaging was used to view the signal (B, E, H, K, N, Q,

T, W, and Z). (C, F, I, L, O, R, U, X, and A’) show overlays of dark-field image over the fluorescent image. All images were taken at 200X.

104 Chapter 3

Comprehensive microarray analysis of Hoxa11/Hoxd11 mutant kidney

development *

Kristopher Schwab A, Heather A. Hartman B, Hung-Chi Liang A,

Bruce J. Aronow C, Larry T. Patterson B, S. Steven Potter A

Division of Developmental Biology A, Nephrology B, and Bioinformatics C,

Children's Hospital Medical Center

3333 Burnet Avenue

Cincinnati, OH 45229

* Published in Dev Biol. 2006 May 15;293(2):540-54.

105 Abstract

The Hox11 paralogous genes play critical roles in kidney development.

They are expressed in the early metanephric mesenchyme and are required for

the induction of ureteric bud formation and its subsequent branching

morphogenesis. They are also required for the normal nephrogenesis response

of the metanephric mesenchyme to inductive signals from the ureteric bud. In this

report, we use microarrays to perform a comprehensive gene expression

analysis of the Hoxa11/Hoxd11 mutant kidney phenotype. We examined E11.5,

E12.5, E13.5 and E16.5 developmental time points. A novel high throughput

strategy for validation of microarray data is described, using additional biological

replicates and an independent microarray platform. The results identified 13

genes with greater than 3-fold change in expression in early mutant kidneys,

including Hoxa11s, GATA6, TGFbeta2, chemokine ligand 12, angiotensin

receptor like 1, cytochrome P450, cadherin5, Lymphocyte antigen 6 complex,

Iroquois 3, EST A930038C07Rik, Meox2, Prkcn, and Slc40a1. Of interest, many

of these genes, and others showing lower fold expression changes, have been

connected to processes that make sense in terms of the mutant phenotype,

including TGFbeta signaling, iron transport, protein kinase C function, growth

arrest and GDNF regulation. These results identify the multiple molecular

pathways downstream of Hox11 function in the developing kidney.

106 Introduction

The kidney is an excellent model system for studying the principles of organogenesis. The developing kidney exhibits many interesting processes, including establishment of an early metanephric field by tissue interactions, budding, reciprocal inductive interactions between ureteric bud (UB) and metanephric mesenchyme (MM), stem cell growth and differentiation, and conversion of mesenchyme into epithelia. In addition, there is branching morphogenesis, apoptosis, fusion (nephrons to collecting ducts), and proximal– distal segmentation along the length of the nephron [for review, see (Davies and

Bard, 1998)]. Metanephric, or adult, kidney development begins in the mouse at around embryonic day 11.0 (E11.0) as the UB grows from the nephric duct and invades the MM. The UB induces the cells of the MM to condense, epithelialize,

and form renal vesicles, each of which develops into a functional nephron

containing a glomerulus, proximal tubule, loop of Henle and distal tubule. Of

particular importance, kidney morphogenesis can be readily studied in organ

culture (Saxen and Lehtonen, 1987).

We have made significant advances in understanding the genetic

regulation of kidney organogenesis (Bouchard, 2004; Yu et al., 2004). Hox genes

play important roles in this process. These transcription factor-encoding genes

often occupy high level positions in the genetic hierarchy of development. It is

interesting that a total of 27 Hox genes show specific domains of expression in

the developing kidney (Patterson and Potter, 2004). Mutations in the Hoxa11,

Hoxc11 and Hoxd11 closely related paralogous genes reveal that they have

107 redundant functions in several aspects ofkidney development (Davis et al., 1995;

Patterson et al., 2001; Wellik et al., 2002). These three genes are expressed in the early MM, and the combined mutation of all three gives a loss of GDNF synthesis and failure of the UB to form (Wellik et al., 2002). A hypomorphic mutant allele combination, with mutations in Hoxa11 and Hoxd11 but not

Hoxc11, results in a MM that does not properly drive branching morphogenesis of the UB, and in turn does not respond correctly to UB signals (Patterson et al.,

2001). These results indicate that the Hox11 paralogs play crucial roles in several stages of kidney development. There are still important gaps, however, in our understanding of the downstream genetic pathways regulated by these Hox genes.

The microarray is a useful tool for gaining deeper insight into the genetic program of kidney development. Microarrays can provide an important gene discovery function, identifying all genes expressed in the developing kidney and cataloging changes that occur over time. In addition, they allow an impartial global view of altered gene expression profiles in mutant developing kidneys.

Instead of looking at just a few selected marker genes by in situ hybridization, it is now possible to conduct an unbiased and universal analysis of gene expression patterns in mutants.

In this paper, we extend the previous microarray studies of normal kidney development, and then use this wild type baseline to analyze the altered gene expression patterns of the Hoxa11/Hoxd11 mutant kidney. We used Affymetrix

MOE430 oligonucleotide microarrays to examine gene expression profiles of the

108 complete normal kidney at E12.5, E13.5, E16.5 and adult. In addition, we determined the gene expression patterns of the E11.5 MM and UB, using both laser capture microdissection and manual microdissection to isolate tissues, thereby identifying over 1500 genes with strong differential expression. These results serve to identify the gene expression networks and signaling pathways active in these kidney primordia. Finally, we performed an extensive microarray dissection of the altered gene expression patterns present in Hoxa11/ Hoxd11 double mutant kidneys. Several developmental time points were examined, including E11.5, E12.5, E13.5 and E16.5. To allow a more robust microarray analysis of the mutant differences we combined independent data from the

Affymetrix and Illumina microarray platforms. The results identify a battery of downstream genes that provide deeper insight into the molecular mechanisms of

Hox11 function in kidney development.

109 Materials and methods

Breeding and genotyping Hoxa11/Hoxd11 mutant mice

Hoxa11 and Hoxd11 mutant mice were previously described (Davis et al.,

1995; Patterson et al., 2001; Small and Potter, 1993). The colony was maintained on a mixed genetic background of four strains of mice (129, C57,

+/− +/− C3H and CF1). Hoxa 11 , Hoxd11 double heterozygous female mice have

uterine defects that severely limit reproductive capacity (Hsieh-Li et al., 1995).

We therefore isolated zygotes from double heterozygote crosses, with

superovulated females, and transferred them to pseudo-pregnant surrogate wild

type CD-1 females (Nagy, 2003). Noon of the day when the vaginal plug was

observed was considered embryonic day 0.5 (E0.5). All mice and embryos were

genotyped as previously described (Patterson et al., 2001). This study focuses

on Hoxa11, Hoxd11 double homozygous mutant mice.

Tissue dissections

Wild type tissues from E12.5 and older were obtained from outbred CD-1

mice. Whole embryonic kidneys and urogenital ridges were dissected in ice-cold

PBS then either frozen at −80°C, or quick-frozen in Tissue-Tek® OCT compound

(Sakura, Torrence, CA) using liquid nitrogen cooled 2methylbutane.

MM and UBs, up to T-shaped stage, were isolated by treatingdissected E11.5

kidneys with 0.5 mg/ml collagenase B (Roche, Indianapolis, IN), in D-MEM

(Invitrogen, Carlsbad, CA) for 30 min at 37°C and carefully dissecting the

mesenchyme from the UB, followed by storage at −80°C.

Laser-capture microdissection

110 E11.5 whole embryos were frozen in OCT. For later time points the kidneys were removed and frozen in OCT. Serial sections (7µ) were made using a Microm HM 550 cryostat (Richard-Allan Scientific, Kalamazoo, MI), collected on Fisher Superfrost plus precleaned slides (Hampton, NH), and stored at

−80°C. Alternate sections were hematoxylin and eosin stained and used to help identify the UB and MM. For LCM, the remaining sections were air dried at room temperature for 3min, acetone fixed for 2min, rinsed in ice cold 1/10 PBS for 3 min and then dehydrated in 75%, 95%, 100%, 100% ethanol, followed by two 5- min rinses with xylene. Laser capture microdissection was performed using the

Arcturus Pixcell II system, according to Arcturus protocols (Mountain View,

California).

RNA isolation and target RNA amplification

Total RNA from wild type and mutant whole kidneys was prepared using the Stratagene Absolutely RNA Nanoprep Kit (La Jolla, CA) and amplified as previously described (Schwab et al., 2003). Target RNA was then hybridized to both the MOE430A and MOE430B Genechips (Affymetrix, Santa Clara, CA).

Microarray analysis of each stage was performed in biological duplicate using either 30 ng or 100 ng of starting total RNA. Each microarray hybridization represented a biological replicate, using an independent biological sample. We pooled 3–9 wild type UB for each sample, and 2–4 MM for each sample. Mutant

E11.5 MM was not pooled. LCM RNA was prepared using the RNeasy Micro Kit

(Qiagen, Valencia, CA), with 30 ng poly-inosine carrier (Epicentre, Madison, WI) added to the RLT buffer. Target RNA was prepared using the TargetAmp™ 2-

111 Round aRNA Amplification Kit 1.0 (Epicentre, Madison, WI), and hybridized to

Affymetrix MOE430_v2 microarrays. To validate the Affymetrix results, total RNA was isolated from HoxA11/ D11 null and normal E13.5 whole kidneys or E11.5

MM, amplified using the Epicentre TargetAmp™ 2-Round aRNA Amplification Kit

(Madison, WI) and hybridized to Sentrix MouseRef-8 Beadchip microarrays

(Illumina, San Diego, CA) containing over 24,000 probes.

Gene expression profile analysis

Affymetrix raw data in the CEL file format was normalized using RMA Express

0.2 (Bolstad et al., 2003)and analyzed using Genespring 7.0. Illumina raw signal data was imported into the Affymetrix MOE430 genome on basis of gene symbol for analysis. Wild type whole developing kidney samples were normalized to adult samples. Normal E11.5 MM samples were normalized to E11.5 UB.

HoxA11/D11 null samples were normalized to the corresponding wild type control. Hierarchical clusters were generated using the Pearson Correlation

Function. All microarray data are available from Signet (http:// cypher.cchmc.org:1104/servlet/GeNet, login as “Guest, select MOE430 genome, data contained in “SPotter/Schwab et al. 2005 folders”) allowing interactive analysis of the data, through other public databases (GEO, GUDMAP), and will be provided upon request.

QPCR validation of microarray changes

Total RNA was obtained and DNAse 1treated from separate E13.5

Hoxa11/ d11 (n = 4) and control kidneys (n =3) using Stratagene Absolutely RNA

Microprep Kit (La Jolla, CA). cDNA was generated using random hexamers

112 according to conventional protocols (Invitrogen, Carlsbad, Ca). The following primers were generated specifically to the sequence obtained from the Affymetrix probe set: Actb (TTGCTGACAGGATGCAGAAG,

ACATCTGCTGGAAGGTGGAC), Cxcl12 (GTCTAAGCAGCGATGGGTTC,

TAGGAAGCTGCCTTCTCCTG), HoxA11s (TGTCCTGGAGGAAGGAGAA,

ATCACCACCATTGGGAGGT), Pdgfrb (AGCAAGAGTGGCAGAGAAGG,

TAATCCCGTCAGCATCTTCC), Slit3 (CGTGGAAGAGGTGGAGAGAC,

AGAGGTTCCATGTGGCTGTT), and Tgfb2 (GAAATACGCCCAAGATCGAA,

TGTCACCGTGATTTTCGTGT) using Primer3 software (Rozen and Skaletsky,

2000). Relative quantitative PCR was performed according to the conventional

SYBR Green protocol (Stratagene) using the Stratagene Mx3000p QPCR system. Dissociation curve and agarose gel analysis of each primer set were used to insure specificity of the amplicon. All data were normalized to an internal

housekeeping control (Actb) and analyzed using the 2(-Delta Delta C(T)) method

(Livak and Schmittgen, 2001).

113 Results

Microarray comparisons of gene expression profiles of wild type E12.5,

E13.5, E16.5 and adult kidneys

We used Affymetrix MOE430 microarrays to examine gene expression patterns of the normal developing kidney at E12.5, E13.5, E16.5 and adult.

These oligonucleotide arrays carry over 48,000 probe sets and monitor expression levels of over 20,000 genes. A high stringency analysis of the data identified 2793 genes that showed a strong change in expression level as a function of developmental time. The heat map in Fig. 1 illustrates the patterns observed. The gene expression profiles of E12.5 and E13.5 kidneys were very similar, while at E16.5 a significant block of genes showed altered expression.

Expression levels were normalized to that of the adult kidney, shown in the right two lanes. All data and gene lists are available from the Signet server (see

Methods).

Gene expression profiles of the E11.5 MM and UB

The MM and UB represent the embryonic precursors of the adult kidney.

At E11.5, the UB has invaded the MM. A series of reciprocal inductive interactions drive the mesenchyme to convert to epithelia and form nephrons, while the bud undergoes branching morphogenesis and forms the collecting ducts.

To better understand the distinct properties of the UB and MM, we defined their gene expression profiles. The UB is a discrete structure and was cleanly purified by manual microdissection following enzymatic treatment to dissociate

114 the surrounding mesenchyme. The MM, with a more poorly defined outer border at this stage of development, was isolated by both manual microdissection and by laser capture microdissection. To provide a UB-MM comparison with great statistical power, we examined a total of four UB and seven MM biological replicates.

The resulting microarray data showed a high degree of reproducibility. The scattergraph in Fig. 2A compares the gene expression patterns of two MM samples purified by laser capture microdissection. If each gene showed exactly the same expression in both samples there would be a single line at 45 degrees.

There is some scatter, with about 1% (526) of the >48,000 probe sets giving more than a 3-fold difference (green lines) in expression level between the two samples. This is likely due to a combination of biological variation, sampling error during LCM, and noise resulting from the two round target amplification procedure required because of the small amount of starting RNA. This noise is eliminated in the analysis by examining multiple biological replicates and requiring consistent change. Fig. 2B compares MM isolated by LCM versus mesenchyme isolated by manual microdissection, showing a similar level of scatter to that seen in Fig. 2A. Interestingly, there was significantly less scatter in pair-wise comparisons of UB samples (Fig. 2C), with less than 0.1% of probe sets (56) showing over a 3-fold difference in expression level. Most striking,

however, was the large number of genes with dramatic differences in expression

when comparing the UB and MM (Fig. 2D).

115 A stringent analysis of the UB and MM microarray data was performed.

Using rigorous criteria requiring an average fold change of at least three, with

Benjamini and Hochberg false discovery rate multiple testing correction, and a ttest of P<0.05, we identified 1518 differently expressed genes. This is illustrated in the heat map of Fig. 3. These numbers are higher than previously reported

(Schwab et al., 2003) because of improved purification protocols, more comprehensive microarrays, and the use of a considerably larger number of microarrays (11vs. 4), lending greater statistical power to the analysis.

Examination of the list of differently expressed genes identifies many interesting genes not previously implicated in early kidney development. In addition, 63 genes that have been previously shown to exhibit restricted UB or MM expression are found on the list of 1518 genes. These genes are included in

Table 1 along with references documenting discrete expression. Of interest, in each case the microarray data is consistent with the previous expression study.

This universal agreement of the microarray data with previous in situ hybridization studies provides an important measure of validation. The UB-MM specific expression patterns of GDNF, Ret, Hoxa11, Hoxc10, Hoxd10, Hoxa10,

Wnt9b, Wnt4, Wnt11, Six2, Wt1, Foxc1, Lhx1 and 50 other genes were all correctly called by the microarrays. In addition, Table 1 lists other selected genes ranked according to fold difference in expression. These genes encode cytokines, receptors, and several categories of transcription factors and growth factors, of potential importance in programming the early functions of these two structures.

116 Analysis of Hoxa11/d11 mutant kidneys

This microarray atlas of gene expression patterns in the normal developing kidney provides a baseline that can be used for the global analysis of altered gene expression patterns in mutants. Previous studies have shown that the Hox11 paralogous genes play important, yet redundant roles in several phases of kidney development (Davis et al., 1995; Patterson et al., 2001; Wellik et al., 2002). To better understand the molecular level perturbations present in the mutant kidneys, we performed an exhaustive microarray analysis. We focused on the Hoxa11/Hoxd11 double homozygous mutant. These mice show a severe kidney phenotype, with reduced branching morphogenesis of the collecting duct system and altered gene expression patterns, as defined by in situ hybridization, during nephrogenesis (Patterson et al., 2001). Further removal of the Hoxc11 gene results in very early arrest of kidney development (Wellik et al.,

2002), precluding the study of later developmental functions of this gene group.

E12.5/E13.5 wild type-Hoxa11/d11 mutant comparison

Our first comparisons of wild type and Hoxa11/d11 mutant used Affymetrix

MOE430 microarrays and examined E12.5, E13.5 and E16.5 developmental time points, each in biological duplicate. As might be expected, the gene expression profiles of the wild type and Hoxa11/d11mutant kidneys were much more similar than seen for the wild type UB-MM comparison. Lowering the stringency of the analysis to find the lower fold gene expression changes present in the mutants resulted in many artifact difference calls, as shown by real time PCR validation

(data not shown).

117 To improve the discriminatory power of the study, we repeated the microarray analysis of the E13.5 wild type and mutant kidneys, this time using the Illumina beaded array platform. This provided additional microarray data, making the statistical comparison stronger, and it added more biological samples, as biological replicates were performed, further helping to remove biological noise. In addition, it gave a new microarray perspective, since an independent microarray system was used. Three normal and two mutant E13.5 kidneys were examined with Illumina. This gave a total of nine microarrays for the E13.5 comparison, with five Illumina and four Affymetrix. These data were analyzed with GeneSpring 7.0, setting the screen parameters at non-parametric test (Welch t test) P < 0.05, using Benjamini and Hochberg multiple testing correction, fold change >1.5, and a minimum raw expression of 400 in at least two samples to remove genes with very low expression levels. This resulted in a list of 467 genes (Supplementary Figure S1).

To further strengthen the analysis, we combined the data from the wild type-mutant comparisons at E12.5 and 13.5. These two time points showed similar wild type gene expression profiles (Fig. 1). A comparable stringency analysis of the E12.5 data was performed, using four Affymetrix MOE430 microarrays, and deleting the multiple testing correction, because of the smaller number of microarrays used. The resulting gene list was used to determine the set of overlapping genes, with different expression levels in both the E12.5 and

E13.5 wild type versus mutant comparisons. This gave a list of 122 probe sets, representing 107 different genes (Table 2), with relative expression levels in the

118 17 microarrays shown in the heat map of Fig. 4. These genes gave consistent expression level differences in both the E12.5 and E13.5 wild type versus mutant comparisons. In addition, many showed similar expression changes in the E16.5 wild type and mutant kidneys (Fig. 4).

It is interesting that the gene showing the greatest change in expression in the Hoxa11/d11 mutants was the Hoxa11 antisense transcript. We have previously shown that the Hoxa11 gene gives rise to both sense and antisense transcripts (Hsieh-Li et al., 1995). The targeting of the Hoxa11 gene created a deletion including the homeobox region (Small and Potter, 1993), which clearly reduced antisense transcription. The Affymetrix microarrays also identified a significant reduction of Hoxa11 sense transcripts in the mutants (−6.3-fold at

E12.5, −6.4 at E13.5 and −5.8 at E16.5). This difference, however, was not

detected by the Illumina microarrays, which reported very low expression of

Hoxa11 even in the wild type (raw signals of 70 to 160), so this gene did not

make the final list. This illustrates that the two microarray platforms show gene

specific detection differences. In general, we observed that both platforms

provided excellent concordant expression level analysis for most genes. But a

few genes were better assayed by one system, and other genes by the other. By

requiring both types of microarrays to observe a change it is clear that some

genuine expression changes, such as the Hoxa11 sense transcript, will be lost.

The Hoxd11 gene was targeted in a different manner, not by deletion but by the

insertion of a Neo selectable marker. Neither microarray system found a change

119 in expression level of this gene, suggesting that the insertion did not significantly alter transcript abundance.

It is also interesting to note that 15 of the genes called differently expressed are represented by duplicate probe sets on the Affymetrix microarrays. That is, the probe set list of 122 includes only 107 different genes.

The fold changes called by these independent probes sets in this combined data set show excellent agreement. These include, in alphabetical order, angiotensin receptor like-1 (fold changes of 2.5, 3.1), carbonic anhydrase 4 (−1.9, −2.1),

elastin microfibril interfacer 1 (2.2, 2.2), EST AI450540 (1.6, 2.2), Fibrillin 1 (2.5,

2.0), GATA6 (3.9, 3.0), high mobility group box 1 (−1.6, −1.8), insulin-like growth

factor binding protein 4 (2.0, 1.8), Lipoma HMGIC fusion partner-like 2 (1.7, 1.6),

mastermind like 1 (2.3, 1.7), pdgf receptor beta (2.6, 2.1), Procollagen, type V,

alpha 1 (2.5, 1.9), Procollagen, type XVIII, alpha 1 (2.3, 2.0), Procollagen, type

XXIII, alpha 1 (2.3, 2.0).

To further test the validity of the resulting gene list we performed real time

PCR for five genes. The results are shown in Fig. 5. For all five genes a

significant difference in gene expression was confirmed, in the direction predicted

by microarrays. The fold change agreement between microarray and real time

PCR was generally excellent (CXCL12, 3.3 vs. 2.8, TGFbeta2, 3.6 vs. 3.3,

pdgfrb, 2.1 vs. 1.5), although for two genes the real time PCR found a much

greater fold change than the microarrays (Slit3, 2.3 vs. 7.8, Hoxa11antisense, −5

vs. −34). These results suggest that by using a combined 17 microarrays,

including two different platforms and examining two closely related

120 developmental stages, it was possible to discern a list of genes with genuine expression differences in wild type and Hoxa11/d11 mutant kidneys.

The microarray data called eight genes with greater than 3fold change in expression in the mutants. These are quite interesting genes, (GATA6,

TGFbeta2, chemokine ligand 12, angiotensin receptor like 1, cytochrome P450, cadherin5, Hoxa11 antisense and Lymphocyte antigen 6 complex), encoding a transcription factor, a growth factor, a chemokine, a receptor and a cell adhesion molecule. The bulk of genes differently expressed, however, showed lower fold changes, in the 1.5–3 range. We have previously shown that certain regions of the mutant kidneys, in particular the two poles, often show relatively normal development (Patterson et al., 2001). The more normal gene expression patterns in these regions would dilute out gene expression differences present in the severely developmentally perturbed ventral–medial part of the mutant kidney. It is likely, therefore, that our results represent underestimates of the fold changes

present in the most abnormal regions.

Comparison of wild type and Hoxa11/d11 mutant E11.5 MM

The E11.5 time point is of particular interest in the analysis of the Hoxa11/d11

mutants. Both Hoxa11 and Hoxd11 are expressed in the E11.5 MM, and at this early stage the MM appears histologically normal. Nevertheless it is functionally deficient. In some mutants with severe phenotypes we have observed that the

UB forms, and penetrates the MM. However, instead of stopping and branching the UB simply grows right through the MM, exiting the other side (Patterson, unpublished observations). The Hoxa11/d11 genes clearly play an important role

121 in communication between the early UB and MM. To perform a global analysis of the altered gene expression pattern in the E11.5 MM we used a strategy similar to that described for the E12.5/E13.5 wild type-mutant comparison. To improve purity of starting material, we used laser capture microdissection to isolate MM from wild type and mutant E11.5 embryos. Following RNA purification and two round in vitro transcription target amplification, we used the Affymetrix MOE430 microarrays to determine gene expression levels. To achieve high throughput validation of the resulting gene list, we then repeated this process, using the

Illumina platform, looking for genes consistently called differently expressed in wild type and mutant E11.5 MM by both systems. Each microarray used an independent biological sample (biological replicate). Again, a total of 17 microarrays were used for the comparison, in this case with 12 Affymetrix and 5

Illumina microarrays. The data were analyzed with GeneSpring 7.0, using similar parameters to those described for the E12.5/E13.5 data, except with a more stringent expression threshold cutoff. In this case we required a minimum raw expression level of at least 600 in two samples, versus the cutoff of 400 used for the E12.5, E13.5 data. Genes with low expression levels give low signal to noise ratios and are a major source of artifacts. In both sets of comparisons, we eliminated the bulk of these false calls by setting stringent minimal expression requirements. This analysis provided a list of 146 genes showing greater than

1.5-fold change in expression in the mutant E11.5 MM. The consistency of these expression differences across the seventeen samples, including both microarray platforms, is illustrated in the series graph of Fig. 6. The gene list, with observed

122 fold changes, is provided in Table 3. We found six genes with greater than 3-fold change in expression. Iroquois 3 (Irx3), encoding a homeodomain transcription factor, was up-regulated 6-fold, and the EST A930038C07Rik was up-regulated

3.5-fold. Down-regulated genes included Meox2 (−3.4) also encoding a homeodomain protein, Hoxa11antisense (−4.4), Prkcn (−4.5) encoding a protein

kinase C, and Slc40a1 (−5.2, −6.1), encoding an iron transporter. In addition

there are a number of interesting genes with smaller fold expression changes in

the 1.5–3 range, some of which are discussed later. Together, these extensive

microarray studies of the altered gene expression patterns present in the

Hoxa11/d11 mutant kidneys provide deeper insight in the normal functions of the

Hox11 group of genes, as discussed further below.

123 Discussion

Microarrays provide a powerful technology for the analysis of both normal and mutant kidney development. The pioneering work of Stuart et al. (2001) used

Affymetrix microarrays to examine normal rat kidney development, identifying several important functional groupings and thousands of specific genes not previously associated with kidney development. This work was subsequently extended to the developing mouse kidney (Schwab et al., 2003). Several recent reports have taken the microarray analysis of kidney development a step further, with Challen et al. (2004) using cDNA arrays to define the genes expressed specifically in the E10.5 uninduced MM, compared to more rostral uninduced intermediate mesoderm, in order to better define the nature ofthe early renal stem cell. Challen et al. (2005) also conducted an extensive cDNA microarray analysis of the gene expression profiles of total developing kidneys at multiple stages, as well as GFP sorted UB at E15.5. In another important study Schmidt-

Ott et al. (2005) manually micro-dissected UB tips and mesenchyme from both mouse (E12.5) and rat (E13.5) and used Affymetrix microarrays to find differentially expressed genes. Of particular interest they identified the cytokine

Clf-1 as a novel regulator secreted by the UB. In this report we extend this growing microarray database of normal kidney development by using the

Affymetrix MOE430 generation microarray to determine gene expression profiles of total kidneys at multiple stages of development.

We also performed an extensive analysis of E11.5 MM and UB gene expression patterns, including laser capture microdissection to insure tissue purity, and using

124 a total of eleven microarrays to provide great statistical power in the analysis.

This resulted in a comprehensive definition of the gene expression states of

these two critical early kidney components. We identified 1518 genes with over

3-fold divergent expression, including growth factors, cytokines, receptors, and

transcription factors that presumably drive the reciprocal inductive interactions

and differential development of these two tissues.

This set of data for normal kidney development then provides the

foundation for the microarray analysis of kidney mutant phenotypes. In this study,

we focused on the analysis of the Hoxa11/Hoxd11 double homozygous mutant

kidneys at early stages of development. At E11.5, the invading UB has branched

once, forming a T-shaped structure, and the MM has not yet formed even the

earliest epithelial nephron precursor, the renal vesicle. It is important to

emphasize that in this study we used laser capture microdissection to purify wild

type and mutant E11.5 MM for microarray analysis. A little later, at E12.5 and

E13.5, the UB has undergone multiple branches and early nephrogenesis is

underway. By looking at these early times, we increased the likelihood of finding

initiating events in the Hoxa11/Hoxd11 mutant disturbed genetic program of

kidney development. These early changes in gene expression in the mutant can

then presumably give rise to further downstream cascade effects at later times,

which might also be interesting, but could be secondary to dramatic changes in

cell differentiation and/or relative abundances of different cell type populations.

We found that the subtle changes in gene expression present in the mutants at early time points were difficult to tease out with microarrays. Indeed,

125 at each developmental stage analyzed, E11.5, and the combined E12.5/E13.5, we found it necessary to use 17 microarrays. Of particular interest, we found that the use of two independent microarray platforms, Affymetrix and Illumina, coupled with independent biological samples, greatly strengthened the study.

A key bottleneck in microarray studies is the validation process. After the microarray data is generated and analyzed a gene list is made. There must, however, be supporting evidence indicating that the gene list is authentic. In this study we found that in our initial comparisons of early wild type and mutant embryonic kidneys, even when using as many as eight Affymetrix microarrays, over half of the genes called differently expressed by microarray were failing to validate by real time PCR. Indeed, we were unable to set a stringency of analysis that clearly distinguished the genes that validated with real time PCR. The use of an additional microarray platform, however, with independent biological replicates, appeared to provide an effective high throughput validation. Genes called differently expressed by both systems consistently validated by real time

PCR. This might be an extremely useful general validation strategy, offering a rapid method for separating true differences from artifacts. Instead of checking for genuine gene expression differences one at a time by real time PCR, one uses a distinct microarray system with new biological samples to confirm changes en masse.

It is interesting that for the wild type-Hoxa11/d11 mutant comparison at both E11.5 and the combined E12.5/E13.5 stages we observed relatively few genes with greater than 3-fold expression change. It was reassuring to find a

126 Hoxa11transcript (Hoxa11 antisense) common to both lists. It was somewhat surprising, however, to find that the two lists of differently expressed genes had very little else in common. Our previous molecular marker studies of these mutant kidneys indicated that the Hox11 genes function in multiple processes at several stages of kidney development, and these distinct gene lists could reflect timing specific variations in function. The Hox11 genes could be driving distinct processes at these two stages. It is also possible that some of the observed differences at E12.5/13.5 could be the result of downstream effects of the earlier changes.

The gene showing the largest fold change in the Hoxa11/d11 mutant kidney at E11.5 was the homeobox gene Irx3, which was up-regulated 6-fold. It is quite interesting that earlier experiments have shown that Irx3 is also a downstream target of both Hoxa9 and Hoxa10 in hematopoietic cells (Ferrell et al., 2005). The Hox9, 10 and 11 paralogous groups are all very closely related

Abd-b type Hox genes, and it is not surprising to see overlap in their downstream targets. In yet another study, looking at development of the hindbrain, it was shown that over expression of Hoxa3, a more distant relative of the Hox11 genes, caused repression of Irx3 transcription (Guidato et al., 2003). These previous connections between Hox genes and Irx3 support the conclusion that

Irx3 is a true downstream effector of Hoxa11/d11 in the developing kidney. It is also interesting that these different Hox genes would share this same target in these different developing systems, although again this is perhaps not surprising in light of the similarities of their DNA binding homeodomains. As multiple targets

127 for multiple Hox genes are unveiled we gain a deeper understanding of Hox functional relationships.

The functions of Irx3 in kidney development remain uncertain. Irx3 has been shown to be capable of regulating competence to respond to hedgehog signaling in the nervous system (Kiecker and Lumsden, 2004). It has also been shown that Irx3 expression can be directed by Wnt signaling in the nervous system (Braun et al., 2003). These observations suggest a possible connection between Irx3 and growth factor signaling in the developing kidney. It has also been previously shown that Irx3 is expressed in the normal developing kidney, but primarily in a more mature structure, the forming glomerulus (Houweling et al., 2001).

The gene with the second greatest fold change in the Hoxa11/d11 mutant kidneys at E11.5 was Slc40a1, an iron exporter also known as Ferroportin1. On two separate probe sets this gene showed down-regulation of −5.2 and −6.1.

This is an especially intriguing observation because of the known importance of iron in kidney development. When growing kidneys in organ culture it is

3+ necessary to add Fe -transferrin to minimal essential media to promote robust

branching morphogenesis of the UB and nephrogenesis of the MM (Landschulz and Ekblom, 1985). Furthermore, another iron transporter, lipocalin, also known as NGAL, has been shown to be an important inducer ofnephrogenesis secreted by the UB (Yang et al., 2003). This established connection between iron levels and nephrogenesis, and our observation that the expression of the Ferroportin1 gene is significantly down-regulated in the mutants, suggests that one molecular

128 mechanism of function of Hoxa11 and Hoxd11 in kidney development might be through the regulation of intracellular iron levels.

Two other genes showed a more than 3-fold down-regulation in the E11.5 mutant MM. Prkcn, a member of the protein kinase C family, showed a 4.5-fold reduction in mutants. It has been previously shown that a number of PKC inhibitors disturb nephron formation and inhibit growth of kidneys in organ culture

(Yang et al., 2003). It is interesting that the inhibition of PKC gives a phenotype resembling that produced by Hoxa11/ d11 mutation.

The homeobox gene Meox2 (also named Gax) was down-regulated 3.4- fold in the E11.5 mutant kidney. Previous studies have demonstrated that over- expression of Meox2 can reduce cell proliferation rates (Fisher et al., 1997; Smith

et al., 1997). This seems somewhat paradoxical, since the mutant kidney shows

reduced Meox2 expression, and reduced size. The Meox2 gene has also been

connected to regulation of cell migration and integrin expression (Perlman et al.,

1999).

Our E12.5/13.5 data provided further evidence that Meox2 is a genuine

downstream target of Hoxa11/Hoxd11 in kidney development. The Affymetrix

microarray data also found a significant decrease in Meox2 expression in the

mutants at this later stage (−3.5-fold change). Real time PCR validation

confirmed this difference at E13.5 (−5.7-fold change, data not shown). The

Illumina microarrays also detected an expression change in this gene (−3.0-fold),

but the observed P value of 0.11 did not make the <0.05 cutoff, so Meox2 was

129 not included in our final list of genes called differently expressed at this later stage.

Several other genes of interest showed altered expression in the E11.5 mutant kidneys, including tropomyosin 2 beta (−2.7-fold, −2.7, and −2.5 on three

separate probe sets), an important component of the cytoskeleton, Wsb1 (−2.6-

fold), a hedgehog inducible ubiquitin ligase (Dentice et al., 2005), and shoc2

(−2.5-fold), a gene involved in fibroblast growth factor receptor signaling (Selfors

et al., 1998). Other genes that showed lower fold changes included Zhx1 (−1.5-

fold) a zinc finger-homeobox gene that is induced by cytokines(Shou et al.,

2004), Sall2 (−1.6), Foxc2 (−2.0), which has been shown to have an important

role in kidney development (Kume et al., 2000a), Rab24(−1.7), a member of the

Ras oncogene family, and Rassf1 (−2.0, −2.1) a Ras associated domain family

member.

At E12.5/13.5 the gene showing the greatest up-regulation encodes the

transcription factor Gata6 (3.6-and 3.0-foldchange on two separate probe sets).

This is particularly interesting because over expression of Gata6 has been

previously shown to cause growth arrest, in both vascular myocytes and

glomerular mesangial cells (Nagata et al., 2000), correlating nicely with the

observed reduced size of the Hoxa11/d11 mutant kidney.

Interestingly, four members of the TGF-beta pathway were also up-

regulated at E12.5/13.5, including Tgfb2 (3.6-fold), Smad6 (2.8), Ltbp1 (2.7), and

fibrillin (2.5, 2.0). Several previous studies have shown the importance of TGF

signaling in kidney development. Tgfb2 has been shown to be expressed in both

130 the MM and UB, and can induce nephrogenesis when added to MM culture along with FGF2 (Plisov et al., 2001). The Tgfb2 knockout mouse shows a renal agenesis phenotype, with incomplete penetrance (Sanford et al., 1997). A conditional MM knockout of Smad4, the downstream activator of TGF-beta signaling, results in abnormal nephrogenesis (Oxburgh et al., 2004). Smad6, an inhibitory Smad, interacts downstream of TGF-beta signals to negatively regulate transcription activation, and interestingly is specifically expressed in the nephrogenic mesenchyme and UB tips of the developing kidney (Vrljicak et al.,

2004). The Fibrillin1 and Ltbp1 genes encode structurally related proteins. TGF- beta is secreted as a latent complex that includes the latent TGF-beta binding protein (Ltbp1) (Annes et al., 2003). The Fibrillin1 protein is a matrix component of the extracellular microfibrils and plays an important role in TGF-beta activation

(Neptune et al., 2003). This observed dysregulation of expression of Tgfb2,

Smad6, Ltbp1 and fibrillin1 suggests an important role for Hox11 genes in the control of TGF-beta signaling in the developing kidney.

Other notable genes up-regulated in the HoxA11/D11mutant E12.5/E13.5 kidneys include, Cxcl12 (3.3-fold), Ldb1 (1.9), Slit3 (2.3), and Robo4 (2.2).

Cxcl12 is a small, c–x–c motif chemokine that is expressed in comma and s- shaped bodies during nephrogenesis and within the mesangium of the maturing glomeruli (Grone et al., 2002). Lbd1 is a cofactor that can interact with many transcription factors, including LIM-homeodomain proteins such Lhx1, which is required for both normal UB branching morphogenesis and developmental progression of the renal vesicle during nephrogenesis (Kobayashi

131 et al., 2005). Lbd1 can increase or decrease transcription factor activity, with correct Lbd1-LIMHD stoichiometry essential for normal development (Neptune et al., 2003). Overexpression of Ldb1 has been associated with a block of differentiation in both immature erythroid cells and mammary epithelial cells

(Visvader et al., 1997, 2001). Slit2 and its receptor Robo2 have been shown to restrict GDNF expression in the intermediate mesoderm thereby restricting the site of the UB formation (Grieshammer et al., 2004). Slit3 and/or Robo4 may have similar roles in repression of GDNF expression. The microarray observed over expression of Slit3 and Robo4 would then be predicted to result in reduced levels of GDNF, and this was indeed found. The level of down-regulation was, however, only 1.3-fold, and therefore not sufficient to make the 1.5-fold cutoff used to make the list of differently expressed genes. These results suggest, but do not prove, that the previously observed loss of GDNF expression in Hox11 triple mutants (Hoxa11, Hoxc11, Hoxd11) (Wellik et al., 2002) might be mediated through increased Slit–Robo expression.

In conclusion, the microarray analysis of a mutant phenotype can be challenging, yet highly rewarding. By examining the early stages of a developmental abnormality, one defines the initiating events, but the differences present between wild type and mutant at this point might be subtle, with many genes showing small changes in expression. In order to obtain convincing results, we found it necessary to use 17 microarrays to compare E12.5/E13.5 wild type and mutant, and then another 17 microarrays to compare E11.5. The initial analysis was performed with Affymetrix microarrays. We found that the use

132 of a second microarray platform, Illumina, with independent biological samples,

provided a high throughput validation that effectively separated genuine

differences from artifacts.

The results of this study identify multiple Hox11 effectors in kidney

development. Many of the genes with altered expression make sense. The gene

with the greatest up-regulation at E11.5, Irx3, has been previously identified as a downstream target of three other Hox genes. In addition, the up-regulation of

Gata6, the perturbation of iron transport, PKC levels, TGF-beta signaling, Lbd1,

Slit and Robo all fit with the observed kidney phenotype of reduced size, reduced branching morphogenesis and altered MM–UB interactions. As more microarray studies are performed on more Hox mutants, in both kidney and other tissues, we will gain deeper insight into the molecular mechanisms of Hox function and the extent of Hox functional overlap.

133 Appendix

Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ydbio.2006.02.023.

134 Acknowledgments

We thank Michael Bennett for help with embryo block preparation, discussions regarding RT-PCR validations, and careful reading of the manuscript. We thank Kristen Saletel and Shawn Smith for generation of microarray data. This work was supported by NIH grant 1RO1DK61916.

135 References

Annes, J.P., Munger, J.S., Rifkin, D.B., 2003. Making sense of latent TGFbeta

activation. J. Cell Sci. 116, 217–224.

Basson, M.A., Akbulut, S., Watson-Johnson, J., Simon, R., Carroll, T.J., Shakya,

R., Gross, I., Martin, G.R., Lufkin, T., McMahon, A.P., Wilson, P.D., Costantini,

F.D., Mason, I.J., Licht, J.D., 2005. Sprouty1 is a critical regulator of

GDNF/RET-mediated kidney induction. Dev. Cell 8, 229–239.

Bolstad, B.M., Irizarry, R.A., Astrand, M., Speed, T.P., 2003. A comparison of

normalization methods for high density oligonucleotide array data based on

variance and bias. Bioinformatics 19, 185–193.

Bouchard, M., 2004. Transcriptional control of kidney development.

Differentiation 72, 295–306.

Brandenberger, R., Schmidt, A., Linton, J., Wang, D., Backus, C., Denda, S.,

Muller, U., Reichardt, L.F., 2001. Identification and characterization of a novel

extracellular matrix protein nephronectin that is associated with integrin

alpha8beta1 in the embryonic kidney. J. Cell Biol. 154, 447–458.

Braun, M.M., Etheridge, A., Bernard, A., Robertson, C.P., Roelink, H., 2003. Wnt

signaling is required at distinct stages of development for the induction of the

posterior forebrain. Development 130, 5579–5587.

Brunskill, E.W., Witte, D.P., Yutzey, K.E., Potter, S.S., 2001. Novel cell lines

promote the discovery of genes involved in early heart development. Dev. Biol.

235, 507–520.

136 Candia, A.F., Hu, J., Crosby, J., Lalley, P.A., Noden, D., Nadeau, J.H., Wright,

C.V., 1992. Mox-1 and Mox-2 define a novel homeobox gene subfamily and

are differentially expressed during early mesodermal patterning in mouse

embryos. Development 116, 1123–1136.

Carroll, T.J., Park, J.S., Hayashi, S., Majumdar, A., McMahon, A.P., 2005. Wnt9b

plays a central role in the regulation of mesenchymal to epithelial transitions

underlying organogenesis of the Mammalian urogenital system. Dev. Cell 9,

283–292.

Challen, G.A., Martinez, G., Davis, M.J., Taylor, D.F., Crowe, M., Teasdale, R.D.,

Grimmond, S.M., Little, M.H., 2004. Identifying the molecular phenotype of

renal progenitor cells. J. Am. Soc. Nephrol. 15, 2344–2357.

Challen, G., Gardiner, B., Caruana, G., Kostoulias, X., Martinez, G., Crowe, M.,

Taylor, D.F., Bertram, J., Little, M., Grimmond, S.M., 2005. Temporal and

spatial transcriptional programs in murine kidney development. Physiol.

Genomics 23, 159–171.

Cho, E.A., Patterson, L.T., Brookhiser, W.T., Mah, S., Kintner, C., Dressler, G.R.,

1998. Differential expression and function of cadherin-6 during renal epithelium

development. Development 125, 803–812.

Davies, J.A., Bard, J.B., 1998. The development of the kidney. Curr. Top. Dev.

Biol. 39, 245–301.

Davis, A.P., Witte, D.P., Hsieh-Li, H.M., Potter, S.S., Capecchi, M.R., 1995.

Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11. Nature 375,

791–795.

137 Dentice, M., Bandyopadhyay, A., Gereben, B., Callebaut, I., Christoffolete, M.A.,

Kim, B.W., Nissim, S., Mornon, J.P., Zavacki, A.M., Zeold, A., Capelo, L.P.,

Curcio-Morelli, C., Ribeiro, R., Harney, J.W., Tabin, C.J., Bianco, A.C., 2005.

The Hedgehog-inducible ubiquitin ligase subunit WSB-1 modulates thyroid

hormone activation and PTHrP secretion in the developing growth plate. Nat.

Cell Biol. 7, 698–705.

Durbec, P., Marcos-Gutierrez, C.V., Kilkenny, C., Grigoriou, M., Wartiowaara, K.,

Suvanto, P., Smith, D., Ponder, B., Costantini, F., Saarma, M., et al., 1996.

GDNF signalling through the Ret receptor tyrosine kinase. Nature 381, 789–

793.

Enomoto, H., Araki, T., Jackman, A., Heuckeroth, R.O., Snider, W.D., Johnson

Jr., E.M., Milbrandt, J., 1998. GFR alpha1-deficient mice have deficits in the

enteric nervous system and kidneys. Neuron 21, 317–324.

Fan, Q.W., Kadomatsu, K., Uchimura, K., Muramatsu, T., 1998. Embigin/ basigin

subgroup of the immunoglobulin superfamily: different modes of expression

during mouse embryogenesis and correlated expression with carbohydrate

antigenic markers. Dev. Growth Differ. 40, 277–286.

Ferrell, C.M., Dorsam, S.T., Ohta, H., Humphries, R.K., Derynck, M.K., Haqq C.,

Largman, C., Lawrence, H.J., 2005. Activation of stem-cell specific genes by

HOXA9 and HOXA10 homeodomain proteins in CD34+ human cord blood

cells. Stem Cells 23, 644–655.

Fisher, S.A., Siwik, E., Branellec, D., Walsh, K., Watanabe, M., 1997. Forced

expression of the homeodomain protein Gax inhibits cardiomyocyte

138 proliferation and perturbs heart morphogenesis. Development 124, 4405–

4413.

Grieshammer, U., Le, M., Plump, A.S., Wang, F., Tessier-Lavigne, M., Martin,

G.R., 2004. SLIT2-mediated ROBO2 signaling restricts kidney induction to a

single site. Dev. Cell 6, 709–717.

Grone, H.J., Cohen, C.D., Grone, E., Schmidt, C., Kretzler, M., Schlondorff, D.,

Nelson, P.J., 2002. Spatial and temporally restricted expression of chemokines

and chemokine receptors in the developing human kidney. J. Am. Soc.

Nephrol. 13, 957–967.

Guidato, S., Prin, F., Guthrie, S., 2003. Somatic motoneurone specification in the

hindbrain: the influence of somite-derived signals, retinoic acid and Hoxa3.

Development 130, 2981–2996.

Hatini, V., Huh, S.O., Herzlinger, D., Soares, V.C., Lai, E., 1996. Essential role of

stromal mesenchyme in kidney morphogenesis revealed by targeted disruption

of Winged Helix transcription factor BF-2. Genes Dev. 10, 1467–1478.

Houweling, A.C., Dildrop, R., Peters, T., Mummenhoff, J., Moorman, A.F.,

Ruther, U., Christoffels, V.M., 2001. Gene and cluster-specific expression of

the Iroquois family members during mouse development. Mech. Dev. 107,

169–174.

Hsieh-Li, H.M., Witte, D.P., Weinstein, M., Branford, W., Li, H., Small, K., Potter,

S.S., 1995. Hoxa 11 structure, extensive antisense transcription, and function

in male and female fertility. Development 121, 1373–1385.

139 Itaranta, P., Lin, Y., Perasaari, J., Roel, G., Destree, O., Vainio, S., 2002. Wnt-6

is expressed in the ureter bud and induces kidney tubule development in vitro.

Genesis 32, 259–268.

Kalatzis, V., Sahly, I., El-Amraoui, A., Petit, C., 1998. Eya1 expression in the

developing ear and kidney: towards the understanding of the pathogenesis of

Branchio-Oto-Renal (BOR) syndrome. Dev. Dyn. 213, 486–499.

Kiecker, C., Lumsden, A., 2004. Hedgehog signaling from the ZLI regulates

diencephalic regional identity. Nat. Neurosci. 7, 1242–1249.

Kispert, A., Vainio, S., McMahon, A.P., 1998. Wnt-4 is a mesenchymal signal for

epithelial transformation of metanephric mesenchyme in the developing

kidney. Development 125, 4225–4234.

Klein, G., Langegger, M., Goridis, C., Ekblom, P., 1988. Neural cell adhesion

molecules during embryonic induction and development of the kidney.

Development 102, 749–761.

Kobayashi, A., Kwan, K.M., Carroll, T.J., McMahon, A.P., Mendelsohn, C.L.,

Behringer, R.R., 2005. Distinct and sequential tissue-specific activities of the

LIM-class homeobox gene Lim1 for tubular morphogenesis during kidney

development. Development 132, 2809–2823.

Kreidberg, J.A., Sariola, H., Loring, J.M., Maeda, M., Pelletier, J., Housman, D.,

Jaenisch, R., 1993. WT-1 is required for early kidney development. Cell 74,

679–691.

140 Kume, T., Deng, K., Hogan, B.L., 2000a. Minimal phenotype of mice

homozygous for a null mutation in the forkhead/winged helix gene, Mf2. Mol.

Cell. Biol. 20, 1419–1425.

Kume, T., Deng, K., Hogan, B.L., 2000b. Murine forkhead/winged helix genes

Foxc1 (Mf1) and Foxc2 (Mfh1) are required for the early organogenesis of the

kidney and urinary tract. Development 127, 1387–1395.

Labastie, M.C., Catala, M., Gregoire, J.M., Peault, B., 1995. The GATA-3 gene is

expressed during human kidney embryogenesis. Kidney Int. 47, 1597–1603.

Landschulz, W., Ekblom, P., 1985. Iron delivery during proliferation and

differentiation of kidney tubules. J. Biol. Chem. 260, 15580–15584.

Lim, K.C., Lakshmanan, G., Crawford, S.E., Gu, Y., Grosveld, F., Engel, J.D.,

2000. Gata3 loss leads to embryonic lethality due to noradrenaline deficiency

of the sympathetic nervous system. Nat. Genet. 25, 209–2012.

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data

using real-time quantitative PCR and the 2(-Delta Delta C(T)) method.

Methods 25, 402–408.

Loughna, S., Hardman, P., Landels, E., Jussila, L., Alitalo, K., Woolf, A.S., 1997.

A molecular and genetic analysis of renalglomerular capillary development.

Angiogenesis 1, 84–101.

Majumdar, A., Vainio, S., Kispert, A., McMahon, J., McMahon, A.P., 2003. Wnt11

and Ret/Gdnf pathways cooperate in regulating ureteric branching during

metanephric kidney development. Development 130, 3175–3185.

141 Miyamoto, N., Yoshida, M., Kuratani, S., Matsuo, I., Aizawa, S., 1997. Defects of

urogenital development in mice lacking Emx2. Development 124, 1653–1664.

Miyazaki, Y., Oshima, K., Fogo, A., Hogan, B.L., Ichikawa, I., 2000. Bone

morphogenetic protein 4 regulates the budding site and elongation of the

mouse ureter. J. Clin. Invest. 105, 863–873.

Moore, M.W., Klein, R.D., Farinas, I., Sauer, H., Armanini, M., Phillips, H.,

Reichardt, L.F., Ryan, A.M., Carver-Moore, K., Rosenthal, A., 1996. Renal and

neuronal abnormalities in mice lacking GDNF. Nature 382, 76–79.

Muller, U., Wang, D., Denda, S., Meneses, J.J., Pedersen, R.A., Reichardt, L.F.,

1997. Integrin alpha8beta1 is critically important for epithelial– mesenchymal

interactions during kidney morphogenesis. Cell 88, 603–613.

Nagata, D., Suzuki, E., Nishimatsu, H., Yoshizumi, M., Mano, T., Walsh, K., Sata,

M., Kakoki, M., Goto, A., Omata, M., Hirata, Y., 2000. Cyclin A downregulation

and p21(cip1) upregulation correlate with GATA-6-induced growth arrest in

glomerular mesangial cells. Circ. Res. 87, 699–704.

Nagy, A., 2003. Manipulating the Mouse Embryo: A Laboratory Manual. Cold

Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Neptune, E.R., Frischmeyer, P.A., Arking, D.E., Myers, L., Bunton, T.E.,

Gayraud, B., Ramirez, F., Sakai, L.Y., Dietz, H.C., 2003. Dysregulation of

TGF-beta activation contributes to pathogenesis in Marfan syndrome. Nat.

Genet. 33, 407–411.

Nishinakamura, R., Matsumoto, Y., Nakao, K., Nakamura, K., Sato, A.,

Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Scully, S., Lacey, D.L., Katsuki,

142 M., Asashima, M., Yokota, T., 2001. Murine homolog of SALL1 is essential for

ureteric bud invasion in kidney development. Development 128, 3105–3115.

Oliver, G., Wehr, R., Jenkins, N.A., Copeland, N.G., Cheyette, B.N., Hartenstein,

V., Zipursky, S.L., Gruss, P., 1995. Homeobox genes and connective tissue

patterning. Development 121, 693–705.

Oxburgh, L., Chu, G.C., Michael, S.K., Robertson, E.J., 2004. TGFbeta

superfamily signals are required for morphogenesis of the kidney mesenchyme

progenitor population. Development 131, 4593–4605.

Patterson, L.T., Potter, S.S., 2004. Atlas of Hox gene expression in the

developing kidney. Dev. Dyn. 229, 771–779.

Patterson, L.T., Pembaur, M., Potter, S.S., 2001. Hoxa11 and Hoxd11 regulate

branching morphogenesis of the ureteric bud in the developing kidney.

Development 128, 2153–2161.

Perlman, H., Luo, Z., Krasinski, K., Le Roux, A., Mahfoudi, A., Smith, R.C.,

Branellec, D., Walsh, K., 1999. Adenovirus-mediated delivery of the Gax

transcription factor to rat carotid arteries inhibits smooth muscle proliferation

and induces apoptosis. Gene Ther. 6, 758–763.

Pichel, J.G., Shen, L., Sheng, H.Z., Granholm, A.C., Drago, J., Grinberg, A., Lee,

E.J., Huang, S.P., Saarma, M., Hoffer, B.J., Sariola, H., Westphal, H., 1996.

Defects in enteric innervation and kidney development in mice lacking GDNF.

Nature 382, 73–76.

143 Piper, M., Georgas, K., Yamada, T., Little, M., 2000. Expression of the vertebrate

Slit gene family and their putative receptors, the Robo genes, in the developing

murine kidney. Mech. Dev. 94, 213–217.

Plisov, S.Y., Yoshino, K., Dove, L.F., Higinbotham, K.G., Rubin, J.S., Perantoni,

A.O., 2001. TGF beta 2, LIF and FGF2 cooperate to induce nephrogenesis.

Development 128, 1045–1057.

Quaggin, S.E., Schwartz, L., Cui, S., Igarashi, P., Deimling, J., Post, M., Rossant,

J., 1999. The basic-helix–loop–helix protein pod1 is critically importa nt for

kidney and lung organogenesis. Development 126, 5771–5783.

Rozen, S., Skaletsky, H., 2000. Primer3 on the WWW for general users and for

biologist programmers. Methods Mol. Biol. 132, 365–386.

Ryan, G., Steele-Perkins, V., Morris, J.F., Rauscher III, F.J., Dressler, G.R.,

1995. Repression of Pax-2 by WT1 during normal kidney development.

Development 121, 867–875.

Sanchez, M.P., Silos-Santiago, I., Frisen, J., He, B., Lira, S.A., Barbacid, M.,

1996. Renal agenesis and the absence of enteric neurons in mice lacking

GDNF. Nature 382, 70–73.

Sanford, L.P., Ormsby, I., Gittenberger-de Groot, A.C., Sariola, H., Friedman, R.,

Boivin, G.P., Cardell, E.L., Doetschman, T., 1997. TGFbeta2 knockout mice

have multiple developmental defects that are non-overlapping with other

TGFbeta knockout phenotypes. Development 124, 2659–2670.

Saxen, L., Lehtonen, E., 1987. Embryonic kidney in organ culture. Differentiation

36, 2–11.

144 Schmidt-Ott, K.M., Yang, J., Chen, X., Wang, H., Paragas, N., Mori, K., Li, J.Y.,

Lu, B., Costantini, F., Schiffer, M., Bottinger, E., Barasch, J., 2005. Novel

regulators of kidney development from the tips of the ureteric bud. J. Am. Soc.

Nephrol. 16, 1993–2002.

Schnabel, C.A., Godin, R.E., Cleary, M.L., 2003. Pbx1 regulates nephrogenesis

and ureteric branching in the developing kidney. Dev. Biol. 254, 262–276.

Schuchardt, A., D'Agati, V., Larsson-Blomberg, L., Costantini, F., Pachnis, V.,

1994. Defects in the kidney and enteric nervous system of mice lacking the

tyrosine kinase receptor Ret. Nature 367, 380–383.

Schwab, K., Patterson, L.T., Aronow, B.J., Luckas, R., Liang, H.C., Potter, S.S.,

2003. A catalogue of gene expression in the developing kidney. Kidney Int. 64,

1588–1604.

Seifert, R.A., Alpers, C.E., Bowen-Pope, D.F., 1998. Expression of platelet-

derived growth factor and its receptors in the developing and adult mouse

kidney. Kidney Int. 54, 731–746.

Selfors, L.M., Schutzman, J.L., Borland, C.Z., Stern, M.J., 1998. soc-2 encodes a

leucine-rich repeat protein implicated in fibroblast growth factor receptor

signaling. Proc. Natl. Acad. Sci. U. S. A. 95, 6903–6908.

Shou, Z., Yamada, K., Kawata, H., Yokoyama, O., Miyamoto, K., 2004. A

mechanism of induction of the mouse zinc-fingers and homeoboxes 1 (ZHX1)

gene expression by interleukin-2. Biochem. Biophys. Res. Commun. 314, 885–

890.

145 Small, K.M., Potter, S.S., 1993. Homeotic transformations and limb defects in

Hox A11 mutant mice. Genes Dev. 7, 2318–2328.

Smith, R.C., Branellec, D., Gorski, D.H., Guo, K., Perlman, H., Dedieu, J.F.,

Pastore, C., Mahfoudi, A., Denefle, P., Isner, J.M., Walsh, K., 1997. p21CIP1-

mediated inhibition of cell proliferation by overexpression of the gax

homeodomain gene. Genes Dev. 11, 1674–1689.

Stark, K.L., McMahon, J.A., McMahon, A.P., 1991. FGFR-4, a new member of

the fibroblast growth factor receptor family, expressed in the definitive

endoderm and skeletal muscle lineages of the mouse. Development 113, 641–

651.

Stark, K., Vainio, S., Vassileva, G., McMahon, A.P., 1994. Epithelial

transformation of metanephric mesenchyme in the developing kidney regulated

by Wnt-4. Nature 372, 679–683.

Stuart, R.O., Bush, K.T., Nigam, S.K., 2001. Changes in global gene expression

patterns during development and maturation of the rat kidney. Proc. Natl.

Acad. Sci. U. S. A. 98, 5649–5654.

Tessarollo, L., Nagarajan, L., Parada, L.F., 1992. c-ros: the vertebrate homolog

of the sevenless tyrosine kinase receptor is tightly regulated during

organogenesis in mouse embryonic development. Development 115, 11–20.

Virtanen, I., Laitinen, L., Korhonen, M., 1995. Differential expression of laminin

polypeptides in developing and adult human kidney. J. Histochem. Cytochem.

43, 621–628.

146 Visvader, J.E., Mao, X., Fujiwara, Y., Hahm, K., Orkin, S.H., 1997. The LIM-

domain binding protein Ldb1 and its partner LMO2 act as negative regulators

of erythroid differentiation. Proc. Natl. Acad. Sci. U. S. A. 94, 13707–13712.

Visvader, J.E., Venter, D., Hahm, K., Santamaria, M., Sum, E.Y., O'Reilly, L.,

White, D., Williams, R., Armes, J., Lindeman, G.J., 2001. The LIM domain

gene LMO4 inhibits differentiation of mammary epithelial cells in vitro and is

overexpressed in breast cancer. Proc. Natl. Acad. Sci. U. S. A. 98, 14452–

14457.

Vrljicak, P., Myburgh, D., Ryan, A.K., van Rooijen, M.A., Mummery, C.L., Gupta,

I.R., 2004. Smad expression during kidney development. Am. J. Physiol.:

Renal. Physiol. 286, F625–F633.

Wellik, D.M., Hawkes, P.J., Capecchi, M.R., 2002. Hox11paralogous genes are

essential for metanephric kidney induction. Genes Dev. 16, 1423–1432.

Yang, J., Mori, K., Li, J.Y., Barasch, J., 2003. Iron, lipocalin, and kidney epithelia.

Am. J. Physiol.: Renal. Physiol. 285, F9–F18.

Yoshino, K., Rubin, J.S., Higinbotham, K.G., Uren, A., Anest, V., Plisov, S.Y.,

Perantoni, A.O., 2001. Secreted Frizzled-related proteins can regulate

metanephric development. Mech. Dev. 102, 45–55.

Yu, J., Carroll, T.J., McMahon, A.P., 2002. Sonic hedgehog regulates

proliferation and differentiation of mesenchymal cells in the mouse

metanephric kidney. Development 129, 5301–5312.

Yu, J., McMahon, A.P., Valerius, M.T., 2004. Recent genetic studies of mouse

kidney development. Curr. Opin. Genet. Dev. 14, 550–557.

147 Figures and tables

148 Table 1 continued.

149 Table 2.

150 Table 3.

151

Figure 1. Heat map showing 2793 genes clustered according to expression during kidney development. Developmental times include E12.5, E13.5, E16.5 and adult. Expression levels were normalized to the adult kidney. Red shows elevated, blue indicates reduced, and yellow shows unchanged expression. Only genes with greater than 3-fold change are included. Complete gene lists and analysis tools are available online (see Methods).

152

Figure 2. Scattergraphs of E11.5 UB and MM microarray data. (A) Biological replicates of E11.5MMpurified by LCM. (B) Comparison of manually dissected and LCM isolated E11.5 MM. (C) Biological replicates of manually dissected

E11.5 UB. (D) Comparison of UB and MM samples. Note the good reproducibility of MM data, whether isolated by manual microdissection or LCM, and the still higher reproducibility of the UB data. The MM-UB comparison, in contrast, shows striking differences. Green lines mark the 3-fold difference boundary. Each dot represents one probeset that has been called by the GCOS software as

“Present” in each microarray, except in D in which at least one probeset must be called “Present.”

153

Figure 3. Heat map showing 1518 genes clustered according to differential expression in E11.5 UB and MM. The MM expression levels were normalized to

UB. The upper blue cluster represents 800 genes with decreased expression in the MM. The lower red cluster shows 718 genes with higher expression levels in the MM. Note the high reproducibility of the expression differences across all samples.

154

Figure 4. Heat map showing 122 genes clustered according to differential expression in wild type and Hoxa11/Hoxd11 mutant kidneys. Expression levels were examined at E12.5, E13.5 and E16.5. Genes were selected by average fold change >1.5 at both E12.5 and E13.5, using both Affymetrix and Illumina microarray platforms (see text for details). All mutant expression levels were normalized to the corresponding microarray platform's wild type expression. Red indicates increased, blue indicates decreased and yellow shows unchanged expression. Most of the differences seen at E12.5 and E13.5 were also present at E16.5

155

Figure 5. QPCR validation of microarray predicted gene expression differences.

Cxcl12, HoxA11s, Pdgfrb, Slit3, and Tgfb2 were examined. Separate E13.5 mutant and wild type kidney samples were assayed in sample groups of at least three. Graph shows mutant fold change normalized to wild type group according to QPCR (blue bars) or microarray data (red bars). The numbers agreed well, except for HoxAlls (HoxA11 antisense), and Slit3, where the QPCR called larger fold changes than the microarrays.

156

Figure 6. Log series graph of 146 genes differentially expressed between

HoxA11/D11 null and wild type E11.5 MM. This graph shows the consistent change in expression for these genes using both the Affymetrix and Illumina platforms. Genes were selected on statistical significance and an average fold change of >1.5 regardless of platform (see text for details). Gene in red are up- regulated and those in blue are down-regulated, as normalized to wild type.

157 Chapter 4

Pygo1 and Pygo2 roles in Wnt signaling in mammalian kidney development *

Kristopher R. SchwabA, Larry T. PattersonB, Heather A. HartmanB, Ni SongC,

Richard A. LangC, Xinhua LinA, and S. Steven PotterA.

Children's Hospital Medical Center

Divisions of Developmental BiologyA, NephrologyB, and OphthalmologyC

Children’s Hospital Medical Center,

3333 Burnet Avenue

Cincinnati, OH 45229, USA

* Submitted to Dev Biol. August 18, 2006

158 Abstract

In Drosophila the Pygopus gene encodes an essential component of the Armadillo

(β-catenin) transcription factor complex of canonical Wnt signaling. To better understand

the functions of canonical Wnt signaling in kidney development we made targeted

mutations in the two mammalian orthologs, Pygo1 and Pygo2. Each mutation deleted

over 80% of coding sequence, including the critical PHD domain, and almost certainly

resulted in null function. Pygo1 homozygous mutants were viable and fertile, with no

detected developmental defects. Pygo2 homozygous mutants died shortly after birth, with

a phenotype including lens agenesis, growth retardation, and in some cases exencephaly

and cleft palate. Double Pygo1/Pygo2 homozygous mutants showed no apparent synergy

in phenotype severity. The Bat-gal transgene reporter of canonical Wnt signaling showed

reduced, but not absent, expression in Pygo1-/-/Pygo2-/- mutants, with tissue specific variation in level of diminution. The Pygo1 and Pygo2 genes both showed widespread expression in the developing kidney, with elevated levels in the stromal cell compartment.

Confocal analysis of the double mutant kidneys showed perturbed branching morphogenesis of the ureteric bud and an expansion of the zone of condensed mesenchyme capping the ureteric bud. Nephron formation, however, proceeded normally.

Microarray analysis showed changed expression of a limited number of genes, including

Cxcl13, Slc5a2, Klk5, Ren2 and Timeless. The relatively mild phenotype observed in the kidney, as well as other organ systems, indicated a striking evolutionary divergence of

Pygopus gene function. In mammals the Pygo1/Pygo2 genes are not required for canonical Wnt signaling, but rather function as quantitative transducers, or modulators, of signal intensity.

159 Introduction

Wnt signaling is of critical importance in several stages of kidney development. Mutual inductive interactions between the ureteric bud and metanephric mesenchyme drive nephrogenesis (Saxen and Sariola, 1987). The ureteric bud synthesizes Wnt9b, which is essential for induction of the mesenchyme to form nephrons (Carroll et al., 2005). Wnt4 is made by the induced metanephric mesenchyme and is also required for nephrogenesis (Stark et al 1994). Furthermore, Wnt11, secreted by the ureteric bud tips, participates in a positive feedback loop promoting GDNF expression by the metanephric mesenchyme (Majumdar et al, 2003). Mutations in Wnt9b or Wnt4 result in a dramatic block in nephron formation, while Wnt11 mutants show a significant reduction in nephron number. It is interesting to note that Wnt4 and Wnt11 have been shown to signal, at least in some cases, through noncanonical pathways

(Du et al., 1995; Maurus et al., 2005; Pandur et al., 2002), while there is evidence indicating that Wnt9b activates canonical Wnt signaling (Carroll et al., 2005).

Genetic studies in Drosophila identified the pygopus (pygo) and legless

(lgs) genes as critical components of canonical Wnt signaling (Belenkaya et al.,

2002; Kramps et al., 2002; Parker et al., 2002; Thompson et al., 2002). Pygo and

Lgs interact with β-catenin during the formation of the canonical transcriptional

complex and are required for accumulation of β-catenin in the nucleus (Townsley

et al., 2004). Lgs binds the central armadillo repeats of β-catenin while Pygo

interacts with Lgs, mediating activation of Wnt targets (Kramps et al., 2002;

Stadeli and Basler, 2005). While experimental evidence suggests Lgs functions

160 solely as an adaptor, Pygo functions in the activation of Wnt target transcription

(Belenkaya et al., 2002; Kramps et al., 2002). Indeed, the N-terminal domain of

Pygo is required for Wnt transcriptional activation, while the PHD motif is required for Pygo’s association with Lgs (Kramps et al., 2002; Stadeli and Basler,

2005). Additionally, a putative nuclear localization signals (NLS) was identified within the N-terminal domain of Pygo suggesting a possible role of nuclear importation of β-catenin (Belenkaya et al., 2002; Thompson et al., 2002). These studies demonstrate the crucial requirement of Pygo for canonical Wnt signaling during Drosophila development.

Two orthologs of Drosophila pygo, Pygo1 and Pygo2, have been identified in mammals suggesting a conservation of pygo function in canonical Wnt signaling during evolution (Kramps et al., 2002; Belenkaya et al., 2002; Li et al.,

2004). In this report we generated targeted mutations of Pygo1 and Pygo2 to determine their functions in mammalian development. The resulting double homozygous mutant embryos showed a reduction in canonical Wnt signaling as measured by Wnt reporter transgene expression. Development remained, however, surprisingly normal, with survival to birth and few apparent defects in most organ systems. Our phenotypic analysis focused on the kidney, showing reduced branching and altered morphology of the ureteric bud, as well as expansion of the zone of condensed mesenchyme, yet relatively normal nephron formation, as measured by histology, confocal analysis, in situ hybridizations and microarray analysis. These results provide deeper insight into the roles of the

161 Pygo genes in Wnt signaling in mammals, and the roles of canonical Wnt signaling in nephrogenesis.

162 Materials and Methods

Targeting of Pygo1 and Pygo2

Pygo1 and Pygo2 genes were each targeted by flanking the critical coding regions of the third exon containing the PHD motif with LoxP sequences.

Targeting constructs were made using the Flp/Cre vector previously described

(Bell et al., 2003), which carries a neomycin resistance gene flanked by FRT sequences, and three unique restriction sites for subcloning the two blocks of homology for driving recombination, and the critical region to be flanked by LoxP sequences. For each construct the three genomic segments required for subcloning were made by PCR from RI ES cell DNA. We confirmed the resulting targeting constructs by sequencing.

For PygoI the genomic sequences used for PCR were: 5’ forward,

GTGAAGGAGAGATGGATAAGTATG, 5’ back, TAGACCCTAACCACCTACAAG, exon forward, GGTTAGGGTCTATGTGCTGG, exon back,

TCACCAAATCTCTGTTCTACAC, 3’ forward, TGTGTAGAACAGAGATTTGGTG,

3’ back, CAGTGAAGAAAGAGGGTCAG. For Pygo II the genomic sequences used for PCR were: GCCTGGGTTGCTTGTCTTCTG and

CCACCTTACTTGTGTGTGAGGATACATAC, CCAAGTCCCAGCATCTCTTAC and CCAGTCATACCAGCAACAAG, and exon sequences

TGGGTGCTGGGAACAGAAC and CAACAACAACAGAAGACAAGC.

Linearized constructs were electroporated into RI ES cells, and resulting targeted ES cells used for C57/Bl6 blastocyst injections according to standard

163 protocols. Resulting chimeras were mated with Swiss Black mice, and the targeted stocks maintained on a mixed 129/Swiss Black background.

Germline null alleles of both Pygo1 and Pygo2 were generated by mating

heterozygous floxed mice with the CMV-Cre mice (Schwenk et al., 1995). The

sequences of the primers used for genotyping PCR were: Pygo2 null allele,

forward (F) CCTGGATTCTTGTTGCTGGTATG, reverse (R)

AAGGTATTTGGTGCTCCGAGGG; Pygo2 WT or floxed allele, F

TGTCTTGATGACAGCGTTTAGCC, R AGATTCAGTAAGCTGAGCCTGGTG;

Pygo1 null allele, F AGTTTGAAATAGCGACGAGTTTGAG, R 5’-

CACTTCTGCCCCTCTCTTTGC; Pygo1 WT or floxed allele, F

AAGCGTGCCTCATCTCCATCCCTAAG, R GCCCTCCCCGACGTTTATATTG.

Noon of the day vaginal plugs were observed was designated E0.5.

Confocal Microscopy

Kidneys were dissected, fixed in paraformaldehyde, treated with methanol,

and washed with PBS containing Tween-20 (PBT) prior to treatment with

lectins and antibodies. PBT was used for incubations of the tissues with

fluoroscein-conjugated Dilichos biflorus aggutinin (DBA, Vector) while PBT

containing goat serum was used for incubations with the antibodies. The

primary antibodies were anti-WT1 (c-19, Santa Cruz), anti-Uvomorulin

(e-cadherin, Cdh1, Sigma), and anti-Cited1 (Neomarkers). The secondary

antibodies were Alexa 555-conjugated anti-rabbit and Alexa 633-conjugated

anti-rat antibodies (Molecular Probes).

164 The tissues were imaged with a Zeiss laser scanning microscope equipped with an Argon (488nm) and two HeNe lasers (543nm and 633nm).

Approximately 2 micron thick optical sections were obtained every 5 microns to a depth of at least 65 microns. The sections began at the surface of the kidney and were on a plane tangential to it. Two Z-stack series were obtained, one from each of the two kidneys of each embryo. Ureteric bud tips identified by section tracing were counted within a defined area of the confocal image.

In Situ Hybridizations

Whole mount in situ hybridizations were performed as previously described (Patterson et al., 2001). Riboprobes to Wnt11 and Wnt7b were described previously (Patterson et al., 2001). The Wnt9b riboprobe was provided by T. Carroll (Carroll et al., 2005).

Pygo1 and Pygo2 antibody production

To generate anti-human Pygo2 (anti-hPygo2) and anti-mouse Pygo1 (anti- mPygo1) polyclonal antibodies, we subcloned cDNA by PCR corresponding to amino acid residues 80-327 of human Pygo 2 or amino acid residues 76-263 of mouse Pygo1 into pGEX4T1 (Amersham). The PCR fragments of hPygo 2 and mPygo 1 lack both NHD and PHD conserved regions of hPygo2 and mPygo1.

GST-hPygo2 and GST-mPygo1 proteins were expressed in bacteria, purified, and injected into rabbit for antibody production by proteintech group inc.

(Chicago). The rabbit antisera of anti-mPygo1 and anti-hPygo2 were initially allowed to bind to the GST affinity matrix to remove any antibodies against GST.

The anti-hPygo2 and anti-mPygo1 antisera were then separated from the GST

165 affinity matrix and allowed to bind to the GST-hPygo2 or GST-mPygo1 affinity columns, respectively. The bound antibodes were eluted with elution buffer.

Bat-gal transgene reporter assay of canonical Wnt signaling.

X-Gal staining of both embryos and developing kidneys was performed as previously described (Maretto et al., 2003). Care was taken to reduce endogenous β-Galactosidase activity within the developing kidney by increasing pH of the X-gal staining solution to 8.0. Changes in transgene β-Gal expression

were quantitated using a β-Gal ELISA (Enzyme-linked immunoassay) kit (Roche,

Indianapolis, IN), normalizing according to total protein. Each genotype was

represented by a sample size of 4 except the non-transgenic (n=5), Pygo1 +/-;

Pygo2 +/+ (n=6), Pygo1 -/-; Pygo2 +/- (n=8), and Pygo1 -/-; Pygo2 -/- (n=6) groups.

Microarray analysis

Total RNA was isolated from E18.5 kidneys dissected from normal and

Pygo1/Pygo2 compound null embryos using the Stratagene Absolutely RNA

Microprep kit (La Jolla, CA). 300 ng of total RNA was amplified and labeled

using the TargetAmp 1-Round Aminoallyl-aRNA Amplification Kit (Epicentre,

Madison, WI). Amplified RNA was hybridized to Sentrix Mouse-6 expression

Beadchip microarrays (Illumina, San Diego, CA) providing coverage of over

47,000 genes and ESTs as previously described (Schwab et al., 2006).

Microarray analysis of Pygo1/Pygo2 null and normal wild type kidneys was

performed in biological triplicate. Raw signal intensities of each probe were

obtained from Beadstudio Data Analysis software (Illumina, San Diego, CA) and

166 imported into Genespring GX 7.3 (Agilent Technologies, Palo Alto, CA). Genes were selected on the basis of greater than two-fold average fold change and statistical significance (P-value < 0.05). Previously described Wnt target genes were obtained from http://www.stanford.edu/~rnusse/ pathways/targets.html.

QPCR validation of microarray results

Total RNA from E18.5 Pygo1/Pygo2 null and control kidneys (both represented in duplicate and distinct from the kidneys used for microrray analysis) was purified using Stratagene Absolutely RNA Microprep Kit, including

DNAse1 treatment (La Jolla, CA). cDNA was generated using random hexamers according to conventional protocols (Invitrogen, Carlsbad, Ca). The following primers were generated to include the sequence obtained from the Illumina probe: Actb (TTGCTGACAGGATGCAGAAG, ACATCTGCTGGAAGGTGGAC);

Aldh1a7 (CCAGAAAGTGGTGTTTTGCT, GAGTTACAGAGAGCTTGCACCTT);

Col8a1 (GCAGACAGGCATCTTCACCT, TGTGTACATCATGGGCTCGT); Csrp1

(CAGCATAGCCCAGGGTAGAG, TGGGCAAGGTAGTGAAGGTT); Klk5

(GCAGCATTACACCCGTCATA, TTGCCTCCATCCATATCTCC); Picalm

(GGGAGGGAACAGAAATCCTT, GCACCGATCAACAGTGCAG); Pygo2

(TTCTGGGAACTTGTGCACTG, AACTTCCTCCAGCCCATTTT); Ren2

(TTGTGAGCTGTAGCCAGGTG, TGTGCACAGCTTGTCTCTCC) and Tia1

(TGATTGAAGGGCTACTAGAGTGGT, AGCCCCAGAGGACAATTTTT) using

Primer3 software (Rozen and Skaletsky, 2000). Relative quantitative PCR was performed according to the conventional SYBR Green protocol (Stratagene) using the Stratagene Mx3000p QPCR system. Dissociation curve and agarose

167 gel analysis of each primer set were used to insure specificity of the amplicon. All data were normalized to an internal housekeeping control (Actb) and analyzed using the 2(-Delta Delta C(T)) method (Livak and Schmittgen, 2001).

168 Results

Pygo1 and Pygo2 targeted mutations

We targeted both the Pygo1 and Pygo2 genes to allow the study of possible functional redundancies. We inserted LoxP sequences in both genes flanking critical coding regions including the PHD domains. The resulting targeted mice were mated with transgenic CMV-Cre mice (Schwenk et al., 1995) to drive germline LoxP recombination, making the null mutant alleles that were used for this study. PCR confirmed the deletion of the bulk of the coding sequences for both genes, including 89% of coding sequence for Pygo1 and 87% of coding sequence for Pygo2 (see Materials and Methods for details). Previous studies in

Drosophila have shown that even a single missense mutation in the PHD domain can eliminate Pygo function in Wnt signaling (Belenkaya et al., 2002).

Pygo1 and Pygo2 mutant phenotypes.

Pygo1 homozygous null mice were viable and fertile with no developmental defects detected. This was surprising given the importance of

Pygo genes in Wnt signaling in Drosophila, and the observed expression of the mouse Pygo1 gene in, for example, the brain, limbs, kidney and branchial arches, during development (Gray et al., 2004; Li et al., 2004; Yu J, Valerius MT,

McMahon AP, GUDMAP, www.gudmap.org).

The Pygo2 homozygous null mice survived to birth, but with rare exception died shortly thereafter. The gut, heart and limbs developed without detected abnormality (data not shown) despite known requirements for Wnt signaling. The

Pygo2 mutants did, however, show growth retardation, lens agenesis and a

169 kidney phenotype with high penetrance, as well as exencephaly, and cleft palate with incomplete penetrance.

Double homozygous mutant Pygo1 and Pygo2 mice gave a phenotype

similar to that of single Pygo2 nulls (Fig. 1 and Table 1). There was no

significant synergism of developmental abnormalities in the double mutants,

arguing against functional redundancy as an explanation of the relatively mild

phenotypes seen in single mutants. Together these results suggest that Pygo2 is

needed for the proper development of a limited number of structures, while

Pygo1 is not necessary for normal development.

Expression of Pygo1 and Pygo2 in the developing kidney

We used immunofluorescence to define the expression patterns of the

Pygo1 and Pygo2 genes in the developing kidney. Both genes were widely

expressed, showing nuclear localization of encoded proteins in the ureteric bud,

capping mesenchyme, and stromal cells (Fig. 2). The Pygo1 gene showed a

distinctly elevated expression in stromal cells, and to a lesser degree this was

also seen for Pygo2, but some level of nuclear staining was detected in

essentially all cells of the developing kidney for both genes.

Confocal analysis of Pygo1/Pygo2 mutant kidneys

Given the crucial roles of Wnt signaling during renal development,

morphological analysis of Pygo2 null and Pygo1/Pygo2 double null kidneys was

performed. Histological sections of mutant kidneys revealed that nephrogenesis

was not severely affected. The full complement of developing kidney structures,

including ureteric bud tips, collecting ducts, early developing nephrons, and

170 maturing nephrons including glomeruli, were identified in both Pygo2-/- and

Pygo1/Pygo2 double homozygous mutant kidneys by standard histological

methods (data not shown).

In order to better visualize renal development, confocal sectioning of

E18.5 Pygo1/Pygo2 null kidneys was performed using a panel of antibodies to

identify the primary developing structures of the embryonic kidney (Fig. 3). Wt1

(red) and Cited1 (red) are both expressed in the capping metanephric

mesenchyme around the ureteric bud tips (Pritchard-Jones et al., 1990). WT1

also stains renal vesicles and glomerular anlage. Cdh1, also known as E-

cadherin (blue), was used to identify epithelial structures including the branching

ureteric bud and nascent nephrons of the developing kidney (Vestweber et al.,

1985). Dolichos biflorus (DBA) lectin (green) was used to selectively stain

ureteric bud-derived structures (Fan et al., 1998).

The metanephric mesenchyme underwent normal nephrogenesis in the

double mutant kidneys (Fig. 3B,D). Expression of both Wt1 and Cited1 (red) was

not perturbed in the Pygo1/Pygo null embryos. Additionally, Cdh1-staining

nephrons (blue) were identified connecting to the ureteric bud tips (green) and

extending into the medulla of the kidney, in normal fashion. Intermediate

structures of nephrogenesis, including renal vesicles, comma shaped bodies and

S-shaped bodies also appeared normal.

The ureteric bud tips (green) of Pygo1/Pygo2 null kidneys, however, were

dilated and misshaped compared to littermates with at least one wild type Pygo2

allele (Fig. 3). Pygo1/Pygo2 double homozygous mutant kidneys also had an

171 approximately 25% decreased number of ureteric bud tips per area compared to littermates with at least one wild type copy of Pygo2 (Table 2). No significant difference in ureteric bud tip density was seen between Pygo2+/- and Pygo2+/+ embryos (P-Value = 0.33). In addition we observed a thickening of the capping mesenchyme, surrounding the ureteric buds (Fig. 3). Measurements showed this zone of condensed mesenchyme to be about 30% wider in the Pygo2-/- mice

compared to wild type and Pygo2+/- (Table 2).

In situ hybridizations

We examined expression of three Wnt genes in Pygo mutants. Wnt7b is

expressed in the stalks (Kispert et al., 1996), Wnt11 is expressed in the tips

(Majumdar et al., 2003), and Wnt9b is expressed in stalks and weakly in the tips

of the branching ureteric bud (Carroll et al., 2005). Previous work had shown the

presence of a positive feedback loop between Wnt11 synthesis in the ureteric

bud and metanephric mesenchyme response (Majumdar et al., 2003). Altered

Wnt synthesis by the ureteric bud could therefore reflect a disrupted metanephric

mesenchyme Wnt response. All three of these Wnt genes, however, showed

similar expression levels in Pygo+/+ and Pygo1/Pygo2 double mutant kidneys

(Fig. 4). In addition these in situ hybridization patterns confirmed the confocal

microscopy results, showing a relatively normal ureteric bud branching

morphogenesis, but with a reduced number of tips per surface area (Wnt11

pattern), or tip density, in the Pygo1/Pygo2 double homozygous mutants.

Reduced canonical Wnt signaling in Pygo1/Pygo2 mutant mice.

172 The Bat-gal transgene reporter of canonical Wnt signaling (Maretto et al.,

2003) was used to examine changes in Wnt signaling in the Pygo mutant mice.

Both Pygo2-/- and Pygo1/Pygo2 double null E10.5 embryos showed a decrease

of canonical Wnt signaling (Fig. 5). There was, however, tissue specific variability

in the degree of Wnt signaling reduction, with the somites, for example, showing

a strong reduction, while the mutant telencephalon still showed robust Wnt

signaling. These results suggest that the mammalian Pygo genes are important

modulators of canonical Wnt signaling in some developing systems.

We also examined Bat-gal reporter expression in the developing

urogenital system of Pygo mutants. The results suggested that the Pygo2 gene is required for canonical Wnt signaling in the nephric duct. Both Pygo2 null and

Pygo1/Pygo2 double null E10.5 embryos showed an absence of reporter

expression in the nephric duct, while control littermates showed strong

expression (Fig. 6A-C). Interestingly, however, even in Pygo1/Pygo2 double null

mutants the nephric duct did form and give rise to the ureteric bud outgrowth,

which showed reduced but not absent Bat-gal reporter expression (Fig. 6A-C).

At E13.5 the Pygo2 gene appeared to play a major role, and the Pygo1

gene a minor role, in canonical Wnt signaling in the ureteric tree, as measured by

Bat-gal expression. A negative control kidney, without the Bat-gal transgene,

showed minimal background X-gal staining (Fig. 6D), while a Bat-gal transgenic

kidney with at least one wild type Pygo2 gene showed strong X-gal staining in

the ureteric tree (Fig. 6E). In contrast, Pygo1+/-/Pygo2-/- E13.5 kidneys showed

very weak reporter expression (Fig. 6F), suggesting a significant loss of

173 canonical Wnt signaling. Homozygous loss of the Pygo1 gene alone, however, had a small effect on reporter expression (Fig. 6G-H). Homozygous mutation of both the Pygo1 and Pygo2 genes gave a more dramatic reduction of Bat-gal

expression than loss of Pygo2 alone (Fig. 6F,L).

The Pygo1, Pygo2 genes were also required for Bat-gal reporter

expression in the paramesonephric (Mullerian) ducts. In Pygo2-/- mice there was

a significant loss of reporter expression (data not shown), while Pygo1-/-/Pygo2+/- and Pygo1+/-/Pygo2+/- mice showed similar levels of Bat-gal expression (Fig.

6G,H,J), and double homozygous mutants showed loss of X-gal staining in the paramesnophric ducts (Fig. 6K).

At E18.5 Bat-gal reporter analysis of the Pygo mutants also identified a significant decrease in canonical Wnt-signaling in the cortical ureteric branches and renal pelvis of the developing kidney (Figure 6M-O). Cortical X-gal staining was seen in the ureteric branches of a Pygo1 +/-/ Pygo2 +/- kidney (Fig. 6M, left), but was completely absent in the cortex of a Pygo1+/- /Pygo2 -/- kidney (Fig. 6M,

right). Bisection revealed a significant loss of X-gal staining cells in the collecting

ducts and renal pelvis of the Pygo2 null kidney compared to the control littermate

(Fig. 6N). Side by side comparison of Pygo1 +/+/Pygo2 +/- (Fig. 6O, left), Pygo1-/-

/Pygo2+/- (Fig. 6O, middle), and Pygo1-/-/Pygo2-/- (Fig. 6O, right) E18.5 kidneys

shows the primary role of the Pygo2 gene in canonical Wnt signaling in the ureteric tree and its derivatives.

In order to validate and quantify the Bat-gal reporter expression changes

in the Pygo2 null and Pygo1/Pygo2 nulls, we performed ELISA measurements of

174 transgene specific β-galactosidase levels in E18.5 kidneys (Fig. 7). Loss of the

Pygo2 gene (Pygo1+/-/Pygo2-/- and Pygo1-/-/Pygo2-/-) gave greater than 90%

reduction in Bat-gal expression. Loss of Pygo1 alone (Pygo1-/-/Pygo2+/+) did not

result in a significant change. Interestingly, however, the Pygo1-/-/Pygo2+/- showed only 50% of wild type Bat-gal expression, suggesting a minor contribution by Pygo1 in canonical Wnt signaling. Although Bat-gal expression

was decreased in the E18.5 Pygo1-/- /Pygo2+/- kidney, confocal analysis of kidneys with this genotype revealed no dilated ureteric tips or significant changes in ureteric tip number per area compared to Pygo1-/-/Pygo2+/+ kidneys (data not shown). Collectively, these Bat-gal reporter results suggested an important role for the Pygo1 and Pygo2 genes in canonical Wnt signaling during development of the ureteric tree of the kidney.

Interestingly however, even in wild type mice the Bat-gal reporter showed no expression in the developing metanephric mesenchyme, or metanephric mesenchyme-derived structures, such as renal vesicles, S-shaped bodies, tubules, and glomeruli. This has been reported previously, and was interpreted to indicate the absence of canonical Wnt signaling in the metanephric mesenchyme

(Maretto et al., 2003). Results from other studies, however, argue for the presence of canonical Wnt signaling in the metanephric mesenchyme. For example, ureteric bud expression of Wnt1, thought to act through canonical Wnt signaling, can rescue Wnt9b mutants and induce nephrogenesis of the metanephric mesenchyme (Carroll et al., 2005). This suggests that the Bat-gal transgene might not accurately report canonical Wnt signaling in the metanephric

175 mesenchyme. To resolve this question we incubated E11.5 metanephric mesenchyme in LiCl, which activates canonical Wnt signaling through inhibition of GSK3, and also functions as an inducer of nephrogenesis in kidney organ culture. We observed that LiCl treated metanephric mesenchyme did undergo nephrogenesis, as expected, but failed to show Bat-gal expression (data not shown), suggesting that this transgene is not a true reporter of canonical Wnt signaling in the metanephric mesenchyme of the developing kidney.

Microarray analysis of Pygo1/Pygo2 null kidneys

To further examine possible perturbation of Wnt signaling in the metanephric mesenchyme of Pygo1/Pygo2 mutant kidneys we used microarrays to perform a global analysis of gene expression changes. E18.5 wild type and

Pygo1-/-/Pygo2-/- kidneys were examined in biological triplicate. The gene expression profiles of the wild type and mutant kidneys showed relatively few differences. Only 45 genes, from over 45,000 probe sets total, were identified as significantly changed using a relatively low stringency screen of the data, with a

P-value cutoff of < 0.05, and fold change >2 (Table 3). Notably, both Pygo1 and

Pygo2 were identified as significantly down regulated (-5.3 and -3.9 fold, respectively) in the mutant kidneys. Other genes with significant decreases in mutant kidneys included Cxcl13 (-2.3), Slc5a2 (-2.8), and Slco1a4 (-2.1). Slc5a2 is expressed in the proximal tubules of the adult nephron and was implicated in autosomal recessive renal glucosuria characterized loss of glucose uptake by the nephron (van den Heuvel et al., 2002). The organic anion transporter, Slco1a4, is also highly expressed in the tubules of the adult kidney (van Montfoort et al.,

176 2002). These results suggest that the Pygo1/Pygo2 genes might play a role in

nephron maturation and subsequent function.

Notable genes up regulated in the Pygo1/Pygo2 null kidney included Klk5

(2.8), Klk6 (2.7), Ren2 (2.2), and Timeless (2.4). Klk5 and Klk6 are members of

the Kallikrein family of trypsin-like serine proteases. Kallikreins have diverse

functions in cancer, tissue remodeling, and regulation of blood pressure

(Borgono et al., 2004). Ren2, a homolog of the endopeptidase Renin 1, activates

the renin-angiotensin system increasing blood pressure (Cheng et al., 2002).

Disruption of the renin-angiotensin system during development results in

congenital abnormalities of the ureter and collecting duct system (Niimura et al.,

2006). Timeless, a transcription factor involved in the regulation of circadian

rhythms, and expressed in the branching ureteric tips, has also been shown to

regulate ureteric branching morphogenesis (Li et al., 2000).

We were particularly interested in the possible altered expression of

known Wnt targets in the Pygo1-/-/Pygo2-/- mutant kidneys, as a measure of

disrupted Wnt signaling. A list of known Wnt targets was compiled from

http://www.stanford.edu/ ~rnusse/pathways/targets.html. Both Pygo1 and Pygo2

probe sets were included as references to illustrate significant down regulation.

Remarkably, the Wnt target genes showed very few differences in expression

between wild type and mutant kidneys (Fig. 8). Only the expression of Ccnd1 and

Wisp1 were significantly changed, with an expression level increase of less than

1.5-fold in Pygo1/Pygo2 null kidneys. This result was surprising since Ccnd1, encoding cyclin D1, as well as Wisp1, a Wnt-inducible protein, have been

177 identified as up-regulated targets of Wnt signaling (Shtutman et al., 1999; Xu et al., 2000). These results indicate an absence of dramatic changes in expression of known Wnt signaling target genes in the Pygo1/Pygo2 mutant kidneys.

We used quantitative real time PCR, with independent biological samples, to validate the microarray results. Nine genes were tested, and for seven we did observe a significant fold change in the same direction predicted by the microarrays (Table 4). It is not surprising that two of the nine genes did not validate, considering the relatively low stringency used in screening the microarray data.

178 Discussion

In Drosophila the Pygopus gene is a key mediator of Wnt signaling. In one study twelve distinct measures of Wg signaling in Pygo mutants were performed, including analysis of leg, wing and eye imaginal discs. In two cases there was a significant reduction of Wg signaling and in ten cases a complete block (Parker et al., 2002). A second study in Drosophila examined the effects of Pygo mutation on cuticle patterning, midgut constriction, CNS and cardiac development and concluded that “Pygo is an essential component in the Wg signal transduction pathway…” (Belenkaya et al., 2002).

Given these results in Drosophila it was surprising to see the relatively mild phenotypes of mice with targeted Pygo1, Pygo2, or double Pygo1/Pygo2

mutations. Pygo1 mutants were normal and fertile, while Pygo2 mutants

developed to birth and showed limited abnormalities. Furthermore, there was no

detected synergism in the double mutant phenotype. One possible explanation of

these unexpected results would be a failure of the gene targeting to eliminate

function of the Pygo1 and Pygo2 genes. The Pygopus deletion alleles described

in this report, however, are almost certainly functional nulls. In Drosophila it has

been shown that the PHD domain is absolutely required for Pygopus function in

Wg signaling. The PHD domain is 60 amino acids with seven cysteines and a

histidine, predicted to chelate two zinc ions. PHD domains are found in diverse

proteins, including transcription factors, and have been implicated in chromatin

remodeling and protein-protein interactions (Belenkaya et al., 2002). In

Drosophila the pygoF107 allele, with a single missense mutation converting amino

179 acid 802 in the PHD domain from cysteine to tyrosine, loses Wnt signaling function in both embryogenesis and imaginal disc development (Belenkaya et al.,

2002). The Pygo1/Pygo2 mutant alleles made in this report carried deletions of the entire PHD domains, as well as most other coding sequences. For the Pygo1 gene the coding region for 372 out of 417 total amino acids was deleted, and for the Pygo2 gene we deleted coding for 354 out of a total of 405 amino acids. It is therefore very unlikely that the relatively mild phenotypes observed were the result of hypomorphic mutations.

In this study we focused our phenotypic analysis on the developing kidney, where Wnt signaling has been shown to be of critical importance in several stages of nephrogenesis. Wnt9b is made by the ureteric bud and induces the metanephric mesenchyme to undergo nephrogenesis (Carroll et al., 2005).

Wnt4 is downstream of Wnt9b (Carroll et al., 2005), is made by the metanephric mesenchyme, and is also required for nephrogenesis (Stark et al 1994). In addition, Wnt11 is produced by the ureteric bud tips and induces GDNF expression in the metanephric mesenchyme (Majumdar et al, 2003).

In this report we describe a novel Wnt function in kidney development.

The Bat-gal transgene reporter showed the presence of canonical Wnt signaling in the ureteric bud and its derivatives in the developing kidney. Further, in the

Pygo2 mutants this signal was lost, indicating abrogation of Wnt signaling. In addition we observed a resulting reduction in ureteric tip density and altered morphology of the ureteric tree in mutants, indicating a role for canonical Wnt signaling in branching morphogenesis of the ureteric bud. While the simplest

180 interpretation is direct Wnt signaling to the ureteric bud, it remains possible that the observed abnormalities are the result of indirect effects, with altered Wnt signaling to the metanephric mesenchyme then perturbing mesenchyme to ureteric bud signaling.

We also observed in Pygo1/Pygo2 mutants an expansion of the zone of

thickened mesenchyme that caps the ureteric bud. Nevertheless, the mutant

metanephric mesenchyme formed nephrons normally. This was surprising, since

Wnt9b signaling from the ureteric bud to the metanephric mesenchyme is

thought to induce nephrogenesis via canonical Wnt signaling (Carroll et al.,

2005). There are two possible explanations. Either Wnt9b signaling is not

through the canonical Wnt pathway, and/or the Pygo1/Pygo2 genes are not

required for canonical Wnt signaling. The generally mild phenotype observed in

multiple organ systems of mutants would seem consistent with a reduced role for

Pygo1/Pygo2 in Wnt signaling in mammals. Nevertheless, it is interesting to

consider the evidence in support of a role for canonical Wnt signaling in kidney

development. It was shown that Wnt1 expression by the ureteric bud can rescue

the Wnt9b mutant phenotype, and Wnt1 is generally thought to function by

canonical Wnt signaling (Carroll et al., 2005). It should be noted, however, that

there is some evidence that Wnt1 can direct both canonical and noncanonical

signaling (Ziemer et al., 2001). In addition, the Wnt9b mediated stabilization of β-

catenin in responsive tissue culture cells was observed to be much weaker than

for well-studied canonical Wnt ligands (Carroll et al., 2005), again consistent with

possible Wnt9b noncanonical signaling.

181 It is also interesting that the Bat-gal transgene reporter failed to detect canonical Wnt signaling in the developing wild type metanephric mesenchyme.

The failure of the transgene to report signal even in LiCl treated metanephric mesenchyme however, which should have activated canonical Wnt signaling, argues that the Bat-gal reporter is not a reliable reporter in this tissue. The reason for this tissue specific inactivity is unclear, but it seems reasonable to suppose that not all Wnt targets (including this transgene reporter) would be activated in all tissues with canonical Wnt signaling. The ability of LiCl to induce nephrogenesis in metanephric mesenchyme might indicate a role for canonical

Wnt signaling, since LiCl has been shown to stabilize β-catenin through GSK3 inhibition. But LiCl has also been shown to activate Wnt signaling through noncanonical pathways (Le Floch et al., 2005). In summary, the evidence in favor of an important role for canonical Wnt signaling in kidney development is persuasive, but not conclusive. The original intent of the experiments of this study was to help clarify this situation. In mammals there are only two Pygopus

genes, unlike the large number of Frizzled receptors and Wnt ligands. Further, in

Drosophila the Pygopus gene is highly dedicated to canonical Wnt signaling,

unlike many other components of the pathway, such as β-catenin, which have

essential functions outside of Wnt signaling. The Pygopus genes in mammals

therefore appeared to represent a genetic bottleneck that could be exploited to

define canonical Wnt signaling function in the developing kidney.

The results in this report, however, indicate a striking evolutionary

divergence of Pygopus function between Drosophila and mammals. In

182 Drosophila the Pygopus gene is required for canonical Wnt signaling, while in mammals the Pygo1/Pygo2 genes play a smaller role in canonical Wnt signaling.

The Bat-gal transgene reporter of canonical Wnt signaling showed reduced but not absent signal in Pygo1/Pygo2 mutant embryos, with tissue specific variation in level of diminution. Further, the microarray results showed little evidence of altered expression of canonical Wnt targets in the Pygo1/Pygo2 mutant kidneys.

In addition, the kidneys were not unique in showing a mild phenotype in

Pygo1/Pygo2 mutants. Indeed, organogenesis generally proceeded without

detected abnormality, with few exceptions. These results suggest that the

mammalian Pygopus genes are mere modulators of canonical Wnt signaling, and

not essential components.

There is evidence, however, that even in mammals the Pygopus genes play a role in canonical Wnt signaling. For example the RNAi knockdown of hPygo2 in human embryonic kidney HEK 293 cells gives an approximately 80% reduction in canonical Wnt signal as measured by Topflash (Wnt reporter) response to LiCl (Mosimann et al., 2006). Conversely, overexpression of hPygo1 or hPygo2 in HEK 293 cells could stimulate canonical Wnt signaling by over 30 fold (Kramps et al., 2002). In yet another study in HEK 293 cells it was shown that Pygopus protein plays a role in the nuclear accumulation of β-catenin

(Krieghoff et al., 2006). Further, 82% of ovarian cancer samples showed elevated hPygo2 expression and knockdown of hPygo2 expression in cancer cell lines halted growth (Popadiuk et al., 2006).

183 In summary, the mammalian Pygopus genes are quantitative transducers of Wnt signaling. The HEK 293 cell culture results cited above, and the reduced

Bat-gal transgene reporter expression in the Pygo1/Pygo2 knockout mice described in this report, do indicate an evolutionarily conserved function in canonical Wnt signaling. In mammals, however, the phenotypic effects of

Pygopus mutation are much milder than in Drosophila. In the kidney, for example, we observed a canonical Wnt signaling function in driving normal branching morphogenesis of the ureteric bud, yet nephron formation occurred normally in Pygo1/Pygo2 mutants. Perhaps the simplest explanation is that in mammals other genes show partial functional redundancy with the Pygopus genes. The β-catenin transcription factor complex includes a large and growing

number of proteins (http://www.stanford.edu/~rnusse/wntwindow.html), some of

which may share the nuclear localization and/or transcription activation functions

of Pygo1/Pygo2 in mammals. The identities and roles of these Pygopus

redundant genes remain to be determined.

184 Acknowledgements

L. McClain provided essential technical assistance and we thank P. Groen and D. Witte for initial histology and embryological analyses. We thank T. Caroll for providing a Wnt9b riboprobe construct. We thank H. Liang for generation of

microarray data. This work was supported by grant DK61916 from the National

Institutes of Health (S.S.P.).

185 References

Adachi, S., Jigami, T., Yasui, T., Nakano, T., Ohwada, S., Omori, Y., Sugano, S.,

Ohkawara, B., Shibuya, H., Nakamura, T., and Akiyama, T. (2004). Role

of a BCL9-related beta-catenin-binding protein, B9L, in tumorigenesis

induced by aberrant activation of Wnt signaling. Cancer Res 64, 8496-

501.

Belenkaya, T. Y., Han, C., Standley, H. J., Lin, X., Houston, D. W., and

Heasman, J. (2002). pygopus Encodes a nuclear protein essential for

wingless/Wnt signaling. Development 129, 4089-101.

Bell, S. M., Schreiner, C. M., Waclaw, R. R., Campbell, K., Potter, S. S., and

Scott, W. J. (2003). Sp8 is crucial for limb outgrowth and neuropore

closure. Proc Natl Acad Sci U S A 100, 12195-200.

Borgono, C. A., Michael, I. P., and Diamandis, E. P. (2004). Human tissue

kallikreins: physiologic roles and applications in cancer. Mol Cancer Res

2, 257-80.

Carroll, T. J., Park, J. S., Hayashi, S., Majumdar, A., and McMahon, A. P. (2005).

Wnt9b plays a central role in the regulation of mesenchymal to epithelial

transitions underlying organogenesis of the mammalian urogenital system.

Dev Cell 9, 283-92.

Cheng, Z. J., Tikkanen, I., Vapaatalo, H., and Mervaala, E. M. (2002). Vascular

effects of COX inhibition and AT1 receptor blockade in transgenic rats

harboring mouse renin-2 gene. J Physiol Pharmacol 53, 597-613.

186 Davies, J. A., and Garrod, D. R. (1995). Induction of early stages of kidney tubule

differentiation by lithium ions. Dev Biol 167, 50-60.

Dressler, G. R. (2006). The Cellular Basis of Kidney Development. Annu Rev

Cell Dev Biol.

Du, S. J., Purcell, S. M., Christian, J. L., McGrew, L. L., and Moon, R. T. (1995).

Identification of distinct classes and functional domains of Wnts through

expression of wild-type and chimeric proteins in Xenopus embryos. Mol

Cell Biol 15, 2625-34.

Fan, Q. W., Kadomatsu, K., Uchimura, K., and Muramatsu, T. (1998).

Embigin/basigin subgroup of the immunoglobulin superfamily: different

modes of expression during mouse embryogenesis and correlated

expression with carbohydrate antigenic markers. Dev Growth Differ 40,

277-86.

Gray, P. A., Fu, H., Luo, P., Zhao, Q., Yu, J., Ferrari, A., Tenzen, T., Yuk, D. I.,

Tsung, E. F., Cai, Z., Alberta, J. A., Cheng, L. P., Liu, Y., Stenman, J. M.,

Valerius, M. T., Billings, N., Kim, H. A., Greenberg, M. E., McMahon, A. P.,

Rowitch, D. H., Stiles, C. D., and Ma, Q. (2004). Mouse brain organization

revealed through direct genome-scale TF expression analysis. Science

306, 2255-7.

Hedgepeth, C. M., Conrad, L. J., Zhang, J., Huang, H. C., Lee, V. M., and Klein,

P. S. (1997). Activation of the Wnt signaling pathway: a molecular

mechanism for lithium action. Dev Biol 185, 82-91.

187 Henderson, J. R., Macalma, T., Brown, D., Richardson, J. A., Olson, E. N., and

Beckerle, M. C. (1999). The LIM protein, CRP1, is a smooth muscle

marker. Dev Dyn 214, 229-38.

Kispert, A., Vainio, S., Shen, L., Rowitch, D. H., and McMahon, A. P. (1996).

Proteoglycans are required for maintenance of Wnt-11 expression in the

ureter tips. Development 122, 3627-37.

Kramps, T., Peter, O., Brunner, E., Nellen, D., Froesch, B., Chatterjee, S.,

Murone, M., Zullig, S., and Basler, K. (2002). Wnt/wingless signaling

requires BCL9/legless-mediated recruitment of pygopus to the nuclear

beta-catenin-TCF complex. Cell 109, 47-60.

Krieghoff, E., Behrens, J., and Mayr, B. (2006). Nucleo-cytoplasmic distribution of

beta-catenin is regulated by retention. J Cell Sci 119, 1453-63.

Mosimann, C., Hausmann, G., and Basler, K. (2006). Parafibromin/Hyrax

activates Wnt/Wg target gene transcription by direct association with beta-

catenin/Armadillo. Cell 125, 327-41.

Lan, Y., Ryan, R. C., Zhang, Z., Bullard, S. A., Bush, J. O., Maltby, K. M., Lidral,

A. C., and Jiang, R. (2006). Expression of Wnt9b and activation of

canonical Wnt signaling during midfacial morphogenesis in mice. Dev Dyn

235, 1448-54.

Le Floch, N., Rivat, C., De Wever, O., Bruyneel, E., Mareel, M., Dale, T., and

Gespach, C. (2005). The proinvasive activity of Wnt-2 is mediated through

a noncanonical Wnt pathway coupled to GSK-3beta and c-Jun/AP-1

signaling. Faseb J 19, 144-6.

188 Li, B., Mackay, D. R., Ma, J., and Dai, X. (2004). Cloning and developmental

expression of mouse pygopus 2, a putative Wnt signaling component.

Genomics 84, 398-405.

Li, Z., Stuart, R. O., Qiao, J., Pavlova, A., Bush, K. T., Pohl, M., Sakurai, H., and

Nigam, S. K. (2000). A role for Timeless in epithelial morphogenesis

during kidney development. Proc Natl Acad Sci U S A 97, 10038-43.

Livak, K. J., and Schmittgen, T. D. (2001). Analysis of relative gene expression

data using real-time quantitative PCR and the 2(-Delta Delta C(T))

Method. Methods 25, 402-8.

Majumdar, A., Vainio, S., Kispert, A., McMahon, J., and McMahon, A. P. (2003).

Wnt11 and Ret/Gdnf pathways cooperate in regulating ureteric branching

during metanephric kidney development. Development 130, 3175-85.

Maretto, S., Cordenonsi, M., Dupont, S., Braghetta, P., Broccoli, V., Hassan, A.

B., Volpin, D., Bressan, G. M., and Piccolo, S. (2003). Mapping Wnt/beta-

catenin signaling during mouse development and in colorectal tumors.

Proc Natl Acad Sci U S A 100, 3299-304.

Maurus, D., Heligon, C., Burger-Schwarzler, A., Brandli, A. W., and Kuhl, M.

(2005). Noncanonical Wnt-4 signaling and EAF2 are required for eye

development in Xenopus laevis. Embo J 24, 1181-91.

Niimura, F., Kon, V., and Ichikawa, I. (2006). The renin-angiotensin system in the

development of the congenital anomalies of the kidney and urinary tract.

Curr Opin Pediatr 18, 161-166.

189 Pandur, P., Lasche, M., Eisenberg, L. M., and Kuhl, M. (2002). Wnt-11 activation

of a non-canonical Wnt signalling pathway is required for cardiogenesis.

Nature 418, 636-41.

Parker, D. S., Jemison, J., and Cadigan, K. M. (2002). Pygopus, a nuclear PHD-

finger protein required for Wingless signaling in Drosophila. Development

129, 2565-76.

Parr, B. A., Cornish, V. A., Cybulsky, M. I., and McMahon, A. P. (2001). Wnt7b

regulates placental development in mice. Dev Biol 237, 324-32.

Patterson, L. T., Pembaur, M., and Potter, S. S. (2001). Hoxa11 and Hoxd11

regulate branching morphogenesis of the ureteric bud in the developing

kidney. Development 128, 2153-61.

Popadiuk, C. M., Xiong, J., Wells, M. G., Andrews, P. G., Dankwa, K., Hirasawa,

K., Lake, B. B., and Kao, K. R. (2006). Antisense suppression of pygopus2

results in growth arrest of epithelial ovarian cancer. Clin Cancer Res 12,

2216-23.

Pritchard-Jones, K., Fleming, S., Davidson, D., Bickmore, W., Porteous, D.,

Gosden, C., Bard, J., Buckler, A., Pelletier, J., Housman, D., and et al.

(1990). The candidate Wilms' tumour gene is involved in genitourinary

development. Nature 346, 194-7.

Rozen, S., and Skaletsky, H. (2000). Primer3 on the WWW for general users and

for biologist programmers. Methods Mol Biol 132, 365-86.

Saxen, L., and Sariola, H. (1987). Early organogenesis of the kidney. Pediatr

Nephrol 1, 385-92.

190 Schwab, K., Hartman, H. A., Liang, H. C., Aronow, B. J., Patterson, L. T., and

Potter, S. S. (2006). Comprehensive microarray analysis of

Hoxa11/Hoxd11 mutant kidney development. Dev Biol 293, 540-54.

Schwenk, F., Baron, U., and Rajewsky, K. (1995). A cre-transgenic mouse strain

for the ubiquitous deletion of loxP-flanked gene segments including

deletion in germ cells. Nucleic Acids Res 23, 5080-1.

Shtutman, M., Zhurinsky, J., Simcha, I., Albanese, C., D'Amico, M., Pestell, R.,

and Ben-Ze'ev, A. (1999). The cyclin D1 gene is a target of the beta-

catenin/LEF-1 pathway. Proc Natl Acad Sci U S A 96, 5522-7.

Shu, W., Jiang, Y. Q., Lu, M. M., and Morrisey, E. E. (2002). Wnt7b regulates

mesenchymal proliferation and vascular development in the lung.

Development 129, 4831-42.

Stadeli, R., and Basler, K. (2005). Dissecting nuclear Wingless signalling:

recruitment of the transcriptional co-activator Pygopus by a chain of

adaptor proteins. Mech Dev 122, 1171-82.

Stark, K., Vainio, S., Vassileva, G., and McMahon, A. P. (1994). Epithelial

transformation of metanephric mesenchyme in the developing kidney

regulated by Wnt-4. Nature 372, 679-83.

Thompson, B., Townsley, F., Rosin-Arbesfeld, R., Musisi, H., and Bienz, M.

(2002). A new nuclear component of the Wnt signalling pathway. Nat Cell

Biol 4, 367-73.

191 Townsley, F. M., Cliffe, A., and Bienz, M. (2004). Pygopus and Legless target

Armadillo/beta-catenin to the nucleus to enable its transcriptional co-

activator function. Nat Cell Biol 6, 626-33.

van den Heuvel, L. P., Assink, K., Willemsen, M., and Monnens, L. (2002).

Autosomal recessive renal glucosuria attributable to a mutation in the

sodium glucose cotransporter (SGLT2). Hum Genet 111, 544-7. van Montfoort, J. E., Schmid, T. E., Adler, I. D., Meier, P. J., and Hagenbuch, B.

(2002). Functional characterization of the mouse organic-anion-

transporting polypeptide 2. Biochim Biophys Acta 1564, 183-8.

Wodarz, A., and Nusse, R. (1998). Mechanisms of Wnt signaling in development.

Annu Rev Cell Dev Biol 14, 59-88.

Xu, L., Corcoran, R. B., Welsh, J. W., Pennica, D., and Levine, A. J. (2000).

WISP-1 is a Wnt-1- and beta-catenin-responsive oncogene. Genes Dev

14, 585-95.

Yoshino, K., Rubin, J. S., Higinbotham, K. G., Uren, A., Anest, V., Plisov, S. Y.,

and Perantoni, A. O. (2001). Secreted Frizzled-related proteins can

regulate metanephric development. Mech Dev 102, 45-55.

Ziemer, L. T., Pennica, D., and Levine, A. J. (2001). Identification of a mouse

homolog of the human BTEB2 transcription factor as a beta-catenin-

independent Wnt-1-responsive gene. Mol Cell Biol 21, 562-74.

192 Figures and tables

193

194 195

196

Figure 1. Pygo1/Pygo2 mutant embryos. (A) E18.5 Pygo1-/-/Pygo2+/-

embryos, with only one wild type Pygo2 allele, appeared normal. (B,C) double

homozygous mutant Pygo1-/-/Pygo2-/- embryos were smaller, with an eye defect including absent or rudimentary lens and folded pigmented retina (arrows). A small percentage of Pygo1/Pygo2 null embryos also displayed exencephaly (C).

197

Figure 2. Pygo1 and Pygo2 expression in the E18.5 developing kidney.

Immunofluorescence was used to determine expression patterns of the Pygo1 and Pygo2 proteins in the cortex of the E18.5 kidney. Both Pygo1 and Pygo2

(red) were localized in the nucleus, as expected and showed widespread expression, with elevated levels in the stromal cell compartment. Epithelial cells, primarily ureteric bud in these sections, were labeled blue using E-cadherin antibody.

198

Figure 3. Confocal analysis of Pygo1/Pygo2 mutant E18.5 kidneys. Pygo1-/-

/Pygo2-/- mutant E18.5 kidneys (B and D) and normal littermates kidneys (A and

B) were stained using antibodies to Cited1 (C and D) or Wt1 (A and B), both expressed in the condensing metanephric mesenchyme, and colored as red,

Cdh1, a general marker of epithelia (blue), and DBA lectin staining the ureteric tree (green). Confocal Z-sections were obtained every 5 microns for 75 to 80 microns. A decreased number of ureteric tips per area were observed in the

Pygo1/Pygo2 null kidneys (C and D) compared to control littermates (A and B).

In addition, ureteric bud tips were often more dilated in Pygo1/2 null mutants compared to controls.

199

Figure 4. Expression of Wnt7b, Wnt9b, and Wnt11 in E18.5 Pygo1/Pygo2 compound null kidneys. Whole mount in situ of hybridizations of E18.5 wild type kidneys (A - C) and Pygo1-/-/Pygo2-/- mutant kidneys (D - F) with the ureteric bud derivative markers: Wnt7b (A and D), Wnt9b (B and E), and Wnt11 (C and

F). The mutant kidneys showed normal expression patterns for the ureteric stalk markers Wnt7b and Wnt9b, and a reduced density of ureteric tips as measured by Wnt11. All images were taken at 32X.

200

Figure 5. Reduced canonical Wnt signaling in Pygo2 and Pygo1/Pygo2 mutant embryos. E10.5 embryos, all with the Bat-gal transgene reporter of canonical Wnt signaling. (A) Pygo1+/+/Pygo2+/- embryos showed normal X-gal staining in the developing brain, pharyngeal pouches, otic vesicle, apical ectodermal ridges of the fore and hind limb buds, and in somites. (B) Pygo1+/-

/Pygo2-/- embryos, with loss of both Pygo2 alleles, showed reduced but not absent Bat-gal reporter expression in many developing structures, including pharyngeal pouches, otic vesicle, and somites. (C) Pygo1-/-/Pygo2+/- embryo, with

mutation of both Pygo1 alleles, but one wild type Pygo2 allele, showed normal

Bat-gal expression. (D) Double homozygous mutant Pygo1-/-/Pygo2-/- embryos

201 Figure 5 continued. still showed some remaining Bat-gal expression, suggesting residual canonical

Wnt signaling. Embryos in panels A and B were from the same litter and were processed in parallel, while embryos in C and D were from a separate litter, also processed in parallel, and were slightly more developmentally advanced. A and

B were taken at 20X, while C and D were taken at 16X.

202

Figure 6. Pygo2 is required for Bat-gal reporter expression in ureteric bud-

derived structures of the developing kidney. X-Gal staining of Bat-gal

transgenic E10.5 (A - C), and E13.5 urogenital tracts (E-L), and E18.5 kidneys

(M-O). (A) Pygo -/-/Pygo2+/+ (Left) and Pygo1+/-/Pygo2-/- (Right). Note the loss of reporter activity in the nephric duct (black arrowhead) and reduction of staining in

203 Figure 6 continued. the ureteric bud (white arrow) of the Pygo2 null embryo (Right). (B) Pygo -/-

/Pygo2 +/- E10.5 embryo with Bat-gal reporter activity in the nephric duct (black

arrowhead) and ureteric bud (white arrow). (C) Pygo1-/-/ Pygo2 -/- embryo, with reporter expression lost in the nephric duct and reduced in the ureteric bud (white arrow). (D) Control background X-gal staining of a urogenital tract from an E13.5 embryo without the Bat-gal transgene (Non-Tg), showing absence of endogenous Beta-galactosidase activity. (E) Pygo +/-/Pygo2+/+, with Bat-gal

reporter activity in the ureteric compartment of the developing kidney, including the ureteric tips, ureteric tree, and ureter. (F) Pygo1+/-/ Pygo2 -/-, with marked reduction of Bat-gal reporter activity in the ureteric compartment. (G) Pygo +/-

/Pygo2 +/-, with reporter expression in the paramesonephric duct (white

arrowhead) and ureteric tree. (H,I) Pygo1-/-/Pygo2 +/-, with ventral view (H)

showing reporter expression in the paramesonephric duct (white arrowhead), and

dorsal view (I) showing ureteric tree expression in the kidney. (J) Pygo1 +/-/Pygo2

+/-, a control processed in parallel with K and L, with expression in ureteric tree and paramesonephric duct. (K,L) Pygo1 -/-/Pygo2 -/-, reporter activity was lost in the paramesonephric duct (white arrowhead), and in the ureter and ureteric compartment of the developing kidneys (dashed circles), except for a few faintly staining cells (K). (M,N) Pygo1+/-/Pygo2+/- (left) and Pygo1-/-/Pygo2-/- (right) E18.5 kidneys. The kidneys in N were bisected. Bat-gal reporter expression was seen in the ureteric tree components of the cortex and medulla of the double heterozygotes but was almost completely lost in the double homozygous

204 Figure 6 continued. mutants. (O) Pygo1+/-/Pygo2 +/+ (left), Pygo1-/-/Pygo2+/- (middle), and Pygo -/-/

Pygo2-/- (right) E18.5 kidneys. Reporter activity was present in the ureteric compartments of the Pygo1 +/-/Pygo2 +/+ (left) and Pygo1-/-/Pygo2 +/- (middle) kidneys, but lost in the Pygo1 -/-/Pygo2 -/- (right) kidney. Magnifications: (A-C)

32X, (D-F) 63X, (G-I) 40X, (J-L) 50X, (M and N) 10X, and (O) 12.5X.

205

Figure 7. Quantitative analysis of Bat-gal reporter expression in

Pygo1/Pygo2 E18.5 kidney extracts. Transgene specific Beta-galactosidase was quantified by ELISA analysis. Pygo1+/-/Pygo2+/+, Pygo1+/-/Pygo2+/- and

Pygo1+/-/Pygo2+/+ kidneys all showed similar levels of Bat-gal expression. In the

Pygo2 hetereozygote, Pygo1 homozygous mutants, however, reporter expression was reduced by about 50%. In Pygo2 homozygous mutants Bat-gal expression was uniformly low, with or without wild type Pygo1 alleles. Each experimental group included a sample size of at least four.

206

Figure 8. Gene expression changes of common Wnt signaling targets in the E18.5 Pygo1/Pygo2 null kidney. Microarray analysis was performed in triplicate on wild type and Pygo1/2 compound null E18.5 kidneys. Possible Wnt

207 Figure 8 continued. targets were selected from those compiled at http://www.stanford.edu/

~rnusse/pathways/targets.html. An initial gene list of 82 Wnt targets was created and then reduced to a total of 40 genes using an expression level restriction requiring the raw expression intensity to be greater than 100 in at least 3 samples. Pygo1 and Pygo2 probes were included to demonstrate significant

changes in expression levels.

208 Chapter 5

Microarray analysis of focal segmental glomerulosclerosis (FSGS) *

Kristopher Schwab2, David P. Witte3, Bruce J. Aronow1, Prasad Devarajan4,

S. Steven Potter2, and Larry T. Patterson4

Division of Bioinformatics1, Developmental Biology2, Pathology3, and

Nephrology4

Children's Hospital Medical Center

3333 Burnet Avenue

Cincinnati, OH 45229

* Published in Am J Nephrol. 2004 Jul-Aug;24(4):438-47.

209 Abstract

Background: Focal segmental glomerulosclerosis (FSGS) is a leading cause of chronic renal failure in children. Recent studies have begun to define the molecular pathogenesis of this heterogeneous condition. Here we use oligonucleotide microarrays to obtain a global gene expression profile of kidney biopsy specimens from patients with FSGS in order to better understand the pathogenesis of this disease. Methods: We extracted RNA from renal biopsy samples of 10 patients with the diagnosis of FSGS and from 5 control kidney samples, and produced labeled cRNA for hybridization to Affymetrix human

U133A microarrays. Results: We identified a gene expression fingerprint for

FSGS that contained 429 of 22,283 possible genes, each with a p < 0.01, using

RMA normalization, Welch t test, and at least a 1.8-fold change in 5 of the 10 patients examined. We also found gene expression differences in samples from subsets of patients who had either nephrotic syndrome or renal insufficiency.

This screen identified many genes and genetic pathways that have already been implicated in the pathogenesis of FSGS. In addition, we found changes in gene expression in genetic pathways that have not been studied in FSGS.

Conclusions: Oligonucleotide DNA microarray analysis of renal biopsy specimens identified a gene expression fingerprint in samples from a heterogeneous population of patients with FSGS. The genes and genetic pathways identified in this study can be compared to results of similar studies of other diseases to examine specificity and used to study the pathogenesis of

FSGS.

210 Introduction

Focal segmental glomerulosclerosis (FSGS) was initially described by

Rich [1] in 1957 among children who had died of nephrotic syndrome. The lesion is focal, with only some glomeruli affected, and segmental, with only restricted regions of a single glomerulus affected. FSGS represents the most common acquired cause of chronic kidney failure in children [2], and there is evidence that it is increasing in frequency [3–5]. The disease is particularly prevalent among

African Americans [6, 7]. The prognosis is generally considered poor, with a low response rate to treatment and a gradual progression to end-stage renal disease.

FSGS is a heterogeneous disease, with multiple causes and varied responses to therapy. There is evidence that the primary defect is often in the podocyte. Mutations in genes expressed in the podocyte, NPHS1 [8], NPHS2 [9] and ACTN4 [10], have been associated with FSGS. There is also evidence for a humoral substance in some patients causing rapid recurrence of proteinuria and

FSGS in newly transplanted kidneys. The typical first line of therapy for FSGS consists of treatment with prednisone, but approximately 75% of patients are steroid resistant.

In this report, we present the results of a microarray analysis of FSGS.

Microarrays allow the simultaneous determination of the expression levels of tens of thousands of genes. The resulting molecular fingerprints can distinguish between dissimilar diseased tissues and in theory could be used to define specific molecular subclasses of FSGS. The molecular fingerprints, then, could be correlated with therapeutic response or prognosis and be used to tailor

211 therapy. Microarrays can also provide deeper insight into the disease process by identifying novel pathways in the disease process. This in turn could suggest new therapeutic strategies.

212 Materials and Methods

Patient Selection and Biopsy Samples

We identified 10 patients (1.4–15.8 years of age) who had proteinuria and whose renal biopsy samples had shown FSGS. Renal biopsies had been performed because the patients’ clinical presentations were atypical for minimal change nephrotic syndrome. The primary reasons were age at onset or resistance to corticosteroid treatment. For histologic diagnosis, the tissue had been processed by standard procedures for light, immunofluorescence and electron microscopy. Archived, frozen (–80 ° C) tissues remaining after evaluation of renal pathology were processed for total RNA isolation. These specimens contained cortical tissue with glomeruli. Normal kidney controls were obtained from histologically normal tissue flanking surgically excised Wilms’ tumors. Care was taken to select cortical regions comparable to those represented in the needle biopsy specimens. Duplicate analysis was performed on 5 patient and 3 control tissue samples to assess RNA quality and microarray reproducibility. Procedures were conducted with approval from the Institutional

Review Board.

RNA Purification and Target Amplification

Total RNA was prepared from the FSGS needle biopsy samples using the

Stratagene Absolutely RNA Nanoprep kit after removing excess TissueTek. Five of the FSGS samples were bisected and the RNA separately purified and amplified. This tested reproducibility both for different sections of a needle biopsy and for the RNA purification and amplification procedures. Total RNA was

213 prepared from normal kidney tissue using RNAzol (Tel-Test, Friendwood, TX,

USA). Previous studies have shown no detectable microarray hybridization pattern differences resulting when comparing guanidinium isothiocyanate solution and silica column RNA purification procedures [11].

Microarray Analysis of Focal Segmental Glomerulosclerosis

Purified total RNAs were amplified using two rounds of in vitro transcription as previously described [12]. Total RNA (50 or 100 ng) for each sample was first reverse transcribed into cDNA using a T7 promoter-dT primer

[5’GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG-(T)24 ], then converted to double-stranded cDNA and amplified with an in vitro transcription reaction using T7 RNA polymerase. The products were then reverse transcribed into cDNA again, using random hexamer primers. A final in vitro transcription reaction using the Bioarray High Yield RNA transcript labeling kit (Enzo Life

Sciences, Farmingdale, NY) was performed producing biotinylated cRNA for microarray hybridization.

Gene Expression Profile Analysis

Amplified, biotinylated cRNA samples were hybridized to Affymetrix human U133A microarrays according to standard procedures as described by

Affymetrix. These microarrays contain 22,283 gene probe sets. Data were normalized using RMA (robust microarray analysis) Express [13]. Normalization is a mathematical technique used to reduce discrepancies in hybridization patterns that might result from variables in target amplification, hybridization conditions, staining or probe array lot. Normalizations standardize the data to

214 facilitate identification of genuine gene expression differences. GeneSpring 6

(Silicon Genetics) software was used to identify genes differently expressed in

FSGS samples. Data were further subdivided to identify genes differently expressed in samples from patients with hypoalbuminemia (serum albumin ! 2.5 g/dl) or with existing renal insufficiency (estimated creatinine clearance ! 80 ml/ min/1.73 m2).

215 Results

Patient Profile

The selected patients represented a diverse population that differed in clinical presentation and pathological findings (table 1). Importantly, patients had variable degrees of hypoalbuminemia and renal dysfunction. Also, variable degrees of interstitial inflammation and foot process fusion were noted on renal biopsy, but tubular atrophy was nearly universal.

Microarray Quality

To control for data quality, we performed multiple replicates and examined the resulting data for reproducibility and indications of initial RNA integrity. Five of the

10 patient samples were analyzed in replicate, by bisecting each needle biopsy and processing both halves in parallel. Similarly, the cortical region of 3 control samples was examined in biological replicate after dividing the original sample.

The resulting microarray data showed a high level of reproducibility. The scattergrams of figure 1 illustrate the high correlation between gene expression levels of replicates. Perfect reproducibility would give a single 45° diagonal line, with each gene showing the same hybridization intensity in each replicate analysis of a single sample. Figure 1a shows the results of a technical replicate, where two 100-ng aliquots of a single RNA sample were processed in parallel and hybridized to microarrays. A comparison of the resulting patterns showed

284 genes with an over two-fold difference in hybridization intensity, 47 genes with an over threefold difference and 18 genes with an over four-fold difference.

Figure 1b shows a scattergram of a biological replicate, where a single needle

216 biopsy specimen was bisected and the two halves processed in parallel. In this case, biological noise was added to technical noise resulting in somewhat greater variation, with 508, 87 and 44 genes showing two-, three- and four-fold changes, respectively. It is important to note that this technical and biological noise is largely random and is almost entirely eliminated by requiring that changes be reproducible. The scattergram in figure 1c shows the large number of gene expression differences observed in comparing normal and FSGS tissue.

The microarray data also provide a measure of the quality of the starting RNA.

Degraded RNA results in premature termination of the initial reverse transcription reaction at the positions of the RNA breaks. This in turn gives relative overrepresentation of the 3’ ends of mRNA sequences and false difference calls when comparing microarrays with different degrees of RNA degradation. The

Affymetrix gene chips include quality control probe set arrays to monitor relative

3’ to 5’ representations. We observed excellent 3’ to 5’ ratios (1.4–3), indicating high quality target amplification.

Altered Gene Expression Patterns in FSGS Kidneys

Is microarray analysis of needle biopsy tissue capable of detecting distinct gene expression patterns in FSGS kidneys? The primary lesion is in the glomerulus, a relatively small proportion of total kidney tissue. In addition, only a fraction of glomeruli are affected. Furthermore, variation in the needle biopsy sampling might yield differences in the representation of affected glomeruli. And finally, FSGS is known to be a heterogeneous disease, raising the possibility that although the microscopic appearances of FSGS kidneys are similar, underlying

217 molecular events might vary considerably from one patient to the next. In this case, there might not be consistent changes in gene expression associated with all FSGS kidneys.

In light of the above considerations, it was interesting to find a dramatic and consistent microarray signature for FSGS. The gene expression profiles of all FSGS patients were compared to controls. Performing a one-way ANOVA parametric test, with the p value set to 0.01 and using the Benjamini and

Hochberg false discovery rate correction, 2,763 genes were found to differ significantly in levels of expression when FSGS tissue was compared to control.

A series graph illustrating the FSGS and control kidney expression levels for these genes showed considerable consistency (fig. 2a). By limiting the list of genes to those with a p < 0.001, we identified 534 genes (fig. 2b). These series graphs provide a visual display of the distinct FSGS gene expression profiles defined by the hybridization patterns of all 23 Affymetrix U133A microarrays used in this study.

The average of replicate pairs was used for further analysis of the microarray data to give equal weight for each patient and control sample. Four hundred and twenty-nine genes (221 upregulated and 208 downregulated) were identified using three filters, the Welch t test (p < 0.01), the Benjamini and

Hochberg false discovery rate correction, and a greater than 1.8-fold change in at least 5 out of 10 patients (fig. 3b). The results demonstrate that microarrays identify a distinct gene expression profile in renal biopsies of patients with FSGS

(see Appendix for Web access to supplemental data).

218 Functional Pathways Altered in FSGS

FSGS Upregulated

Cell Cycle and Proliferation. Several interesting patterns of altered gene expression patterns emerged. For example, 15 genes associated with cell proliferation were upregulated in FSGS. These included proliferating cell nuclear antigen, which was previously noted to be upregulated in FSGS [15], and the cell cycle-associated genes CDC34, CDKN2C, CDK5RAP3 and CCNG2 (cyclin G2).

Also upregulated were the DNA replication and cell division associated genes

ORC5L, PRC1 (protein regulating cytokinesis), the SMC5L gene associated with chromosome segregation, CETN3 (centriole-associated centrin 3), PCNT1

(centrosomal pericentrin), and PSPHL, a regulator of cell proliferation.

Immune Response. IFI27 and IFI44 (interferon-inducible protein 27 and

44), 2 genes associated with response to viral infection, were upregulated,

consistent with reports associating viral infections and FSGS [16, 17].

Receptors and Ligands. One gene of interest, neuralin 1, which is a

chordin-like inhibitor of BMP signaling, was upregulated in FSGS. TGFβ

superfamily signaling has been well studied in FSGS and in renal development,

so it will be interesting to learn to what degree this gene can modulate the

signaling pathway. Another gene of particular interest observed to be

upregulated in FSGS was Nov. This gene encodes a growth factor that is also

overexpressed in nephroblastoma and is negatively regulated in podocytes by

WT-1 [18]. This suggests that the loss of WT-1 expression in FSGS podocytes,

observed in this report and elsewhere, could lead to enhanced Nov expression.

219 IGFBP6, another gene with growth factor activity, was also expressed at higher

levels in FSGS kidneys. Also elevated in the FSGS kidney was PTGER3, a

prostaglandin receptor, which is normally expressed in the thick ascending limb

of the loop of Henle [19]; IL18 (IL1-Α), which encodes a ligand for the IL1

receptor; ELTD1, a latrophilin-related, G-protein coupled receptor, and ITB3BP

(integrin ί3 binding protein coreceptor), which has been implicated in the

downregulation of the urokinase-type plasminogen activator receptor [20].

Transcription Factors. Eleven transcription factor genes, including RB1

(retinoblastoma 1), MYCBP (MYC binding protein), NMI (N-Myc interactor), the

general transcription factors TAF11 and GTF2E2 were also upregulated, as were

other less well characterized transcription factors.

Other. Ten genes involved in RNA processing or splicing showed

increased expression. In addition several protease inhibitor genes were

activated, including SPINK1, RECK, SLP1, and LXN. Activation of another

protease inhibitor, plasminogen activator inhibitor type-1, has been described

and thought to be associated with the pathogenesis of FSGS [21]. Further, there

were 23 FSGS upregulated genes encoding signaling proteins, mostly GTP

coupled. Finally of note, SMURF2, an ubiquitin ligase that targets the

degradation of SMAD2, was upregulated and like neuralin 1 would be expected

to modulate TGFβ signaling.

FSGS Downregulated

Proteases. Several interesting functional groupings of genes were

downregulated in FSGS. Seven proteases (FLJ10613, KIAA0570, plasminogen,

220 tripeptidyl aminopeptidase, cathepsin L2, dipeptidyl peptidase) and one protease inhibitor (Serpin A1) were downregulated. This is consistent with the upregulation of multiple protease inhibitors in FSGS noted above, and suggests a general reduction of protease activity. This has been a relatively unexplored area in the pathogenesis of FSGS.

Receptors and Ligands. Thirteen receptor genes showed reduced expression in FSGS. Two of these, the activin receptor ALK4 and the TGFβ receptor TGFBR3, are receptors of the TGFβ receptor superfamily, a family important in FSGS. Another downregulated receptor gene of note was the somatostatin receptor 2. It is interesting to note one report in which treatment of a pituitary adenoma with the somatostatin analogue, octreotide acetate, reduced proteinuria in the patient who also had FSGS [22]. A similar beneficial effect was observed for somatostatin treatment of chronic renal failure in subtotal nephrectomized rats [23]. Other receptors with reduced expression in FSGS included EGF receptor, insulin receptor, vitamin D receptor, PPAR-α·, CUBN, and the orphan nuclear receptors NR4A1 and MINOR.

Transcription Factors. The FSGS samples showed 25 downregulated transcription factor genes, including Jun, Fos, Foxo1A, Hmx1, Egr1, Bhlhb2,

Dux4, Tcf8, and Cutl1. A complete listing is available in the Web supplement

(see Appendix).

Transporters. It is also interesting to note that 15 genes encoding proteins with transporter function were down-regulated in FSGS, while essentially no transporter genes in the 429 gene list were upregulated. A few transporters with

221 reduced expression included aquaporin 7, sodium/ nucleoside cotransporter, organic ion transporter 3, potassium chloride cotransporter KKCC3, L-type amino acid transporter 2, and other transporters associated with renal tubule function.

These changes in transporter expression may be associated with the tubular dysfunction that frequently accompanies FSGS [24].

Other. Nonmuscle myosin IIA (MYH9) was significantly reduced in expression in FSGS. Mutation of MYH9 has been associated with Fechtner syndrome, podocyte abnormalities and proteinuria [25]. Finally, while 15 genes

associated with cell proliferation were upregulated, only 1, CIR61, was

downregulated, showing internally consistent results with a trend toward

upregulation in genes involved in cell proliferation.

Gene Expression Patterns Associated with Nephrotic Syndrome

The above data demonstrate the presence of a clear gene expression

signature for FSGS. We next asked if the different clinical subgroups of FSGS

patients would show distinct gene expression patterns. Microarray data for

patients with and without hypoalbuminemia were compared. Differences were

found, but far fewer than those observed for FSGS versus control comparisons.

A one-way ANOVA parametric test using a p value cutoff of 0.01 produced a list

of 527 genes with different expression between nephrotic and nonnephrotic

kidneys. Using a Welch t test (p < 0.01) combined with a greater than 1.8-fold

change in at least 2 out of 4 patients, we reduced the list to 151 genes, with 80

showing reduced and 71 showing elevated expression in nephrotic samples (fig.

3c).

222 Nephrotic Syndrome Upregulated

Extracellular Matrix. Thrombospondin expression was elevated in nephrotic kidneys. Thrombospondin has been associated with activation of TGFβ signaling, and subsequent extracellular matrix (ECM) synthesis in fibrotic renal disease. Consistent with an elevation of ECM synthesis pathways, we also observed increased expression of the collagen genes Col1A1, Col1A2, Col3A1,

Col4A1 and Col4A2. It is of particular interest to note that increased TGFβ signaling in vivo has previously been shown to induce aberrant collagen type IV isotype (4A1 and 4A2) synthesis in the glomerular basement membrane of the kidney [26]. We also observed increased expression of the versican gene in nephrotic samples. Versican is an ECM component synthesized by mesangial cells, and may represent a ligand for L-selectin [27].

Receptors and Ligands. Ligands, TNSF10 and TNSF15, and a receptor,

TNFRSF1A, of the TNF pathway were upregulated. Several studies have connected TNF-· and nephrotic syndrome one of which showed elevated serum levels of TNF-· in nephrotic FSGS patients [28].

Immune Response. We found increased expression of 7 genes from the general category of immune response. Two of these, GBP1 and IFI16, are genes known to be induced by interferon, and 2 are components of the complement pathway

(C1QB and C2). Another gene of some interest was CD163, a macrophage marker antigen that may represent detection of transdifferentiated epithelial cells as found in other studies [29]. Alternatively, it could simply represent invasion of cells of the mononuclear phagocyte system.

223 Nephrotic Syndrome Downregulated

Transcription Factors. Interesting genes with previous connections to kidney development and disease showed reduced expression in nephrotic

FSGS. One downregulated transcription factor gene was GATA3.

Haploinsufficiency of this gene causes HDR syndrome, which includes renal

malformations [30]. Other transcription factor genes with reduced expression

were HEY1, ZNF204 and PKNOX1.

Extracellular Matrix. Of interest, we found no reduction in expression of

ECM genes in the nephrotic samples as compared to 7 that had increased

expression.

Immune Response. Only 3 genes in the immunity functional class showed

lower expression. One of these, β-defensin-1, is expressed in the collecting ducts

and the loops of Henle [31]. Defensins are part of the innate immune system,

have antimicrobial activity, and have been shown to be downregulated by TNFα

[32]. The transferrin receptor gene also showed lowered expression. The

transferrin receptor is expressed by mesangial cells, and has been suggested to

also provide immunoglobulin IgA receptor function [33].

Bioactive Peptides. Several genes of the kallikrein pathway, kallikrein (KLK1),

kininogen (KNG), and kallistatin (KAL), were significantly downregulated in

samples from patients with nephrotic syndrome. These genes play an important

role in the regulation of vasodilation/constric-tion and in the control of sodium

secretion.

Altered Gene Expression Profiles with Renal Insufficiency

224 We performed a similar comparison of microarray hybridization patterns from patients without renal failure to those who had renal insufficiency at the time of renal biopsy. The analysis showed significant changes in gene expression associated with renal insufficiency, but similar to nephrotic syndrome, far fewer than seen in the comparison of FSGS to controls. Using statistical criteria similar to those described for the nephrotic comparison, we identified 72 upregulated and 93 downregulated genes in renal failure (fig. 3d).

Renal Insufficiency Upregulated

Cell Death. Only 4 upregulated genes were directly related to cell death or

survival. Two of these, clusterin and TOSO, a regulator of Fas-induced

apoptosis, are associated with improved cell survival, while the other two,

caspase 1 and caspase 2, are key mediators of apoptosis. Caspase 1 has also

been associated with an inflammatory response.

Receptors and Ligands. Only 4 receptors showed upregulation, and

interestingly 3 of them (OPRM1, CCR7, and CHRM2) belong to the rhodopsin-

like GPCR superfamily. The CCR7 receptor is expressed in mesangial cells, while its ligand SLC is expressed in podocytes. The chemokine receptor has been implicated in mesangial cell proliferation, migration and regeneration [34], but has not yet been studied in FSGS. The CHRM2 muscarinic acetylcholine receptor is expressed in MDCK and HEK293 cells and has been associated with the regulation of intracellular calcium levels. The fourth receptor, discoidin

domain receptor, has been shown to regulate the growth and adhesion of

mesangial cells [35].

225 IL1RN, which encodes an interleukin 1 receptor antagonist, was upregulated as well. Interleukin 1 strongly promotes inflammation, and this antagonist can reduce the inflammatory response.

Extracellular Matrix. Laminin-5 (LAMC2, kalinin) expression was upregulated in FSGS patients with renal insufficiency. It has been associated with wound repair and has been shown to be upregulated in renal ischemic injury and repair [36].

Signal Transduction. Seven genes involved in signal transduction were upregulated. One of these, PKD2, is mutated in polycystic kidney disease.

Others included TNK1, a nonreceptor tyrosine kinase, MAPK3, a mitogen- activated protein kinase, CISH, a cytokine inducer of signaling, PRKG1, a c-

GMP-dependent protein kinase, CABP5, a calcium-binding protein, and GNG4, a

receptor-coupled GTPase.

Renal Insufficiency Downregulated

Cell Death. There were also several genes of interest downregulated in

renal insufficiency. Three were related to cell death, including Bcl-2-related

MCL1, TIA1, and GADD45B, which has been associated with the hypertonic

stress phenotype of renal inner medullary cells [37].

Extracellular Matrix. Several genes encoding ECM and cell surface

proteins were reduced in expression. Among these was NPHS1, which can result in FSGS when mutated. Also repressed were glypican 3, an ECM component involved in Wnt signaling, and syndecan 2, which is expressed in mesangial cells and can also modulate growth factor signaling.

226 Bioactive Peptides. Of interest, several genes of the kallikrein system were also downregulated, similar to what was previously noted for nephrotic patients, only different components. For renal insufficiency, there was repression of kininogen KLKB1, kallikrein precursor KLK7, and kallikrein 6 preprotein KLK6.

Receptors and Ligands. Seven genes in the receptor and ligand category were repressed in renal insufficiency. This gene list included pleiotrophin, previously identified as a mesenchymally derived factor that drives branching morphogenesis [38], natriuretic peptide receptor C (NPR3), pleckstrin-like insulin receptor substrate 2 (IRS2), PDGF receptor-like (PDGFRL), endothelial specific selectin ligand (EMCN), tyrosine kinase receptor B (NTRK2) and orphan nuclear receptor NR4A1.

RNA Splicing. Also repressed were 6 genes involved in RNA splicing

(SFRS3, SFRS11, SFRS12, SFPQ, SRPK2, and LUC7A), suggesting large-scale changes in RNA processing.

Signal Transduction. Nine of the downregulated genes in renal insufficiency fell into the functional category of signal transduction. This gene list included mitogen activated kinase kinase kinase 5 (MAPK4K5), CDC-like kinase

(CLK1), c-AMP-dependent protein kinase (PRKACB), rho-associated coiled-coil

protein kinase (ROCK1), calmodulin1 (phosphorylase kinase, CALM1), rho-

GTPase binding CDC42EP3, GTP exchange factor AKAP13, C8FW and

phosphoinositol 3-phosphate-binding protein-2 (KIAA1686).

Transcription Factor. Finally, 12 genes involved in the regulation of

transcription showed lower transcript abundance. Several of these encoded zinc

227 finger transcription factors. Of particular interest are tumor suppressor genes

(WT1 and PLAGL1). WT1 encodes a zinc finger transcription factor that is

normally expressed in podocytes and is associated with mesangial sclerosis

when mutated. PLAGL1 also encodes a zinc finger protein and is involved in the

p53-mediated induction of apoptosis. A third zinc finger transcription factor gene,

TncRNA (MEA1), has been predicted to be a tumor suppressor. Another zinc

finger transcription factor gene, Jumonji (JMJD1) was reduced in renal failure, as

was EGR1, which is predicted to encode a zinc finger transcription factor. Two

other reduced expression genes were the homeobox genes HOXA9 and MEIS2.

228 Discussion

Microarray studies of normal and diseased tissues can provide a rapid and complete gene expression fingerprint that is not possible by other standard practices. In this report, we demonstrate that the FSGS kidney generates a distinct molecular signature when analyzed by microarrays. Thousands of genes were identified that had altered expression levels and represented multiple compartments within the kidney. This procedure offers promise for molecular classification of FSGS types, for future correlation of molecular class with therapeutic response, and for discovery of molecular mechanisms involved in

FSGS disease progression.

It was not apparent a priori that microarray analysis of FSGS needle biopsy samples would be informative. The primary lesion is limited to glomeruli, which represent a small fraction of cortical tissue. Furthermore, only a fraction of glomeruli are affected, and it would be expected that different biopsies might by chance carry different numbers of glomeruli. It was therefore somewhat surprising to find that the FSGS kidneys consistently gave a very unique gene expression profile, easily distinguished from control kidneys. The large number of differently expressed genes in FSGS samples suggests that widespread changes occurred throughout the kidney well beyond the lesioned glomerulus. This is supported by the detection of genes that are specific for tubules (i.e. transporter genes) in addition to those for the glomerulus (i.e. NPHS1). It is therefore unlikely that all changes in gene expression are specific for FSGS. Only with further study

229 of glomerular disease will we be able to identify a specific FSGS gene expression profile.

A significant fraction of the genes found in this microarray analysis had been previously implicated in kidney disease and specifically in glomerular lesions (WT1, NPHS1, and MYH9). This serves to validate the approach and suggests that the results provide an important gene discovery function, perhaps identifying new genetic elements important in the cause and progression of

FSGS.

It is also important to note the strong internal consistency of changes in expression that help validate this approach. Changes in expression of genes involved in one molecular pathway often increased or decreased together and were not randomly distributed.

This study serves to justify a more exhaustive microarray analysis of

FSGS and other glomerular diseases. The technology is robust, high quality data can be generated, and there is considerable reproducibility in the data generated

from the two parts of a bisected single needle biopsy. FSGS samples give gene

expression profiles clearly distinguishable from normal kidney samples. Further,

the results suggest considerable power in defining molecular subtypes of FSGS.

Although the patient series in this study was limited in size, distinct gene

expression fingerprints emerged for nephrotic versus nonnephrotic patients, and

renal insufficiency versus nonrenal failure patients. The results suggest that in

the future more exhaustive libraries of FSGS microarray data will allow a detailed

categorization of patients that can then be related to therapeutic response.

230 Appendix

The interactive Web site containing microarray data can be found at http://genet.chmcc.org/. Log in as Guest and select HGU133. The interactive gene trees can be accessed in the Experiments folder under

FSGS_SchwabEtAl_2004. Select FSGS_SchwabEt-Al_2004, select graph from the pull-down menu, and enter display to obtain a graph of the data. Gene lists are found by sequentially opening the folders Gene Lists,

FSGS_Schwab_EtAl_2004, and 92403. The master tree is found by sequentially opening the folders Gene Trees, SPotter, and FSGS. Data sets are available for download and use with GeneSpring or similar software. We will also provide electronic data upon request.

231 Acknowledgements

We thank Rachel Veldkamp and Jean Snyder for assistance in the initial phases of this project. This work was supported by grants DK61916 (S.P.) and

DK02702 (L.P.). Reprint requests to Larry Patterson, MD, Nephrology and

Hypertension, Cincinnati Children’s Hospital Medical Center, 3333 Burnet

Avenue, Cincinnati, Ohio 45229, USA; E-Mail [email protected].

232 References

1 Rich AR: A hitherto undescribed vulnerability of the juxtamedullary glomeruli in

lipoid nephrosis. Bull Johns Hopkins Hosp 1957;100: 173–186.

2 Schnaper HW: Idiopathic focal segmental glomerulosclerosis. Semin Nephrol

2003;23:183– 193.

3 Haas M, Meehan SM, Karrison TG, Spargo BH: Changing etiologies of

unexplained adult nephrotic syndrome: A comparison of renal biopsy findings

from 1976–1979 and 1995– 1997. Am J Kidney Dis 1997;30:621–631.

4 Bonilla-Felix M, Parra C, Dajani T, Ferris M, Swinford RD, Portman RJ, Verani

R: Changing patterns in the histopathology of idiopathic nephrotic syndrome in

children. Kidney Int 1999;55:1885–1890.

5 Srivastava T, Simon SD, Alon US: High incidence of focal segmental

glomerulosclerosis in nephrotic syndrome of childhood. Pediatr Nephrol

1999;13:13–18.

6 Bakir AA, Bazilinski NG, Rhee HL, Ainis H, Dunea G: Focal segmental

glomerulosclerosis. A common entity in nephrotic black adults. Arch Intern

Med 1989;149:1802–1804.

7 Tucker JK: Focal segmental glomerulosclerosis in African Americans. Am J

Med Sci 2002;323: 90–93.

8 Kestila M, Lenkkeri U, Mannikko M, Lamer-din J, McCready P, Putaala H,

Ruotsalainen V, Morita T, Nissinen M, Herva R, Kashtan CE, Peltonen L,

Holmberg C, Olsen A, Tryggvason

233 K: Positionally cloned gene for a novel glomerular protein – nephrin – is

mutated in congenital nephrotic syndrome. Mol Cell 1998;1:575– 582.

9 Boute N, Gribouval O, Roselli S, Benessy F, Lee H, Fuchshuber A, Dahan K,

Gubler MC, Niaudet P, Antignac C: NPHS2, encoding the glomerular protein

podocin, is mutated in autosomal recessive steroid-resistant nephrotic

syndrome. Nat Genet 2000;24:349–354.

10 Kaplan JM, Kim SH, North KN, Rennke H, Correia LA, Tong HQ, Mathis BJ,

Rodriguez-Perez JC, Allen PG, Beggs AH, Pollak MR: Mutations in ACTN4,

encoding alpha-actinin-4, cause familial focal segmental glomerulosclerosis.

Nat Genet 2000;24:251–256.

11 Ojaniemi H, Evengard B, Lee DR, Unger ER, Vernon SD: Impact of RNA

extraction from limited samples on microarray results. Biotechniques

2003;35:968–973.

12 Baugh LR, Hill AA, Brown EL, Hunter CP: Quantitative analysis of mRNA

amplification by in vitro transcription. Nucleic Acids Res 2001;29:E29.

13 Bolstad BM, Irizarry RA, Astrand M, Speed TP: A comparison of normalization

methods for high density oligonucleotide array data based on variance and

bias. Bioinformatics 2003;19:185–193.

14 Schwartz GJ, Brion LP, Spitzer A: The use of plasma creatinine concentration

for estimating glomerular filtration rate in infants, children, and adolescents.

Pediatric Clinics of North America 1987;34:571–590.

15 Geleilete TJ, Costa RS, Dantas M, Coimbra TM: Alpha-smooth muscle actin

and proliferating cell nuclear antigen expression in focal segmental

234 glomerulosclerosis: Functional and structural parameters of renal disease

progression. Braz J Med Biol Res 2001;34:985–991.

16 Moudgil A, Nast CC, Bagga A, Wei L, Nurmamet A, Cohen AH, Jordan SC,

Toyoda M: Association of parvovirus B19 infection with idiopathic collapsing

glomerulopathy. Kidney Int 2001;59:2126–2133.

17 Tanawattanacharoen S, Falk RJ, Jennette JC, Kopp JB: Parvovirus B19 DNA

in kidney tissue of patients with focal segmental glomerulosclerosis. Am J

Kidney Dis 2000;35:1166– 1174.

18 Martinerie C, Chevalier G, Rauscher FJ 3rd, Perbal B: Regulation of nov by

WT1: A potential role for nov in nephrogenesis. Oncogene 1996;12:1479–

1492.

19 Breyer MD, Davis L, Jacobson HR, Breyer RM: Differential localization of

prostaglandin E receptor subtypes in human kidney. Am J Physiol

1996;270:F912–F918.

20 Hapke S, Gawaz M, Dehne K, Kohler J, Marshall JF, Graeff H, Schmitt M,

Reuning U, Lengyel E: beta(3)A-integrin downregulates the urokinase-type

plasminogen activator receptor (u-PAR) through a PEA3/ets transcriptional

silencing element in the u-PAR promoter. Mol Cell Biol 2001;21:2118–2132.

21 Hamano K, Iwano M, Akai Y, Sato H, Kubo A, Nishitani Y, Uyama H, Yoshida

Y, Miyazaki M, Shiiki H, Kohno S, Dohi K: Expression of glomerular

plasminogen activator inhibitor type 1 in glomerulonephritis. Am J Kidney Dis

2002;39:695–705.

22 Yoshida H, Akikusa B, Saeki N, Hasegawa S, Iesato K, Yamamoto S,

235 Murotani N, Wakashin M, Shimada T: Effect of pituitary microsurgery on

acromegaly complicated nephrotic syndrome with focal segmental

glomerulosclerosis: Report of a rare clinical case. Am J Kidney Dis

1999;33:1158–1163.

23 Uemasu J, Tokumoto A, Godai K, Kawasaki H, Hirayama C: Effects of chronic

administration of somatostatin analogue SMS 201-995 on the progression of

chronic renal failure in subtotal nephrectomized rats. Exp Clin Endocrinol

1990;96:97–104.

24 Futrakul P, Yenrudi S, Futrakul N, Sensirivatana R, Kingwatanakul P,

Jungthirapanich J, Cherdkiadtikul T, Laohapaibul A, Watana D, Singkhwa V,

Futrakul S, Pongsin P: Tubular function and tubulointerstitial disease. Am J

Kidney Dis 1999;33:886–891.

25 Ghiggeri GM, Caridi G, Magrini U, Sessa A, Savoia A, Seri M, Pecci A,

Romagnoli R, Gangarossa S, Noris P, Sartore S, Necchi V, Ravazzolo R,

Balduini CL: Genetics, clinical and pathological features of glomerulonephritis

associated with mutations of nonmuscle myosin IIA (Fechtner syndrome). Am

J Kidney Dis 2003;41:95–104.

26 Chai Q, Krag S, Miner JH, Nyengaard JR, Chai S, Wogensen L: TGF-beta1

induces aberrant laminin chain and collagen type IV isotype expression in the

glomerular basement membrane. Nephron Exp Nephrol 2003;94:e123– e136.

27 Kawashima H, Li YF, Watanabe N, Hirose J, Hirose M, Miyasaka M:

Identification and characterization of ligands for L-selectin in the kidney. 1.

Versican, a large chondroitin sulfate proteoglycan, is a ligand for L-selectin. Int

236 Immunol 1999;11:393–405.

28 Suranyi MG, Guasch A, Hall BM, Myers BD: Elevated levels of tumor necrosis

factor-alpha in the nephrotic syndrome in humans. Am J Kidney Dis

1993;21:251–259.

29 Bariety J, Bruneval P, Hill GS, Mandet C, Jacquot C, Meyrier A:

Transdifferentiation of epithelial glomerular cells. J Am Soc Nephrol

2003;14:S42–S47.

30 Van Esch H, Groenen P, Nesbit MA, Schuffenhauer S, Lichtner P,

Vanderlinden G, Harding B, Beetz R, Bilous RW, Holdaway I, Shaw NJ, Fryns

JP, Van de Ven W, Thakker RV, Devriendt K: GATA3 haplo-insufficiency

causes human HDR syndrome. Nature 2000;406: 419–422.

31 Schnapp D, Reid CJ, Harris A: Localization of expression of human beta

defensin-1 in the pancreas and kidney. J Pathol 1998;186:99– 103.

32 Nitschke M, Wiehl S, Baer PC, Kreft B: Bactericidal activity of renal tubular

cells: The putative role of human beta-defensins. Exp Nephrol 2002;10:332–

337.

33 Moura IC, Centelles MN, Arcos-Fajardo M, Malheiros DM, Collawn JF,

Cooper MD, Monteiro RC: Identification of the transferrin receptor as a novel

immunoglobulin (Ig)A1 receptor and its enhanced expression on mesangial

cells in IgA nephropathy. J Exp Med 2001; 194:417–425.

34 Banas B, Wornle M, Berger T, Nelson PJ, Cohen CD, Kretzler M, Pfirstinger

J, Mack M, Lipp M, Grone HJ, Schlondorff D: Roles of SLC/CCL21 and CCR7

in human kidney for mesangial proliferation, migration, apoptosis, and tissue

237 homeostasis. J Immunol 2002;168: 4301–4307.

35 Curat CA, Vogel WF: Discoidin domain receptor 1 controls growth and

adhesion of mesangial cells. J Am Soc Nephrol 2002;13:2648– 2656.

36 Zuk A, Matlin KS: Induction of a laminin isoform and alpha(3)beta(1)-integrin

in renal ischemic injury and repair in vivo. Am J Physiol Renal Physiol

2002;283:F971–F984.

37 Chakravarty D, Cai Q, Ferraris JD, Michea L, Burg MB, Kultz D: Three

GADD45 isoforms contribute to hypertonic stress phenotype of murine renal

inner medullary cells. Am J Physiol Renal Physiol 2002;283:F1020–1029.

38 Sakurai H, Bush KT, Nigam SK: Identification of pleiotrophin as a

mesenchymal factor involved in ureteric bud branching morphogenesis.

Development 2001;128:3283–3293.

238 Figures and tables

239

Figure 1. Representative scatter graphs of FSGS and normal samples demonstrate reproducibility of gene expression. Scatter graphs allow for the crosswise comparison of two microarray hybridization patterns in order to observe both the reproducibility and variation of gene expression profiles. Each dot represents a probe set that is called ‘P’ (present) by the MAS 5.0 software in at least one of the two samples compared. Each probe set is plotted on the log graph according to the MAS 5.0 signal value. The lines represent the boundaries of a threefold expression change so that probe sets outside of this boundary display a threefold or higher change in signal between the two samples plotted. a

Technical replicate of normal kidney tissue. Two 100-ng aliquots from a single control RNA preparation were used to generate two, labeled target RNAs, which were hybridized to two U133A microarrays. The resulting data are highly reproducible, with only 47 genes showing an over threefold difference. b

Biological replicate of FSGS biopsy. Needle biopsy of FSGS patient No. 4 was bisected, and then RNA and labeled target were prepared separately and hybridized to two U133A microarrays. With the addition of biological noise, there is more scatter than in the technical replicate, but still a very high level of

240 Figure 1 continued. reproducibility, with 87 genes showing an over threefold difference. c Crosswise

comparison of control (C) and FSGS samples (P). Notice the increased scatter,

illustrating the presence of many gene expression differences.

241

Figure 2. Series graphs of FSGS and normal samples demonstrate the expression pattern of genes that were upregulated or downregulated compared to control. a One-way ANOVA test found 2,763 differently expressed genes in the

FSGS samples with p < 0.01. b One-way ANOVA test found 534 differently expressed genes in FSGS samples with p < 0.001. The gene expression levels

242 Figure 2 continued. from each microarray hybridization were normalized using RMA Express, which examines hybridization signals of each gene from all microarrays of an experiment and determines relative expression levels. Differently expressed genes are plotted. Biological duplicates were obtained and plotted for 5 of the 10 patient (P) samples and 3 of the 5 control (C) samples. Duplicates were assigned the same subscript number. A total of 23 microarray hybridizations were performed. Individual genes were color-coded according to expression levels in the first control (C1) sample. Red marks genes with higher expression in the control, blue marks genes with lower expression, and yellow indicates genes with intermediate expression levels. Note the consistently reversed expression levels in control and patient samples.

243

Figure 3. Gene tree heat map shows distinct gene expression profile fingerprint for FSGS samples. a Hierarchal cluster analysis of 722 genes that were differently expressed in any of the 3 individual comparisons (b–d). The dendrogram on the left shows the relationship of clusters of similarly expressed genes. b Comparison of 429 differently expressed genes in samples from FSGS patients (P) and controls (C). c Comparison of 151 differently expressed genes in

samples from patients with hypoalbuminemia (†) and those patients without

(albumin 1.4 g/dl vs. 3.5 g/dl, p ! 0.05). Genes were identified using a Welch t test

(p < 0.01) combined with an at least 1.6-fold change in samples from patients

with hypoalbuminemia. d Comparison of 165 differently expressed genes from

samples of patients with renal insufficiency (*) and those patients without

2 2 (creatinine clearance 45 ml/ min/1.73m vs. 116 ml/min/1.73m , p < 0.05). Genes

244 Figure 3 continued. were identified with a Welch t test (p < 0.01) combined with an at least 1.6-fold change in at least 2 out of 3 renal failure patients. Blue represents low-level gene expression relative to the controls, red indicates high expression, and yellow indicates intermediate expression levels.

245 Chapter 6

General Discussion

246 Microarray expression analysis of kidney development

Microarray technology has undoubtedly revolutionized the field of molecular biology. The previously discussed studies have demonstrated the usefulness of this technology in describing the transcriptional state or

“transcriptome” of the kidney in both normal and abnormal developmental states.

These developmental expression profiles have identified thousands of genes specifically expressed within the embryonic kidney compared to the adult kidney.

Likewise, comparison of the two initial metanepheric components of the

E11.5 kidney, the metanephric mesenchyme and ureteric bud, yielded thousands of genes differentially expressed identifying the specific gene expression profiles of these developing structures. In order to obtain pure E11.5 metanephric mesenchyme populations, two different procedures were used: a refined manual micro-dissection technique and a laser capture micro-dissection technique optimized for the preservation of total RNA. Extremely pure ureteric bud samples were micro-dissected using the same manual techniques. The quality of the data generated in the metanephric mesenchyme and ureteric bud using these new techniques represent a vast improvement over our first microarray study. Overall, these microarray expression studies greatly expand our knowledge of gene expression in the developing metanephric kidney which was first performed by

Stuart and colleagues (2001).

Many of the genes identified in these studies included previously described genes identified as important regulators of kidney development through both knockout studies and expression screening using in situ

247 hybridization. The identification of previously identified genes demonstrates the validity this microarray approach in describing renal gene expression profiles during development. Additionally, other developmental genes identified in these comparisons have been targeted in mice. Although these gene knockouts display no significant kidney phenotype, the expression studies suggest a role in kidney development which may be masked by functional redundancy or compensation.

Additionally, hundreds of the developmental genes identified are uncharacterized genes or ESTs representing new candidate genes involved in renal development.

The ability to manipulate renal development in organ culture offers an attractive system to assay gain or loss of gene function using a variety of different protocols (Saxen and Sariola, 1987). The developmentally expressed genes identified in these microarray studies could be functional analyzed using methods to reduce gene expression in kidney organ culture, such as RNAi (Elbashir et al.,

2001a; Elbashir et al., 2001b). This approach could provide a high through-put assay to analyze the developmental function of the hundreds of genes identified in these microarray studies. Our attempts to use this technology proved to be unsuccessful using common transfection techniques, probably due to the difficulty in penetrating the kidney in organ culture. Other transfection methods, such as the lentiviral RNAi expression system, could successfully reduce gene expression in kidney organ culture (Abbas-Terki et al., 2002).

Also, these developmental expression studies of the kidney are actively undergoing expansion within renal development community. In addition to laser capture micro-dissection, other techniques can used to isolate specific cells types

248 for microarray analysis, such as fluorescent-activated cell sorting (FACS), which was recently employed in the isolation of metanephric mesenchyme expressing green fluorescent protein (GFP) driven by the Sall1 enhancer elements

(Takasato et al., 2004). Also, the GUDMAP (A Molecular Atlas of Genitourinary

Development - www.gudmap.org) project aims to broaden our understanding of the development of the gentiourinary system by defining the expression patterns of genes during kidney development using microarray expression technology and in situ hybridization. The techniques performed in the previous developmental studies will be used to identify the complete gene expression profile of specific renal structures throughout development, such as the early renal vesicles, maturing tubule segments, and glomeruli. The molecular characterization of these developing structures will deepen our understanding of the genetic pathway involved in the patterning and differentiation providing the signaling and transcriptional inputs required for normal kidney morphogenesis.

Microarray expression analysis of Hoxa11/Hoxd11 null mutants

In addition to the study of the normal kidney developmental, microarray technology can be used to analyze targeted mice possessing renal defects. The loss of Hoxa11/Hoxd11 during kidney development results in defects in both the ureteric bud and metanephric mesenchyme compartments due to the loss of

Gdnf expression (Patterson et al., 2001; Wellik et al., 2002). The microarray analysis of these mutants revealed many genes significantly changed throughout development representing possible Hoxa11/Hoxd11 targets.

249 Notably, the Hoxa11/Hoxd11 null early metanephric mesenchyme displays

a significant decrease in Meox2 which encodes homeodomain-containing

transcription factor expressed in mesenchyme. Although a renal phenotype has

not been reported within the Meox2 targeted mouse, functional redundancy or

compensation by other factors may mask the role this gene plays in renal

development (Mankoo et al., 2003). Interestingly, Meox2 is expressed specifically

in the nephrogenic mesenchyme condensing around the ureteric bud tips

suggesting that a subtle nephrogenic phenotype may be present within the

Hoxa11/Hoxd11 null embryo (Yu et al., GUDMAP, www.gudmap.org). Also,

Meox2 is down regulated throughout later kidney development within the double

mutants, again suggesting a nephrogenic phenotype in the Hoxa11/Hoxd11 null

kidney. Molecular analysis of these mutants using known nephron segment

markers may reveal a subtle nephron patterning defect in these mice. If no

nephron defect is identified, functional redundancy of other Hox genes may be

masking a nephron phenotype. Further study of the Meox2 null embryonic kidney is needed to ascertain its probable function in the developing nephron.

Additionally, Irx3, an Iroquois homeodomain transcription factor, is up- regulated by 6-fold within the E11.5 Hoxa11/Hoxd11 metanephric mesenchyme

(Yu et al., GUDMAP, www.gudmap.org). Irx3 expression can be identified within both the early and late developing nephron suggesting a role in the patterning of the nephron. The role of Irx3 in renal development has not been described, but the microarray data suggests negative regulation by the Hox11 paralogs within the metanephric mesenchyme. Experimental evidence in hematopoeitic cells has

250 identified Irx3 as a downstream target of both Hoxa9 and Hoxa10 (Ferrell et al.,

2005). Again, further study of Irx3 is needed to understand its possible function in renal development.

Surprisingly, the gene lists generated in this study of the Hoxa11/Hoxd11 null E11.5 metanephric mesenchyme displayed very little overlap with the mutant analysis of embryonic kidneys conducted at later stages of development. The disparity between these two gene lists suggests that the Hox11 paralogs may play two roles, the activation of Gdnf and regulation of nephron development.

Alternatively, as the Hoxa11/Hoxd11 null kidney continues to develop, the disruptions in normal renal patterning become more and more significant, especially within the mid-ventral containing undefined mesenchyme. The significant changes in transcription of the mutant embryonic kidney may represent the uninduced metanephric mesenchyme which becomes a significant portion of the kidney. Therefore, the analysis the Hoxa11/Hoxd11 kidney at these later stages may represent the expression profile of a dysplastic kidney rather than representing true targets of the Hox11 paralogs. It is critical to identify direct targets of the Hox11 paralogs, not indirect expression changes occurring at later developmental stages, to illuminate the genetic pathways involved in renal morphogenesis.

Overall, the microarray analysis of the Hoxa11/Hoxd11 targeted embryo represents one of the most thorough microarray studies of a targeted strain possessing abnormal renal morphogenesis. Past experiments have relied on the use of immortalized metanephric mesenchyme cell lines to adequately represent

251 the gene expression profile of the developing kidney (Valerius et al., 2002a;

Valerius et al., 2002b). The incorporation of the latest microarray technology, sensitive amplification procedures, and cutting-edge micro-dissection techniques were crucial to the microarray analysis of the Hoxa11/Hoxd11 targeted phenotype. Future experiments in which multiple Hox genes with similar renal expression profiles are inactivated using gene targeting may allow us to discern the function of these genes in nephrogenesis by bypassing the functional redundancy of these genes.

Analysis of Pygo1/Pygo2 null mutants

The function of Pygo1 and Pygo2 in renal development was investigated

using a variety of molecular tools, including the Bat-gal transgenic reporter of

canonical Wnt signal, confocal microscopy, whole mount in situ hybridization,

and microarray analysis. These studies have identified the requirement of Pygo2

in ureteric tip morphogenesis during renal development, as well as other

developing structures, while Pygo1 is dispensable for development and survival.

Interestingly, a function of Pygo1 in modulating Bat-gal reporter activation was

identified. When Pygo2 is reduced to heterozygosity on the Pygo1 null

background Bat-gal activation in the E18.5 kidney is reduced by about fifty

percent compared to Pygo2 wild type embryos on the Pygo1 null background.

The complete loss of Pygo2 expression results in the striking loss of Bat-gal

reporter activation in the E18.5 kidney suggesting a role for Pygo2 in mediating

Wnt signaling within the ureteric compartment. Interestingly, the expression of

known Wnt targets was not perturbed in the Pygo1/Pygo2 null E18.5 kidney,

252 although these genes may not represent true Wnt targets in the kidney.

Alternatively, the Bat-gal transgene may not adequately represent Wnt signaling in the developing kidney.

These experiments raise questions whether canonical or noncanonical

Wnt signaling is required for nephrogenesis, given that Pygo1/Pygo2 null embryos have no identifiable defects in nephrogenesis and the Bat-gal transgenic kidney lacks reporter expression within the metanephric mesenchyme or nephrons. Alternatively, Pygo function may not be required in some forms of canonical Wnt signaling within the kidney since Bat-gal reporter expression is not completely lost in the developing structures of the E10.5 embryo.

To determine if canonical Wnt signaling is required for nephrogenesis, we attempted to conditionally delete Beta-catenin within metanephric mesenchyme using Cre / loxp recombination. Unfortunately, the transgenic Cre strain (Pax3-

Cre) used in these experiments resulted in complete loss of the posterior embryonic development rendering study of renal development impossible (data not shown) (Li et al., 2000). The loss of posterior morphogenesis in the conditional deletion of β-catenin resembles the combined knockout of both Lef1 and Tcf1 demonstrating the necessity of β-catenin in canonical Wnt signaling

(Galceran et al., 1999). The generation of a metanephric mesenchyme-specific

Cre strain would greatly aid this experiment since the Pax3-Cre strain possesses

general expression in posterior tissues of the embryo (data not shown).

Clearly, further analysis of the Wnt signaling pathway is needed in the

renal morphogenesis. Loss of Pygo1/Pygo2 results in a mild ureteric tip defect

253 and a remarkable reduction of canonical Wnt signaling within the ureteric tree demonstrated by the Bat-gal transgene while nephrogenesis is not severely affected. Gene targeting of other β-catenin transcriptional complex members, such as Bcl9 and Bcl9l, will greatly aid in our understanding of the role of canonical Wnt signaling in kidney development. Also, the inactivation of other members within the transcriptional complex will also aid our understanding of canonical Wnt signaling during kidney development.

Focal segmental glomerulosclerosis

FSGS has become a common disease in pediatric nephrology resulting in a rapid progression to renal failure. Additionally, the disease is described as heterogeneous due to variable response to treatment of patients and the unknown insult resulting in the characteristic segmental lesion. Our study aimed to understand the gene expression changes underlying the pathology of the

FSGS. Notably, the analysis of ten FSGS kidney biopsies identified hundreds of genes differentially expressed compared to normal kidney tissue. Genes identified in the study include those previously identified in both FSGS and other kidney diseases. Additionally, new candidate genes were identified which may operate in the pathogenesis of the disease. Surprisingly, this analysis identified a significant decrease of the specific podocyte genes, Wt1 and NPHS1, within the

FSGS samples even though the glomeruli populate a small percentage of the total volume of the kidney.

The microarray analysis was extended to the study of FSGS patients possessing either nephrotic syndrome identified by loss of serum albumin, or

254 renal failure identified by decreased creatine clearance. The expansion of this study to hundreds of FSGS patients could provide an excellent diagnostic tool to further classify FSGS based on the molecular expression profile. Further classification of FSGS disease states would also provide the means to predict disease progression and response to renal protective therapies.

Additionally, the laser capture micro-dissection of individual renal structures of the FSGS kidney will provide an even greater resolution of the molecular changes involved in this renal disease than kidney biopsies. The gene expression changes of the FSGS glomerulus could provide new information in the signaling pathways involved in gomerular injury and potential new diagnostic and therapeutic targets. As previously mentioned, if the sample sizes of these experiments are increased, the microarray analysis of laser captured micro- dissected glomeruli could provide a new clinical classification of FSGS based on molecular expression changes in addition to histological analysis. Also, the microarray analysis could be extended to the tubules of the nephron allowing the investigation of the downstream gene expression changes resulting from the primary lesion within the glomerulus.

Overall, this study provides the clinician with an expanded view of the molecular changes occurring in the FSGS kidney, providing hundreds of possible new candidate genes involved in renal pathologies. These results validate the potential use of expression microarray technology as a diagnostic tool in the molecular characterization of different FSGS cases, hopefully providing new

255 insight into the pathology of the disease, predicting the progression to renal failure, and determining the most effective therapy to pursue.

256 References

Abbas-Terki, T., Blanco-Bose, W., Deglon, N., Pralong, W., and Aebischer, P.

(2002). Lentiviral-mediated RNA interference. Hum Gene Ther 13, 2197-

201.

Elbashir, S. M., Harborth, J., Lendeckel, W., Yalcin, A., Weber, K., and Tuschl, T.

(2001a). Duplexes of 21-nucleotide RNAs mediate RNA interference in

cultured mammalian cells. Nature 411, 494-8.

Elbashir, S. M., Lendeckel, W., and Tuschl, T. (2001b). RNA interference is

mediated by 21- and 22-nucleotide RNAs. Genes Dev 15, 188-200.

Ferrell, C. M., Dorsam, S. T., Ohta, H., Humphries, R. K., Derynck, M. K., Haqq,

C., Largman, C., and Lawrence, H. J. (2005). Activation of stem-cell

specific genes by HOXA9 and HOXA10 homeodomain proteins in CD34+

human cord blood cells. Stem Cells 23, 644-55.

Galceran, J., Farinas, I., Depew, M. J., Clevers, H., and Grosschedl, R. (1999).

Wnt3a-/--like phenotype and limb deficiency in Lef1(-/-)Tcf1(-/-) mice.

Genes Dev 13, 709-17.

Li, J., Chen, F., and Epstein, J. A. (2000). Neural crest expression of Cre

recombinase directed by the proximal Pax3 promoter in transgenic mice.

Genesis 26, 162-4.

Mankoo, B. S., Skuntz, S., Harrigan, I., Grigorieva, E., Candia, A., Wright, C. V.,

Arnheiter, H., and Pachnis, V. (2003). The concerted action of Meox

homeobox genes is required upstream of genetic pathways essential for

257 the formation, patterning and differentiation of somites. Development 130,

4655-64.

Patterson, L. T., Pembaur, M., and Potter, S. S. (2001). Hoxa11 and Hoxd11

regulate branching morphogenesis of the ureteric bud in the developing

kidney. Development 128, 2153-61.

Saxen, L., and Sariola, H. (1987). Early organogenesis of the kidney. Pediatr

Nephrol 1, 385-92.

Stuart, R. O., Bush, K. T., and Nigam, S. K. (2001). Changes in global gene

expression patterns during development and maturation of the rat kidney.

Proc Natl Acad Sci U S A 98, 5649-54.

Takasato, M., Osafune, K., Matsumoto, Y., Kataoka, Y., Yoshida, N., Meguro, H.,

Aburatani, H., Asashima, M., and Nishinakamura, R. (2004). Identification

of kidney mesenchymal genes by a combination of microarray analysis

and Sall1-GFP knockin mice. Mech Dev 121, 547-57.

Valerius, M. T., Patterson, L. T., Feng, Y., and Potter, S. S. (2002a). Hoxa 11 is

upstream of Integrin alpha8 expression in the developing kidney. Proc Natl

Acad Sci U S A 99, 8090-5.

Valerius, M. T., Patterson, L. T., Witte, D. P., and Potter, S. S. (2002b).

Microarray analysis of novel cell lines representing two stages of

metanephric mesenchyme differentiation. Mech Dev 110, 151-64.

Wellik, D. M., Hawkes, P. J., and Capecchi, M. R. (2002). Hox11 paralogous

genes are essential for metanephric kidney induction. Genes Dev 16,

1423-32.

258