EXPRESSION OF RET, GDNF AND GDNFR-a IN HUMAN

DEVELOPMENT AND DISEASE

STACEY M. IVANCHUK

A thesis submitted to the Department of Pathology in conforming with the requirements for

the degree of Master of Science

Queen's University

Kings ton, Ontario, Canada

April 1997

copyright 8 Stacey M. Ivanchuk, 1997

National Library Bibliothhue nationale du Canada

The author has granted a non- exclusive licence allowing the National Library of Canada to reproduce, loan, distri'bute or sell copies of hismer thesis by any means and in any form or formaf making this thesis available to interested persom-

The author retains ownership of the copyright in bidher thesis. Neither the thesis nor substantial adacts &om it may be printed or otherwise reproduced with the author's permission.

L'auteur a accorde me licence non

forme cpe ce soit pour mettre des exemplaires de cette th&e a la disposition des persomes ht&e&es.

L'auteurcoaserve IapropriM & b i t d'auteur qyi proege sa t h h . Ni la t h b ni des extmb substantiels & ceile-ci ne doivent &re imprimes ou autrement reproduits sans son

Abstract

The RET protooncogene encodes a receptor tyrosine kinase required for kidney

development. RET is predicted to fuaction in the reciprocal inductive interactions that give

rise to the mature kidney. To investigate the role for RET in human kidney development,

we used RT-PCR analyses to examine RET expression in a panel of fetal kidney RNAs

ranging in gestational age h m 7-5-24 weeks and in adult kidney. We detected expression

of RET and its multiple 5' and 3' coding variants in each RNA sample. RET expression

was developmentally regulated with expression levels 6- to 7-fold higher in the early

gestational ages dative to adult A similar pattern was observed for two alternative RET

transcripts generated by exon skipping at the 5' end of the gene. The expression of a third

&anscrip& RER16, which lacks exons 3-5, was distinct with relatively higher expression

in early gestational ages and a more rapid decrease in expression. RETU6 expression was

highest through a period of rapid bifurcation of the ureteric bud raising the possibility that

RM16 has a role in this process. We detected expression of three alternatively spliced

RET 3' variants ( R E D , REnl and RET43) throughout human kidney development.

RE19 was the most abundant variant in all samples. In contrast, RETSI expression was

almost undetectable early in development but significantly upregulated by 9 weeks

gestation. RETSl expression levels were approximately 35% those of RE19 in later

gestational ages. These results suggest a role for RETSL distinct from that for RET9 during

human kidney development*

The ligand for RET was recently identified as a complex consisting of GDNF and

GDNFR-a. We detected expression of both GDNF and GDNFR-a throughout our panel

of fetal kidney RNA. Developmental regulation of GDNFR-a expression was identified.

The expression pattem was distinct from that identified for RM during human kidney

development and may be indicative of a role for GDNFR-a in regulating RET activity.

In our investigation of RET, GDNF and GDNFR-a expression in the renal

tumours, Wilms' tumour and renal cell carcinoma, no consistent pattem of expression of

the three transcripts was discerned although RET was consistently detected in both tumour

tY pes-

Acknowledgements

This thesis is dkdicated to the memory of my mother and guardian angel, Bernice

(Bunny) Lvanchuk, who taught me the importance of dedication and determination. Her

spirit has found a home in me and she guides me through all the ups and downs that We

brings. I thank my father, Morris, and brother, Bryce, for supporting me in so mauy ways

throughout the course of this degree. Their constant love and encouragement has been my

rock.

I want to thaok Dr. Lois Mulligan for giving me the opportunity to pursue a

graduate degree in her lab. There have been many times when I have wondered what she

saw in me. She bas been a remarkable teacher and mentor but above al l else, a friend- I

don't know what I would have done without her constant encouragement and faith in me.

She has helped cultivate my love of science and research. It has been both a privilege and a

pleasure to work for her.

The people I worked with every day made a l l the hard times so much more

bearable. Shirley Myers has been a wonderful resource and a "running" motivator! Dr.

Harriet Feilotter is, and will always be, my favorite post-doc. I thank her for all her help

and the "Ieaf of hope" which I have kept with me through the course of writing up. I would

also like to thank Lee Fraser, a fellow graduate in the lab, in whom I found someone who

shares my taste in music and Love for sports. Thanks, too, to Nicolina Zakova, a late

addition to our lab.

I would like to thank my committee members, Drs. David Li;Licrap and Peter Greer,

for their constant support, helpful advice and wful discussions. Thanks also to Dr. Jim

Gerlach and Ms. Leah Young for all the help with PCR techniques. To Carla Cuthbert, my

favorite graduate student, I am forever indebted for she has made me feel that I am capable

of pursuing a career in science.

v

Final thanks go to Drs. Robert Hofstra and Marc Billaud for sharing their data prior

to publication and to the Queen's University Department of Pathology for their financial

support over the past three years.

Table of Contents

Abstract

Acknowledgements

Table of Contents List of Figures List of Tables List of Abbreviations

Chapter 1: Introduction

1.1 The REI proto-oncogene 1.2 RET Expression 1.3 The RET Ligand Complex

1.3.1 Glial Cell Line-Derived Neurotrophic Factor (GDNF) 1.3.2 GDNF Receptor-a (GDNER-a)

1.4 The RET Knockout Mouse 1.5 Kidney Development 1.6 Molecules Involved in Kidney Development 1.7 Alternative Splicing 1.8 RET in Disease

1.8.1 RET Expression in Tumours and Tumour Ceil Lines 1.8.2 R l Z Rearrangements in Papillary Thwoid Carcinoma 1.8.3 RET in Inherited Syndromes 1.8 -4 RET Mutations in Sporadic Tumours

Research Objectives

Chapter 2: Materials and Methods

Oligonucleotides RNA . Tissues and Patient Samples

Hirschsprung Patient Samples Fetal Kidney RNA and Renal Tissue Samples

Total RNA Extraction Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

First S m d cDNA Synthesis PCR Using cDNA Templates Semi-Quantitative RT-PCR

Southern Blotting and Hybridization Southern Blotting Oligonucleotide Labeling Hybridization

Sequencing DNA Template Preparation Sequencing Reactions

Chapter 3: Expression of RET and its Multiple Splice Forms in Human Developing Kidney

Introduction

Resuits REl Expression in Human Kidney Development Quantitation of RET Expression During Human Kidney Development RET Alternative Splice Variants Quantitation of RTr Alternative Splice Vadaats

Discussion Materials and Methods

Growth of 'IT Cells M A Extraction Fetal Kidney, Adult Kidney and TT RNA Analyses Semi-Quantitative Reverse Transcription-Polymerase Chain Reaction Southern Blotting

Chapter 4: Expression of RET 3' Splicing Variants During Human Kidney Development

Introduction Results

Expression of RET 3' Alternatively Spliced Transcripts Quantitation of RET 3' Alternatively Spliced Variants

Discussion Materials and Methods

Fetal Kidaey, Adult Kidney and 'IT RNA Analyses Semi-Quantitative Reverse Transcription-Polymerase Chain Reaction

Chapter 5: Expression of Genes Encoding the RET Ligand Complex Components GDNF and GDNFR-a During Human Kidney Development

Introduction Results

Expression of GDNF in Human Fetal Kidney Expression of GDNFR-a During Human Kidney Development Quantitation of GDNFR-a in Human Fetal Kidney

Discussion Materials and Methods

Reverse Transcription-Polymerase Chain Reaction Analyses Semi-Quantitative Reverse Tcanscription-Polymerase Chain Reaction

Chapter 6: AnaIyses of RET, GDNF and GDNFR-u Expression in Human Disease

Introduction Results

Expression of RET, GDNF and GDNFR-a in Wilms' Tumour Expression of RET, GDNF and GDNFRa in Rend Cell Carcinoma GDNF Mutation Analysis in Hirschsprung Patients

Discussion RET/GDNF/GDNFR-a in Wilms' Turnours RET/GDNF/GDNFR-a Expression in Renal Cell Carcinoma

GDNF Mutation Analysis in Hirschspmng Disease Materials and Methods

Reverse Transcription-Polymerase Chain Reaction Conditions Genomic Sequencing

Chapter 7: Discwion

RET in Human Kidney Development RET 5' AIternatively Spliced Variants in Development GDNF and GDWR-a in Human Kidney Development RET 3'Altematively Spliced Variants in Development RET, GDNF and GDNFR-a in Human Disease

Summary

References

Appendix

Curriculum Vita

List of Fi y r e s

Figure

Schematic of the RET Receptor Tyrosine Kinase

Schematic of the RET 3' Alternative Coding Variants

Schematic of the REl Exon Structure and its Relationship with the RET 5' Alternatively Spliced Variants

The E2ETIGDNFiGDNFR-a Signaling Complex

Kidney Induction Cascade

Determination of Conditions for Semi-Quantitative RT-PCR Analysis of O v e d RlTExpression in Human Fetal and Adult Kidney

Expression of RET mRNA Forms Generated by Alternative Splicing at the 5' End of the Gene in Humao Fetal and Adult Kidney

Developmental Expression of RET5' Splice Variants in Human Fetal Kidney

RET 3' Isoforrn Interactions with Downstream Effector Molecules

Schematic of RET 3' Exons and the Locations of Primers Used in RT-PCR Analysis

Expression of REI 3' Coding Variants in Human Fetai and Adult Kidney

Determination of Conditions for Semi-Quantitative RT-PCR Analysis of RET 3' Coding Variant Expression

Developmental Expression of RET 3' Codiag Variants in Human Fetal Kidney

Expression of GDNF and GDNFR-a in Human Fetal and Adult Kidney

Determination of Conditions for Semi-Quantitative RT-PCR Analysis of GDNFR-a Expression

Developmental Expression of GDNFR-a in Human Fetal Kidney

Page

6.1 RET, GDNF and GDNFR-a Expression in W i ' Tuxnoun

6.2 RET, GDNF and GDNFR-a Expression in Renal Cell Carcinoma

6.3 GDNF Mutation in a Hirschpmg (HSCR) Patient

List of Tables

Table

Relative Expression Levels of RET in the Human Fetal Kidney

RET 5' Splice Vadant Expression Relative to Full Length RET Expression in Human Fetal Kidney

REN6 Expression Relative to FL RETExpression in Human Fetal Kidney

RETSl Expression Relative to RE19 Expression in Human Fetal Kidney

Expression of GDNFR-a in Developing Human Fetal Kidney

Histology and REZ: GDNF and GDNFR-a Expression in W h ' Tumours

RET, GDNF and GDNFR-a Expression in Renal Cell Carcinoma

Page

35

Abbreviations

bek BMP b~ BWS

C cDNA CJD CNS C-tenninus "c DEPC DNA dNTP dpc dT m EDTA EGFR ENS

FFI FGF FGFR FISH M C

G GDNF GDNFR-a GPI GUSB

HC1 HSCR

KC1 KGF KGFR

Id LMP LOH

adenosine triphosphate avian m ~ e l ~ b I a s t ~ ~ i s virus reverse trans&iptase adenosine triphosphate

basic mrobiast growth factor receptor bone morphogenic protein base pair Beckwith-Wiedemann syndrome

cytosine complementary deoxyn'bonucleic acid Creutzfeldt-lakob disease central nervous system carboxy terminus degrees Celsius

diethyl pyrocarbonate deoxyribonuclek acid deoxynucleoside triphosp hate days post-coitum deoxythymine dithiothreitol

ethylenediamhetetraacetic acid epidermal growth factor receptor enteric nervous system

fatal familial insomnia fibroblast growth factor fibroblast growth factor receptor fluorescence in sim hybridization familial medullary thyroid carcinoma

guanosine trip hosphate glial cell linederived neumtrophic factor glial cell linederived neurotmphic factor receptor-alpha glycosylp hosphatidylinositol beta-glucuronidase

hydrogen chloride Hinchsprung disease

immunoglobulin

potassium chloride keratinocyfe growth factor keratinocyte growth faftor receptor

limb deformity Low melting point loss of heterozygosity

M MAPK MEN MM Mrn2 min mRNA MTC

N NaOH NF N G F R P ~ ~ NK at N-terminus

P PCR PNK PNS m PTC PY

RA RAR-a RB RCC RET RNA RNase R7K RT-PCR

SCG sd SDS SH SLF SSC

T TBE TGF-P TK TPE

molar mitogen activated protein base multiple endocrine neoplasia metanephric mesenchyme r n ~ a i u m chloride minute mssenger ribonucleic acid medullary thyroid carcinoma

normal sodium hydroxide neurofibromatosis low-affinity nerve growth factor receptor normal kidney nucleotide - amino-terminus

phosphate polymerase chain reaction polynucleotide kinase peripheral nervous system phosphostyrosine binding papillary thyroid carcinoma phosphotyrosine

retboic acid retinoic acid receptor-alpha retinoblastoma renal cell carcinoma rearranged in transfection ninucieic acid ribonucIease receptor tyrosine kiaase reverse transcription-polymerase chain reaction

superior cervical ganglia standard deviation sodium dodecyl sulphate Src-homology steel factor saline sodium citrate

thymine Tris-borate EDTA transforming growth factor-beta tumour kidney Tris-phosphate EDTA

untranslated region volts von Hippel-Lindau weeks

Chapter 1

Introduction

1 .I The RET Proto-oncogene

RET (REarranged in Transfedon) was first isoiated as a dominant transforming

gene by transfection of turnour DNA into fibrobIasts 2. The gene was found to be

activated by recombination between two unlinked segments of DNA. The 5' portions were

different, however, the 3' segments of the isolated transforming genes were identical and

corresponded to the tyrosine kinase domain of a novel gene, RET 3. The gene consists of

2 1 exow that span 58 kb of DNA 4-7. Fluorescence in sim hybridization (FISH) analyses

were used to assign RET to chromosome 10q11.2

RETencodes a receptor tyrosine kinase (RTK) (Figure 1.1) 3. The yosine kinase

domain of RET is split by an insertion sequence of 27 amino acids. In addition to the

b a s e domain and short transmembrane domain cbaractedsac of RTK family members,

RET consists of a large extracellular domain comprising a signal peptide, a region of

cadherin homology and a cysteine residue-rich domain 9- lo. The cysteine-rich domain

contains 16 cysteine residues in a stretch of 120 amino acids. The translation start site and

amino 0-terminal signal sequence both lie within RET exon 1 5. Twelve possible N-

linked glycosylation sites are predicted in the exmcellular domain of the protein 97 10.

Five RET transcripts of 7.0,6.0,4.6,4.5 and 3.9 kb are detected by northern blot

analysis 11- 12. The differences in size are due to alterations in the 3' end of the gene 6. 12.

Splicing of RE;T 3' exons generates transcripts which encode several different carboxy (0-

termini. AU transcripts share. the first 19 exons after which sequences diverge. Splicing

may not occur at the end of exon 19 in which case the coding sequence continues into

intron 19 sequence coding for a further nine codons (Figure 1.2) 12. These transcripts

encode proteins designated RET9. Alternatively, exon 19 may be spliced to exon 20 or

exon 2 1 generating RET aanscripts with a novel 5 1 (RETS 1) or 43 (RET43) amino acid C-

terminus respectively (Figure 1.2) 12. W T 5 1 manscripts generate a 1 1 14 amino acid

Figure 1.1 Schematic of the RET receptor tyrosine kinase. The characteristic domains

contained within the extra- and intracellular domains of the protein are indicated

Signal Peptide

Cad he tin Homology

Region

Cysteine Rich Region

Transmembrane Domain

Kinase lnterkinase

Domain

Extracellular Domain

lntracellular Domain

Figure 1.2 Schematic of the RET 3' alternative coding variants. All transcripts share the

k t 19 exons of RET . The coding sequence for each isoform is the same up to and

including sequence for tyrosine residue 1062 (Y1062) after which the sequences diverge.

Splicing may not occur at the 3' and of exon 19 in which case the sequence reads into

intron 19 (RE19). Alternatively, RETexon L9 may be spliced to exon 20 (RETS1) or exon

21 (RET43). Coding regions are given as boxes and non-coding regions are given as solid

Lines. Dashed lines are used to indicate splicing. Polyadenylation sites are indicated by

(AM-

RET protein while RE79 transcripts generate a 1072 amino acid isoform 12. RET43 is

predicted to encode a protein of 1106 amino acids.

Alternative splicing also occurs at the 5' end of RET. Multiple transcripts are

possible as a result of exon skipping (Figure 1.3) 13. Exon 2 may be spliced to exon 3

resulting in transcripts that encode the complete extracellular domain. When exon 2 is

spliced to either exon 4 or 6, the resulting transcripts encode proteins that lack portions of

the region of cadherin homology 13. When exon 2 is spliced to exon 5. transcripts

encoding truncated products result due to a frameshift and consequent premature stop in

exon 5 13. In vino translation of proteins from the cDNAs corresponding to the RET 5 '

alternative splice variants has been described, however. it is not yet clear whether these

isoforms are functional in vivo,

1.2 RET Expression

In rodents. northern blot, in situ hybridization and immunohistochemistry analyses

have been used to determine the dismbution of RETexpression. During development, RET

transcripts have been found predominantly in subsets of cells of the centraI nervous (CNS)

and peripheral nervous systems (PNS) and of the excretory system l l v 14-16- Within the

CNS* RET expression has been localized to the motor neuron Lineages of the spinal cord

and hindbrain 14. Expression has also been detected in subsets of neuroretinal cells. Within

the PNS, RET expression has been localized to the sensory ganglia of the head and neck.

the dorsal root ganglia and its prrcursors. the cranial ganglia and the vagal neural crest and

myenteric ganglia of the gut I4l6. Expression in the embryonic kidney has been detected in

the early pro- and mesooephric structures as well as in the metanephros which matures into

the functional kidney lCL6. REZexpression in the developing kidney has been found to be

strongest early in development and to subsequently decmase with increasing gestational age

14, 15.

Figure 1.3 Schematic diagram of the RET pmto-oncogene and its protein products. A] The

relationship of RET coding exons and the corresponding protein domains. B] Predicted

products of alternatively spliced RET transcripts with exon skipping of 5' exom.

FL=M length

2/4=lPET exon 2 spliced to exon 4

US=RETexon 2 spliced to exon 5

2/6=-RETexon 2 spliced to exon 6

Additional norneural structures that express RET during embryogenesis include

salivary glands, spleen and lymph nodes 11- 16. RET expression has also been detected

in the embryonic liver. however, the period of expression is brief (125-14.5 days post-

coitum (dpc)) 149 1s.

In the adult, RETexpression has been found in subsets of PNS and C N S neurons

16, 17. Tmmunohistochemistry has been used to detect RFT expression in the sympathetic

ganglia and neuroepitheLium-derived astrocytes and cortical neurons 17. Northern blot and

in situ hybridization analyses have been used to detect low levels of RET expression in

adult lung, heart. spleen, lymph nodes, thymus and small intestine while stronger

expression has been detected in adult brain and salivary glads 14* 16. These analyses have

failed to detect RET expression in the adult Liver, kidney or thyroid 1 1 - 14. However, the

more sensitive reverse transcription-polymerase chain reaction (RT-PCR) has been used to

detect RET expression in the latter two tissues 13- Is.

1.3 The RET Ligand Complex

1.3.1 Gliai Cell Line Derived Neurotrophic Factor (GDNF)

The ligand for RET has recently been identified as a multicomponent complex.

Members of this complex include glial cell line-derived neurotrophic factor (GDNF) and

GDNF receptor-a (GDNFR-a) (F~gure 1.4) 19-22. Lin et al. 23* t4 identified GDNF as a

neurotrophic factor with the ability to sustain dopaminergic neurons. GDNF is synthesized

as a preproprotein containing an N-terminal signal sequence required for secretion. The

mature human 134 amino acid protein is a glycosylated disulfide-linked homodimer.

GDNF has seven conserved cysteine residues in the same relative spacing found in

memka of the TGF-P family 23. Mature GDNF shares less than 20% homology with any

family member 24.

The human GDNF gene has been mapped to chromosome 5p 12-p 13.1 and

consists of two exons 23. The use of an alternative splice site within the first exon results

Figure 1.4 The RETfGDNFfGDNFR-a signaling complex. GDNFR-a, a GPI-linked cell

surface protein, is required to mediate GDNF homodimer binding to RET. Whether the

GDNF dimer binds one or two molecdes of GDNFR-a is not known, thus, both

possibilities axe illustrated. GDNF in complex with GDNFR-a is capable of activating

RET, resulting in receptor autophosphorylation (indicated by P enclosed in a circle).

GDNF Ihnnnf imcr

in the deletion of 78 nucleotides affecdng sequence encoding the potential secretion signal

and the consensus sequence for proteolytic processing which are both absent in the rnature

GDNF protein 23. Thus, both the full length and alternatively spliced mRNAs encode the

same mature protein and the signiticance of these isoforms is unclear *6* 2'.

GDNF is widely expressed in many non-neural as well as neural tissues 26+ 2'. In

the developing murine embryo, GDNF is first detected in the neuroectoderm. Low levels of

GDNF expression have been detected throughout the C N S from early embryonic ages to

adulthood. Highest levels of GDNF expression during development have been detected in

skin. whisker pad, kidney, stomach and testis 28. Expression levels in the gut and kidney

decrease after birth similar to RET expression 28- 29. In comparison, notable increases in

GDNF expression Levels have been observed in adult liver, lung, testis and ovary 2'- 28.

Recently, a GDNF-related molecule has been identified. Human neurmrin has 42%

amino acid homology with human GDNF 30. 31. Both GDNF and neurmrin can stimulate

the MAP kinase signaling pathway in cultured sympathetic neurons and support the

survival of sympathetic as well as sensory neurons in vim0 3 While GDNF has been

shown to interact with RET and stimulate RET activation, neurturin's ability to bind to RET

has not yet been investigated

1.3.2 GDNF Receptor-a

GDNFR-a is the other identified member of the RET multicomponent receptor

complex. This glycosyLphosphatidylinosito1 (GPO-Linked protein is a membrane-bound

molecule required to mediate GDNF binding to RET (Figure 1.4) Z1- 22. Human GDNFR-

a encodes a 465 amino acid protein with highly hydrophobic amino 0- and cvboxy (C)-

termini characteristic of GPI-linked proteins (reviewed in 32). GDNFR-a contains 3

potential glycosylation sites and 30 cysteine residues 21. The spacing of the cysteine

residues is not related to that found in any extracellular cysteine-rich domain reported for

known receptors 21.

GDNFR-a has been mapped to l0q25-26. In the rat, GDNFR-a expression has

been detected in both neural and non-nets.rai t issws in a pattern similar to both RET and

GDNF. Expression has been detected in the dopaminergic, spinal motor and subsets of

dona1 root ganglia neurons, in developing nephrons and in embryonic smooth and striated

muscles of the enteric nervous system (gut, esophagus, stomach) 21- 22. Expression has

also been detected in the retina, thalamus, pons, medulla oblongata, pituitary and urogenital

trac to

GDNFR-a is required to mediate GDNF binding to RET. In the absence of

GDNFR-a, GDNF binding to RET and autophosphorylation of the receptor are minimal

21. 22. Treanor et ol. 22 have proposed that a GDNF hornodimer binds to GDNFR-a

forming a composite ligand that is capable of binding to and activating RET.

Recently, RElTL2 has been identified as a GDNFR-a family member 33. Human

RETL2 shares 49% amino acid homology with human GDNFR-a. Common features of

both RETL2 and GDNFR-a are a hydrophobic N-terminus indicative of a signal sequence

and a hydrophobic C-terminus indicative of a GPI-linkage motif 21- 22. 33. Both GDNFR-

a and RETL2 are expressed in embryonic brain, lung, kidney and intestine. However,

GDNFR-a is expressed at relatively higher levels in the kidney and intestine while RETL2

is expressed at relatively higher levels in the lung 33. Sanicola et a[. 33 have shown that, in

the absence of RET, GDNFR-a binds to GDNF with greater aftinty than does RETL2.

This suggests that RETL2 requires RET to facilitate binding to GDNF. Consistent with

this, RETL2 forms a high affinity complex with GDNF in the presence of RET.

1.4 The RET Knockout Mouse

While mice heterozygous for functional RET appear normal, RET knockout

(4) mice develop to term and die 16-24 hours after birth 34. Dissection of the RET -1- mice

revealed absent or rudimentary kidneys. Histological analysis of the mutant kicky

rudiments has shown severe dyspiasia characterized by reduced ureteric bud branchings

and absence of a mature collecting duct system as well as large areas of undifferentiated

mesenchyme 34. Some RET-f- mice contain blind-ending ureters with no renal tissue and

others display a complete absence of the ureter and kidney as either uni- or bilateral renal

agenesis. The histology of the kidney rudiments in the RET-/- animals is consistent with

failure of the ureteric bud to grow and bifurcate properly 35.

In addition to remi dysmorphology, the phenotype of the RET-I- mice is

characterized by absence of neurons of the myenteric plexus from the small and large

intestines, esophagus and stomach 34- 36. Enolase and peripherin, enteric neuron markers,

have not been detected in the myenteric and submucosal plexi of the small and large

intestines in the RET-I- mice 34. RET-I- mice also display complete absence of superior

cervical ganglia (SCG), the most anterior ganglia of the sympathetic chain, which are

believed to be derived from a pool of precursor cells that also gives rise to the enteric

nervous system (ENS) 34. 36. Anatomical abnormaiities have not been detected in other

parts of the developing nervous system that normally express RD' including cranial, dorsal

root and sympathetic ganglia as well as spinal cord motor neurons and brain 34.

1.5 Kidney Development

The mammalian kidney develops through a series of reciprocal inductive

interactions between two distinct cell types, the ureteric bud epithelium and the metanephric

mesenchyme (MMJ 35. Interactions between these two cell types result in the growth and

branching of the ureteric bud as well as in the condensation and epithelialization of the MM

(Figure 1.5) (reviewed in 35 and 37). As the kidney matures, the ureteric bud develops into

the collecting duct system while the MM gives rise to both the nephrons and the connective

tissues.

One of the first signs of kidney induction is the formation of the ureteric bud off the

Wolftian duct 35. The evagination of the ureteric bud is under the control of signals from

Figure 1.5 The kidney induction cascade. The two cell lineages that give rise to the kidney

are the ureteric bud epitheiium and the metanephric mesenchyme (MM). Arrows are used to

indicate progressive events in kidney development The ureteric bud gives rise to the

collecting duct system while cells derived from MM give rise to the nephrons and

connective tissue.

Requirement

Differentiation Lr' Growth and Bifurcation

Stromagenic

Condensation Mesenchyme-Epithelium

Transition

Nep hrons I Oifferentiaring I I ColIecting Duct System I

Induction

Adapted from Bard et al., 1994

Duct Bikat ion Ceases

the MM (Figure 1.5). In the young embryo, the MM is fmt visible as a discrete zone of

blastema 3-4 cells thick 35- 37. The blastema is induced to divide and enter a stem cell phase

by signals from the newly formed ureteric bud (reviewed in 38). The stem cells of the MM

give rise to progeny which are either stromagenic or nephrogenic (reviewed in 38 and 39).

Further induction signals derived from the ureteric bud stimulate the MM nephrogenic cells

to aggregate, epithelialize and form nephrons and the stromagenic cells to differentiate into

connective tissue. Signals derived h m the MM supplement ureteric bud growth and lead

to bud bihucation (Figure 1.5) (reviewed in 38). At the more advanced stages of kidney

organogenesis, these inductive interactions take place only in the nanow region around the

outer edge of the kidney referred to as the nephrogenic zone 35. Early in kidney

development, the metanephric blasted cells are separated from the Wolffian duct by a

greater physical distance than the MM stem cells are fkom the ureteric bud epithelia 35. This

suggests that the earlier inductive interactions represent long range signaling in comparison

to subsequent inductive interactions where the two cell types are in close proximity.

In humans, the nephrogenic cells fit begin to differentiate into the various cell

types of the mature nephron around 8 weeks gestation (reviewed in 35 and 37). From 8

until 14-15 weeks, nephrons attach to tubules within the zone of wteric bud anterior

growth. During this period, the rate of ureteric bud bifurcation is rapid (reviewed in 35 and

37). Undifferentiated MM covers the entire ureteric bud during the initial two divisions of

the bud (reviewed in 35 and 37). As more branches are formed and the bud elongates,

undifferentiated mesenchyme advances with the actively growing tips of the ureteric bud.

This constitutes the nephrogenic zone. Around 15 weeks gestation. the rate of ureteric bud

bifurcation slows through to 19-20 weeks (reviewed in 37). Branching of the bud rarely

occurs between 20 and 32 weeks gestation and anterior extension at this time is minimal.

1.6 Molecules involved in Kidney Development

In the mouse, formation of the kidney is initiated around 10.5-1 1.5 dpc when the

ureteric bud emerges near the caudal end of the nephric duct and grows dorsally (reviewed

in 38). As the ureteric bud invades the MM and begins to branch, RET expression is

observed throughout the bud 14- 15. With the progression of kidney development, RET

expression is restricted to the growing tips of the ureteric bud within the nephrogenic zone.

The requirement for RE' for the development of the ureteric bud has been confirmed in

organ culture experiments 40. Ureteric bud and MM tissues isolated fiom RET-/- and

normal control rnurine embryos can be co-cultured in vino. m e n RET-I- bud and MM

tissues are cocultuced, the ureteric bud fails to grow in the majority of cases and any

degree of branching is rare 40. Additionally, differentiation of the MM is not observed In

co-cultwes of normal ureteric bud tissue with RET-I- MM tissue, the mutant mesenchyme

is capable of inducing growth and branching of the ureteric bud *. However, when

REW- ureteric bud tissue is co-cultured with MM tissue from a normal embryo, branching

of the ureteric bud does not occur 40. These results suggest that the failure of ureteric bud

development to proceed in RET-I- embryos was due to the absence of functional RET in the

bud epithelium preventing response of the ureteric bud to mesenchyme-derived signals.

Maturation of the ureteric bud and metanephric mesenchyme are interdependent and

expression studies have implicated more than 50 molecules in the regulation of induction

andlor early or later differentiation in the mammalian kidney (reviewed in 38 and 39) Many

of the molecules identified as essential for proper kidney development are absent in the

adult organ indicating developmental regulation of expression (reviewed in 39). Assigning

specific roles to the molecules involved in mammalian kidney development has proven to

be difficult. Targeted disruption of more than 30 genes known to be expressed during

kidney development has produced a "helpful" phenotype. characterized by some degree of

renal dysmorphology. in only 6 knockout mice (reviewed in 38 and 39). Aside fiom gene

knockouts. however, the effects of deregulated gene expression or mutations that alter the

n o d function of the gene product have provided insight into the roles of other molecules

involved in kidney development.

Knockout mice that fail to show early signs of kidney development include the

WT- I-/-. RAR-ay-I-. and GDNF-/- mice suggesting roles for these molecules in

kidney induction. The WI-14- mouse is characterized by a lack of ureteric bud outgrowth

41. Targeted disruption of the RARa and y g e n s in mice results in a similar renal

abnormality 42. WT-1, RARa and RARy are expressed in the uninduced MM suggesting

that both play a role in kidney inductive events 4245. The phenotype of the RET-/- mouse,

described in section 1.4, includes a variable degree of failure of the ureteric bud to invade

the MM, grow and branch as well as large areas of undifferentiated mesenchyme 34. When

GDNF h c t i o n is abolished, the mice resemble the RET-I- mice in terms of renal

dysmorphology confirming that the RET signaling pathway is required for proper kidney

development 46-48. A role for the limb deformity (14 gene, mutations in which are

responsible for limb deformity in mice, in kidney induction is suggested by the phenotypes

of mice with recessive mutations in id. Mice homozygous for any of the id mutations have

rend dysplasia characterized by the absence or delay of ureteric bud outgrowth 4?

Molecules implicated in early kidney differentiation include bone morphogenic

protein (BMP)-7 and PAX-2. BMP-74- mice exhibit lethal kidney hypoplasia 50. These

mice develop normally until L2.5 dpc, however, fewer nephrons develop than in the wild

type mice and ureteric bud branching is stopped prematurely. Deregulation of PAX-2

expression, normally spatially and tempomlly restricted in the developing kidney. results in

failure to develop ureteric buds and mature nephric tubules Normally, PAX-2

expression is quickly downregulated in the developing kidney suggesting that properly

timed activation of PAX-2 expression is essential 51.

Molecules involved in subsequent differentiation of the maturing kidney include

WNT4 and WT-1 (reviewed in 39). WNT-4 transcripts are expressed in induced regions

of the developing kidney undergoing extensive differentiation during the mesenchyme-to-

epithelium transition and a role for the molecule in this process has been suggested 52.

Normally, WNT-4 is expressed in induced aggegates of mesenchymal cells and their

derivatives 52. The kidneys of WNT-41- mice display mesenchymal condensations but few

of these condensations epitheliahe or express such markers for late differentiation as PAX-

8. Branching of the ureteric bud remains unaffected 52. The upregulation of WT-I

coincides with downregulation of PAX-2 in cells differentiating into glomerular structures

53. WT-1 expression peaks in glomerular epithelium, one of the latest stages of kidney

differentiation 43* 44. Ryan et ol. 54 have shown that WI-l is in fact a repressor of PAX-2

expression in the mesenchymal cells which coexpress the two.

I. 7 Alternative Splicing

Alternative splicing at both the 5' and 3' ends of RET can result in multiple

different transcripts which translate into distinct proteins. W e little is known about the

function of the RET 5' isoforms (discussed above in section 1-2). in vitro studies have

suggested functional differences between RET9 and RET51, two of the 3' isoforms.

Activated RETS I constructs tramfected into a phaeochromocytoma cell line (PC 12) induce

more prominent neurite outgrowth than activated RET9 constructs 55. Binding assays have

suggested differences in the abilities of the EUX 3' isoforms to interact with the Src-

homology 2 (SEI2) and phosphotyrosine binding (PTB) domains of adapter molecules Shc

and Grb2 56. 57. This raises the possibility that RET9 and RET51 have different

downstream signaling capabilities .

Alternative splicing events occur in many developmentally important genes in

addition to RET. Individual members of the Trk family of RTKs use differential splicing of

alternative exons to generate transcripts encoding tmcated receptors that lack the entire

catalytic domain (reviewed in 58). The in vivo functional sigruficance of these truncated

Trk isoforms is unclear, however, roles in the active tramport of Ligaods or receptor

clearing have been suggested 59. TrkC, one of the Trk f d y members, additionally

encodes a family of full length receptors with catalytic domain insertions which coder

differential signaling properties to these receptors 60-62. This is similar to RE' where

alterations at the C - t e e of the RET isofonns affect interactions with downstream

effector molecules.

The fibroblast growth factor receptor (FGFR) f ' y consists of four

trammembrane receptor tyrosine kinases which have immunogIobulin (Ig) repeats in their

extracellular domains (reviewed in 63). Expression of the FGFRs is upregulated upon

cellular differentiation 64. The use of alternative splice sites within the sequence encoding

the extracellular domains of the FGFRs gives rise to muidpie protein isoforms with

differential ligand binding specificities 65-67. Two variants of FGFR-2 are generated from

differential use of exons encoding the third extracellular immunoglobulin domain. The

resulting receptors, keratinocyte growth factor receptor (KGFR) and bek (basic FGFR),

display different affinities for KGF/FGF-7 and basic FGFEGF-2 respectively 68. 69. Eariy

in development, the overall expression level of KGFR is consistently higher than that of

bek 70. Shi et al. 71 identified mutually exclusive expression patterns of the FGFR-2

variants which were tissue-specific suggesting different requirements for FGFR-2 variants

during development.

Due to exon skipping, similar to that seen in RET, the single id gene gives rise to

four major mRNAs (I-IV) that are expressed in two distinct patterns during development

72. Transcripts encoding isoforms EIII each encode a basic amino terminus but differ with

respect to coding exons. These isoforms are coordinately expressed during development in

the nervous system and kidney but not in the limb bud 72- 73. Isofom I transcripts contain

the full exon complement. Isoform I1 transcripts lack exons 2 and 3 while isoform III

-scripts lack exons 2 ,3 and 4 and, as a result of being translated out of frame, result in a

truncated product. Isoform IV differs £?om I-III in that it contains an acidic amino terminus

and is expressed in both the embryonic kidney and limb bud 73.

1-8 RET in Disease

1.8.1 RET Expression in Twnours und Turnour Cell lines

Turnours of neural crest origin have been shown to express RET at high levels.

Santoro er al. 74 found RET was consistently expressed in medullary thyroid carcinomas

and phaeochromocytow (see section 1.5). High levels of RET expression have

been reported in surgically resected neumblastomas and in human neuroblastoma cell lines

3 9 759 76- However, no mutations in RET have been identified in cases of neuroblastoma

examined 77.

RET expression has also been detected in promyelocytic and monocytic leukemia

cell lines 3 and papillary thyroid carcinomas 78-80- The latter are discussed in section 1.5.1.

1.8.2 RET Rearrangements in Papillary Thyroid Carcinoma

The RET protooncogene is found rearranged and constitutively active in a

proportion of papillary thyroid carcinomas (PTC) 78-82. The rearrangements occur in vivo

as tumour-specific events. The characterized recombination events result in the fusion of

the sequence encoding the RET tymsine kinase domain with the 5' sequences of different

genes. These 5' sequences contain elements responsible for RET expression in a cell type

where it is not normally expressed The chimeras encode molecules with Ligand-

independent dimerization ability. The first of these rearrangement events to be characterized

involved a paracentric inversion of the long arm of chromosome LO, inv( lO)(q 1 1.2q2 1) 83,

which resuited in the fusion of RET C-terminus encoding sequence with the H4 locus

(DIOS170) 81. Rearrangements involving the gene encoding the RIa regulatory subunit of

protein b a s e A 79 and ELI3 80- 82, a ubiquitously expressed gene localized on

chromosome 10, have since been described. Together, rearrangements in RET have been

recognized in up to 35% of papillary thyroid carcinoma cases 82.

1.8.3 RET in Inherited Syndromes

Germline mutations in RET are responsible for its conversion to a dominant

oncogene in the hereditary cancer syndrome multiple endocrine neoplasia (MEN 2)

subtypes 2A and 2B 8487. MEN 2A is- the most common MEN 2 subtype encompassing

greater than 90% of cases of MEN 2. Patients can present with medullary thyroid

carcinoma (MTC), phaeochromocytorna and/or hyperthyroidism (reviewed in 8 8).

However, penetrance is incomplete and 30% of patients have no symptoms of m C ,

phaeochromocytoma or hyperparathyroidism by the age of 70 (reviewed in 88). RET

mutations have been identified in 95% of MEN 2A families. Affected family members have

germhe rnissense mutations in sequence encoding one of five cysteine codons (codon

609,6 L 1,618 or 620 in exon 10; codon 634 in exon 11) 8" 89* These same RET

mutations have been identified in 85% of cases of familial medullary thyroid carcinoma

(FMTC) where MTC is the only clinical feature 86* 899 Two novel mutations associated

with FMTC have been identified. These mutations result in amino acid substitutions at

codons 768 (exon 13) and 804 (exon 14) and are predicted to modify the kinase activity of

RET 9 1-93

Clinical presentation of the MEN 2B subtype is similar to that for MEN 2A.

However, MTC and phaeochromocytoma develop much earlier in MEN 28 patients,

parathyroid involvement is rare and patients may present with developmental abnormalities

including ganglioneuromatosis, myelhated corneal nerves and marfanoid habitus (reviewed

in 88). Approximately 5% of all MEN 2 cases are the 2B subtype. A point mutation at RE'

codon 9 18 (exon 16) has been identified in 95% of MEN 2B patients 94-96.

HSCR occurs in 115000 Live births and can be sporadic or familial 97. Mutations in

the RET protosncogene have also been identified in 10-4045 of cases of Hirschsprung

disease (HSCR). a congenital abnormality characterized by the absence of sympathetic

neurons in the hindgut (reviewed in 88 and 98). The RET mutations associated with HSCR

are inactivating and result in loss or abrogation of RET function 99. Rarely, HSCR patients

have chromosome lOq11.2 deletions that result in complete loss of a single RET d e l e loo*

l . More common are mutations in RET which are found throughout the entire gene.

Frameshift and nonsense mutations are predicted to result in RET protein truncations 102-

107. Missense point mutations can also cause RET inactivation n- 99 (M. Billaud, personal

communication). No genotype-phenotype correlation has been identified to date. These data

suggest that haploinsufficiency for RET results in the disease phenotype.

I .8.4 RET Mutations in Sporadic Tumours

Somatic mutations of RET have been described in a proportion of sporadic MTCs

and phaeochromocytomas (reviewed in 88). The frequency of RET mutations in sporadic

MTC varies from 23 to 86% depending on sample size and population studied 94- 95- 108-

113- The majority of reported mutations affect RET codon 918, however, infkquent

mutations of codons 634.768 and 883 have also been described 95- 1099 1 13-1 16 (reviewed

in 88). In sporadic phaeochromocytomas. RET mutations have been identified in

approximately 10% of tumours 95- 1 17. 18. Reported mutations affect codons 620, 630,

634 as well as codon 9 18 9% 1 19.

RESEARCH OBJECTIVES

The objectives of this project were:

11 To identify RETexpression in developing human kidney ranging in gestational age from

7.5 through 24 weeks and to quantitate RETexpression in these samples.

21 To determine the expression patterns of both the 3' and 5' RET alternatively spliced

variants in human kidney development and quantitate relative expression levels of these

alternatively spliced variants.

31 To characterize the expression of GDNF and GDNFR-a, genes encoding members of

the multicomponent RET receptor complex, in developing human kidney.

41 To investigate a role for RET signaling complex molecules in disease by i] examining

expression of RET, GDNF and GDNFR-a in a panel of Wilms' turnours and renal cell

carcinomas and ii] by screening a panel of HSCR patients for GDNF mutations.

Materials and Methods

Oligonucleotides

The sequences of the primers used in PCR analyses are listed below. Primers were

synthesized either by Cortec (Kingston, Ontario) or by Genosys (The Woodlands, TX).

For reverse primers, the position of the most 3' base is given first.

The primers used for PCR amplification of the RET gene were 5- 84:

Forward primers used in RET 5' alternatively spliced variant expression analyses

consisted of the terminal 10 nucleotides of RET exon 2 and the first 10 nucleotides of the

exon to which exon 2 is spliced (any of exons 3.4'5 or 6) 5* 13* 84:

p25F exon ZS; 477 nt 5'-AGTGTCCGCAGACACCGTGG-3 ' ESA exon 5; 1203 nt 5'-GTACGGTCGCCCGCACGAAm-3'

Additional primers used in semiquantitative reverse transcription-polymerase chain

reaction (RT-PCR) anaiyses of RETexpression were designed using the program OIigo 18.

120,

The primers used for amplification of the RET 3' alternatively spliced transcripts

were according to Myers et al. 6. Nucleotide position start sites correspond to genomic

RET sequence:

The primers used in PCR analyses of GDNF expression were 121:

GDXlF exon 1; 3 nt 5'-GAATTATGGGATGTCGTG-3' GDN7R exon 2; 240 nt 5'-ATCTGGTGACCITITCAGTC-3 ' GDX3R exon 2; 45 1 nt 5'-CATCGCAAGAGCCGCTGCAG-3'

The primers used in PCR analyses of GDNFR-a expression were:

Primers specific for the P-glucuronidase gene (GUSB) used in semiquantitative

RT-PCR analyses were 18:

GUSB3 exon 2: 53 1 nt 5'-ACTATCGCCATCAACAACACACI'CACC-3' GUSBS exon 3; 725 nt 5'-GACGGTGATGTCATCGATGT-3'

RNA, Tissues and Samples

Hirscbsprung Patient Samples

DNA was extracted from the peripheral blood of a collection of HSCR patients by

Dr. Lois Mulligan as described in Mulligan et aL 122.

Fetal Kidney RNA and Renal Tissue Samples

Human fetal kidney samples ranging in gestational age from 8 through 24 weeks

were obtained from therapeutic abortions conducted at Children's Hospital in Montreal,

Quebec. Additional human fetal kidney sampies were obtained from the Central Laboratory

for Human Embryology at the University of Washington (Seattle, WA). All human fetal

kidney samples were collected under ethically approved protocols issued by the host

institution. Wilms' turnour samples were obtained from either the Divisions of Hematology

and Oncology, Children's Hospital Medical Center in Cincinnati, OH or the Children's

Hospital of Philadelphia, PA under ethically approved protocols 123. Three of six Wilms'

tuxnow samples used in analyses were obtained fkom mouse xenografts. The remaining

three samples were obtained from primary tumours. Renal tumour and corresponding

normal tissues were obtained following nephrectomies conducted at the Kingston General

Hospital with the help of Drs. Iain Young and Sandy Boag. Tissues were snap h z e n in

liquid nitrogen and stored at -800C. Human tumour analyses described in chapter 6 were

performed as a project which had separate and additional ethical approval.

Total RNA Extraction

RNA was extracted from human fetal kidney ranging in gestational age from 8

through 24 weeks by Dr. Lois Mulligan essentially by the method of Chomczynski and

Sacchi 1z4. Snap-frozen tissues were pulverized using a mortar and pestle. Ground tissue

or tissues too small to grind were transferred to a dounce homogenizer. One millilitre

TRIzol (Gibco-BRL, Gaithersburg, MD) was added per lOOmg tissue and the sample was

homogenized. Samples were transferred to sterile lOmL polypropylene tubes and incubated

at room temperature for a minimum of 5 minutes. 0.2 volumes chloroform was added to

the sample. Samples wen centrifuged in a Sowall RC-SB superspeed model (Dupont.

Wilmington, VA) at 10000 rpm for 15 minutes (40C) using Sorvall fixed angle rotor type

SS-34. The upper aqueous layer containing RNA was transferred to a new lOmL

polypropylene tube. 0.5 volumes isopropanol was added to precipitate the RNA. Samples

were centrifuged at 1OOOO rpm for 10 minutes (40C). The pellet was washed with 1

volume ethanol. Samples were then centrifuged at 8000 rpm for 5 minutes (4OC). The

ethanol was decanted and the pellets air dried for 5 to 7 minutes. Care was taken not to

allow the pellets to dry too long as this would make resuspension difficult, The pellets were

resuspended in 25pL diethyl pyrocarbonate @EPC)-treated water per 1mL TRIzoi. The

EWA concentration was determined by specaophotomeUy using an Ultraspec JII

spectrophotometer (Pharmacia, Baie d'Urf', Quebec). Absorbance readings at 2 6 h and

28Onm wavelengths were recorded RNA concentration was determined according to the

formula 125:

(absorbance at 260nrn)(40pg/mL,)(dilution factor)

where 40pg/mL is an estimate of the amount of RNA required to produce an absorbance

reading of 1. Integrity of RNA was c o w e d by elecnophoresis on ethidium bromide-

stained 1% agarose gels.

Reverse Traascription-Polymerase Chain Reaction

First strand cDNA synthesis

One pg of RNA was suspended in LOW DEPC-treated water, heat denatured at

700C for 5 minutes and cooled on ice briefly. F i t strand cDNA was generated fkom RNA

templates by incubating at 420C for 1 hour in 50mM Tris-HCl (pH8.3)- 8mM MgC12,

3OmM KCI, lOmM dithiothreitol (DTT). 2mM deoxynucleotide triphosphates (dNTPs)

(Pharmacia, Baie d'Urf6. Quebec), 7 units avian myeloblastosis virus reverse transcriptase

(AMV-RT) (Promega, Madison, WE) and 10 units human placental RNase inhibitor

(Pharmacia, Baie d'UrE, Quebec). 2jM random hexamer primers (Pharmacia, Baie

d'Urf6, Quebec) were used for RT-PCR analyses of RET and RET 5' alternatively spliced

variants expression. 2pM oligo dTls primer (Genosys, The Woodlands, TX) were used in

all other analyses. The enzyme was inactivated by incubating at 650C for 10 minutes then

freezing at -200C. Negative control reactions were performed identically but lacked RNA.

PCR Using cDNA Template

One tenth of a first s m d cDNA reaction was used in each PCR. Amplfication was

performed in solution containing lOmM Tris-HCl (pH8.3). 50mM KCl, 0.01% gelatin.

2OOpM d N T P s , 1 -5 units Taq DNA polymerase (Gibco-BRL, Gaithersburg. MD). 0.75-

1.75mM MgC12, 1p.M of both forward and reverse primers and deionized water to a

volume of to@. Samples were laye~d with light mineral oil and amplified in an

HybaidlOmnigene automated DNA t h e d cycler (Interscience. Markham, Ontario) using

1 min at 950C to denature, 1 rnin at 550C to anneal primers and 1 minute at 720C to extend

repeated for 40 cycles. Reactions were concluded with an extension period of 720C for LO

minutes. PCR products were eiectmphoresed on 2% agarose gels stained with ethidium

bromide and visualized under ultravioiet Light. LOO base pair molecular weight standard

(Pharmacia, Baie d'Urf6. Quebec) was used for band size estimation.

Semi-Quantitative RT-PCR

Experimental details for individual quantitative studies ace given in chapters 3-5.

Briefly, 0.1p.M [PPI-ATP labeled fonvard primer was included in PCRs to pennit liquid

scintillation of incorporated counts. cDNA templates were subjected to increasing numbers

of PCR cycles. PCR products were separated on 2% agarose gels and appropriate bands

were excised. Measurements of counts per minute (cpm) obtained fkom liquid scintillation

were plotted as a hction of cycle number to determine the range of cycles for which

amplification was hear.

Two-fold serial dilutions of cDNA template were subjected to cycles of PCR within

the linear range to determine amplification efficiencies. Where plots of cpm versus dilution

factor were linear, the conditions wen conside~d reliable for quantitation analyses.

Southern Blotting and Hybridization

Southern Blotting

PCR products from amplifcation of RET 5' alternatively spliced variants were

electrophoresed on 2% agarose gels in 1X TPE buffer at 120V for 2 hours. Gels were

denatured in 0.4N NaOH for 30 minutes. Gels were placed onto a sheet of 3MM filter

paper (Ahlstrom Filtration, Philadelphia, PA) wetted with 0.4N NaOH. The 3MM paper

was laid over a glass plate support such that the ends soaked in a reservoir of 0.4N NaOH.

A piece of Hybond N+ nylon membrane (Amersham, Arlington Heights, IL) cut to fit the

gel was wetted in 0.4N NaOH and placed directly on the gel. Care was taken to prevent air

pocket formation between the gel and the membrane. A single sheet of 3MM paper cut to fit

the gel was soaked in 0.4N NaOH and placed on the membrane. Multiple sheets of

adsorbent paper were cut to size and placed onto the 3MM paper. Capillary blotting

proceeded for a minimum of 16 hours. The nylon membrane was then washed briefly in

2X SSC and allowed to air dry prior to use in hybridization protocols.

Oligonucleotide Labeling

40pmol of oligonucleotide was radiolabeled in a reaction conraining 30pCi [y 32Pj -

ATP, 50mM Tris-HCl (pH 7.6), lOmM MgC12, 5mM DTT. LOOmM spennidine-HC1.

1OOmM EDTA (pH 8.0) and LO units T4 polynucleotide kinase (New England Biolabs,

Beverly, MA). The reaction was incubated at 37% for 1 hour then heated to 650C to

inactivate the enzyme. Oligonucleotide probes were ethanol precipitated prior to use by

adding two volumes ice cold 100% ethanol, 1/10 volume of ammonium acetate and 1pg

tRNA to the radiolabeled mixture. The samples were mixed well and cooled at -700C for 1

hour. Reactions were then centrifuged at 15000 rpm for 15 minutes (40C) using a Hettich

Mikroliter model D7200 microcenaifuge (Tuttlingen, Germany). The pellet was washed in

two volumes cold 7096 ethanol and the mixture centrifuged for 5 minutes at 15000 rpm

(4OC). The pellet was air dried then dissolved in deionized water.

Hybridization

Filters were placed in plastic hybridization bags and pre-hybridized at 37OC for 3

hours in 6X SSC, 0.1% SDS, 5X Denhardt's solution and lOOpg/m.L sheared and

denatured heterologous DNA. Re-hybridization solution was removed and a fresh aliquot

of the same solution containing the radiolabeled probe added. Hybridization proceeded

overnight at 370C in a shaking water bath. Following hybridization, blots were rinsed in

solution containing 2X SSC and 0.18 SDS then washed for 15 minutes in 2X S SC/O. 1%

SDS solution at room temperature. An additional wash in O.1X SSC/O. 1% SDS for 15

minutes at room temperature! was followed by washing in 0.1X SSCIO. 1 B SDS at 650C

for up to 30 minutes. The filters were exposed to Dupont Ulaavision G autoradiographic

6I.m (Wilmington, VA) with Dupont Lightning Plus intensifying screens (Wilmington, VA)

at -8OOC.

Sequencing

DNA Template Preparation

DNA templates for sequencing reactions were derived from PCR products

electrophoresed on 2% low melting point agarose gels (Gibco-BRL, Gaithenburg, MD).

Appropriate bands were excised, placed in sterile 1.5mL microcentrifuge tubes and heated

at 7OOC until agarose melted completely. PCR products were purified using the WizardIM

PCR preps DNA purification system (Promega, Madison, WI) according to manufacturers'

instructions.

Sequencing Reactions

AU sequencing reactions were performed using the 8hq cycle sequencing kit

(Amenham, Arlington Heights, IL) according to the manufacturer's instructions for use

with [f*P]-ATP end-labeled primer. Primers were labeled as previously described. Cycle

conditions for the termination reaction consisted of 950C for 30 seconds followed by 7 P C

for 120 seconds repeated for 30 cycles. Sequencing reactions were analyzed on 5%

denaturing acrylamide gels (see Appendix). Biorad's Sequi-Gen GT System sequencing

apparatus (Hercules, CA) was used to cast gels. 0.025% ammonium persulfate and

0.0 15% TEMED (Fisher Chemicals, FairIawn, VA) were included to polymerize gels.

Denaturing gel loading buffer was added to samples as 53% reaction volume, samples were

heated at 950C for 5 minutes and maintained at 850C prior to loading. Sequencing

reactions were electrophoresed in 1X TBE at a constant voltage of 2000V for 1.5-2 hours.

Gels wereblotted to 3MM paper and dried using Biorad's model 583 gel drier (Hercules,

CA). Dried gels were exposed to Dupont Ultravision G autoradiographic film (Wilmington,

VA) at -800C with Dupont Lightning Plus intensifying screens (Wi ig ton , VA)

Chapter 3

Expression of RET and its Multiple Splice Forms in Developing Human Kidney

1 performed all the analyses of RET and RET 5' splice variant expression in human kidney

development. The results of this work have been accepted for publication in Oncogene 18.

Introduction

The mammalian kidney develops through a series of reciprocal inductive

interactions between two cell groups that share a common mesodemal origin: ureteric bud

epithelium and metanephric mesenchyme 35. Upon induction. the ureteric bud an

outgrowth of the Woffian duct, grows and branches repeatedly to produce the collecting

duct system of the kidney while undifferentiated metanephric mesenchyme condenses and

epithelializes, forming glomeruli and proximal and distal convoluted tubules. A number of

molecules, including WT-141, WNT-4 52 and Pax-2 I26 and members of the Hox family

(reviewed in 39), have been implicated in induction of these differentiating events, as

described in Chapter 1 (see Figure 1.5). Recent studies have suggested that the RET proto-

oncogene may also be essential to these processes. RE'Tencodes a receptor tyrosine kinase

with roles in migration, development and survival of neural crest cells and their derivatives

14-16- 127. Multiple RET transcripts, generated by alternative splicing of RET exons

encoding the extracellular Ligand-binding and C-terminal domains, are expressed in cells

and tissues derived from the neural crest, branchial arches and ureteric bud I*-14.

Several pieces of evidence have suggested the involvement of RET in kidney

induction. Fist, in situ hybridization studies in murine embryos have identified RET

expression in the kidney as early as 8.5 dpc with highest levels in that portion of the

Woffian duct that will evaginate to produce the ureteric bud 14. RET expression is later

localized to the ureteric bud epithelium and the branching tips of the bud. Levels of RET

peak at 11.5 dpc then decrease with time. Second. the phenotype of the RET-I- mice

indicates that RET is essential to kidney development RET -1- mice display kidney

agenesis or severe dysgenesis characterized by reduced branching of the ureteric bud.

absence of a mature collecting tubule system and large areas of undifferentiated

mesenchyme 34. Third, ureteric bud tissue fiom RET+f+ mice is able to grow and branch

in culture in the presence of REI+/+ or -1- rnetanephric mesetlchymal tissue. However,

ureteric bud tissue from RET-/- mice fails to branch in the presence of either REW- or

RET+/+ mesenchyme suggesting that RET 4- bud is unable to transduce a mesenchyme-

derived signal 34-40. Finally, several recent studies have identified a ligand for RET which

is also implicated in kidney morphopnesis. The RET ligand is a multimeric complex

composed of GDNF, a soluble molecule, and GDNFR-a, a GPI-linked cell surface

receptor without an intraceIlular domain which must interact with RET to trmsduce signals

intraceilularly 21*2. GDNF -I- mice have phenotypic characteristics similar to their

RET-I- counterparts: kidney agenesis or severe dysgenesis and absence of the enteric

ganglia 46-48. Consistent with RET's proposed role as a transducer of mesenchyme-derived

signals in kidney induction, GDNF expression has been localized to the metanephric

mesenchyme 29.

Although murine studies have suggested developmental regulation of RET in

kidney organogenesis, the expression of RET in human kidney development has not been

investigated. In this study, we examined RET expression in a panel of human fetal kidney

RNAs using semiquantitative reverse transcription-polymerase chain reaction (RT-PCR).

We have found that RET expression is developmentally regulated with highest expression

in the earliest gestational ages decreasing with time and with lowest levels of RET being

detected in adult kidney. In addition, we show that the developmental expression patterns

are dissimilar amongst the various RET alternative transcripts. Our data suggest that an

alternatively spliced form of RET. which encodes a product lacking a portion of the

extracellular ligand-binding domain (Figure 1.3), has a higher prevalence and perhaps a

greater sigruficance at the earliest developmental stage of the human kidney.

Results

RET Erpression in Human Kidney Development

We examined the expression of RET in a panel of RNAs prepared from human fetal

kidneys ranging in gestational age from 8 to 24 weeks and from adult kidney. RNA from

medullary thyroid carcinoma cell line TI; known to express JET at high levels, was used

as a positive control 74- 128. Owing to limited quantities of RNA available for analysis, RT-

PCR was used to investigate RET expression. Initial analyses were conducted using

primers CRTl7S and CRT17A which amplify RET sequence coding for the region just

within the RET tyrosine base domain 9. This 122 bp amplicon is present in all

transmembrane RET isoforms and would thus serve as an indicator of overall RET

expression. RT-PCR of total RNA identified an amplification product of the predicted size

for a l l ages of fetal kidney examined as well as for adult kidney and the TT cell Line (data

not shown). Initial inspection suggested that the level of RET amplification for adult kidney

was relatively lower than the levels observed for our fetal kidney samples and the IT cell

he, suggesting a developmentally regulated RET expression pattern. This is consistent

with previous mouse and rat studies which showed that levels of RET expression decrease

with time during kidney development 1.14-16.

Quantitation of RET EApression During Human Kidney Development

To investigate relative REl expression levels during human kidney organogenesis,

we developed a semiquantitative RT-PCR assay that compared relative expression level of

RET and a housekeeping gene, P-glucumnidase (GUSB), in a given RNA sample using

multiplex PCR. Qualitative RT-PCR is used to determine the presence or absence of gene

expression whereas quantitative RT-PCR is used to determine how many target molecules

are present in a given sample. Semiquantitative RT-PCR lies somewhere in between the

two and is used to investigate the expression level of a target gene relative to a standard

reference gene. The accepted term for assays investigadng the comparison of relative

expression levels is semiquantitative RT-PCR 13'. Previous studies have shown that

GCrSB is consistently expressed in kidney throughout the examined developmental stages

129- 130. PCR was performed with primers CRTlTS and CRT17A to amplify RET and

GUSB3 and GUSBS to amplify a GUSB product of 195 bp. A LO: 1 molar ratio of RET

primers to GUSB primers proved to be optimal for co-amplification of both products. PCR

amplification proceeds exponentially; therefore, for reliable quantitation of product, PCR

conditions must be selected so that quantitation is pedormed in the linear phase of

arnplitication where concentrations of product are proportional to starting levels of target

131. Experiments were conducted to determine the kinetics of arnpiif?cation over a range of

cycle numbers to select conditions for which amplification was linear. End-labeled forward

primers were included in PCRs and the quantity of PCR product was represented by the

amount of radioactive incorporation. Template cDNAs were amplified for 12, 16, 20, 24,

28, 32 or 36 cycles. The relevant products were quantitated by separation on 2% agarose

gels, excision of the appropriate bands and measurement of incorporated counts (cpm). We

found that amplification was linear for both GUSB and RET between 18 and 26 cycles

(Figure 3.1). To ensure that amplification occurred with e q d efficiency regardless of

amount of starting template, we performed PCR using a two-fold serial dilution of TI'

cDNA for 22, 25 or 26 cycles. A linear relationship between input cDNA and amount of

PCR product was identified for each of GUSB and RET when PCR proceeded for 25

cycles (Figure 3.1).

In our analyses, we chose to investigate RETexpression in six kidney RNA

samples that spanned the range of gestational ages available: 7.5,8.5, 10.5, 14, 18 and 24

weeks gestation as well as adult kidney. The reliability of the conditions established for

quantitation of transcripts in the TI' cell Line was confumed for fetal kidney. Quantitation of

RET expression relative to GUSB expression during human kidney development was

performed at 25 cycles of PCR. Semiquantitative assays were repeated five times for each

of the human fetal kidney samples as well as for human adult kidney and mean expression

Figure 3.1 Determination of conditions for semi-quantitative RT-PCR analysis of o v e d

RET expression in human fetal and adult kidney. A] Kinetics of simd*laeous ampLification

of RET and 8-glucuronidase (GUSB). Amounts of REl and GUSB product amplified in a

single multiplex reaction were quantitated by liquid scintillation counting of radioactivity

incorporated in PCR products which had been resolved on 2% agarose gels and excised.

Counts (measured in cpm) were plotted relative to the number of ampLification cycles. The

exponential reaction was linear between 22 and 26 cycles. At 26 cycles, rates of

amplification for both RET and GUSB products decreased and approached plateau B]

Confirmation of amplification efficiencies of RET and GUSB primers used in multiplex

PCR for semiquantitative analyses of overall RET expression. A two-fold serial dilution of

TI' cDNA was subjected to 22, 25 or 26 cycles of PCR. The results for 25 cycles are

shown. Linear results were obtained when counts were plotted relative to dilution factor for

amplification of both RET and GUSB transcripts indicating comparable amplification

efficiencies for these primer pairs.

4 b

too0

, . b

a m - 100 a* -

LO me- GUSB S I

4 b . b

I I 1 I m I 1 ~ 1 ~ 1 I

16 18 20 22 24 26 28 30 32 34 Number of PCR Cycles

Two-Fold Serial Dilution Factor

with standard deviation calculated. The expression of REl relative to GUSB was assessed

by comparison of incorporated counts, where RET product was expressed as a fiaction of

GUSB product (Table 3.1). RETexpression levels were highest in the 7.5 week kidney

but lower in the 14 week sample by 3.7 fold RET expression differed between 14 week

and 18 week kidney by 1.5 fold. Expression was only slightly Lower in 24 week kidney.

differing from expression in 18 week kidney by about 2%. RET level in adult kidney was

1.2-fold lower than those in 24 week kidney and almost 7-fold lower compared to 7.5

week kidney. Thus, RETexpression appeared to decrease with increasing gestational age.

RET Alternative Splice Variants

Our initial analyses indicated that RET was expressed in a developmentally

regulated fashion in the human kidney. Although this may reflect overall reduction in RET

expression, it could equally represent differences in expression of the many RET mRNA

forms generated by differential splicing. Alternative RET transcripts in which exon 2 may

be juxtaposed to any of exons 3, 4, 5 or 6, resulting in a shorter putative RET ligmd

binding domain, have been reported (Figure 1.3) 13. Using the sample panel described

above, we investigated the expression of these variant RET transcripts during human

kidney development In order to distinguish the various RET -As generated by

alternative splicing of RET 5' exons, RT-PCR analyses were conducted using primer

combinations that ampLified each independently and exclusively (Figure 3.2). Speciticity

was ensured by using a forward primer that spanned the splice junction of exon 2 and the

downstream exon and a reverse primer within the downstream exon. Amplicoos predicted

for full length (FL) transcript and those lacking exon 3 (2/4), exons 3 and 4 (215) and

exons 3,4 and 5 ( 2 6 ) were 210 bp, 262 bp, 196 bp and 148 bp respectively.

PCR amplification specific for each aanscript was performed using the panel of

fetal and adult kidney samples descnid above. Products were resolved on agarose gels.

Southern blotted and probed with an end-labeled oligonucleotide specific for the 3' exon

Table 3.1. RET expression ~lat ive to GUSB expression in human fetal kidney.

Quantitation of RET relative to GUSB was based on multiplex RT-PCR analyses. GUSB

expression levels have been normalized to 1-00 in al l samples- Means and standard

deviations (sd) for five repeats are given in the table.

Sample Mea-sd

7.5 week 6.68+0.05

8.5 week 3.83k0.27

10.5 week 2.53&0.20

14 week 1 -83+0. 1 1

18 week 1 -29H.02

24 week 1 -25fl.02

Adult 1 *05+0.0 1

Figure 3.2. Expression of RETmRNA forms generated by alternative splicing at the 5' end

of the gene in human fetal and adult kidney. PCR amplification of the four RET mRNA

forms was conducted using primer pairs described in Chapter 2. The predicted product size

for each amplification, as well as the exon location of the primers used to amplify the

specific products, is indicated. PCR products were resolved on 2% agarose gels, gels were

Southern blotted and membranes were hybridized to a second labeled primer corresponding

to the 3' exon.

present in each amplified product. A single PCR product corresponding to each of the

predicted sizes was observed in all samples of fetal and adult kidney indicating that each

RET 5' spliced variant is expressed from 8 weeks through 24 weeks gestation (Figure

3 -2).

Quuntitation of RETAItemolive Splice Varimts

We addressed the possibility of variation in expression of RET 5' alternative splice

variants using a semiquantitative RT-PCR assay similar to that described above. W e

compared the expression level of each RET alternative transcript to that of FL RET in our

panel of human kidney RNA by pairwise amplifications using three primers in each

reaction: a single forward primer in exon 2 (X2S) and two reverse primers, one in exon 3

(CRT7A) and the other in one of exons 4 (X4A), 5 (E5A) or 6 (E6A). In combination with

a f o m d primer in exon 2, the latter primers would amplify alternative splice variants 2/4,

215 and 2/6 respectively. Although, in theory, the forward primer and any of the reverse

primers would amplify full length transcripts as well, these Longer products were not

detected.

As described above, optimal RT-PCR conditions for comparison of expression of

the different RET transcripts were chosen based on experiments which defined the linear

range of amplification. Amplification was found to be linear between 22 and 30 cycles for

all primer combinations (data not shown). Quantitative PCR analyses were performed at 28

cycles of amplification which was both within the hear range for all primer combinations

and permitted visualization of even the least abundant transcript upon electrophoresis. To

ensure that amplification occurred with equal efficiency regardless of amount of starting

template, we performed PCR using two-fold serial dilutions of 14, 18 and 24 week human

fetal kidney samples as well as adult kidney and TT cell line cDNAs.

Initially, we investigated RET expression in RNA samples From 14, 18 and 24

week as well as adult kidney. RT-PCR experiments were performed at 28 cycles of

amplification and the expression levels of alternative RET transcripts were calculated as

fractions of ET RET. Quanatation of each alternative RET =dpt in a sample was

repeated twice for each of three independent cDNA preparations and the data summarized in

Table 3.2. We found that FL, RET was the most highly expressed RET transcript in each

sample analyzed The expression Levels of the RERI4 and REZ2/5 W p t s were both

much lower than those of FL, RET . These aaDmipts maintained a level of expression

approximately 113 that of FL, RET at al l stages of development

The REN6 transcript was the least abundant transcript in all samples tested, being

expressed.at 0.22fl.02 times the levels of K RET in 14 week kidney and at lower levels

in 18 week (O.18fl.02) and 24 week kidney (O.l&O.O3) (Table 3.2). In adult kidney, the

expression level of W 6 was approximately 116 that of FL RET. Our data suggested that

expression of the RETU6 transcript was higher in eadier stages of human kidney

development relative to other RET 5' splice variants and decreased more rapidly through to

24 weeks gestation.

In order to confirm this observation, we compared the expression of FL RET and

W 6 in our full panel of human fetal kidney RNAs. The expression level of the R E W 6

transcript relative to FL, RET was calculated as described above. We found that RE7216

was expressed at higher levels in the eariiest gestational ages of human kidney we

examined (0.26&@02 in 8 week kidney), approaching the relative levels of -4 and

REZ2/S seen throughout our tested samples (0.15H.02 in 24 week and 0.149.02 in adult

kidney) (Table 3.3, Figure 3.3).

Discussion

The mature metanephric kidney arises from inductive interactions between epithelial

(ureteric bud) and mesenchymd (metanephric mesenchyme) cells. It is clear that this

complex process involves a number of biochemical signals rather than a single molecular

event. The RET RTK and members of its Ligand complex, GDNF and GDNFR-a, are

likely to play roles in both kidney induction and morphogenesis. Using semiquantitative

Table 3.2. RET 5' splice variant expression relative to full length RET (FL, RET)

expression in human fetal kidney. Quantitation of multip1e RET alternative transcripts

dative to FL RET was based on muitiplex RT-PCR analyses. Means and standard

deviations were calculated from the results of six individual assays.

RE71u4 RE;IZ/S W 6

Sample FL REl Mkanksd Neanksd Meauksd

14 week 1 .OO 0.32H.04 0.3 1H.04 0.22H.03

18 week 1 .OO 0.3SH.05 0.3 1H.06 0.20_+0.02

24 week 1 .OO 0.3m.05 0.33fl.05 0.15@.02

Addt 1 .OO 0.32H.05 0.29fl.05 0 .1w.02

'IT ceUs 1 .OO 0.3 1H.04 0.3 1fl.04 O.l4&0.02

Table 3.3 Quantitation of RETZ16 relative to FL RET based on muItiplex RT-PCR

analyses. Means were calculated based on six individual v a t s of the assay.

-6

Sample FL RET MeankSD

8 week 1-00 0 . 2 w - 0 2

12 week 1 .OO 0.25H.04

13.5 week 1 .00 0.25fl.05

14 week 1.00 0.22&0.03

14.5 week 1.00 0.23+0.04

15 week 1.00 0.23a.03

15.5 week 1.00 0.2 1fl.02

16 week 1.00 0.20+_0.03

16.5 week 1.00 0.2 1M.03

18 week 1.00 0.20+0.02

20 week 1.00 0.18~0.04

21 week 1.00 0.19k0.02

22 week 1 .00 0. I7fl.02

23week 1 .OO 0.15fl.02

24 week 1.00 0.15+0.02

Adult Kidney 1.00 O.lck_O.O2

TT cells 1.00 O.lS+0.02

Figure 3 -3. Developmental expression of RET 5' splice variants in human fetal kidney.

A] Levels of alternatively spliced RM transcripts were calculated as fractions of the FL

RET transcript amplified in the same RT-PCR reaction for a given sample. After

electrophoresis of PCR products, quantitation was achieved by scintillation counting of

radioactivity incorporated in PCR products, made possible by including 0.1w of end-

labelled s e w strand primer in reactions. B] Rome of developmental expression of

RETU6 transcript generated by splicing exon 2 to exon 6. Expression levels of RETU6

were calculated as fractions of FL RET as described above.

0 12 14 16 18 20 22 24

Gestational Age of Kidney (weeks)

rn FLRET RETU4

A RETUS RETU6

Gestational Age of Kidney (weeks)

RT-PCR methodology, we have shown that RET is expressed in human fetal kidney

during the 8 to 24 week period of gestation. Further, RETexpression is highest early in

development but decreases through to 24 weeks gestation relative to expression of a

housekeeping gene control. This is consistent with the RETexpression pattem observed in

mice and rats where RETexpression has been localized to the ureteric bud in early kidney

development but later appears in the actively growing tips of the bud branches lCL6.

The relatively high RET expression level early in development coincides with a

stage of kidney morphogenesis in which the ureteric bud undergoes rapid bifiucation (fkom

8 until 14 weeks human gestation). A period of rapid branching is initiated at 8 weeks and

continues until 14- 15 weeks when the rate of branching slows. During this period, cells of

the metanephric mesenchyme condense, epithelialite and begin to differentiate to form the

nephrons while the ureteric bud continues to branch and bud cells proliferate (reviewed in

37). While the ureteric bud is active fkom the time of its development through to 36 weeks

intrauterine Life, the role of the bud varies during the process of kidney development.

During this period, the bud induces condensation and epithelialization of the surrounding

mesenchyme, resulting in nephrons which attach to tubules that form within the zone of

bud growth. Branching seldom occurs between 20 and 32 weeks (reviewed in 37). Given

the relatively higher expression level of RET we observed in 7.5 week human fetal kidney

compared to 24 week fetal kidney, it is possible that RET's significance changes according

to the stage of renal development. The RET expression pattern suggests a more significant

involvement of RET in earlier developmental stages, however, whether it functions in cell

differentiation, proliferation or both requires further investigation.

The developmentally regulated pattern of M T expression in human kidney

suggests that RE' may have varying roles or signiscame at different stages of kidney

development. It is possible that this effect is simply mediated by the relative level of overall

RET expression. However, it is also possible that differences in expression of the various

RET alternatively spliced transcripts &ect RET's contribution to the process of kidney

development. Multiple isoforms of RET arise as a result of alternative splicing events 6. 12+

L3. These RET isoforms may have different roles in human renal development, perhaps by

recognizing different RET ligands or binding the same Ligand with different affinities. We

have shown that all RE15' spliced variants are expressed in the human kidney throughout

the 8 to 24 week gestational period but that the relative levels of these alternative RET

forms vary (Figure 3.3). FL RET was most abundant in all samples while -4 and

RET 2/5 transcript expression levels were consistently about 113 that of FL RET.

Interestingly, however, REW6 expression levels varied with respect to all other RET

splice forms, being relatively more abundant in the early kidney developmental stages

(0.26fl.02 in 8 week sample) and decreasing more markedly to low adult levels

(0.15fl.02) (Figure 3.3).

Although all RET mRNA forms are expressed at relatively high levels during early

kidney development and decrease with time, the variation in RETU6 expression is most

striking. In neural crest-derived cells, RET is predicted to function in both differentiation

and proliferation processes. Activated fonns of RET have the ability to induce

differentiation in PC 12 cells 132-134. In contrast, activated RET is transforming in N W T 3

cells and results in proliferation 81- 134. 135. In human kidney, it is not clear which role is

chiefly attributable to RET. However, it is conceivable that different RET isoforms may

have different roles. Lorenzo et al. examined the relative levels of RET 5' alternatively

spliced transcripts in a panel of hunours of neural crest origin 13. In these analyses,

RE12/6 expression was approximately 1/10 FL RET expression, similar to the expression

levels we detected in adult kidney. Our data show that the high expression level of RE2216

coincides with the period of rapid bud bifiucation and growth of the ureteric bud. If

R E W 6 was simply a marker of rapidly dividing cells, we would expect to see similar

levels of this transcript in developing fetal kidney and nunours. However, our data suggest

that levels of RETU6 are comparable in neural crest-derived tumours and cell lines and in

human adult kidney, tissues with very different proliferative profdes. In fact, RER/6 was

2.5-fold more abundant in 8 week human fetal kidney than in either human adult kidney or

the TT cell line. These data suggest that the RER16 product may have a non-proliferative

role in human kidney morphogenesis.

Many RTKs generate multipie isoforms by alternative splicing. When such events

occur near the 5' end of the gene, ligand-binding domains can be affected resulting in

altered ligand-binding specificities. In the case of fibroblast growth factor receptor type 2,

for example, alternative exon usage in the immunoglobulin-like loop region results in

receptor variants with different Ligand-binding affinities 6** 69- These transcripts are

expressed in a tissue- and developmental stage-specific fashion 64* 70- The relatively

high expression of RERI6 in early human renal development may represent a requirement

for this specific RJX isofom during that time. The RETU6 transcript encodes a molecule

that lacks 239 amino acids spanning four potential N-linked glycosylation sites, seven

cysteine residues and the region of cadherin homology which are a l l present in FL RET

(Figure 1.3). Ligand binding is thought to involve all or part of this region 9. Given the

decreased Length of this region in -6, it is possible that this molecule binds a different

ligand than its full length counterpart or that it binds the same ligand with differential

affinity.

Recently, the RET ligand was identified as a novel form of multicomponent Ligand

complex, consisting of GDNF and GDNFR-a 21. 22. GDNF is a soluble molecule which

forms disulfide-linked dimers that bind to a novel GPI-linked protein, GDNFR-a. The

GDNFIGDNFR- a complex then binds to RET which dirnerizes. resulting in

autophosphorylation and downstream signaling 21. 33. We have identified expression of

both GDNF and GDNFR a in human fetal kidney ranging in gestational age from 7.5 to

24 weeks (discussed in Chapter 5). Expression of these transcripts during kidney

development is consistent with their role as RET ligands. The phenotype of RET-I- and

GDNF-f- mice and in vitro explant studies have clearly shown that both RET and GDNF

are required for inductive events during kidney development. Interestingly, the differences

in phenotype between RET-I- and GDNF-/- mice have suggested that another RET ligand

may exist. In REF/- mice, the superior cenical ganglion neurons are absent 349 36 while

they are only moderately reduced in GDW-/- mice 46-4*. In other respects. these mice have

very similar phenotypes. The RETU6 transcript is predicted to encode a molecule that

differs fiom its full length counterpart in that it lacks a portion of the protein thought to be

involved in Ligand-binding. Thus, RETU6 is a logical candidate as a receptor for an

alternative ligand.

Our data have shown that multiple RET 5' splicing variants are expressed in a

developmentally regulated fashion during human kidney development (Figure 3.3). 'The

relatively high expression levels and clear developmental differences in expression of these

variants suggest that they represent fuactional isofonns and not tolerated artifacts of

splicing. Interestingly, RET2/6, the RET isoform with the shortest extracellular domain,

had the strongest developmental variation in expression. REW6 lacks part of the predicted

region of ligand binding present in FL, RET which raises the possibility that it interacts with

a different Ligand with relevance to mammalian kidney development. Thus, metanephric

kidney formation may be dependent on more than one role for RET: the

RETIGDNFIGDNFEt-a interaction and another as yet uncharacterized signaling molecule.

Materials and Methods

Growth of 77' Cells

The medullary thyroid carcinoma cell line 'IT was grown at 370C in humidified air

in L-L5 medium (Gibco-BRL Life Technologies) supplemented with 10% fetal calf serum

(Wisent) and gentamycin (50 pg/ml). Cells were grown as monolayers and replated

weekly.

RNA Eitraction

Total RNA was isolated fiom TT cells using the TRIzol method according to

manufacturers' instructions as described in Chapter 2 (Gibco-BRL Life Technologies).

Total RNA was isolated &om adult kidney tissue and feral kidney tissue ranging in

gestational age Earn 8 weeks to 24 weeks essentially by the method of Chomczynski and

Sacchi 124-

Feral Kidney, Adult Kidney and lT RNA Analyses

One pg of total RNA was heated at 7WC for 5 minutes and cooled on ice briefly.

First strand cDNA was generated from these templates by incubating in 50mM Tris-HCl

(pH 8.3), 8mM MgC12. 3 W KCl, lOmM dithiothreitol 0. 2mM deoxynucleoside

hiphosphates (dNTE%), 7 units AMV reverse transcriptase (Promega). 2 units human

placental RNase inhibitor (Pharmacia) and either Z p M random hexamer primers

(Pharmacia) or 2p.M oligo dT15 primer (GENOSYS) at 420C for 1 hour. To inactivate the

enzyme, samples were incubated at 650C for 10 minutes and then fkozen at -200C. PCR

amplification of cDNA templates was pedonned in lOmM Tris-HCl (pH 8.3). 50mM KCl.

0.0 1% gelatin. 1pM of each primer. 2OOw dNTPs and 1.5 units Toq DNA polymerase

(Gibco-BRL Life Technologies) with 0.75-1.5mM MgCI2 depending on primers being

used. Unless otherwise stated, PCR amplification was carried out for 40 cycles of 95oC 1

min/S50C 1 min/72OC 1 min.

Semi-quuntifative PCR

himers used in semiquantitative RT-PCR analyses are described in Chapter 2. For

~uantitation experiments, RT-PCR was performed in 10pL volumes using the PCR

conditions described above but including O.1p.M 32P-end-labeled sense-strand primers and

varying the amounts of cold primers used and MgC12 buffcr concentrations. Specifically, in

multiplex PCR to amplify RET and GUSB products, 1pM RET primers (CRT17S,

CRT17A) and 0.1pM GUSB primers (GUSB3, GUSBS) were used in 1SmM MgClz

buffer conditions. To amplify the REl alternative splice variants, 1p.M concentrations of all

primers were used in pairwise amplification experiments (CRVAK2SIX4A,

CRT7A/X2S/ESA, CRT7A/X2S/E6A) in buffer containing 0.75mM MgC12. Analyses to

determine the hear range of amplification wen performed using 1/10 of a cDNA reaction

as described above in PCR for 12, 16, 20, 24, 28, 32 or 36 cycles. PCR products were

separated on 2% agarose gels and appropriate bands were excised. Quantitation of PCR

products was determined by liquid scintillation of incorporated radioactivity. Plots of

incorporated counts (cpm) versus cycle number were examined to define the range of PCR

cycle numbers for which amplification was linear (Figure 3.1). Two-fold serial dilutions of

cDNA were subjected to cycles of PCR within the hear range to confirm the number of

cycles which would amplify with similar efficiency regardless of amount of starting

template. PCR products were separated on 2% agarose gels, appropriate bands were

excised and Liquid scintillation counting was used to measure incorporated radioactivity

(measured in cpm). If the relationship of cprn versus dilution factor was bear, the

conditions were considered reliable for quantitation purposes (Figure 3.1). For quantitation

of RET relative to GUSB, 25 cycles was chosen for al l funher studies while for

quantitation of RET alternative splice variants relative to FL RET, 28 cycles was used for

al l fuaher experimentation.

Southern Blomkg

PCR products were resolved on 2% agarose gels and transferred to Hybond N+

membranes (Amenham). OLigonucleotide probes were labeled in a reaction containing

30pCi **P]-ATP. 10 units of T4 polynucleotide kinase (Gibco-BRL Life Technologies),

SOmM Tris-HC1 (pH 7.6), lOmM MgC12, 5mM D'IT, lOOmM spermidhe-HC1, lOOmM

EM'A (pH 8.0) at 370C for 30 minutes. Probes were ethanol precipitated prior to use.

Blots were prehybridized and hybridized at 37OC as described in Lorenzo et al. '3 and

washed in 0. 1XSSCIO. 1 %SDS at 650C.

Chapter 4

Expression of RET 3' Splicing Variants During Human Kidney

Development

I performed all the analyses of RET 3' coding variant expression in deveioping human

kidney. The results of these experiments have been submitted to the journal Oncogene for

publication.

Introduction

The development of the mature mammalian kidney requires a series of reciprocal

inductive interactions between ureteric bud epithelium and metanephric mesenchyme.

Mouse metanephric kidney development begins around 10.5-1 1 dpc when induction of the

WoEl5an duct produces a diverticulum, the ureteric bud 35. Metanephric mesenchyme, a

dense blastema of mesenchyme in the vicinity of the W o E a n duct, induces the ureteric bud

to grow and branch repeatedly giving rise to the collecting duct system. At the same time,

the ureteric bud induces undifferentiated metanephric mesenchyme to condense,

epithelialize and uldmately differentiate to form the glomeruli and proximal and distal

convoluted tubules. The precise mechanisms of the differentiation events are unclear.

However, a number of molecules, including WT-I 41, W-4 52, PAX-2 126 and HOX

family members (reviewed in 39), have been implicated in the induction of these processes.

In addition, recent studies suggest a requirement for the RET receptor tyrosine kinase in

kidney induction. RET is expressed in cells and tissues derived from the neural crest,

branchid arches and kidney 12- 14. In situ hybridization studies have localized RET

expression to the ureteric bud in embryonic rnurine kidney i4. Consistent with RE'

involvement in tbe induction of kidney morphogenesis is the phenotype of the RET -/-

mouse which includes kidney agenesis or severe dysgenesis 34. In v i m tissue mixing

experiments using ureteric bud and mesenchyme tissues from RET-I- murine embryos and

normal control murine embryos coatirm that hctional RET expression in the ureteric bud,

but not in rnetanephric mesenchyme. is required for the induction of bud growth and

branching 34- 40. RET's proposed role as a transducer of mesenchyme-derived signals in

kidney induction is supported by the localized expression of GDNF, which encodes a

soluble member of the RET ligand complex, to the rnetanephric mesenchyme 20. 29.

Phenotypic similarities between REF/- and mice null for GDNF, including kidney

ageaesis or severe dysgenesis, have recently been descnid confirming the requirement for

both RET and GDNF in kidney induction 4648.

RET encodes multiple transcripts resulting b m alternative splicing of 5' or 3 '

exons 6- 12- 13- Splicing of exon 2 to any of exons 3, 4, 5 or 6 at the 5' end of RET

generates transcripts predicted to encode proteins with altered extracellular domains. As

described in Chapter 3. expression of RET 5' alternatively spliced variants is

developmentally regulated during human kidney organogenesis 18.

Alternative splicing at the 3' end of RET results in transcripts encoding RET

isoforms with different C-tennind amino acids 6- 12. RET exon L9 is present in all

transcripts, however, Merentid splicing at the 3' end of exon 19 results in transcripts

where exon 19 is unspliced spliced to exon 20 or spliced to exon 21 (Figure 1.2). These

transcripts encode RET isofonns with distinct 9 (RET9)' 5 1 (RETS I) or 43 (RET 43) C-

tenninal amino acids respectively. Multiple polyadenylation sites and 3' UTRs associated

with these three coding variants have also been identified 6. 12. h this chapter. we have

confined ourselves exclusively to the discussion of the tbRe different coding sequences

irrespective of the associated UTR. The sequences of the proteins encoded by the RET 3 '

alternatively spliced transcripts diverge at amino acid 1063 12. The final common amino

acid for all three isoforms is a tyrosine (Y1062) which has been shown to be

phosphorylated on RET activation 136. Alternative splicing places Y 1062 in different amino

acid contexts in the three RET isoforms coderring it with different binding capabiIities 9.

In the sequence context of RETS 1, pY 1062 can interact with the Shc-PTB domain (Figure

4.1). While pY1062 in the sequence context of RET9 can also interact with the Shc-PTB,

the same pY has a higher affinity for the ShcSH.2 domain F~gure 4.1) (M. Billaud,

personal communication) 57. RETS 1 contains two additional tyrosine residues (Y 1090 and

Y lO96) not present in either of RET9 or RET43 6. 12. Y lO96 has been identified as an

autophosphorylation site 136 able to interact with Grb2-SH2 domains directly 4.1)

56. Y 1090 has not been demonstrated to become phosphorylated on RET activation.

Functional evidence that these differences in effector binding alter cell phenotype

has been shown by Rossel et al. who observed more prominent neurite outgrowth in PC 12

cells transfected with activated RET5 1 compared to activated RET9 constructs 55. The data

are conflicting with respect to the relative transforming potentials of the RET9 and RET5 1

isoforms, however, in the only study to compare the transforming abilities of RET9 and

RETS 1 directly, no differences were observed 55. The muscript encoding RET43 has only

recently been identified and the hction and interactions of this isofom have not yet been

investigated. However, we have identified expression of the RE143 transcript in human

thyroid and kidney and in neural crestderived hunours 6. 18.

In this study, we examined the expression of RET 3' alternatively spliced variants

in a panel of human fetal kidney RNAs using semiquantitative reverse-transcriptase

polymerase chain reaction (RT-PCR). We observed expression of transcripts representing

RE79 and RET43 throughout our sample panel. RE19 was the most abundant of the 3'

splicing variants in all samples examined while RET43 levels were not high enough for

accurate quantitation (42% of RET9 expression levels). Interestingly, we observed very

low levels of RET5 1 expression in the earliest gestational ages then a Ffold increase in the

Figure 4.1 RET 3' isofonn interactions with downstream effector molecules. The C-termini

of RET9 and RETS1 are shown. The amino acid sequence immediateiy upstream and

downstream of a phosphotyrosine (pY) can confer afbity for specific downstream

effectors. RET9 and W S 1 share amino acid sequence upstream of Y 1062 (LPN). The

presence of this amino acid sequence N-terminal to pY1062 suggests that the tyrosine is

capable of interacting with the PTB domain of Shc upon phosphorylation. However,

Y1062 in the sequence context of RET9 downstream sequence (YGRI) has greater affinity

for the SH2 domain of Shc. The YGRI sequence is bolded in RET9 to indicate a preference

Shc-SH2 domain interaction at Y 1062. RET51 stronger affsnity for Shc-PTB at Y 1062 and

has an additional Grb2-SH2 domain interaction site at Y1096 (YANW), one of the two Y

residues not present in RET9.

relative expression level of this transcript by 9 weeks gestation. Our data suggest that the

RET 3' variant encoding the RETS 1 isoform is not required for early kidney inductive

events. However, RETS 1 expression may be important for differentiation events

responsible for shaping the mature humao kidney.

Results

Expression of RET 3' AltemutiveIy Spliced Transcripts

We examined the expression of RET 3' alternatively spliced transcripts. which

differ with respect to the C-terminal amino acids they encode, in a panel of RNAs prepared

from human fetal kidneys ranging in gestational age from 7.5 through 24 weeks and human

adult kidney. RNA from a medullary thyroid carcinoma cell he , TT, known to express

RET at high levels, was used as a positive control T4. Given the Limited quantities of RNA

available for analysis, RT-PCR was used to investigate the expression of these transcripts.

For each analysis, primer pairs were selected to specifically amplify one of the RET 3'

alternatively spliced coding variants. A single fornard primer, CRT 148. that corresponds

to REl exon 19 sequence, was used in combination with reverse primer KRT14D.

KRT20A or KRT3B (Figure 4.2) to arnpliry transcripts corresponding to RED, RETS 1 or

RET43 C-terminal amino acids respectively 6. PCR products were electrophoresed on

ethiciium bromide-stained 2% agarose gels. We identified the (109 bp) and RET43

amplicons (177 bp) in a l l ages of human fetal kidney examined as well as in adult kidney

and the TI' cell line control (Figure 4.3). However, the ET" 1 amplicon (88 bp), which

corresponds to transcripts in which RET exon 19 is spliced to exon 20 (Figure 1.2), was

detected at much lower levels in the 7.5 week human fetal kidney, the earliest age available

to us, relative to the 12 week fetal kidney, the next gestational age analyzed (Figure 4.3).

This observation was independently confirmed by the reactions described above using

primers CRT14B (exon 19) and KRT3B (exon 21). In addition to the 177 bp product

representing the exon 191exon 21 splice which corresponds to RE143, a PCR product of

Figure 4.2 Schematic of RET 3' exons and primers used in RT-PCR analyses. The intmn

and exon organization of the RET 3' end is represented Coding sequence is indicated by

boxes. Non-coding sequence is indicated by lines. The approximate locations of primers

relative to the RET 3' exons are indicated with arrows.

Figure 4.3 Expression of RET 3' coding variants in human fetal and adult kidney. RT-PCR

analyses were used to detect expression of the tbree RET 3' coding variants in fetal kidney

ranging in gestational age from 8-24 weeks (wk). Primer pairs used in these analyses are

described in Chapter 2. The size of the predicted amplification product for each variant is

indicated. A] Expression of R E D , B] expression of RETS 1 and c} expression of RET43.

Positive (+ve) and negative (-ve) controls are indicated.

353 bp was also produced (Figure 4.3). We have shown previously that this product

results from splicing of exon 19 to exon 20 6 . This mRNA encodes a RETSL aanscript

with a 3' UTR extending through a non-coding exon 21. Although the 177 bp amphcon

corresponding to RET43 was detected in al l fetal kidney samples, the 353 bp product

corresponding to RET51 was present only in samples ranging in gestational age from 12

through 24 weeks. Our data suggested that RET51 was absent in early kidney

developmental stages and that the initiation of RETSl expression occuned after 7.5 and

before 12 'weeks gestation.

Qmtitation of RET 3' Alternatively Spliced Varirts

To investigate the ~ I a t i v e expression levels of the three RET 3' coding variants duriag

human kidney organogenesis and to characterize the early expression of REn 1, we used a

semiquantitative RT-PCR assay simiIar to that previously described in quantitation of RET

5' spliced variants 18. We compared the expression levels of RE73 1 and RET43 relative to

REIP, the most abundant of these transcripts, within a single RNA sample using multiplex

PCR A single fornard primer in exon 19 (CRT MB) was used in combination with reverse

primers in intron 19, exon 20 and exon 21 to amplify RE79, RETS1 and RET43 transcripts

respectively in a single reaction w~gure 4.2). Initial experiments were conducted to

determine the kinetics of amplification over a range of cycle numbers in order to select

conditions for which amplification was linear. End-labeled forward primers were included

in PCRs and the quantity of PCR product was represented by the amount of radioactive

incorporation. Template cDNAs were amplified for 16, 20, 24, 28, 32, 36 or 40 cycles.

PCR products were separated on 2% agamse gels, the appropriate bands excised and

incorporated counts measured by liquid scintillation. We found that amplification was Linear

for all products between 20 and 26 cycles of PCR (Figure 4.4). To ensure that

amplification proceeded with equal efficiency regardless of amount of starting template. we

performed PCR using a two-fold serial diIution of TT cDNA for 22. 24 and 26 cycles. A

Figure 4.4. Determination of conditions for semiquantitative RT-PCR analysis of RET 3'

coding variant expression. A] Kinetics of simultaneous ampiification of RED. RET5 1 and

RET43. Amounts of the three transcripts amplified in a single multipIex reaction were

quantitated by liquid scintiuation counting of radioactivity incorporated in PCR products

which had been resolved on 2% agarose gels and excised. Counts (measured in cpm) were

plotted relative to the number of amplification cycles. The exponential reaction phase was

hear between 20 and 26 cycles. Above 26 cycles, rates of ampLification for all RET 3 '

coding variant products decreased and approached plateau. B] Confirmation of

amplification efficiencies of RET primer pairs used in multiplex PCR. The results for 24

cycles are shown. Linear results were obtained when counts were plotted relative to

dilution factor for amplification of all RET 3' variant transcripts. Similar slopes for the

linear plots indicate comparable efficiencies for all primer pairs. Differences in slopes were

accounted for in final calculations using the formula (Mi/')(caIcuIated value) 131.

2 3 Two-Fold Serial Dilution Factor

linear relationship between input cDNA and amount of PCR product was identified for each

of RED, RElSl and RET43 when PCR proceeded for 24 cycles (Figure 4.4).

In our initial analyses, we chose to investigate relative expression levels of RET 3'

alternatively spliced transcripts in five human kidney RNA samples that spanned the range

of gestational ages available (75, 14, 18 and 24 weeks) as well as adult kidney.

Quantitation of RET5 1 relative to during human kidney development was performed

at 24 cycles of PCR. RET43 expression was not detected at quantifmble levels in the

human kidney samples (4% of expression levels). Thus, RET43 expression was

not considered huther in these analyses. For each of these kidney samples, the reliability of

PCR conditions for quantitation was confirmed by establishing a hear relationship

between input cDNA and PCR product. Two-fold serial dilutions of sample cDNA were

amplified for 24 cycles of PCR. The expression of RETS l relative to RED was assessed

by comparison of incorporated counts, where RET5 1 product was expressed as a fraction

of product flable 4.1). Semiquantitative assays were repeated five times for each of

7.5, 14, 18 and 24 week human fetal kidney samples as well as for human adult kidney

and mean expression with standard deviation calculated. No signiticant differences in the

relative expression level of RET5 1 to R E D were observed for the 14, 18 and 24 weeks

gestation samples and these did not vary greatly from the relative levels observed in adult

kidney (Table 4.1). In a l l cases, RET51 expression was approximately 114 to 1/3 that of

RE79 expression. Interestingly, our analyses did not detect appreciable expression of the

RE25 1 transcript in the 7.5 weeks gestation kidney sample (Table 4.1).

We next broadened these analyses to examine the expression of RETS 1 relative to

RED in fetal kidney samples spanning 7.5-14 weeks gestation using the same semi-

quantitative RT-PCR analysis. Expression of RET5 1 was found to be comparable in 7.5

and 8.5 week human fetal kidney samples at 5&3% and 6&5% that of RE29 respectively

(Table 4.1, Figure 4.5). However, by 9 weeks gestation, the expression level of REn 1

had increased seven-fold to 35&8% that of RE29 (Table 4.1, F i g w 4.5). Our &ta suggest

Table 4.1. RE151 expression relative to RE29 expression in human fetal kidney.

Quatitation of the RET 3' alternatively spliced coding variants during human kidney

development based on semiquantitative multiplex RT-PCR analyses. RETS 1 expression

levels are given as a fiaction of RElP expression levels. RE79 expmssion levels have been

normalized to 1.00 in all samples. Means and standard deviations of five repeats are given.

Sample RElP Meanksd

7.5 week 1.00 0.05+0.03 8.5 week L .OO 0.06&0.05 9 week 1.00 0.35&0.03

10.5 week 1.00 0.2m.04 12 week 1.00 0.3w.08

14 week 1.00 0.36+0.06 18 week 1 .OO 0.32+0.05

24 week 1.00 0.3 1+0.04

adult 1 .OO 0.2m.04

Figure 4.5 Developmental expression of RET 3' coding variants in human fetal kidney.

Expression levels of RETSl transcripts relative to RET9 transcripts amplified in the same

mulitplex RT-PCR reaction are given for each sample. Quantitation was performed as

described in text. Mean values and standard deviations shown in Table 4.1 are plotted.

Age of Fetal Kidney (weeks)

developmental regulation of expression of the RE15 1 M p t over the 7.5-24 week

period in the human fetal kidney. Further, we have shown that RE751 is expressed at very

low levels early in gestation but represents approximately 114 of RET transcripts by 10.5

weeks gestation in the human kidney.

Discussion

The human kidney is fairIy well-defined by 8 weeks gestation and begins to

function by 11 weeks gestation (reviewed in 37). We have shown that RET expression is

higher at earlier stages of kidney development (7.5 weeks gestation) relative to later stages

(14-24 weeks gestation). RET expression levels were shown to decrease with increasing

fetal age through to 24 weeks gestation 18. In this study, we have shown that the three RET

3' coding variants, R E D , RETSl and RET43, are expressed during human kidney

development in samples ranging in gestational age from 7.5 through 24 weeks as well as in

adult kidney. However, we observed that RETS1 expression is very low in the earliest

gestational ages (7.5-8.5 weeks) relative to R E D expression in the same samples. While

we did not detect significant RETS1 expression in the 7.5 and 8.5 week fetal kidney

samples, a considerable increase in RE75 1 expression was observed by 9 weeks gestation,

the next available age of human fetal kidney, in all later gestational ages examined (9

through 24 weeks) and in adult kidney (Tables 4.1 and 4. 2). The absence of R E T S 1

expression until approximately 9 weeks gestation in human kidney raises the possibility

that RETSl has a role in kidney development distinct from those of RET9 and RET43. By

8 weeks gestation, the permanent kidneys have taken on a recognizable form and 3-5

branchings of the ureteric bud have occurred (reviewed in 37). The renal pelvis and the

major and minor calyces, ail of which are products of the ureteric bud, begin to take shape

by 10 weeks gestation around the time we see initiation of RETSl expression. Our data

indicate that the absence of RETS 1 expression during the early inductive events in the

human kidney is followed by a notable increase in RET51 expression by the time early

differentia~g events begin.

There are considerable data in the Literature suggesting that the functions of the

RET9 and RETS1 isoforms are not totally redundant. Atti6 et al. lo* have identified a

missense mutation in exon 20 which is, thus, present in the RET51 isoform but not in the

other RET 3' coding variants in a patient with HSCR. As described in chapter 1, HSCR is

a congenital abnormality characterized by the absence of sympathetic neurons in the

hindgut Inac t i v a ~ g RET mutations are identified as the underlying cause of 1040% of al l

HSCR cases 97. Thus, RET51 appears to be required for the induction of the myenteric

nerve plexus but not for kidney induction. Consistent with this phenotype, Rossel et al.

identified more prominent neurite outgrowth in PC12 cells transfected with activated

RETSI compared to activated RET9 although both isoforms induced some degree of

neurite differentiation 55. Taken together, these data might suggest that RETS1 may

activate downstream differentiation pathways in addition to or perhaps in preference to

mitogenic pathways.

The RETS1 isoform contains two additional tyrosine residues not present in the

RET9 or RET43 isoform (Y1090 and Y1096) 9. Y1096 has been identified as an

autophosphorylation site in RET 136. Direct binding between the Grb2-SH2 interaction

domain and the at pY 10% of RET5 1 has now been demonstrated 56. W three isoforms

include the tyrosine residue at 1062 which can be autophosphorylated on RET activation.

Splicing and sequence divergence at amino acid residue 1063 places this tyrosine in three

different binding contexts in the 3' RET isoforms 9. AU three sites share a consensus

sequence for interaction with the Shc-PTB domain (Figure 4.1) 13'. PreIiminary studies

have identified differences in the abilities of RET9 and RETS1 to interact with the Shc-PTB

domain at Y1062 57 (M. Billaud, personal commUtLication). In comparison to RET9, the

stronger binding of RETSI to the Shc-PTB domain is probably due to a higher affinity of

Y1062 for Shc-PTB in the sequence context of RETSL. Y1062 has been found to be a

major Shc-SII2 domain interaction site in RET9 while RETSl has distinct Shc-SH2 and

PTB interaction sites 56.

The differences between RET9 and Rl3"SI affinities for Grb2-SH2 and Shc-SH2

and -PTB interaction domains suggest that these two RET isoforms may differ with respect

to their downstream signaling pathways. Thus, the expression of these two isoforms could

modulate the relative activation of these pathways. We have shown that the RET 5 '

alternatively spliced transcript, RERI6, has highest expression early in human kidney

development (8 weeks gestation) and decreases through to 24 weeks gestation 18. We

suggested that RET may have varying roles or significance at different stages of kidney

development affected by differences in expression of the various RET 5' altemativeiy

spliced forms. The data described in this chapter may also suggest that the RET 3'

alternatively spliced variants affect kidney morphogenic processes. The dramatic increase in

the relative expression of RETS 1 compared to RE19 and RET43 between 8.5 and 9 weeks

gestation in the human kidney implicates a role for RET51 in human kidney development

that is distinct from either that of RET9 or RET43 (Table 4.1, Figure 4.5).

Materials and Methods

Fetal Kidney. Adult Kidney and IT RNA Analyses

The RNA extraction methods used are described in Chapter 3. Primer sequences

used in these analyses are given in Chapter 2. One l g of total RNA was heated at 7 0 0 ~ for

5 minutes and briefly cooled on ice. First strand cDNA was synthesized by incubating

RNA templates and PCR amplification performed under conditions descriid in Chapter 3.

Initial amplification analyses to identify expression of the individual RET 3' altemativeiy

spliced variants were performed using 1 min at 950C I 1 min at 550Cll min at 72OC 1 min

for 40 cycles followed by a final extension of 72% for 10 minutes.

Semi-Quantitative RT-PCR

RT-PCR experiments were performed in LOW volumes using the RT-PCR

conditions described but including 0. lpM 32Pend-labeled sense strand primer. To amplify

the three RET 3' alternatively spliced coding variants in a single multiplex PCR reaction,

1p.M each of NT forward primer (CRT14B) and reverse primers (KRT14D, KRT2OA and

KRT3B) was used Analyses to determine the linear range of amplification were performed

using 1/10 of a cDNA reaction, described above, in PCR for 16, 20, 24, 28, 32, 36 or 40

cycles. PCR products were separated and visualized on ethidium bromide-stained 2%

agarose gels and appropriate bands were excised. Quantitation of PCR products was

performed by Liquid scintillation of incorporated radioactivity. Plots of incorporated counts

(cpm) versus cycle number were used to define the range of PCR cycle numbers for which

amplification was linear (Figure 4.4). Two-fold serial dilutions of cDNA were subjected to

cycles of PCR within the Linear range to co&m the number of cycles for which primers

would amplify with similar efficiencies regardless of amount of starting template (Figure

4.4). PCR products were quantitated as described above. If the relationship of cpm versus

dilution factor was hear, the conditions were considered reliable for quantitation purposes.

Based on the preliminary experiments, 24 cycles of PCR was selected for semiquantitative

RT-PCR analysis of RET 3' alternatively spliced transcript expression in human fetal and

adult kidney RNA samples.

Chapter 5

Expression of genes encoding the RET ligand complex components GDNF and GDNFR-a daring human kidney development

I performed all the analyses of GDNF and GDWR-a expression in human kidney.

Introduction

Mammalian kidney development quires a series of reciprocal inductive

interactions between two distinct cell types, the ureteric bud epithelium and rnetanephric

mesenchyme 35. Induction is followed by the development of tubular epithelium from the

mesenchyme and the proliferation and branching of the ureteric bud. Roles for glial cell

linederived neurotrophic factor (GDNF) and RET. its receptor. in them events are

suggested by the phenotypes of the knockout mice, both of which display renal

dysmorphology characterized by retarded growth and reduced branching of the ureteric bud

as well as large areas of undifferentiated mesenchyme (described in Chapter 1) 34* 46-48.

GDNF receptor-a (GDMR-a), a cell surface protein, is required to mediate high affinity

binding of GDNF to RET 21.22- 33.

GDNF is a distant relative of the transforming growth factor (TGF)-P superfamily

based on the presence of seven conserved cysteine residues spaced identically in nerve

growth factor, plateletderived growth factor and TGF-f3 23. Human GDNF consists of

two exons and encodes a 211 amino acid pmtein precursor- Mature GDNF forms

glycosylated, disulfide-linked hornodimers to impart biological activity 24. The use of an

alternative splice site at nucleotide (nt) 69 of the human GDNF cDNA sequence results in

the deletion of 78 at corresponding to nt 70-148 (amino acids 24-50} and leads to a shorter

GDNF mRNA 239 26-27. The protein sequence deleted as a result of alternative splicing is

in the preprotein sequence and is not predicted to Sect the amino acid sequence of the

mature GDNF protein. GDNF is highly expressed in the gut prior to innervation by the

neural crestderived enteric neuroblasts and in the undifferentiated metanephric

mesenchymal cells in the nephrogenic zone of the developing kidney 20v 29. The phenotype

of GDNF -/- mice, which includes complete rend agenesis, characterized by lack of

ureteric bud or failure of the bud to grow and invade the rnetanephric mesenchyme,

suggests that GDNF plays a role in kidney induction 4&48. There are striking similarities

between the phenotypes of the GDW-/- and REW- mice. Not only do these mice have

similar degrees of renal dysplasia, they also both lack entenc neurons 34* 464%

The expression of RFT at the cell surface is not sufficient for high m t y binding

of GDNF 19-22. Binding assays have indicated a requirement for GDNFR-a in GDNF

binding to RET 21. 22. GDNFR-a, a novel GPI-linked protein, was originally isolated

from expression libraries prepared from mRNA isolated from retinal cell cultures enriched

for photoreceptors and rat embryonic midbrain enriched for GDNF-responsive

dopaminergic neurons 21- 22. Northern analyses have shown high levels of GDNFR-a

expression in embryonic kidney and intestine, sites of RET and GDNF expression 21- 227

33. Low levels of GDNFR-a expression have been detected in fetal brain, lung and other

components of the central and peripheral nervous systems.

We have shown that RET expression levels are highest during early human kidney

development then decrease with increasing embryonic age (described in Chapter 3) 18. In

this study, we examined expression of the genes encoding the RET ligand complex

components, GDNF and GDNFR-a, in developing human kidney using RT-PCR analyses

similar to those described in Chapters 3 and 4. In each of our fetal kidney samples. we

detected GDNF as well as GDNFR-a expression. Using a semiquantitative RT-PCR

analysis, we observed developmentally regulated expression of GDNFR-a in fetal kidney.

Our data is consistent with a role for REWGDNFI GDNFR-a signaling in human kidney

morphogenesis.

Results

Expression of GDNF in Human Fetal K idnq

We examined the expression of GDM; in a p a d of RNAs prepared from human

fetai kidneys ranging in gestational age b m 75 through 24 weeks as well as in adult

kidney. RNA from a Wi' tumour found to express GDNF at high levels (discussed in

Chapter 6) was used as a positive control. As described in Chapter 3, RT-PCR analyses

with a primer pair selected to amplify across the exodmtmn boundary were used to

investigate GDNF expression. A primer pair was selected that would amplify both full

length GDNF and the alternative splicing v&ant in a single reaction. PCR products were

electrophoresed on ethidium bromide-stained 2% agarose gels. We identified two predicted

products of 448 bp and 370 bp, corresponding to full length and alternatively spliced

GDNF, respectively, in all ages of human fetal kidney examined and in adult kidney

(Figure 5.1). Based on the intensities of the GDNF PCR products relative to one another,

expression appeared to be stronger between 8.5 and 12 weeks gestation compared to other

ages. RT-PCR was performed using GUSB-specific primers described in Chapter 2 to

determine if the differences we observed were due to variability in RNA integrity. The

intensities of the GUSB products were similar in all samples indicating minimal variability

in integrity across samples (Figure 5.1).

Erpression of GDNFR-a During Human Kidney Development

We used RT-PCR to investigate the expression of GDNFR-a in human kidney

development. PCR was performed using RAN0 and R A W primers, designed to amplify

products that crossed an exodintron boundary, to amplify a GDNFR-a product of 208 bp.

RT-PCR was performed on our panel of human fetal kidney RNAs ranging in age from 7.5

through 24 weeks as well as adult kidney. RNA extracted from a medullary thyroid

carcinoma cell line, TT, treated with retinoic acid (RA) for 24 hours was used as a positive

control (TT+RA). GRNFR-a expression was not detected in RNA samples fkom untreated

Figure 5.1. Expression of GDNF and GDNFR-a in human fetal and adult kidney. RT-

PCR analyses were used to detect GDNF and GDNFR-a expression in developing human

kidney. Similar analyses were used to detect GUSB expression in these samples as an

indicator of sample integrity. Rimer pairs used in these analyses are described in Chapter

2. The sizes of the predicted amplification products are indicated Gestational age of the

kidney sample is given in weeks (wk). A] GDNF expression, B] GDNFR-a expression

and C] GUSB expression. Positive (we) and negative control reactions are indicated (-ve).

'IT cells (data not shown). PCR products were electmphoresed and visualized on ethidium

bromide-stained 2% agarose gels. We identifled amplicons of the predicted size

comsponding to GDNFR-a throughout our panel of human fetal kidneys as well as in

adult kidney (Figure 5.1).

Quantitation of GDNFR-a in Hwnan Fetal Kidney

To investigate the reiative expression levels of GDATFR-a duriag the various stages

of human -kidney development, we used a semiquantitative RT-PCR assay to compare

expression levels of GDNFR-a to those of a housekeeping gene, P-glucuronidase

(GUSB). In order to avoid variations that might arise due to inter-sample differences in

RNA quality, we determined GDNFR-a expression relative to GUSB expression within a

sample. In addition to a 208 bp GDNFR-a product, a 195 bp GUSB product was

amplified in the reactions. A 5: 1 molar ratio of GDNFR-a primers to GUSB priwrs was

used to obtain similar amplitication of both products. To determine PCR conditions that

would be reliable for quantitation of products, preliminary experiments similar to those

described in Chapters 3 and 4 were conducted Template cDNAs were amplified over

increasing cycles of PCR, products were electrophoresed on 2% agarose gels and

appropriate bands excised. Incorporated counts (cpm), representative of the amount of

amplified product, were measured using liquid scintillation. After determiaing that

amplification proceeded Linearly between 20 and 26 cycles of PCR (Figure 5.2).

amplification efficiencies were confirmed using a two-fold serial dilution of TT+RA cDNA

for 22, 24 or 26 cycles as described previously in Chapters 3 and 4. At 24 cycles of

amplification. a linear relationship between input cDNA and amount of PCR product was

identified for each of GUSB and GDNFR-a (Figure 5.2).

We investigated GDNFR-a expression in RNA exuacted from human fetal kidneys

aged 7.5, 12.5, 14, 18 and 24 weeks, spanning the range of gestational ages available, as

well as in adult kidney. The reliability of conditions established for our control cell h e was

Figure 5.2. Determination of conditions for semiquantitative RT-PCR analysis of

GDNFR-a expression. A] Kinetics of simultaneous a~~~~lif?cation of GDNFR-a and

GUSB. Amounts of G M R a and GUSB product amplified in a singe muitiplex reaction

were quantitated by liquid scintillation counting of incorporated radioactivity. Counts

(measured in cpm) are plotted relative to the number of amplification cycles. The

exponential reaction was linear between 20 and 26 cycles. At 26 cycles, rates of

amplification for both GDNFR-a and GUSB decreased and approached piateau. B]

Confirmation of amplification efficiencies of G D N F R a and GUSB primers used in

multiplex PCR for semiquantitative analyses of GDNFR-a expression. A two-fold serial

dilution of rrtinoic acid-treated TI' (TT+RA) cDNA was subjected to 22,24 or 26 cycles of

PCR. The results for 24 cycles are shown. Linear results were obtained when counts were

plotted relative to dilution factor for amplification of both GDNFRa and GVSB transcripts

indicating comparable amplification efficiencies for these primer pairs.

GUSB GDNFR-a

Number of PCR Cycles

GDNFR-a GUSB

0 1 2 3 4 5 Two-Fold Serial Dilution Factor

confkmed for each human kidney sample by subjecting two-fold serial dilutions of cDNA

to 24 cycles of PCR A linear relationship between input cDNA and amount of PCR

product, determined by liquid scintillation of incorporated counts, coafirmed equivalent

amplification efficiencies for GDNFR-a and GUSB in each kidney sample.

Comparison of incorporated counts was used to assess GDNFR-a expression

levels relative to those of GUSB. The amount of GDNFR-a product was expressed as a

fraction of the amount of GUSB product. M e w of five repeats of the assay are given in

Table 5.1 with calculated standard deviations. We found that GDNFR-a expression levels

were relatively high in human fetal kidney aged 7.5 through 14 weeks compared to later

gestational ages. Expression leveis were approximately 2-fold lower in the 24 week feral

kidney sample compared to earlier ages (Table 5.1. Figure 5.3). Expression levels in adult

kidney were approximately 3-fold lower than those observed for early ages of fetal kidney

(Table 5.1, Figure 5.3). Our data suggest that GRNFR-a expression decreases during later

stages of human kidney development.

Discussion

In this study, RT-PCR analyses were used to investigate GDNF and GDNFR-a

expression in human kidney development. We observed expression of both mRNAs in the

earliest gestational ages available to us (7.5 weeks) through to 24 weeks gestation as well

as in adult kidney. GDNF expression appeared to be reduced in the adult kidney relative to

expression in the early gestational ages. Our data are consistent with reports of GDNF and

GDNFR-a expression in rodents 2 2 27-29. GDNF expression has been found to be lower

in the adult relative to embryonic and neonatal rodent kidneys 2*. In rodents, GDNFR-a

expression in the developing kidney has not yet been directly compared with expression in

the adult kidney. However, high levels of GDNFR-a in the adult kidney relative to

expression levels in other organs have been described 21.

Table 5.1. Expression of GDNFRa in developing human kidney. GUSB expression

levels have ken normalized to 1.00 for a l l samples. GDM;R-a expression levels are given

as fractions of GUSB expression levels amplified in the same reaction. Means of five

repeats and calculated standard deviations are given.

Sampie Meanksd

7.5 weeks 1.1 1a.02

12.5 weeks 0.9 lH.03

14 weeks 0.9 1H.20

18 weeks 0.82H.23

24 weeks 0.49M -04

adult 0.3&0.06

Figure 5.3 Developmental expression of GDNFR-a in human fetal and adult kidney. Semi-

quantitative RT-PCR was used to investigate GDNFR-a expression levels. Ratios of

GDNFR-a expression to GUSB expression within a sample are given for fetal kidney

samples ranging in gestational age fiom 7.5 through 24 weeks. Points plotted on the graph

are the means of five individual assays. Error bars for each sample represent calculated

standard deviations.

6 8 1-0 12 14 16 18 20 22 24 26 Gestational Age of Kidney (weeks)

In Chapter 3, we reported developmental regulation of RET expression in human

fetal kidney '8. Here, we show that GDNFR-a also undergoes developmental regulation

over the same period. Our data show that GDNFR-a expression levels are relatively higher

in the early gestational ages (7.5- 14 weeks) then decrease through to 24 weeks gestation in

human kidney. The Lowest levels of GDNFR-a expression were detected in adult kidney.

GDMR-a has been shown to mediate the association of GDNF with RET 21. 22. 33.

Tremor et al. 22 suggest that a disulfide-linked GDNF dirner bound to one or two

molecules of GDNFR-a forms a complex which subsequently binds to and activates RET

(Figure 1.4). In the absence of GDNFR-a expression, GDNF is unable to induce

autophosphorylation of RET 21- 22. In vino assays have shown that overexpression of

GDM;R-a greatly increases GDNF binding to RET and, hence, RET autophosphorylation

21- 22. Given that GDNFR-a expression levels modulate GDNF binding to RET, the

higher GDNFR-a expression in early ages of human fetal kidney relative to expression in

adult kidney which we observed is consistent with GDNFR-a expression as a means of

regulating RET activity during kidney development.

Another means of regulating RET activity would be through receptor binding to

alternative ligand components. The multiple GDNF and GDNFR-a family members

identified to date are potential modulators of RET activity. Recently, Sanicola et aL 33

reported the isolation of RETL2 from a rat embryonic kidney cDNA Library. The RETL.2

protein shares 4 9 1 homology to rat GDNFR-a but, unlike GDNFR-a, is capable of

binding GDNF only in the presence of RET. GDNFR-a and R E Z E are expressed in a

wide variety of embryonic and adult tissues including kidney. This raises the possibility

m y or all of these molecules affect RET activity. Sanicola et al. 33 have proposed that the

dissociation of the REWGDNF/RET complex would release GDNF from the cell surface

while dissociation of the GDNFR-or/GDNF/RET complex would retain GDNF bound to

GDNFR-a. In this scenario, RETL2-expressing cells might display greater sensitivity to

changing local concentrations of GDNF than those expressing only RETLZ. Neurmrin has

been identified as a relative of GDNF 31. The protein shares 42% homology with the

mature GDNF protein. Similar to GDNF. neurturin is expressed in a wide range of

neonatal and adult tissues including brain and kidney. The ability to promote survival in

neuronal populations, which neurturin shares with GDNF, suggests that the two proteins

are capable of acting through common signaling pathways 31. However, whether neurturin

is able to interact with RET has not yet been investigated.

High levels of GDNF expression have been observed in embryonic rat kidney 2'-

29. GDNF expression in the developing kidney decreases after embryonic day 16 and

significantly lower levels of GDNF expression are detected in the postnatal kidney 2'- 28.

No expression was detected in the adult kidney 27v 28. I . humans. we have shown that

GDNF expression occurs throughout the various stages of kidney development. Lower

levels of expression in the adult kidney are suggested by the intensity of the PCR product.

Thus, a similar developmental tread in GDNF expression in the developing kidneys of

rodents and humans is suggested, however, quantitative analysis of GDNF expression in

the developing human kidney is required.

The expression of genes encoding the fust two identified RET ligand complex

components. GDNF and GDNFR-a, in human kidney development is consistent with a

role for the RET signaling complex during kidney morphogenesis. We have identified a

downward trend in GDNFRa expression with increasing fetal age in the human kidney.

The developmental pattern of GDNFR-a expression in the kidney might act as a

mechanism for regulatory control of RET activation during development in this organ.

Materials and Methods

Reverse-Trc~nscnption- Polymerase Chain Reachon Analyses

RNA was extracted using the methods described in Chapter 3. Primers sequences

used in RT-PCR analyses are given in Chapter 2. One pg of total RNA was heated to 70oC

for 5 minutes and briefly cooled on ice. First strand cDNA was synthesized by incubating

RNA templates under conditions descn'bed in Chapter 3. PCR amplification of cDNA

templates was performed in lOmM Tris-HCl (pH 8.3), 50mM KCI, 0.01% gelatin, 2 0 w

d N T P s , lpM each primer, 1.5 units Taq DNA polymerase (Gibco-BRL Life Technolo@es)

and 0.75m.M MgC12 for ampiification of GDNF products or 1.5m.M MgC12 for

amplification of GDNFR-a products. The PCR cycling conditions used are described in

Chapter 3.

Semi-Quantitative RT-PCR

RT-PCR experiments were performed in 10p.L volumes using the RT-PCR

conditions described but including O.1p.M 32~end-labeled sense strand primer. Optimal

conditions for co-amplification of GDNFR-a and GUSB in a single multiplex PCR

reaction consisted of LpM of each GDNFR-a primer and 0.2pM of each GUSB primer.

The hear range of amplification was determined as described in Chapters 3 and 4. Two-

fold serial dilutions of cDNA were subjected to cycles within the linear range to confirm the

number of cycles for which primers would amplify with similar efficiencies regardless of

amount of starting template. PCR products were quantitated as described above. If the

relationship of cpm versus dilution factor was hear, the conditions were considered

reliable for quaatitation purposes. Based on the preliminary experiments, 24 cycles of PCR

was selected for semiquantitative RT-PCR analysis of G D N F R a expression in human

fetal and adult kidney RNA samples.

Chapter 6

Analyses of RPT, GDNF and GDNFR-a Expression in Human Disease

I performed all analyses of RET, GDNF and GDNFR-(x expression in Wilms' tumours and rend cell carcinomas. The d t s of G D W mutation screening in Hirschsprung disease appear in Human Molecular Genetics 5:2023-2026 (1996) and are the result of a collaborative effort with Ms. Shirley Myers. Sequencing analyses of GDNF were performed equally by Ms. Myers and myself.

Introduction

The E T proto-oncogene encodes a receptor tyrosine kinase required for enteric

neurogewsis and kidney development 9- 349 40. Expression of RET is restricted both

spatially and temporally in the developing nervous and excretory systems, suggesting a role

for this receptor in the regulation of proliferation and/or differentiation during development

11* 14-169 18. RET's involvement in tumourigenesis and in developmental anomalies has

been well documented. RET is rearranged and constitutively activated in a proportion of

papillary thyroid carcinomas 78-82. Germline point mutations of RET are associated with

the majority of multiple endocrine neoplasia types 2A and 2B and familial meduliary

thyroid carcinoma, all of which are dominantly inherited cancer syndromes characterized by

multiple tumours of neuroectodermal origin (reviewed in 88). These mutations convert

RET into an activated oncogene 87. Mutations in RET have also been described in 1040%

of cases of HSCR, a congenital abnormality characterized by aganglionosis of the lower

gut 989 102- 10% los- 106. It has been shown that RET mutations associated with HSCR are

inactivating and result in abrogation of biological activity of the receptor 9 9 ~ 1 3 ~ .

Recently, a ligand for RET has been identified as a multicomponent complex

consisting of GDNF and GDNFR-a 21- 22. 33. While GDM is a soluble molecule,

GDNFR-a is a GPI-Linked cell surface protein. GDNF is not sufficient for high affiaity

binding to RET aud has a requirement for GDNFR-a. The pattem of GDNF expression

during development of the nervous and excretory systems is similar to that of RET 14* 159

27.29. Further, the phenotypes of the GDNF -/-and RET-/- mice are similar including renal

dysplasia and lack of enteric neurons 349 46-48. These observations suggest that

disturbances in any pan of the RET activation complex may have similar consequences.

Phenotype associations have previously been reported for mutations in c-KTr and Steel

factor (SLF), genes encoding a complementary receptodgrowth fanor system (reviewed in

139).

Reciprmal inductive interactions between ureteric bud epithelia and metanepbric

mesenchyme are required to signal cell proliferation aud differentiation in the developing

kidney 3? It has bem suggested that defccts in mechanisms which halt cell proLiferatiou

and/or induce mesodefmal differentiation can result in Wilms' turnour m, the most

common paediatric solid -our of the kidney and one of the most common solid tumours

of childhood (reviewed in 140-143). The majority of WTs are sporadic although 1% of

cases are familial (reviewed in 143). Histological analysis of WTs have indicated a

resemblance to inappropriate development Turnours may be triphasic containing compact

areas of blastema, more differentiated epithelial components and/or saomal elements. One

of these elements often predominates '44. WTs are believed to be derived from cells of the

MM that normally differentiate into the epithelial components of the nephron (reviewed in

141-143).

The molecular genetics of WT ae complex. however, three genes have been

implicated in tumourigenesis. 3040% of sporadic twnours are associated with loss of

heterorygosity (LOH) on chromosome 1 1 p 145. Mutations inactivating WT- 1, which has

been mapped to 1 lp 13, have been identified in a subset of cases with genetic susceptibility

to WT and in 1045% of sporadic tumows t46-148. One-third of WTs display LOH

restricted to 1 lpl5.5 14% 150. This same chromosomal region has been implicated in

Beckwith-Wiedemann syndrome (B WS), a congenital disorder characterized by generalized

overgrowth and an increased risk of childhood tumours including Wilms' tumour l49-151.

A BWS gene wodd be expected to impact on WT development Recently, mutations in

p57kip2, an imprinted gene that maps to chromosome 1 lp 15.5. have been identified in a

subset of BWS patients L52. Expression of pS7kip2 is reduced in WTs. however, p 5 7 W

has not been investigated for mutations in WT 152. A familial \KT predisposition gene,

FWTI, has been lccaiized to chromosome 17q12-21 153. Several candidate genes are

located in this interval including neurofibomatosis-l (IVF-I) and insulin-like growth factor

binding protein (IGFBP)-4, however, the gene associated with this phenotype has yet to be

defined 154.

Rend cell carcinoma (RCC) is the most commoa malignancy of adult kidney and

accounts for 2 8 of al l cancers (reviewed in 155). Most cases of RCC are sporadic,

however, rare familial forms are characterized by autosod dominant inheritance, young

age at diagnosis and bilateral and multifocal turnours. Von Hippel-Lindau (m) disease is

an inherited cancer syndrome charactexkd by high incidence of RCC (reviewed in 156).

The VHL gene lies at 3~25-26 157. Chromosome 3p a b n o d t i e s including

rearrangements and LOH are the most common occurrence and are amongst the earliest

abnormalities identitied in both sporadic RCC and MIL-associated renal turnours t58-160.

Frequent point mutations or transcriptional inactivation through hypermethylation of the

remaining VHL dele are found in RCC resulting in tumourigenesis l61- 162. Other

chromosomal abnormalities associated with RCC include LOH at chromosomes 6q, tip, 9,

LOq and 14q 158. 161-1% These are likely to represent secondary events in -our

progression 1%

Ultrasmcturd and cell surface protein expression studies suggest that RCC derives

from proximal tubule epithelium of the renal cortex (reviewed in 155). During

organogenesis, these cells, as well as giomerular and distal convoluted tubule epithelial

cells, originate from the MM, a cell mass that proliferates and differentiates in response to

signals h m the ureteric bud 35. Continued proliferation of the MM and its early derivatives

is dependent upon induction. This may suggest that disruption of genes with roles in the

promotion of proliferation or in induction could contribute to RCC.

Rcviously, we investigated the expression of RET during human kidney

morphogenesis l8 (described in Chapter 3). In this study, we examined RET expression as

well as expression of GDNF and GDNFR-a in a series of renal tumoun. RET expression

was detected in each WT but GDNF and GDNFR-a were not. In contrast, RET and

GDNFRa expression were detected in each RCC sample and GDNI; expression was

always absent.

We also investigated HSCR patients for mutations in GDNF. Our analysis revealed

a de novo mutation in one HSCR patient suggesting that GDNF mutations may conmbute

to the HSCR phenotype.

Results

Expression of RET, GDNF and GDNFR-a in Wilms' Tumour

We investigated RET' G D and GDNTRa expression in a panel of six WTs and

one WT cell line. DW2. Three tumour samples were obtained from primary tunours

(WT20, WT21, W 6 3 ) (Tabie 6.1). The remaining three tumour samples were obtained

from mouse xenografts of primary WTs (WT26, WT27A. WT19) flable 6.1) lu. AU

tumours in our panel were sporadic and had favourable histology according to Beckwith's

criteria 14'? Histological analyses identified the tumours in the panel as predominantly

blasternal. predominantly epithelial differentiated, predominantly smrnal differentiated or

epithelial mixed (Table 6.1). RT-PCR analyses were performed on RNA extracted from the

tumour samples and from DW2 cells using primer pain designed to cross an exodintron

boundary. Using RET-specific primers CRT4S and CRT4B, an amplicon of the predicted

size (187 bp) was detected in all six WTs and in the DW2 cell line (Figure 6.1). This result

was surprising given that the WT cell of origin is not predicted to express RET.

Investigation of GDNF and GDNFR-a expression in this WT panel was conducted

using RT-PCR analyses similar to those described in Chapter 5. The GDNF-specific

primers selected to amplify products that spanned an identified alternative splicing site were

GDXlF and GDN7R 26. 27. GDNF expression was not detected in one of six WTs

(WT27A) nor in the DW2 cell line (Figure 6.1). We detected two GDNF products of the

Table 6.1. Histology and RET, GDNF and GDNFR-a gene expression in Wilms' turnour.

AU tumours were classified as Wilms' tumours based on Beckwith criteria. Rimary

turnours and nude mouse xenografts are indicated Gene expression is indicated by + while

absence of gene expression is indicated by -. NA is used to denote lack of information.

WT Tumour Histol~gicai RET GDNF GDNFR-a Sam~le Source Classification Expression Expression Expression

WT20 primary stromal + + + differentiated

WT2 1 primary epithelial + + + differentiated

WT26 xenograft mixed + + + WT27A xenograft blasted + - + WT29 xenograft mixed + + + WT763 primary NA + + + DW2 cell line NA + - -

Figure 6.1 RET, GDNF and GDNFR-a expression in Wilms' turnour 0. PCR

amplification of cDNA templates was performed using primer pairs described in Chapter 2.

The predicted product size for each amplification is indicated A] RET expression.

B] GDNF expression and C] GDNFR-a expression. Positive (+ve) and negative (-ve)

controls are indicated.

RET

predicted sizes, 238 bp and 160 bp, in al l other samples including a 14 week human fetal

kidney RNA sample used as a positive control (Figure 6.1). There was no difference in the

level of expression of the two GDNF-specific PCR products nor in the ratio of expression

of the two products amongst samples.

RT-PCR analyses were also used to examine GDNFR-u expression in our panel of

tumours. Using GDNFR-a-specific primes, a product of the predicted size (208 bp) was

amplified in all six WTs and detected upon electrophoresis and visualization on agarose

gels (Figure 6.1). The intensities of detected PCR produfts appeared to be equivalent

across these samples. In comparison, a barely detectable GDNFR-a-specific PCR product

was amplified in the DW2 cell Line which was previously shown to express RET but not

GDNF (Table 6.1, Figure 6. L).

Expression of RET. GDNF and GDNFR-a in Renal Cell Carcinoma

In order to address the possibility of RET involvement in adult onset renal turnours,

we investigated the expression of genes encoding the RET/GDNF/GDNFR-a complex

components in 11 cases of renal cell carcinoma (RCC) and in the corresponding normal

kidney tissues. Tumour and matched normal kidney cortex tissue samples were obtained

following nephrectomies performed at the Kingston General Hospital. Tumow samples

were dissected from the surgically removed kidneys taking care to minimize contaminating

nonnal tissue. We used the same RT-PCR strategies described above for examination of

RET and GDNFR-a expression in WTs to investigate expression of these transcripts in

RCC and matched normal kidney. GDXlF and GDX3R primers were used for GDNF

expression analyses in RCC. For al l analyses, human fetal kidney aged 12 weeks was used

as a positive control. Using RlT-specific primea CRT4S and CRT4B in PCR, the

predicted 187 bp product was detected in each normal kidney sample and in each of the I 1

RCC tumour samples (Table 6.2). The results from a subset of these tunours and matched

normal kidney samples are shown in Figure 6.2. In each case, the intensity of the PCR

Table 6.2. RET, GDNF and GDNFR-a expression in renal ceU carcinoma. RT-PCR analyses

were used to investigate gene expression in a series of 11 RCCs (TK) and matched normal kidney

samples (NK). Expression of GLISB, a low-level house-keeping gene. was used to confirm

sample integrity. +, H, ttc and ++++ are arbitmy units used to denote expression levels where

+ represents low levels and cm represents high levels of expression baxd on the intensity of the

PCR product. I indicates absence of expression.

RET GDNF GDNFR-a GUSB

Sample NK TK NK TK NK TK NK TK

Figure 6.2 RET, GDNF and GDNFR-a expression in renal cell carcinoma (RCC). PCR

ampiification of cDNA templates prepared from RCC C I X ) and matched normal kidney

cortex (NK) samples was performed using primer pain described in Chapter 2. The

predicted product size for each amplification is indicated A] RET expression. B] GDNF

expression, C] GDNFR-a expression and D] GUSB expression. Positive (+ve) and

negative (-ve) controls are indicated.

product was greater in the normal kidney compared to the tumour sample which may

suggest differences in RET expression between the two. The ecpivalent intensities of

GUSB PCR products generated from these same samples suggested that the differences in

intensity of the RETPCR products was not due to differences in RNA quality (Figure 6.2).

GDNF and GDNFR-a expression was examined in our RCC samples. PCR

analyses failed to detect GDNF expression throughout our panel of RCCs while expression

was observed for each matched normal kidney cortex sample (Table 6.2, Figure 6.2). We

detected relatively high GDNFRa expression in 1111 1 RCC and comparable expression

levels in the matched normal kidney samples (Table 2 Figure 6.2).

GLlNF Mutation Analysis in Hirschspmg Patients

We used direct genomic sequencing to analyze DNA from 16 sporadic and 20

familial cases of HSCR. PCR was used to amplify GDNF exons 1 and 2 in separate

reactions and the products purified. Complete sequencing of GDNF exon 1 did not identify

any sequence variants amongst our samples. However, two sequence variants were

identified within exon 2 of GDNF in two different HSCR patient samples. The first

sequence variant was identified in a patient with totai colonic aganplionosis associated with

a cytogeneticdy detectable deletion of 10q 11.2-21 -2 which spanned the RET locus 101.

Both parents were healthy and had normal karyotypes suggesting that the disease

phenotype was associated with the deletion. The GDNF sequence variation involved a G-

>A substitution at base pair 429 (exon 2) (data not shown). The change was predicted to

destroy an RraI restriction site in the GDNF sequence. This was confirmed by restriction

enzyme digestion of the amplified patient DNA (Figure 6.3). The substitution was not

detected in 35 other HSCR patients nor in 301 normal controls 16* (R. Hofstra, personal

communication). Since neither an amino acid change nor an alteration in splicing resulted

from this substitution, it most likely represents a very rare polymorphism.

Figure 6.3 GDNF mutation in a Hirschsprung (HSCR) patient. a) Sequence from GDNF

exon 2 in patient HS4-3 and a n o d control (N). HS4-3 is heterorygous for the n o d

(ACA) and mutant (TCA) sequence at codon 154. b) Restriction e v e digest analysis of

GDNF exon 2 PCR products with Hinff and HincII in family HS4. Digestion of the 478

bp product with Hinfl produces bands of 36 and 115 bp in the absence of the T154S

mutation. In the presence of the mutation, novel bands of 195 and 168 bp are generated by

cleavage of the 363 bp hgment. The 478 bp fragment is not digested by H i n d in the

absence of the mutation. When T154S is present, novel bands of 303 and 175 bp are

generated. A 100 bp DNA marker is indicated (M).

A C G T A C G T

H i n f I H i n c I I

The second change was detected in a patient with sporadic long segment HSCR and

no detectable mutations of RET or odw HSCR disease susceptibility loci. The sequence

variation was a substitution of an A for a T at base pair 460 in exon 2 (Figure 6.3). This

resulted in an amino acid change from threonine to serine at codon 154. The change was

not detected in 35 other HSCR patients nor in 301 normal controls 165 (Hofstra, personal

communication). Novel Hinfl and HincII restriction sites were predicted to result from the

substitution. The presence of these sites was confirmed upon digestion of amplified GDNF

products from the patient (Figure 6.3). The restriction e n y m cut sites were not detected in

either parent suggesting a de novo mutation of GDNF in the patient. Paternity was

coofirmed using three microsatellite markers h m chromosome LO. These data may

indicate a causative relationship between the GDNF mutation and the HSCR disease

phenotype in this individual.

Discussion

RET/GDNF/GDNFR-a in Wilms' Turnour

In this study, we investigated the expression of RET and the genes encoding

GDW and GDNFR-a, RET's recently identified ligand complex components, in a panel

of six WTs and one WT cell line, DW2. We found that REZ was expressed throughout our

p d of WTs and in the DW2 cell line. This result was intriguing given that WTs are

believed to originate from r e d blastemal cells, mesenchymal cells that give rise to

nephrons when induced by the ureteric bud 144. Normally, RET is not expressed in this

cell type 14- 349 40. While unlikely, it is possible that a small amount of contaminating

normal cells was responsible for the strong RET expression we detected in the primary

tumour samples. However, this answer cannot suffice for the mouse xenografied samples

where the possibility of contaminating normal human tissue is eliminated. Our results show

that RET expression becomes deregulated in WT cells. Whether this is a direct effkct of

transformation or a secondary effect of deregulation of other molecules is not known. Other

genes such as T M have been shown to be deregulated, presumabiy as a secondary event,

in Wilm's tumourigenesis 166. Like RET, TrkC is normally expressed in the ureteric bud

epithelium 167. 168. While TrkC expression is not detected in the minduced ureteric bud

nor in the uninduced blastema and differentiated structures derived fiom MM, its

expression is associated with more advanced stages of collecting duct differentiation in cells

derived f?om the ureteric bud epithelium 16% 16% TrM= mRNA and protein expression

have, nonetheless, been detected in WTs where expression has been localized to the

epithelial tubule elements '66. In this study, we were not able to determine whether RET

expression occurs in a i l WT cell types or is localized to subsets of cells with specific

histological features. However, we have shown that RET and 5' splicing variants of RET

are a l l strongly expressed in a panel of WTs with histologies ranging from predominantly

blastemal to predominantly stmmal-differentiated.

Four of the five tumours in our panel for which we have histological data are

predominantly composed of a population of cells with a differentiated phenotype (Tabie

6.1). A f~ turnour, WT27A, is described as predominantly blastemal. Interestingly, this

tumcur expresses both RET and GDNFR-a but not GDNF. Given that WTs arise from

MM cells, GDNF expression would be predicted in a l l WTs. Absence of GDNF

expression in WT27A must represent a loss of gene expression in these tumour cells.

Neither GDNF nor GDNFR-a expression was detected in the DW2 cell line which

expressed RET. No clear pattern of expression of the three components could be discerned

for WT. The existence of three genes that predispose to WT has been described and at least

a fourth is predicted (reviewed in 143). Given that mutations in any one of these genes can

result in a similar phenotype, it is possible that the different mutated genes affect an

overlapping set of transcriptional regulatory pathways thus explaining RET expression in

all turnours investigated but the absence of GDNF and GDNFR-a expression in select

WTs Is** 153- 16% The variable expression of RET, GDNF and GDNFR-a in WTs

may represent different t u m o ~ etiologies which result in expression and/or repression of

unique subsets of mRNAs.

Kidney development proceeds via interaction between two different cell types, the

ureteric bud epithelium, where RET is expressed, and the metanephric mesenchyme, where

GDNI; is expressed 14- 2% 35. In contrast, tumours arise from a single cell type. The

expression of each of RET, GDNF and GDNFR-a in subsets of WT indicates that these

genes are all expressed in the tumour cell of origin unlike the localization of expression in

kidney development Differences in the pattern of gene transcription between differentiated

WT structures and the equivalent structures in fetal kidney are not uncommon. For

example, there is not a point during kidney development when WT- 1, PAX-2 and PAX-8

are a l l expressed at high levels in the same cell type although al l are expressed in the

condensed mesenchyme and/or its derivatives in normal developing kidney 44- 171-

PAX-2 and PAX-8 are also expressed at low levels in the ureteric bud epithelium where

W7'-1 expression is absent. Once differentiation of the MM commences, PAY-2 and PAX-

8 transcription attenuates while WT-1 expression is upregulated rapidly s4. l72. However,

many WTs display elevated expression of both PAX-2 and PAX-8 in highly differentiated

epithelial structures 172 173. This observation is consistent with deregulation of PAX-2 and

PAX-8 expression in Wilms' tumourigenesis compared to the tight control of expression

during kidney development. Similarly, RET, GDNF and/or GDNFR-a expression may be

deregulated in W s resulting in expression in a cell type where the gene(s) is not normally

expressed It is equally possible that disruptions in WT genes are not only capable of

activating but also of repressing transcription of genes such as GDNF which are normally

expressed in the cell thus explaining absence of GDNF expression in WT27A.

RET, GDNF and GDNFR-a Expression in K C

RET expression was identitied in 1111 1 RCC and corresponding matched normal

kidney cortex samples. The expression levels in the tumours appeared to be slightly lower

than in the normal samples. The results of RT-PCR amplification of GUSB, a low-level

house-keeping gene, confirmed that this observation was not just a reflection of differences

in RNA quality between normal and tumour samp1es 6.2). GDNF expression was

only detected in the matched normal controls at very low levels. GDNFRa expression was

detected in all tumours and matched normal samples at relatively high Levels compared to

GDNF expression in these same samples.

The presumed cell of origin of RCC, proximal tubule epithelium, is a derivative of

the metanephric mesenchyme and might not be predicted to express RET or GDNF

(reviewed in 155 and 174). Normally, RET is expressed in ureteric bud epithelium and

derivatives while GDNF is expressed early in uninduced MM and early derivatives but not

in the more differentiated cell types derived from MM 14- 29. Studies delineating cell type-

specific expression of GDNFRa in the normal kidney show that it is normally expressed

at least in some of the same cells as RET, adjacent to GDNF-expressing cells 22. The

absence of relatively high expression levels of REI or GDNF in RCCs may simply be a

reflection of the tumour cell of origin which might not be expected to express either

transcript. In fact, RET expression levels in RCC samples appeared to be similar to

expression levels in the matched normal kidney tissues. This observation raises the

possibility that the trace expression we observed in tunour samples was due to the

presence of residual normal tissue contaminating the huwur sample.

Expression of GDNFR-a has been shown to be relatively high in the normal

kidney 21- 2 2 33. Given that localization of GDNFRa expression in the kidney appears to

mimic that of RET, co-expression of RET and GDNFR-a in the same cells is not

surprising. Their co-expression in RCC is, however, surprising given that the tumour cell

of origin is not predicted to express RET. It is possible that transcription of these genes is

affected by mutations in a gene predisposing for RCC. Analyses of sporadic RCCs have

revealed loss of one VHL dele in 95-9796 of turnours with kquent point mutations or

hypermethylation of the remaining allele 175-17? VHL is involved in the inhibition of gene

transcription and is ubiquitously expressed in the rnetanephric, ureteric and stromal

components of the developing kidney 178. There appear to be several factors that contribute

to the initiating or regulating events of VHL in normal organogenesis as well as in

pathogenesis. It is possible that any of these might also affkct RET and GDNFR-a

expression in these same cells causing them to be expressed when they are not n o d y

expressed.

Reports of GDNFR-a expression in cells that do not express RET, such as

Schwann cells, suggest that GDNFR-a expression in RCC might not be unexpected 22. In

sinc hybridization and immunocytochemical analyses indicate that GDNFR-a expression is

more widespread than RET expression in the developing kidney 22. This provides an

alternative explanation for the relatively high GDNFR-a expression levels we detected in

both RCC and matched normal kidneys compared to low levels of RET. We cannot rule out

the possibility that GDNFR-a has the capacity to interact with molecules other than GDNF

and RET and hence stimulate different downstream pathways that are involved in RCC

tumourigenesis.

The absence of GDNF expression in RCC most Likely reflects the tumour cell of

origin which is not normally predicted to express GDNF. Similar low levels of RET

expression were detected in both RCCs and matched n o d kidney samples. GDNFR-a

expression was relatively higher in both turnour and normal kidney samples compared to

RET expression. RET and GDNFR-a expression in these tumours may be the result of

disruption of regulatory pathways during the course of tumourigenesis.

GDNF Mutation Analysis m Hirschpmng Disease

We screened a panel of 36 sporadic HSCR patients for mutations in GDNF, a gene

that encodes one of the RET ligand complex members. Within our panel, we identified one

patient with a very rare polymorphism, a conservative substitution of G->A at GDNF base

pair 429. HSCR disease in this patient was previously found to be associated with a

deletion of chromosome l0q 1 1.2-2 1.2 which spanned the RET locus lo I . This same

polymorphism has been reported by R Hofstra in a patient with a familial form of HSCR

(personal communication).

There are reports in the literatwe of polyrnotphisms in a gene modulating

susceptibility to disease. For example, Creutzfeld-Mob disease (UD) occurs in both

sporadic and f e d forms 179- 180. Familial CTD is associated with the presence of

mutations in the pnon protein gene (PRNP). In addition to pathologic mutations of PRNP,

homozygosity at the site of a common poIymorphism at codon 129 is implicated in the

development of sporadic and infectious CJD The disease phenotype of familial

CJD is also modulated by this same polymorphism 182. Mutations in codon 178 of PRNP

can either result in fatal familial insomnia (FFI), a dementing illness. or familial CTD 183-

l86. The genotypic basis for the difference between FFI and CJD associated with this

PRNP mutation Lies in the codon 129 polymorphism IB7. Individuals with a mutation in

PRNP at codon 178 and homozygous for valine at codon 129 develop CJD whereas

individuals with the same mutation in codon 178 and homozygous for methionine at codon

129 develop FFI. In light of the fact that polymorphisms can affect disease phenotype. it is

possible that GDNF sequence variations such as the one we detected have bearing on the

expression or severity of the HSCR phenotype in patients with RET or other HSCR

mutations without being causative for the disease. It is, however, difficult to predict what

the effects of a sequence variation that does not lead to an amino acid change could be on

the GDM: protein.

In a second patient with sporadic long segment HSCR, we found a de novo

mutation of GDNF at base pair 460 resulting in substitution of s e ~ e for threonine. The de

now occurrence of the mutation was c o n w e d using restriction enzyme digests which

identified a cut site generated by the mutation in the patient but not in either parent. The fact

that both RET-I- and GDNF-I- mice displayed absence of enteric ganglia suggested that

inactivation of GDNF could result in a similar phenotype as RET inactivation 34- 46-48.

Thus, any gene encoding a member of the RET multicompouent Ligand complex could

represent a candidate gene for HSCR disease. Similar phenotypes are attniutable to

mutations in genes encoding either the c-KlT receptor or its corresponding ligand SLF

188. Interactions between the products of these genes influence various developmental

processes including survival, proliferation andor differentiation of germ cells, pigment

cells and hematopoietic cells (reviewed in 139). In the mouse, mutations at either the white-

sponing/c-KIT 189 or SLF 190 locus result in almost identical defects including dominant

white-sponing, sterility and severe macmcytic anemia (reviewed in 139). Mutations in

either c-KFT or SLF control the human piebald phenotype by resulting in a reduction in

tyrosine kinase activity and failure of mlawcytes to thrive and reach the skin during

embryogenesis 191- 192- It is possible that mutations in GDNF affect the HSCR phenotype

similar to mutations in RET. Whether or not mutations and/or polymorphisms in GDNFU-

a can also contribute to the HSCR phenotype remains to be established.

We and others have shown that the occurrence of GDNF mutations in HSCR

patients is very infrequent 165. 193. Angrist et al. 193 identified one familial GDNF

mutation in a HSCR patient with a known de novo RET mutation while Salomon a al. 165

identified mutations in patients with either RET mutations or trisomy 21, a chromosomal

anomaly known to predispose to HSCR disease. Their results suggested that GDNF

mutations were neither necessary nor sufficient for the genesis of the disease phenotype.

The de novo case of HSCR we report here is associated with a de novo mutation of GDNF

in the absence of any RET mutations or nisomy 2 1. In addition, the patient does not have

any other associated HSCR disease phenotypes such as deafness or pigmentary

abnormalities which might suggest the involvement of another known HSCR locus such as

EDN3 or EDNRB 98.1949 195. While the GDNF mutation we have identified in this patient

could be purely coincidental with the occurrence of HSCR, it is equally possible that the

TIS4S mutation has functional significance for GDNF. A better understanding of this

awaits analysis of the functional domains of GDNF.

The ligand for RET consists of a multicomponent complex which includes GDNF

and GDNFR-CY. The possibility exists that mutations in GDNFR-a can contribute to or are

responsible for the HSCR disease phenotype. It is equally possible that other RT3 ligand

complex members exist but have not yet been elucidated or that GDNF and GDNFR-a

family members exist that are also capable of interacting with REX Any of these might also

contriiute to the genesis of HSCR disease.

Miterids and Methods

Reverse Trunscription and Polymerase Chain Reaction Conditions

Total RNA was extracted h m RCC and matching normal tissues and fiom Wilms'

tumour samples and first strand cDNA synthesis was performed as described Chapter 3.

Sequences of primers used in PCR amplification of cDNA templates are given in Chapter

2. PCR was essentially as described in Chapter 3 with the exception of the MgClz

concentrations included in reactions. 1 SmM, 0.75mM, and 1 SmM concentrations of

MgQ were included in PCRs to amplify RET, GDNF and GDNFR-a sequences

respectively. PCR proceeded for 40 cycles of 950C for 1 min/550C for 1 W 2 0 C for 1

min. Cycling concluded with a final extension of 72% for 10 minutes.

Genomic Sequencing

DNA was extracted fiom peripheral blood using standard protocols 122.

Sequencing analyses were as described in Ivanchuk et al. l21.

Chapter 7

Discussion

RET in Hwnan Kidney Development

Proliferation and differentiation events are both required for maturation of the

human kidney. Inductive stimuli from the ureteric bud cause the rnetanephric mesenchyme

to proliferate and differentiate into nephrogenic and stromagenic cells (reviewed in 38 and

39). Under the stimulus of the metanephric mesenchyme, the ureteric bud grows and

branches 35. Expression studies in rodents have shown high levels of RET expression

early in development and much lower levels or absent expression in adult animals 1416.

We have identified a similar pattern of RETexpression in the developing human kidney 18.

Overall, RET expression was found to be approximately 7-fold higher at the earliest stage

of human kidney development investigated (7.5 weeks) compared to RETexpression in the

adult kidney. Around 6 weeks gestation, the ureteric bud is predicted to invade the

metanephric mesenchyme, establishing contact between the two populations of cells and

facilitating reciprocal inductive interactions. Branching morphogenesis and mesenchymal

differentiation. the results of reciprocal inductive interactions, begin around 6 weeks

gestation in humans and continue through to 32-34 weeks gestation (reviewed in 37).

Ureteric bud branching is rapid between 8 and 14-15 weeks gestation. RET expression was

found to decrease 1.5-fold by 8.5 weeks gestation, which is close to the time at which

rapid ureteric bud bifurcation is initiated (8 weeks). RET expression was found to decrease

considerably around 14 weeks gestation, when the rate of bud branching slows, and

approach adult kidney expression levels. The relatively high level of RET expression

observed in human fetal kidney aged 7.5 weeks compared to expression in adult kidney is

consistent with a role for RET early in kiduey development.

In the embryonic murine kidney, RET expression is found in epithelial cells

throughout the uninduced ureteric bud l6l6. W~th increasing gestational age, RET

expression localizes to the branching tips of the ureteric bud within the nephrogenic zone,

the site of reciprocal inductive interactions between the ureteric bud and the metanephric

mesenchyme 14- 15. The observed expnssion pattern is consistent with the phenotype of

the RET4- mice and suggestive of a role for this receptor in kidney development. REI-I-

mice display variable degrees of kidney dysgenesis characterized by growth failure and

reduced branchings of the ureteric bud as well as large areas of undifferentiated

metanephric mesenchyme 34. Th majority of RET-/- mice (67%) develop ureteric buds

indicating that R E ? signaling is not the first step in metanephric kidney induction 34.

However, in approximately one-half of these cases, the bud fails to invade the metanephric

mesenchyme. In comparison, the ureteric bud fails to form altogether in 100% of WT-I -/-

mice 41. W e molecules such as WT-1 are involved in ureteric bud formation, others such

as RET are necessary for interadom between the ureteric bud and metanephric

mesenchyme (Figure 1 5 ) .

RET 5' Alternatively Splicing Variants

Alternative splicing variants of RET are generated when exon 2 is spliced to any of

exons 3, 4, 5 or 6 (Figure 1.3) 13. The resulting transcripts differ from transcripts

encoding full length RET in that they lack sequence encoding portions of the RET

extraceUular domain. In the case where exon 2 is spliced to exon 5, a soluble product is

predicted due to a premature stop codon that resuits from translation of exon 5 out of h e

13. The expression levels of these variant RET mRNAs were found to vary during the

different stages of human kidney development 18. The expression levels of the REZW4 and

REZ2/5 alternatively spliced variants were consistently one-third the level of FL RET

expression. These results suggest that levels of these transcripts vary proportionally with

changing levels of overall RET expression during kidney development. The expression

pattern of the third transcript, RER16, was markedly different from those of the others.

Expression of -6 was found to be highest in our earliest sample, 8 weeks gestation,

the point in kidney development when rapid division of the ureteric bud is initiated

(reviewed in 37). Exp~ssion levels remained relatively high through to 14 weeks gestation

after which RETu6 expression levels were found to decrease (Figure 3.3). Decreased

expression of RETU6 coiacides with a period in kidney morphogenesis when the rate of

ureteric bud division slows and bifiucation is infkquent (14-20 weeks) (reviewed in 37).

Around 24 weeks gestation, expression levels of R E N 6 were found to approach those of

adult kidney. The high levels of RETZ16 expression during a period of rapid branching of

the ureteric bud and the corresponding decrease in expression when bifurcation slows may

suggest that the RET isoform encoded by this transcript is significant to the branching

process.

The distinct expression profile of REW6 relative to the other RET 5' alternatively

spliced variants during human kidney morphogenesis suggests that it is capable of

modulating a different subset of RET activities. These activities could be mediated at the

level of ligand binding. The alterations in the extracellular domain that result fiom

alternative splicing at the 5' end of RET result in novel RET isoforms which may have

distinct speciticity or affiaity for a RET ligand. For example, RER16 encodes a product

which lacks the entire region of cadherin homology present in the full length RET protein

13. If this protein is capable of binding the GDNWGDNFR-a complex, it is likely that its

affiaity for the complex differs from that of full length RET. It is equally possible that

RET2/6 binds an alternate RET ligand. The identification of new GDNF and GDNFR-a

family members has raised the possibility that Ligand relatives are also capable of binding to

RET. Neurturin was identified as a relative of GDNF based on sequence homology and a

similar ability to support neuronal survival in 31. Although these results suggest that

neumuin may act through signaling pathways similar to those of GDNF, the ability of

neurturin to interact with RET has yet to be investigated. Human RETL2 has been

identified as a relative of GDNFR-a and displays 49.146 homology to human GDNFR-a

33. Human -2. l i k GDNFR-a, can facilitate GDNF-dependent phosphorylation of

RET. In the absence of RET, GDNF has been shown to bind to GDNFR-a expressing

cells but to bind poorly to RETL2 expressing cells 3'. The differences in GDNF binding

observed for GDNFR-a and RETU are due to different requirements for RET to faciltate

complex formation: RETL2 binding to GDNF appears to be dependent on the presence of

RET whereas GDMR-a binding to GDNF is RET-independent.

As GDNFR-a and REIZ2 an both expressed in the embryonic kidney, it is

possible that either or both of these molecules Narc capable of modulating RET activity

during human kidney development 33. It will be interesting to leam if the various GDNF

and GDNFR-a family members are capable of interacting not only with RET but also with

its distinct isoforms. Little is known about the functions of these RET isoforms. In vine

translation of proteins fiom the corresponding cDNAs has been described 13 (M. Billaud,

personal communication). When these constructs were transfated into a kidney cell Line,

proteins were generated and those isoforms predicted to retain the transmembrane domain

were found to localize at the cell surface (M. Billaud, personal communication). This

suggests that the mechanisms are in piace for translocation of these RET isoforms to the

cell surface at least in this cell type.

GDNF and GDNFR-a in Human Kidney Development

The renal agenesis observed in the RET-1- and G D W - mice indicates the

importance of both of these molecules in kidney development 3 4 v 4 6 4 . It has been shown

that GDNF requires GDNFR-a for high affinity binding to the RET receptor 21. 22, 33. In

the absence of GDNFR-a, GDNF binding to RET and RET autophospholylation is

minimal. Chapter 5 describes the expression of RET ligand complex components in the

developing human kidney. The data are consistent with GDNF and GDNFR-a expression

during kidney development in rodents suggesting that the pathway is conserved in humans

22, 29.

A developmental expression pattern for GDNFR-a in human kidney

morphogenesis has been observed. GDNFR-a expression was found to be higher in the

early fetal kidney ages then decrease with increasing gestational age. Comparison of the

downward trends in RET and GDNFR-a expression during the come of human kidney

development indicates they are dissimilar. GDNFR-a expression levels were found to be

~latively consistent through the early ages of fetal kidney (7.5-14 weeks) whereas RET

expression was found to decrease dramatically between 7.5 and 8.5 weeks. A significant

decrease in GDNFR-a expression was not noted until 18 weeks gestation. These results

indicate that RET and G D N F R a expression are not coordinately regulated in the

developing human kidney. In light of the f a that GDNFR-a is required for high affinity

binding of GDNF to RET, it is clear that the reguiation of GDNFR-a expression during

human kidney development may act as a mechanism for regulating RET activation. In v i m

experiments have shown that overexpression of GDNFR-a augments GDNF binding to

RET and, hence, RET autophosphorylation 21- 22. This suggests that RET activity can be

regulated at the level of GDNFR-a. At early stages of kidney development, where both

GDNFR-a and RETexpression levels are high and GDNF is present, one might expect to

see increased levels of RET activation. When REI expression decreases, GDNFR-a

expression levels are maintained During this period of reduced RET expression, the high

levels of GDNFR-a could, in the presence of GDNF, be important for the maintenance of

RET activation. Regulation of RET activity at the level of GDNFR-a could be important in

tissues such as the kidney which co-express GDNFR-a family members such as EETL2

which is also capable of interacting with RET and GDNF. Sanicola et aL 33 suggest that

dissociation of the RETt2/GDNF/RET complex would release GDNF from the surface

while dissociation of the GDNFR-WGDNFIRET complex would leave GDNF bound to

GDNFRu on the cell surface and thus free to reassociate with RET. In the absence of the

ability to retain GDNF at the cell surface, RETL2 expressing cells would be more sensitive

to fluctuations in the local concentration of free GDNF.

RET 3' Alternatively Sjdiced Variants

RET appears to have roles in both protiferation and differentiation. A role for RET

in proliferation is suggested at both the in vim0 and in vivo levels. In vivo evidence for

RET's role in proliferation comes from MEN 2 patients. In these individuals, hyperplasia

of the calcitonin-producing C-cells of the thyroid is associated with activating mutations in

RET 8% 196. At the in vitro level, transfections of chimeric constructs resulting in

constitutive RET activation or of constructs with activating RET mutations both result in

tyrosine phosphorylation correlated with potent transformation and mitogenic activities 8'-

133. 197- 198. Conversely, the expression of constitutively activated RET in PC12 cells

results in morphological changes such as neurite outgrowth and induction of nemfilament

mRNA expression both of which are associated with differentiation in this cell type 5% 134-

198. The RET 3' isoforms appear to have varying effects on proliferation and

differentiation. While there are no functional data to suggest the role of the RET43 isoform,

there are data to suggest different roles for RET9 and RET5 1. Interestingly, we observed

RETSI was expressed at very low levels during the early stages of human kidney

development (7.5-8.5 weeks) and expression was upregulated around 9 weeks gestation

(Figure 4.5). Compared to RET5 1 expression. expression was relatively constant in

all ages of human fetal kidney examined There is an absolute requirement for RET early in

kidney development and our expression data suggest that RET9 possibly in conjunction

with RET43. satisfies this requirement. RElSl expression is upregulated at a later stage

when ureteric bud bifurcation is rapid which may indicate an important role for this isoform

in branching morphogenesis. Interestingly, the situation is different in the development of

the ENS. The identification of a R.1-specific mutation in a HSCR patient shows that

RETSI is absolutely required for the development of neurons in the lower gut 105. It is

possible that requirements for REn 1 expression during embryogenesis are tissue-specific.

Further studies are required to examine the details of RET9 and RET5 1 functions in kidney

development However, our data indicate that the two proteins have distinct roles in the

process.

RET9, RET51 and RET43 have in common the first 1062 amino acids but differ

with respect to their C-termini beginning at residue 1063. As a result of sequence

divergence at Y1062, RETSl contains an additional two tyrosine residues (Y 1090,

Y1096), not found in either of RET9 or RET43, one of which can be autophosphorylated

(Y lO96) lo- 12,136. While both RET9 and RETS1 isofonns display certain similarities with

respect to -interactions with downstream effector molecules, differences between the two

isoforms in their interactions have also been observed 5 6 s7 (M. Billaud, personal

communication). RET9 and RETS 1 differ with respect to their relative binding affinities for

the SH2 and PTB domains of the Shc adapter protein 57. Although Y1062 is the last

common amino acid shared by both RJ39 and RETS1. as a result of alternative splicing

after sequence encoding Y1062, this tyrosine residue is placed in different amino r i d

contexts in the translated products. M o u s l y , it was shown that the amino acid sequence

immediately upstream or downstream of pY residues confers preference for downstream

effector interactions (reviewed in 199). While Y 1062 has a greater affinity for Shc-SH2 in

the sequence context of RET9, in the sequence context of RETS I it has a greater affinity for

Shc-PTB interaction domain 56- 57 (M. Biilaud, personal communication). As a result of

the additional two tyrosine residues present in RET51, only RETSl contains a Grb2

interaction site (Y 106) and, hence, is capable of binding Grb2 directly via the SH2

interaction domain 56.57- In addition to binding to activated receptors. Grb2 is capable of

simultaneously binding to phosphorylated Shc in vivo (reviewed in 199). This enables

Grb2 to form multiple activated complexes with other SH2-containing proteins as well as

with activated receptors.

Upon phosphorylation of the receptor, both RET9 and RET5 1 have been shown to

associate with downstream molecules that interact with effectors of the RAS activation

pathway 56.57~ 200. The S hc and Grb2 adapter proteins are two such molecules. Both S hc

and Grb2 interact with phosphorylated RET via SH2 interaction domains while Shc is

additionally capable of interacting with RET via its PTB domain 56. Thus, RET 3' isoforms

differ in their downstream interaction and, potentially, these differences contribute to

distinct abilities of REIP and RETS1 to elicit downstxeam signaling responses. As

descn id above, our data show that RET5 1 expression is considerably lower in early fetal

kidney relative to expression at 9 weeks gestation and beyond. These results indicate that

RET51 is not required during the earliest stages of human kidney development d i k e

RET9 which is consistently expressed at relatively higher levels throughout kidney

organogenesis. The EZET5 1 isofom-specific mutation associated with HSCR supports the

theoxy of distinct roles for the RET 3' isoforms during development 105. The mutation in

this patient affects an amino acid residue immediately downstream of Y 1062 (M1064T). in

vitro, the mutation disrupts the interaction of N X 5 1 with the ITB domain of Shc 57.

Given that Shc phosphorylation is Linked to the RAS pathway, disruptions in Shc

phosphorylation are likely to disrupt activation of signaling through the RAS pathway.

Differences in expression of the RET 3' alternatively spliced variants such as those

identitied in the developing human kidney could be an important modulator of activation of

these pathways in these tissues.

The extremely low levels of REnl expression in the earlier stages of kidney

development investigated compared to 9 weeks gestation is interesting. RET9 and RETSl

isoforms appear to use distinct interactions to modulate downstream signaling pathways.

RAS activation can result in either proliferation or differentiation effects and it is not clear

which results from RET activation during the various stages of human kidney development

(reviewed in 201). The regulated expression of the EtET 3' isoforms in the developing

kidney could be an important determinant in proliferation versus differentiation both of

which are precisely timed events during development. The unique expression of REnl

relative to RE19 during human kidney development suggests non-compiementary roles for

these proteins. The RET51 isofom-specific HSCR mutation suggests that functional

RETS1 is required at a specifk time during development to ensure proper development.

The upregulation of RE151 expression we Qtected in a 9 week fetal kidney sample

coincides with a period of rapid ureteric bud bifurcation. Potentially, RE15 1 contributes to

the control of this process by affecting key downstream pathways responsible for

transcription of molecules involved in differentiation events. The relatively high levels of

RE19 expression throughout human kidney development suggest that RET9 has a role in

kidney morphogenesis that is distinct from that of RET5 1.

RET, GDNF and GDFNR-a in Human Disease

The mature human kidney consists of a number of different cell types each of which

is derived h r n either ureteric bud epithelium or metanephric mesenchyme 35. During the

early formative stages, the cells that give rise to the mature kidney express any of RET,

GDNF andlor GDNFR-a 14- 21. "- 29. RET is expressed in the ureteric bud epithelium

and its derivatives that give rise to the collecting duct system 14. While GDNFR-a

expression overlaps with RET expression in cells of the nephrogenic zone based on

irnmunocytochemistry, similar studies indicate that GDNFR-a may be expressed in a

broader range of cells 22. GDNF expression in metanephric mesenchyme is restricted to

undifferentiated cells which give rise to progeny with the potential to become either

nephrogenic or stromagenic cells (Figure 1.5) 29- 34. Given the cell typespecific

expression of these genes, some of the results of our investigation of RET, GDNF and

GDNFR-a expression in renal tumom are surprising. Renal cell carcinoma is thought to

arise from proximal tubule epithelial cells, highly specialized cells derived from

metanephric mesenchyme (reviewed in 174). We found Low Levels of GDNF expression in

adult kidney cortex samples but not in any of the I1 RCC samples dissected from the same

kidneys. Absence of GDNF expression in RCC is consistent with the tumour cell of origin.

The lack of expression in cells derived from metanephric mesenchyme following epithelial

conversion is predicted to be required to limit ureteric bud branching to the more distal

regions of the organ, preserving the radial patterning *02. Sanchez et al. 48 have shown

very Little GDNF is expressed in the tubular epithelium of the developing kidney suggesting

that GDAF expression is downregulated after epithelial differentiation.

RET expression appeared to be higher in n o d kidney samples relative to RCC

samples based on the intensities of the corresponding PCR products. Low levels of RET

expression are normally maintained in the adult kidney. The RET expression detected in

RCC might reflect the presence of contaminating normal cells which had been dissected

with the tumour sample. Alternatively, RETexpression in these tumours may be the result

of deregulated expression in this cell type. Mutations in genes responsible for RCC may

affect pathways involved in control of RET expression resulting in RET expression in a cell

type where it is normally absent.

GDNFR-a expression levels appeared similar in the RCC and matched normal

adult kidney cortex samples based on the intensities of the PCR products. GDNFR-a

expression in these tumours is not necessarily surprising. GDNFR-cx is expressed in cell

types which do not express RET. For example, GDNFR-a is expressed in Schwann cells

whereas RET is not 22. In situ hybridization has been used to show that GDNFR-a is

expressed between the inner and outer smooth muscle layers adjacent to and possibly

within the ENS; RET is only present in the ENS 22. Thus, GDNFR-a expression in RCC

may reflect the normal pattern of expression in the cell of origin and, thus, may not be

unusual whereas Rmexpression in RCC is unexpected.

Unlike RCC, expression of each of RET, GDNF and G D F N R a was detected in

the majority of WTs investigated. RET expression was consistent amongst tumour

samples. Single nunours expressed one of either GDNF or GDNFR-a in addition to RET

indicating that the collective expression of RET and the genes encoding its Iigmd

components is a common but not a consistent feature of WT. Expression studies have

shown that GDNFR-a and RET are both expressed in developing nephrons adjacent to

sites of GDNF expression 22. This indicates that the three components are not normally

expressed together in the same cell during kidney development. The WT cell of origin is

thought to be a derivative of the MM (reviewed in 140, 142 and 143). Given the tunour

cell of origin, GDNF expression in these tumours is not unexpected. GDNF is normally

expressed in both uninduced metanephric mesenchyme and its early derivatives 22.29. Both

cell types can be elements of WT 144 As with RCC, however, RET expression in WTs is

unexpected given the tumow cell of origin. There are no data to suggest that RET

represents one of the predicted genes responsible for WT. As mutations in s e v e d different

genes can result in an identical WT phenotype, it is Uely that several pathways may be

responsible for this phenotype. In this sense, RET expression in WT may be the result of

mutations in genes which, while not directly involved in RET expression, deregulate

pathways that are. G D N F M expression in WTs may be explained similarly.

RET mutations are found in 1040% of HSCR patients (reviewed in 88). Mutations

at other known HSCR loci account for an additional 3 4 % of cases indicating that a

significant proportion of HSCR disease is due to mutations in unidentified genes 98. 194.

195*203- 204. We identified a mutation in GDNF in a patient with HSCR that did not have

any mutations in other HSCR genes. The phenotype of GDNF-/- mice has suggested

GDNF is a HSCR candidate gene. Mice that lack functional GDNF are devoid of neurons

of the myenteric plexus which is derived from the vagal neural crest 4648. The absence of

the ENS in GDNF-I- mice was previously noted in RET-1- mice 34. Our data suggest that

mutations in either the receptor, RET, or its cognate ligand, GDNF, can result in a similar

phenotype. It has been found that mutations in either c-KIT or SLF, a complementary

receptor and ligand, result in similar phenotypes in mice and humans (reviewed in 139). It

is interesting that we have identified a GDNF mutation in only one of 36 HSCR patients

121. Similar studies have identified GDM:mutations in only four of 279 individual HSCR

patients 165. 193. Givea the low kquency of GDNF mutations in HSCR patients, it

appears that GDNF is not a major HSCR susceptibility locus.

In addition to a GDNF mutation associated with HSCR, we have identified a rare

GDNF polymorphism in a case of HSCR associated with deletion of a single copy of the

RET gene. The possibility exists that the polymorphism has a modulating effect on the

disease phenotype in this individual, however, any genotype-phenotype correlation has yet

to be proven. A situation where a polymorphism can affect disease phenotype is the prion

protein gene (PRNP). Whether individuals are homo- or heterozygous for a polymorphism

at codon 129 of PRNP affects susceptibility to sporadic or infectious CJD 179- 1%

Additionally, individuals bearing mutations at codon 178 of PRNP can present with either

f d a l CID or familial fatal insomnia depending on the nature of the polymorphism at

codon 129 182. It is unclear if the GDNF polymorphism we detected contributes to disease

phenotype in this individual, however, the possibility does exist that polymorphisms

contribute to the expression and/or severity of the HSCR disease phenotype.

Summary

The mature kidney arises from a series of reciprocal inductive interactions between

the ureteric bud epithelium and the MM. The RET proto-oncogene encodes an EUK

required for kidney development. We have identified developmental regulation of RET in

human kidney development. The datively high levels of RET expression in early

gestational ages compared to later in development indicate a role for RET in the early

inductive processes. In addition, we have identified variation in the expression levels of

alternatively spliced RETcoding variants over the course of kidney development Although

the functions of the isoforms these transcripts encode have not yet been fully characterized,

the developmental regulation of expression indicates roles for these proteins during specific

phases of kidney development.

GDNF and GDNFR-cx were recently identified as components of a ligand for RET.

GDNF is required for branching of the ureteric bud and differentiation of the MM in kidney

development similar to RET. We have identified expression of both GDM; and GDNFR-a

in the developing human kidney. In addition, we have shown that GDNTR-a expression is

developmentally regulated in this organ. Relatively high levels of GDNFR-a expression

are maintained when RET expression levels are downregulated. Given that GDNFR-a is

required for high affinity binding of GDNF to RET, our data suggest that regulation of

GDNFRa expression may be a means for regulating RET activity.

Abrogation of RET function contniutes to developmental anomalies. Inactivating

mutations in RET account for 104% of HSCR. We have identified a de n o w mutation in

GDNF suggesting that mutations in either the Iigand (GDNF) or the receptor (RET) can

result in a similar disease phenotype.

The expression of each of RET, GDNT and GDNFR-a in subsets of WTs is

interesting. Given that the WT cell of origin is believed to be derived fkom the MM, RET is

not predicted to be expressed in these tumours. Deregulation of transcriptional control

pathways during the course of tumourigenesis may result in RET expression as well as in

GDNFR-a expression in these tumours. The same can be said of RET and GDNFR-a

expression in RCC. Whether or not the expression of these genes has any tunour-specific

significance has not been investigated.

References

1. Takahashi, M., J. Ria , and G. M. Cooper. 1985. Activation of a novel human

transforming gene, ret, by DNA rearrangement. Cell. 4258 L -588.

2. Kuneida, T., M. Matsui, N. Nomura, and R. Ishizaki. 199 1. Cloning of an activated

human ref gene with a novel 5' sequence fused by DNA remgement . Gene. 107:323-

328.

3. Takahashi, M., and G. Cooper. 1987. ret transforming gene encodes a fusion protein

homologous to tyrosine kinases. Mol Cell Biol. 7: 1378-1385-

4. Itoh, F., Y. Ishizaka, T. Tahira, M. Yamamoto, A. Miya, K. Imai, A. Yachi, S. Takai,

T. Sugimwa, and M. Nagao. 1992. Identification and analysis of the ref proto-oncogene

promoter region in neuroblastoma cell lines and medullary thyroid carcinomas from

MENZA patients. Oncogene. 7: 120 1- 1206.

5. Kwok, J. B. J., E. Gardner, J. P. Warner, B. A. I. Ponder, and L. M. Mulligan. 1993.

Structural analysis of the human ref protmncogene using exon trapping. Oncogene.

8~2575-2582.

6. Myers, S., C. Eng, B. Ponder, and L. Mulligan. 1995. Characterization of RET proto-

oncogene 3' splicing variants and polyadenylation sites: a novel C terminus for RET.

Oncogene. 1 1:2039-2045.

7. Pasini, B., R. M. W. Hofstra, L. Yin, R. Bocciardi, G. Santanaria, P. M.

Grootscholten, I. Ceccherini, G. Patrone, M. Priolo, C. H. C. M. Buys, and G. Romeo.

1995. A physical map of the human RETproto-oncogene. Oncogene. 1 1: 1737-1743.

8. Ishizaka, Y., F. Itoh, T. Tahira, I. Jkeda, T. Sugimura, J. Tucker, and A. Fertitta.

1989. Human ret proto-oncogene mapped to 1Oq 1 1.2. Oncogene. 4: 15 19- 152 1.

9- Takahashi, M., Y. Buma, T. Iwamoto, Y. Inaguma, H. Ikeda, and H. Hiai. 1988.

Cloning and expression of the ret protooncogene encoding a tyrosine kinase with two

potential transmembrane domains. Oncogene. 357 1-578.

LO. Takahashi. M., Y. Buma, and H. Hi& 1989. Isolation of ref protooncogene cDNA

with an amino-terminal signal sequence. Oncogene. 4:805-806.

1 1. Tahira, T., Y. IshiAa, T. Sugimura, and M. Nagao. 1988. Expression of proto-ret

mRNA in embryonic and adult rat tissues. Biochem Biophys Res Cornmun. 153: l29O-

1295.

12. Tahira, T., Y. Ishizaka, F. Itoh, T. Sugimura, and M. Nagao. 1990. Characterization

of ret protooncogene mRNAs encoding two isoforms of the protein product in a human

neuroblastoma cell line. Oncogene. 5:97-102.

13. Lorenzo, M. J., C. Eng, L. M. MulIigan, T. J. Stonehouse, C. S. Healey, B. A. J .

Ponder, and D. P. Smith. 1995. Multiple mRNA isoforms of the human RET proto-

oncogene generated by alternative splicing. Oncogene. 10: 137701383.

14. Pachnis, V., B. Mankoo, and F. Costantini. 1993. Expression of the c-ref proto-

oncogene during mouse embryogenesis. Development. 1 19: 1005- 10 17.

15. Avantaggiato, V., N. Dathan, M. Grieco, N. Fabien. D. Lanaro, A. Fusco, A.

Simeone, and M. Santoro. 1994. Developmental expression of the RET protooncogene.

Cell Growth and Differentiation. 5:305-3 1 1.

16. Tsuzuki, T., M. Takahashi, N. Asai, T. lwashita, M. Matsuyama, and J. Asai. 1995.

Spatial and temporal expression of the ret protooncogene product in embryonic, infant and

adult rat tissues. Oncogene. 10: 19 1- 198.

17. Nakamura, T., Y. Ishizaka, M. Nagao, M. Ham, and T. Ishikawa. 1994. Expression

of the ret proto-oncogene product in human normal and neoplastic tissues of neural crest

origin. J Pathol. l72:255-260.

IS. Ivanchuk, S. M., C. Eng, W. K. Cavenee, and L. M. MuIligan. 1997. The expression

of RET and its multiple splice forms in developing human kidney. Oncogene. In press.

19. Trupp, M., E. Arenas, M. Fainzilber, A.-S. Nilsson, B.-A. Sieber, M. Grigoriou, C.

Kilkenny, E. Salazar-Grueso, V. Pachnis, U. Arum&, H. Sariola, M. Saarma, and C. F .

Ibiiiez. 1996. Functional receptor for GDNF encoded by the c-ref proto-oncogene. Nature.

38 1 :785-789.

20. Durbec, P., C. V. Marcos-Gutierrez, C. Kllkenny, M. Grigoriou K. Wartiowaara. P.

Suvanto, D. Smith, B. Ponder, F. Costaathi, M. Saarma, H. Sariola, and V. Pachnis.

1996. GDNF signalling through the Ret receptor tyrosine kinase. Nature. 38 1 :789-793.

21. ling, S., D. Wen, Y. Yu, P. L. Hol~t, Y. Luo, M. Faag, R. Tamir, L. Antonio, 2.

Hu, R. Cupples, J.-C. Louis, S. Hu, B. W. Mtrock, and G. M. Fox. 1996. GDNF-

induced activation of the Ret protein tyrosine kinase is mediated by GDNFR-a, a novel

receptor for GDNF. Cell. 85: 1 1 13- 1 124.

22. Tremor, I. J. S., L. Goodman, F. de Sauvage, D. M. Stone, K. T. Podsen, C. D.

Beck, C. Gray, M. P. Armanini, R. A. Pollock, F. Hefii, H. S. Phillips, A. Goddad M.

W. Moore, A. Buj-BeLIo, A. M. Davies, N. Asai, IM. Takahashi, R- Vanden, C. E.

Henderson, and A. Rosenthal. 1996. Characterization of a multicomponent receptor for

GDNF. Nature. 382:80-83.

23. Lin. L.-F. H., D. H. Doherty, J. D. Lile, S. Bektesh, and F. Collins. 1993. GDm a

&a1 cell hederived neurotmphic factor for midbrain dopaminergic neurons. Science.

260: 1130-1 132,

24. Lin, L., T. Zhang, F. Collins, and L. Armes. 1994. Purification and initial

characterization of rat B49 glial cell line-derived neurotrophic factor. J Neurochem. 633758-

768.

25. Schindelhauer, D., S. Schaenhauer, T. Gasser, A. Steinkasserer, and T. Meitinger.

1995. The gene coding for glial cell line-derived neurotrophic factor (GDNF) maps to

chromosome Sp 12- 13.1. Genornics. 28:605-607.

26. Suter-Crazzolara, C., and K. Unsicker. 1994. GDNF is expressed in two forms in

many tissues outside the CNS. NeuroRepon. 5:2486-2488.

27. Choi-Lundberg, D. L., and M. C. Bohn. 1995. Ontogeny and distribution of glial cell

line-derived neurotmphic factor (GDNF) mRNA in rat. Devel Brain Res. 85230-88.

28. Trupp, M., M. Ryden, H. Jornvall, H. Funakoshi, T. Timmush E. Arenas, and C.

Ibfiez. 1995. Peripheral expression and biological activities of GDNF, a new neurotrophic

factor for avian and mammalian peripheral neurons. J Cell Biol. 130: 137- 148.

29. HeUmich, H. L., L. Kos, E. S. Cho, K A. Mahon, and A. Zimmer. 1996. Embryonic

expression of glial cell-he derived neurotrophic factor (GDNF) suggests multiple

developmental roles in neural differentiation and epithelial-mesenchymal interactions. Mech

Devel. 54:95- 105.

30. Unsicker, K. 1996. GDW. a cytokioe at the interface of TGF-Bs and neurotrophins.

Cell Tissue Res. 286: 175- 178.

31. Kotzbauer. P. T., P. A. Lampe, R. 0. Heuckeroth. J. P. Golden, D. J. Creedon, E.

M. Johnson, and I. Milbrandt. 1996. Neurnuin, a relative of glial-cell-line derived

neurotrophic factor. Nature. 384:467-470.

32. Wnfriend, S ., and K. Kodulrula. 1995. How glycosyIphosphatidylinositol-anchored

membrane proteins are made. Ann Rev Biochem. 64563-59 1.

33. Sanicola, M., C. Hession, D. Worley, L. Walus. P. Carmillo, C. Ehrenfels, S .

Robinson, G. larworski, H. Wei, R. Tizard, A. Whitty, R. B. Pepinsky, and R. Cate.

submitted. Two accessory proteins mediate glialcell-line derived neurotrophic factor

(GDNFJ-dependent RET signalling via distinct mechanism. Pruc Nut1 Acad Sci, USA.

submitted.

34. Schuchardt, A., V. D'Agati, L. Larsson-Blomberg, F. Costantini, and V. Pachnis.

1994. Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase

receptor Ret. Nature. 3673380-383.

35. Saxen, L. 1987. Organogenesis of the kidney. Vol. 19. P. W. Barlow, P. B. Green,

and C. C. White, editors. Cambridge University Ress, Cambridge. 1-87.

36. Durbec, P. L., L. B. Larsson-Blomberg, A. Schuchardt, F. Costantini, and V.

Pachnis. 1996. Common origin and developmental dependence on c-ret of subsets of

enteric and sympathetic neuroblasts. Development. 122349-358.

37. Potter, E. L. 1972. Normal and A b n o d Development of the Kidney. Year Book

Medical Publishers Inc., Chicago. 305 pp.

38. Bard, I., J. Davies, I. Karavanova, E. Lehtonen, H. Sariola, and S. Vainio. 1996.

Kidney development: the inductive interactions. Cell Devel Bid. 7: 195-202.

39. Bard, J. A., J. E. McCo~eu, and l. A. Davies. 1994. Towards a genetic basis for

kidney development. Mech Devel. 48:3- 1 1.

40. Schuchardt, A., V. D'Agati, V. Pachnis, and F. Costanthi. 1996. Remi agenesis and

hypodysplasia in ret-k- mutant mice result from defects in ureteric bud development.

Development. 122: 19 19- 1929.

41. Kreidberg, J. A., H. Sariola, I. M. Loring, M. Maeda, J. PeUetier, D. Housmm, and

R. Jaenisch. 1993. WT- 1 is required for early kidney development. Cefi. 74:67969 1.

42. Mendelsohn, C., D. Lohnes, D. Decimo, T. Lutkin, M. Lemeur. P. Chambon and M.

Mark. 1994. Function of the retinoic acid receptors (RARs) during development.

Development. 120:2749-277 1.

43. Armstrong, J. F., P.-J. K, W. A. Bickmore, N. D. Hastie, m d J. B. L. Bard. 1992.

The expression of the Wilms' turnour gene WT-1 in the developing mammalian kidney.

Mech Devel 40: 85-97.

44. Pritchard-Jones, K.. S. Heming, D. Davidson, W. Bickmore, D. Porteous, C.

Gosden, J. Bard, A. Buckler, J. Pelletier, D. Housman, V. van Heyningen, and N.

Hastie. 1990. The candidate Wilms' tumor gene is involved in genitourinary development.

Nature. 346: 194-197.

45. D a m , K., R. Heyman, K. Umesono, and R. Evans. 1993. Functional inhibition of

retinoic acid response by dominant negative retinoic acid receptor mutants. Proc Narl Acad

Sci. 90:2989-2993.

46. Moore, M. W., R. D. Klein, I. Faciiias. H. Sauer, M. Annanini, H. Phillips, L. F.

Reichardt, A. M. Ryan, K. Carver-Moore, and A. Rosenthaf. 1996. Renal and neuronal

abnormalities in mice lacking GDNF. Nature. 382:76-79.

47. Pichel, I. G., L. Shea, H. 2- Sheng, A*-C. Gcanhoh, J. Drago, A. Grinberg, E- I.

Lee, S. P. Huang, M. Saarma, B. J. Hoffer, H. Sariola, and H. Westphal. 1996. Defects

in enteric innervation and kidney development in mice lacking GDNF. Nature. 382:73-76.

48. Shchez, M. P., I. Silos-Santiago, J. Frisen, S. A. Lira, and M. Barbacid 1996.

Renal agenesis and the absence ofenteric neurons in mice lacking GDNF. Nutzwe. 382:70-

73.

49. Maas, R., R. Zeller, R. Woychik, T. Vogt, and P. Leder. 1990. Disruption of formin-

encoding transcripts in two limb dflormity alleles. Nature. 346:853-855.

50. Dudley, A., K. Lyons, and E. Robertson. 1995. A requirement for bone

morphogewtic protein-7 during development of the mammalian kidney and eye. Genes d

Devel. 9:2795-2807.

5 L. Dressler, G.. J. Wilkinson, U. Rothenpieler, L. Patterson. L. Wiams-Simons, and

H. Westphal. 1993. Deregulation of Pax-2 expression in transgenic mice generates severe

kidney abnormalities. Nature. 362:65-67.

52. Stark, K., S. Vainio, G. Vassileva, and A. P. McMahon. 1994. Epithelial

transformation of metanephric mesenchyme in the developing kidney regdated by Wnt-4.

Nature. 372:679-683.

53. Dressler, GOT U. Deutsch, K. Chowdhury, H. Homes, and P. Gruss. 1990. Pax2, a

new murine paired-box containing gene and its expression in the developing excretory

system. Deve[opment. 109:787-795.

54. Ryan, G., V. Steele-Perkins, J. F. Morris, F. J. Rauscher, and G. R Dressler. 1995.

Repression of Pax-2 by WTl during normal kidney development. Development. 12 13367-

875.

55. Rossel, M., A. Pasini, S. Chappuis, 0. Geneste, L. Fournier, I. Schuffenecker, M.

Takahashi, L. A. van Grunsven, J. L. Urdiales, B. B. Rudkin. G. M. Lenoir, and M.

Billaud. 1997. Distinct biological properties of two RET isoforms activated by MEN

2A/FMTC and MEN 2B mutations. Oncogene. 14:265-275.

56. Bod lo , M. G., G. Pelicci, E. Arighi, L. DeFilippis, A. GECO, I. Bongarzone, M.

G. Rinetti, P. G. Pelicci, and M. A. Pierotti. 1994. The oncogenic versions of the ret and

trk tyrosine kinases bind Shc and Grb2 adaptor proteins. Oncogene. 9: 166 1- 1668.

57. Lorenzo, M., G. Gish, C. Houghton, T. Stonehouse, T. Pawson, B. Ponder, and D.

Smith. 1997. RET alternative splicing influences the interaction of activated RET with the

SH2 and PTB domains of Shc, and the SH2 domain of Grb2. Oncogene. 14:763-77 1.

58. van der Geer, P., and T. Hunter. 1994. Receptor protein-tyrosine kinases and their

signal transduction pathways. Ann Rev Cell Biol. l O : Z 1-337.

59. Klein, R., D. Conway, L. F. Parada, and M. Barbacid. 1990. The trkB tyrosine

kinase gene codes for a second neurogenic receptor that lacks the catalytic kinase domain.

Cell. 6 1 3647-656.

60. Lambde, F., R Klein, and M. Barbacid 1991. trlcC, a new member of the trk family

of tyrosine protein kinases, is a receptor for neurotrophin-3. Cell. 66:967-979.

61. Tsoulfas, P., D. Soppet, E. Escandon, L. Tessarollo, J. Mendoza-Ramirez, A.

Rosenthal, K. Nikolics, and L. Parada. 1993. The rat trkC locus encodes multiple

neurogenic receptors that exhibit differential response to neurotrophin-3 in PC 12 cells.

Neuron. 10:975-990.

62. Valemela, D., P. Maisonpierre, D. Glass, E. Rojas, L. Nunez, Y. Kong, D. Gies, T.

Stin. N. Ip, and G. Yancopoulos. 1993. Alternative forms of rat TrkC with different

functional capabilities. Neuron. 10:963-974.

63. Partanen, J., S. Vainikka, J. Korhonen, E. Armstrong, and K. Alitalo. 1992. Diverse

receptors for fibroblast growth factors. Prog Growth Factor Res. 4:69-83.

64. McDonald, F. J. 1994. Developmentally Regulated Expression of Fibroblast Growth

Factor Receptor Genes and Splice Variants by Murine Embryonic Stem and Embryonal

Carcinoma Cells, Devel Genet. 15: 148- 154.

65. Johnson, D., I. Lu, H. Chen, S. Werner, and L. Williams. 1991. The human

fibroblast growth factor receptor genes: A common structural arcangemem underlies the

mechanism for generating receptor forms that differ in their third immunoglobulin domain.

Mol Cell Biol. 1 1 :4627-4634.

66. Werner, S., D. Duan, d. V. C, K, Peters, D. Johnson, and L. Williams. 1992.

Differential splicing in the extracellular region of £%roblast growth f ~ o r receptor 1

generates receptor variants with different Ligand-binding specificities. Mol Cell Biol. 1282-

88.

67. Avivi, A., A. Yayon, and D. Givol. 1993. A novel form of FGF receptor 3 using an

alternative exon in the immunoglobulin domain III. FEBS Letfers. 330:249-252.

68. Miki. T., D. P. Bottaro, T. P. Fleming, C. L. Smith, W. L. Burgess. A. M.-L. Chan,

and S. A. Aaronson. 1992. Determination of ligand-binding specificity by alternative

splicing: Two distinct growth fictor receptors encoded by a single gene. Proc Nati Acod

Sci USA. 89:246-250.

69. Yayon, A., Y. Zimrner, G. Guo-Hong, A. Avivi, Y. Yarden, and D. Givol. 1992. A

confined region confers ligand binding specificity on fibroblast growth factor receptors:

Implications for the origin of the immunoglobulin fold. EMBO J. 1 1 : 1885- l89O.

70. On-Urtreger, A., M. T. Bedford, T. Burakova, E. Arman, Y. Zimmer, A. Yayon. D.

Givol. and P. Lonai. 1993. Developmental localization of the splicing alternatives of

fibroblast growth faftor receptor-;? (FGFR2). Devel Biol. 158:475-486.

71. Shi, D.-L., C. Lauoay, V. Fromentoux, J.-J. Feige, and I.-C. Boucaut. 1994.

Expression of fibroblast growh factor receptor-2 splice variants is developmentally and

tissue specifically regulated in the amphiiian embryo. Devel B i d 164: 173-182.

72. Jackson-Grusby, L., A. Kuo. and P. Leder. 1992. A variant Limb deformity transcript

expressed in the embryonic mouse limb defiws a novel formin. Genes and Devel. 6:29-37.

73. Chan, D., A. Wynshaw-Boris, and P. Leder. 1995. Formin isofom are differentially

expressed in the mouse embryo and are required for nonnai expression of fgf4 and shh in

the limb bud. Development. 121:3 15 1-3 162.

74. Santoro, M., R. Rosiui, M. Grieco, M. T. Berlingieri, G. L. C. D'Amato, V.

deFranciscis, and A. Fusco. 1990. The RET protooncogene is consistently expressed in

human pheochromocytomas and thyroid medullary carcinomas. Oncogene. 5: 1595-1 598.

75. Tahira, T., Y. Ishizaka, F. Itoh, M. Nakayasu, T. Sugimura and M. Nagao. 1 99 1.

Expression of the ret proto-oncogene in human neurob1astoma cell lines and its increase

during neuronal differentiation induced by retinoic acid. Oncogene. 6:2333-2338.

76. Nagao, M., Y. Isbaka, A. Nakagawara, K Kobo, M. Kuwano, T. Tahim F. Itoh,

I. Ikeda, and T. Sugimara. 1990. Expression of ref proto-oncogene in human

neuroblastomas. Jup J Canc Res. 8 1 :309-3 12.

77. Hofstra, R. M. W., N. C. Cheng, C. Hansen, R. P. Stulp, T. Stelwagen, N. Clausen,

N. Tommerup, H. Caron, A. Westerveld, R. Versteeg, and C. H. C. M. Buys. 1996. No

mutations found by RM mutation scanning in sporadic and hereditary neurobiastoma

Hum Genet. 97:362-364.

78. Ishizaka, Y., H. Shima, T. Sugimura, and M. Nagao. 1992. Detection of

phosphorylated r e t m oncogene product in cytoplasm. Oncogene. 7: 144 1- 1444.

79. Bonganone, I., N. Montini, M. G. Borrello, C. Carcano, G. Ferraresi, E. Arighi. P.

Mondellini, G. Della Porta, and M. A. Pierotti. 1993. Molecular characterization of a

thyroid tumor-specific transforming sequence formed by the fusion of ref tyrosine kinase

and regulatory subunit RIa of cyclic AMP-dependent protein kinase A. Mol Cell Biol.

13:358-366.

80. Santoro, M., N. Dathan, M. Berlingieri, I. Bongarzow, C. Paulin, M. Grieco, M.

Pierotti, G. Vecchio, and A. Fusco. 1994. Molecular characterization of RET/PTC3; a

novel rearranged version of the RET proto-oncogene in a human thyroid papillary

carcinoma. Oncogene. 9:509-5 16.

81. Grieco, M., M. Santoro, M. T. Berlingieri, R. M. Meiillo, R. Donghi, I. Bongmone,

M. A. Pierotti, G. Della Porta, A. Fusco, and G. Vecchio. 1990. PTC is a novel

rearranged form of the ret protooncogene and is frequently detected in vivo in human

thyroid papillary carcinomas. Cell. 60:557-563.

82. Bonganone, I., M. Butti, S. Coronelli, M. BorrelIo, M. Santoro, P. Mondellini, S.

Pilotti, A. Fusco, G. Della Porta, and M. Pierotti. 1994. Frequent activation of ret

protooncogene by fusion with a new activating gene in papillary thyroid carcinomas. Cimc

Res. 542979-298s.

83. Pierotti, M. A., M. Santoro, M. B. Jenkins, G. Sozzi, I. Boagarzone, M. Grieco, M.

Monzini, M. Miono, M. A. Kemnann, A. Fusco, I. D. Hay, G. DeUa Porta, and G .

Vecchio. 1992. Characterization of an inversion on the long arm of chromosome 10

juxtaposing Diosizo and Ret and creating the oncogenic sequence RetlPTC. P roc Natl Acod

Sci USA. 89: 16 16 - 1620.

84. Mulligan, L. M., I. B. I. Kwok, C. S. Healey, M. I. Elsdon, C. Eng, E. Gardner, D.

R. Love, S. E. Mole, I. K. Moore, L. Papi, M. A. Ponder, H. Telenius, A. Tunnacliffe,

and 8. A. J. Ponder. 1993. Germ-line mutations of the RET proto-oncogene in multiple

endocrine neoplasia type 2A. Namre. 363:458-460.

85. Mulligan, L. M., C. Eng, C. S. Healey, D. Clayton, J. B. I. Kwok, E. Gardner, M.

A. Ponder, A. Frilling, C. E. Jackson, H. Lehnert, H. P. H. Neumann, S. N. Thibodeau,

and B. A. J. Ponder. 1994. Specific mutations of the RET protwncogene are related to

disease phenotype in MEN 2A and FMTC. Nature Genet. 6:70-74.

86. Donis-Keller, H., S. Dou, D. Chi, K. M. Carlson, K. Toshima, T. C. Lairmore, J. R.

Howe, I- F. Moley, P. Goodfellow, and S. A. Wells. 1993. Mutations in the RET proto-

oncogene are associated with MEN 2A and FMTC. Hum MoL Genet. 2 :MI-856.

87. Santoro. M., F. Carlomagno, A. Romano, D. P. Bottaro, N. A. Dathaa, M. Grieco,

A. Fusco, G. Vecchio, B. Matoskova, M. H. Kraus, and P. P. Di Fiore. 1995. Activation

of RETas a dominant transforming gene by germline mutations of MEN2A and MEN2B.

Science. 267:38 1-383.

88. Eng, C., and L. M. Mulligan. 1997. Mutations of the RET protooncogene in the

multiple endocrine neoplasia type 2 syndromes, related sporadic nunours and Hirschspruag

disease. H m Mutat. 9:97- 109.

89. Eng, C.. D. Clayton. I. Schuffenecker, G. Lenoir, G. Cote, R. F. Gagel, H.-K Ploos

van Amstel, C. J. M. Lips, I. Nishisho, S.4. Takai, D. J. Marsh, B. G. Robinson, K.

Frank-Raue, F. Raue, F. Xu, W. W. No& C. Romei, F. Pacini, M. Fink, B. Niederie, J .

Zedeaius, M. Nordenskjold, P. Komminoth, G. Hendy, H. Gharib, S. T h l i e a u , A.

Lacroix, A: Frilling, B. A. J. Ponder, and L. M. Mulligan. 1996. The relationship between

specific RET pmto-oncogene mutations and disease phenotype in multiple endocrine

neoplasia type 2: International REl Mutation Consortium. J M . 276: 1575- 1579.

90. Mulligan, L. M., D. I. Marsh, B. G. Robinson, I. Schuffenecker, J. Zedenius, C. J .

M. Lips, R. F. Gagel, S.4. Takai. W. W. NOH. M. Fink, F. Raue, A. Lacroix, S. N.

Thibodeau. A. Frilling, B. A. J. Ponder, C. Eng, and International RET Mutation

Consortium. 1995. Genotype-phenotype cordation in multiple endocrine neoplasia type 2:

report of the International RET Mutation Consortium. J Intern Med. 238:343-346.

91. Bolino, A., I. Schuffenecker, Y. Luo, M. Seri, M. Silengo, T. Tocco, G. Chabrier,

C. Houdent, A. Murat, M. Schlumberger, J. Toumiaire, G. M. Lenoir, and G. Romeo.

1995. RET mutations in exons 13 and 14 of M C patients. Oncogene. 10:2415-2419.

92. Boccia, L. M., J. S. Green, C. Joyce, C. Eng, S. A. M. Taylor, and L. M. Mulligan.

in press. Mutation of RET codon 768 is associated with the M C phenotype. Clin Genet.

in press.

93. Eng, C., D. P. Smith, L. M. Mulligan, C. S. Healey, M. J. Zvelebil, T. J.

Stonehouse, M. A. Ponder, C. E. Jackson, M. Waterfield, and B. A. J. Ponder. 1995. A

novel point mutation in the tyrosine kinase domain of the RET proto-oncogene in sporadic

medullary thyroid carcinoma and in a family with FMTC. Oncogene. lO:509-5 13.

94. Hofsaa, R. M. W., R. M. Landsvater, I. Ceccherini, R. P. Stulp. T. Stelwagen, Y.

Luo. B. Pasini, J. W. M. Hoppener, H. K. Ploos van Amstel, G. Romeo, C. L M. Lips,

and C. & C. M. Buys. 1994. A mutation in the RET proto-oncogene associated with

multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature.

367:375- 376.

95. Eng, C.. D. P. Smith, L. M. Mulligan, M. A Nagai, C. S. Healey, M. A. Ponder, E.

Gardner, G. F. W. Scheumaan, C. E. Jackson, A. Tunnacliffe, and B. A. J. Ponder.

1994. Point mutation within the tyrosine kinase domain of the RET protooncogene in

multiple endocrine neoplasia type 2B and related sporadic ttunours. Hum Mol Genet.

3 :237-24 1 ;

96. Carlson, K, M., S. Dou, D. Chi, N. Scavarda, K. Toshirna, C. E. Jackson, S. A.

Wells, P. J. Goodfellow, and H. Donis-Keller. 1994. Single missense mutation in the

tyrosine kinase catalytic domain of the ret proto-oncogene is associated with multiple

endocrine neoplasia type 2B. Proc Noll Acad Sci USA. 9 L : 1579- 1583.

97. Passarge. E. 1967. The genetics of Hirschsprung's disease. Evidence for

heterogeneous etiology and a study of sixty-three families. AL Engl. I. Med. 276: 138- 143.

98. Chakravarti, A. 1996. Endothelin receptor-mediated signaling in Hirschsprung disease.

Hum Mol Genet. 5303-307.

99. Pasini, B-, M. G. BorreUo, A. Greco, I. Bonganone, Y. Luo, P. Mondek, L.

Alberti, C. Mirancia, E. Arighi. R. Bocciardi, M. Seri, V. Baroae, M. T. Radice. G.

Romeo, and M. A Pierotti. 1995. Loss of function effkct of RET mutations causing

Hirschspmg disease. Nature Genet. 10:35-40.

100. Luo, Y., I. Ceccherini, B. Pasini, I. Matera, M. P. Bicocchi, V. Baroae, R.

Bocciardi, H. KMriiiinen, D. Weber, M. Devoto, and G. Romeo. 1993. Close Linkage

with the RET proto-oncogene and boundaries of deletion mutations in autosomal dominant

Hirschsprung disease. Hum Mol Genet. 2: 18034808.

101. Fewtrell, M. S., P. K. H. Tam, A. H. Thornson, M. Fitchett, J. Currie, S. M.

Huson, and L. M. Mulligan. 1994. Hinchsprung's disease associated with a deletion of

chromosome 10(q 1 1.2): a further Iink with the neurocristopathies? J Med Genet.

3 1:325-327.

102. Edery, P., S. Lyomet, L. M. Mulligan, A. Pelet, E. Dow, L. Abel, S. Holder, C.

Nihoul-F6k&6, B. A. I. Ponder, and A. Munnich. 1994. Mutations of the RET proto-

oncogene in Hirschsprung's disease. Nature. 367:378-380.

103. Romeo. G., P. Ronchetto. Y. Luo, V. Barme, M. Seri. I. Ceccherini, B. Pasini, R.

Bocciardi, M. Lerone, H. Kiiiiriiiinen, and G. Martucciello. 1994. Point mutations

affecting the tyrosine kinase domain of the RET proto-oncogene in Hirrchsprung's disease.

Nature. 367:377- 378.

104. Luo. Y., V. Barone, M. Seri, A. Bolino. R. Bocciardi, I. Ceccherini, B. Pasini, T.

Toco, M. Lerone, S. Cywes, S. Moore, I. M. Vanderwinden. M. J. Abramowicz, U.

Kristofferson, L. T. Lanson, M. Silengo, G. Martuciello, and G. Romeo. 1994.

Heterogeneity and low detection rate of RET mutations in Hirschsprung disease. Eur J

Hum Genet. 2:272-280.

105. Attie, T., A. Pelet, P. Edery, C. Eng, L. M. Mulligan. I. Arniel, L. Boutran, C.

Beldjord, C. Nioul-Feltit& A. Munnich, B. A. J. Ponder, and S. Lyomet 1995.

Diversity of RET proto-oncogene mutations in f d a l and sporadic Hirschsprung disease.

Hum Mol Genet. 4: 138 1- 1386.

L06. Angrist. MOT S. Bok, B. Thiel. E. G. Puffenberger, R. M. Hofstra, C. H. C. M.

Buys, D. T. Cass, and A. Chakravarti. 1995. Mutation analysis of the RET receptor

tyrosine kinase in Hirschsprung disease. Hwn Mol Genet. 4:82 1-830.

107. Yin, L. 1994. Heterogeneity and low detection rate of RET mutations in

Hirschsprung disease. Eur J Hum Genet. 2:272-280.

108. Blaugrund. J. E., M. M. Johns, Y. J. Eby, D. W. Bd, S. B. Baylin. R. H. Hruban,

and D. Sidransky. 1994. RET proto-oncogene mutations in inherited and sporadic

medullary thyroid cancer. Hum Mol Genet. 3: 1895- 1897.

109. Eng, C., L. M. Mulligan, D. P. Smith, C. S. Healey, A. Frilling, F. Raue, H. P. H.

Neumam, R. Pfragner, A. Behmel. M. I. Lonnzo, T. I. Stonehouse. M. A. Ponder. and

B. A. 3. Ponder. 1995. Mutation of the RET protooncogene in sporadic medullary thyroid

carcinoma. Genes Chrom Cane. l2:209-2 12-

1 10. Zedeaius, J., G. WaIlin, B. Hamberger, M. Nordenskjold, G. Weber, and C.

Larsson. 1994. Somatic and MEN 2A de novo mutations identified in the RET proto-

oncogene by screening of sporadic ma. Hum Mol Genet. 3:1259-1262.

1 1 1 . Zedenius, I., C. Larsson, U. Bergholm, I. BovCe, A. Svensson, B. Hallengren, L.

Grimelius, M. Backdahl, G. Weber, and G. Wallin. 1995. Mutations of codon 9 18 in the

RET proto-oncogene correlate to poor prognosis in sporadic medullary thyroid carcinomas.

J Clin Endacrinol Metab. 80:3088-3090.

112. Komminoth, P., E. K. Kunz, X. Matias-Guiu, 0. Hion, G- Christiansen, A.

Colomer, I. Roth, and P. U. Heitz. 1995. Analysis of the RET protooncogene point

mutations distinguishes heritable from nonheritable medullary thyroid carcinomas. Cmcer.

76:479-489.

113. Marsh, D. J., D. L. Learoyd, S. D. Andrew, L. Krishnan, R. Pojer, A.-L.

Richardson, L. Delbridge, C. Eng, and B. G. Robinson. 1996. Somatic mutations in the

RET proto-oncogene in sporadic medullary thyroid carcinoma Clin Endocrinol. 44249-

257.

114. Dou, S., D. Chi, K. M. Carlson, I. A. Moley, S. A. Wells, and H. Donis-Keller.

1994. RET protooncogene mutations associated with sporadic cases of medullary thyroid

carcinoma. F@h International Workshop on Multiple Endocrine Neoplasia:73.

115. Donis-Keller, H. 1995. The RlT proto-oncogene and cancer. J Intent Med. 238:3 19-

325.

116. Romei, C., R. Elisei, A. Pinchera, I. Ceccherhi, E. Molinaro, F. Mancusi, E.

Martino, G. Romeo, and F. Pacini. 1996. Somatic mutations of the RET proto-oncogene

in sporadic medullary thyroid carcinoma are not to exon 16 and are associated

with tumor recurrence. J CZin Endocrinul Metab. 8 1 : 16 19- 1622.

117. Lindor, N. M,, R- Honchel, S. Khosla, and S. N. Thrideau. 1995. Mutations of

the RET protwncogene in sporadic pheochromocytomas. J CIin Endocri1101 Metab.

80:627-629.

118. Kornminoth, P., E. Kunz, 0. Hiort, S. SchrWer, X. Matias-Guiu, G. Christiansen,

J. Roth, and P. U. Heitz. 1994. Detection of RET protooncogene point mutations in

paraffin-embedded pheochmmocytoma specimens by nonradioactive single-strand

conformation polymorphism analysis and direct sequencing. Am J Path. 145922-929.

1 19. Beldjord, C., F. Desclaux-Arramond, M. Raffiin-Sanson, J. C. Corvol, Y. De

Keyser, J. P. Luton, P. F. Plouin, and X. Bertagaa 1995. The RE;T protooncogene in

sporadic pheochromocyt~mas: fkquent Multiple Endocrine Neoplasias 2-Like mutations

and new molecular defects. J Clin Endo Metab- 80:2063-2068.

120. Rychlik, W., and R. E. Rhoads. 1989. A computer program for choosing optimal

oligonucleotides for filter hybridization, sequencing and in vino amplification of DNA.

Nucleic Acids Research. 17 (2 1):8543-855 1.

121. Ivanchuk, S. M., S. M. Myers, C. Eng, and L. M. Mulligan. 1996. De novo

mutation of GDNF, ligand of the REWGDNFR-a receptor complex in Hirschsprung

disease. Hum Mol Genet. 52023-2026.

122. Mulligan, L. M., G. J. Matlashewski, H. J. Scrable, and W. K. Cavenee. 1990.

Mechanisms of p53 loss in human sarcomas. Pruc. Natl. Acad Sci. USA. 87:5863-5867.

123. Koufos, A., M. F. Hansen, B. C. Lampkin, M. L. Workman, N. G. Copeland, N.

A. Jenkins, and W. K. Cavenee. 1984. Loss of alleles at loci on human chromosome 11

during genesis of Wilms' turnour. Nature. 309: 170- 172.

124. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA isolation by

guanidiniium thiocyanate-phenol-chlom fom extraction. Analytical Biochemistry. 162: 156-

159.

L25. Sambrwk, I., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A laboratory

manual. Cold Spring Harbor Laboratories, Cold Spring Harbor, New York

126. Rothenpieler, U. W., and G. R Dressier. 1993. P a - 2 is required for mesenchyme-

to-epithelium conversion during early kidney development Development. 1 l9:7 11-720.

127. Nakamura, T., Y. Ishizaka, M. Nagao, M. Hara, and T. Ishikawa. 1993. Expression

of the ret protwncogene in normal and neoplastic tissues. . Submitted,

128. Leong, S. S., I. S. Horoszewicz, K. Shimaoka, M. Friedman, E. Kawinski, M. I.

Song, R. Zeigel, T. M. Chu, S. B. Baylin, and E. A. Mirand. 1983. In Advances in

Thyroid Neoplasia M. Andreoli, F. Monaco, and I. Robbins, editors. Educational Italia,

Rome. 95-108.

129. Bracey, L. T., and IS. Paigen. 1987. Changes in translational yield regulate tissue-

specific expression of fbglucuronidase. Proc Natl Acad Sci USA. 84:9020-9024.

130. Bevilacqua, A., R. P. Erickson, and V. Hieber. 1988. Antisense RNA inhiiits

endogenous gene expression in mouse preimplantation embryos: lack of double-stranded

RNA "melting" activity. Proc Natl Acad Sci USA. 8583 1-835.

131. Morrison, C.. and F. Gannon. 1994. The impact of the PCR plateau phase on

quantitative PCR Biochim Biophys Acta. 12 l9:493-498.

132. Bunone, G., M. G. Borrello, R. Picetti, I. Bonganone, F. A. Peverali, V. de

Franciscis, G. Della Valle, and M. A. Pierotti. 1995. Induction of RET protwncogene

expression in neuroblastoma cells preceded neuronal differentiation and is not mediated by

protein synthesis. E*p CeN Res. 2 l7:92-99.

133. Asai, NOT T. Iwashita, M. Matsuyama, and M. Takahashi. 1995. Mechanism of

activation of the ref proto-oncogene by multiple endocrine neoplasia 2A mutations. Mol

Cell BioL 15:1613-1619.

134. Califano, D., C. Monaco, G. De Vita, A. DIAlessio, N. A. Dathan, R. Possenti, G.

Vecchio, A, Fusco, MI Santoro, and V. de Franciscis. 1995. Activated RETmC

oncogene elicits immediate early and delayed response genes in PC12 cells. Oncogene.

1 l:lO7-L 12.

135. Santoro. M., R. U Melillo, M. Grieco, M. T. Bcrlingieri, G. Vecchio, and A.

Fusco. 1993. The TRK and RET tyrosine k i n a oncogenes cooperate with ras in the

neoplastic transformation of a rat thyroid epithelial cell line. CeN Growth Di# 4:77-84.

136. Liu X., Q. C. Vega, R A. Decker, A. Pandey, C. A. Worby, and J. E. Dixon.

1996. Oncogenic RE' receptors display different autophosphorylation sites and substrate

binding specificities. J Biol Chem. 27 1 5309-53 12.

137. Songyang, Z., K. L. Carraway, M. J. Eck, S. C. Harrison, R. A. Feldman, M.

Moharnmadi, I. Schlessinger, S. R Hubbard, D. P. Smith. C. Eng, B. A. J. Ponder, B.

I. Mayer, and L. C. Cantley. 1995. Cataytic specificity of protein-tyrosine kinases is

critical for selective signalling. Nature. 373 :S36-539.

138. Iwashita, T., H. Murakami, N. Asai, and M. Takahashi. 1996. Mechanisms of Ret

dysfunction by Hirschsprung mutations affecting its extracellular domain. Hum Mol Genet.

5: 1577- 1580.

139. Momson-Graham, K., and Y. Takahashi. 1993. Steel Factor and c-Kit Receptor.

From Mutants to a Growth Factor System. BioEssays. 1577-83.

140. Huff, V., and G. Saunders. 1993. Wilms tumor genes. Biochim Biophys Acta.

1 155:295-306.

141. Re, G. G., D. I. Hazen-Martin, D. A. Sens, and A. I. Gamin. 1994.

Nephroblastoma (Wilms' Tumor): a Model System of Aberrant Renal Development. Sem

Diag Pathol. 1 1 : 126-1 35.

142. Junien, C., and I. Henry. 1994. Genetics of Wilms' tumor: A blend of aberrant

development and genomic imprinting. Kidney International. 46: 1264- 1279.

143. Tay, J. 1995. Molecular genetics of Wilms' tumor. J Paediatr Child Health. 3 1 :379-

383.

144. Beckwith, I. B. 1986. W i ' tumor and other renal tumors of childhood. In

Pathology of Neoplasia in Children and Adolescents. M. Fiegold, editor. I B Lippincott,

Philadelphia. 3 13-332.

145. Wadey, R., N. Pal, B. Buckle. E. Yeomans, I. Pritchard, and J. Cowell. 1990. Loss

of heterozygosity in W i i ' tumor involves two distinct regions of chromosome 11.

Oncogene. 5 9 1-907.

146. Call, K. M., T. Glaser, C. Ito, A. I. Buckler, I. Pelletier, D. A. Haber, E. A. Rose.

A. Kral, H. Yeger, W. H. Lewis, C. Jones, and D. E. Housman. 1990. Isolation and

characterization of a zinc finger poiypeptide gene at the human chromosome 11 Wilms'

tumor locus. Celt. 60509-520.

147. Gessler, M., A. Poustka, W. K. Cavenee, R L. Neve, S. H. Orkin, and G. A. P.

Bruns. 1990. Homozygous deletion in W i turnours of a zinc-finger gene identified by

chromosome jumping. Nature. 343:774-778.

148. Haber, D. A., and D. E. Houseman. 1992. Role of the WT I gene in Wilms' turnour.

Cancer Surveys. 12: 105- 1 17.

149. Koufos, A., P. Grundy, K. Morgan, K. A. Aleck, T. Hadro, B. C. Lampkin, A.

Kalbakji, and W. K. Cavenee. 1989. Familial Wiedemann-Beckwith syndrome and a

second Wilrns tumor locus both map to 1 lp15.5. Am JHum Genet. 4471 1-719.

150. Ping, A. I., A. E. Reeve, D. I. Law, M. R. Young, M. Boehnke, and A. P.

Feinberg. 1989. Genetic linkage of Beckwith-Wiedemam syndrome to 1 lp 15. Am J Hum

Genet. U:720-723.

151. Wiedemann, H. R. 1983. Turnours and hemihypertrophy associated with

Wiedemann-Beckwith syudrome. Eur J Ped. 141: 129.

152. Hatada, I., H, Ohashi, Y. Fukushima, Y. Kaneko, M. Inoue, Y. Komoto, A. Okada,

S. Ohishi, A. Nabetani, H. Morisaki, M. Nakayama, N. Niikawa, and T. Mukai. 1996.

An imprinted gene ~ 5 7 ~ ~ ~ 2 is mutated in Beckwith-Wiedemann syndrome. Nature Genet.

14: 171-173.

153 Rahman, N., L. Arbour, P. Tonin, J. Renshaw, J. Pelletier, S. Baruchel, K.

Pritchard-Jones, M. R. Stratton, and S. A. Narod. 1996. Evidence for a familial WiIms'

tumour gene ( M I ) on chromosome 17q 12-2 1. Nature Genet. l3:46 1-463.

154. Tonin, P., E. Ehrenborg, G. Lenoir, J. Feunteun. H. Lynch, K. Morgan. H. Zazzi,

A. Vivier, M. Pollak, and H. Huynh. 1993. The human insulin-like growth factor-binding

protein 4 gene maps to chromosome region 17q12q21.1 and is close to the gene for

hereditary breast-ovarian cancer. Genomics. 1 8 :4 14-4 17.

155. Motzer, R. J., N. H. Bander, and D. M. Nanus. 1996. Renal Cell Carcinoma New

Engl J Med* 335:865-875.

156. Maher, E. R 1994. Von Hippel-Lindau Disease. Eur J Canc. 30: 1987- l9W.

157. Latif, F., K Tory, I. Gnarra, M. Yao, F.-M. Duh, M. L. Orcutt, T. Stackhouse, I.

Kuunin, W. Modi, L. Geil, L. Schmidt, F. Zhou, H. Li, M. H. Wei, F. Chen, G. Glenn,

P. Choyke, M. M. Walther, Y. Weng, D.4. R. &an, M. Dean, D. Glavac, F. M.

Richards, P. A. Crossey, M. A. Ferguson-Smith. D. LePaslier, I. Chumakov. D. Coben.

A. C. Cbinault, E. R. Maber, W. M. Linehan, B. Zbar, and M. I. Lerman. 1993.

Identification of the von Hippel-Lindau disease tumor suppressor gene. Science. 160 13 17-

1320.

158. Anglad, P., K. tory, H. Brauch, G. H. Weiss, F. Latif, M. J. Merino, M. I.

Lerman, B. Zbar, and W. M. Lineban. 1991. Molecular analysis of genetic changes in the

origin and development of renal cell carcinoma Canc Res. 5 1 : 107 1 - 107%

159. Tory, K., H. Brauch, M. Linehan, D. Barba, E. Oldfield, M. Filling-Katz, B . Seizinger, Y. Nakamura, R. White, F. Marshall, M. Lerman, and 8. Zbar. 1989. Specific

genetic change in tumors associated with von Hippel-Lindau disease. J Natl Cmtc Inst.

8l:lO97-llOl.

160. Kovacs, G., and S. Frisch. 1989. Clonal chromosome abnormaIities in tumor cells

from patients with sporadic renal cell carcinomas. Canc Res. 49:6S 1-659.

161. Seizinger, B. R., G. A. Rouieau, L. J. Ozelius, A. H. Lane, G. E. Farmer, I. M.

Lamiel, J. Haines, J. W. Ywns, D. Collins, and J. Majoor-Krakauer. 1988. Von Hippel-

Lindau disease maps to the region of chromosome 3 associated with renal cell carcinoma.

Nature. 3 32:268-269.

162. Crossey, P. A., E. R. Maher. M. H. Jones, F. M. Richards, F. Latif, M. E. Phipps,

M. Lush, K Foster, K. Tory, I. S. Green, B. Oostra, J. R. W. Yates, W. M. Linehan,

N. A. Affara, M. Lerman, B. Zbar, Y. Nakamura, and M. A. Ferguson-Smith. 1993.

Genetic Linkage between Von Hippel-Lindau disease and three microsatellite

polymorphism refmes the localization of the VHL locus. Hum Mol Genet. 2 (3):279-282.

163. Mocha, R., J. Ishikawa, M. Tsutsumi, K. Hikiji, Y. Tsukada, S. Kamidono, S.

Maeda, and Y. Nakamura. 1991. Melotype of renal cell carcinoma Cmu: Res. 51:820-

823.

164. Morita, R., S. Saito, J. Ishikawa, 0. Ogawa, 0. Yoshida, K. Yamakawa, and Y.

Nakamura. 1991. Common regions of deletion on chromosomes Sq, 6q and 10q in renal

cell carcinoma. Canc Res. 5 158 17-5820.

165. Salomon, R., T. Attie, A- Pelet, C. Bidaud. C. Eng, J. Amiel, S. Samacki, 0.

Godet, C. Ricour, C. Nioul-Fekktti, A. Munnich, and S. Lyomet. 1996. Germline

mutations of the RET ligand, GDNF, are not sufficient to cause Hirschsprung disease.

Nature Genet. 14:345-347.

166. Donovan, M. J., B. Hempstead, L. I. Huber, D. Kaplan, P. Tsoulfas, M. Chao, L.

Parada, and D. Schofield. 1994. Identification of the newotrophin receptors p75 and trk in

a series of Wilms' tumors. Am J Pathol. 145:792-80 1.

167. Durbeej, M., S. Soderstrom, T. Ebendal, T. Birchmeier, and P. Ekblom. 1993.

Differential expression of neurotmphin receptors during renal development. Development.

119:977-989.

168. Tessarollo, L., P. Tsoulfas, D. Martin-Zanca, D. Gilbert, N. Jenkins, N. Copeland,

and L. Paracia, 1993. ukC, a receptor for neurotrophin-3, is widely expressed in the

developing nervous system and in non-neural tissues. Development. 1 18:463-475.

169. Huff, V., H. Miwa, D. A. Haber, K. M. call, D. Housman, L. C. Strong, and G. F.

Saunders. 199 1. Evidence for WT1 as Wlms tumor (WT) gene: -genic germinal

deletion in bilaterd WT. Am J Hum Genet. 48:997-1003.

170. Maw, M., P. E. Grundy, L. J. Millow, M. R. Eccles, R. S. Dunn, P. I. smith, A. P.

Feinberg, D. J. Law, M. C. Paterson, P. E. Telzerow, D. F. Callen, A. D. Thompson, R.

I. Richards, and A. E. Reeve. 1992. A third Wilms' tumor locus on chromosome 16q.

Cane Res:3094-3098.

171. Eccles, M. R., L. J. Wallis, A. E. Fidler, N. K. Spurr, P. I. Goodfellow, and A. E.

Reeve. 1992. Expression of the PAX2 gene in human fetal kidney and Wilms tumor. Ceil

Growth Diff. 3~279-289.

172. Poleev, A., H. Fickenscher, S. Mundlos, A. Winterpacht, B. Zabel, A. Fidler, P.

Gruss, and D. Plachov. 1992. PAX8, a human paired box gene: isolation and expression

in developing thyroid kidney and Wi' tumor. Development. 1 l6:6 1 1-623.

173. Eccles, M. R., K. Yun, A. E. Reeve, and A. E. Fidler. 1995. Comparative in siru

hybridization analysis of PAX2, PAX8 and WTI gene transcription in human fetal kidney

and Wilms' tumors. Am J Puthol. 146:40-45.

174. Weiss, L. M., A. B. Gelb, and L. J. Medeiros. 1995. Adult renal epithelial

neoplasms. Am J Clin Pathol. 103:624-635.

175. Gnarra, I. R., K. Tory, Y. Weng, L. Schmidt, M. & Wei, H. Li, F. Latif, S. Liu,

F. Chen, F.-M. Duh, I. Lubensky, D. R. Duao, C. Florence, R. Pouatti, M. M. Walther,

N. H. Bander, H. D. Grossman, H. Brauch, S. Pomer, J. D. Brooks, W. B. Isaacs, M. I.

Lerman, B. Zbar, and W. M. Linehan. 1994. Mutations of the VHL tumour suppressor

gene in renal carcinoma. Nature Genet. 7:85-90.

L76. Foster, K., A. Prowse, A. van den Berg, S. Fleming, M. M. Hulsbeek, P. A.

Crossley, F. M. Richards, P. Cairns, N. A. Affrara and M. A. Ferguson-Smith. 1994.

Somatic mutations of the von Hippel-Lindau disease tumour suppressor gene in non-

farnilid clear cell rend carcinoma, Hum Mol Genet. 3:2l69-2 173.

177. Heman, J. G., F. Latif. Y. Weng, M. I. Lennan, B. Bar, S. Liu. D. Samid, D . 4 .

R. Duan, I. R. Gnarra, W. M. Linehan, and S. B. Baylin. 1994. Silencing of the VHL

tumor-suppressor gene by DNA methylation in rend carcinoma Pruc Nut1 Acad Sci USA.

9 139700-9704.

178. Kessier. P. M., S. P. Vasavada, R. R. Rackley, T- Stackhouse, F.-M. Dub, F. Latif,

M. I. Lerman, B. Zbar, and 8. R. G. Wfiams. 1995. Expression of the Von Hippel-

Lindau Tumor Suppressor Gene, VHL, in Human Fetal Kidney and During Mouse

Embryogenesis. Mol Med. 1 :457-466.

179. Palmer, M., A. Dryden, I. Hughes, and J. Collinge. 1991. Homozygous prion

protein genotype predisposes to sporadic Creuafeldt-Jakob disease. Nature. 3S2:3480-34 1.

180. Brown, P.. L. G. Goldfarb, C. I. Gibbs, and D. C. Gajdusek. 1991. The phenotypic

expression of different mutations in transmissible familial Creutzfeldt-Jakob disease. Eur J

Epidemiol. 7:469-476.

181. Brown, P., L. Cemenkova, L. G. G~ldfarb~ W. R. McCombie, R. Rubinstein, R.

G. Will, and M. Pocchiari. 1994. Iatrogenic Creuafeldt-Jakob disease: an example of the

interplay between ancient genes and modem medicine. Neurology. W 2 9 1-293.

182. Windl, 0.. M. Dempster, I. P. Estibeiro, R Lathe, R. de Silva, T. Esmonde, R.

Will. A. Springbett, T. A. Campbell, K C. L. Sidle. M. S. Palmer, and J. Collinge.

1996. Genetic basis of Creutzfddt-Jakob disease in the United Kingdom: a systematic

analysis of predisposing mutation and aUelic variation in the PRNP gene. Him Genet.

98:259-264.

183. Medori. EL, H. J. Tritschler, A. LeBIanc, F. Villare, V. Manetto, H. Y. Chen, R.

Xue, S. Leal, P. Montagna, and P. Cortelli. 1992. F a d familial insomnia, a prion disease

with a mutation at codon 178 of the prion protein gene. New Engl J Med. 326:444-449.

184. Medori, R., P. Montagna, H. I. Tritschler, A. LeBlanc, P. Cortelli, P. Tinuper, E.

Lugaresi, and P. Gambetti. 1992. Fatal familial insomnia: a second kindred with mutation

of prion protein gene at codon 178. Neurology. 42:669-670.

185. Petersen, R. B., M. Tabaton, L. Berg, S. Schrank, R. M. Torack, S. Leal, I. Mien,

C. Vital, B: Deleplanque, and W. W. Pendiebury. 1992. Analysis of the prion protein gene

in thdatuic dementia. Neurology. 42: 1859- 1863.

186. Goldfarb, L. G.. M. Haltia, P. Brown, A. Nieto, I. Kovanen, W. R. McCombie, S.

Trapp, and D. C. Gajdusek. 1991. New mutation in scrapie amyloid precursor gene (at

codon 178) in F i s h Creutzfeldt-lakob kindred. Luncet. 337:425.

187. Petersen, R. B., L. G. Goldfarb, M. Tabaton, P. Brown, L. Monari, P. Cortelli, P .

Montagna, L. Autilio-Gambetti, D. C. Gajdusek, E. Lugaresi, and P. Gambetti. 1994. A

Novel Mechanism of Phenotypic Heterogeneity Demonstrated by the Effect of a

Polymorphism on a Pathogenic Mutation in the PRNP (Rion Protein Gene). Mol

Neurobiol. 8:99- 103.

188. Williams, D. E., J. Eisenmann, A. Baird, C. Rauch, K. Van Ness, C. J. March, L .

S. Park, U. Martin, D. Y. M~~hizuki , H. S. Bosweil, G. S. Burgess, D. Cosman, and S .

D. Lyman. 1990. Identification of a ligand for the c-kit proto-oncogene. Cell. 63: 167474.

189. Chabot, B., D. A. Stephenson, V. M. Chapman. P. Besmer, and A. Bernstein.

1988. The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps

to the mouse W locus. Nature. 335%-89.

190. Zsebo, K. M., D. A. Williams, E. N. Geissler, V. C. Broudy, F. H. Martin, H. L.

Atkins, E2.X Hsu, N. C. Birkett, K H. Okino, D. C. Murdock, F. W. Jacobsen, K. E.

Langley, K. A. Smith, T. Takeishi, B. M. Cattanach, S. I. Galli, and S. V. Suggs. 1990.

Stem cell fa~tor is encoded at the Sl locus of the mouse and is the Ligand for the c-kit

tyrosiw kinase receptor. Cell. 63 :2 13-224.

19 1. Geibel, L. B ., and R. A. Spritz. 199 1. Mutation of the KIT (mastktem cell growth

factor receptor) protwncogene in human piebaldism, Proc Natl Acod Sci USA. 88:8696-

8699.

192. FIeischma~l~l, R. A., D. L. Saltman, V. Stastny, and S. Zneimer. 199 1. Deletion of

the c-kit proto-oncogene in the human developmental defect piebald trait. Pruc Nati A c d

Sci USA. 88: 10885-10889.

193. Angrist, M., S. Bolk, M. Halushka, P. A. Lapchak, and A. Chakravarti. 1996.

Germline mutations in glid cell line-derived neurotrophic factor (GDNF) and RET in a

Hirschspmg disease patient. Nature Genet. l M 4 1-343.

194. Edery, P., T. Attie, J. Amiel, A. Pelet, C. Eng, R. M. W. Hofstra, H. MarteIli, C.

Bidaud, A. Munnich, and S. Lyonnet. 1996. Mutation of the endothelin-3 gene in the

Waardea burg-Hirschsprung (Shah- Wwden burg syndrome). Nature Genet. L 2:442-444.

195. Hofstra, R. M. W., J. Osinga, G. Tan-Sindhunata, Y. WU, E.4. Karnsteeg, R. P.

Stulp, C. van Ravenswaaij-Arts, D. Majwr-Krakauer, M. Angrist. A. Chakravarti. C.

Meijers, and C. H. C. M. Buys. 1996. A homozygous mutation in the endothelin-3 gene

associated with a combined Waardenburg type 2 and Hirschsprung phenotype (Shah-

Waardenburg syndrome). Nature Genet. 12:445-447.

196. Mulligan, L. M.. E. GXdne~, B. A. Smith, C. G. P. Mathew, and B. A. J. Ponder.

1993. Genetic events in tumour initiation and progression in multiple endocrine aeopiasia

type 2. Genes Chrum Cane. 6: 166-177.

L97. Romano. A*, W. T. Wong, M. Santoro, P. J. Winh, S. S. Thorgeirsson, and P. P.

DiFiore. 1994. The high transforming potency of erbB-2 and ret is associated with

phosphorylation of paxiilin and a 23 kDa protein. Oncogene. 9:2923-2933.

198. Borrello, M. G., D. P. Smith, B. Pasini. I. Bonganone. A. Greco, M. I. Lorenzo,

E. Arighi, C. Mirancia, C. Eng, L. Alberti, R. Bocciardi, P. Mondellini, L. Scopsi, G.

Romeo, B. A. i. Ponder, and M. A. Pierotti. 1995. RET activation by germline MEN2A

and MEN28 mutations. Oncogene. 11:2419-2427.

199. Pawson, T. 1995. Protein modules and signalling networks. Nature. 373573-580.

200. Pandey, A., D. F. Lazar, A. R. Sdtiel, and V. M. Dixit. 1994. Activation of the Eck

receptor protein tyrosine kinase stimulates phosphatidylinositol 3-kinase activity. J Bioi

Chem. 269330 154-30157.

201. Mama, H., and A. Burgess. 1994. Regulation of the Ras signalling network.

BiuEssuys. 16:489-496.

202. Vega, Q., C. Worby, M. Lechner, J. Dixon, and G. Dressier. 1996. Glial cell Iine-

derived neurotrophic factor activates the receptor tyrosine base REX and promotes kidney

morphogenesis. Proc Natl Acad Sci USA. 93: 10657- LO66 1.

203. Puffenberger, E., E. buffman, S. Bok, T. Matise, S. Washington, M. Angrist, I.

Weissenbach, K. Garver, M. Mascari, R. Ladda, S. Slaugenhaupt and A. Chakravarti.

1994. Identity-by-descent and association mapping of a recessive gene for Hirschsprung

disease on chromosome 13q22. Hum Mol Genet. 3: 12 17- 1225.

204. Puffenberger, E. G., K. Hosoda, S. S. Washington, K. Nakao, D. de Wit, M.

Yanagisawa, and A. Chakravarti. 1995. A missense mutation of the endothelin B receptor

gene in multigenic Hirschsprung's disease. Cell. 79: 1257- 1266.

Appendix 1

Solutions

LOOX Denhardt's

2SX SSC

LOX TBE

LOX TPE

sequencing gel mix

2% bovine serum albumin 2% (wh) Ficoll400 2% (w/v) po~yvinylpyro~do~

3.7M sodium chloride 0.4M uisodium citrate-2H20 *pH 7.4

0.9M Tris 0.9M boric acid 2- EDTA

0.9M Tris 0.0 1% orthophosphoric acid 2OmMEDTA

7M urea 5% acrylamide 0.25% (w/v) NN methyl bis acrylamide 1X TBE

denaturing GLB 95% deionized formamide 20mM EDTA 0.05% bromophenol blue 0.05% xylene cyan01

lOX Sigma GLB 40% sucrose O.1M EDTA 0.5% sodium dodecyl sulphate 0.0556 bromophenol blue

Rehybridization Buffed 6X SSC Hybridization Buffer 0.1% SDS

5X Denhardt's solution 100pg/mL sheared, denatured heterologous DNA

10X PNK Buffer ~OIIIM Tris-HC1 (pH 7.6) MgQ

5mM IYIT lOOmM spermidhe-HCl L O W EDTA (pH 8 -0)

10X RT Buffer

10X PCR Buffer lOmM Tris-HCl (pH 8.3) SOmM KCI 0.0 1 % gelatin 0.75- 1 .75mM MgClz