University of Groningen

Functional and genetic studies in MEN2 and Hirschsprung disease Moreira Alves, Maria Manuela

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2010

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Moreira Alves, M. M. (2010). Functional and genetic studies in MEN2 and Hirschsprung disease. Moreira Alves, Maria Manuela.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 27-09-2021 FUNCTIONAL AND GENETIC STUDIES IN MEN2 AND HIRSCHSPRUNG DISEASE MARIA MANUELA MOREIRA ALVES

SCG10 KIF1Ba KIF58

EDNRB ------PHOX2B RET NRG1 GDNF GFRa SOX10 KBP EDN3 Functional and Genetic Studies in MEN2 and Hirschsprung disease

Maria Manuela Moreira Alves Maria Manuela Moreira Alves

Functional and Genetic Studies in MEN2 and Hirschsprung disease.

Thesis University of Groningen with summary in Dutch.

The studies described in this thesis were financially supported by the Groningen University Institute for Drug Exploration (GUIDE).

Printing of this thesis was financially supported by: Rijksuniversiteit Groningen, University Medical Center Groningen; Groningen University Institute for Drug Exploration (GUIDE).

Cover Designed by Optima Grafische Communicatie. Printed by Wehrmann Print Service.

© 2010 M.M. Moreira Alves. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means without permission of the author.

ISBN: 978-90-367-4603-8 Stellingen behorende bij het proefschrift Functional and Genetic Studies in MEN2 and Hirschsprung disease

1. KBP connects microtubule dynamics to HSCR development (chapter 2, this thesis). 2. The study of complex disorders will benefit from systems biology. 3. SCGlO may link KBP to the RET and EDNRB signalling pathways (chapter 2, this thesis). 4. The presence of common and rare mutations can explain the genetic basis of complex disorders. 5. Different RET mutations trigger specific downstream signalling pathways (chapter 5, this thesis). 6. Patients with Medullary and Papillary Thyroid Carcinomas will benefit from mutation specific therapies (chapter 6, this thesis). 7. The fact that one HSCR associated RET variant is commonly found worldwide suggests a protective effect of this mutation for the carriers. 8. "The important thing in science is not so much to obtain new facts as to discover new ways of thinking about them."

William Bragg

9. "Science is wonderfully equipped to answer the question "How?" but it gets terribly confused when you ask the question "Why?"

Erwin Chargaff

11. "Try and fail, but don't fail to try."

Stephen Kaggwa

Maria Alves, 10 November 2010

RIJKSUNIVERSITEIT GRONINGEN

Functional and Genetic Studies in MEN2 and Hirschsprung disease

Proefschrift

ter verkrijging van het doctoraat in de Medische Wetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op woensdag 10 november 2010 om 14.45 uur

door

Maria Manuela Moreira Alves geboren op 10 augustus 1981 te Santo Tirso, Portugal Promotor: Prof. dr. R.M.W. Hofstra

Copromotor: Dr. B.J.L. Eggen

Beoordelingscommissie: Prof. dr. I. Ceccherini Prof. dr. T.P. Links Prof. dr. M.P. Peppelenbosch "Pedras no caminho, guardo-as todas. Um dia vou construir um castelo ..."

FernandoPessoa Paranimfen: Ana Gracinda Carvalho Jan Osinga To Pai, Vo and Vo

Contents

Chapter 1- General introduction and aims of the thesis 11

PartI - Hirschsprung disease

Chapter 2 - KBP interacts with SCGlO, linking Goldberg-Shprintzen 37 syndrome to microtubule dynamics and neuronal differentiation

Chapter 3 - Mutations in SCG10 are not involved in Hirschsprung disease 61

Chapter 4 - Hirschsprung disease as a model for complex disorders 69

PartII - MEN2

Chapter 5 - Genome wide kinome profiling of the RET oncogene 89 identifies MEN2A -, MEN2B - and FMTC - specific signalling pathways

Chapter 6- The effects of four different tyrosine kinase inhibitors 113 on Medullary and Papillary thyroid cancer cells

Chapter 7 - Summarizing discussion and future perspectives 131

Abbreviations 143 References 147 Summary 170 Nederlandse samenvatting 173 Resumo 177 Acknowledgements 181

Chapter 1

General Introduction and Aims of the thesis Chapter 1

General introduction The neural crest is a transient and multipotent embryonic structure in vertebrates that gives rise to most of the peripheral nervous system and to several non-neural cell types, including smooth muscle cells of the cardiovascular system, pigment cells in the skin, craniofacial bones, cartilage and endocrine, paraendocrine, and connective tissues (Le Douarin, 1999). Although neural crest cells undergo developmental restrictions with time, at least some of these cells have the capacity to self-renew and display a developmental potential nearly only topped by embryonic stem cells. Intriguingly, such neural crest-derived stem cells {NCSCs) are not only present in the embryonic neural crest, but also in various neural crest-derived tissues in the foetal and even adult organism (Shakhova and Sommer, 2010). Several neural crest-related disorders have been reported. Due to their nature, they are called neurocristopathies and they can be caused by genetic abnormalities in specific genes. Hirschsprung disease (HSCR) and Multiple endocrine neoplasia type 2 (MEN2) are two of these neurocristopathies. HSCR is a developmental disorder characterized by the absence of enteric ganglia in the intestinal tract (Parisi and Kapur, 2000), whereas MEN2 is a dominantly inherited cancer syndrome affecting neuroendocrine organs (Hansford and Mulligan, 2000). HSCR and MEN2 have distinct properties and in each case different tissues are affected, however they can both be caused by genetic abnormalities in a specific proto-oncogene: the REarranged during Transfection (RETI gene. In this thesis, we focus on HSCR and MEN2 and we combine genetic and functional approaches in an attempt to understand the protein networks involved in the development of these disorders.

RET gene and encoded protein RET was first described in 1985 during transfection of NIH-3T3 cells with human lymphoma DNA {Takahashi et al., 1985). In the transformed cells, two unlinked human

12 Introduction

DNA segments were recombined resulting in a chimaeric gene. As this chimaeric gene was not present in the transfected DNA, the recombination step must have taken place during the transfection experiment and it was therefore called REarranged during Transfection (RET) (Takahashi et al., 1985). RET is located on the long arm of chromosome 10 (10q11.2), has a length of 55 Kb and contains 21 exons (Pasini et al., 1995). It encodes a receptor tyrosine kinase, which is expressed in neural crest-derived cell lineages and is essential for the early development Cadherin Domain of the kidney, spermatogenesis, and Cysteine Rich Domain the sympathetic, parasympathetic and enteric nervous systems (Pachnis et al., 1993; Schuchardt et al., 1994). By alternative splicing, Tyrosine kinase Domain 1 three different RET isoforms are Tyrosine kinase formed and they can be Domain 2 distinguished from each other by RET9 the length of their C-terminus: 9 RET43 RET 51 (RET9), 43 (RET43) and 51 (RET51) Figure 1 - Schematic representation of the RET receptor. amino acids (Fig. 1). This results in RET proteins of 1072, 1106 and 1114 amino acids in length, respectively. Several observations suggested that different isoforms have distinct tissue-specific effects during embryogenesis, with RET 51 and RET 9 being the most prevalent isoforms in vivo (Myers et al., 1995; de Graaff et al., 2001). Like other receptor tyrosine kinases, RET contains an extracellular region, a transmembrane region and an intracellular tyrosine kinase region (Fig. 1). The extracellular region is composed of a cadherin-like domain that induces and stabilizes conformational changes needed for interaction with ligands and co-receptors

13 Chapter 1

(Schneider, 1992), and by a cysteine-rich domain located in proximity to the plasma membrane. This domain plays a critical role in the establishment of intramolecular disulphide bonds which are important for the tertiary structure of RET (Hansford and Mulligan, 2000). Moreover, the extracellular region also contains several glycosylation sites necessary for membrane anchoring (Takahashi et al., 1991). Based on the glycosylation state, RET localization alternates between the cell membrane, cytoplasm and endoplasmatic reticulum. When RET is totally glycosylated, it is in its mature form and is present in the cell membrane. When RET is not glycosylated and therefore in its immature form, it is present only in the endoplasmatic reticulum and the cytoplasm (Takahashi et al., 1993). The intracellular region contains two tyrosine kinase subdomains that are involved in numerous intracellular signal transduction pathways initiated by phosphorylation on one of the tyrosine residues (Fig. 1).

RETactivation and downstream signalling pathways RET plays a central role in several intracellular signalling cascades that regulate cell survival, differentiation, proliferation, migration, and chemotaxis (de Groot et al., 2006). These cascades are initiated upon RET activation, a process that is unique among receptor tyrosine kinases as it requires a multi-protein complex instead of common direct receptor/ligand activation (Plaza-Menacho et al., 2006). The RET activation complex includes a soluble ligand of the Glial cell-line-derived neurotrophic factor (GDNF) family, a glycosyl-phosphatidylinositol membrane anchored co-receptor of the GDNF family receptor (GFRa 1-4), and RET itself {Manie at al., 2001). The activation process starts when GDNF or one of its homologues, neurturin, artemin or persephin, binds to GFRa (Jing et al., 1996; Airaksinen et al., 1999). After the ligand and co-receptor have bound to each other, this complex binds to RET, which by then has been recruited to the lipid rafts. By association of two RET molecules to the ligand-co-receptor complex, RET becomes activated (Manie et al., 2001; de Groot et al., 2006; Plaza-Menacho et al., 2006).

14 Introduction

Once activated, specific tyrosine residues of the RET tyrosine kinase domain become phosphorylated and initiate specific downstream signalling pathways. The phosphorylated residues act as docking sites for adaptor proteins that link receptor activation to the main signal transduction pathways. Several phosphorylation sites have been identified in RET and these include: tyrosine 687 (Y687), serine 696 (S696), Y752, Y791, Y806, Y809, Y826, Y864, Y900, Y905, Y928, Y952, Y981, Y1015, Y1029, Y1062, Y1090 and Y196. The last two sites are only present in the long RET isoform, RET51 (Manie et al., 2001; Liu et al., 1996; Santoro et al., 2004; lchihara et al., 2004). Phosphorylation of S696 is necessary for activating the RAC1-guanine nucleotide exchange factor (GEF) and for lamellipodia formation. Interestingly, activation of Y687 induces the opposite effect on lamellipodia formation. These cytoskeletal rearrangements are regulated via cyclic AMP/protein kinase A (PKA)­ dependent signalling (Fukuda et al., 2002). RET Y752 and Y928 activate a latent transcription factor involved in several types of cancer when aberrantly activated, called signal transducer and activator of transcription 3 {STAT3) {Schuringa et al., 2001; Fukuda et al., 2002). Phosphorylation of Y905 stabilizes the active conformation of the kinase and thus facilitates auto-phosphorylation of tyrosine residues located on the C-terminus (Iwashita et al., 1996). Furthermore, it interacts with the growth factor receptor­ bound (GRB) docking proteins 7 /10. Tyrosines Y806, Y809 and Y900 were described as complementing the function of Y905 and thus also interact with GRB7 /10 {Kawamoto et al., 2004). Phosphotyrosine residue Y981 was described as the major binding site for the chicken Rous sarcoma virus oncogene (Src), and is therefore the responsible residue for Src activation. Src is known to play a role in neuronal survival, but has also been described as part of the oncogenic RET signalling by mediating signalling between RET and focal adhesion kinase (FAK). FAK is an important regulator of tumour formation and cell migration and it was reported to be necessary for the invasion and metastatic behaviour of cancer cells (Encinas et al., 2004; Panta et al., 2004; Mclean et al., 2005).

15 Chapter 1

Y1015 is the binding site for phospholipase C-y (PLC-y), which triggers protein kinase C (PKC) activation. PKC is know to exert a double effecton RET: it can stimulate RET phosphorylation, but it can also downregulate RET and its downstream signalling pathways exerting a negative feedback loop that controls RET activity (Borrello et al., + 1996). Furthermore, PLC-y is involved in the release of Ca2 via generation of inositol­ triphosphate (IP3) (Andreozzi et al., 2003). Ca2+ ions are known to play a decisive role in RET transport to the cell membrane and in facilitating binding of RET ligands for receptor activation, however the exact role of Ca2+ influx triggered by RET is not yet clear (van Weering et al., 1998; Nozaki et al., 1998). Y1062 is known to be the docking site of several adaptor proteins namely, Src homolog collagen (She), Src homology 2 domain-containing transforming protein C (ShcC), insulin receptor substrate (IRS) 1/2, fibroblast growth factor receptor substrate 2 (FRS2), downstream of kinase (DOK) 1/4/5, enigma extracellular signal regulated kinase 5 (ERKS), mitogen activated protein kinase (MAPK), phosphoinositide-dependent kinase 1 (PDK1), cyclin-dependent kinase 5 (CDKS), SH3 (a protein-protein interaction domain), multiple ankyrin repeat domains 3 (Shank3) and PKC isoforms. In general, the signalling pathways activated by phosphorylation of Y1062 include: the phosphatidylinositol 3-kinase/protein kinase B (Pl3/Akt) pathway initiated via recruitment of She, ShcC, IRSl/2 or Shank3; the MAPK and the rat sarcoma oncogene (RAS)/ERK pathways triggered by binding of Shank3, DOK 4/5 and She or FRS2 to the Grb2/ son of sevenless multi-protein scaffold (SOS) complex; and the c-Jun N-terminal kinase (JNK) pathway initiated by binding of DOK1 (Segouffin­ Cariou and Billaud, 2000; Pelicci et al., 2002; Mellilo et al., 2001; Schuetz et al., 2004; Kurokawa et al., 2001; Mellilo et al., 2001; Grimm et al., 2001; Xia and Karin, 2004; Kennedy and Davis, 2003). The Pl3K/Akt pathway is involved in cell survival, enhanced cell-cycle progression and RET-mediated transformation (Segouffin-Cariou and Billaud, 2000; Murakami et al., 1999). The RAS/ERK and MAPK pathways are responsible for cellular differentiation and proliferation (Watanabe et al., 2002), and

16 Introduction the JNK pathway plays a role in cell proliferation, cell survival, cell death, DNA repair and metabolism (Xia and Karin, 2004; Kennedy and Davis, 2003). Y1096, only present in the RETSl isoform, serves as a docking site for GRB2 which activates the RAS/ERk and Pl3K pathways (Lorenzo et al., 1997). Finally, the activated downstream signalling pathways initiated by phosphorylation of the residues Y791, Y826, Y864, Y952, Y1029 and Y1091 still need to be delineated (Kawamoto et al., 2004; Liu et a/.,1996). An overview of all the proteins involved in activation of the different signalling pathways initiated via RET phosphorylation is shown in figure 2.

Lipid rafts

Motllity PL::1 ----- OPJ ___ c,:.::::::::. r "l--,:='"''"�-1-I �...,._____ 1--....'.'.c�ProlifeAitlon,'.::'�=•:._' r-- •Jl":IC--.l�o( Celtsuf'Yfval nnl(1-j l'IA� -- r.•K ProlifeAitlon, Dilferenli.ilion

on••• ·····-___1=: ..... J Prolift!Aition, t__ tVrt.P< -t-----t DIHerenll.Ulan

l:.nam�-+---1�--_J-�,:-•.,:-.,.,....=,"""1 TAinS9()rt

�ts.:e---- Glllt,'?: • -� . 1 Motility ,��rt:�� ORHs---r Prolifu�tion, ,.,r.,.o r,r.,.-;t Dilferenlf.itlon, CDI<: Cellsurvival Ao:Cou,ifc,rm•

Figure 2 - Overview of signalling pathways triggered by phosphorylation of specific RET tyrosin kinase residues (Adapted from de Groot et al., 2006).

17 Chapter 1

RETassociated syndromes Several disease-causing germline mutations have been found in RET. Roughly these mutations can be classified into two types: inactivating and activating mutations/rearrangements. Inactivating mutations, result in a loss of RET function and have been found to be associated with a developmental disorder called Hirschsprung disease (Parisi and Kapur, 2000). The activating mutations/rearrangements on the other hand, lead to a constitutively active RET receptor and, as a consequence, are involved in the development of the cancer syndrome Multiple Endocrine Neoplasia type 2 and Papillary Thyroid Cancer (Hansford and Mullingan, 2000; Takahashi et al., 1985). These disorders will be discussed in more detail in the next section.

Hirschsprung disease (HSCR) HSCR is an abnormality of the enteric nervous system characterized by the absence of enteric neurons along a variable length of the gut. It is a developmental disorder that is caused by a failure in the migration of neural crest stem cells (NCSCs) into the intestinal tract or due to a failure in the survival, proliferation, or correct development of NCSCs once they reach the gut (Heanue and Pachnis 2007). In either case, an aganglionic segment occurs in which peristalsis is absent (Fig. 3).

Normal colon HSCR affected colon

Figure 3 - Hirschsprung disease (Adapted from 2003 Nucleus Communications).

18 Introduction

Based on the length of the aganglionic segment, three forms of HSCR can be distinguished. In 80% of the cases the aganglionic segment does not extend beyond the upper sigmoid and this is called short-segment HSCR (SS-HSCR). In 15% of the cases aganglionosis extends proximal to the sigmoid and is called long-segment HSCR (LS-HSCR). When the absence of NCSCs affects the whole bowel, total intestinal aganglionosis (TIA) is observed. This is a rare form of HSCR accounting for only 5% of the cases (Passarge, 1967; Holschneider, 1982; Garver et al., 1985; Chakravarti and Lyonett, 2001) (Fig. 4).

Normal

Short segment HSCR

Long segment HSCR

Total intestinal aganglionosis .- -- ) -:.:..:.-rr7 / Figure 4 - Representation of the three types of HSCR classified based on the length of aganglionic segment (Adapted from Heanue and Pachnis, 2007).

HSCR is a relatively common disorder with an occurrence of approximately 1:5000 live births and is thereby the main cause of intestinal obstruction (Passarge, 1967; Goldberg, 1984; Spouge and Baird, 1985; Garver et al., 1985; Russell et al., 1994; Ikeda and Goto, 1984). However, there is significant racial variation in the incidence of

19 Chapter 1 this disease. Among all ethnic groups, HSCR is most often found in Asians (2.8 per 10,000 life births), whereas Hispanics show the lowest incidence of this disease (1.0 in 10,000 life births) (Torfs, 1998). HSCR also shows a low, sex-dependent penetrance and is seen more frequently in males than in females (4:1) (Badner et al., 1990). Interestingly, the male:female ratio is significantly higher for S-HSCR (4.2-4.4) than for L-HSCR (1.2-1.9) (Torf, 1998; Badner et al., 1990; Amiel and Lyonnet, 2001). HSCR can be a Mendelian (heterogeneous) disease but, in most cases, it is a complex non-Mendelian disorder. For both situations several genes have been identified but RET is the most important one. Several RET mutations have been identified in HSCR patients including large deletions, micro-deletions and insertions, nonsense, missense and splicing mutations (Edery et al., 1994; Romeo et al. 1994). These findings, together with mice studies showing that inhibition of RET expression resulted in a HSCR phenotype, confirmed RET as a major gene involved in HSCR (Schuchardt et al., 1995; Uesaka et al., 2008). However, mutations in the coding region (CDS) of this gene are only identified in 50% of familial and 15% of sporadic HSCR cases (Angrist et al., 1995; Attie et al., 1995, Hofstra et al., 2000). Nevertheless, linkage studies performed in sib-pairs and multi-generational HSCR families showed that in the majority of patients, association to the RET locus was found, even when CDS mutations in RET had not been identified (Attie et al., 1995; Bolk et al., 2000, Gabriel et al., 2002). Based on this result, it is clear that non-coding RfTvariants could also play a major role in HSCR development. Several studies performed by different groups identified the presence of a common disease associated haplotype in the 5' region of the RET locus in patients of European, American and Asian descent, confirming the idea that non-coding RET variants do play an important role in disease development (Borrego et al., 2003, Fitze et al., 2003, Sancandi et al., 2003; Burzynski et al., 2004; Pelet et al., 2005). This haplotype spans approximately 27 Kb and includes 4Kb of the 5' UTR, exon 1, intron 1 and exon 2, which pointed to the presence of possible mutations in either the promoter region or intron 1. In total, four presumed pathogenic single nucleotide

20 Introduction polymorphisms (SNPs) were identified in this region. Two of these SNPs (rs10900296 and rs10900297) were located in the promoter region just upstream of the RET transcription starting site (Fitze et al., 2003), and were shown to reduce the binding affinity of the transcription factor TTF-1 (Thyroid Transcription Factor 1) (Garcia­ Barcelo et al., 2005). They were also shown to reduce promoter activity in luciferase reporter assays, but this effect was found to be cell-line-dependent (Griseri et al., 2005). The two other SNPs are located in intron 1. They have been considered as disease-causing variants based on association studies, comparative genomics and functional assays: rs2435357 and rs2506004 (Burzynski et al., 2005; Emison et al., 2005; Emison et al., 2010; Sribudiani et al., unpublished data). Both these SNPs are in close proximity to each other and are located in highly conserved regions among different mammalian species. Mutant alleles of these SNPs showed the highest values for association in HSCR patients and the region containing both SNPs showed reduced promoter activity in luciferase assays (Emison et al., 2005). Additional studies performed by Grice et al., and Fisher et al., in mice and zebrafish respectively, confirmed the relevance of this region in regulating RET expression (Grice et al., 2005; Fisher et al., 2006). Several transcription factors were predicted to bind to these two regulatory regions, however no functional data was available to support these predictions. Recently, however, two independent studies were able to identify the binding of two transcription factors to these regulatory regions. SOXlO binding was reported to be disturbed by SNP rs2435357 (Emison et al., 2010), and the heterodimer NXF2/ARNT was shown to bind to the regulatory region where SNP rs2506004 is located (Sribudiani et al., unpublished data). Furthermore, both these SNPs were shown to reduce RET activation, confirming the importance of these two regulatory regions in HSCR development. Taken together, these data show that, in almost all HSCR cases, RET plays a minor or major role in disease development. Besides RET, ten other genes have been implicated in the development of HSCR: GDNF, NTN, EDNRB, EDN3, SOXlO, ECE-1, ZFHX1 B, PHOX2B, KIAA1279/KBP and NRGl (Angrist et al., 1996; Salomon et al., 1996; Doray et al., 1998; Puffenberger et

21 Chapter 1 al., 1994; Hofstra et al., 1996; Edery et al., 1996; Pingault et al., 1998; Hosftra et al., 1999; Cacheux et al., 2001; Amie! et al., 2003; Brooks et al., 2005; Garcia-Barcelo et al., 2009). GDNF and NTN encode for the RET ligands, GDNF and Neurturin (NTN), which are necessary for receptor phosphorylation and activation of downstream signalling pathways. EDNR, EDN3 and ECE-1, encode for the endothelin receptor type B, endothelin 3 and endothelin-converting enzyme 1, respectively, and they are part of the endothelin signalling pathway. SOX10, ZFHXlB and PHOX28 encode for transcription factors involved in RET and/or EDNRB activation. KIAA1279/KBP encodes for the kinesin-binding protein which has recently been involved in neuronal differentiation (Alves et al., 2010). Finally, NRGl, recently identified (Garcia-Barcelo et al., 2009), encodes for a receptor protein that belongs to the Erb B family of tyrosine kinase. Despite the fact that all these genes have been identified as playing a role in HSCR disease, they only account for 20% of all the cases, with most of them being associated to syndromic forms of HSCR (Amiel et al., 2008). This suggests that other, unidentified genes or loci are likely to be involved in the development of this disease.

Syndromic HSCR In 70% of the cases HSCR occurs as an isolated trait, while 30% of the cases are associated with other congenital malformations (Amiel and Lyonnet, 2001). Chromosomal abnormalities are found in 12% of cases, trisomy 21 being the most frequent (90%), while congenital anomalies are found in 18% of HSCR patients (Amiel et al., 2008). Among the eleven genes described as playing a role in HSCR, seven have been directly associated to the development of congenital malformations: EDNRB, EDN3, ECE-1, SOX10, PHOX28, ZFXHl and KIAA1279/KBP. Mutations in EDNRB and EDN3 are involved in an autosomal recessive form of Waardenburg-Shah syndrome (WSS). WSS is characterized by sensorial deafness, pigmentary abnormalities of the skin, eyes and hair, and by HSCR disease (Omenn and

22 Introduction

McKusick, 1979; Shah et al., 1981). A dominant form of WSS has also been identified in some cases where additional neurological features are present. However, in this dominant form, mutations in SOXlO are responsible for the phenotype (Pingault et al., 1998). An ECEl mutation was identified in a single patient with HSCR, craniofacial and cardiac defects (Hofstra et al., 1999). Congenital Central Hypoventilation syndrome (CCHS), also called Ondine's curse, is a life-threatening disorder characterized by failure of the autonomic control of breathing, especially during sleep (Weese-Mayer et al., 2005). HSCR is variable feature and is present in approximately 20% of CCHS cases. The association of HSCR with CCHS is called Haddad syndrome (Trochet et al., 2005; Haddad et al., 1978; Verloes et al., 1993). PHOXZB was identified as the disease-causing gene with de novo heterozygous mutations, and a mutation in this gene has been identified in more than 300 CCHS patients (Amiel et al., 2003; Weese-Mayer et al., 2003; Matera et al., 2004; Trang et al., 2005). Mutations in ZFXHl are involved in Mowat-Wilson syndrome (MWS), a multiple congenital anomaly characterized by microcephaly, epilepsy, a facial gestalt and severe mental retardation (Mowat et al., 1998). HSCR is a variable feature and is present in 40-70% of the cases (Zweier et al., 2005). Heterozygous de novo deletions encompassing ZFHXl or truncating mutations within this gene have been found in over 100 cases (Wakamatsu et al., 2001; Cacheux et al., 2001; Amiel et al., 2001). Additionally, rare splicing and missense mutations have also been reported (de Pontual et al., 2007). Goldberg-Sprintzen syndrome (GOSHS) is an autosomal recessive disorder characterized by polymicrogyria, microcephaly, facial dysmorphic features, mental retardation and HSCR (Goldberg and Shprintzen, 1981). Homozygous nonsense mutations in KIAA1Z79/KBP were identified as being the disease-causing factor (Brooks et al., 2005).

23 Chapter 1

Multiple Endocrine Neoplasia type 2 (MEN2) MEN2 is a dominantly inherited cancer syndrome that accounts for approximately 25% of all the medullary thyroid carcinoma (MTC) cases. MEN2 can be divided into three subtypes: MEN2A, MEN2B and Familial Medullary Thyroid Carcinoma (FMTC), based on the clinical description and tissues affected (Hansford and Mulligan, 2000). All subtypes are characterized by the presence of medullary thyroid carcinoma (MTC), a tumour of the parafollicular C-Cells, which are responsible for the secretion of calcitonin by the thyroid gland (Sipple, 1961). However, each MEN2 subtype also has its own characteristic features. In MEN2A, pheochromocytoma (PC) is observed in 50% of the cases and hyper-parathyroidism (HPT) in 15-30% of the cases (Sipple, 1961). MEN2B patients also develop PC (in approximately 50% of the cases), but instead of HPT, these patients develop a more complex clinical phenotype that includes ganglioneuromas on the tongue, lips and eyelids, intestinal ganglioneuromas, thickened corneal nerves as well as a marfanoid habitus. It is the most aggressive form of MEN2 with the earliest age of onset (Carney et al., 1976; de Groot et al., 2006). FMTC is characterized by the development of MTC as the only clinical manifestation and normally several family members are affected. In general, patients with FMTC tend to develop detectable tumours at a later age than patients with MEN2A and MEN2B and they also tend to have a more favourable prognosis (Yip et al., 2003). Activating mutations can occur in specific domains of the RET protein and, depending on their location, they are associated with either MEN2A, MEN2B or FMTC (Fig. 5). RET mutations associated to MEN2 result in a ligand-independent activation of the encoded RET protein, which leads to a constitutive phosphorylated receptor and consequently to uncontrolled activation of downstream signalling pathways. Mutations located in the cysteine-rich domain of the extracellular region of RET (codons 609, 611, 618, 620, 630 and 634), are generally involved in the development of MEN2A. These mutations result in the replacement of a critical cysteine residue by another amino acid leading to the loss of an intramolecular disulphide bond. As a

24 Introduction consequence, a "free" cysteine residue becomes available for the formation of an intermolecular disulphide bond (with another mutated RET monomer), resulting in a constitutively activated receptor (Plaza-Menacho et al., 2006). Mutations located in the intracellular tysosine kinase domain are associated with the development of either MEN2B (codons 883 and 918) or of FMTC (codons 768, 790, 791, 804 and 891). The way these mutations affect RET activation is completely different from MEN2A mutations as they signal independently of GDNF as monomeric oncoproteins. However, it has been shown that GDNF can further enhance MEN2B activity (Plaza-Menacho et al., 2005; Plaza-Menacho et al., 2006). Independent of the type and location of the mutation, the end result is always the same: continuous RET phosphorylation that leads to a constant activation of downstream signalling pathways and ultimately to uncontrolled tumour growth.

MEN2A FMTC MEN2B

Codon 609 611 618 620 Cadhenn Domain 630 Cysteme Rich Domain 634

768 790 791 804 883 891 RET43 918 RET51

Figure 5 - Overview of the location of known missense mutations in RET (Adapted from Plaza-Menacho et al., 2006).

25 Chapter 1

Papillary Thyroid Carcinoma {PTC) PTC is the most common cause of thyroid cancer, representing approximately 80% of all thyroid cancer cases. PTC occurs in the follicular epithelial cells of the thyroid and exposure to radiation is a major risk factor for its development (Sherman, 2000). Several genetic alterations have been reported to be associated with PTC, including oncogenic activation of RAS (Fagin, 2002), v-RAF murine sarcoma viral oncogene homolog B1 (BRAF) (Xing, 2005), hepatocyte growth factor receptor (MEn (Wasenius et al., 2005), and chromosomal alterations that affect NTRK1 and RET (Pierotti et al., 1998). Specific rearranged forms of RET, among which is the same rearrangement originally found by Takahashi et al. (Takhashi et al., 1985), were detected in PTC and account for a small percentage (S-30%) of cases, with the majority being caused by exposure to external radiation. In these rearrangements, the intracellular kinase domain of RET is recombined with the N-terminus of a gene normally expressed in the follicular epithelial cells that is capable of ligand-independent dimerization. The RET/PTC proteins do not have a transmembrane domain, and thus are not integral membrane proteins, but RET is constitutively active. Several genes have been described as being rearranged with RETforming different hybrid oncogenes. The most prevalent variants are RET/PTC1 and RET/PTC3, which result from the rearrangement of RET with H4 and Elel, respectively (Nikiforov, 2002).

Signalling pathways involved in HSCR and MEN2

HSCR associated signallingpath ways

Several genes have been implicated in HSCR development. Although a first look at the proteins encoded by these genes shows no apparent connection between them, the majority of these genes can be divided into three groups: those involved in RET activation pathways (RET, GDNF, NTN), those involved in EDNRB pathways (EDNRB, EDN3, ECE-1), and transcription factors that can affect both RET and/or EDNRB pathways (SOX10, ZFXH1B, PHOX2B).

26 Introduction

Effect of RET signalling on NCSCs Besides RET itself, mutations have also been identified in the coding sequences of GDNF (Angrist et al., 1996; Salomon et al., 1996) and NTN (Doray et al., 1998), both encoding for RET ligands necessary for receptor activation. One of the main purposes of the RET signalling pathway is to promote migration of NCSC precursors into the foregut (Pachnis et al., 1998; Young et al., 2001). For that, GDNF works as a chemoattractant promoting migration of RET­ positive NCSCs into the foregut. In the earliest stages of gut colonization (E9.0), high levels of GDNF are observed in the stomach ahead of the migrating front of NCSCs. Through time, GDNF expression shifts from the migration front to the streams of migrating cells, probably exerting a proliferative and survival effect rather than a migratory effect (Natarajan et al., 2002). Gdnr1- mice showed an approximately 50% reduction in the number of NCSCs, which was probably due to reduced NCSC proliferation and delayed NCSC migration (Gianino et al., 2003; Shen et al., 2002; Flynn et al., 2007). Furthermore, equivalent aganglionosis phenotypes were observed in mice and zebrafish in the absence of Ret, Gdnf or Gfral, supporting the importance of this pathway for gut colonization (Schuchardt et al., 1994; Shepherd et al., 2004; Sanchez et al., 1996; Pichel et al., 1996; Moore et al., 1996; Cacalano et al., 1998; Shepherd et al., 2001).

Effect of EDNRB signalling on NCSCs Mutations in EDNRB, EDN3 and ECE-1 have been identified and these genes encode for components of the EDNRB pathway. EDNRB is a G-protein coupled receptor that is activated by endothelins such as EDN3. However, for this activation to take place, EDN3 needs to be first cleaved in an active peptide and this is done by the endothelin converting enzyme 1 (ECE-1). Ednrb is expressed primarily in migrating NCSCs, whereas Edn3 is expressed in the midgut and hindgut during early phases of NCSCs migration, and at high levels in the caecum and proximal colon as NCSCs colonize the terminal gut regions (Barlow et al., 2003; Leibl et al., 1999). This pattern of expression

27 Chapter 1 suggests that EDN3-EDNRB signalling is involved in regulating the normal migration of NCSC. Consistent with this model, Edn3 and Ednrb mutant mice exhibit delays in NCSC migration (Barlow et al., 2003; Lee et al., 2003; Ro et al., 2006). Furthermore, several studies have suggested an additional role for EDN3-EDNRB signalling in maintaining enteric progenitors in a proliferative state, because NCSC numbers are reduced in Edn3 mutant mice (Barlow et al., 2003; Bondurand et al., 2006), and NCSC differentiation is inhibited in the presence of EDN3 (Bondurand et al., 2006; Nagy and Goldstein, 2006; Wu et al., 1999). These data implicate EDN3-EDNRB signalling in several aspects of the Enteric Nervous System (ENS) development, as well as in coordinating normal colonization of the gut by NCSCs.

RET and EDNRB interaction Interaction between the RET and EDNRB signalling pathways were described first in a genetic study and afterwards using an animal model. A genome-wide association study using 43 trios from a genetically isolated Mennonite population reported the joint associations between the inheritance of an EDNRB mutation and a certain RET allele in patients with HSCR (although no RET mutations were identified), suggesting epistasis between RET and EDNRB (Carrasquillo et al., 2002). This idea was supported using animal models: mice homozygous for the recessive hypomorphic allele of Ednrb (EdnrbS or Piebald, which only rarely exhibit aganglionosis) and heterozygous for a Ret null mutation (Ret1-) showed high frequencies of aganglionosis (Carrasquillo et al., 2002; Mccallion et al., 2003). Similar genetic studies showed that, whereas Ref1151 and Edn315115 (lethal spotting) mice display colonic aganglionosis, combinations of these mutant alleles led to almost complete intestinal aganglionosis (Barlow et al., 2003). Together, these studies suggest a genetic interaction between the RET and EDNRB signalling pathways. Since the effects of EDN3 on both proliferation and chemoattraction could be mimicked by PKA inhibitors, it was suggested that PKA might serve as a connection link between RET and EDNRB signalling (Barlow et al., 2003). Furthermore, a critical role for PKA was also suggested by the NCSC migration

28 Introduction defects and aganglionosis phenotypes observed when the putative PKA binding site of RET, serine 697, was mutated (Asai et al., 2006). However, it is unclear whether these phenotypes are related to a role of PKA in mediating RET or EDNRB signalling. Parallel to these findings, GDNF and EDN3 were seen to have opposite effects during the development of the ENS. While GDNF promoted the proliferation of migratory NCSCs and encouraged differentiation (Hearn et al., 1998; Barlow et al., 2003), EDN3 was shown to modulate GDNF activity to inhibit differentiation (Hearn et al., 1998; Wu et al., 1999). These results led to a model in which EDN3 acts to balance the effect of GDNF to prevent premature differentiation of NCSCs and maintain these cells in a proliferative state, thus generating appropriate numbers of NCSCs to fully colonize the gut (Hearn et al., 1998; Wu et al., 1999). However, how the balance between RET and EDNRB signalling is achieved is still unknown.

Transcription factors In the third group of genes implicated in HSCR development, those encoding transcription factors were described as mutated in HSCR patients: SOXlO, PHOX2B and ZFXH18. Interestingly, mutations in these genes were never found in patients presenting with HSCR as an isolated trait, but they are always associated with other congenital anomalies involved in rare recessive syndromes. SOXlO encodes a transcription factor that can regulate expression of both RET and EDNRB (Pattyn et al., 1999; Zhu et al., 2004). Indeed, regulatory binding regions have been found in RET and EDNRB that, once mutated, led to an impairment in SOXl0 binding and, as a consequence, to a decrease of RET expression (Heanue and Pachnis, 2007; Emison et al., 2010). PHOX2B is also a transcription factor that is expressed in several classes of differentiating neurons of both the peripheral and central nervous systems. Mice with a homozygous disruption of Phox2b showed no enteric ganglia and no RET expression which suggests that PHOX2B might play a regulatory role in RET expression {Pattyn et al., 1999; Brunet and Pattyn, 2002).

29 Chapter 1

Finally, ZFXH1 encodes for a transcriptional repressor which interacts with several members of the SMAD family. Expression of ZFXHl was detected in early neuronal development in Xenopus and homozygous mutant Zfxhl mice showed a complete lack of NCSC precursors (Van de Putte et al., 2003). However, the exact role of ZFXHl in HSCR development is unknown.

Oncogenic RET signalling The observation that different RET mutations result in different disease phenotypes might be explained by signalling through diverse downstream pathways triggered via differential activation of the mutated RET receptors. However, not much is known about these specific signalling pathways initiated by oncogenic RET receptor. There might be subtle differences in protein conformation when RET is activated by either ligand-binding or MEN2 mutations that can lead to the initiation of different signalling pathways. Furthermore, studies have shown that wild-type RET and mutated RET receptors display differences in phosphorylation of docking sites, and are thus likely to activate downstream pathways specific for a disease phenotype (Liu et al., 1996; Arighi et al., 2004; Chiariello et al., 1998). This indeed seems to be the case. Mutations associated to MEN2A and MEN2B have been described as having a marked impact on downstream Akt activation when compared to the wild-type receptor (Freche et al., 2005). Akt is part of the Pl3 kinase/Akt pathway which is responsible for cell survival, enhanced cell-cycle progression, and RET-mediated transformation. Interestingly, this pathway seems to be more active in MEN2B than in MEN2A (Segouffin-Cariou and Billaud, 2000; Murakami et al., 1999), a difference that might be due to enhanced phosphorylation at tyrosine 1062 (Y1062) by MEN2B mutations. Since Y1062 is the docking site for Shank3, an adaptor protein involved in activating of not only the Pl3 kinase pathway but also the Ras/MAP kinase pathway, this difference in phosphorylation levels probably contributes to the differences between the MEN2A and MEN2B phenotypes (Salvatore et al., 2001). Further evidence for the differences in oncogenic signalling triggered by the various mutation types is provided by a strong association of the JNK pathway with

30 Introduction

RET/MEN2B. Just like the Pl3K/AKT and RAS/MAPK pathways, Y1062 is also the phosphorylation site involved in JNK signalling activation. Thus, higher phosphorylation levels of RET on Yl062 seen in MEN2B mutants lead to an activation of the JNK pathway (Murakami et al., 2002; Grimm et al., 2001). STATs are a family of latent transcription factors involved in activating many genes induced by cytokines and growth factors (Schindler et al., 1995). However, STATs can also contribute to cancer development when aberrantly activated (Schindler et al., 1995). In particular, STAT3 has been reported to play a role in many types of cancer, including MTC. The MEN2A mutant RET C634R has been reported to enhance oncogenic activation of STAT3 via a process independent of Janus tyrosine kinases (JAKs) and c-Src (Schuringa et al., 2000). In contrast, FMTC mutants, RET Y791F and S891A, need the involvement of c-Src and JAKs to constitutively activate STAT3 (Plaza-Menacho et al., 2005). These data suggest that STAT3 activation by MEN2A­ and FMTC-associated mutations is triggered via different signalling routes. Not all RET mutant receptors are sensitive to GDNF stimulation and this characteristic may also influence oncogenic signalling. MEN2A mutations such as RET C634R are responsive to GDNF, while MEN2A/HSCR and FMTC/HSCR mutations, such as RET C620R, do not respond to the same stimulation (Arighi et al., 2004). A similar influence of GDNF was seen for MEN2B-associated mutations, such as RET M918T, since they are highly responsive to GNDF stimulations (Plaza-Menacho et al., 2005). These findings agree with the idea that differences in signalling pathways activated by different RET mutations, combined with differences in GDNF responsiveness of these receptors, could be the explanation for the different phenotypes (Jhiang, 2003).

HSCR and MEN2

HSCR-associated RET mutations are inactivating mutations, disabling expression and activation of this receptor (Pasini et al., 1995), while MEN2 mutations result in a constitutively activation of RET (Santoro et al., 1995). Nevertheless, HSCR can be

31 Chapter 1 found in association with MEN2A and FMTC in patients with a mutation in codons 609, 611, 618 and 620 (Takahashi et al., 1999; Sijmons et al., 1998). These mutations seem to have a dual effect on RET: they can constitutively activate RET by the formation of covalent dimers, but they can also result in a drastic reduction of RET expression at the plasma membrane. As a consequence, uncontrolled proliferation of the C-cells of the thyroid occurs in MEN2, and apoptosis of enteric neurons is detected in HSCR. This double effect is likely to be due to differences in sensitivity to GDNF. MEN2A and FMTC mutations, such as RET C634R, are normally responsive to GDNF stimulation, but HSCR-MEN2A mutations such as RET C620R are not (Arighi et al., 2004). Considering that insensitivity to GDNF makes cells more prone to apoptosis, it might explain the features seen in all HSCR-associated RET mutations (Mograbi et al., 2001). Furthermore, as GDNF functions as a chemoattractor for the NCSCs, the lack of mutant RET receptors at the membrane might also be implicated in this process. Further studies are needed to confirm this hypothesis. Based on this dual effect of RET, it is tempting to suggest that understanding the protein network involved in MEN2 migth bring new insights into HSCR and vice versa. Previous studies have shown that, indeed, this could be the case. Homozygous mutations on leucine 1061 of RET were involved in HSCR development due to disruption of the docking site of She on tyrosine 1062 (Geneste et al., 1999). She is an adaptor protein involved in activating the MAP kinase pathway. Since this pathway is known to be involved in cell proliferation and differentiation, it is not surprising that it plays a role during enteric neurogenesis. However, it is also not surprising that it is involved in oncogenic RET signalling (Schuringa et al., 2000, Plaza-Menacho et al., 2007). These observations suggest that the signalling pathways involved in MEN2 and HSCR triggered by RET activation should be combined and seen as a whole. In this way, a full picture of the RET protein network will become available and might help in identifying new players involved in the development of HSCR and MEN2.

32 Introduction

Aims and outline of this thesis The aim of the work described in this thesis was to obtain a better understanding of the protein network involved in the development of Hirschsprung disease (HSCR) and Multiple Endocrine Neoplasia type 2 (MEN2). Thus, this thesis conveniently divides into two parts. In the first part we focus on HSCR, a complex inherited disorder for which several genes have been identified. However, since these known genes explain no more than 30% of all HSCR cases, it is likely that other genes are involved in the development of this disease. Giving that HSCR is frequently associated with congenital malformations present in rare recessive syndromes, studying the genes involved in these syndromes may reveal new insights into the HSCR-related protein network and help in identifyingnew candidate genes. In chapter 2 we studied the function of the protein encoded by KIAA1279/KBP, the gene involved in Goldberg-Sphrintzen syndrome, a rare autosomal recessive syndrome characterized by HSCR, among other features. We found that KBP interacts with SCGlO, a stathmin-like protein, and is involved in neuronal differentiation. Because SCGlO was identified as a susceptibility gene for HSCR development in a mouse model, we decided to investigate the occurrence of SCGlO mutations in a group of HSCR patients. The results of this study are reported in chapter 3. To conclude the first part of the thesis, we have reviewed what is known about HSCR in chapter 4. We describe the aetiology of HSCR, its genetic heterogeneity, and the polygenic nature of this disease. Moreover, we discuss challenges and expectations from technological advances in the quest to better understand the protein networks involved in HSCR development and hope these will help identify candidate genes involved in the polygenic inheritance of this disorder. In the second part of this thesis, we focus on MEN2. In chapter 5 we aimed to gain more insights into the signalling pathways associated with different types of MEN2: MEN2A, MEN2B and FMTC. Kinome profile analysis of the most common mutations associated with each type of MEN2 were performed in an attempt to

33 Chapter 1 identify specific signalling pathways triggered by MEN2A, MEN2B and FMTC mutations. In chapter 6 the effect of four different tyrosine kinase inhibitors in the treatment of medullary and papillary thyroid carcinoma was assessed, in an attempt to find more effective therapeutic options for patients with a MEN2A or a MEN2B mutation, and with a RET/PTC rearrangement. Finally, in chapter seven, the results described in the previous chapters are summarized and discussed, and future perspectives are presented.

34 PART I

Chapter 2

KBP interacts with SCGlO, linking Goldberg­ Shprintzen syndrome to microtubule dynamics and neuronal differentiation

Maria M. Alves, Grzegorz Burzynski, Jean-Marie Delalande, Jan Osinga, Annemieke van der Goat, Amalia M. Dolga, Esther de Graaff, Alice S. Brooks, Marco Metzger, Ulrich L. M. Eisel, lain Shepherd, Bart J.L. Eggen and Robert M.W. Hofstra

Human Molecular Genetics, 2010, 19(18): 3642-51. Epub 2010 Jul 9. Chapter 2

Abstract Goldberg-Shprintzen syndrome is a rare clinical disorder characterized by central and enteric nervous system defects. This syndrome is caused by inactivating mutations in the Kinesin Binding Protein (KBP) gene, which encodes a protein of which the precise function is largely unclear. We show that KBP expression is up­ regulated during neuronal development in mouse cortical neurons. Moreover, KBP-depleted PC12 cells were defective in NGF-induced differentiation and neurite outgrowth, suggesting KBP is required for cell differentiation and neurite development. To identify KBP interacting proteins, we performed a yeast two­ hybrid screen and found that KBP binds almost exclusively to microtubule­ associated or related proteins, specifically SCGlO and several kinesins. We confirmed these results by validating KBP interaction with one of these proteins: SCGlO, a microtubule destabilizing protein. Zebrafish studies further demonstrated an epistatic interaction between KBP and SCGlO in vivo. To investigate the possibility of direct interaction between KBP and microtubules, we undertook co-localization and in vitro binding assays, but found no evidence of direct binding. Thus, our data indicates that KBP is involved in neuronal differentiation and that the central and enteric nervous system defects seen in Goldberg-Shprintzen syndrome are likely caused by microtubule-related defects.

Keywords: KBP, SCGlO, kinesins, microtubules, neuronal development, axonal growth.

38 KBP interacts with SCGlO

Introduction Goldberg-Shprintzen syndrome (GOSHS) is an autosomal recessive disorder characterized by polymicrogyria, mental retardation, microcephaly, facial dysmorphisms and, in most cases, by Hirschsprung disease (Goldberg and Shprintzen, 1981; Brooks et al., 2005). Homozygosity mapping studies showed linkage to chromosome 10 and subsequent sequence analysis of all 35 genes in this region led to the identification of truncating mutations in the KIAA1279 gene (Brooks et al., 2005). KIAA1279 encodes a protein with two tetratrico peptide repeats whose function is still uncertain. Based on the GOSHS clinical phenotype, it is clear that mutations in KIAA1279 are associated with both central and enteric nervous system defects. Subsequent studies by Wozniak and co-workers led to KIAA1279 being named Kinesin Binding Protein (KBP) due to its interaction with the motor domain of two related kinesin-like proteins, KIFlC and KIF1Ba. (Wozniak et al., 2005). The interaction with KIF1Ba.suggested that KBP plays a role in mitochondria localization and distribution (Wozniak et al., 2005). More recently, KBP has been implicated in the regulation of neuronal microtubules organization, and in axonal growth and maintenance (Lyons et al., 2008). This finding stemmed from the characterization of a kbp mutant zebrafish, the first animal model to be described for GOSHS (Lyons et al., 2008). However, the mechanism by which KBP exerts its effect on microtubules is still unknown. Microtubules are dynamic structures that provide mechanical support for the shape of cells, and a track along which molecular motors, kinesins and dyneins move organelles from one part of the cell to another (Howard and Hyman, 2003). To perform these functions, a cell must carefully control the assembly and orientation of its microtubule cytoskeleton. Several proteins have been described that modulate microtubule dynamics and they can be divided into two groups: microtubule stabilizing proteins (MAPs) and microtubule destabilizing proteins (Andersen, 2000). Although much is known about the role of MAPs, relatively little is known about the molecular machinery that controls the microtubule destabilizing family of proteins,

39 Chapter 2 especially with respect to neuronal development and regulation (Duncan and Goldstein, 2006). Here we describe the identification of several KBP-interacting proteins, all of which are implicated in microtubule transport and microtubule dynamics, and show that KBP is necessary for proper neuronal differentiation and neuronal development.

Materials and Methods

Constructs mKBP and hKBP constructs: pCMV-mKBP is an image clone ordered from Geneservice (5350638) with the full length mouse cDNA cloned into pCMVSport 6.1. The pCMV­ hKBPkiaa is derived from the human image clone (Geneservice - 455085), which was recloned into the pOTB7 vector. hKBP was cloned into pCMVsport6.1 using EcoRI-Xhol restriction sites. pGBKT7-mKBP construct was generated by Notl-Sa/1 digestion of pCMV-mKBP, followed by ligation with pGBKT7 vector. pCDNA-HA-mKBP was generated by EcoRI-Notl digestion of pGBKT7-mKBP followed by ligation with pDNA3-HA. pCDNA-HA-hKBP construct was generated by EcoRI-Xhol digestion of pCMV-hKBP followed by ligation with pCDNA3-HA. pEGFP-Nl­ hKBP construct was generated by hKBP PCR amplification from pCDNA-HA-hKBP using the primers KBP forward {5'-CGTGATGTGTACAGTATGGCGAACGTTCCG-3') and KBP reverse {5'-CGTGACGTCTAGAGGATTAAGTCAGGGCCATCTTG-3'). BsrGI-Xbal restriction sites included respectively in KBP forward and KBP reverse primers, were used to clone hKBP into pEGFP-Nl vector. pcDNA3-Myc-SCG10 construct: mSCGlO cDNA was obtained from the yeast two-hybrid screen and amplified using the primers SCGlO forward {5'­ CGCGGAATTCACAATGGCTAAAACAGCAATGGC-3') and SCGl0 reverse (5'­ CATGGCGGCCGCTGCTTCAGC CAGAC- 3'). EcoRI-Notl restriction sites, included respectively in the SCGlO forward and SCGl0 reverse primers (underlined sequences), were used to clone mSCGlO into pCDNA3-Myc vector.

40 KBP interacts with SCGlO

pS UPER-shKBP construct: The target sequence used for designing the shRNA was based on the rat KBP sequence: GCTCACGCTTACTATGACA. Two oligos were designed containing the target sequence in its sense and antisense form (bold) separated by the hairpin loop sequence and flanked by BamHI, and Hindlll, overhangs at the S'and 3'ends respectively (underlined). Forward sequence (5'­ GATCCCC GCTCACGCTTACTATGACACTTCCTGTCA TGTCATAGTAAGCGTGAGC TTTTTGGA�-3') and reverse sequence (5' -AGCTTTTCCAAAAA GCTCACGCTTACTATGACA TGACAGGAAG TGTCATAGTAAGCGTGAGC GG§_-3'). BamHI­ Hindlll restriction sites were used to clone the shRNA into pSuper vector.

Cell culture and transfections HEK293, NIH-3T3, Hela and NlE/115 cells were cultured in DMEM high medium containing 4.5 g/L of glucose, L-glutamine and pyruvate {Gibco), supplemented with 10% fetal calf serum (BioWhittaker) or with 10% newborn calf serum (Gibco) (for NIH- 3T3), and 1% penicillin/streptomycin (Gibco), at 37°C and 5% CO2. For transient transfections, 300.000 cells of HEK293, Hela, 3T3 and NlE/115 were seeded in 6-well plates. Transfections were performed 24 hours after seeding using jetPei (Promega).

Culture of primary cortical neurons Primary cortical neurons were prepared from embryonic brains (ElS-16) of C57Bl/6J mice. After mechanical dissociation, neurons were plated on 6-well plates previously coated with poly-D-lysine (2 µg/ml). Neurobasal medium (GIBCO) supplemented with 827-supplement (GIBCO), 0.5 mM glutamine (BioWhittaker), 1% penicillin/streptomycin (Gibco) and 2.5 µg/ml amphotericin B (Sigma-Aldrich), was used as culture medium. After 48 h, cells were treated with 10 µM cytosine arabinoside (Sigma-Aldrich) and incubated for another 48 h to inhibit non-neuronal cell growth.

41 Chapter 2

Yeast two-hybrid screen The yeast two-hybrid screen was performed using the Matchmaker 3 System (ClonTech) according to the manufacturer's instructions. Briefly, pGBKT7-mKBP was transformed into AH109 Saccharomyces cerevisiae yeast strain, which was then mated with the Y187 strain pre-transformed with an eleven days old mouse library (ClonTech). Clones that survived the quadrupole selection (-His, -Ade, -Leu, -Trp) and that stained positive for �-galactosidase activity were isolated. The DNA of these clones was recovered and after electroporation into KC8 cells, was subsequently analyzed by digestion with Hindlll. Clones that showed the expected two bands after digestion were used for repeating mating experiments with Y187 yeast strain transformed with empty pGBKT7 and pGBKT7 mKBP to eliminate false-positives. Clones able to grow in a triple dropout medium (-Leu/-Trp/-His) only in the presence of mKBP were subsequently transformed into JM101 cells. The DNA isolated from this strain was sequenced and the results were blasted for matches.

Cell lysates, co-immunoprecipitation assays and Western blot analysis Cell lysates were prepared as followed: cells were washed with ice cold PBS and incubated with lysis buffer (200 mM NaCl, 20mM Tris-HCI pH 7.8, 1% Triton X-100 and protease inhibitors) for 30 min on ice. Cell lysates were collected by scraping and cleared by centrifugation at 14000 rpm for 10 min in a pre-cooled (4°C) centrifuge. Supernatants were stored at -80°C before they were processed further for SDS-PAGE followed by Western blot analysis. For immunoprecipitation, lysates were incubated with HA-coupled sepharose beads (Roche) ON at 4°C. Precipitates were washed with lysis buffer and protein was eluted in loading buffer. For western blot analysis the following antibodies were used: KBP (affinity purified rabbit polyclonal directed against a peptide encompassing amino-acids 24-38 of mKBP, Eurogentec), SCGlO (rabbit polyclonal kindly provided by Ellen Coffey, Turku Centre for Biotechnology, Finland), HA (3F10; Roche), tubulin (B-5-1-2, Sigma), actin (C4, MP Biomedicals). Secondary antibodies used for immunodetection were goat anti-rabbit lgG-HRP (Santa

42 KBP interacts with SCGlO

Cruz Biotechnology), rabbit anti-rat lgG-HRP (DakoCytomatin), and goat anti-mouse lgG-HRP (Bio-Rad).

Microscopy and image analysis For immunofluorescence analysis NlE/115, 3T3 and Hela cells were cultured on poly­ D-lysine coated cover slips. 2% PFA was used as fixative agent for NlE/115 and Hela cells and 3T3 cells were fixed in ice cold methanol for 3 min at -20°C. Cells were made permeable with 1% BSA and 0,1% Triton X-100 in PBS. Endogenous SCGl0 was detected using a a-SCGlO antibody (clone LS/1, Antibodies Incorporated) followed by a goat a-mouse Cy3 secondary antibody (Zymed). HA-KBP was detected using a a-HA antibody (3Fl0; Roche) followed by goat a- rabbit Alexa-fluor 488 secondary antibody (Molecular probes). The Golgi protein giantin was detected using a a-giantin antibody (Abeam) followed by a goat a- rabbit Alexa-fluor 488 secondary antibody (Molecular probes) or a goat a- rabbit Cy3 secondary antibody (Jackson lmmunoResearch). Mitochondria staining was detected using MitoTracker (lnvitrogen). For microtubule staining, 3T3 cells were cultured on poly-D-lysine coated cover slips and fixed in ice cold methanol for 3 min at -20°C. Cells were permeabilized with 0.1% Triton X-100 and 1% BSA in PBS. Endogenous tubulin was detected using a­ tubulin antibody (B-5-1-2, Sigma), followed by an anti-mouse lgG conjugated antibody. Fluorescent images were made using a Leica TCS SP2 (AOBS) microscope.

In vitro polymerization of microtubules and high-speed fractionation NIH-3T3 cells were lysed in tubulin buffer (Cytoskeleton) containing 0.1% Triton X- M 100. Lysates were centrifuged at 100,000 X g for 1 h in an OptimaT MAX-E- Beckman Coulter ultracentrifuge. Supernatant fractions were incubated for 30 min at 37°C with purified, taxol-stabilized microtubules, which were generated according to the manufacturer's indications (Cytoskeleton™). Microtubules were pelleted down by centrifugation (100,000 X g for 1 hour) and both supernatant and pellet fractions were checked by Western blotting.

43 Chapter 2

PC12 culture, differentiation and transfection PC12 cells were cultured in RPMI 1640 (Gibco) supplemented with 10% horse serum (Gibco), 5% fetal calf serum (Gibco) and 1% penicillin/streptomycin (Gibco), at 37°C and 5% CO2. For transient transfections, 50,000 PC12 cells were seeded in 12-well culture plates coated with laminin (Millipore). 24h after, cells were transfected with pEGFP­ Nl vector alone or in combination with pSuper-shKBP vector using lipofectamine 2000 (lnvitrogen). 24 h after transfection, cells were cultured in low serum concentration for 12 h at 37°C and 5% CO2. P-NGF (R&D systems) was then added to the culture medium and 48 h after green cells were counted and analyzed for the presence of neurites.

Fish and embryos Zebrafish were kept and bred under standard conditions at 28.5°C (Westerfield, 2000). Embryos were staged and fixed at specific hours post-fertilization (hpf) as described elsewhere (Kimmel et al., 1995). To better visualize in situ hybridization results, embryos were grown in 0.2 mM 1-phenyl-2-thiourea (Sigma) to inhibit pigment formation (Westerfield, 2000).

Antisense oligonucleotide (Morpholinos) injections Splice blocking scg10a, scg10b and kbp morpholinos (SBMO) were designed to complement the sequences corresponding to the splice donor site at the predicted exon2/exon3 junction for scg10a, exonl/exon2 for scg10b and exon3/exon4 for kbp. scglOa SBMO: S'TTTCATTTCCTACCTTCAAACTCGC3', scglOb SBMO: 5'TTCTCCCACTTACCTTCAAACTCAC3', kbp SBMO: 5' AAGAGAGGGATGTTACCTTCTAATC3'. The MOs (Table Sl) were diluted in sterile filtered water over a concentrations range of 1-5 µg/µI. Approximately 1 nl of diluted MO was injected at the 1-2 cell stage using a gas-driven microinjection apparatus to determine the effects

44 KBP interacts with SCGlO of knocking down single genes or combinations of them. The standard control MO from Gene Tools was injected as a negative control for the MO experiments. A p53 MO (5'GCGCCATTGCTTTGCAAGAATTG3') was injected in combination with other MOs to eliminate potential apoptotic side effectsof MOs injections (Plaster et al., 2006).

Whole mount in situ hybridization and double-labeling immunohistochemistry in zebrafish Digoxigenin labeled riboprobes that complement scglOa, scglOb and kbp mRNAs were generated by linearization of pCR TOPO II vectors that contained partial ORFs of the genes. SCGlOa and SCGl0b plasmids were linearized with EcoRV (New England Biolabs) and subsequently transcribed with SP6 polymerase (Promega). KBP plasmid was linearized with Sall (New England Biolabs) and transcribed with T7 polymerase (Promega). Embryos were collected and processed for whole-mount in situ hybridization as previously described (Thisse et al., 1993). Digoxigenin-labeled probes were detected using a standard NBT/BCIP staining reaction. For immunohistochemistry, FoxD3-GFP transgenic embryos (Gilmour et al., 2002; Lister et al., 2006) were fixed in 4% PFA. To obtain frozen sections, tissues were cryoprotected overnight in a 20% sucrose/PBS solution and placed for lh at 37°C in a 20% sucrose/5% gelatin solution. Afterwards, tissues were rapidly frozen in isopentane pre-cooled in liquid nitrogen to -60°C. Frozen sections were cut at 14 µm, collected on Superfrost Plus microscope slides (BDH Laboratories), air dried, and stored at -20°C. Prior to labeling, sections were placed in warmed PBS (37°C) for 15 min in order to remove the sucrose/gelatine, and then rinsed in PBS (2X 5 min). Histochemical staining was done using a polyclonal rabbit anti-GFP antibody (Molecular Probes) as previously described (Delalande et al., 2008). Sections were then stained with DAPI (5 min). In order to observe any co-localization of KBP RNA and FoxD3-GFP, bright field in situ images were converted into grayscale, inverted and then inserted into the red channel before being overlaid with FITC images using Adobe Photoshop software.

45 Chapter 2

Results

1) KBP plays a role in neuronal development and neuronal differentiation Goldberg-Shprintzen syndrome is a developmental disorder in which both the central and enteric nervous systems are affected. To determine KBP expression levels in neuronal cells, we isolated ElS mouse cortical neurons and cultured them for 6 days. Cells were lysed each day from the second day of the experiment onwards, and protein levels were analyzed by Western blotting (Fig. lA). Our data shows that there is an up-regulation of KBP expression levels in primary cortical neurons that seems to go hand-in-hand with neuronal differentiation. In situ hybridization performed in zebrafish confirmed expression of KBP both in the central and enteric nervous system (Fig. 1B). A detailed examination of the expression pattern in the developing gut revealed that KBP is only expressed at low levels in migrating ENCSCs during the initial period of gut colonization (Fig. lC). Taken together, these mouse and zebrafish results suggest that KBP plays a role during the development of the central and enteric nervous systems. To determine if KBP is required for neuronal maturation, we reduced KBP expression in PC12 cells using a sh RNA expressing vector (Fig. 2A). We used PC12 cells because they have the ability to differentiate into neuronal-like cells in the presence of nerve growth factor (�·NGF) (Greene and Tischler, 1982). After co-transfection of PC12 cells with shRNA expressing vector targeting KBP mRNA and a GFP expressing plasmid to monitor the transfected cells, GFP expressing cells were counted (approx. 4,000 cells/condition) and morphologically analyzed for the presence of neurites (Fig. 2B). We consider any extension that could be seen from the cellular body a neurite and the presence of one small extension was enough to consider the cell as differentiated. Our results show that the number of PC12 cells that differentiate and produce neurites when KBP expression was impaired was significantly lower (p< 0.05) than for the control (Fig. 2C). These data indicate that KBP strongly contributes to the NGF-induced neuronal differentiation of PC12 cells.

46 KBP interacts with SCGlO

A) B )�----�-----24 hof �----48 hof � -- b"" KBP Aclln

C)

Figure 1 - KBP expression during neuronal development and neuronal maturation. (A) Western blot analysis of mouse primary cortical neurons cultured for 6 days showed that KBP expression levels increase during neuronal development. (B) Whole-mount in situ hybridization of KBP in zebrafish. At 20-somite stage KBP is expressed ubiquitously (B'). At 24 hpf KBP expression is restricted to the anterior central nervous system (CNS) and to the anterior gut (B"). This pattern of expression is maintained at 48 hpf with cranial ganglia (cg) also expressing KBP (B"'). (C) KBP expression in the hindbrain, cranial ganglia and the gut. Transverse sections of whole-mount KBP in situ hybridization/GFP immunohistochemistry performed in Tg(foxd3:gfp ) zebrafish at 48 hpf showed a strong KBP expression in the ventral neural tube (nt), the posterior lateral line ganglia (pllg), the developing liver (L) and gut (g) (Cl, Cl'). A magnified view of the gut shows that GFP/Foxd3 positive enteric neural crest cells (encc) express only low levels of KBP (Figure C2, C2').

A) C)

PC12+shKBP PC12

DPC12 a PC12+shKBP B)

• p < 0,05

47 Chapter 2

Figure 2 - KBP is required for PC12 differentiation. (A) Western blot analysis of KBP expression in PC12 cells transfected with GFP alone or in combination with a shRNA expression vector targeting KBP mRNA. (B)

Morphology of �-NGF treated PC12 cells expressing a GFP vector alone or in combination with the shKBP vector was analysed using a fluorescence microscope. (C) The number of green differentiated cells was counted in the absence and in the presence of the shKBP and showed to be significantly lower when KBP expression was reduced (p<0.05). Scale bars 15 µm.

2) Yeast two-hybrid interactions of KBP KBP contains two tetratricopeptide repeats which mediate protein-protein interactions (Blatch and Lassie, 1999), suggesting that KBP likely binds to other proteins. To identify KBP interacting partners, a pre-transformed Ell mouse cDNA library was screened using a yeast two-hybrid screen. The Ell mouse library was used as this is a critical stage in brain development and coincides with the stage at which the colon is being colonized by neural crest cells. From our yeast two-hybrid screen we selected 69 positive clones which we isolated and sequenced. 19 of these 69 clones encoded for the same protein, SCGlO, a stathmin-like protein. In addition, 6 kinesins were found among these inserts, namely KIFSB, KIFSC, KIF7, KIF2C, KIF3A and also, KIFCl. A blast search using the cDNA of the kinesins found revealed that they all interact with KBP via their motor domain. None of the kinesins previously described to interact with KBP (Wozniak et al., 2005), KIFlC and KIFlBa, were found in our screen. Additionally, three of these 69 clones encoded for tubulin, the main constituent of microtubules (Table 1) . Taken together, these results strongly suggest a role for KBP in microtubule organization and stability. Moreover, the fact that all these kinesins were obtained in this screen reinforced the idea that KBP is involved in kinesin-mediated microtubular transport and is indeed a kinesin binding protein. Since the majority of clones obtained encoded for microtubule associated or microtubule interacting proteins, we conclude that KBP might be involved in microtubule-related events. In order to confirm the validity of this result, we decided to further study KBP interaction with the most frequently found clone, SCGl0.

48 KBP interacts with SCGlO

7 clones KIFSB 1 clone KIFSC 1 clone KIF7 2 clones KIF2C 1 clone KIF3A 6 clones KIFCl 3 clones Tubulin ex7 19 clones Superior cervical ganglia, neural specific 10 - SCG10

Table 1. Yeast two-hybrid results for the Kines in binding protein (KBP)

3) KBP interacts with SCG10 SCG10 is a member of the stathmin family of proteins, all of which bind to tubulin to act as a sequestering agent and promote microtubule disassembly (Bondallaz et al., 2006). SCG10 is a membrane associated neuronal protein that is largely expressed during development and whose expression correlates with neurite outgrowth (Bondallaz et al., 2006). It possesses a unique N-terminal domain that is critical for membrane binding, is responsible for SCG10 localization to the Golgi complex and is important in the targeting of SCG10 to the growth cones (Nixon et al., 2002; Di Paolo et al., 1997). To confirm the specificity of the KBP interaction with SCG10 we performed a co-immunoprecipitation assay. HA-KBP was expressed in Human Embryonic Kidney cells (HEK293) alone or in combination with Myc-tagged SCG10. Expression of exogenous Myc-SCG10 was necessary as HEK293 cells do not endogenously express SCG10. Precipitation of Myc-SCG10 was only observed in the presence of HA-KBP (Fig. 3A), confirming an interaction between KBP and SCG10. This interaction was further supported by co-localization studies in a mouse neuroblastoma cell line, NlE-115. This cell line was chosen due to its ability to differentiate in neuronal like cells when incubated in the presence of low serum concentrations. Using this cell line we evaluated GFP-KBP and SCG10 cellular distribution in the cell. SCG10 is enriched in the

49 Chapter 2

Golgi complex and in the growth cones as expected (Di Paolo et al., 1997), while KBP shows a cytoplasmic distribution in the cell that overlaps with SCGlO specifically in the Golgi complex (Fig. 3B). To confirm KBP and SCGl0 co-localization in the Golgi complex we used a specific antibody against a Golgi protein, giantin (Linstedt and Hauri, 1993), as a Golgi marker (Fig. 3B). These results support our yeast two-hybrid data and confirm the cellular interaction between KBP and SCGl0.

4) Epistatic interaction between KBP and SCGlO in Zebrafish To further validate KBP-SCGlO interaction, distribution of the mRNA transcripts for both these proteins was evaluated and epistasis was determined using a zebrafish model. The zebrafish genome contains one KBP gene and two SCG10 orthologs (Lyons et al., 2008; Burzynski et al., 2009). To determine the spatial expression pattern of the zebrafish scglOa, scglOb and to compare it to kbp orthologs during embryogenesis, matched whole-mount in situ hybridizations were performed using riboprobes for all three genes. Discrete patterns of expression were detectable as early as 16 hours post-fertilization (hpf). At this stage, expression of both scglOa and scglOb genes is restricted to the posterior lateral line (PLL) ganglia and Rohan-Beard sensory neurons in the spinal cord (Fig. 4A). By contrast, expression of kbp is initially rather diffused and ubiquitous in the whole embryo (Fig. lB'). From 24 hpf kbp's expression becomes predominantly restricted to the anterior CNS and this pattern of expression persists throughout all the stages examined in the present study. kbp transcript is present in all brain regions, as well as in the anterior and posterior lateral line (ALL and PLL) ganglia (Fig. lC, 4A). SCGlO expression at 24 hpf can be detected within the anterior CNS regions including ventral telencephalon and diencephalon, all hindbrain rhombomeres, the ALL and PLL and primary sensory neurons of the spinal cord (Fig. 4A).

50 KBP interacts with SCGlO

A) HA·KBP + My<:· IP: HA Myc-SCGlO HA-KBP SCGlO

PD rT ,o n- ,0 FT ,0 " II"'·� . HA·KBP

Myc · SCGlO

B)

Figure 3 - KBP interacts with SCGlO. (A) HA pulldown performed in HEK293 cells co-expressing HA-KBP alone or in combination with Myc-SCGlO showed that Myc-SCGlO precipitates only in the presence of HA-KBP. (I-input, FT- flow through, PD- pull down). (B) Confocal images of NlE/115 cells showed that GFP-KBP co-localizes with SCGlO in the Golgi complex and in the neurites. Giantin staining was performed to confirm KBP and SCGlO co­ localization in the Golgi complex. Scale bars 20 µm.

At 48 hpf, expression of the SCGlO genes becomes more abundant but like the pattern of kbp expression, they are primarily restricted to the anterior CNS and cranial ganglia. Further in development (72 and 96 hpf) SCGlO transcripts remain expressed in the same CNS regions and are also expressed in the enteric neurons (Burzynski et al., 2009). The pattern of expression of both SCGlO orthologs in zebrafish is very similar, although scglOa transcript is much more abundant (Burzynski et al., 2009). At 48 hpf expression of scglOa, scglOb and kbp overlap in the forebrain region, retina,

51 Chapter 2 optic tectum, trigeminal, vagal ganglia, ALL and PLL ganglia, and in the hindbrain (Fig. 4A). To determine if there is an epistatic interaction between KBP and SCGlO, 1-2 cell stage zebrafish embryos were injected with splice blocking morpholinos (SBMO) targeted against scg10a, scg10b and kbp transcripts (Draper et al., 2001). Injections were carried out either separately or in combination. In addition, combined injections were performed with a p53 morpholino (MO) to exclude potential cytotoxic side effects of MOs used (Robu et al. , 2007). The concentration of scglO and kbp MOs were titrated to the sub-threshold doses so that when injected separately they resulted in no apparent phenotype (Fig. 48). By contrast, when the MOs were injected in combination, perturbation in development became apparent. In the case of scglOa and scg10b double injections, morphants displayed a general delay in development, smaller heads and eyes and axial defects (Fig. 48). This result suggests that the scg10 genes are functionally redundant and this is consistent with their almost identical patterns of expression. Injections of scglOa, scglOb and kbp MOs together resulted in a much more severe phenotype. Triple morphants were significantly smaller with severely malformed body axis, very small head and eyes, necrotic areas in the brain and heart edema. Injections with p53 MO reduced the cell death in the retina and optic tectum areas but did not change the general morphology of the morphants, confirming that the observed morphant phenotype is not the result of offsite effects of the MOs (Fig. 48) (Robu et al., 2007). These results strongly suggest an epistatic interaction between kbp and scglO genes and support the results of the yeast two­ hybrid screen and co-immunoprecipitation assays.

A) kbplntcral kbp dor»I scgl0.1 b1

52 KBP interacts with SCGlO

B) I.Dug "f/Ob . 0.5ug HJ!__Oit•b•ll1r•rsJ Cl,Jftlltul "''°". -- -! ..,rii�= .iiris=fil:::::!lo

]-,ti\lii.;:;;;;;;;;;::;:'?l_.::;,;,; ;;;;:=� -----... .,.00 --=-- l Figure 4 - Epistatic interaction of KBP with SCG10. (A) Whole-mount in situ hybridizations performed in zebrafish at 24 and 48 hpf show expression patterns of scglOa, scg10b and kbp orthologues. (B) Splice blocking morpholinos (SBMO) targeted against scglOa, scg10b and kbp transcripts were injected in zebrafish embryos at 1-2 cell stage. When injected individually hardly any phenotype was seen. scg10a and scglOb double injections, lead to a general delay in development. However, when kbp SBMO was injected in combination with scglOa and scglOb, morphants showed a severe phenotype with malformed body axis, very small head and eyes, necrotic areas in the brain and heart edema. A p53 morpholino was combined to exclude potential cytotoxic side effects of the morpholinos used.

5) KBP interaction with microtubules Based on the yeast two-hybrid results and on the direct interaction between KBP and SCGlO, a microtubule destabilizing protein, we hypothesize that KBP plays a role in microtubule organization or microtubule stability and dynamics. Furthermore, the recent study characterizing a kbp mutant zebrafish showed that the loss of KBP function leads to a disorganization of axonal microtubules and to an improper orientation of microtubules along the axonal axis (Lyons et al., 2008). In an attempt to further explain this microtubule effect, we investigated whether there is a direct interaction between KBP and microtubules by determining the co-localization of GFP­ KBP with tubulin in mouse fibroblast cells (NIH-3T3). We detected no co-localization between KBP and tubulin (Fig. SA). To rule out the possibility that the interaction of KBP with microtubules is very transient and thus difficult to detect in co-localization studies, in vitro microtubule binding assays were performed. Taxol-stabilized microtubules were purified and mixed with cytosolic cell extracts of NIH-3T3 cells expressing KBP. Microtubules were pelleted by high speed centrifugation and the binding of KBP to the microtubules was determined by co-sedimentation. We found

53 Chapter 2 that KBP remains in the supernatant both in the absence and presence of microtubules, suggesting that KBP does not directly associate with microtubules (Fig. SB).

A)

B) Microtubules Mlcrotubules + KBP KBP

SP SP SP KBP

Tubulln

Figure 5 · KBP association with microtubules. (A) Confocal images of NIH-3T3 cells expressing GFP-KBP and stained with an anti-tubulin antibody showed that KBP does not co-localize with microtubules. (B) Microtubule in vitro binding assay showed that KBP does not precipitate with Taxol-stabilized microtubules when subjected to high speed centrifugation (P- pellet, SP- supernatant). Scale bars 20 µm.

6} KBP has a cytoplasmic localization During the co-localization studies performed we noticed that KBP was distributed throughout the cytoplasm. Previous studies performed in NIH-3T3 cells reported that KBP co-localizes with mitochondria and that it plays a role in mitochondria distribution by interaction with KIFlBa. (Wozniak et al., 2005). To reassure ourselves that the GFP tag was not influencing KBP localization in the cell, we expressed HA­ human KBP in Hela cells. Using an anti-HA antibody and a specific marker for mitochondria {MitoTracker), we repeatedly observed that KBP is distributed throughout the cytoplasm and is not specifically localized to mitochondria {Fig. GA). To

54 KBP interacts with SCGlO further support this observation, we co-expressed a modified Bicaudal D2 (BICD2-N) construct in Hela cells with the HA-human KBP construct (Fig. 6B). BICD2 is a motor­ adaptor protein involved in the transport of various cargoes by dynein-mediated transport in Drosophila and mammals. Previous studies have shown that in the absence of its C-terminal, BICD2-N-terminal strongly binds to dynein and impairs its normal function leading to disruption of the retrograde distribution of membranous organelles (Hoogenraad et al., 2001). When we over-expressed BICD2-N in Hela cells mitochondria became aggregated, however, we saw no change in KBP localization (Fig. 6B). Since NIH-3T3 cells were also used by us with no change in the diffuse localization of KBP in the cell (Fig. SA), maybe a difference in the experimental procedure used could explain the differences in co-localization seen by us and by the previous study (Wozniak et al. , 2005). However, based on our results, we conclude that KBP has a cytoplasmic localization and does not co-localize with mitochondria.

A)

B)

Figure 6 - KBP has a cytoplasmic localization. (A) Confocal images of Hela cells expressing HA-hKBP and stained with a mitochondrial marker (Mitotracker) showed that KBP has a cytoplasmic distribution and does not co­ localize with mitochondria. (B) Co-expression of HA-hKBP and BicD2-N constructs in Hela cells confirms KBP

cytoplasmic localization. Scale bars 20 µm.

55 Chapter 2

Discussion Homozygous nonsense mutations in the KIAA1279/KBP gene encoding the Kinesin Binding Protein (KBP) lead to Goldberg-Shprintzen syndrome (GOSHS). Affected patients have undetectable levels of KBP and show both central and enteric nervous system defects (Brooks et al., 2005). Based on the clinical phenotype, KBP is required for normal neuronal development, but its precise function is largely unclear. Subsequent to our initial identification of KBP mutations underlying the GOSHS phenotype, two independent studies have investigated the function of the KBP protein. The first study showed that KBP has a mitochondrial localization and that it interacts with the a isoform of KIFlB, increasing its motility. The same study showed that a reduction in the expression levels of KBP led to mitochondrial aggregation, suggesting that KBP regulates mitochondria distribution in the cell by controlling KIFlBa activity (Wozniak et al., 2005). The second study characterized a zebrafish kbp mutant, which had reduced axonal growth and disruption of axonal microtubules (Lyons et al., 2008). Axonal degeneration and mislocalization of mitochondria were also observed in the neurons at later stages of development, suggesting that KBP is a regulator of the neuronal cytoskeleton (Lyons et al., 2008). Although both studies shed some light on KBP function, they only partially explain the role of KBP in GOSHS. Our results show that KBP plays a role in neuronal development and is necessary for neuronal differentiation. Using zebrafish as a vertebrate model, we clearly show that KBP is strongly expressed in the central nervous system (CNS) during embryonic development. KBP is also expressed in some peripheral nervous system (PNS) structures, such as the cranial ganglia, although its expression is low in the migrating enteric precursors. Our study also shows that KBP interacts with several microtubule related proteins, including the stathmin SCGlO, a microtubule destabilizing protein. Based on this interaction, we suggest that KBP potentially modulates the proper organization of microtubules. We also show that KBP interacts with several kinesins, as previously described (Wozniak et al., 2005), further confirming that KBP is indeed a kinesin binding protein and suggesting that KBP may

56 KBP interacts with SCGlO regulate multiple aspects of intracellular transport along microtubules. Furthermore, we can not exclude the possibility that KBP may be involved in signalling pathways mediated by primary cilium especially because one of the kinesins that we found in our yeast two hybrid screen, KIF7, was shown to be a cilia-associated protein (Liem et al., 2009). Taken together, these results provide further insights into the cellular mechanisms that are perturbed in GOSHS patients and that underlie the clinical phenotype.

KBP, SCG10 and microtubules Microtubules are dynamic structures composed of a and � tubulin dimers. These structures are necessary for many processes in the cell, including transport and maintenance of neuritic processes (Tanaka et al., 1991; Rochlin et al., 1996; Dent and Gertler, 2003; Manna et al., 2007). Microtubule dynamics are characterized by stages of catastrophe alternated with stages of polymerization; this process is called dynamic instability. Although dynamic instability occurs both at the plus and minus ends of purified microtubules in vitro, it is more pronounced at the plus ends (Manna et al., 2007). Several proteins are known to regulate microtubule dynamics, including KBP (Lyons et al., 2008). Our results show that KBP interacts with SCGlO, a stathmin-like protein previously reported to be involved in microtubule dynamics (Di Paolo et al., 1997; Grenningloh et al., 2004). Moreover, during differentiation SCGl0 enhances neurite outgrowth, a phenomenon that is believed to depend on microtubule dynamics (Grenningloh et al. 2004; Riederer et al., 1997). Our results strongly suggest that KBP's role in microtubule organization/dynamics is not a result of its direct interaction/association with microtubules but is instead due to its interaction with SCGlO. As KBP interacts with several kinesins, a possible function for KBP would be as an adaptor molecule for the transport of SCGl0 to the growth cones via kinesin related transport. However, our

57 Chapter 2 yeast two-hybrid screen showed that KBP's interaction with the kinesins is via the kinesin motor domain. Knowing that the motor domain is responsible for the docking of kinesins to the microtubules, it is very unlikely that KBP is directly involved in the transport of SCGlO to the growth cones. However, KBP might be involved in the process that precedes the transport of SCGl0 to the growth cones. Additionally, KBP might regulate SCGlO activity, leading to a change in its microtubule destabilizing properties and may thus be involved in the maintenance of microtubule dynamics. Another possibility is that KBP might mediate SCGlO interaction with tubulin. Consistent with this, we observed an interaction between KBP and tubulin in our yeast two-hybrid study. Potentially, KBP could control the levels of free tubulin available for microtubule polymerization. If either of these mechanisms is correct, KBP will affect neurite extension and axonal growth, and this is in line with our data and the analysis of the zebrafish kbp mutant (Lyons et al., 2008). Furthermore, as the expression pattern of SCGlO is restricted to neuronal tissue, any perturbation of the KBP-SCGl0-tubulin interaction will potentially result in neuronal development defects consistent with the GOSHS clinical phenotype. Finally, our observation that KBP is required for neuronal maturation of PC12 cells is consistent with our biochemical studies that showed an interaction between KBP and SCGlO.

KBP implications in GOSHS and HSCR GOSHS is a rare but severe genetic disorder characterized by central and enteric nervous system defects. Based on our results, we propose that the GOSHS phenotype is caused by a reduction in neuronal differentiation and perturbed microtubule dynamics due to a deregulation of SCGlO activity. The data we now present supports previous findings (Lyons et al., 2008). Since GOSHS patients frequently have Hirschsprung disease (HSCR), KBP interacting proteins could bring some new insights about HSCR development. Currently, the genetic basis of this condition can only be explained in approximately 20% of all the

58 KBP interacts with SCGlO cases, with RET being the major gene involved (Amiel et al., 2008). SCGlO has been previously identified as a down-regulated gene in a RET mouse model for HSCR (Heanue and Pachnis, 2007) and we have shown that SCGlO interacts with KBP in vitro and in vivo. Both these facts suggest that SCGlO might also play a role in HSCR development, but it is still not clear what mechanism is involved. SCGlO activity is known to be controlled by two post-translational modifications: palmitoylation and phosphorylation (Antonsson et al., 1998; Lutjens et al., 2000). The former is responsible for growth cones targeting of SCGlO and the latter is responsible for controlling SCGl0 activity. As RET modulates several signalling pathways, it is possible that RET, via one of its downstream pathways, controls SCGlO activity, linking this gene directly to HSCR. Further studies are required to determine if there is any association between RET and SCGlO and a possible involvement of SCGl0 in HSCR.

Conclusions Our study shows that KBP is necessary for neuronal development and neuronal maturation. Furthermore, we show that KBP interacts with several microtubule­ associated and microtubule-related proteins in vitro and in vivo, being likely involved in microtubule organization/stability. Our results provide new insights into perturbed cellular mechanisms that lead to GOSHS and suggest that KBP has an important role in modulating SCGlO function during neuronal development.

Acknowledgements The authors would like to thank: Susanne Kooistra and Loes Drenth-Diephuis for technical help during the yeast two-hybrid screen; Myrrhe van Spronsen and Dr. Casper Hoogenraad for technical assistance with the BICD2-N construct and the pSuper-shKBP vector; Prof. Eleanor Coffey for kindly providing us with the SCGlO antibody; Dr. Ben Giepmans for providing the NlE/115 cells and for technical suggestions; Klaas Sjollema for microscope assistance, and Jackie Senior for editing

59 Chapter 2 the manuscript. This work was supported by the Graduate School of Medical Sciences (GUIDE) and the Jan Kornelis de Cock Stichting grant to M.M.A.

60 Chapter 3

Mutations in SCG10 are not involved in Hirschsprung disease

Maria M. Alves, Jan Osinga, Joke B.G.M. Verheij, Marco Metzger, Bart J.L. Eggen, Robert M.W. Hofstra

Submitted Chapter 3

Abstract Hirschsprung disease (HSCR) is a congenital malformation characterized by the absence of enteric neurons in the distal part of the colon. Several genes have been implicated in the development of this disease that together account for 20% of all cases, implying that other genes are involved. Since HSCR is frequently associated with other congenital malformations, the functional characterization of the proteins encoded by the genes involved in these syndromes can provide insights into the protein-network involved in HSCR development. Recently, we found that KBP, encoded by the gene involved in a HSCR-associated syndrome called Goldberg­ Shprintzen syndrome, interacts with SCGlO, a stathmin-like protein. To determine if SCGlO is involved in the ethiology of HSCR, we determined SCGlO expression levels during development and screened 85 HSCR patients for SCGlO mutations. We showed that SCGlO expression increases during development but no germline mutation was found in any of these patients. In conclusion, this study shows that SCGlO is not directly implicated in HSCR development. However, an indirect involvement of SCG10 cannot be ruled out as this can be due to a secondary effect caused by its direct interactors.

Keywords: SCGlO, Hirschsprung disease, mutation, screening.

62 SCG10 mutation screening in HSCR

Introduction Hirschsprung disease (HSCR) is an abnormality of the enteric nervous system characterised by the lack of ganglia along a variable length of the gut. It is a developmental disorder that arises due to failure in migration of enteric neural crest cells into the intestinal tract or due to a failure in survival, proliferation, or correct development of enteric neurons once they reach the gut (Heanue and Pachnis, 2007). In either case an aganglionic segment occurs in which contraction and relaxation is absent. The length of the aganglionic segment can vary and HSCR patients are classified into short segment (S-HSCR, 80%), long segment (L-HSCR, 15%), or total colonic aganglionosis (TCA, 5%) (Chakravarti and Lyonnet, 2001). HSCR is a common genetic disorder with an estimated incidence of 1:5000 live births and is more frequent in males than in females (4:1), a difference most prominent in S-HSCR (Badner et al., 1990). Several genes have been implicated in the development of this disease, being RETthe most important one (Edery et al., 1994). RET (REarranged during Transfection) encodes a receptor tyrosine kinase which is expressed in neural crest derived lineages, playing a pivotal role during development of the enteric nervous system (Heanue and Pachnis, 2007). Mutations in the coding region of RET are responsible for 50% of familial HSCR cases and 15% of sporadic ones (Sancandi et al., 2000). However, they only account for 20% of HSCR cases, suggesting that other genes are involved. Since HSCR is frequently associated with other congenital malformations present in rare recessive syndromes, the study of the proteins involved in these syndromes can bring new insights about HSCR development. Such syndrome, in which HSCR is frequently observed, is Goldberg-Shprintzen syndrome (GOSHS), a rare autosomal recessive disorder. The gene responsible for this syndrome is KBP and recently we identified the Superior Cervical Ganglia Neural Specific-10 or SCGlO as the major interacting protein of KBP (Brooks et al., 2005; Alves et al., 2010). SCGlO is a stathmin­ like protein involved in microtubule dynamics (Grenningloh et al., 2004). But unlike other stathmins, SCGlO acts in two ways to promote microtubule dynamics: it

63 Chapter 3 stabilizes microtubules at their plus end and promotes microtubule catastrophe at their minus end (Manna et al., 2007). Moreover, SCGlO is known to play an important role in neuronal differentiation by enhancing neurite outgrowth, a phenomenon that is believed to be dependent on microtubule dynamics (Grenningloh et al., 2004). Also, SCG10 has been previously identified as a down-regulated gene in a RET mouse model for HSCR (Heanue and Pachnis, 2006) and we have shown that SCGlO interacts with KBP in vitro and in vivo (Alves et al., 2010). All together these facts suggest that SCGlO might also play a role in HSCR development. In this paper we evaluate the role of SCG10 in HSCR development by determining SCGlO expression levels during the process of gut colonization and by screening a set of isolated non-syndromic HSCR patients without RET mutations for the presence of SCG10 mutations.

Material and Methods

Culture of enteric neural crest stem cells (ENCSCs) ENCSCs were prepared from embryonic gut (jejunum to rectum) of C57BL/6 mice El0.5, 11.5, 12.5, 13.5, 14.5 and 15.5. The gut was mechanically dissected and then tissues were incubated in collagenase/dispase enzyme solution (collagenase XI [750 U/ml; Sigma-Aldrich] and dispase II [250 µg/ml; Roche] in PBS for approximately 5 minutes at 37°C. Digested tissue was triturated until a homogenous cell suspension was created. Cell suspensions were washed and seeded on 35-mm petri dishes 2 previously coated with fibronectin (2 µg/cm , Sigma-Aldrich). DMEM-F12 medium (PAA) supplemented with N2 supplement (lnvitrogen), B27-supplement (lnvitrogen), 1% penicillin/streptomycin (PAA), 20 ng/ml fibroblast growth factor (FGF, Peprotech) and 20 ng/ml epidermal growth factor (EGF, Peprotech), was used as culture medium. Every two days half of the medium was replaced with new medium containing freshly added FGF and EGF. The enteric neural crest-derived cells were kept in culture as neurospheres-like bodies (NLBs) for 14 days.

64 SCG10 mutation screening in HSCR

RNA extraction and RT-PCR After 14 days in culture, NLBs were collected and washed with PBS. RNA extraction was performed using the RNeasy mini kit0 (Qiagen). First strand cDNA was originated from 2ug of RNA using the Ready-To-Go You-Prime First-Strand Beads0 kit (GE Healthcare) and pdN(6) as first-strand primer. First-strand cDNA generated by this method was directly used for RT-PC. The primers used for amplification were SCGlO forward (S'-ACAATGGCTAAAACAGCAATGGC-3') and SCGlO reverse (S'­ TGCTTCAGCCAGAC-3'), Actin forward (S'-ATATCGCTGCGCTGGTCGTC-3') and Actin reverse (S'-AGGATGGCGTGAGGGAGAGC -3').

Patients All patients included in this study were diagnosed with HSCR disease and showed no mutations in the RET gene. All patients gave their written informed consent for the study. In total, 85 patients were screened. From these 85 patients 43 were S-HSCR, 11 were L-HSCR and for 31 cases the length of the segment affected was unknown.

Mutation screening of SCGlO Genomic DNA was isolated from peripheral blood lymphocytes by use of standard methods. Sequencing of exons 1-5 of SCG10 was performed using 5 sets of primers (for primers and PCR conditions see table 1). DNA amplification was performed using song of genomic DNA in 30µ1 PCR reactions containing: lOX reaction buffer, lOµM primer pair mix, 25mM dNTPs and 1 unit Taq Polymerase (Amersham). PCR products were amplified and purified (ExoSap it - GE Healthcare). Direct sequencing of exons 1- 5 was performed using the 3730 DNA Analyser (Applied Biosystems). Using the software Mutation Surveyor (Version 3.23, SoftGenetics LLC) sequences were aligned and compared with consensus sequences.

65 Chapter 3

Primer Sequences bp Size(bp) Tm °C PCR Seq.prim

SCGl0 lF TCTAGCACGGTCCCACTCTG 20 174 58 lF SCG10 1R AGGTAGAGCCGACGGAGAAC 20 SCGl0 2F ACCTGGCAATATTCACTCTG 20 340 58 2F SCGl0 2R TAGACACGGCAAGTCAATAG 20 SCGl0 3F CTCCCGGAATAACAACGCTAC 21 444 58 3F SCGlO 3R ACATGTTGGCATGGCACAG 19 SCGlO 4F CCGTTATTCTGCTAGGTTTG 20 327 58 4F SCGlO 4R TCAGGCATATGGAAGTTCAC 20 SCGl0 SF TAGACACCAAACTGGGTTAC 17 236 58 SF SCGlO SR ATCCTGATATCGCATGATCC 20

Table 1: PCR conditions.

Results

Expression of SCGlO in ENCSCs To determine SCGlO involvement in the development of the enteric nervous system, SCG10 expression levels were analysed in enteric neural crest stem cells (ENCSCs). Because gut colonization occurs between day 9 and 15.S of mouse development, ENCSCs were isolated from El0.5 until ElS.S mouse and maintained for 2 weeks in an undifferentiated state. In these ENCSCs, SCGlO expression was detected in the earliest stage analysed, El0.5, and an increase in expression levels was observed in the following stages (Fig. 1). Our data shows that there is an up-regulation of SCG10 expression levels in ENCSCs.

10.5 11.5 12.5 13.5 14.S 15.5 B ----� SCGlO Actin

66 SCG10 mutation screening in HSCR

Figure 1-SCGlO expression levels during the development of the enteric nervous system. Expression studies in mouse enteric neural crest cells show that RNA levels of SCGlO are upregulated in the initial stage of gut colonization (11.S) and remain constant until the whole process is finished by ElS.5.

Mutation screening of SCGlO To further evaluate the involvement of SCGlO in HSCR development, we screened blood DNA from 85 patients diagnosed with isolated, non-syndromic HSCR. No germline mutation was detected in SCGl0.

Discussion Hirschsprung disease (HSCR) is a developmental disorder in which the process of gut colonization is disturbed. To date, 11 genes have been reported to be involved in HSCR development: RET, GDNF, NTN, EDNRB, EDN3, SOXlO, ECE-1, ZFHXlB, PHOX2B, KBP and NRGl (Edery et al., 1994; Brooks et al., 2005; Angrist et al., 1996; Doray et al., 1998; Puffenberger et al., 1994; Hofstra et al., 1996; Pingault et al., 1998; Hofstra et al., 1999; Cacheux et al., 2001; Amiel et al., 2003; Garcia-Barcelo et al., 2009). However, all together, they only explain approximately 20% of all HSCR cases. To better understand HSCR development, the study of rare syndromes that are characterized by the presence of HSCR can be of major importance, as determining the protein network connected to the mutated proteins can help identifying candidate genes. In this study we focus on such a candidate gene, SCGl0, encoding a microtubule destabilizing protein, recently described to interact with KBP and believed to be important in the development of Goldberg-Shprintzen syndrome and consequently, in HSCR (Alves et al., 2010). Furthermore, SCGlO has been previously identified as a downregulated gene in a RET mouse model for HSCR (Heanue and Pachnis, 2006) reinforcing the idea that SCGlO might also play a role in HSCR development. Our results showed that SCGlO expression is clearly up-regulated during the process of gut colonisation, which suggests that SCGlO migth play a role in the development of the enteric nervous system. However, SCGlO screening in a set of 85 patients diagnosed with isolated, non-syndromic HSCR showed that no mutation

67 Chapter 3 was present. Taken together, these results suggest that although SCGlO seems necessary for the development of the enteric nervous system, it is not directly implicated in HSCR. A possible explanation for this result is that SCGlO involvement in HSCR can be a consequence of a disruption in the normal protein network necessary for proper development. SCGlO activity is known to be controlled by two post­ translational modifications: palmitoylation and phosphorylation (Lutjens et al., 2000). The first one is responsible for growth cone targeting of SCGlO and the second one is responsible for controlling SCGlO activity. As RET is a tyrosine kinase receptor known to be involved in several signalling pathways, it is possible that RET, via one of its downstream pathways, controls SCGlO activity, linking this gene directly to HSCR. Further studies are required to determine if there is any association between RET and SCGlO and a possible involvement of SCGlO in HSCR development.

Acknowledgements: The authors would like to thank Jackie Senior for editing the manuscript.

Ethical Statement: All patients included in this study gave their written informed consent. Ethical approval was obtained from the Medisch Ethische Toetsings commissie (METc).

68 Chapter 4

Hirschsprung disease as a model of complex disorders

Maria M. Alves, Bart J.L. Eggen, Gerard J. te Meerman and Robert M.W. Hofstra

Submitted Chapter 4

Abstract Genome-wide association studies (GWAS) have tried to identify the inherited risk factors contributing to the development of complex diseases. Although many disease­ associated variants have been found, only a small fraction of the total genetic risk for these diseases can be explained. The unexplained heritability is due partly to the diseases' multifactorial origin, and partly to a lack of power to identify associations with high frequency genetic variation. Rare and stronger mutations may also play a role in specific diseases, such as Hirschsprung (HSCR) disease. In this review, we focus on HSCR disease as an example of a complex disorder. We discuss what is known about HSCR disease, how new technological advances can be used to support the hypothesis of a risk contribution from rare penetrant mutations and how these advances can contribute to their identification.

Keywords: Hirschsprung disease, genes, GWA studies, next generation sequencing, gene-networks, systems biology

70 HSCR as a complex disorder

Introduction Complex diseases are common and a major contributor to disability and death worldwide. They are thought to arise from multiple predisposing factors, both genetic and non-genetic, and the joint effects of these factors are thought to be of key importance in the disease development (Lander and Schork, 1994; Kiberstis and Roberts, 2002). Understanding genetically complex (polygenic) diseases has become a major topic in the field of human genetics, since finding mutations that, in concert give rise to a disease has proven to be extremely difficult. Most of the attention has been given to large-scale association studies and in particular, to the progress made in the Human Genome Project and the HapMap project, which have made genome-wide association studies (GWAS) feasible (Witkowski, 2010). It is now possible to scan the entire genomes of patients and controls for more than 1 million DNA markers (i.e. single nucleotide polymorphisms or SNPs) and test these for genetic association. SNPs showing a considerable difference between cohorts of affected and non-affected individuals and those that are independently replicated can be considered genetic risk factors for the disease under study. This hypothesis-free approach has led to the identification of many novel, disease-susceptibility loci for a large number of multifactorial diseases. All these studies claim to have found new loci or genes and many results have already been replicated in independent populations. However, it is important to realize that GWAS use SNPs that have high heterozygosity and are less sensitive to find effects of low frequency variation. The results are therefore biased towards loci or variants that are common (occur frequently), meaning that the specific loci or variants do occur at high frequencies. Furthermore, all the recent studies make it clear that the common variants identified do not explain more than 20% on average of the total attributable risk for these complex diseases (i.e. they have low odd ratios (OR) of 1.2-1.5), which is substantially less than the heritability in fact observed for most diseases. Yang and colleagues (Yang et al., 2010) have put forward arguments that the missing heritability can be explained as a consequence of associations that

71 Chapter 4 are too small to be detectable and the bias due to the SNP composition of the genotyping arrays (Ioannidis et al., 2006). In addition to the valid arguments raised by Yang et al, we propose an additional explanation: besides many common weak mutations, complex diseases are also caused by rare, and stronger, mutations. In this review, we focus on Hirschsprung disease (HSCR) as an example of a complex disorder and we discuss what is known about HSCR, and what is expected from technological advances in detecting rare causative mutations. Furthermore, we discuss how these technologies can be used to find new candidate genes and rare mutations.

Box 1. HSCR - the disease

HSCR is a congenital disorder characterized (pathologically) by an absence of enteric neurons in the distal part of the gut. This absence leads to tonic contraction of the affected segment, intestinal obstruction that results in a failure to pass the first stool within 48 hours after birth, vomiting, and massive distension of the proximal bowel (also called megacolon) or neonatal enterocolitis. Patients are diagnosed with one of three forms: the short-segment form {S-HSCR, approximately 80% of cases) when the aganglionic segment does not extend beyond the upper sigmoid; the long-segment form (L-HSCR) when aganglionosis extends proximal to the sigmoid (Amiel et al., 2008); or total colonic aganglionosis or even total intestinal HSCR, which occur in a small minority of patients or (Amiel et al., 2008). One in three affected children also has other congenital anomalies, although a syndrome diagnosis is only established in a minority of these cases (Brooks et al., 2005). The enteric neurons that form the enteric nervous system (ENS) originate mainly from vagal neural crest cells that invade the foregut and migrate in a rostral to caudal direction to colonize the entire foregut, midgut, caecum, and hindgut. These vagal neural crest cells give rise to most of the enteric nervous system. Sacral neural crest cells also make a contribution but 1t is much smaller. These sacral cells migrate in the opposite direction, caudal to rostral, to colonize the colon (Burns and Douarin, 1998; Anderson et al., 2006). It is believed that HSCR can be caused by any failure in the migration, proliferation, differentiation or survival in the intestinal wall of the progenitors of the enteric neurons, the cells that are called entenc neural crest-derived cells or ENCCs (Heanue and Pachnis 2007).

HSCR and genetics The situation as seen in HSCR, a congenital The situation as seen in HSCR, a congenital disorder characterized by an absence of enteric neurons in the distal part of the gut, resembles many other complex diseases (for a more detailed description of the clinical features of HSCR see Box 1). We see a small proportion of familial cases,

72 HSCR as a complex disorder whereas the majority of cases are sporadic. Furthermore, HSCR in a substantial proportion of cases is associated with other congenital malformations and a syndrome diagnosis can be established in some cases (Moore, 2006; Amiel et al., 2008). The familial (syndromic) cases show a Mendelian mode of inheritance (both dominant and recessive), whereas the sporadic cases are believed to have a complex, non-Mendelian mode of inheritance (Badner et al., 1990). Since around 1990, most attention has been paid to the rare familial and syndromic cases: linkage analysis was performed in large multi-generational HSCR families, sib-pair analysis in smaller nuclear families, homozygosity mapping in consanguineous HSCR families, and genetically isolated populations were used to search for inherited ancestral alleles or shared haplotypes putatively carrying mutations (for a review see Brooks et al., 2005). Besides these extended family-based approaches, candidate genes were also identified based on natural mouse models, created mouse models and gene networks. Although these approaches successfully identified mutations in 11 genes, only -20% of the cases were explained by the mutations found (see Box 2 for an overview of the genes identified). In the last decade, most studies have focused on association, using both a case-control design and a trio design, and almost all screening isolated HSCR cases (Borrego et al, 2003; Fitze et al., 2003; Sancandi et al., 2003; Burzynski et al., 2004; Pelet et al., 2005). The association studies focused mainly on the major HSCR gene, RET. Haplotype sharing was found for a region of approximately 27 Kb, including 4 Kb of the 5'UTR, exon 1, intron 1 and exon 2 of RET. It is believed that one or more SNPs within intron 1 are involved in disease development (Emison et al., 2005, 2010; Burzynski et al., 2005). One GWAS was recently performed in Chinese cases of sporadic HSCR in an attempt to find additional loci that contribute to the development of this disease (Garcia-Barcelo et al., 2009). As expected a strong association to RET was found. However, two additional SNPs located in intron 1 of Neuregulin-1 gene (NRGl) were also found to be strongly associated, pointing to NRGl as a plausible candidate gene. Its involvement was corroborated by the identification of coding sequence mutations in NRGl. Under an additive model, the authors showed that there was significant association between

73 Chapter 4

RET and NRGl with an increase in the OR for the RET risk genotype in the presence of a heterozygous mutation in NRGl (Garcia-Barcelo et al., 2009).

Box 2. HSCR-associated genes

Genetic dissection of HSCR was successful in identifying 11 genes, with RET as the most important one (Edery et al., 1994; Romeo et al., 1994). RET encodes for a receptor tyrosine kinase involved in the activation of several downstream pathways. Besides RET, ten other genes have been implicated in the development of HSCR· GDNF, NTN, EDNRB, EDN3, SOX10, ECE-1, ZFHXlB, PHOX28, KIAA1279/KBP and NRGl (Puffenberger et al., 1994; Angrist et al., 1996; Salomon et al., 1996; Hofstra et al., 1996; Edery et al., 1996; Doray et al., 1998; Pingault et al., 1998; Hosftra et al., 1999; Cacheux et al., 2001; Amie! et al., 2003; Brooks et al., 2005; Garcia-Barcelo et al., 2009). GDNF and NTN encode for the RET ligands, GDNF and Neurturin (NTN), necessary for receptor phosphorylation and activation of downstream signalling pathways. EDNR, EDN3 and ECE-1, encode for the endothelin receptor type B, endothelin 3, and endothelin-converting enzyme 1 respectively and they are part of the endothelin signalling pathway. SOXlO, ZFHXlB and PHOX28 encode for transcription factors involved in RET and/or EDNRB activation. KIAA1279/KBP encodes for the kinesin-binding protein which has recently been involved in neuronal differentiation (Alves et al., 2010). Finally, NRGl, recently identified (Garcia-Barcelo et al., 2009), encodes for a receptor protein that belongs to the Erb B family of tyrosine kinases. For detailed information about these genes and their encoded proteins, see the recent review by Tam and Garcia-Barcelo (2009)

Can GWAS find the missing heritability? It is hypothesized that complex, late onset diseases are driven by the combined effect of many mutations, each of which may have only a small effect (Roopers, 2007). Patients are therefore believed to carry large numbers of mutations. If this hypothesis is true, it means these mutations must be very common in the general population and they must therefore be hundreds - if not thousands - of generations old. Finding only a small number of loci could be due to a lack of power in the study, so that enlarging the case cohorts should lead to additional loci being found. Many consortia are now collecting 10,000 samples or more, in anticipation of finding these loci. In a recent study on risk factors for coronary artery disease more than 100,000 samples were analysed to find 59 new loci (Teslovich et al., 2010). However, these enlargements are not without risk. To achieve a very large case cohort, consortia members need to combine their separate cohorts, which might have been diagnosed or ascertained in

74 HSCR as a complex disorder slightly different ways. Moreover, small genetic variations within or between cohorts can have strong effects on outcome and replicating results (Ioannidis et al., 2006). Although genotyping is very reliable, a small fraction of the SNPs still produce non­ Mendelian genotypes for known and unknown reasons. Due to the high number of SNPs tested, there is a possibility that false signals are being generated that are indistinguishable from true signals. Although expanding the cohorts will certainly help, the loci that will be identified will have an even lower effect than the ones that have been found already. Therefore we think that expanding the cohorts will probably not result in the identification of SNPs with a high attributable risk for these complex diseases. In addition, it should be recalled that association only suggests causation and inferring causality from a statistical association is not straightforward.

Effect of selection wilh 1% disadvantage Number of generations lo 90% elimination Number of generations to 99% elimination

t00.000

� 10.000 .2� -� 1000 1000 1 000 I

0 100 ! 100 100 � !� 0.010 10�������� 200 400 600 800 1000 0 10 1 00 10 00 Number ol genorallons SUMval disadvantage in % survival disadvantageIn ,c.

Figure 1 - Alleles that are associated with a decreased reproductive fitness will reproduce at a lower rate compared to alleles with normal fitness. Once a mutation has occurred, survival is in first instance mainly determined by chance, because the probability that a specific allele is retained over many generations is very low. Once alleles have drifted to a substantial number, adverse or positive selection comes into play. The number of descendants of a specific allele many generations later is highly variable, and follows an exponential distribution with the property that the standard deviation is equal to the mean. The probability that an allele that confers a reduced fitness survives over longer periods compared to the allele with normal fitness is mainly determined by selection, where even a low disadvantage will eliminate an allele over time. Most complex diseases lead only to a slight decrease in relative fitness which translates into a survival disadvantage dependent on the allele frequency and the disease incidence. An example calculation illustrates the effects to be expected. For a complex disease with a 2% incidence associated with a dominant allele with 20% allele frequency, giving a relative risk of disease of 1.5 compared to homozygotes for the non-disease allele and a reduced fitness of 90% for affected individuals, a survival disadvantage can be calculated. From this the expected number of generations where the allele frequency is reduced to 1% of its original value can be computed as 6777 generations. If the

75 Chapter 4 fitness is 50% reduced, the expected number of generations for a reduction to 1% of its original frequency is 1343. If the relative risk is 2.0, this number becomes 773 generations.

Will epistasis give the answers and is it related to the idea that combinations of mutations can influence each other in such a way that they enhance the effect of the mutations? Although epistasis exists, the combined effects of rare variants will be very difficult to detect and, moreover, cannot contribute strongly to the attributable risk because the specific combinations with increased risk over the additive risks will be relatively rare compared to the individual alleles. Was it realistic, from an evolutionary point of view, to assume that GWAS could identify many loci that would explain a large part of the genetic risk, or could moderate or strong predispositions be expected in the GWAS? Mutations that only have a marginal influence on reproductive success will already be selected against in a fairly efficient way. A mutation that has a selective disadvantage of only 1 %, will be reduced to 1/e = 36% in 100 generations, and to 12% after 200 generations (Figure 1). Taking this selection pressure into account and considering the findings from GWAS in complex genetic diseases, it seems inevitable that most of the common loci can only contribute marginally to the disease risk. Expecting very strong effects fom common loci seems unrealistic, or the mutation should provide the carriers with a selective advantage. Of course, one could argue that for late-onset diseases, the selective advantage is less important compared to early-onset polygenic diseases. However, many late-onset diseases, e.g. multiple sclerosis or dementia, also have early-onset forms and thus cannot be considered evolutionarily neutral. If we consider alleles with a selective advantage, HSCR provides us with a good example. The previously mentioned HSCR-associated allele gives homozygous carriers (50% of patients are homozygous for this variant), a 20 times higher chance of developing HSCR. As HSCR is a congenital disease, this allele can only exist when it gives carriers a selective advantage (Emison et al., 2005, 2010). The strongest argument for a selective advantage for heterozygotes for the RET mutation is that the same mutation is present in all the investigated populations. This points towards a very early origin,

76 HSCR as a complex disorder which is inconsistent with an evolutionary disadvantage. However, most data point towards a selective pressure driving the frequency of the mild and strongly associated mutations down to the rare variant level. Given these considerations, it seems realistic to conclude that GWAS will eventually identify only a small part of the total genetic risk. We therefore propose a complementary hypothesis, namely that besides many weak mutations, rare and strong mutations must contribute to complex diseases.

Rare variants In clinical practice, rare variants are encountered fairly often. These variants are observed in genes that are suspected because of their functional properties, possibly involved in a causal pathway, or based on previously identified associations. The number of rare variants present in a population is very high. Simple arithmetic shows that under neutral conditions (no selection pressure) the contribution to the mutational load of each breeding generation is the same after the point of coalescence, where we have a single ancestor (for most human DNA some 100-500 generations ago). The hypothesis that these commonly occurring rare variants are involved in complex diseases has some supporting evidence. Two of the early studies were described by Frikke-Schmidt and colleagues (Frikke-Schmidt et al., 2004) and Cohen and colleagues (Cohen et al., 2004). They sequenced the ABCAl gene, which is known to cause a Mendelian form of low HDL-cholesterol levels (Tangier disease), in individuals taken from the lower and upper 1-5% tails in a population-based study. Non-synonymous rare sequence variants were found to be significantly more common in individuals with low HDL-cholesterol than in those with high HDL­ cholesterol, leading to major phenotypic effects being proved to contribute significantly to low plasma HDL-cholesterol levels in the general population. These rare DNA variants could not be detected by GWAS, but needed extensive, direct DNA sequence analysis.

77 Chapter 4

How to identifythese rare disease-associated variants

Next Generation Sequencing Over the past few years, large-scale sequencing has been revolutionized by the development of several so-called next-generation sequencing (NGS) technologies. The major strengths of NGS is the fact that tens of millions PCR products can be sequenced in parallel, generating several hundreds of megabases to tens of gigabases of sequencing data per run and making it possible to sequence entire human genomes. Several studies have completed whole human genome sequencing and resulted in important scientific findings and new insights into human genetic variations (Levy et al., 2007; Wheeler et al., 2008; Mckernan et al., 2009; Pushkarev et al., 2009). For complex disorders, including HSCR, application of NGS has still not been reported. It is definitely a very promising approach to find rare coding or non-coding variants in identified and in other, unknown genes. In spite of these advantages, one could also argue that since genome-wide sequencing generates a huge amount of information, it might be hard - even almost impossible - to extract the relevant information. This is especially the case because the specificity is still not optimal, although the sensitivity for detecting mutations is very high, because most deviations from the reference sequence in humans amount to false-positives (Bloksberg, 2008). Undoubtedly this will improve over time as artefacts due to DNA processing are better recognized and eliminated. In addition, the effects of variations in non-coding regions are still difficult to explain and these regions are more likely to have neutral or weak effects on phenotypes, even in well conserved non-coding sequences (Ahituv et al., 2007; Chen et al., 2007). An alternative approach is to sequence all protein-coding regions, also called exome sequencing. This approach has been successfully applied for the detection of recessive (Ng et al., 2009, 2010) and de novo mutations (Hoischen et al., 2010). With this alternative approach, only 5% of sequencing results will be generated when compared to whole genome sequencing. Furthermore, it detects not only changes in protein-coding sequences, but also splice acceptor and donor sites;

78 HSCR as a complex disorder these represent an additional class of sequences that are enriched for highly functional variation and are also targeted by this method (Ng et al., 2010).

Filtering next generation sequencing data for candidate genes If exome sequencing is the method of choice, the number of non-synonymous coding and flanking intronic rare variants that will be detected will be very large (thousands of such variants can be expected, see for instance Ng et al., 2009, 2010). A first step in the selection procedure might be a statistical one. One could determine which genes contain more mutations in cases than in controls (for an example of such an approach see Johansen et al., 2010). The power to detect an excess of rare variants depends on differences in the mutational load of cases and controls. We assume that cases and controls are sampled under the null hypothesis that they come from the same population. The power of such studies is high. For instance, when analysing two groups of 500 people for a stretch of 100,000 bases, and a time to coalescence of 1000 generations, we expect the number of mutations seen in the controls to be : E(mutations) = mutation rate per generation per base x number of generations to coalescence x number of people observed. Being (#) the number of people observed and (") the number of cases detected: E(mutations) = 2.5X10-8 # x 8 ,, 10 x 500 = 1250. If we assume that the mutation rate in the descent tree for cases 8 for this particular stretch of bases is 10% higher (2.75x10- ), then we would expect to see 1375 mutations in the case group. The excess of 125 mutations is 3.5 X the expected standard deviation using a Poisson model and indicates that there is substantial power to detect such increases. Variants with a higher relative risk will improve the power of such a study, due to the possibility of linking cases through haplotype sharing. It would be interesting in such kind of studies to make a comparison between the expression levels of the highly mutated genes in cases and controls. One would expect to find loss-of-function mutations, which is consistent with lower expression of the highly mutated genes (in cases).

79 Chapter 4

Another option to reduce the number of variants is to focus on mutations in candidate disease genes. As complex diseases are probably caused by mutations in genes that participate in specific protein-networks, identification and analysis of the gene-networks is essential in understanding the genetic aspects of a complex disease. This idea, that disease-associated proteins are interconnected, can be nicely illustrated by the proteins involved in HSCR. In Box 3 we show that for HSCR almost all the disease-associated proteins interact which each other directly or indirectly. We therefore believe that identifying and analyzing these protein-clusters is also essential in understanding a complex disease. A systems biology approach, in which protein networks or protein interactions are searched for, using a number of a layers of information (often 'omics' datasets) might be the best choice to select these candidate disease genes. These layers of information could consist of genetic data (association or linkage data), expression data, protein-network data, or proteomics data.

Box 3. Interactions between the known HSCR p.:ithways Several genes have been implicated in HSCR development (Box 2). Although a first look at the proteins encoded by these genes shows no apparent connection between them, the majority of these genes can be divided ,nto three groups: those involved in RET activation p.:ithways (RET, GDNF, NTN), t11ose involved in EDNRB pathways (EDNRB, EDN3, ECE-1), and transcription factors that can affect both RET and/or EDNRB pathways (SOXlO, ZFXHlB, PHOX2B). There is increasing evidence to show that the proteins encoded by these genes are interconnected. A genome-wide association study, using 43 trios from a genetically isolated Mennonite population, reported statistically significant joint transmission of RET .:ind EDNRB alleles (C.:irrasquillo et al., 2002). Furthermore, mice homozygous for the recessive hypomorphic allele of Ednrb (EdnrbS or Piebald) and heterozyr:1ous for a Ret null mutation (Ree ·i showed high frequencies of aganelionosis. These findings are suggestive of epistasis between EDNRB and RET (Carrasquillo et al., 2002), and this idea was fu rther supported by work on RET/EDNRB mouse crosses 1 (McCall ion et al., 2003). Simililrst udies showed that, whereas Ret" ;,: and Edn3;,;i·. (lethal spotting) mice display colonic aganglionosis, combinations of these mutant alleles lead to ulmost complete intest1nul aganglionosis (Barlow el al., 2003).

80 HSCR as a complex disorder

Parallel to these findings, it was observed that GDNF and EDN3 have opposite effects during the development of the enteric nervous system: GDNF promotes the proliferation of migriltory NCSCs and encourages differentic1tion (Hearn et al., 1998; Barlow et al., 2003), while EDN3 has been shown to modulate GDNF activity to inhibit differentiation (Hearn et al., 1998; Wu et al., 1999). Since the effects of EDN3 on both proliferation and chemoattraction could be mimicked by Protin Kinase A (PKA) inhibitors, it was suggested that PKA might serve as u connection link between RET and EDNRB signillling (Barlow et al., 2003). A critical role for PKA was also suggested by the NCSC migration defects and aganglionosis phenotypes observed when the putative PKA-binding site of RET, serine 697, was mutated (Asai et al., 2006). The transcription factor SOXlO is involved in the regulation of both RET and EDNRB expression (Pattyn et al., 1999; Zhu et al., 2004, Heanue und Pachnis, 2007). PHOX28 is also a transcription factor that is expressed in several classes of differentiating neurons of both the peripheral and central nervous systems. Mice with a homozygous disruption of PhoxZb showed no enteric ganglia and no RET expression which suggests that PHOX2B might play a regulatory role in RET expression (Pattyn et al., 1999; Brunet et al., 2002). KBP might also link to the other proteins involved in HSCR development due to its interaction with SCGlO {Alves et al., 2010). SCGlO activity can be controlled by phosphorylation of specific serine residues by PKA (Togano et al., 2005). Possibly PKA is not only the missing link between the RET und the EDNRB pathway, but it might also connect (indirectly) these two pathways to KBP (Fig. 2).

Mesenchyme

ECEl NRTN

Neurobtast

Figure 2 - Scl1ernatic representation of possible interactions between HSCR-related genes.

81 Chapter4

A good example of such a systems biology approach has been described for Parkinson's disease, as explained in Box 4. A similar systems biology approach might provide insights into HSCR disease, pinpointing possible disease-associated genes. In such a study for HSCR we could include: four genomic regions, loci identified by sib­ pair analysis, identity-by-descent (IBD) mapping and linkage analysis in 3p21, 4q31.3- q32.3, 9q31, 16q23 and 19q21 (Bolk et al., 2000; Gabriel et al., 2002; Carrasquillo et al., 2002; Brooks et al., 2005; Brooks et al., 2006); many potential loci as identified by GWAS (on data of the International HSCR Consortium, Garcia-Barcelo et al., 2009); the known protein-networks for RET, EDNRB, KBP, ErbB and the genes whose expression is (partly) regulated by the identified transcription factors (see Box 2); and expression profiling studies of neural crest stem cells isolated from the gut (Heanue and Pachnis, 2006 and Iwashita et al., 2003). Further analysis of these expression datasets and comparable datasets currently being generated with new bioinformatic tools such as the TSR-finder (transcriptional system regulation) might reveal new and as yet unknown interactions. TSR-finder, a bioinformatics tool developed in-house (Fehrmann et al., 2008) is based on the correlation structure of expression variation in over 17,000 published arrays. It can detect clusters of genes and thereby identify systematic changes of expression in gene-clusters with a minimal, but consistently, low expression.

How to prove the pathogenic nature of the presumed disease associated rare variants? When the best candidate genes are selected, one can start analysing the identified rare variants in these genes both by in silica methods and by functional assays. In silica predictions can be applied in a disease- and gene-independent manner. The functional assays, on the other hand, are gene- and disease-dependent. Below we will briefly discuss possible approaches to study newly identified, mutant HSCR genes.

82 HSCR as a complex disorder

Box 4. Systems biology helps in understanding Parkinson's disease A successful example of this type of approach came from the study of Parkinson's disease (PD). Previous studies had reported that genetic variability in the axon guidance pathway might be a contributing factor to the cause of PD (Livesey and Hunt, 1997). Furthermore, GWAS had identified the most significantly associated SNP for PD susceptibly in SEMA5A (Maraganore et al., 2005), a gene known to encode a protein that interacts with several other proteins from the axon guidance pathway (Chilton JK, 2006). These two results reinforced the idea that the axonal guidance pathway was involved in the development of PD. To confirm this result, Lesnick et al., employed bioinformatics methods to analyse a whole-genome association dataset for SNPs that were within brain - expressed, axon­ guidance pathway genes (Lesnick et al., 2007). They identified 128 such genes and found 1,460 SNPs based on the SNP dataset in 117 of those genes. By using a multi-stage process to predict PD susceptibility, they found that of the 1,460 SNPs identified, 183 were individually associated to PD susceptibility. Furthermore, by using an expression profiling dataset they found that 13 of these genes were differentially expressed in at least one of the three regions of the brain affected by PD (eight in the substantia nigra, one in the putamen, and four in the caudate) compared to controls (Lesnick et al., 2007).

In silica prediction of newly identifiedrare variants

To verifythe putative pathogenic nature of the identified coding variants (amino acid changes), one can start by checking online (locus-specific) databases and by analysing these variants by in silico prediction programs such as PolyPhen, Sift or Grantham. As coding variants might also disturb or create a splice-site, splice-site prediction programs such as Netgene2, Splicesitefinder, and Splicesite predict should be checked. For the non-coding variants one might look for the disruption or creation of regulatory sequences such as transcription factor binding sites by using, for instance, Matslnspector. As non-coding variants might also disrupt or create spice-sites they should be analysed by the splice-site prediction programs as well.

Functional assays for newly identified HSCR mutated genes

If the in si/ico analysis predicts the coding variant to be possibly pathogenic, the variant should be further analysed using functional assays. These assays can be highly gene-specific but, in the case of HSCR, one might also try a more general approach, such as in vitro differentiation assays using neural crest stem cells (NCSCs), as

83 Chapter 4 described by Metzger et al. (2009). It would be even better to determine whether the new (mutated) HSCR candidate genes are involved in gut colonization by NCSCs, by for instance grafting of NCSCs carrying the mutation on aganglionic explants. All these experiments should, of course, be carried out by comparing wild-type and mutant gene constructs.

If the in silica assays predict that a non-coding variant disturbs a transcription factor binding site, several assays can be performed to determine the effect of the identified variant on gene expression. Promoter studies could be performed to determine the effect of the mutation on promotor activity. Electrophoresis mobility shift assays could be used to study the potential effect of transcription factor binding, complemented by supershift assays if the transcription factor binding to the identified sequence element can be predicted. Alternatively, if no functional supershift antibodies are available, DNA pull-down assays could be performed.

If the outcomes are still in favour of an involvement in disease development, these assays might be followed by the development of animal models, which would give greater insight into how the newly identified genes contribute to the disease's development (Telslovich et al., 2010).

Conclusion

Unravelling the genetics of complex, polygenic diseases is proving extremely difficult. Combinations of common and rare variants probably contribute to disease development and the identification of the disease-associated variants is far from easy. We have outlined a way to find rare disease-associated variants, making use of a statistical selection procedure followed by a systems biology approach, in combination with next-generation exome sequencing. To validate the findings, we discussed the need for performing functional analysis of the variants in the selected genes.

84 HSCR as a complex disorder

Acknowledgments The authors would thank to thank Jackie Senior for editing the manuscript.

85 86 PART II

Chapter 5

Genome-wide kinome profiling of the RET oncogene identifies MEN2A-, MEN2B- and FMTC-specific signalling pathways

Maria M. Alves, Kaushal Parikh, Lydia Visser, Esther Kleibeuker, Christine van der Werf, Saskia Berndt, Jan Osinga, Maikel P. Peppelenbosch, Bart J.L. Eggen, Robert M.W. Hofstra

Submitted Chapter 5

Abstract RET (REarranged during Transfection) is a proto-oncogene that encodes a receptor tyrosine kinase, which is expressed in neural crest-derived cell lineages. Specific RET mutations result in a ligand-independent activation of the encoded RET protein and cause the inherited cancer syndromes Multiple Endocrine Neoplasia types 2A and 2B (MEN2A, MEN2B) and Familial Medullary Thyroid Carcinoma (FMTC). As different RET mutations result in different disease phenotypes, we hypothesize that these mutations have different effects on downstream signalling pathways. To test this hypothesis, HEK293 cell lines stably expressing RET wild type and the most common mutations found in MEN2A (RET609, RET620, RET634), MEN2B (RET918) and FMTC (RET791, RET891) under control of the Tet-On system, were generated and analysed on a peptide array consisting of 1024 different kinase substrates. Our results showed that there are significant differences between cells expressing RET mutants and wild­ type RET. All mutants showed activation of the canonical RET signalling pathways via Pl3 Kinase/Akt, MAP kinase and the Rae pathways. However, we were also able to identify specific alterations for some of these mutants. We saw an increase in FAK phosphorylation levels for the RET918 mutant, and detected a decrease in receptor internalization for RET891. We also observed that mutations associated with the same phenotype exhibit different signalling properties, as small differences in activation levels of shared pathways were detected. In conclusion, our results showed that differences in the signalling pathways between wild-type RET and different types of MEN2 do exist, confirming the hypothesis that different RET mutations have specific effects on downstream signalling pathways. Furthermore, it also shows that differences in clinical phenotype are not only due to differential activation of canonical RET signalling pathways, revealing a hitherto unrecognised role of non­ canonical RET signalling in clinical phenotype.

Keywords: MEN2, RET, mutations, kinome profile, signalling pathways

90 Kinome profiling of RET wild-type vs. RET mutants

Introduction RET (REarranged during Transfection) is a proto-oncogene that encodes a receptor tyrosine kinase which is expressed in neural crest-derived cell lineages (Marcos and Pachnis, 1996). RET plays a central role in several intracellular signalling cascades that regulate cellular survival, differentiation,proliferation, migration, and chemotaxis (de Groot et al., 2006). These cascades are initiated upon RET activation, a process that is unique among receptor tyrosine kinases as it requires a multi-protein complex instead of common, direct receptor/ligand activation (Plaza-Menacho et al., 2006). The RET activation complex is formed in two steps. The first step starts when a soluble ligand of the glial cell-line derived neurotrophic factor (GDNF) family, (GDNF or one of its homologues, neurturin, artemin or persephin), binds to a glycosyl­ phosphatidylinositol membrane anchored co-receptor of the GDNF family receptor (GFRa 1-4) (Jing et al., 1996; Airaksinen et al., 1999). The second step consists of interaction of this complex with RET, leading to its autophosphorylation on specific tyrosine residues (Manie at al., 2001). The complex formation and activation takes place in the lipid rafts (Manie et al., 2001). Specific mutations in RET result in a ligand-independent activation of the encoded RET protein, giving rise to the development of the cancer syndrome Multiple Endocrine Neoplasia type 2 (MEN2). MEN2 is a dominantly inherited cancer syndrome that can be divided into three subgroups: MEN2A, MEN2B and Familial Medullary Thyroid Carcinoma (FMTC), based on the clinical features and tissues affected (Hansford et al., 2000). Activating mutations can occur in distinct domains of the RET protein and, depending on their location, they are associated with either MEN2A, MEN2B or FMTC. Mutations located in the cysteine-rich domain of the extracellular part of RET (codons 611, 618, 620, 630 and 634), are generally involved in the development of MEN2A. Mutations located in the intracellular tyrosine kinase domain are associated with the development of MEN2B (codons 883 and 918) or of FMTC (codons 768, 790, 791, 804 and 891).

91 Chapter 5

The observation that specific RET mutations result in different disease phenotypes, might be explained by alterations in signalling through diverse downstream pathways. Moreover, these altered signalling profiles might be shared between mutations associated with the same phenotype. If true, identifying the specific signalling pathways associated to different RET mutations and consequently to different phenotypes, can provide important clues to the design specific kinase inhibitors for treating MEN2 subtypes. Previous studies have already explored this effect showing that there are indeed differences in the downstream signalling via activation of these mutants (Iwashita et al., 1999; Schuringa et al., 2001; Plaza-Menacho et al., 2005; Plaza­ Menacho et al., 2007). To further investigate this hypothesis we looked for protein phosphorylation profiles specific for the most common mutations found in MEN2A (RET609, RET620, RET634), MEN2B (RET918) and FMTC (RET791, RET891), using a peptide array consisting of 1024 different kinase substrates. We have previously shown that the kinome profiling is a powerful tool to generate descriptions of cellular signalling and to elucidate the pathways involved in different pathogenesis (Parikh et al., 2010; Parikh et al., 2009; Diks et al., 2004). With this approach, we aim to find a more significant relationship between specific mutations and different MEN2 phenotypes. Moreover, we will try to explain these differences based on downstream signalling pathways, providing new insights into the RET signalling network.

Materials and Methods

Expression vectors pTRE2hyg-RETWT was generated by PCR amplification of RETWT from pRC-CMV­ RETWT (Plaza-Menacho et al., 2005) using the primers RETwtF and RETwtR (Table 1). fcoRV-Notl restriction sites included respectively in RETwt forward and RETwt reverse

92 Kinome profiling of RET wild-type vs. RET mutants primers (underlined sequences), were used to clone RETWT into pTRE2-hyg vector (ClonTech). pTRE2hyg-RETC609Y, pTRE2hyg-RETC620R, pTRE2hyg-RETC634R, pTRE2hyg­ RETY791F, pTRE2-RETS891A and pTRE2hyg-RETM918T were generated from the pTRE2hyg-RETWT vector by site-directed mutagenesis according to the manufacturer's instructions (QuickChange site-directed mutagenesis kit, Stratagene) using the primers listed in Table 1. Following mutagenesis, the entire RET insert was checked by sequencing.

Primer name Sequence RETwtF CATGTAGCGGCCGCATGGCGAAGGCGACGTCCGG RETwtR CGCAGATATCGGACAGCGGTGCTAGAATCTAGT RET609F CTGGCTATGGCACCTACAACTGCTTCCCTGAG RET609R CTCAGGGAAGCAGTTGTAGGTGCCATAGCCAG RET620F GAGGAGAAGTGCTTCCGCGAGCCCGAAGAC RET620R GTCTTCGGGCTCGCGGAAGCACTTCTCCTC RET634F CCACTGTGCGACGAGCTGCGCCGCACGGTGATCGCAGCC RET634R GGCTGCGATCACCGTGCGGCGCAGCTCGTCGCACAGTGG RET791F CACATGTCATCAAATTGTTTGGGGCCTGCAGCCAGGA RET791R TCCTGGCTGCAGGCCCCAAACAATTTGATGACATGTG RET891F CGGAAGATGAAGATTGCGGATTTCGGCTTGTC RET891R GACAAGCCGAAATCCGCAATCTTCATCTTCCG RET918F CGGATTCCAGTTAAATGGACGGCAATTGAATCCCTTTTTG RET918R CAAAAAGGGATTCAATTGCCGTCCATTTAACTGGAATCCG

Table 1. List of primers.

Cell culture and transfection HEK293-Tet-On cells (ClonTech) were cultured in DMEM high medium containing 4.5 g/L of glucose, L-glutamine and pyruvate (Gibco), supplemented with 10% fetal calf

93 Chapter 5 serum (Tet-On system approved FBS, ClonTech), 1% penicillin/streptomycin (Gibco) and 100 µg/ml Geneticin (Gibco). To generated HEK-Tet-On RET stable cell lines, 1.500.000 HEK293-Tet-On cells were seeded in 10 cm dishes (Greiner) and transfected with 5 µg RET wild-type, RETC609Y, RETC620R, RETC634R, RETY791F, RETS891A and RETM918T expression vectors, using jetPei (Promega). 24 hours after transfection, cells were washed and cultured in HEK293-Tet-On medium containing 100 µg/ml of the selection agent, hygromycin (Gibco). Massive cell death was observed 5 days after selection and 10 days after selection, clones were picked and screened for the presence of RET by Western blotting. All cells were cultured at 37°C and 5% CO2.

Stimulations, cell lysates, and Western blot analysis Before lysis, cells were incubated for 24 hours with 1 µg/ml doxycicline (ClonTech) after which they were stimulated for 20 minutes with GDNF (20 ng/ml, Prepotech) and GFRa 1 (100 ng/ml, R&D systems) in serum free medium. Stimulations were terminated by an ice-cold phosphate buffered saline wash and cell lysates were prepared by incubation with lysis buffer (m-Per, Thermo Scientific) containing protease (Thermo Scientific) and phosphatase (Thermo Scientific) inhibitors, for 30 min on ice. Cell lysates were collected by scraping and cleared by centrifugation at 14,000 rpm for 10 min in a pre-cooled (4°C) centrifuge. Supernatants were stored at - 80°C before they were processed further for SOS-PAGE followed by Western blot analysis. The following antibodies were used: RET (H-300), p-RET (Y1062) (Santa Cruz Biotechnology); p-ERK, p-JNK, p-p38, p-mTOR, p-p70S6 kinase, p-PDK1, p-Akt, p-PAK2, p-FAK (Cell Signaling); and actin (C4, MP Biomedicals). Secondary antibodies used were goat anti-rabbit lgG-HRP and goat anti-mouse lgG-HRP (Bio-Rad).

Proliferation assays HEK-Tet-On RETWT, HEK-Tet-On RET609, HEK-Tet-On RET620, HEK-Tet-On RET634, HEK-Tet-On RET791, HEK-Tet-On RET891 and HEK-Tet-On RET918 were seeded in 96-

94 Kinome profiling of RET wild-type vs. RET mutants well plates (10,000 cells/well) containing complete medium used for culture of HEK­ Tet-On RET cells (previously described). Doxycycline was added 24h after and cells were incubated for 48 and 96 hours. Cell proliferation was measured using a cell proliferation kit (MTTassay, Roche) according to the manufacturer's instructions.

Softagar colony formation assays 2500 cells of HEK-Tet-On and HEK-Tet-On RET mutants, RET609, RET620, RET634, RET918, and RETWT, were mixed with 1.5 ml of 0.7% agarose (Eurogentec) and 1.5 ml of DMEM containing 20% FBS, 2 µg/ml doxycycline, 200 ng/ml of GFRa. 1 and 40 ng/ml of GDNF. 1ml of this mixture was platted in triplicate in 6-well plates previously coated with 1 ml of 0.5% agar (Becton, Dickinson (BD)) and DMEM medium containing 20% FBS. Cells were fed twice a week with 500 µI of culture medium and incubated for 25 days in an incubator at 37°C and 5% CO2. Each well was stained with 0.5 ml of 0.005% Crystal Violet for 2 hours and the cells were analysed in a stereo microscope (Leica-MZ75).

Kinome array analysis The protocol of the kinome array is described in detail on the website (http://www.pepscan presto.com/files/PepCh ip%20Kinase%20Lysate%20Protocol_vs. pdf).

Data acquisition and statistical analysis of the PepChip array After drying, the glass slides were exposed to a phosphor-imager plate for 72 hours. Acquisition of the peptide array was performed using a phosphor-imager (Storm; GE Healthcare). The level of incorporated radioactivity, which corresponds to the phosphorylation status, was quantified by array software (ScanAlyze; Eisen Software) and as described elsewhere (Diks et al., 2004; Diks et al., 2007; Lowenberg et al., 2005; van Baal et al., 2006). Different kinase activities of the lysates from the RET wild-type, RET609, RET620, RET634, RET791, RET891 and RET918 cell lines were

95 Chapter 5 determined by significant fold change ratios of the combined values of phosphorylated peptides, as previously described (Parikh et al., 2009; Fuhler et al., 2010).

Flow cytometryanalysis After doxycycline incubation and stimulation with GDNF and GFRa 1, HEK293-Tet-On cells stably expressing RETWT, RET791 and RET891 were collected by resuspension in PBS and transferred to a FACs tube. Cells were washed with PBS containing 1% BSA and centrifuged for 5 min at 1000 rpm in a pre-cooled centrifuge. An antibody recognizing the extracellular domain of RET (H-300) was added (1:50) and cells were incubated for 1 hour at 4°C. Cells were washed twice and incubated for 30 min with a goat a- rabbit FITC secondary antibody (Jackson Labs). Cells were washed twice and fixed with 4% paraformaldehyde. Flow cytometry was performed using the FACS Calibur (BD Biosciences), and data was analysed using Flowjo software.

Results

Propertiesof the different RET mutants To study the signalling properties of the different RET mutations, we used a system in which RET expression can be induced by doxycycline. As the host cell line, we used Human Embryonic Kidney cells (HEK293). Since HEK293 does not endogenously express RET, the use of this cell line provided us with a clean background to study the effect of each mutation on downstream pathways without interference of the endogenous protein. Seven double-stable cell lines were created expressing the most common MEN2A (RET C609Y, C620R, C634R), MEN2B (RET M918T) and FMTC (RET Y791F, S891A) mutations. To analyse the levels of RET expression and RET activation of the different mutants, cells were incubated with doxycycline in order to induce RET expression. After 24 hours, GDNF and GFRa were added and cells were incubated for 20 minutes.

96 Kinome profiling of RET wild-type vs. RET mutants

This time point was chosen because RET has been reported to reach a peak in phosphorylation between 15-30 min after GDNF/GFRa.stimulation (Richardson et al., 2006). Doxycycline induced RET expression and RET phosphorylation on tyrosine 1062 in all clones. However, in some clones, RET was expressed in the absence of doxycycline (Fig. lA). RET mutants harbouring the same type of mutation with similar levels of RET expression differed in their levels of activated RET. For MEN2A mutants, RET609 displayed the lowest levels of RET phosphorylation, while RET620 and RET634 showed similar, high levels of RET phosphorylation. For FMTC mutants, levels of RET and phosphorylated RET were equivalent for RET791 and RET891, and they were the mutants showing the lower levels of RET expression and RET phosphorylation. For MEN2B mutation RET918, RET expression levels were similar to MEN2A mutants, however, the level of Y1062 phosphorylation was higher for the MEN2B mutant (Fig. lA). Next, the effect of mutant RET expression on the cellular proliferation rate was determined. After two days of doxycycline induction almost no change was seen between the mutants and the wild-type cell lines. However, after four days both the MEN2A and the MEN2B mutants showed higher proliferation rates compared to RET wild-type (RETWT). For the FMTC mutants, there was almost no increase in proliferation after four days' incubation (Fig. 1B). To measure the capacity of these cells to proliferate, anchorage-independent, soft agar assays were performed. Cells were cultured for 25 days and analysed for the presence of foci. Interestingly, the highest number of foci were observed for the most aggressive RET mutations (de Groot et al., 2006) (RET609, RET620, RET634 and RET918) (Fig. le). For FMTC mutants, although the number of foci was higher that of wild-type RET expressing cells, fewer colonies were observed compared to MEN2A or MEN2B mutants (Fig. le). To summarize, RET mutants harbouring a MEN2B or a MEN2A mutation have higher RET phosphorylation levels, increased cell proliferation, and anchorage­ independent growth properties compared to FMTC mutants and wild-type RET. These

97 Chapter 5

results are in agreement with the disease phenotype associated with the different mutations in which the most aggressive phenotype is observed for MEN2B and MEN2A mutations.

A) MEN2A FMTC MEN2B 609 620 634 791 891 918 WT Tet on + + + + + + + + RET --- P-RET Actin

B) MJTassiy

OB •2d1','I 061,2 84davi 0,4 0�

609 620 634 791 891 511 W1 T1111,0tnu C) 620 791 . .,.... .,, ... • 114 • ,.•• ' . . •- • • Tet on • •

Figure 1-Different propertiesof RETTet-On stable mutants. A) Western blot analysis of the HEK293-Tet-On RET stable cell lines expressing RET wild-type (WT) and six different RET mutations involved in the development of MEN2A (RET609, RET620, RET634), FMTC (RET791, RET891) and MEN2B (RET918) shows that RET is expressed at higher levels when doxycycline is present (+). Phosphorylated RET is almost exclusively detected in the presence of doxycycline. (B) Cell proliferation of HEK293Tet-On RETWT and the different mutants was analysed after stimulation with GDNF, GFRa and doxycycline for two and four days. After two days of culturing almost no change in proliferation was seen between the mutants and the WT cell lines. However, after four days both MEN2A and MEN2B mutants showed higher proliferation rates compared to the WT. For the FMTC mutants, there was almost no increase in proliferation after four days of incubation. C) Soft agar assays performed for the different RET expressing cell lines, showed that the MEN2A and the MEN2B mutants are the ones with stronger invasive properties in comparison with the FMTC mutants and RETWT. Hek293-Tet-On cells were used as a negative control, since they do not endogenously express RET.

98

791 Kinome profiling of RET wild-type vs. RET mutants

MEN2A, MEN2B and FMTC RET mutants show distinct kinome profiles that are different from the wild-type RET. To gain insight into the differences in signal transduction properties of mutant RET receptors, the kinome of cells expressing different RET mutations, MEN2A mutants (RET609, RET620, RET634), MEN2B mutant (RET918) and FMTC mutants (RET791, RET891) was determined. We analysed all cell lines in the presence (+) and absence (-) of doxycycline (Fig. lA). Each kinome was compared with cell lines expressing wild­ type RET. All cell lines were activated with GDNF and GFRa. The profiles were generated by incubating each cell lysate onto an array exhibiting 1024 different kinase substrates spotted in triplicate, with 33P-gamma-ATP. In each experiment, the amount of protein incubated in each array was corrected for relative RET expression based on the Western blotting analysis (Fig. lA), in such a way that the amount of RET protein used for the arrays was nearly the same for all the different cell lines. The arrays incorporated substantial amounts of radioactivity and in figure 2A, a representative phosphor-imager scan of the generated kinome profiles for RET918 in the presence and absence of doxycycline is shown. Kinome profiles were generated in duplicate, with the average Pearson product moment correlation coefficient between the replicates greater than 0.85 in all experimental groups. From qualitative comparison of the four experimental groups, it appeared that the profiles of the mutants were substantially different from wild-type RET expressing controls, with correlation values not exceeding 0.4 (Fig. 2B). On the other hand, the profiles of the mutant RET cells grouped together were surprisingly similar, with FMTC and MEN2B being the most different from each other (Fig. 2B). Clustering analysis of the kinome profiles according to Johnson (Johnson MS, 1967) confirmed this trend, with all the mutants clustering together and away from the wild-type receptor (Fig. 2C). Additionally, for tyrosine phosphorylation we determined patterns of the different mutants by Western blotting using a general phosphotyrosine antibody (p­ Y). We observed that, although there are many bands in common between the

99 Chapter 5

mutants and the wild-type samples, each mutation has its own specific pattern of phosphorylation, supporting the idea that different mutations have different signalling properties resulting in specific phosphorylation patterns that are distinctfrom the wild-type receptor (Fig. 2D).

A) B)

R' ,.038 JOO R1 •0.2l JOO ,.. ,.. ,00 ,oo

i 1,0 • 1,0 . ,.-·n. I .,, - ,.; ' t-, 100 •;.�.. •. ... . ' � t_·_.:�,:._t.�_: -; w 50 �·.J ,.1;��-� ,:=.,;:i �·.r,,.'!"r- -�- • I •:�:� ;i'� -�'::! -:·c ·�, 100 200 JOO 100 200 lOO �·",;h:....;.,. ..I..J :�-J 100 ,00 . MEN1A (609920.al'I MEN28 (918) FMTC!89t 191) -.. :.;... ;,.-....1 ' •"" -- -­

! --1• •! I 300 1 300 R'""'/J.79 R 77 ,50 1-::·•I:ri" -i.:,. ·-··1, 250 . �: .:•. 916- 918t

50

100 200 300 100 � .>00 100 200 300 MEN28 (11111) MEN2A(609.620.634) MEN2A (609,820 63-4)

C) D) 609 620 634 791 891 918 WT Tet on + - + - +

p-Y

Figure 2 - Kinome profile of RET wild-type and different RET mutants. A) Representative phosphor-imager scans of the peptide arrays after incubation with lysates of HEK293-Tet-On RET918 in the presence (+) and absence (·) of doxycycline. B) Qualitative comparison of the kinome profiles of the four experimental groups studied showed substantial differences between the mutants and wild-type RET, with correlation values not exceeding 0.4. On the other hand, the profiles of the different RET mutants were surprisingly similar, with FMTC and MEN2B being the most different from each other, and MEN2A and FMTC the most similar. C) Clustering of kinome profiles reveals that MEN2A and FMTC mutants are the ones with a higher similarity to each other, while MEN28 presented a more distinct profile. D) Western blot analysis of the phosphorylation status of the different mutants in the presence and absence of doxycycline, using a p-Y antibody showed clear differences in the phosphorylation patterns when RET expression is switch on. Furthermore, different RET mutants showed different phosphorylation patterns in comparison to other mutants and to the wild-type.

100 Kinome profiling of RET wild-type vs. RET mutants

Extracellular mutations lead to activation of the Map kinase, Pl3 kinase and RAC pathways To gain insight into the signalling pathways activated by the RET receptor containing extracellular mutations (RET 609, RET620, RET634), we first compared the kinome of these mutant cell lines with the wild-type RET kinome in the presence (+) and absence (-) of doxycycline. Significant changes in phosphorylation of 37 peptides specific for RET609+ were observed after correcting for phosphorylation changes in the RET609-, RETWT- and RETWT+ (Fig. 3A). Only nine peptides showed changes in phosphorylation both in the RETWT+ and RET609+ cells (after correction for changes in RETWT- and RET609-). Similarly, the RET620+ showed 39 differentially phosphorylated peptides as compared to RETWT-, RETWT+ and RET620- (Fig. 3A). RET620+ and RETWT+ only shared eight peptides in common. Finally, RET634+ showed unique phosphorylation of 37 peptides with nine peptides shared with the RETWT + (Fig. 3A). One of the most prominent effects observed with all MEN2A mutants was the consistent phosphorylation changes in MAP kinase, Phosphoinositide 3-kinase (Pl3 kinase) and Rae pathways. A list of peptides can be found in Supplementary table 1. The activity of a number of altered kinases involved in these pathways was verified by Western blotting. Consistent with the kinome array data, the Western blot analysis showed an increased phosphorylation of p38 and ERK kinases for the RET620 and RET634 mutants compared with RETWT (Fig. 3B, 3E). For the RET609 mutant, although there was an increase in ERK phosphorylation, no effect was observed on p38 kinase, as predicted by the kinome arrays (Fig. 3B, 3E). The kinome arrays also indicated a significant change in the Pl3 kinase pathway. By Western blotting we observed that the RET609 mutant presented an increase in phosphorylation of members of this pathway, namely Akt, PDKl and a slight increase in mTOR. The RET620 and the RET634 mutants also showed an increase in the levels of PDKl and mTOR phosphorylation and a slight increase on Akt and p70S6 kinase compared to the wild-type (Fig. 3C, 3E). Differential activation of the Rae pathway was also observed on the kinome arrays and Western blot analysis of its downstream substrate p21

101 Chapter 5 activated kinase 2 (PAK2) confirmed these observations. Both RET620 and RET634 showed an increase in phosphorylation for PAK2, while RET609 showed a decrease in PAK2 phosphorylation (Fig. 3D, 3E).

A) B) Tet on MEN2A609 609 620 634 WT +

J7 p-p38

13 39 p-ERK

Actin D)

+ + + + + ... 609 620 634 WT Tot on

p-PAK2 C) Actin E) 609 620 634 WT Tet on "' p-Akt ,. aWT Gf.U p·PDKl C!lO an.i p-mTOR artto"

p-p70

Actm

Figure 3 - Signalling pathways associated with MEN2A mutations. A) Venn diagrams showing similarities and differences (in numbers of peptides) in the kinome profile of different MEN2A mutations: RET609, RET620 and RET634. B) Western blot results confirmed activation of the Map kinase pathway, C) the Pl3 kinase pathway, and D) the Rae pathway for all the MEN2A mutants in comparison to RETWT. E) Semi-quantification of the band intensity corresponding to individual signalling proteins as determined with lmageJ software. The intensity value obtained for each band was divided by the intensity of the �-actin band from the same lysate to obtain a ratio.

The ratio for each protein was then subtracted from the ratio obtained when no doxycycline stimulation was present. This value was compared between the different mutants and the wild-type. The figure shows a trend for increased phosphorylation levels for all the mutants compared to the wild-type. HEK-Tet-On cells were used as a negative control as they do not express RET. (+) and (-) correspond respectively to the presence or absence of doxycycline stimulation.

MEN2B mutation leads to a specific activation of FAK The kinome profile of the most aggressive mutation, RET M918T, is qualitatively different from the kinome of the wild-type RET in the presence (+) and in the absence

102 Kinome profiling of RET wild-type vs. RET mutants

(-) of doxycycline, as expected. Significant changes in phosphorylation of 51 peptides were identified for the RET918+ after correction for phosphorylation changes in the RET918-, RETWT- and RETWT+. Moreover, only six peptides showed changes in phosphorylation in both RETWT+ and RET918+ cells (Fig. 4A). Similar to the MEN2A mutants, phosphorylation changes in the MAP kinase, the Pl3 kinase and the Rae pathways were consistently observed for the RET918+ mutant. Moreover, a change in phosphorylation of the Focal Adhesion Kinase (FAK) and Aurora Kinase B were also seen for the MEN2B mutant (Supplementary table 2). FAK is involved in cell migration cell adhesion and proliferation (Cai et al., 2008), and Aurora Kinase B is a serine/threonine kinase required for chromosome segregation and cytokinesis (Mountzios et al., 2008).

A) B) MEN2B 918 WT Te ton + + + p-p38

27 p-ERK

p-JNK

C) 918 WT Te ton + • + • + D) WT Te ton p-Akt + + 918 -- p-PAKZ

Actin F) _ ...... p-p70

Actin

E) •wr 918 WT Tet on D918 Dltton

p·pl8 p-DIK ,-.IHl p-Akt p-POKl p-n1T111 p-pll ,-,AJJ •-,u:

Figure 4 - Signalling pathways associated with MENZB mutations. A) Venn diagrams showing similarities and differences (in number of peptides) in the kinome profile of the MEN2B mutant, RET918, compared to the wild­ type. B) Western blot results confirmed activation of the Map kinase pathway, C) the Pl3 kinase pathway, and D) the Rae pathway for the MEN2B mutant in comparison to RETWT. E) Specific increases in the levels of phosphorylated FAK on tyrosine 397 {Y397) were confirmed by Western blot for RET918. F) Semi-quantification

of the band intensity corresponding to individual signalling proteins as determined with lmageJ software. The

103 Chapter 5 intensity value obtained for each band was divided by the intensity of the 13-actin band from the same lysate to obtain a ratio. The ratio for each protein was subtracted from the ratio obtained when no doxycycline stimulation was present. This value was compared between the different mutants and the wild-type. The figure shows a trend for increased phosphorylation levels for all the mutants compared to the wild-type. HEK-Tet-On cells were used as a negative control as they do not express RET. (+) and (-) correspond respectively to the presence or absence of doxycycline stimulation.

By Western blotting we were able to confirm that the phosphorylation levels of the MAP kinases p38, ERK and JNK were up-regulated in the RET918 mutant compared to the RETWT (Fig. 4B, 4F). An increase in phosphorylation of the kinases involved in the Pl3 kinase pathway (Akt, PDKl, mTOR and p7056 kinase) and the Rae pathway (PAK2) was also observed (Fig. 4C, 4D, 4F). A change in the levels of phosphorylated FAK was exclusively detected for the RET918+ and this change was confirmed by Western blotting after comparison to RETWT (Fig. 4E, 4F).

FMTC mutant RET 891 is more resistant to receptor internalization For the FMTC mutants, RET Y791F and RET 5891A, 18 and 26 peptides were phosphorylated exclusively for the RET791+ and RET891 + respectively, after corrections for phosphorylation changes in the RET791-, RET891-, RETWT- and RETWT+. RET791+ and RET891+ shared only six and nine peptides respectively, with the RETWT+ (Fig. SA and supplementary table 3). As observed for the MEN2A and MEN2B mutants, a marked change in the MAP kinase, Pl3 kinase and Rae pathways was detected in the kinome arrays of the RET791 and RET891 cell lines. For the MAP kinase pathway, phosphorylation of p38, ERK and JNK was analysed by Western blotting. Phosphorylation levels of these three kinases were up-regulated for both RET791 and RET891 mutants (Fig. SB, SE). For the Pl3 kinase pathway only Akt was altered in the arrays. As observed for the MEN2A and MEN2B mutants, a marked change in the MAP kinase, Pl3 kinase and Rae pathways was detected in the kinome arrays of the RET791 and RET891 cell lines. For the MAP kinase pathway, phosphorylation of p38, ERK and JNK was analysed by Western blotting. Phosphorylation levels of these three

104 Kinome profiling of RET wild-type vs. RET mutants kinases were up-regulated for both RET791 and RET891 mutants (Fig. SB, SE). For the Pl3 kinase pathway only Akt was altered in the arrays. However, since PDKl, mTOR and p7056 kinase are also affected by the activation of Akt, the phosphorylation levels of all these kinases were analysed. We observed that for both RET791 and RET891 mutations, there was an increase in phosphorylation of Akt, PDKl and mTOR compared to the wild-type. However, only the RET891 mutant showed an increase in phosphorylation of the p7056 kinase (Fig. SC, SE). An alteration of the Rae pathway, specifically of PAK2, was also detected in the arrays for both the FMTC mutants. However, we were only able to confirm an up-regulation of PAK2 phosphorylation for the RET791 mutation (Fig. SD, SE).

A) B) 791 891 WT 891 + + + + FMTC Tet on p·p38

p-ERK

p·JNK � 20 � Actin

D) 791 891 WT Tet on + + C) p-PAK2

791 Tet on 891 WT Aton + + + p·AKT E)

Figure S - Signalling pathways associated with FMTC mutations. A) Venn diagrams showing similarities and differences (in numbers of peptides) in the kinome profile of different FMTC mutations: RET791 and RET891. B) Western blot results confirmed activation of the Map kinase pathway, C) the Pl3 kinase pathway, and D) the Rae pathway for the FMTC mutants in comparison to RETWT. E) Semi-quantification of the band intensity corresponding to individual signalling proteins as determined with lmageJ software. The intensity value obtained for each band was divided by the intensity of the �-actin band from the same lysate to obtain a ratio. The ratio for each protein was subtracted from the ratio obtained when no doxycycline stimulation was present. This value

105 Chapter 5 was compared between the different mutants and the wild-type. The figure shows a trend for increased phosphorylation levels for the all the mutants compared to the wild-type. HEK-Tet-On cells were used as a negative control as they do not express RET. (+) and (-) correspond respectively to the presence or absence of doxycycline stimulation.

Previous studies in tyrosine kinase receptors such as EGFR, have shown that aminoacids changes from serine to alanine can lead to an abrogation of ligand­ induced internalization of the receptor, while no effect was observed when tyrosine residues were mutated (Oskvold et al., 2003). Knowing that one of the FMTC mutants has a serine to alanine mutation (RET 5891A), we decided to evaluate the effect of alanine changes on RET receptor tyrosine kinase internalization. The recycling properties of the RET receptor were measured by flow cytometry. We observed that while the wild-type receptor resurfaced twice in one hour, the RET891 mutant only recycled once (Fig. GA and 6B). Interestingly, RET791 containing a tyrosine to phenylalanine alteration, displayed no changes in its recycling pattern compared to the wild-type RET, recycling twice in one hour (Fig. 6A and 6B). This result showed that serine to alanine mutations in the RET receptor lead to a decrease in endocytosis comparable to what is found in the EGF receptors. Furthermore, we observed that RET891 is more resistant to internalization than RET791 and RETWT, and thus stays longer in the membrane.

Discussion Several groups have studied how different RET mutations can give rise to a constitutive active receptor and which signalling pathways are activated by these mutated receptors, in an attempt to find specific signalling pathways that can reveal how different RET mutations result in different disease phenotypes (Iwashita et al., 1999; Schuringa et al., 2001; Plaza-Menacho et al., 2005; Plaza-Menacho et al., 2006). However, most of these studies have focussed on one or a specific set of RET mutations, and almost no comprehensive comparison between the different mutant RET proteins was available. We have therefore compared the most common

106 Kinome profiling of RET wild-type vs. RET mutants mutations found in MEN2A (RET609, RET620, RET634), MEN2B (RET918) and FMTC (RET791, RET891), using an enzymatic approach that can quantify phosphorylation of a wide spectrum of kinase substrates. With this approach, not only activated kinases are revealed, but also active substrates - we refer to this collectively as "the kinome"

(Parikh et al., 2009; Parikh et al., 2010).

A) an ,.,. B) r:t-J-= -•• I 60 aH�= ...� 50

40

D � 3 ,a

10

•::1•-1 ,.,,.,,,. .� 5' 10' 20' JD' 60' !���� ,,/\ ' Tim, (mrn) 1-.!. :;1·;1 .' \\\ i 1--' I,. �-�! \ ·1 1't ··· \.:::>---::- ,,, ' 1 , --, nuunc.,, ",; · ,,114 a.-1...... I.l"El ,,� .,,n. .... �....

Figure 6 - RET891 is more resistant to receptor internalization. A) Mean fluorescence intensity {MFI) values showed that RET891 recycles only once (at 30 min) in 1 hour in contrast to RETWT and RET791 (10 min and 30 min). B) Graphic representation of mean MFI values obtained for RETWT, RET791 and RET891, confirm a reduction in receptor internalization for RET891 compared to RET791 and RETWT.

Analysis of the kinome profile and the phosphorylation levels of each mutant showed that different mutations lead to activation of the canonical RET pathways but also, as expected, to differentactivation patterns of specific kinases compared to RET wild-type. We observed that the RET918 mutant presented the most distinct profile compared to RET wild-type (Fig. 2B, 4A). Interestingly, we also observed that the

107 Chapter 5

MEN2A and the FMTC mutants were highly similar (Fig. 2B). This result is in agreement with the fact that the distinction between MEN2A and FMTC phenotype is somewhat disputed and that FMTC should be considered to be a milder form of MEN2A (Marx, 2005). Previous studies have shown that activation of the RET receptor induces downstream signalling via RAS/ERK kinase, the Pl3 kinase/Akt, the p38 kinase, PLC-y,

JNK and STAT pathways (Santoro et al., 2004, Arighi et al., 2005, Kodama et al., 2005). For all the mutants studied, the kinome analysis revealed an increased activation of this canonical signalling route MAP kinase, Pl3 kinase and Rae pathways. This increase was expected, as the MAP kinase pathway is involved in cell proliferation and differentiation, the Pl3 kinase pathways is involved in cell survival and proliferation, and the Rae pathway is associated with cytoskeletal modifications and cell motility (Fukuda et al., 2002; Kodama et al., 2005). However, not all mutants showed the same levels of phosphorylation of these three pathways. It should be noted that activation of p38, p70S6kinase and PAK2 was less pronounced for RET609, and PAK2 activation was less pronounced for RET891, when compared to the other mutants and to RETWT. These results reveal that mutations associated to the same phenotype do show variations in activation of downstream signalling pathways. The focal adhesion kinase (FAK) is an intracellular tyrosine kinase that has been shown to be an important integrator of adhesion-dependent and growth factor­ dependent signalling (Sieg et al., 2000), and is involved in several cellular responses including cell migration, adhesion and proliferation (Cai et al., 2008). Previous studies have reported a direct effect of wild-type RET on FAK phosphorylation (Murakami et

al., 1999, You et al., 2001). Furthermore, RET was reported to signal through FAK in medullary thyroid cancer cells (Panta et al., 2004). In our study, we showed that FAK activation at tyrosine 397 is specifically up-regulated in MEN2B. Since FAK is involved in cytoskeletal changes that lead to an increase in invasive properties, this result is in agreement with our soft agar assays (Fig. lC) and the more severe phenotype observed for patients with this mutation. The kinome profile of RET918 also identified

108 Kinome profiling of RET wild-type vs. RET mutants a specific up-regulation of the Aurora kinase B, a serine/threonine kinase required for chromosome segregation and cytokinesis (Mountzios et al., 2008). Aurora kinase B was previously detected to be over-expressed in a variety of human tumour types, e.g. endometrial carcinomas (Kurai et al., 2005) and it has been shown to correlate with the level of genomic instability within a tumour, suggesting a role for this kinase in the acquisition of genetic alterations critical for neoplastic transformation (Smith et al., 2005). In addition, Aurora kinase B expression was detected in anaplastic thyroid carcinoma (Wiseman et al., 2007). Although further studies are necessary to confirm a potential effect of Aurora kinase B in the development of MEN2B, the fact that this kinase was significantly up-regulated in our arrays for RET918 could also account for the more severe phenotype seen in MEN2B. Serine to alanine substitutions in EGF tyrosine kinase receptor have been reported to result in an abrogation of ligand-induced internalization of the receptor (Oskvold et al., 2003). Since the S891A FMTC mutant included in this study had a serine to alanine substitution, we investigated the effect of this change on RET internalization. We showed that RET891 is more resistant to internalization and remains on the membrane longer compared to RETWT and RET791. This result showed that serine to alanine mutations in the RET receptor can lead to a decrease in endocytosis comparable to the ones found in the EGF receptors. Furthermore, it can also explain the slight decrease in activation of the MAP kinases (ERK, p38 and JNK) observed for RET891 in comparison with RET791, since receptor internalization was reported to be required for MAP kinase activation (Richardson et al., 2006). A schematic representation summarizing all the changes detected on the kinome arrays for the different mutations included in this study can be seen in Figure 7. This figure shows the differences in the signalling pathways between RET wild-type and different mutants involved in the development of MEN2A, MEN2B and FMTC. Identification of specific signalling pathways for the different mutants can be used to develop specific therapeutic approaches for treatment of MEN2A, MEN2B and FMTC. Our observation that p38 phosphorylation was up-regulated in both

109 Chapter 5

MEN2B and FMTC mutants compared to MEN2A and the wild-type receptor, suggests that p38 might be a good candidate target for the design of specific therapies for treating these subtypes of MEN2. To date, there are no reports of the use of p38 inhibitors in thyroid cancer. However, they have been used for treating inflammatory diseases (Kumar et al. , 2003) as well as cancer (Sun and Sinicrope, 2005), with high selectivity, resulting in almost no side-effects for the patients. In conclusion, our results showed that differences between RET wild-type and different RET mutants do exist that can help explaining the oncogenic phenotypes observed in MEN2. Furthermore, we show that, not only RET mutants associated to different MEN2 subtypes activate specific signalling pathways, but also different mutations responsible for the same phenotype do not necessarily share the same signalling pathways, suggesting that different mutations have their own specific signalling pathways. Taken together, these results yield new insights into the RET protein network that might be helpful in developing specific therapeutic approaches for treating MEN2A, MEN2B and FMTC.

A) 629, 762 731

510, 605, 627, 485, 766, 955, l62, 8ll, 953, 4B8, 115, 950 957

61

114

95. 669,485, 332 887, 413

511, l4) B87

proliferation lamellipodia survival, survival formation• m1grat1on prohferatlon

110 Kinome profiling of RET wild-type vs. RET mutants

----�--.------. 707

I ·- ,I ·- - I I .,,. I I 6» 1015

I I I I I I �--i---!I�------i decreased apoptosis lamellipod1a & increased protein cell cycle prohferatton survival formation . migration synthesis

C)

863, 678, 207 879, 725

2.10

462

678

cell proffleration,cell survival, cell migration mtegrm activation

111 Chapter 5

Figure 7 . Provisional signal transduction schemes of the signalling pathways active in MEN2A, MEN2B and FMTC compared to wild-type. Signalling pathways observed in: A) MEN2A, B) MEN28, and C} FMTC. Blue circles represent specifically activated kinases for each of the MEN2 subgroups and pink circles represent common activated pathways between the different subgroups. Peptide numbers provided correspond to the substrate location on the peptide array. Unmarked circles correspond to kinases/substrates that were not directly identified on the array, but are well known downstream/upstream targets of the respective pathways.

Acknowledgements We would like to thank Jackie Senior for editing the manuscript. This work was supported by the Graduate School of Medical Sciences (GUIDE) and a Jan Kornelis de Cock Stichting grant to M.M.A.

Supplementary tables can be found online by following the appropriate links.

Supplementary table 1: https :// spreadsheets0.google.com/ ccc ?key=t2gK9al2am 1_9a9H R BqOOPQ&a uth key= CMeghKMB&hl=en&authkey=CMeghKMB#gid=0

Supplementary table 2: https://spreadsheets.google.com/ccc?key=0Anl335bprQuCdHVDVIFxSE9NVTVfTXINS DZ3UzhNZ3c&authkey=CLCstYcJ&hl=en#gid=0

Supplementary table 3: https://spreadsheets.google .com/ ccc? key=0An1335 bprQuCd FltcF pFQTFiTH lzM UN mN zZlbUs5bEE&authkey=CLiAh9wl&hl=en#gid=0

112 Chapter 6

The effects of four different tyrosine kinase inhibitors on medullary and papillary thyroid cancer cells

Hans H.G. Verbeek", Maria M. Alves", Jan-Willem B. de Groot, Jan Osinga, John T.M. Plukker, Thera P. Links, and Robert M.W. Hofstra

# These authors contributed equally.

Submitted Chapter 6

Abstract Context: Medullary and papillary thyroid carcinoma (MTC and PTC) are two types of thyroid cancer that can originate from activating mutations or rearrangements in the RET gene. Therapeutic options are limited in recurrent disease but since RET is a tyrosine kinase (TK) receptor involved in cellular growth and proliferation, treatment with a TK-inhibitor might be promising. Several TK-inhibitors have been tested in clinical trials, but it is unknown which inhibitor is most effective and if there is any specificity for particular RET mutations. Objective: We aimed to compare the effect of four TK-inhibitors (axitinib, sunitinib, vandetanib and XL184) on cell proliferation, RET expression and autophosphorylation, and ERK activation in cell lines expressing a MEN2A (MTC-TT) mutation, a MEN2B (MZ-CRC-1) mutation, and a RET/PTC (TPC-1) rearrangement. Design: The three cell lines were cultured and treated with the four TK-inhibitors. Effects on cell proliferation, RET and ERK expression and activation were determined. Results: XL184 and vandetanib, most effectively inhibited cell proliferation, RET autophosphorylation in combination with a reduction of RET expression, and ERK phosphorylation in MTC-TT and MZ-CRC-1 respectively. TPC-1 cells showed a decrease in RET autophosphorylation after treatment with XL184, but no effect was observed on ERK activation. Conclusion: There is indeed specificity for different RET mutations, with XL184 being the most potent inhibitor in MEN2A and PTC, and vandetanib the most effective in MEN2B in vitro. No TK-inhibitor was superior for all the cell lines tested, indicating that mutation-specific therapies could be beneficial in treating MTC and PTC.

Keywords:MTC, PTC, tyrosine kinase inhibitors, XL184, vandetanib.

114 Effectof TK-inhibitors on MTC and PTC cells

Introduction Medullary thyroid carcinoma (MTC) is a rare neuroendocrine tumor which originates from the parafollicular calcitonin-producing cells (C-cells) of the thyroid. MTC usually occurs sporadically (75% of the cases) or as part of a familial cancer syndrome (25% of cases). The familial form occurs as one of the two multiple endocrine neoplasia (MEN) type 2 syndromes (MEN2A and MEN2B) or as familial medullary thyroid cancer (FMTC). These familial syndromes are caused by a germline mutation in RET, while a somatic RET mutation is found in 23-69% of the sporadic MTC patients (de Groot et al., 2006; Hofstra et al., 1994; Eng and Mulligan, 1997). Papillary thyroid carcinoma (PTC) is the most common cause of thyroid cancer, representing approximately 80% of all thyroid cancer cases. PTC occurs in the follicular epithelial cells of the thyroid and exposure to radiation is a major risk factor for developing PTC. A small percentage of PTCs (5-30%) develops due to RET rearrangements, although this percentage is higher - over 60% - in patients exposed to radiation and in children (de Groot et al., 2006; Nikoforov, 2008; Greco et al., 2010). Curative treatment for MTC is surgery and consists of a total thyroidectomy and extensive lymph node dissection, including a central compartment dissection. Patients with sporadic MTC usually present with a palpable solitary thyroid nodule. However, when the initial diagnosis is made, cervical and mediastinal node metastases are already present in 50% of sporadic MTC patients (Maley and Fialkowski, 2007). Recurrent disease is often seen in sporadic patients and until recently, re-operation was the only option, because no effective adjuvant therapies were available (Kloos et al., 2009). Young MEN2 patients with a known RET mutation undergo prophylactic thyroidectomy and have high cure rates. However, when the diagnosis is made at a later age, similar to sporadic patients, metastases are frequently present and recurrent disease often occurs (Modigliani et al., 1998). The main treatment in patients with PTC also consists of a total thyroidectomy and, depending on the extent of the disease lymph node dissection,

115 Chapter 6 which can be followed by thyroid ablation with radioactive iodine (Cooper et al., 2006; Pacini et al., 2006). However, there is no effective adjuvant treatment available for PTC patients with radioiodine refractory disease. New systemic therapies are therefore needed for both recurrent MTC and PTC. With RET being the gene involved in a subset of both MTC and PTC, it is logical to consider the RET receptor as an important target for systemic therapy. RET is expressed in all tumor cells and is involved in tumor growth (de Groot et al., 2006). In MTC patients, RET mutations can affect the extracellular or the intracellular domain of this receptor. In both cases, RET is continuously autophosphorylated on tyrosine residues, which results in a constant activation of downstream signalling pathways that ultimately lead to tumor growth. In PTCs, the intracellular domain of RET (which has tyrosine kinase activity) is fused to the N-terminus of an expressed gene. RET/PTC does not have a transmembrane domain, and thus is not an integral membrane protein, but RET is continuously activated (Plaza-Menacho et al., 2006). Therefore, inhibition of RET phosphorylation and its downstream pathways could be of great value. Small molecule inhibitors that selectively inhibit tyrosine kinases (TK) have been proven to be effective in the treatment of several neoplastic diseases (Druker et al., 2001; Demetri et al., 2002; McArthur et al., 2005). A number of these clinically useful inhibitors target tyrosine kinase receptors that belong to the same family group of proteins as RET (Schlumberger et al., 2008; Ye et al., 2010) Recently, it has been proven that some of these inhibitors, such as imatinib, can inhibit RET activity (Cohen et al., 2002; de Groot et al., 2006). However, high concentrations of imatinib are needed to achieve clinically relevant anti-tumor effects and resistance may develop. Indeed, the first clinical trials in patients with MTC with imatinib showed little effect (de Groot et al., 2007; Frank-Raue et al., 2007). Other TK­ inhibitors such as sorafenib (Nexavar, Bay 43-9006) and vandetanib (ZD6474) have already been tested in vitroand have proven to be more effective than imatinib in the treatment of MTC (Plaza-Menacho et al., 2007; Carlomagno et al., 2002). Additionally,

116 Effectof TK-inhibitors on MTC and PTC cells several TK-inhibitors such as sorafenib, motesanib, gefitinib, vandetanib, axitinib, sunitinib, and XL184 have been evaluated in phase II clinical trials (Table 1). In a large number of patients (25-81%), a stable disease state can be established and some patients even show a partial response (2-33%) (de Groot et al., 2007; Frank-Raue et al., 2007; Lam et al., 2010; Schlumberger et al., 2009; Pennell et al., 2008; Wells et al., 2010; Robinson et al., 2010; Cohen et al., 2008; Cohen et al., 2008; Kurzrock et al., 2010). For an extensive review see Sherman SI (Sherman, 2009). As most studies have focused on one particular TK-inhibitor and have not looked for mutation specificity, it is hard to compare these compounds for the different patient groups. We therefore set out to compare the efficiency of four recently developed TK-inhibitors - XL184, vandetanib, sunitinib and axitinib - in order to determine if there is mutation specificity. We studied the effect of these inhibitors on the proliferation of three cell lines: MTC-TT reported to be derived from a sporadic MTC expressing a C634W RET mutation; MZ-CRC-1 derived from a patient with metastatic sporadic MTC expressing an M918T RET mutation; and TPC-1 derived from a patient with PTC expressing a RET/PTC-1 rearrangement. We also determined the effect of the most efficient TK-inhibitors on RET expression levels, autophosphorylation and activation of the downstream signalling pathways, specifically the MAPK/ERK pathway.

Molecular targets No. of Tumor Stable disease Refs patients response %(n) %(n) lmatinib Bcr-Abl, PDGFR a and 24 0 38% (9) 3,4 �. c-Fms, c-Kit, RET

Sorafenib RAF, VEGFR2 and 3, 21 10% (2) 86% (18) 5 PDGFR, RET

Motesanib VEGFR 1, 2, and 3, 91 2% (2) 81%(74) PDGFR, c-Kit, RET 6

1 Gefitinib EGFR 4 0 0 7 Vandetanib VEGFR, EGFR, RET 30 20% (6) 53% (16) 8 19 16% (3) 53% (10) 9

117 Chapter 6

Axitinib VEGFR 1, 2, and 3, 11 18% {2) 27% (3) 10

1 Sunitinib RET, VEGFR, PDGFR 6 D 83% (5) 11

2 XL184 MET, VEGFR-2, RET 37 (35 with 27% (10) 41% (15) 12 measurable disease)

Table 1. Molecular targets and clinical trials performed with tyrosine kinase inhibitors in medullary thyroid 1 carcinoma (MTC) patients. Clinical trial performed in advanced thyroid cancer patients, not only those with MTC; 2 Phase 1 trial. (3. de Groot et al., 2007; 4. Frank-Raue et al., 2007; 5. Lam et al., 2010; 6. Schlumberger et al., 2009; 7. Pennell et al., 2008; 8. Wells et al., 2010; 9. Robinson et al., 2010; 10. Cohen et al., 2008; 11. Cohen et al., 2008; 12. Kurzrock et al., 2010)

Materials and Methods

Cell culture MTC-TT and TPC-1 cell lines were cultured in RPMI medium (Gibco) supplemented with 15% or 10% (for TPC-1) fetal bovine serum (FBS, BioWhittaker), 1% L-glutamine (Gibco) and 1% penicillin/streptomycin (Gibco). MZ-CRC-1 and HEK293 cell lines were cultured in DMEM high medium containing 4.5 g/L of glucose, L-glutamine and pyruvate (Gibco), supplemented with 15% or 10% (for HEK293) FBS and 1% ° penicillin/streptomycin. All cells were maintained at 37 C and 5% CO2 atmosphere.

Tyrosine kinase inhibitors XL184 was provided by Exelixis as a stock solution of 10 mM in DMSO. Vandetanib was provided by AstraZeneca and made up as a stock solution of 10 mM in DMSO. Sunitinib and axitinib were provided by Pfizer and made up as a stock solution of 10 mM in DMSO.

Cell proliferation assays Cells were seeded in 48-well plates at different cell densities, according to their growth rates. MTC-TT and MZ-CRC-1 cells were plated in 200 µI medium at

118 Effect ofTK-inhibitors on MTC and PTC cells concentrations of 4 x 104 cells/ well. TPC-1 and HEK293 cells were plated at a density of 2 x 103 cells/ well. After growing the cells overnight, different concentrations of TK­ inhibitors were added in an extra 100 µI of medium. Proliferation was measured at different time points using a cell proliferation kit (MTT assay, Roche) according to the manufacturer's instructions. The optical density was determined at 590 nm wavelength. To determine the concentration that led to 50% growth inhibition (IC50), a concentration range of all the inhibitors tested was added to the wells and ICSO was determined using linear interpolation at r = 0.5. All the experiments were performed in triplicate.

Cell lysates and Western blot analysis MZ-CRC-1, MTC-TT and TPC-1 cells were seeded in 75 cm2 kimbles (Greiner) and grown overnight. Cells were then treated with 0, IC50 and maximal concentrations of the different TK-inhibitors. These concentrations were determined based on the results of the proliferation assays (Table 2). Cells were grown in the presence of the TK-inhibitors for 0, 2 and 5 days and lysates were prepared as follows: cells were washed with ice-cold PBS and incubated with lysis buffer (m-PER, Thermo Fisher Scientific) containing protease (Thermo Fisher Scientific) and phosphatase (Thermo Fisher Scientific) inhibitors, for 30 min on ice. Cell lysates were collected by scraping and cleared by centrifugation at 14,000 rpm for 10 min in a pre-cooled (4°C) centrifuge. Supernatants were stored at -80°C before they were processed further for SOS-PAGE followed by Western blot analysis. The following antibodies were used: RET (H-300), RET (C-19), p-RET (Y1062) (Santa Cruz Biotechnology); ERK, p-ERK (Cell Signaling) and actin (C4, MP Biomedicals). Secondary antibodies used were goat anti­ rabbit lgG-HRP (Bio-Rad) and goat anti-mouse lgG-HRP (Bio-Rad). All the experiments were performed in duplicates.

119 Chapter 6

MTC-TT MZ-CRC TPC-1 �Compound Sunitinib 1.56 ± 0.50 µM 1.89 ± 0.46 µM 0.37 ± 0.03 µM

Vandetanib 0.47 ± 0.33 µM 0.26 ± 0.07 µM 0.41 ± 0.09 µM

Axitinib 3.43 ± 0.23 µM >SµM 3.49 ± 0.12 µM

XL184 0.06 µM >SµM 0.06 ± 0.02 µM

Table 2. ICSO values of the different tyrosine kinase inhibitors in different cell lines. The cell lines express a MEN2A (MTC·TT) or a MEN28 (MZ-CRC-1) mutation, or a RET/PTC (TPC-1) rearrangement.

RNA extraction and RT-PCR MTC-TTand MZ-CRC-1 cells treated with 0, ICS0 and maximal concentrations (Table 2) of XL184 and vandetanib respectively, were collected after 0, 2 and 5 days. Cells were washed with PBS (Gibco) and RNA extraction was performed using the RNeasy mini kit!Cl (Qiagen). First strand cDNA was originated from 2 µg of RNA using the Ready-To­ Go You-Prime First-Strand Beadsc, kit (GE Healthcare) and pdN (6) as first-strand primer. First-strand cDNA generated by this method was used directly for RT-PCR. The primers used for amplification were: RET forward (S'-CCGTGAAGATGCTGAAAGAG-3');

RET reverse (S' -AGAGGCCGTCGTCAT AAATC-3'); GAPDH forward (S'­ TGAAGGTCGGAGTCAACGGATTTGGT-3'); GAPDH reverse (5'- GCAGAGATGATGACCCTTTTGGCTC -3').

Statistical analysis The statistical analysis was performed using the softwareprogram SPSS 16.0.

Results

Effect of differentTK-inhibitors on cell proliferation We first evaluated the effect of four different inhibitors (XL184, vandetanib, sunitinib and axitinib) on the proliferation of MZ-CRC-1, MTC-TTand PTC-1 cell lines. Cells were

120 Effect of TK-inhibitors on MTC and PTC cells incubated with increasing concentrations of each TK-inhibitor for 1, 4 and 7 days. After these time periods, cell growth was determined using the MTT assay. A dose­ dependent decrease in cell proliferations was seen for all four inhibitors (Fig. lA, 1B and lC). In contrast, for a cell line that does not express RET, Human Embryonic Kidney cells (HEK293), only minor effects on proliferation were observed at concentrations lower than 0.5 µM (Fig. 1D). Concentrations higher than 0.5 µM seemed to have a toxic effect on HEK293 cells, except for XL184. Based on the IC50 values for XL184, vandetanib, sunitinib and axitinib (Table 2), we observed that XL184 was the most effective inhibitor of MTC-TT (IC50 = 0.06 µM) and TPC-1 (IC50 = 0.06 µM) proliferation. However, for MZ-CRC-1, treatment with vandetanib inhibited cell proliferation at lower concentrations (IC50 = 0.26 µM); it also showed a marked effect on TPC-1 (IC50 = 0.41 µM) and MTC-TT (IC50 = 0.47 µM). Sunitinib also seemed to be a potent inhibitor for TPC-1 (IC50 = 0.37 µM) but it was less effective in MTC-TT and MZ-CRC-1 cells. Axitinib seemed to have hardly any effect on MZ-CRC-1 and TPC-1 since higher concentrations were needed to achieve 50% cell death (Table 2). Based on these results, we concluded that XL184 is the most efficient inhibitor for MTC-TT and TPC-1 cells, followed by vandetanib and sunitinib. For MZ-CRC-1, vandetanib is the most efficient inhibitor.

A) B)

MZ-CRC-1 MTC·TT , ...... Sunltinlb --Vandet.inib �Axitinlb -xu84j 125"

1"""

75"

25" "" 0.005 0.05 0.1 0.5 0005 0.05 01 05 Cona!nlratlon (uM) Concentration (uMl

121 Chapter 6

Cl TPC-1 D) HEK293 J 1 �Sunitin1b -Vandetanib ---Axitirub --..u.&4) �Sunitinib --e-vandetanib -.-AKitinib -++-XL184I 125% 100% t-;;;;;;;;iz½_l ��� C :: 75" i 75" � � SO% �t �==::�_,.-+----- 50% 25" 25" o" °" 0.005 oos O 1 0.5 0.005 005 01 li5- Concentration (uM) Concentratton(uM) E) ICSO Concentrations

asunillnib •Vandetamb CAx1timb •xL-184

MTC·TT MZ.CRC TPC-1 HEK293 • Cell line

Figure 1-Effect of XL184, vandetanib, sunitib and axitinib in cell proliferation. Dose response curves of (A) MZ­ CRC-1, (B) MTC-TT, (C) TPC-1 and (D) HEK293 cell lines incubated with different concentrations of XL184, vandetanib, sunitinib and axitinib. (E) Overview of the ICS0 values of the four tyrosine kinase inhibitors for MTC­ TT, MZ-CRC-1 and TPC-1 cell lines. Error bars shown correspond to standard deviation. MZ-CRC-1 express a MEN2B mutation (RET M918T), MTC-TT express a MEN2A mutation (RET C634R), and TPC-1 express a RET/PTC rearrangement. ICS0 = concentration that lead to 50% cell death.

The effect of XL184 and vandetanib on RET autophosphorylation In order to evaluate the effect of the inhibitors with a stronger effect on cell proliferation {XL184 and vandetanib) on RET autophosphorylation, cell lysates of MTC­ TTand TPC-1 were prepared after treatment with IC50 and maximal concentrations {5 µM) of XL184. Cell lysates of MZ-CRC-1 were also analyzed after incubation with IC50 and maximal concentrations (5 µM) of vandetanib. Inhibition of RET autophosphorylation on tyrosine 1062 was observed for the three cell lines tested. For MTC-TT and TPC-1, RET activation started decreasing with maximal concentrations of XL184 after 2 days of treatment. However, 5 days' exposure to IC50 values of XL184 were also sufficient to reduce RET autophosphorylation levels (Fig. 2A and 28). For MZ-CRC-1, vandetanib led to a decrease in RET activation after 2 days' exposure with IC50 levels. Prolonged exposure to IC50 concentrations of vandetanib led to total inhibition of RET autophosphorylation. Moreover, maximal concentrations of

122 Effect of TK-inhibitors on MTC and PTC cells vandetanib resulted in total inhibition of RET activation after only 2 days' exposure (Fig. 2C). To investigate if the observed reduction in RET autophosphorylation was due to decreased levels of RET expression, the membranes were stripped and probed for RET. The total amount of RET in TPC-1 cells remained relatively unchanged after 5 days' exposure to maximal concentrations of XL184, although we observed a slight increase in RET expression levels after 5 days (Fig. 2B). In contrast, MTC-TT cells showed a dose-dependent response to XL184 after 5 days' incubation with the inhibitor. At ICSO values, a marked decreased in RET expression was observed and at maximal concentrations, RET was no longer detectable (Fig. 2A). For MZ-CRC-1, RET expression was markedly affected by ICSO concentrations of vandetanib after 5 days' exposure (Fig. 2C}. Maximal concentrations of this inhibitor had an effect on RET expression after 2 days' treatment and led to total RET inhibition after prolonged exposure (Fig. 2C). Together, these results show that XL184 and vandetanib are potent inhibitors of RET activation in MTC-TT, TPC-1, and MZ-CRC-1 cells.

A) MTC-TT treated with XL184 C) MZ-CRC-1 treated with vandetanib

2 davs 5 davs 2 davs 5 davs C O ic-50 max O ic-50 max C 0 ic-50 max O oc-50 max RET - p-RET

Actin

B) TPC-1 treated with XL184 2 days 5 days

C O ic-50 max O ic-50 max

Figure 2 - Effectof Xl184 and vandetanib on RET expression and autophosphorylation. (A) MTC-TT cells treated

with ICSO concentration and 5 µM of XL184 for 2 and 5 days. (B) TPC-1 cells treated with ICSO concentration and

5 µM of XL184 for 2 and 5 days. (C) MZ-CRC-1 cells treated with ICSO concentration and 5 µM of vandetanib for 2

123 Chapter 6 and 5 days. MTC-TT express a MEN2A mutation (RET C634R), TPC-1 express a RET/PTC rearrangement, and MZ­ CRC-1 express a MEN2B mutation (RET M918T). ICSO " concentration that leads to 50% cell death.

The effect of XL184 and vandetanib on RET expression To determine if the reduction of RET expression seen in MTC-TT and MZ-CRC-1 was due to inhibition of RET transcription, RET expression levels were determined by RT­ PCR after treatment with XL184 {MTC-TT) and vandetanib (MZ-CRC-1) for 2 and 5 days. Treatment of MTC-TT with ICSO and maximal concentration of XL184 had no effect on RET transcription levels (Fig. 3A). However, exposing MZ-CRC-1 for 5 days to maximal concentrations of vandetanib led to a decrease in RET expression (Fig. 3B). Semi-quantification of the band intensity, calculated with lmageJ software and corrected for the intensity of the GAPDH band for each of the cell lines, confirmed that a decrease in RET transcript was observed after prolonged exposure to maximal concentrations of vandetanib {Fig. 3C). These results show that although RET protein levels were reduced in MTC-TT cells, there was no corresponding reduction of RET mRNA. In contrast, in MZ-CRC-1 we showed that maximal concentrations of vandetanib led to reduced levels of RET mRNA and corresponding reduced levels of RET protein.

A) MTC-TT treated with XL184 C)

Dld.O CU!I.IC!oO Clld...... RET Cl5d,O GAPDH .5d.lC50 11......

B) MZ-CRC-1 treated with vandetanib

2 da'Y! S days C O ic-50 max O k:·5D ma,c;

RET GAPDH

Figure 3 - Effects of XL184 and vandetanib on RET mRNA levels. (A) MTC-TT cells treated with ICSO concentration and 5 µM of XL184 showed no effect on RET transcription levels. (B) MZ-CRC-1 cells treated with vandetanib showed a decrease in RET transcription levels after prolonged exposure (S days) to 5 µM of this

124 Effect of TK-inhibitors on MTC and PTC cells inhibitor. (C} Graphical representation of RET bands' intensity after correction for GAPDH levels confirmed that a reduction of RET mRNA levels was only seen for MZ-CRC-1 cells after prolonged exposure to 5 µM of vandetanib. MTC-TT express a MEN2A mutation (RET C634R) and MZ-CRC-1 express a MEN28 mutation (RET M9 18T). ICSO = concentration that leads to 50% cell death. GAPDH was used as a control.

The effect of XL184 and vandetanib on RET downstream signalling pathways RET is involved in the activation of several signalling pathways, one of which is the MAP kinase pathway, which is especially important for cell proliferation, survival and transformation (Plaza-Menacho et al., 2007; Wells and Santoro, 2009). To evaluate the effect of XL184 and vandetanib in ERK activation, we treated MTC-TT, TPC-1 and MZ-CRC-1 cells with ICS0 and maximal concentrations of XL184 (MTC-TT and TPC-1) and vandetanib (MZ-CRC-1). XL184 was able to reduce ERK phosphorylation in MTC-TT cells at ICS0 concentrations after 2 days' exposure. Prolonged exposure to maximal concentrations of XL184 led to total inhibition of ERK activation (Fig. 4A). In contrast, no effect was observed for TPC-1, even when maximal concentrations of XL184 were used for a period of 5 days (Fig. 4B). For MZ-CRC-1, ERK phosphorylation was markedly reduced with ICS0 levels of vandetanib after 5 days' treatment. Maximal concentrations of this inhibitor were able to completely block ERK activation after only 2 days' exposure (Fig. 4C). Interestingly, total inhibition of ERK activation was observed for MTC-TT and MZ-CRC-1 after treatment with XL184 and vandetanib, respectively, and was related to a reduction in ERK expression levels (Fig. 4A and 4C).

A) MTC-TTtreated with XL184 B) TPC-1 treated with XL184

2 days 5 days Zdays 5 days

C 0 ic-50 max O ic-50 max C 0 ic-50 max O ic-50 max

ERK

p-ERK

Actin

125 Chapter 6

C) MZ-CRC-1 treated with vandetanib

2 days 5 days

C O ic-50 max O ic-50 max

ERK

p-ERK

- -· Actin ------Figure 4 - Effects of XL184 and vandetanib on ERK expression and phosphorylation. (A) MTC-TT cells treated with ICSO concentration and 5 µM of XL184 for 2 and 5 days. (Bl TPC-1 cells treated with ICSO concentration and 5 µM of XL184 for 2 and 5 days. (C) MZ-CRC-1 cells treated with ICSO concentration and 5 µM of vandetanib for 2 and 5 days. MTC-TT express a MEN2A mutation (RET C634R), TPC-1 express a RET/PTC rearrangement, and MZ­ CRC-1 express a MEN2B mutation (RET M918T). ICSO = concentration that lead to 50% cell death.

Discussion Effective systemic treatment for patients with recurrent medullary or papillary thyroid carcinoma (MTC, PTC) is still not available, but new therapeutic options are under development. In both MTC and PTC, mutations or rearrangements of the RET gene, encoding a tyrosine kinase receptor, lead to a constitutive phosphorylated receptor that continuously activates its downstream pathways, leading to uncontrolled cell growth and proliferation. Therefore, tyrosine kinase (TK) inhibitors that can selectively block RET and its intracellular pathways, leading to a decrease in tumor growth and invasion, might have potential as new treatment options for recurrent MTC and PTC. Several TK-inhibitors have been developed and some have been evaluated in clinical trials for the treatment of thyroid cancer with positive results (de Groot et al., 2007; Frank-Raue et al., 2007; Lam et al., 2010; Schlumberger et al., 2009; Pennell et al., 2008; Wells et al., 2010; Robinson et al., 2010; Cohen et al., 2008; Cohen et al., 2008; Kurzrock et al., 2010). However, a comparative study of these different inhibitors has never been done and therefore the relative effectiveness of the different inhibitors is hard to assess. Furthermore, nothing is known about specificity of the different TK­ inhibitors for different types of thyroid cancer and RET mutations.

126 Effect of TK-inhibitors on MTC and PTC cells

In this paper we, for the first time, compared four different tyrosine kinase inhibitors: XL184, vandetanib, sunitinib and axitinib, for their effects on RET inhibition. All four TK-inhibitors have been tested in clinical trials and table 1 summarizes the molecular targets and therapeutic outcomes (Table 1). It is interesting to see that XL184 and vandetanib are the two TK-inhibitors which gave the highest percentage of tumor responses in the treatment of MTC (29% and 16-20%, respectively). However, it was unknown if these four inhibitors show specificity for MEN2A, MEN2B or PTC mutations. To evaluate the efficacy and specificity of these inhibitors, we studied their effect on the proliferation of cells harboring a MEN2A mutation (MTC-TT), a MEN2B mutation (MZ-CRC-1) or a RET/PTC rearrangement (TPC-1). Our results showed that all four of the TK-inhibitors investigated are capable of inhibiting cell proliferation to some extent. However, the concentrations of inhibitors that were needed to inhibit cell growth differed greatly and the ICS0 values for some of them were not within the tested concentration range. XL184 was found to be the most potent inhibitor for MEN2A {ICS0 = 60 nM) and PTC (ICS0 = 60 nM) derived cell lines. Effects on MZ-CRC-1 cell proliferation were only seen with higher concentrations of XL184. For the MEN2B cell line, vandetanib proved to be the most potent inhibitor (ICS0 = 260 nM). An interesting observation was that at the ICS0 values found to have an effect on the proliferation of these cell lines, we saw no effect on the proliferation of HEK293, a control cell line that does not express RET. Based on the data generated for these three cell lines we conclude that XL184 and vandetanib specifically target RET induced cell proliferation at concentrations that in vitro have little or no toxic effect. We also determined the effect of the two most potent inhibitors, XL184 and vandetanib, on RET autophosphorylation levels. For MTC-TT and TPC-1, RET phosphorylation on tyrosine 1062 was reduced after exposure to ICS0 concentrations of XL184 for 5 days. MZ-CRC-1 cells showed a decrease in RET activation after 2 days' exposure to ICS0 values of vandetanib, but prolonged exposure led to a total inhibition of RET phosphorylation. Parallel to this effect, we observed that RET

127 Chapter 6 expression was also affected by prolonged exposure to these inhibitors in MTC-TT and MZ-CRC-1. Both cell lines showed a decrease in RET expression levels after exposure for 5 days to ICS0 concentrations of XL184 and vandetanib respectively, but only vandetanib exerted this effect by inhibiting RET transcription. For XL184 we were unable to detect a change in RET transcript levels even after prolonged exposure to maximal concentrations of the inhibitor. It is probable that for XL184 lysosomal or proteasomal degradation is involved, as was described for other inhibitors {Plaza­ Menacho et al., 2007). No effect on RET expression levels was detected for TPC-1, even after prolonged exposure to maximal concentrations of XL184. This means that the effect of XL184 in PTC is only achieved by direct inhibition of RET autophosphorylation and that possible lysosomal or proteasomal degradation may not be effective due to the presence of the fusion protein. Finally, we explored the effect of these drugs in a downstream signalling pathway, the MAP kinase pathway, known to be directly activated by RET, and specifically at the levels of ERK phosphorylation. Previous studies have reported the effect of vandetanib on RET activation and have shown that exposure to this inhibitor led to inhibition of ERK phosphorylation (Carlomagno et al., 2002). For MTC-TT and MZ-CRC-1, exposure to ICS0 levels of XL184 and vandetanib, respectively, induced a marked decrease in ERK phosphorylation. However, prolonged exposure to vandetanib was necessary to achieve this effect for MZ-CRC-1. Interestingly we observed that there was a reduction on ERK expression in MTC-TT and MZ-CRC-1 after treatment with maximal concentrations of XL184 and vandetanib, respectively, suggesting that these inhibitors have a possible effect on ERK transcription. XL184 demonstrated efficacy and inhibition of RET/PTC in the TPC-1 PTC model system. However, no change in ERK phosphorylation was detected, even after prolonged exposure to maximal concentrations of XL184. ERK phosphorylation through other tyrosine kisase receptors and tyrosine kinase effector molecules involved in activation of the MAPK/ERK pathway, e.g. BRAF, are likely to exert a stronger effect in ERK phosphorylation than RET. It is possible that for PTC, a

128 Effect of TK-inhibitors on MTC and PTC cells combination of different inhibitors targeting RET and e.g. BRAF, could circumvent this problem and thus result in an even more effective treatment for this type of cancer. In conclusion, our in vitro results are in general agreement with the outcome of the clinical trials, supporting the idea that XL184 and vandetanib are two potent inhibitors for tumor response in MTC. By studying three cell lines we did observe specificity for the treatment of different RET mutations, with XL184 being the most potent inhibitor for the cell line expressing a MEN2A mutation and for a cell line expressing a PTC rearrangement. Vandetanib proved most effective in a cell line expressing a MEN28 mutation. As no TK-inhibitor was superior for all the cell lines we tested, our results suggest that mutation-specific therapies could be beneficial in the treatment of MTC and PTC. This is an important finding since in clinical trials to date no distinction has been made between the different RET related mutational subtypes (MEN2A/MEN28/PTC). Making such distinctions might change the outcome, in particular in tumor response, and likely improve the outcome of ongoing studies. Our data also suggest that a reanalysis of already performed trails, based on mutation status is more than worthwhile.

Acknowledgements We would like to thank Exelixis, AstraZeneca and Pfizer for providing the TK-inhibitors for investigation and Jackie Senior for editing the manuscript.

129 130 Chapter 7

Summarizing Discussion and Future Perspectives Chapter 7

RET is a tyrosine kinase receptor which plays a crucial role during embryonic development. In the early stages, RET expression is detected in a cranial population of neural crest stem cells (NCSCs) and subsequently throughout the central nervous system, but specifically in motor and catecholaminergic neurons. During development, RET- expressing NCSCs have the ability to migrate caudally via the intestinal mesenchyme to form the enteric nervous system, located in the gut wall of the gastro-intestinal tract. In addition, another group of RET expressing cells gives rise to the development of sensory and autonomic ganglia of the peripheral nervous system, adrenal chromaffin cells and c-cells of the thyroid, and the kidney (Manie et al., 2001; Arighi et al., 2005; de Groot et al., 2006). Specific germline mutations in RET have been reported to lead to an aberrant embryonic development of the NCSCs and are consequently are responsible for two unrelated neural crest disorders: Hirschsprung disease (HSCR), a congenital absence of the enteric neural crest cells in the gastrointestinal tract, and multiple endocrine neoplasia type 2 (MEN2), a dominantly inherited cancer syndrome. In this last section of the thesis, I summarize the functional and genetic work described in this thesis and discuss how it fits into the current body of knowledge on the field. I addition, I also present ideas for future work that needs to be carried out in the future to improve our current understanding of HSCR and MEN2.

Part I - HSCR HSCR is a complex disorder for which several causative genes have been identified. To date, mutations in eleven genes have been described in HSCR patients. Despite the relatively large number of associated genes, the precise mechanism of disease development is still not fully understood, nor do these genes explain all HSCR cases. Since HSCR is frequently associated with other congenital malformations present in rare (recessive or dominant) syndromes, studying the genes involved in these syndromes may provide greater insights into HSCR development and help identify new HSCR candidate genes.

132 Discussion and Future Perspectives

In Chapter 2 the protein network associated with KIAA1279/KBP was studied; this protein is encoded by a HSCR-associated gene and its function is largely unclear (Brooks et al., 2005). Homozygous nonsense mutations in KBP are involved in a rare, recessive syndrome called Goldberg-Shprintzen syndrome (GOSHS) which is characterized by polymicrogyria, microcephaly, facial dysmorphisms, mental retardation and, in most cases, with HSCR (Goldberg and Shprintzen, 1981; Brooks et al., 2005). Based on the clinical phenotype, KBP is required for the normal development of both the central and peripheral nervous system, but how this protein is involved in this process was largely unclear, in other words, the precise function of KBP was unknown. Our results shed some light on its function, by showing that KBP expression is up-regulated during neuronal development in mouse cortical neurons. Moreover, we showed that KBP is necessary for neuronal differentiation and neurite outgrowth of PC12 cells, suggesting that KBP is involved in cell differentiation and neuronal development or plasticity. Parallel to this, we performed a yeast two-hybrid screen to identify KBP­ interacting proteins, and found several kinesin like proteins, supporting the idea that KBP is indeed a kinesin binding protein (Wozniak et al., 2005). In addition, we also found that KBP interacts with SCGlO, a stathmin-like protein previously shown to be involved in microtubule dynamics (Di Paolo et al., 1997; Grenningloh et al., 2004). Additional studies in vitro and in vivo confirmed this interaction and suggested that KBP may play a role in microtubule organization or stability. Previous characterization of a kbp mutant zebrafish had implicated KBP in the regulation of neuronal microtubules organization (Lyons et al., 2008), but we found that KBP does not directly interact with microtubules. This raises the question as to how KBP affects microtubule dynamics? The answer may lie in its interaction with SCGl0 and three possible mechanisms of action are proposed. The first is that KBP might be involved in the palmitoylation process that precedes the transport of SCGl0 to the growth cones. Since SCGl0 needs this post-translational modification to be

133 Chapter 7 transported and to exert its microtubule destabilizing action (Lutjens et al., 2000), KBP depletion would result in an impairment of SCGl0 transport and, consequently, in an alteration of its microtubule destabilizing action. The second mechanism could be that KBP regulates SCGlO activity. SCGl0 alternates between stages of activation/inactivation, triggered by dephosphorylation/phosphorylation of its serine residues that are necessary for microtubule dynamics (Antonsson et al., 1998). The third possible mechanism is that KBP might mediate SCGlO interaction with tubulin. Consistent with this, we observed an interaction between KBP and tubulin in our yeast two-hybrid screen. Thius, KBP could potentially control the levels of free tubulin available for microtubule polymerization. Either way, KBP depletion will affect microtubule dynamics and, consequently, neurite extension and axonal growth, which is in line with the neuronal development defects seen in the GOSHS clinical phenotype. However, the obvious question that arises from our work is what do these results tell us about HSCR? Impairments in the survival, proliferation and migration of NCSCs have been suggested as possible mechanisms for HSCR development (Heanue and Pachnis, 2007). Our study showed that KBP is involved in neuronal differentiation, suggesting that a failure in differentiation of NCSCs into enteric neurons after colonization of the intestinal tract could also play a role in HSCR development. Thus, we propose a new mechanism based on cell differentiation, in addition to cell survival, proliferation and migration. Interestingly, our data suggest that SCGlO might be the possible link between KBP and the other genes involved in HSCR development. SCGlO activity can be controlled by phosphorylation of specific serine residues by the Protein Kinase A (PKA) (Togano et al., 2005). PKA was previously implicated in HSCR as a possible link between the RET and EDNRB signalling pathways (Barlow et al., 2003). Accordingly it is tempting to postulate that PKA migth also be the link between KBP, SCGlO, RET and EDNRB. A scheme of all the genes involved in HCSR and of their possible interacting network is shown in Figure 1.

134 Discussion and Future Perspectives

GDNF

EDN3

EONR& si1nallln1

cytoplasm

Proliferation Migration D Differentiation

• ENCCs

Figure 1 - Schematic representation of possible interactions between HSCR related genes (Adapted from Tam and Garcia-Barcelo, 2009).

In Chapter 3, the involvement of SCGlO in HSCR development is evaluated. Previous studies have identified SCG10 as a down-regulated gene in a RET mouse model for HSCR (Heanue and Pachnis, 2006), and we have shown that SCG10 interacts with KBP. In addition, previous studies performed in zebrafish have already shown that SCGlO is expressed during the development of the central and enteric nervous systems (Burzynski et al., 2009), supporting the idea that SCGlO is involved in HSCR development. To determine if SCGlO is expressed in NCSCs during colonization of the intestinal tract, E10.5 to E15.5 mouse NCSCs were collected and analysed for SCGlO

135 Chapter 7 expression. We detected SCGl0 expression in NCSCs from all embryonic stages, with a slight increase in SCG10 levels in the initial stages {El0.S to Ell.S). To further evaluate SCG10 involvement in HSCR, 85 sporadic, non-syndromic HSCR patients, in whom no mutation in RET had been found, were screened for the presence of SCG10 mutations. However, no mutations in SCG10 were found in any of these patients. Based on this result, we concluded that SCG10 is not directly involved in HSCR development. However, there is still a possibility that SCG10 might play an indirect role in HSCR due to a secondary effect caused by disruptions in its protein network. SCG10 activity is known to be controlled by two post-translational modifications: palmitoylation and phosphorylation (Antonsson et al., 1998; Lutjens et al., 2000). As RET is a tyrosine kinase receptor known to be involved in several signalling pathways, it is possible that the RET receptor, via one of its downstream pathways, controls SCG10 activity, linking this gene directly to HSCR. Further studies are required to further elucidate any potential role of SCG10 in HSCR development. Complex disorders, such as HSCR, are not caused by mutations in a single gene, but rather arise from the combined action of several genes and risk-conferring variants. However, the genes contributing to the development of such disorders are difficult to find since they normally exert only a small effect on disease risk (Kiberstis and Roberts, 2002). Thus, there is now a search for new and more sophisticated technologies that might identify these genes/rare variants. In Chapter 4, recent approaches that might be helpful in identifying new candidate genes for complex disorders are discussed. Using HSCR as an example, we reviewed what is known about the disease, what we have learned from genome-wide association studies, and how next-generation sequencing in combination with a systems biology approach might help us in to unravel the complex inheritance pattern of HSCR.

PartII - MEN2

Activating mutations in RET are the main cause of MEN2, a dominantly inherited cancer syndrome that can be divided into three subtypess: MEN2A, MEN2B and

136 Discussion and Future Perspectives

Familial Medullary Thyroid Carcinoma (FMTC), based on the clinical description and tissues affected (Hansford and Mulligan, 2000). The observation that different RET mutations result in different disease phenotypes, suggests that specific signalling pathways are activated by the mutated RET receptor. Therefore, identifying the signalling pathways triggered by the different RET mutations might yield new insights into the RET protein network and could be used for the development of specific therapeutic approaches for treating MEN2A, MEN2B and FMTC. In Chapter 5, we tried to identify the specific downstream signalling pathways using a powerful tool called kinome profiling, to generate descriptions of cellular signalling networks involved in the different pathogenesis. With this approach, phosphorylation patterns specific for the most common mutations found in MEN2A (RET609, RET620, RET634), MEN2B (RET918) and FMTC (RET791, RET891) were analysed, using a peptide array consisting of 1024 differentkinase substrates. Based on our results, we confirmed that there are indeed differences between RET mutants and RET wild type. But, interestingly, all the mutants we analysed showed an enhanced activation of canonical RET signalling pathways: MAP kinase, Pl3Kinase/Akt and Rae pathways. This up-regulation was expected since the MAP kinase pathway is involved in cell proliferation and differentiation, the Pl3 kinase pathways is involved in cell survival and proliferation and the RAC pathway is associated with cytoskeletal modifications and cell motility (Fukuda et al., 2002; Kodama et al., 2005). Furthermore, it suggests that the differences between RET mutants and the wild-type receptor are small and likely to result from exacerbated activation of pre-existing signalling pathways. Specific alterations in signalling properties were identified in MEN2B- and FMTC-associated mutants. For RET918 a specific increase in activation of the focal adhesion kinase (FAK) was observed. FAK is an intracellular tyrosine kinase involved in several cellular responses including cell migration, adhesion and proliferation (Cai et al., 2008). Previous studies have reported a direct effect of wild type RET on FAK phosphorylation (Murakami et al., 1999, You et al., 2001). Furthermore, RET was

137 Chapter 7 reported to signal through FAK in medullary thyroid cancer cells (Panta et al. , 2004). Since FAK is involved in cytoskeletal changes that lead to an increase in invasive properties, this increase in FAK phosphorylation levels for RET918 might explain the more severe phenotype observed for patients with MEN2B. Besides FAK, we also found a specific activation of the Aurora kinase B, a serine/threonine kinase required for chromosome segregation and cytokinesis (Mountzios et al., 2008), in the RET918 mutant. Aurora kinase B was previously reported to be over-expressed in a variety of human tumour types, e.g. endometrial carcinomas (Kurai et al., 2005), and it has been shown to correlate with the level of genomic instability in a tumour, suggesting that Aurora B contributes to the acquisition of genetic alterations critical for neoplastic transformation (Smith et al., 2005). Although Aurora kinase B expression was detected in anaplastic thyroid carcinoma (Wiseman et al., 2007), this is the first time that this kinase has been found to be associated with MEN2. Furthermore, the increase in phosphorylation of this kinase observed specifically for the RET918 mutation can also account for the more severe phenotype seen in MEN2B compared to the other mutants and to the wild-type receptor, but further studies are necessary to determine the effect of the Aurora kinase Bin MEN2B. For RET891, one of the FMTC mutants analysed, a specific alteration in receptor endocytosis was detected. We observed that while RET wild-type recycles twice in one hour after ligand activation, RET891 recycles only once. This difference could be explained by the fact that RET891 is not responsive to ligand activation

(Plaza-Menacho et al., 2005). Neither was a response to ligand activation observed for

RET791 (Plaza-Menacho et al. , 2005), but no difference in receptor internalization was detected for this mutant when compared to the wild-type. Previous studies have shown that a serine to alanine substitution in EGF tyrosine kinase receptors result in an abrogation of ligand-induced internalization of the receptor (Oskvold et al., 2003), but no effect of similar substitutions has been described for the RET receptor. Given that RET891 is the only RET mutant where a serine to alanine substitution occurs, this difference might explain the changes observed in the pattern of RET internalization.

138 Discussion and Future Perspectives

Accordingly, RET891 is more resistant to internalization than RET791 and RET wild­ type and is retained on the membrane longer after ligand stimulation. Together, these data show that there are differences between RET mutants and wild-type and provide further evidence of the presence of signalling profiles specific for different RET mutations. Our results also yielded new insights into the RET protein network, which might be helpful in developing specific therapeutic approaches for treating MEN2A, MEN2B and FMTC. There is an ongoing search for better and more efficient therapeutic approaches for treating recurrent medullary and papillary carcinoma. Since the oncogenic activation of the RET receptor plays a major role in the development of these carcinomas, the search for efficient tyrosine kinase (TK) inhibitors that could inhibit RET and the activation of its downstream signalling pathways is being conducted by several pharmaceutical companies. Several TK-inhibitors have been developed in order to specifically block the constitutive activation of RET and some of them have already been tested in clinical trials. Unfortunately, for the majority of these inhibitors, the concentrations needed to suppress RET oncogenic signalling are too high to be administered without severe side-effects for the patients. In chapter 6, we focus on four new TK-inhibitors: XL184, vandetanib, sunitinib and axitinib, and their effects on the treatment of medullary and papillary thyroid carcinoma. In order to compare the efficiency of XL184, vandetanib, sunitinib and axitinib the effect of these inhibitors on the proliferation of three cell lines: MTC-TT (MEN2A, RETC634R); MZ-CRC-1 (MEN2B mutation, RETM918T), and TPC-1 (RET/PTC-1 rearrangement) was analysed. Furthermore, we evaluated their effect on the levels of RET expression, RET phosphorylation and activation of downstream signalling pathways, specifically on the MAPK/ERK pathway. We observed that the cell lines tested have different sensitivities to the various inhibitors. For the cells lines expressing a MEN2A (MTC-TT) and a RET/PTC-1 rearrangement (TPC-1), XL184 was the most efficient inhibitor, while vandetanib was a more potent inhibitor for MZ-CRC-1 cells expressing a MEN2B mutation. In addition,

139 Chapter 7 we were able to show that both XL184 and vandetanib led to a decrease in the levels of RET and ERK phosphorylation in MTC-TT and MZ-CRC-1 cells, respectively. Interestingly, clinical trials for these four TK-inhibitors have been performed and have shown that XL184 and vandetanib are the two TK-inhibitors with the highest percentage of tumour responses in the treatment of medullary thyroid carcinoma (Wells et al., 2010; Robinson et al., 2010; Cohen et al., 2008; Cohen et al., 2008; Kurzrock et al., 2008). Our results are in agreement with the outcome of the clinical trials, reinforcing the idea that XL184 and vandetanib are potent inhibitors for the treatment of MTC. However, we also showed that there is specificity for different RET mutations, with XL184 being the most potent inhibitor in MEN2A, and vandetanib the most effective in MEN2B. For TPC-1 cells, although XL184 demonstrated efficacy and inhibition of RET/PTC, no change in ERK phosphorylation was detected. Thus, it is likely that other tyrosine kisase receptors and tyrosine kinase effector molecules involved in activation of the MAPK/ERK pathway, e.g. BRAF, are likely to exert a stronger effect in ERK phosphorylation than RET. It is possible that for PTC, a combination of different inhibitors targeting RET and e.g. BRAF, could circumvent this problem and thus result in an even more effectivetreatment for this type of cancer. The most important finding of this study was the mutation specificity of the different inhibitors. It has a direct bearing on the RET-related clinical trials since, so far, no distinction has been made between the different RET-related mutational subtypes (MEN2A/MEN2B/TPC) in the trails performed.

Future Perspectives The work described in this thesis provides a better understanding of the development of two neural crest-related disorders, HSCR and MEN2. It has, however, raised five important questions that should be addressed in future studies to fully evaluate the effect of these findings. The first question arose from the functional studies performed for KBP. Although, we were able to associate KBP to microtubule dynamics based on its

140 Discussion and Future Perspectives interaction with SCGlO, the mechanism by which KBP exerts this effect is still unknown. We reported that KBP does not directly interact with microtubules, which suggests that KBP might be able to control the microtubule destabilizing activity of SCGl0. As described, this could happen in three possible ways: 1) KBP might be involved in the palmitoylation process that regulates SCGlO transport to the growth cones; 2) KBP could play a role in SCGlO dephosphorylation/phosphorylation that controls SCGlO activation; 3) KBP might mediate SCGl0 interaction with tubulin. To test these hypotheses, KBP expression should be knocked down in neuronal cells and 1) SCGlO localization should be determined; 2) SCGl0 phosphorylation levels should be checked; 3) SCGl0 microtubule destabilizing effects should be analysed. If, in one of these situations, KBP depletion results in a deregulation of SCGlO activity, a microtubule defect is to be expected, linking KBP to microtubule dynamics. A second question concerns the involvement of KBP in HSCR development. Our results showed that KBP is necessary for neuronal differentiation. However, we still cannot exclude KBP being incolved in the survival, proliferation and migration of NCSCs. To further investigate these effects, it would be of tremendous value to generate a KBP-/- mouse in which the process of gut colonization by NCSCs could be followed and analysed for the presence of abnormalities. A third question is related to the mechanism by which SCGl0 is involved in HSCR. Although SCGlO had been proposed as a candidate HSCR gene, and we found a direct interaction between SCGlO and KBP, the exact mechanism by which SCGlO plays a role in HSCR still remains to be elucidated. One possible option is that RET controls SCGlO activity via direct interaction. To investigate this hypothesis, co­ immunoprecipitation assays to detect RET-SCGlO interaction, and analysis of SCGlO phosphorylation levels in the presence of specific RET TK-inhibitors, might help clarify the role of SCGlO in HSCR. Another possible mechanism is that RET controls SCGlO activity indirectly, via activation of one of its downstream signalling pathways. Since SCGlO activity is known to be controlled by PKA, and serine 696 (in mice 697) of RET is known to be the docking site for PKA, we suggest that SCGlO activity could be

141 Chapter 7 affected by an alteration of PKA activation due to a mutation of RET serine 696. To investigate this hypothesis, SCGlO and PKA activation should be analysed in the presence of this RET mutation. Since there is already a mouse model containing a homozygous RET697 mutation available, it could be used to study the precise effect of RET in SCGlO activation. A fourth question comes from the study of different RET mutations. Although we found activation of specific pathways for different mutations, these changes were studied using cell lines and thus in an in vitro setting. Ideally, these pathways should be validated in patient material. This would further allow us to identify true targets, since it is possible that some of these pathways are not replicated in the three­ dimensional setting of a tissue. Identifying these true targets could then lead to creating animal models that would provide further mechanistic insights into the development of the different subtypes of MEN2. In addition, this could lead to the development of specific therapeutic strategies for treating MEN2A, MEN2B and FMTC. The final question is related to the use of TK-inhibitors in the clinic. We showed that different RET mutations have different sensitivities to the TK-inhibitors tested, suggesting that mutation-specific therapies could be beneficial in treating MTC and PTC. Thus, future treatment strategies should take into consideration patients' genetic make-up along with any functional data available for the different TK­ inhibitors, in order to provide patients with the most efficient treatment available. Furthermore, our data suggest that re-analysing earlier trials based on their mutation status would be more than worthwhile.

142 Abbreviations Abbreviations

List of abbreviations ALL- Anterior Lateral Line AKT/PKB - Protein kinase B BICD2-N - Modified Bicaudal 02 containing only the N-terminal region BRAF -v-RAF murine sarcoma viral oncogene homolog Bl BSA - Bovine Serum Albu mine c-AMP - Cyclic adenosine monophosphate CCHS - Congenital Central Hypoventilation syndrome CDK5 - cyclin dependent kinase 5 CDS - coding sequence CNS - Central nervous system DAPI - Diamidino phenylindole DMEM - Dulbecco's Modified Eagle Medium DNA - Deoxyribonucleic acid Dox - doxycycline DOK - Downstream of kinase Dox - doxycycline El0.5 - E15.5 - Embryonic day 10.5 - 15.5 ECE-1 - endothelin-converting enzyme 1 EDN3 - endothelin 3 EON RB - endothelin receptor type B ENCSCs - Enteric neural crest stem cells ENS - Enteric Nervous System Erb B - Epidermal growth factor receptor ERK - Extracellular signal-regulated protein kinase FACs - Fluorescence-activated cell sorting FAK - Focal adhesion kinase FITC- Fluorescein lsothiocyanate FMTC - Familial Medullary Thyroide Carcinoma FRS2 - fibroblast growth factor receptor substrate 2 GDNF - Glial cell-line derived neurotrophic factor GFP - Green fluorescent protein GFRa.- Glycosyl - phosphatidylinositol membrane anchored co-receptor alpha GOSHS - Goldberg-Shprintzen syndrome GWAS - Genome - wide association studies HEK293 - Human embryonic kidney cells HeLa - Henrietta Lacks cell line

144 Abbreviations hpf- hours post-fertilization HSCR - Hirschsprung disease IP3 - inositol-triphosphate IRS - insulin receptor substrate JAKs -Janus tyrosine kinases JNK- c-Jun N-terminal kinase KBP - Kinesin binding protein KIF - Kinesin family of proteins LS-HSCR - long-segment HSCR MAPs - Microtubule associated proteins MAPK- Mitogen-activated protein kinase MENZ - Multiple Endocrine Neoplasia type 2 MET - hepatocyte growth factor receptor MTC - Medullary Thyroid Carcinoma mTOR - mammalian target of rapamycin MO - Morpholino MWS - Mowat-Wilson syndrome NlE-115 - Mouse neuroblastoma cell line NaCl - Sodium chloride NBT/BCIP-Nitro Blue Tetrazolium/Bromo Chloro lndolyl Phosphate NCSCs - Neural Crest Stem Cells �-NGF - Nerve growth factor � NGS - Next - generation sequencing NIH-3T3 - Mouse embryonic fibroblast cell line NRGl - Neuregulin 1 NTN - Neurturin NTRKl - Neurotrophic tyrosine kinase, receptor, type 1 NXF2/ARNT- Nuclear export factor 2 / aryl hydrocarbon receptor nuclear translocator ON - Overnigth OR - odd ratio ORFs - Open Reading Frames PS3 - tumor protein 53 PAK2 - p21 activated kinase 2 PBS - Phosphate Buffered Saline PC - pheochromocytoma PC12 - Rat pheochromocytoma cells PCR - Polymerase chain reaction

145 Abbreviations

PDKl - phosphoinositide-dependent kinase 1 PEI - Polyethylenimine PFA - Paraformaldehyde PHOX2B - paired-like homeobox 2b, neuroblastoma paired-type homeobox protein Pl3K - phosphoinositide 3 kinase PKA - Protein kinase A PKC - Protein kinase C PLC-y- Phospholipase C-y PLL - Posterior lateral line PNS - Peripheral nervous system PTC - Papillary Thyroid Carcinoma RAS - Rat sarcoma oncogene RET - Rearranged during Transfection RNA - Ribonucleic acid RPMI - Roswell Park Memorial Institute SBMO - Splice Blocking Morpholino SCGl0 - Superior cervical ganglia, neural specific 10 SDS-PAGE - Sodium dodecyl sulfate - Polyacrylamide gel electrophoresis Shank3 - multiple ankyrin repeat domains 3 She - Src homolog collagen ShcC - Src homology 2 domain-containing transforming protein C shRNA - Short hairpin ribonucleic acid SNPs - Single nucleotide polymorphisms SOS - Son of sevenless multi-protein SOXlO - SRY (sex determining region Y)-box 10 Src - chicken Rous sarcoma virus oncogene SS-HSCR - Short-segment HSCR STAT3 - Signal transducer and activator of transcription 3 TIA - Total intestinal aganglionosis TK -Tyrosine kinase Tris-HCI - Tris(hydroxymethyl)aminomethane- Hydrochloride TTF -1 - Thyroid Transcription Factor 1 UTR - Untranslated region WSS - Waardenburg-Shah syndrome WT - wild-type ZFHXlB - Zinc finger homeobox 1B

146 References References

References

1. Airaksinen MS, Titievsky A, Saarma M. (1999) GDNF family neurotrophic factor signaling: four masters, one servant? Mol Cell Neurosci. 13:313-25.

2. Ahituv N, Zhu Y, Visel A, Holt A, Afzal V, et al. (2007) Deletion of ultraconserved elements yields viable mice. PLoS Biol. 5: e234.

3. Airaksinen MS, Titievsky A, Saarma M. (1999) GDNF family neurotrophic factor signaling: four masters, one servant? Mol Cell Neurosci. 13: 313-325.

4. Alves MM, Burzynski G, Delalande JM, Osinga J, van der Goot A, et al. (2010) KBP interacts with SCGlO, linking Goldberg-Shprintzen syndrome to microtubule dynamics and neuronal differentiation. Hum Mol Genet. 19: 3642-51.

5. American Thyroid Association Guidelines Task Force, Kloos RT, Eng C, Evans DB, Francis GL, et al. (2009) Medullary thyroid cancer: management guidelines of the American Thyroid Association. Thyroid. 19: 565-612.

6. Amiel J, Espinosa-Parrilla Y, Steffann J, Gosset P, Pelet A, et al. (2001) Large-scale deletions and SMADIPl truncating mutations in syndromic Hirschsprung disease with involvement of midline structures. Am J Hum Genet. 69:1370-7.

7. Amie! J, Lyonnet S. (2001) Hirschsprung disease, associated syndromes, and genetics: a review. J Med Genet. 38:729-39.

8. Amiel J, Laudier B, Attie-Bitach T, Trang H, de Pontual L, et al. (2003) Polyalanine expansion and frameshift mutations of the paired-like homeobox gene PHOX2B in congenital central hypoventilation syndrome. Nat Genet. 33: 459-61.

9. Amie! J, Sproat-Emison E, Garcia-Barcelo M, Lantieri F, Burzynski G, et al. (2008) Hirschsprung disease, associated syndromes and genetics: a review. J Med Genet. 4S: 1-14.

10. Andersen SS. (2000) Spindle assembly and the art of regulating microtubule dynamics by MAPs and Stathmin/Op18. Trends Cell Biol. 10: 61-267.

11. Anderson RB, Anderson RB, Newgreen DF, Young HM. (2006) Neural crest and the development of the enteric nervous system. Adv Exp Med Biol. 589: 181-96.

12. Andreozzi F, Melillo RM, Carlomagno F, Oriente F, Miele C, et al. (2003) Protein kinase Calpha activation by RET: evidence for a negative feedback mechanism controlling RET tyrosine kinase. Oncogene. 22: 2942-9.

13. Angrist M, Kauffman E, Slaugenhaupt SA, Matise TC, Puffenberger EG, et al. (1993) A gene for Hirschsprung disease (megacolon) in the pericentromeric region of human chromosome 10. Nat Genet. 4: 351-6.

148 References

14. Angrist M, Bolk S, Thiel B, Puffenberger EG, Hofstra RM, et al. (1995) Mutation analysis of the RET receptor tyrosine kinase in Hirschsprung disease. Hum Mol Genet. 4: 821-30.

15. Angrist M, Bolk S, Halushka M, Lapchak PA, Chakravarti A. (1996) Germline mutations in glial cell line-derived neurotrophic factor (GDNF) and RET in a Hirschsprung disease patient. Nat Genet. 14: 341-4.

16. Antonsson B, Kassel DB, Di Paolo G, Lutjens R, Riederer BM, et al. (1998) Identification of in vitro phosphorylation sites in the growth cone protein SCGl0. Effect of phosphorylation site mutants on microtubule-destabilizing activity. J. Biol. Chem. 273: 8439-8446.

17. Arighi E, Popsueva A, Degl'lnnocenti D, Borrello MG, Carniti C, et al. (2004) Biological effects of the dual phenotypic Janus mutation of ret cosegregating with both multiple endocrine neoplasia type 2 and Hirschsprung's disease. Mol Endocrinol. 18: 1004-17.

18. Arighi E, Borrello MG, Sariola H. (2005) RET tyrosine kinase signaling in development and cancer. Cytokine Growth Factor Rev. 16: 441-67.

19. Asai N, Fukuda T, Wu Z, Enomoto A, Pachnis V, et al. (2006) Targeted mutation of serine 697 in the Ret tyrosine kinase causes migration defect of enteric neural crest cells. Development. 133: 4507-16.

20. Attie T, Pelet A, Edery P, Eng C, Mulligan LM, et al. (1995) Diversity of RET proto-oncogene mutations in familial and sporadic Hirschsprung disease. Hum Mol Genet, 4: 1381-6.

21. Badner JA, Sieber WK, GarverKL, Chakravarti A. (1990) A genetic study of Hirschsprung disease. Am J Hum Genet. 46: 568-80.

22. Barlow A, de Graaff E, Pachnis V. (2003) Enteric nervous system progenitors are coordinately controlled by the G protein-coupled receptor EDNRB and the receptor tyrosine kinase RET. Neuron. 40: 905-16.

23. Blatch G.L., Lassie M. (1999) The tetratricopeptide repeat: a structural motif mediating protein­ protein interactions. Bioessoys. 21, 932-939.

24. Bloksberg LN. (2008) Next-Generation Sequence Data Analysis. In Next-Generation Genome Sequencing. Towards Personalized Medicine. M. Janitz, editor. Wiley-Blackwell.

25. Bolk S, Pelet A, Hofstra RM, Angrist M, Salomon R, et al. (2000) A human model for multigenic inheritance: phenotypic expression in Hirschsprung disease requires both the RET gene and a new 9q31 locus. Proc Natl Acad Sci USA. 97: 268-73.

26. Bondallaz P, Barbier A, Soehrman S, Grenningloh G, Riederer BM. (2006) The control of microtubule stability in vitro and in transfected cells by MAPlB and SCGl0. Cell Motil. Cytoskeleton. 63: 681-695.

149 References

27. Bondurand N, Natarajan D, Barlow A, Thapar N, Pachnis V. (2006) Maintenance of mammalian enteric nervoussystem progenitors by SOXlO and endothelin 3 signalling. Development. 133: 2075- 86.

28. Borrego S, Wright FA, Fernandez RM, Williams N, Lopez-Alonso M, et al. (2003) A founding locus within the RET proto-oncogene may account for a large proportion of apparently sporadic Hirschsprung disease and a subset of cases of sporadic medullary thyroid carcinoma. Am J Hum Genet. 72: 88-100.

29. Borrello MG, Alberti L, Arighi E, Bongarzone I, Battistini C, et al. (1996) The full oncogenic activity of Ret/ptc2 depends on tyrosine 539, a docking site for phospholipase Cgamma. Mol Cell Biol. 16: 2151-63.

30. Brooks AS, Bertoli-Avella AM, Burzynski GM, Breedveld G.J., Osinga J. et al. (2005) Homozygous nonsense mutations in KIAA1279 are associated with malformations of the central and enteric nervous systems. Am J Hum Genet. 77: 120-126.

31. Brooks AS, Oostra BA, Hofstra RM. (2005) Studying the genetics of Hirschsprung's disease: unraveling an oligogenic disorder. Clin Genet. 67(1):6-14.

32. Brooks AS, Leegwater PA, Burzynski GM, Willems PJ, de Graaf B et al. (2006) A novel susceptibility locus for Hirschsprung's disease maps to 4q31.3-q32.3. J Med Genet. 43: e35.

33. Brunet JF, Pattyn A. (2002) Phox2 genes - from patterning to connectivity. Curr Opin Genet Dev. 12: 435-40.

34. Burns AJ, Douarin NM. (1998) The sacral neural crest contributes neurons and glia to the post­ umbilical gut: spatiotemporal analysis of the development of the enteric nervous system. Development. 125: 4335-47.

35. Burzynski GM, Nolte IM, Osinga J, Ceccherini I, Twigt B, et al. (2004) Localizing a putative mutation as the major contributor to the development of sporadic Hirschsprung disease to the RET genomic sequence between the promoter region and exon 2. Eur J Hum Genet. 12: 604-12.

36. Burzynski GM, Nolte IM, Branda A, Bos KK, Osinga J, et al. (2005) Identifying candidate Hirschsprung disease-associated RET variants. Am J Hum Genet. 76: 850-8.

37. Burzynski GM., Delalande JM, Shepherd I. (2009) Characterization of spatial and temporal expression pattern of SCGl0 during zebrafish development. Gene Expr. patterns. 9: 231-237.

38. Cacalano G, Farinas I, Wang LC, Hagler K, Forgie A, et al. (1998) GFRalphal is an essential receptor component for GDNF in the developing nervous system and kidney. Neuron. 21: 53-62.

39. Cacheux V, Dastot-Le Moal F, Kaariainen H, Bondurand N, Rintala R, et al. (2001) Loss-of-function mutations in SIPl Smad interacting protein 1 result in a syndromic Hirschsprung disease. Hum Mol Genet. 10: 1503-10.

150 References

40. Cai X, Lietha D, Ceccarelli DF, Karginov AV, Rajfur Z, et al. (2008} Spatial and temporal regulation of focal adhesion kinase activity in living cells. Mal Cell Biol. 28: 201-214.

41. Carlomagno F, Vitagliano D, Guida T, Ciardiello F, Tortora G, et al. (2002} ZD6474, an orally available inhibitor of KOR tyrosine kinase activity, efficiently blocks oncogenic RET kinases. Cancer Res. 62: 7284-7290.

42. Carney JA, Go VL, Sizemore GW, Hayles AB. (1976} Alimentary-tract ganglioneuromatosis. A major component of the syndrome of multiple endocrine neoplasia, type 2b. N Engl J Med. 295: 1287-1291.

43. Carrasquillo MM, Mccallion AS, Puffenberger EG, Kashuk CS, Nouri N, et al. (2002} Genome-wide association study and mouse model identify interaction between RET and EDNRB pathways in Hirschsprung disease. Nat Genet. 32: 237-44.

44. Chakravarti A, Lyonnet S. (2001) Hirschsprung disease. In The metabolic and molecular bases of inherited disease. E.Scriver, editor. McGraw-Hill, New York. 6231-6255.

45. Chen CT, Wang JC, Cohen BA. (2007} The strength of selection on ultraconserved elements in the human genome. Am J Hum Genet. 80: 692-704.

46. Chiariello M, Visconti R, Carlomagno F, Melillo RM, Bucci C, et al. (1998) Signalling of the Ret receptor tyrosine kinase through the c-Jun NH2-terminal protein kinases (JNKS): evidence for a divergence of the ERKs and JNKs pathways induced by Ret. Oncogene. 16: 2435-45.

47. Chilton JK (2006) Molecular mechanisms of axon guidance. Dev Biol 292: 13-24.

48. Cohen EE, Needles BM, Cullen KJ, Wong SJ, Wade JL, et al. (2008) Phase 2 study of sunitinib in refractory thyroid cancer. J Clin Oneal 26: 2008: 6025

49. Cohen EE, Rosen LS, Vokes EE, Kies MS, Forastiere AA, et al. (2008) Axitinib is an active treatment for all histologic subtypes of advanced thyroid cancer: results from a phase II study. J Clin Oneal 26: 4708-4713.

50. Cohen JC, Kiss RS, Pertsemlidis A, Marcel YL, McPherson R. et al. (2004) Multiple rare alleles contribute to low plasma levels of HDL cholesterol. Science. 305: 869-72.

51. Cohen MS, Hussain HB, Maley JF. (2002) Inhibition of medullary thyroid carcinoma cell proliferation and RET phosphorylation by tyrosine kinase inhibitors. Surgery. 132: 960-967.

52. Cooper OS, Doherty GM, Haugen BR, Kloos RT, Lee SL, et al. (2006} Management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 16: 109-142.

53. Delalande JM, Barlow AJ, Thomas AJ, Wallace AS, Thapar N, et al. (2008) The receptor tyrosine kinase RET regulates hindgut colonization by sacral neural crest cells. Dev. Biol. 313: 279-292.

151 References

54. Demetri GD, van Mehren M, Blanke CD, Van den Abbeele AD, Eisenberg B, et al. (2002) Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors. N Engl J Med. 347: 472- 480.

55. Dent E.W., Gertler F.B. (2003) Cytoskeletal dynamics and transport in growth cone motility and axon guidance. Neuron. 40, 209-227.

56. Di Paolo G., Lutjens R., Osen-Sand A., Sobel A., Catsicas S., Grenningloh G. (1997) Differential distribution of stathmin and SCGlO in developing neurons in culture. J. Neurosci. Res. 50, 1000- 1009.

57. Di Paolo G., Lutjens R., Pellier V., Stimpson S.A., Beuchat M.H., Catsicas S., Grenningloh G. (1997) Targeting of SCGl0 to the area of the Golgi complex is mediated by its NH2-terminal region. J. Biol. Chem. 272: 5175-5182.

58. Diks SH, Kok K, O'Toole T, Hammes OW, van Dijken P, et al. (2004) Kinome profiling for studying lipopolysaccharide signal transduction in human peripheral blood mononuclear cells. J Biol Chem. 279: 49206-49213.

59. Diks SH, Parikh K, van der Sijde M, Joore J, Ritsema T, et al. (2007) Evidence for a minimal eukaryotic phosphoproteome? PLoS One. 2: e777.

60. Doray B, Salomon R, Amiel J, Pelet A, Touraine R, et al. (1998) Mutation of the RET ligand, neurturin, supports multigenic inheritance in Hirschsprung disease. Hum Mal Genet. 7: 1449-52.

61. Draper BW, Morcos PA, Kimmel CB. (2001) Inhibition of zebrafish fgf8 pre-mRNA splicing with morpholino oligos: a quantifiable method for gene knockdown. Genesis. 30: 154-156.

62. Druker BJ, Talpaz M, Resta DJ, Peng B, Buchdunger E, et al. (2001) Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N Engl J Med. 344: 1031-1037

63. Duncan JE, Goldstein LS. (2006) The genetics of axonal transportand axonal transport disorders. PLoS Genet. 2: e124.

64. Edery P, Lyonnet S, Mulligan LM, Pelet A, Dow E, et al. (1994) Mutations of the RET proto­ oncogene in Hirschsprung's disease. Nature. 367: 378-80.

65. Edery P, Attie T, Amie! J, Pelet A, Eng C, et al. (1996) Mutation of the endothelin-3 gene in the Waardenburg-Hirschsprung disease (Shah-Waardenburg syndrome). Nat Genet. 12: 442-4.

66. Emison ES, Mccallion AS, Kashuk CS, Bush RT, Grice E, et al. (2005) A common sex-dependent mutation in a RET enhancer underlies Hirschsprung disease risk. Nature. 434: 857-63.

67. Emison ES, Garcia-Barcelo M, Grice EA, Lantieri F, Amiel J, et al. (2010) Differential contributions of rare and common, coding and noncoding Ret mutations to multifactorial Hirschsprung disease liability. Am J Hum Genet. 87: 60-74.

68. Encinas M, Crowder RJ, Milbrandt J, Johnson EM Jr. (2004) Tyrosine 981, a novel ret autophosphorylation site, binds c-Src to mediate neuronal survival.J Biol Chem. 279: 18262-9.

152 References

69. Eng C, Mulligan LM (1997) Mutations of the RET proto-oncogene in the multiple endocrine neoplasia type 2 syndromes, related sporadic tumours, and Hirschsprung disease. Hum Mutat. 9: 97- 109

70. Fagin JA. (2002) Minireview: branded from the start-distinct oncogenic initiating events may determine tumor fate in the thyroid. Mal Endocrinol. 16: 903-11.

71. Fehrmann RS, de Jonge HJ, Ter Elst A, de Vries A, Crijns AG, et al. (2008) A new perspective on transcriptional system regulation (TSR): towards TSR profiling. PLoS One. 3: e1656.

72. Fitze G, Appelt H, Konig IR, Gorgens H, Stein U, et al. (2003) Functional haplotypes of the RET proto-oncogene promoter are associated with Hirschsprung disease (HSCR). Hum Mal Genet. 12: 3207-14.

73. Fisher S, Grice EA, Vinton RM, Bessling SL, McCallion AS. (2006) Conservation of RET regulatory function from human to zebrafish without sequence similarity. Science. 312: 276-9.

74. Flynn B, Bergner AJ, Turner KN, Young HM, Anderson RB. (2007) Effect of Gdnf haploinsufficiency on rate of migration and number of enteric neural crest-derived cells. Dev Dyn. 236: 134-41.

75. Frank-Raue K, Fabel M, Delorme S, Haberkorn U, Raue F. (2007) Efficacy of imatinib mesylate in advanced medullary thyroid carcinoma. Eur J Endocrinol. 157: 215-220.

76. Freche B, Guillaumot P, Charmetant J, Pelletier L, Luquain C, et al. (2005) Inducible dimerization of RET reveals a specific AKT deregulation in oncogenic signaling. J Biol Chem. 280: 36584-91.

77. Frikke-Schmidt R, Nordestgaard BG, Jensen GB, Tybjaerg-Hansen A (2004) Genetic variation in ABC transporter Al contributes to HDL cholesterol in the general population. J Clin Invest. 114: 1343-53.

78. Fuhler GM, Baanstra M, Chesik D, Somasundaram R, Seckinger A, et al. (2010) Bone marrow stromal cell interaction reduces Syndecan-1 expression and induces kinomic changes in myeloma cells. Exp Cell Res. 316: 1816-28.

79. Fukuda T, Kiuchi K, Takahashi M. (2002) Novel mechanism of regulation of Rae activity and lamellipodia formation by RET tyrosine kinase. J Biol Chem. 277: 19114-21.

80. Gabriel SB, Salomon R, Pelet A, Angrist M, Amie! J, et al. (2002) Segregation at three loci explains familial and population risk in Hirschsprung disease. Nat Genet. 31: 89-93.

81. Garcia-Barcelo M, Ganster RW, Lui VC, Leon TY, So MT, et al. (2005) TTF-1 and RET promoter SNPs: regulation of RET transcription in Hirschsprung's disease. Hum Mal Genet. 14: 191-204.

82. Garcia-Barcelo MM, Tang CS, Ngan ES, Lui VC, Chen Y, et al. (2009) Genome-wide association study identifies NRGl as a susceptibility locus for Hirschsprung's disease. Proc Natl Acad Sci U S A. 106: 2694-9.

153 References

83. Garver KL, Law JC, Garver B. (1985) Hirschsprung disease: a genetic study. Clin Genet. 28: 503-8.

84. Geneste 0, Bidaud C, De Vita G, Hofstra RM, Tartare-Deckert S, et al. (1999) Two distinct mutations of the RET receptor causing Hirschsprung's disease impair the binding of signalling effectors to a multifunctional docking site. Hum Mol Genet. 8: 1989-99.

85. Gianino S, Grider JR, Cresswell J, Enomoto H, Heuckeroth RO. (2003) GDNF availability determines enteric neuron number by controlling precursor proliferation. Development. 130: 2187- 98.

86. Gilmour DT, Maischein HM, Niisslein-Volhard C. (2002) Migration and function of a glial subtype in the vertebrate peripheral nervous system. Neuron. 34: 577-588.

87. Goldberg RB, Shprintzen RJ. (1981) Hirschsprung megacolon and cleft palate in two sibs. J Craniofac Genet Dev Biol. 1: 185-9.

88. Goldberg EL. (1984) An epidemiological study of Hirschsprung's disease. Int J Epidemiol. 13: 479- 85.

89. de Graaff E, Srinivas S, Kilkenny C, D'Agati V, Mankoo BS, et al. (2001) Differential activities of the RET tyrosine kinase receptor isoforms during mammalian embryogenesis. Genes Dev. 15: 2433-44.

90. Greco A, Miranda C, Pierotti MA. (2010) Rearrangements of NTRKl gene in papillary thyroid carcinoma. Mol Cell Endocrinol. 321: 44-49.

91. Greene LA, Tischler AS. (1982) PCl2 pheochromocytoma cultures in neurobiological research. Adv. Cell. Neurobiol. 3, 373-414.

92. Grenningloh G, Soehrman S, Bondallaz P, Ruchti E, Cadas H. (2004) Role of the microtubule destabilizing proteins SCGlO and stathmin in neuronal growth. J. Neurobiol. 58, 60-69.

93. Grice EA, Rochelle ES, Green ED, Chakravarti A, Mccallion AS. (2005) Evaluation of the RET regulatory landscape reveals the biological relevance of a HSCR-implicated enhancer. Hum Mol Genet. 14: 3837-45.

94. Grimm J, Sachs M, Britsch S, Di Cesare S, Schwarz-Romond T, et al. (2001) Novel p62dok family members, dok-4 and dok-5, are substrates of the c-Ret receptor tyrosine kinase and mediate neuronal differentiation. J Cell Biol. 154: 345-54.

95. Griseri P, Sachetti T, Puppo F, Lantieri F, Ravazzolo R, et al. (2005) A common haplotype at the 5' end of the RET proto-oncogene, overrepresented in Hirschsprung patients, is associated with reduced gene expression. Hum Mutat. 25: 189-95.

96. de Groot JW, Links TP, Plukker JT, Lips CJ, Hofstra RM. (2006) RET as a diagnostic and therapeutic target in sporadic and hereditary endocrine tumors. Endocr Rev. 27: 535-60.

154 References

97. de Groot JW, Plaza Menacho I, Schepers H, Drenth-Diephuis LI, Osinga J, et al. (2006) Cellular effects of imatinib on medullary thyroid cancer cells harboring multiple endocrine neoplasia Type 2A and 2B associated RET mutations. Surgery. 139: 806-814.

98. de Groot JW, Zonnenberg BA, van Ufford-Mannesse PQ, de Vries MM, Links TP, et al. (2007) A phase II trial of imatinib therapy for metastatic medullary thyroid carcinoma. J Clin Endocrinol Metab. 92: 3466-3469.

99. Haddad GG, Mazza NM, Defendini R, Blanc WA, Driscoll JM, et al. (1978) Congenital failure of automatic control of ventilation, gastrointestinal motility and heart rate. Medicine (Baltimore). 7: 517-26.

100. Hansford JR, Mulligan LM. (2000) Multiple endocrine neoplasia type 2 and RET: from neoplasia to neurogenesis. J Med Genet. 37: 817-27.

101. Heanue TA, Pachnis V. (2006) Expression profiling the developing mammalian enteric nervous system identifies marker and candidate Hirschsprung disease genes. Proc Natl Acad Sci U S A. 103: 6919-24.

102. Heanue TA, Pachnis V. (2007) Enteric nervous system development and Hirschsprung's disease: advances in genetic and stem cell studies. Nat Rev Neurosci. 8: 466-79.

103. Hearn 0, Murphy M, Newgreen D. (1998) GDNF and ET-3 differentially modulate the numbers of avian enteric neural crest cells and enteric neurons in vitro. Dev Biol. 197: 93-105.

104. Hofstra RM, Landsvater RM, Ceccherini I, Stulp RP, Stelwagen T, et al. (1994) A mutation in the RET proto-oncogene associated with multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature. 367: 375-376.

105. Hofstra RM, Osinga J, Tan-Sindhunata G, Wu Y, Kamsteeg EJ, et al. (1996) A homozygous mutation in the endothelin-3 gene associated with a combined Waardenburg type 2 and Hirschsprung phenotype (Shah-Waardenburg syndrome). Nat Genet. 12: 445-7.

106. Hofstra RM, Valdenaire 0, Arch E, Osinga J, Kroes H, et al. {1999) A loss-of-function mutation in the endothelin-converting enzyme 1 (ECE-1) associated with Hirschsprung disease, cardiac defects, and autonomic dysfunction. Am J Hum Genet. 64: 304-8

107. Hofstra RM, Wu Y, Stulp RP, Elfferich P, Osinga J, et al. (2000) RET and GDNF gene scanning in Hirschsprung patients using two dual denaturing gel systems. Hum Mutat. 15: 418-29.

108. Hoischen A, van Bon BW, Gilissen C, Arts P, van Lier B, et al. (2010) De nova mutations of SETBPl cause Schinzel-Giedion syndrome. Nat Genet. 42: 483-5.

109. Holschneider AM. (1982) Hirschsprung's congenital megacolon. The concept of physiopathology and therapy. Med Welt. 33: 210-3.

155 References

110. Hoogenraad C.C., Akhmanova A., Howell S.A., Dortland B.R., De Zeeuw C.I., Willemsen R., Visser P., Grosveld F., Galjart N. (2001) Mammalian Golgi-associated Bicaudal-D2 functions in the dynein­ dynactin pathway by interacting with these complexes. EMBO J. 20, 4041-4054.

111. Howard J, Hyman AA. (2003) Dynamics and mechanics of the microtubule plus end. Nature. 422, 753-758.

112. lchihara M, Murakumo Y, Takahashi M. (2004) RET and neuroendocrine tumors. Cancer Lett. 204: 197-211.

113. Ikeda K, Goto S. (1984) Diagnosis and treatment og Hirschsprung's disease in Japan. An analysis of 1628 patients. Ann. Surg. 199: 400-405.

114. Ioannidis JP, Trikalinos TA, Khoury MJ. (2006) Implications of small effect sizes of individual genetic variants on the design and interpretation of genetic association studies of complex diseases. Am J Epidemiol. 164: 609-14.

115. Iwashita T, Asai N, Murakami H, Matsuyama M, Takahashi M. (1996) Identification of tyrosine residues that are essential for transforming activity of the ret proto-oncogene with MEN2A or MEN2B mutation. Oncogene. 12: 481-7.

116. Iwashita T, Kato M, Murakami H, Asai N, lshiguro Y, et al. (1999) Biological and biochemical properties of Ret with kinase domain mutations identified in multiple endocrine neoplasia type 2B and familial medullary thyroid carcinoma. Oncogene. 18: 3919-3922.

117. Iwashita T, Kruger GM, Parda! R, Kiel MJ, Morrison SJ. (2003) Hirschsprung disease is linked to defects in neural crest stem cell function. Science. 301: 972-6.

118. Jing S, Wen D, Yu Y, Holst PL, Luo Y, et al. (1996) GDNF-induced activation of the ret protein tyrosine kinase is mediated by GDNFR-alpha, a novel receptor for GDNF. Cell. 85: 1113-24.

119. Johansen CT, Wang J, Lanktree MB, Cao H, McIntyre AD, et al. (2010) Excess of rare variants in genes identified by genome-wide association study of hypertriglyceridemia. Nat Genet. 42: 684-7.

120. Johnson MS. (1967) Hierarchical clustering schemes. Psychometrika. 32: 241-54.

121. Kawamoto Y, Takeda K, Okuno Y, Yamakawa Y, Ito Y, et al. (2004) Identification of RET autophosphorylation sites by mass spectrometry. J Biol Chem. 279: 14213-24.

122. Kennedy NJ, Davis RJ. (2003) Role of JNK in tumor development. Cell Cycle. 2: 199-201.

123. Kiberstis P, Roberts L. (2002) It's Not Just the Genes. Science. 296: 5568-685.

124. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. (1995) Stages of embryonic development of the zebrafish. Dev. Dyn. 203: 253-310.

156 References

125. Kodama Y, Asai N, Kawai K, Jijiwa M, Murakumo Y, et al. (2005) The RET proto-oncogene: a molecular therapeutic target in thyroid cancer. Cancer Sci. 96: 143-148.

126. Kumar S, Boehm J, Lee JC. (2003) p38 MAP kinases: key signalling molecules as therapeutic targets for inflammatory diseases. Nat Rev Drug Discov. 2: 717-26.

127. Kurai M, Shiozawa T, Shih HC, Miyamoto T, Feng YZ, et al. (2005) Expression of Aurora kinases A and B in normal, hyperplastic, and malignant human endometrium: Aurora B as a predictor for poor prognosis in endometrial carcinoma. Hum Pathol. 36: 1281-1288.

128. Kurokawa K, Iwashita T, Murakami H, Hayashi H, Kawai K, et al. (2001) Identification of SNT/FRS2 docking site on RET receptor tyrosine kinase and its role for signal transduction. Oncogene. 20: 1929-38.

129. Kurzrock R, Cohen EE, Sherman SI, Pfister DG, Cohen RB, et al. (2010) Long-term results in a cohort of medullary thyroid cancer (MTC) patients (pts) in a phase I study of XL184 (BMS 907351), an oral inhibitor of MET, VEGFR2, and RET. J Clin Oneal 28: 15s (suppl; abstr 5502).

130. Lam ET, Ringel MD, Kloos RT, Prior TW, Knopp MV, et al. (2010) Phase II clinical trial of sorafenib in metastatic medullary thyroid cancer. J Clin Oneal. 28: 2323-2830.

131. Lander ES, Schork NJ. (1994) Genetic dissection of complex traits. Science. 265: 2037-48.

132. Le Douarin N, Kalcheim C. The neural crest. 2nd ed. Cambridge: Cambridge University Press, 1999.

133. Lee HO, Levorse JM, Shin MK. (2003) The endothelin receptor-a is required for the migration of neural crest-derived melanocyte and enteric neuron precursors. Dev Biol. 259: 162-75.

134. Leibl MA, Ota T, Woodward MN, Kenny SE, Lloyd DA, et al. (1999) Expression of endothelin 3 by mesenchymal cells of embryonic mouse caecum. Gut. 44: 246-52.

135. Lesnick TG, Papapetropoulos S, Mash DC, Ffrench-Mullen J, Shehadeh L, et al. (2007) A genomic pathway approach to a complex disease: axon guidance and Parkinson disease. PLoS Genet. 3: e98.

136. Levy S, Sutton G, Ng PC, Feuk L, Halpern AL. et al. (2007) The diploid genome sequence of an individual human. PLoS Biol. 5: e254.

137. Liem KF Jr, He M, Ocbina PJ, Anderson KV. (2009) Mouse Kif7 /Costal2 is a cilia-associated protein that regulates Sonic hedgehog signaling. Proc. Natl. Acad. Sci. U.S.A. 106: 13377-13382.

138. Linstedt AD, Hauri HP. (1993) Giantin, a novel conservedGolgi membrane protein containing a cytoplasmic domain of at least 350 kDa. Mol. Biol. Cell. 4: 679-693.

139. Lister JA, Cooper C, Nguyen K, Modrell M, Grant K, Raible DW. (2006) Zebrafish Foxd3 is required for development of a subset of neural crest derivatives. Dev. Biol. 290: 92-104.

157 References

140. Livesey FJ, Hunt SP (1997) Netrin and netrin receptor expression in the embryonic mammalian nervous system suggests roles in retinal, striatal, nigral, and cerebellar development. Mol Cell Neurosci. 8: 417-429.

141. Liu X, Vega QC, Decker RA, Pandey A, Worby CA, et al. (1996) Oncogenic RET receptors display different autophosphorylation sites and substrate binding specificities. J Biol Chem. 271: 5309-12.

142. Lorenzo MJ, Gish GD, Houghton C, Stonehouse TJ, Pawson T, et al. (1997) RET alternate splicing influences the interaction of activated RET with the SH2 and PTB domains of She, and the SH2 domain of Grb2. Oncogene. 14: 763-71.

143. Lowenberg M, Tuynman J, Bilderbeek J, Gaber T, Buttgereit F, et al. (2005) Rapid immunosuppressive effects of glucocorticoids mediated through Lek and Fyn. Blood. 106: 1703- 1710.

144. Lutjens R, lgarashi M, Pellier V, Blasey H, Di Paolo G, et al. (2000) Localization and targeting of SCGl0 to the trans-Golgi apparatus and growth cone vesicles. Eur. J. Neurosci. 12: 2224-2234.

145. Lyons DA, Naylor SG, Mercurio S, Dominguez C, Talbot WS. (2008) KBP is essential for axonal structure, outgrowth and maintenance in zebrafish, providing insight into the cellular basis of Goldberg-Shprintzen syndrome. Development. 135: 599-608.

146. Manie S, Santoro M, Fusco A, Billaud M. (2001) The RET receptor: function in development and dysfunction in congenital malformation. Trends Genet. 17: 580-9.

147. Manna T, Grenningloh G, Miller HP, Wilson L. (2007) Stathmin family protein SCGl0 differentially regulates the plus and minus end dynamics of microtubules at steady state in vitro: implications for its role in neurite outgrowth. Biochemistry. 46, 3543-3552.

148. Maraganore DM, de Andrade M, Lesnick TG, Strain KJ, Farrer MJ, et al. (2005) High-resolution whole-genome association study of Parkinson disease. Am J Hum Genet 77: 685-693.

149. Marcos C, Pachnis V. (1996) The effect of the ret- mutation on the normal development of the central and parasympathetic nervous systems. Int J Dev Biol. Suppl 1: 137S-138S.

150. Marx SJ. (2005) Molecular genetics of multiple endocrine neoplasia types 1 and 2. Nat Rev Cancer. 5: 367-75.

151. Matera I, Bachetti T, Puppo F, Di Duca M, Morandi F, et al. (2004) PHOX2B mutations and polyalanine expansions correlate with the severity of the respiratory phenotype and associated symptoms in both congenital and late onset Central Hypoventilation syndrome. J Med Genet. 41: 373-80.

152. McArthur GA, Demetri GD, van Oosterom A, Heinrich MC, Debiec-Rychter M, et al. (2005) Molecular and clinical analysis of locally advanced dermatofibrosarcoma protuberans treated with imatinib: lmatinib Target Exploration Consortium Study B2225. J Clin Oneal. 23: 866-873.

158 References

153. Mccallion AS, Starnes E, Conlon RA, Chakravarti A. {2003) Phenotype variation in two-locus mouse models of Hirschsprung disease: tissue-specific interaction between Ret and Ednrb. Proc Natl Acad Sci U SA. 100: 1826-31

154. McKernan KJ, Peckham HE, Costa GL, McLaughlin SF, Fu Y, et al. (2009) Sequence and structural variation in a human genome uncovered by shortread, massively parallel ligation sequencing using twobase encoding. Genome Res. 19: 1527-1541.

155. McLean GW, Carragher NO, Avizienyte E, Evans J, Brunton VG, et al. (2005) The role of focal­ adhesion kinase in cancer - a new therapeutic opportunity. Nat Rev Cancer. 5: 505-15.

156. Melillo RM, Santoro M, Ong SH, Billaud M, Fusco A, et al. (2001) Docking protein FRS2 links the protein tyrosine kinase RET and its oncogenic forms with the mitogen-activated protein kinase signaling cascade. Mal Cell Biol. 21: 4177-87.

157. Melillo RM, Carlomagno F, De Vita G, Formisano P, Vecchio G, et al. (2001) The insulin receptor substrate (IRS)-1 recruits phosphatidylinositol 3-kinase to Ret: evidence for a competition between She and IR5-1 for the binding to Ret. Oncogene. 20: 209-18.

158. Metzger M, Caldwell C, Barlow AJ, Burns AJ, Thapar N. (2009) Enteric nervous system stem cells derived from human gut mucosa for the treatment of aganglionic gut disorders. Gastroenterology. 136: 2214-25. el-3.

159. Modigliani E, Cohen R, Campos JM, Conte-Devolx B, Maes B, et al. (1998) Prognostic factors for survival and for biochemical cure in medullary thyroid carcinoma: results in 899 patients. The GETC Study Group. Groupe d'etude des tumeurs a calcitonine. Clin Endocrinol. 48: 265-273.

160. Mograbi B, Bocciardi R, Bourget I, Juhel T, Farahi-Far D, et al. (2001) The sensitivity of activated Cys Ret mutants to glial cell line-derived neurotrophic factor is mandatory to rescue neuroectodermic cells from apoptosis. Mal Cell Biol. 21: 6719-30.

161. Maley JF, Fialkowski EA. (2007) Evidence-based approach to the management of sporadic medullary thyroid carcinoma. World J Surg. 31: 946-956.

162. Moore MW, Klein RD, Farinas I, Sauer H, Armanini M, et al. (1996) Renal and neuronal abnormalities in mice lacking GDNF. Nature. 382: 76-9.

163. Moore SW (2006) The contribution of associated congenital anomalies in understanding Hirschsprung's disease. Pediatr Surg Int 22:305-315.

164. Mountzios G, Terpos E, Dimopoulos MA. (2008). Aurora kinases as targets for cancer therapy. Cancer Treat Rev. 34: 175-182.

165. Mowat DR, Croaker GD, Cass DT, Kerr BA, Chaitow J, et al. (1998) Hirschsprung disease, microcephaly, mental retardation, and characteristic facial features: delineation of a new syndrome and identification of a locus at chromosome 2q22-q23. J Med Genet. 35: 617-23.

159 References

166. Murakami H, Iwashita T, Asai N, Iwata Y, Narumiya S, et al. (1999) Rho-dependent and - independent tyrosine phosphorylation of focal adhesion kinase, paxillin and p130Cas mediated by Ret kinase. Oncogene. 18: 1975-82.

167. Murakami H, Iwashita T, Asai N, Shimano Y, Iwata Y, et al. (1999). Enhanced phosphatidylinositol 3-kinase activity and high phosphorylation state of its downstream signalling molecules mediated by ret with the MEN 2B mutation. Biochem Biophys Res Commun. 262: 68-75.

168. Murakami H, Yamamura Y, Shimano Y, Kawai K, Kurokawa K, et al. (2002) Role of Dokl in cell signaling mediated by RET tyrosine kinase. J Biol Chem. 277: 32781-90.

169. Myers SM, Eng C, Ponder BA, Mulligan LM. (1995) Characterization of RET proto-oncogene 3' splicing variants and polyadenylation sites: a novel C-terminus for RET. Oncogene. 11: 2039-45.

170. Nagy N, Goldstein AM. (2006) Endothelin-3 regulates neural crest cell proliferation and differentiation in the hindgut enteric nervous system. Dev Biol. 293: 203-17.

171. Natarajan D, Marcos-Gutierrez C, Pachnis V, de Graaff E. (2002) Requirement of signalling by receptor tyrosine kinase RET for the directed migration of enteric nervous system progenitor cells during mammalian embryogenesis. Development. 129: 5151-60.

172. Ng SB, Turner EH, Robertson PD, Flygare SD, Bigham AW, et al. (2009) Targeted capture and massively parallel sequencing of 12 human exomes. Nature. 461: 272-6.

173. Ng SB, Buckingham KJ, Lee C, Bigham AW, Tabor HK, et al. (2010) Exome sequencing identifies the cause of a mendelian disorder. Nat Genet. 42: 30-5.

174. Nikiforov YE. (2002) RET/PTC rearrangement in thyroid tumors. Endocr Pathol. 13: 3-16.

175. Nikoforov YE. (2008) Thyroid carcinoma: molecular pathways and therapeutic targets. Mod Pathol. 21: Suppl 2:S37-43.

176. Nixon AB, Grenningloh G, Casey PJ. (2002) The interaction of RGSZl with SCGl0 attenuates the ability of SCGl0 to promote microtubule disassembly. J. Biol. Chem. 277: 18127-18133.

177. Nozaki C, Asai N, Murakami H, Iwashita T, Iwata Y, et al. (1998) Calcium-dependent Ret activation by GDNF and neurturin. Oncogene. 16(3):293-9.

178. Omenn GS, McKusick VA. (1979) The association of Waardenburg syndrome and Hirschsprung megacolon. Am J Med Genet. 3: 217-23.

179. Oksvold MP, Thien CBF, Widerberg J, Chantry A, Huitfeldt HS et al. (2003) Serine mutations that abrogate ligand-induced ubiquitination and internalization of the EGF receptor do not affect c-Cbl association with the receptor. Oncogene 22: 8509-8518.

180. Pachnis V, Mankoo B, Costantini F. (1993) Expression of the c-ret proto-oncogene during mouse embryogenesis. Development. 119: 1005-17.

160 References

181. Pachnis V, Durbec P, Taraviras S, Grigoriou M, Natarajan D. {1998} Ill. Role Of the RET signal transduction pathway in development of the mammalian enteric nervoussys tem. Am J Physiol. 275: G183-6.

182. Pacini F, Schlumberger M, Dralle H, Elisei R, Smit JW, et al. {2006} European consensus for the management of patients with differentiated thyroid carcinoma of the follicular epithelium. Eur J Endocrinol. 154: 787-803.

183. Panta GR, Nwariaku F, Kim LT. {2004} RET signals through focal adhesion kinase in medullary thyroid cancer cells. Surgery. 136: 1212-1217.

184. Parikh K, Poppema S, Peppelenbosch MP, Visser L. {2009} Extracellular ligation-dependent CD45RB enzymatic activity negatively regulates lipid raft signal transduction. Blood. 113: 594-603.

185. Parikh K, Peppelenbosch MP. (2010} Kinome profiling of clinical cancer specimens. Cancer Res. 70: 2575-2578.

186. Parisi MA, Kapur RP. {2000} Genetics of Hirschsprung disease. Curr Opin Pediatr. 12: 610-7.

187. Pasini B, Hofstra RM, Yin L, Bocciardi R, Santamaria G, et al. {1995} The physical map of the human RET proto-oncogene. Oncogene. 11: 1737-43.

188. Passarge E. {1967} The genetics of Hirschsprung's disease. Evidence for heterogeneous etiology and a study of sixty-three families. N Engl J Med. 276: 138-43.

189. Pattyn A, Morin X, Cremer H, Goridis C, Brunet JF. {1999) The homeobox gene Phox2b is essential for the development of autonomic neural crest derivatives. Nature. 399: 366-70.

190. Pelet A, de Pontual L, Clement-Ziza M, Salomon R, Mugnier C, et al. (2005) Homozygosity for a frequent and weakly penetrant predisposing allele at the RET locus in sporadic Hirschsprung disease. J Med Genet. 42: el8.

191. Pelicci G, Troglio F, Bodini A, Melillo RM, Pettirossi V, et al. (2002) The neuron-specific Rai {ShcC) adaptor protein inhibits apoptosis by coupling Ret to the phosphatidylinositol 3-kinase/Akt signaling pathway. Mol Cell Biol. 22: 7351-63.

192. Pennell NA, Daniels GH, Haddad RI, Ross DS, Evans T, et al. 2008 A phase II study of gefitinib in patients with advanced thyroid cancer. Thyroid. 18: 317-323.

193. Pichel JG, Shen L, Sheng HZ, Granholm AC, Drago J, et al. {1996) Defects in enteric innervation and kidney development in mice lacking GDNF. Nature. 382: 73-6.

194. Pierotti MA, Vigneri P, Bongarzone I. {1998) Rearrangements of RET and NTRKl tyrosine kinase receptors in papillary thyroid carcinomas. Recent Results Cancer Res. 154: 237-47.

195. Pingault V, Bondurand N, Kuhlbrodt K, Goerich DE, Prehu MO, et al. (1998) SOXl0 mutations in patients with Waardenburg-Hirschsprung disease. Nat Genet. 18: 171-3.

161 References

196. Plaster N, Sonntag C, Busse CE, Hammerschmidt M. (2006) p53 deficiency rescues apoptosis and differentiation of multiple cell types in zebrafish flathead mutants deficient for zygotic DNA polymerase deltal. Cell. Death. Differ. 13: 223-235.

197. Plaza Menacho I, Koster R, van der Sloat AM, Quax WJ, Osinga J, et al. (2005). RET-familial medullary thyroid carcinoma mutants Y791F and S891A activate a Src/JAK/STAT3 pathway, independent of glial cell line-derived neurotrophic factor. Cancer Res. 65: 1729-1737.

198. Plaza-Menacho I, Burzynski GM, de Groot JW, Eggen BJ, Hofstra RM. (2006) Current concepts in RET-related genetics, signaling and therapeutics. Trends Genet. 22: 627-36.

199. Plaza-Menacho I, Mologni L, Sala E, Gambacorti-Passerini C, Magee Al, et al. (2007) Sorafenib functions to potently suppress RET tyrosine kinase activity by direct enzymatic inhibition and promoting RET lysosomal degradation independent of proteasomal targeting. J Biol Chem. 282: 29230-29240.

200. Plaza-Menacho I, van der Sluis T, Hollema H, Gimm 0, Buys CH, et al. (2007) Ras/ERKl/2- mediated STAT3 Ser727 phosphorylation by familial medullary thyroid carcinoma-associated RET mutants induces full activation of STAT3 and is required for c-fos promoter activation, cell mitogenicity, and transformation. J Biol Chem. 282: 6415-24.

201. de Pontual L, Pelet A, Clement-Ziza M, Trochet D, Antonarakis SE, et al. (2007) Epistatic interactions with a common hypomorphic RET allele in syndromic Hirschsprung disease. Hum Mutat. 28: 790-6.

202. Puffenberger EG, Hosoda K, Washington SS, Nakao K, deWit D, et al. (1994) A missense mutation of the endothelin-B receptor gene in multigenic Hirschsprung's disease. Cell. 79: 1257-66.

203. Pushkarev D, Neff NF, Quake SR. (2009) Single-molecule sequencing of an individual human genome. Nature Biotech. 27: 847-850.

204. Richardson DS, Lai AZ, Mulligan LM. (2006) RET ligand-induced internalization and its consequences for downstream signaling. Oncogene. 25: 3206-3211.

205. Riederer BM, Pellier V, Antonsson B, Di Paolo G, Stimpson SA, et al. (1997) Regulation of microtubule dynamics by the neuronal growth-associated protein SCGl0. Proc. Natl. Acad. Sci. U SA. 94: 741-745.

206. Ro S, Hwang SJ, Muto M, Jewett WK, Spencer NJ. (2006) Anatomic modifications in the enteric nervous system of piebald mice and physiological consequences to colonic motor activity. Am J Physiol Gastrointest Liver Physiol. 290: G710-8.

207. Robinson BG, Paz-Ares L, Krebs A, Vasselli J, Haddad R. (2010) Vandetanib (100 mg) in patients with locally advanced or metastatic hereditary medullary thyroid cancer. J Clin Endocrinol Metab 95: 2664-2671.

208. Robu ME, Larson JD, Nasevicius A, Beiraghi S, Brenner C, et al. (2007) p53 activation by knockdown technologies. PloS Genet. 3: e78.

162 References

209. Rochlin MW, Wickline KM, Bridgman PC. {1996) Microtubule stability decreases axon elongation but not axoplasm production. J. Neurosci. 16: 3236-3246.

210. Romeo G, Ronchetto P, Luo Y, Barone V, Seri M, et al. (1994) Point mutations affecting the tyrosine kinase domain of the RET proto-oncogene in Hirschsprung's disease. Nature. 367: 377-8.

211. Ropers HH. (2007) New perspectives for the elucidation of genetic disorders. Am J Hum Genet. 81: 199-207.

212. Russell MB, Russell CA, Niebuhr E. (1994) An epidemiological study of Hirschsprung's disease and additional anomalies. Acta Paediatr. 83: 68-71.

213. Salomon R, Attie T, Pelet A, Bidaud C, Eng C, et al. (1996) Germline mutations of the RET ligand GDNF are not sufficient to cause Hirschsprung disease. Nat Genet. 14: 345-7.

214. Salvatore D, Melillo RM, Monaco C, Visconti R, Fenzi G, et al. (2001) Increased in vivo phosphorylation of ret tyrosine 1062 is a potential pathogenetic mechanism of multiple endocrine neoplasia type 2B. Cancer Res. 61: 1426-31.

215. Sancandi M, Ceccherini I, Costa M, Fava M, Chen B, et al. (2000) Incidence of RET mutations in patients with Hirschsprung's disease. J Pediatr Surg. 35: 139-42.

216. Sancandi M, Griseri P, Pesce B, Patrone G, Puppo F, et al. (2003) Single nucleotide polymorphic alleles in the 5' region of the RET proto-oncogene define a risk haplotype in Hirschsprung's disease. J Med Genet. 40: 714-8.

217. Sanchez MP, Silos-Santiago I, Frisen J, He B, Lira SA, et al. (1996) Renal agenesis and the absence of enteric neurons in mice lacking GDNF. Nature. 382(6586):70-3.

218. Santoro M, Carlomagno F, Romano A, Bottaro DP, Dathan NA, et al. (1995) Activation of RET as a dominant transforming gene by germline mutations of MEN2A and MEN2B. Science. 267: 381-3.

219. Santoro M, Carlomagno F, Melillo RM, Fusco A. (2004) Dysfunction of the RET receptor in human cancer. Cell Mal Life Sci. 61: 2954-64.

220. Schindler U, Wu P, Rothe M, Brasseur M, McKnight SL. (1995) Components of a Stat recognition code: evidence for two layers of molecular selectivity. Immunity. 2: 689-97.

221. Schlumberger M, Carlomagno F, Baudin E, Bidart JM, Santoro M. (2008) New therapeutic approaches to treat medullary thyroid carcinoma. Nat Clin Pract Endocrinol Metab. 4: 22-32.

222. Schlumberger MJ, Elisei R, Bastholt L, Wirth LI, Martins RG, et al. (2009) Phase II study of safety and efficacy of motesanib in patients with progressive or symptomatic, advanced or metastatic medullary thyroid cancer. J Clin Oneal. 27: 3794-3801.

223. Schneider R. (1992) The human protooncogene ret: a communicative cadherin? Trends Biochem Sci. 17: 468-9.

163 References

224. Schuchardt A, D'Agati V, Larsson-Blomberg L, Costantini F, Pachnis V. (1994) Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret. Nature. 367: 380-3.

225. Schuchardt A, D'Agati V, Larsson-Blomberg L, Costantini F, Pachnis V. (1995) RET-deficient mice: an animal model for Hirschsprung's disease and renal agenesis. J Intern Med. 238: 327-32.

226. Schuetz G, Rosario M, Grimm J, Boeckers TM, Gundelfinger ED, et al. (2004) The neuronal scaffold protein Shank3 mediates signaling and biological function of the receptor tyrosine kinase Ret in epithelial cells. J Cell Biol. 167: 945-52.

227. Schuringa JJ, Wojtachnio K, Hagens W, Vellenga E, Buys CH, et al. (2001) MEN2A-RET-induced cellular transformation by activation of STAT3. Oncogene. 20: 5350-8.

228. Segouffin-Cariou C, Billaud M. (2000) Transforming ability of MEN2A-RET requires activation of the phosphatidylinositol 3-kinase/AKT signaling pathway. J Biol Chem. 275: 3568-76.

229. Shah KN, Dalal SJ, Desai MP, Sheth PN, Joshi NC, et al. (1981) White forelock, pigmentary disorder of irides, and long segment Hirschsprung disease: possible variant of Waardenburg syndrome. J Pediatr. 99: 432-5.

230. Shakhova 0, Sommer L (2010) Neural crest-derived stem cells, In: StemBook ed. The Stem Cell Research Community, doi/10.3824/stembook.l.51.l. (http://www. stembook. org).

231. Shen L, Pichel JG, Mayeli T, Sariola H, Lu B, et al. (2002) Gdnf haploinsufficiency causes Hirschsprung-like intestinal obstruction and early-onset lethality in mice. Am J Hum Genet. 70: 435- 47.

232. Shepherd IT, Beattie CE, Raible DW. (2001) Functional analysis of zebrafish GDNF. Dev Biol. 231: 420-35.

233. Shepherd IT, Pietsch J, Elworthy S, Kelsh RN, Raible OW. (2004) Roles for GFRalphal receptors in zebrafish enteric nervous system development. Development. 131: 241-9.

234. Sherman SI. (2000) The management of metastatic differentiated thyroid carcinoma. Rev Endocr Metab Disord.1: 165-71.

235. Sherman SI. (2009) Tyrosine kinase inhibitors and the thyroid. Best Pract Res Clin Endocrinol Metab. 23: 713-22.

236. Sieg DJ, Hauck CR, Ilic D, Klingbeil CK, Schaefer E, et al. (2000) FAK integrates growth-factor and integrin signals to promote cell migration. Nat Cell Biol. 2: 249-256.

237. Sipple JH. (1961). The association of pheochromocytoma with carcinoma of the thyroid gland. Am J Med. 31: 163-164.

164 References

238. Sijmons RH, Hofstra RM, Wijburg FA, Links TP, Zwierstra RP, et al. (1998) Oncological implications of RET gene mutations in Hirschsprung's disease. Gut. 43: 542-7.

239. Smith SL, Bowers NL, Betticher DC, Gautschi 0, Ratschiller D, et al. (2005) Overexpression of aurora B kinase (AURKB) in primary non-small cell lung carcinoma is frequent, generally driven from one allele, and correlates with the level of genetic instability. Br J Cancer. 93: 719-29.

240. Spouge D, Baird PA. (1985) Hirschsprung disease in a large birthcohort. Teratology. 32: 171- 177.

241. Sun Y, Sinicrope FA. (2005) Selective inhibitors of MEK1/ERK44/42 and p38 mitogen-activated protein kinases potentiate apoptosis induction by sulindac sulfide in human colon carcinoma cells. Mol Cancer Ther. 4: 51-9.

242. Takahashi M, Ritz J, Cooper GM. (1985) Activation of a novel human transforming gene, ret, by DNA rearrangement. Cell. 42: 581-8.

243. Takahashi M, Buma Y, Taniguchi M. (1991) Identification of the ret proto-oncogene products in neuroblastoma and leukemia cells. Oncogene. 6: 297-301.

244. Takahashi M, Asai N, Iwashita T, lsomura T, Miyazaki K, et al. (1993) Characterization of the ret proto-oncogene products expressed in mouse L cells. Oncogene. 8: 2925-9.

245. Takahashi M, Iwashita T, Santoro M, Lyonnet S, Lenoir GM, et al. (1999) Co-segregation of MEN2 and Hirschsprung's disease: the same mutation of RET with both gain and loss-of-function? Hum Mutat. 13: 331-6.

246. Tam PK, Garcia-Barcelo M. (2009) Genetic basis of Hirschsprung's disease. Pediatr Surg Int. 25: 543-58.

247. Tanaka EM, Kirschner MW. (1991) Microtubule behavior in the growth cones of living neurons during axon elongation. J. Cell Biol. 115: 345-363.

248. Teslovich TM, Teslovich TM, Musunuru K, Smith AV, Edmondson AC, Stylianou IM, et al. (2010) Biological, clinical and population relevance of 95 loci for blood lipids. Nature. 466: 707-13.

249. Thisse C, Thisse 8, Schilling TF, Postlethwait JH. (1993) Structure of the zebrafish snaill gene and its expression in wild-type, spadetail and no tail mutant embryos. Development. 119: 1203- 1215.

250. Togano T, Kurachi M, Watanabe M, Grenningloh G, lgarashi M. (2005) Role of Ser50 phosphorylation in SCGl0 regulation of microtubule depolymerization. J Neurosci Res. 80: 475-80.

251. Torts CP. (1998) An epidemiological study of Hirschsprung disease in a multiracial California population. The Third International Meetings: Hirschsprung disease and related neurocristopathies. Evian, France.

16S References

252. Trang H, Laudier B, Trochet D, Munnich A, Lyonnet S, et al. (2004) PHOX2B gene mutation in a patient with late-onset central hypoventilation. Pediatr Pulmonol. 38: 349-51.

253. Trochet D, O'Brien LM, Goza! D, Trang H, Nordenskjold A, et al. (2005) PHOX2B genotype allows for prediction of tumor risk in congenital central hypoventilation syndrome. Am J Hum Genet. 76: 421-6.

254. Uesaka T, Nagashimada M, Yonemura S, Enomoto H. (2008) Diminished Ret expression compromises neuronal survival in the colon and causes intestinal aganglionosis in mice. J Clin Invest. 118: 1890-8.

255. van Baal JW, Diks SH, Wanders RJ, Rygiel AM, Milano F, et al. (2006) Comparison of kinome profiles of Barrett's esophagus with normal squamous esophagus and normal gastric cardia. Cancer Res. 66: 11605-11612.

256. van de Putte T, Maruhashi M, Francis A, Nelles L, Kondoh H, et al. (2003) Mice lacking ZFHXlB, the gene that codes for Smad-interacting protein-1, reveal a role for multiple neural crest cell defects in the etiology of Hirschsprung disease-mental retardation syndrome. Am J Hum Genet. 72: 465-70.

257. van Weering DH, Moen TC, Braakman I, Baas PD, Bos JL. (1998) Expression of the receptor tyrosine kinase Ret on the plasma membrane is dependent on calcium. J Biol Chem. 273: 12077-81.

258. Verloes A, Elmer C, Lacombe D, Heinrichs C, Rebuffat E, et al. (1993) Ondine-Hirschsprung syndrome (Haddad syndrome). Further delineation in two cases and review of the literature. Eur J Pediatr. 152: 75-7.

259. Wakamatsu N, Yamada Y, Yamada K, Ono T, Nomura N, et al. (2001) Mutations in SIPl, encoding Smad interacting protein-1, cause a form of Hirschsprung disease. Nat Genet. 27: 369-70.

260. Wasenius VM, Hemmer S, Karjalainen-Lindsberg ML, Nupponen NN, Franssila K, et al. (2005) MET receptor tyrosine kinase sequence alterations in differentiated thyroid carcinoma. Am J Surg Pathol. 29: 544-9.

261. Watanabe T, lchihara M, Hashimoto M, Shimano K, Shimoyama Y, et al. (2002) Characterization of gene expression induced by RET with MEN2A or MEN2B mutation. Am J Pathol. 161: 249-56.

262. Weese-Mayer DE, Berry-Kravis EM, Zhou L, Maher BS, Silvestri JM, et al. (2003) Idiopathic congenital central hypoventilation syndrome: analysis of genes pertinent to early autonomic nervous system embryologic development and identification of mutations in PHOX2b. Am J Med Genet A. 123A: 267-78.

263. Weese-Mayer DE, Berry-Kravis EM, Zhou L. (2005) Adult identified with congenital central hypoventilation syndrome-mutation in PHOX2b gene and late-onset CHS. Am J Respir Crit Care Med. 171: 88.

166 References

264. Wells SA Jr, Gosnell JE, Gagel RF, Maley J, Pfister D, et al. (2010) Vandetanib for the treatment of patients with locally advanced or metastatic hereditary medullary thyroid cancer. J Clin Oneal 28: 767-772.

265. Wells SA Jr, Santoro M. (2009} Targeting the RET pathway in thyroid cancer. Clin Cancer Res 15: 7119-23.

266. Westerfield M. (2000) The zebrafish book. A guide for the laboratory use of zebrafish (Dania rerio). Univ. of Oregon Press, Eugene, 4th ed.

267. Wheeler, DA, Srinivasan M, Egholm M, Shen Y, Chen L, et al. (2008) The complete genome of an individual by massively parallel DNA sequencing. Nature 452: 872-876.

268. Wiseman SM, Masoudi H, Niblock P, Turbin D, Rajput A, et al. (2007). Anaplastic thyroid carcinoma: expression profile of targets for therapy offers new insights for disease treatment. Ann Surg Oneal. 14: 719-29.

269. Witkowski J. (2010) Long view of the Human Genome Project. Nature. 466: 921-2.

270. Wozniak MJ, Melzer M, Dorner C, Haring HU, Lammers R. (2005) The novel protein KBP regulates mitochondria localization by interaction with a kinesin-like protein. B.M.C. Cell Biol. 6, 10.1186/1471-2121-6-35.

271. Wu JJ, Chen JX, Rothman TP, Gershon MD. (1999) Inhibition of in vitro enteric neuronal development by endothelin-3: mediation by endothelin B receptors. Development. 126: 1161-73.

272. Xia Y, Karin M. (2004) The control of cell motility and epithelial morphogenesis by Jun kinases. Trends Cell Biol. 14: 94-101.

273. Xing M. (2005) BRAF mutation in thyroid cancer. Endocr Relat Cancer. 12: 245-62.

274. Yang J, Benyamin B, McEvoy BP, Gordon S, Henders AK, et al. (2010) Common SNPs explain a large proportion of the heritability for human height. Nat Genet. 42: 565-9.

275. Ye L, Santarpia L, Gagel RF. (2010} The evolving field of tyrosine kinase inhibitors in the treatment of endocrine tumors. Endocr Rev. 31: 578-599.

276. Yip L, Cote GJ, Shapiro SE, Ayers GD, Herzog CE, et al. (2003). Multiple Endocrine Neoplasia Type 2 Evaluation of the Genotype-Phenotype Relationship. Arch Surg. 138: 409-416.

277. You L, Ebner S, Kruse FE. (2001). Glial cell-derived neurotrophic factor (GDNF)-induced migration and signal transduction in corneal epithelial cells. Invest Ophthalmol Vis Sci. 42: 2496- 2504.

278. Young HM, Hearn CJ, Farlie PG, Canty AJ, Thomas PQ, et al. (2001) GDNF is a chemoattractant for enteric neural cells. Dev Biol. 229: 503-16.

167 References

279. Zhu L, Lee HO, Jordan CS, Cantrell VA, Southard-Smith EM, et al. (2004) Spatiotemporal regulation of endothelin receptor-B by SOXl0 in neural crest-derived enteric neuron precursors. Nat Genet. 36: 732-7.

280. Zweier C, Thiel CT, Dufke A, Crow VJ, Meinecke P, et al. (2005) Clinical and mutational spectrum of Mowat-Wilson syndrome. Eur J Med Genet. 48: 97-111.

168 Summary Nederlandse samenvatting Resumo Summary

Summary

Hirschsprung disease (HSCR) and Multiple Endocrine Neoplasia type 2 (MEN2) are two seemingly unrelated neural crest-associated disorders that can be caused by germ line mutations in RET. HSCR and MEN2 have distinct characteristics and different tissues are affected in each disease. However, in both diseases the basic problem is the disruption of signal transduction pathways required for normal physiological functions. This thesis presents a functional and genetic study on HSCR and MEN2, and aimed at unravelling the protein networks involved in the development of these disorders. Part I of this thesis focuses on HSCR, which is an inherited disorder characterized by the absence of enteric neurons in the distal part of the colon. It can occur as an isolated trait, but is frequently associated with other congenital malformations present in rare, recessive syndromes. The study of the genes involved in these syndromes can yield new insights into the HSCR-related protein network and may help identify new HSCR candidate genes. One of these rare, recessive syndromes is called Goldberg-Shprintzen syndrome, for which mutations in KBP/KIAA1279 were found to be the causal entity. In Chapter 2 the function of the protein encoded by KBP/KIAA1279, the Kinesin-Binding Protein (KBP) was studied. Since the KBP function was largely unclear, we performed expression studies in zebrafish and found that KBP is expressed in the central and enteric nervous system. In addition, we showed that KBP is involved in neuronal differentiation and neurite outgrowth. In order to determine the KBP protein network, a yeast two-hybrid screen was performed. Several kinesin-like proteins were found to interact with KBP, as expected. KBP was also found to interact with SCGlO, a microtubule-destabilizing protein. This interaction was further confirmed by co-immunoprecipitation, co-localization and by epistasis experiments in zebrafih, thus linking KBP to microtubule dynamics. Since a direct interaction between KBP and microtubules could not be found, it is likely that KBP exerts its effect via SCG10 interaction.

170 Summary

In Chapter 3 we evaluated the involvement of SCGlO in HSCR development. Since SCGlO was reported to be a down-regulated gene in a mouse model for HSCR, and we found a direct interaction between SCGlO and KBP in vitro and in vivo, we considered SCGlO to be a good candidate gene for HSCR. To determine if SCGlO is involved in the aethiology of HSCR, we performed expression studies in mouse enteric neural crest stem cells. In addition, 85 patients diagnosed with isolated, non­ syndromic HSCR were screened for the presence of SCGlO mutations. We found that SCGO expression can be detected from the early stages of gut colonization, but there were no germline mutations found in the patients analysed. These results suggest that SCGlO is not directly implicated in HSCR development. However, an indirect involvement of SCGlO in HSCR cannot be ruled out as this could be due to a secondary effect in SCGlO regulation. In chapter 4 we focus on HSCR as a complex disorder. Using HSCR as an example, we reviewed what is known about this disorder, what we learned from the genome-wide association studies, and discussed how next-generation sequencing in combination with a systems biology approach can help in unravelling the complex inheritance of HSCR. Part II of this thesis focused on MEN2, which is a dominantly inherited cancer syndrome that can be divided into three subtypes: MEN2A, MEN2B and Familial Medullary Thyroid Carcinoma (FMTC), based on the clinical description and tissues affected. As different RET mutations result in different disease phenotypes, it is logical to assume that different mutations have different effects on downstream signalling pathways. In chapter 5, the kinome profile of MEN2 mutants was determined. For this, protein phosphorylation profiles specific for the most common mutations found in MEN2A (RET609, RET620, RET634), MEN2B (RET918), and FMTC (RET791, RET891) were determined using a peptide array consisting of 1024 different kinase substrates. All the mutants showed activation of the canonical RET signalling pathways via MAP kinase, Pl3 Kinase/ Akt and Rae pathways. Specific alterations were also identified for

171 Summary some of these mutants. For the RET918 mutant, an increase in FAK phosphorylation levels was seen, and for RET891, a decrease in receptor internalization was detected. We also observed that mutations associated to the same phenotype do not always behave in the same way, as small differences in activation levels of the shared pathways were detected. We concluded that there are indeed differences in the signalling pathways between RET wild-type and different types of MEN2, confirming the hypothesis that different RET mutations have specific effects on downstream signalling pathways. Furthermore, different mutations associated to the same phenotype do not necessarily share the same signalling pathways, which supports the idea that different mutations act through their own specific signalling pathways. In chapter 6, four tyrosine kinase inhibitors were analysed for the treatment of MEN2 and papillary thyroid carcinoma (PTC). For this, the effects of XL184, vandetanib, sorafenib and axitinib on cell proliferation, RET expression and autophosphorylation, and activation of the MAP kinase pathway, namely ERK, were compared in cell lines expressing a MEN2A (MTC-TT) mutation, a MEN2B (MZ-CRC-1) mutation, and a RET/PTC (TPC-1) rearrangement. Our results showed that the most

effective inhibitors in vitro are XL184 and vandetanib, in reducing cell proliferation, RET autophosphorylation in combination with a reduction of RET expression, and ERK phosphorylation, in MEN2A and MEN2B, respectively. For PTC, XL184 was the most efficient inhibitor in reducing cell proliferation and RET autophosphorylation. However, no effect on RET expression and ERK activation was observed for PTC after exposure to XL184. To conclude, there is TK-inhibitor specificity for the different RET mutants and, mutation-specific therapies could therefore be beneficial in treating MEN2 and PTC. Together, our results yielded new insights into HSCR and MEN2 as we unravelled the protein network involved in the development of these disorders.

172 Nederlandse Samenvatting

Nederlandse Samenvatting

De ziekte van Hirschsprung (afgekort als HSCR) en Multipele Endocriene Neoplasieen type 2 (afgekort als MEN2) zijn twee schijnbaar ongerelateerde aandoeningen. Ze kunnen echter warden veroorzaakt door mutaties in een en hetzelfde gen, het RET gen. In beide gevallen gaat het om mutaties die leiden tot een verstoring van de signaaltransductie routes die door RET warden geactiveerd, routes die nodig zijn voor de normale fysiologische functies van de aangedane weefsels. In dit proefschrift beschrijven we functionele en genetische studies over zowel HSCR als MEN2. De meeste studies hebben als doel het ontrafelen van de eiwitnetwerken die betrokken zijn bij de ontwikkeling van deze aandoeningen. Deel I van dit proefschrift richt zich op HSCR. HSCR is een erfelijke aandoening die wordt gekenmerkt door de afwezigheid van neuronen in het laatste deel van de dikke darm. De ziekte kan voorkomen als een geTsoleerde eigenschap (een patient heeft alleen HSCR), maar in ongeveer 20% van de patienten komt de ziekte voor in combinatie met andere aangeboren afwijkingen of is het een onderdeel van een syndroom. In deel 1 richten we ans op een dergelijk zeldzaam syndroom, genaamd Goldberg-Shprintzen. Dit syndroom wordt veroorzaakt door mutaties in KBP. In hoofdstuk 2 wordt de functie van het eiwit waar het gen voor codeert beschreven. Dit omdat de functie grotendeels onbekend is. Wat we wel weten is dat het een interactie kan aangaan met een eiwit uit de kinesine eiwit familie. We hebben gekeken naar expressie van KBP in zebravis embryo's en vonden dat KBP wordt aangemaakt in zowel het centrale als het perifere (enteric) zenuwstelsel. Bovendien konden we laten zien dat KBP betrokken is bij de neuronale differentiatie en neuriet uitgroei. Met het oog op het KBP-eiwit netwerk, hebben we naar eiwit-eiwit interacties gezocht door middel van een 'yeast two-hybride screen'. Verschillende kinesines werden gevonden, zeals verwacht. Daarnaast werd echter oak een interactie van KBP met SCGlO gevonden. SCGl0 is een microtubuli destabiliserend eiwit. Deze interactie werd bevestigd door middel van co-immunoprecipitatie. Oak werden co-localisatie studies en een epistase experiment in zebravissen uitgevoerd.

173 Nederlandse Samenvatting

Alie data koppelen KBP aan microtubuli. Een directe interactie tussen KBP en microtubuli werd niet gevonden. In hoofdstuk 3 evalueerden we de mogelijke betrokkenheid van SCGlO in HSCR. Dit omdat we een eiwit-eiwit interactie vonden tussen SCGl0 en KBP, maar ook omdat SCGlO al eerder aangemerkt was als goede kandidaat voor de ziekte door anderen. Om te bepalen of SCGl0 betrokken is bij het ontstaan van HSCR, werd gekeken of het gen tot expressie komt in de voorloper cellen van het zenuwstelsel in de darm. Deze studie werd gedaan op cellen ge'isoleerd uit muizenembryo darmen. Bovendien werden 85 patienten met ge'isoleerde HSCR gescreend op de aanwezigheid van SCGlO mutaties. We vonden dat SCGl0 inderdaad tot expressie komt in deze voorloper cellen, maar we konden geen (kiembaan)mutatie vinden bij de patienten. Deze resultaten suggereren dat SCGlO niet rechtstreeks betrokken is bij de ontwikkeling van de ziekte. Een indirecte betrokkenheid van SCGlO in HSCR kan echter niet warden uitgesloten. Hoofdstuk 4 is een review waarin wordt beschreven hoe complexe aandoeningen, genetisch gezien, wordt aangepakt. We gebruiken HSCR als voorbeeld. In dit hoofdstuk geven we een overzicht van wat er bekend is over deze aandoening, wat we hebben geleerd uit het verleden en vooral uit de genoom wijde associatie studies, en we bespreken wat een vervolgstrategie zou kunnen zijn voor het verder ontrafelen van deze en andere genetisch complexe aandoeningen. Deel II van dit proefschrift richt zich op MEN2. MEN2 is een dominant erfelijke kanker syndroom. Op basis van de klinische beschrijving (en de weefsels die zijn aangedaan) kan MEN2 in drie subtypen warden ingedeeld: MEN2A, MEN2B en familiale medullaire schildkliercarcinoom (afgekort als FMTC). Omdat verschillende RET mutaties resulteren in verschillende subtypen, is het logisch om te veronderstellen dat verschillende mutaties verschillende effecten op downstream­ signaalroutes hebben. In hoofdstuk 5 beschrijven we de 'kinome' profielen die gevonden zijn voor de verschillende MEN2 mutaties. Een kinome profiel is in feite een verzameling van

174 Nederlandse Samenvatting eiwitten die door RET kunnen warden geactiveerd. Deze bepaling doen we door actieve RET moleculen te hybridiseren met peptides (stukjes eiwitten) die mogelijk door RET (een tyrosine kinase) kunnen warden geactiveerd. Deze eiwitten zijn zogeheten tyrosine kinase substraten. Dit werd gedaan voor de meest voorkomende mutaties die gevonden zijn in MEN2A (RET609, RET620, RET634), MEN2B (RET918) en FMTC (RET791, RET891). We hebben gekeken naar 1024 verschillende kinase substraten. Alie mutanten vertoonden activering van bekende RET signaalroutes zoals de MAP-kinase, Pl3 kinase/Akt en Rae signaalroutes. Daarnaast werd voor een aantal mutaties specifieke veranderingen gevonden. Zo werd voor de RET918 mutant een toename van FAK phosphorylatie gezien, terwijl voor de RET891 mutant een daling van de receptor internalisatie werd gedetecteerd. We vonden ook dat mutaties die geassocieerd zijn met hetzelfde fenotype niet altijd hetzelfde resultaat gaven. Tot slot, bespreken we de gevonden verschillen in de signaalroutes tussen de RET wild-type en de verschillende soorten MEN2 mutaties. Dit alles bevestigt onze hypothese dat verschillende RET mutaties specifieke effecten hebben op de downstream signaleringsroutes. Hoofdstuk 6 laat de analyse zien van vier tyrosine kinase remmers die mogelijk bruikbaar zijn voor de behandeling van MEN2 en papillair schildkliercarcinoom (PTC). Gekeken is naar de effecten van XL184, vandetanib, sorafenib en axitinib op eel proliferatie, RET expressie, RET autophophorylatie, en activering van de MAP-kinase route. Dit werd gedaan in eel lijnen waarin een MEN2A (MTC-TT), een MEN2B (MZ-CRC-1) mutatie, en een RET/PTC (TPC-1) rearrangement aanwezig zijn. Onze resultaten toonden aan dat de meest effectieve inhibitoren XL184 en vandetanib zijn voor respectievelijk MEN2A en MEN2B. Voor PTC bleek XL184 de meest efficiente remmer. Er werd echter geen effect op de RET expressie en op activering van de MAP-kinase route waargenomen voor PTC na blootstelling aan XL184. Ons onderzoek laat zien dat de TK remmers mutatie specifiek lijken te zijn en dus zouden mutatie-specifieke therapieen gunstig kunnen zijn voor de behandeling van MEN2 en PTC.

175 Nederlandse Samenvatting

Onze resultaten hebben geleid tot nieuwe inzichten in de ontwikkeling van zowel HSCR als MEN2.

176 Resume

Resumo

A doenc;a de Hirschsprung (HSCR) e a Neoplasia End6crina Multipla tipo 2 (NEM2) sao duas doenc;as aparentemente diferentes que podem ser causadas par mutac;oes no proto-oncogene RET. A HSCR e a NEM2 tern caracterfsticas distintas e em cada uma destas doenc;as sao afectados 6rgaos diferentes. No entanto, em ambos os casos a disrupc;aodas vias de sinalizac;aonecessarias para o normal funcionamento fisiol6gico constitui o maior problema. Nesta tese, sao apresentados estudos funcionais e geneticos sabre HSCR e NEM2 desenvolvidos com o objectivo de decifrar a rede proteica, "the protein network", envolvida no aparecimento destas duas doenc;as. A primeira parte desta tese focaliza-se na HSCR. A HSCR e uma doenc;a genetica caracterizada pela ausencia de nervos entericos na parte distal do colon. A HSCR pode ocorrer de forma isolada, mas esta frequentemente associada a outras malformac;oes congenitas presentes em sfndromes recessives raros. 0 estudo dos genes envolvidos nestes sfndromes pode trazer novas pistas sabre a rede proteica envolvida no desenvolvimento da HSCR, e pode ajudar a identificar novas genes candidates. Um exemplo destes sfndromes recessives e o sf ndrome de Goldberg­ Shprintzen (GOSHS), em que se descobriu que a principal causa desta patologia sao mutac;oesno gene KIAA1279/KBP. No capftulo 2 e estudada a func;ao da protefna envolvida em GOSHS, a Kinesin Binding Protein (KBP). Tenda em conta que a func;ao da KBP e praticamente desconhecida, a expressao desta proteina foi determinada em zebrafish e descobriu­ se que a KBP e expressa no sistema nervoso central e no sistema nervoso enterico. Adicionalmente, mostrou-se que a KBP esta envolvida na diferenciac;aoneuronal e no crescimento de neurites. Com o objective de determinar a rede proteica da KBP, realizou-se um "yeast two-hybrid screen". Foram identificadas varias cinesinas que interactuam com KBP. Paralelamente, identificou-se que a KBP interactua com a SCGl0, uma protefna envolvida no processo de destabilizac;ao dos microtubulos. Esta interacc;ao foi confirmada par co-imunoprecipitac;ao, co-localizac;ao e determinac;ao de epistase em zebrafish entre KBP e SCGlO. Desta forma, a KBP foi implicada no

177 Resumo processo de estabilizac;ao/destabilizac;ao (dinamica) dos microtubulos. No entanto, nao foi possfvel detectar uma interacc;ao directa da KBP com os microtubulos, o que sugere um envolvimento indirecto da KBP na dinamica dos microtubulos, possivelmente exercido via SCGlO. No capftulo 3, foi determinado o envolvimento de SCGl0 na HSCR. Tenda em conta que a reduc;ao de expressao de SCGl0 foi previamente detectada num modelo animal para HSCR, e que n6s descobrimos que a KBP interactua com a SCGl0 in vitro e in vivo, o SCG10 foi considerado um 6ptimo gene candidate para a HSCR. Com o objective de determinar o envolvimento de SCGl0 na HSCR, a expressao deste gene foi determinada em celulas da crista neural derivadas de rato. Adicionalmente, 85 pacientes diagnosticados com HSCR, foram testados para a presenc;ade mutac;oesno SCG10. Os nossos resultados mostraram que o SCG10 e expresso desde o perfodo inicial de colonizac;ao do intestine, mas nao foi detectada nenhuma mutac;ao neste gene nos pacientes analisados. Estes resultados sugerem que SCGlO nao estara directamente envolvido no desenvolvimento da HSCR. No entanto, o envolvimento indirecto deste gene na HSCR nao pode ser exclufdo, ja que pode ocorrer devido a um efeito secundario desencadeado na regulac;aoda actividade de SCGl0. No capftulo 4, sao analisados os problemas associados ao estudo de doenc;as complexas. Usando a HSCR coma exemplo, realizamos uma revisao da literatura, na qual descrevemos a etiologia desta doeni;:a, o que aprendemos dos estudos de associac;ao gen6mica, e discutimos de que forma a sequenciac;aode pr6xima gerac;ao em combinac;ao com a biologia de sistemas nos pode ajudar a desvendar a complexa rede genetica envolvida no desenvolvimento da HSCR. A segunda parte desta tese incide na NEM2. A NEM2 e uma forma sindr6mica de cancro da tir6ide que pode ser dividida em 3 subtipos: NEM2A, NEM2B e Carcinoma Medular da Tir6ide Familiar (CMTF), dependendo das caracterfsticas clfnicas e dos tecidos afectados. Como diferentes mutac;oes no proto-oncogene RET resultam em diferentes tipos de NEM2, e 16gico prever-se que diferentes mutac;oes desencadeiam diferentes vias de sinalizac;ao.

178 Resumo

No capftulo 5 foi determinado o "kinome profile" das diferentes mutac;oes associadas a NEM2. Para isso, foram determinados os padroes de fosforilac;ao especfficos para as mutac;oes mais comuns encontradas em NEM2A (RET609, RET620, RET634), NEM2B (RET918) e CMTF (RET791, RET891), recorrendo a um "array" contendo 1024 substratos diferentes para cinases. Em todas as mutac;oes analisadas foi detectada uma activac;aodas vias can6nicas de sinalizac;ao associadas a RET: MAP cinase, Pl3 cinase/ Akt e Rae. Adicionalmente, foram identificadas alterac;oes especfficas associadas a algumas mutac;oes. Para o RET918 foi identificado um aumento de fosforilac;ao da cinase FAK, e para o RET891 foi detectada uma diminuic;ao de internalizac;ao do receptor. Pode-se tambem observar que as mutac;oes associadas ao mesmo tipo de NEM2 nem sempre se comportam da mesma forma, ja que foram detectados nfveis diferentes de activac;ao das vias de sinalizac;ao comuns. Em conclusao, as vias de sinalizac;ao associadas ao RET "wild-type" e as diferentes mutac;oes sao distintas, o que esta de acordo com a hip6tese inicialmente proposta: diferentes mutac;oes em RET tern diferentes efeitos na activac;ao das vias de sinalizac;ao. Para alem disso, mutac;oes associadas ao mesmo tipo de NEM2 nem sempre se apresentam da mesma forma, o que reforc;a a ideia de que mutac;oes diferentes activam vias de sinalizac;ao especfficas. 0 capftulo 6 descreve a analise de quatro inibidores de tirosina-cinases (XL184, vandetanib, sunitinib e axitinb) para o tratamento da NEM2 e do Carcinoma Papilar da Tiroide (CPT). Para isso, os efeitos de XL184, vandetanib, sunitinib e axitinb foram avaliados em termos de proliferac;ao celular, expressao e fosforizac;aodo RET, e activac;ao da via de sinalizac;ao da MAP cinase, nomeadamente da ERK, entre as diferentes linhas celulares expressando mutac;oes caracteristicas de NEM2A (MTC-TT), NEM2B (MZ-CRC) e de um rearranjo do RET, RET/PTCl (TPC-1). Os nossos resultados mostraram que os inibidores mais eficazes in vitro foram o XL184 e o vandetanib, uma vez que promoveram a reduc;ao da proliferac;ao celular, dos nfveis de fosforizac;ao do RET em combinac;ao com uma diminuic;ao da expressao deste receptor, e a activac;ao

179 Resumo da ERK, em NEM2A e NEM2B respectivamente. Para o CPT, o inibidor mais eficaz foi o XL184 que provocou uma diminui�ao da prolifera�ao celular e dos nfveis de fosforila�ao do RET. No entanto, nenhum efeito foi detectado nos nfveis de expressao do RET e na activa�ao da ERK no CPT ap6s exposi�ao a XL184. Deste estudo pode-se concluir que os inibidores de tirosina cinase apresentam especificidade para as diferentes muta�oes em RET, e par isso, uma interven�ao terapeutica espedfica para as diferentes muta�oes podera ser benefica no tratamento da NEM2 e do CPT. Em conclusao, os resultados apresentados nesta tese trazem novas perspectivas para o estudo da HSCR e da NEM2, uma vez que contribuem para a compreensao da rede proteica envolvida no desenvolvimento destas doen�as.

180 Acknowledgements

ACKNOWLEDGEMENTS

To accomplish great things, we must not only act, but also dream; not only plan, but also believe." William Bragg

With this sentence I reach the last section of my thesis! When I look back to where everything started, it is hard to believe it was such a long time ago ... Everyone tells you that it goes very fast. But I guess you have to experience it to understand what that means ... During this time, so many things have happened and so many other things have changed ... But as I look back and remember everything I've learned, all the people I've met, and all the places I've seen, it is hard not to smile! So now it is time acknowledge everyone who was part of my journey and express my deep gratitude. Dear Robert, where can I start to thank you for everything you've done for me all these years? I guess it is impossible to actually describe it in words, but I promise I will do my best! I remember the first time we met. I had just arrived from Portugal and you were kind enough to pick me up at the train station. Back then I thought: "See, this is not so bad ... Robert is a nice boss!", and suddenly I had a comforting feeling that everything was going to be ok. And indeed, that was the case! Working with you has made me grow as a scientist and I'm now much more independent and much more confident in my work than ever before. You made me trust my instincts and you were always supportive every time I had a new idea. I really think you are a great boss ... and especially in these last few months, it was amazing to receive all the help and support you gave me. Without your help, I don't think I would ever have made it in time. Thank you for always being there and for being not only my boss but also my friend! It was, it is, and it will always be a great pleasure to share a beer with you, and I am looking forward to our next meeting already! © Dear Bart, thank you for all your help and all your support. Every time I needed it, you were always ready to help me ... Every time I didn't know what to do,

181 Acknowledgements or how to proceed, you were always on the other end of the phone giving me ideas and suggestions. You were always able to make me feel a little more relaxed and a little more confident about tackling new experiments. What else can I say ... you were my expert in the field! © Thank you for also letting me work in your lab in Haren and for never getting tired of explaining something to me. Thank you for always being so patient and for making me understand that, in our field, as in so many other things in life, being patient is a good quality to have! It was great to work with you. Prof. Isabella Cecherinni, Prof. Thera Links and Prof. Maikel Peppelenbosch, thank you all for making up the reading committee for my thesis. I am very grateful that you agreed to read it and provided critical comments. Thank you for your approval for my thesis defence. Dear Yunia, Jan, Christine and Hans, my partners in crime in the RET/Hirschsprung group! Thank you for being such good colleagues and friends. My dear Yunia, thank you for always being there for me and for all your help. We met at the beginning of my PhD and together we've shared so many things ... laughs, worries, work, trips, dreams. It was great to have you as a friend and I wish you all the luck and success for completing your own PhD thesis and for your life. Dear Jan, thank you for always being so nice to me. You were my first roommate and I enjoyed your company in the lab and in the office very much. Thank you for all the help in my projects, for taking care of my cells and for doing all my sequences. Thank you also for being my paranimf. It is great to know you and I am sure we will keep in touch! Dear Christine, it was great to meet you and to work with you! Thank you for all your help, your optimism and supporting words whenever I needed it. I wish you all the best for your internship and your PhD. Dear Hans, it was great to work with you on the TK paper! I know it was a little bit stressful because everything had to be finished as fast as possible, but in the end I think we did a very good job! Thank you for dropping by in the lab, even after your clinical work, to help me every time you could, and for being such a nice guy. It was really great to meet you and I wish you all the best for your internship and for your PhD.

182 Acknowledgements

Dear Susanne, Raj and Loes, thank you for being my second group and for making my stay in Haren a great time! My dear Susanne, it was great to meet you and to work with you. You helped me so much during my yeast two-hybrid experience ... Your knowledge and skills made everything much easier! Thank you for the thousands of times you brought stuff to Groningen for me from Haren! I don't know how I could have made it through those hard times without you ... And of course it was great to hang around after work for some beers in the Buurvrouw! It was a great time! I wish you lots of luck and success in Copenhagen and I'm looking forward to meet you again soon! © Dear Raj, it was great to meet you and to have you as a colleague ! Even when you were teasing me, you were always fun to be with! I wish you lots of luck for the end of your PhD and lots of success in life! Dear Loes, thank you for all your help with the cell work and in the lab. You always made sure I had everything I needed and I really appreciated your help. Dear Saskia, Esther and Annemieke, you girls were great to work with! Thank you for being such good students and for all the help in my projects. I wish you all good luck and success in your lives. Prof. Buys and Prof. Wijmenga, dear Charles and Cisca, thank you for being such good heads of the Genetics Department (past and present) and for always having a friendly word. Dear Helga, Ellen and Dineke, thank you for all your input and all the tips towards getting better results. Helga, thank you so much for all the support and for always being so nice to me. It was good to know that you were in the next room and that I could always count on your help whenever I needed it. Ellen, thank you for all your comments, the critical discussions and for always being so enthusiastic about my work. Dineke, thank you for all the lab discussions about techniques, procedures, cells, antibodies, western blots and so many other things! Dear Jackie, thank you for always being so nice and for always making sure that I continue to improve my English! Thank you also for editing my thesis. It is so much better now after your magic touch!

183 Acknowledgements

To all our collaborators, thank you for all the help, the effort and the fruitful discussions. Dear Uli, Amalia and lvi, thank you for letting me work in your lab and for helping me with the cortical neurons. I learned a lot from you. Dear lain, Greg, and Jean-Marie, It was great to work with you on the KBP paper. Thank you for all the zebrafish work and for always trying to make the paper better. Dear Greg, thank you for always caring and for always being nice to me. You were one of my first friends in Groningen and I will never forget all the fun we had! I wish you all the best for your life and I hope we can meet again. Dear Maikel and Kaushal, it was great to work with you on the kinome paper. Thank you for all your input and insightful comments. Dear Maikel, it was amazing to see how much you know about signal transduction and how many ideas you get from a single kinase! It was and is great to work with you! Dear Kaushal, thank you for performing the arrays and for analysing all the data. It was crazy how much you did in so little time ... It was great to work with you and learn all about kinases and pepchips! Thank you for all your help. Dear Esther (de Graaff), Ben, Myrrhe and Casper, thank you for making my stay in Rotterdam so much nicer. Dear Ben, thank you for letting me work in your lab and for always being so nice to me. Dear Esther, thank you for being my host and for always making sure that I was not alone. Thank you also for all the help during the co-lPS struggle. It was great to meet you! Dear Casper and Myrrhe, thank you for the good suggestions and your help with the co-localization and co-IP experiments. Dear Alice, thank you for all your comments on the KBP paper and for explaining the clinical importance of the lab work. Dear Knut and Jonathan, thank your for all your comments and suggestions. You were always so friendly. Dear Jonathan, it was a pity that in the end our live-tracking experiments didn't work ... Thank you for teaching me how to track vesicles and showing how much fun it is to see the tiny dots move inside the cells! Dear Marco, thank you for sharing your neural crest stem cells expertise with us! It was great to meet you and to work with you on the KBP/SCGlO paper. Dear Ben (Giepmans) and Klaas, thank you for all your help with the immunofluorescence experiments. Dear Ben (Giepmans), thank you for your help and suggestions to improve my stainings and tubulin pull-

184 Acknowledgements downs. Dear Klaas, thank you for your help with the confocal microscope and for always being so nice to me, even when I knocked on your door 20 times a day! © You really helped me a lot. Dear Lydia, thank you for all your help with the last experiments of my PhD work. Thank you for making my FACs experience so easy. © It was great to meet you and to work with you and I hope we will keep in touch! Dear Ivan, thank you for all the suggestions and tips about RET stimulations. It was pity we didn't have a chance to work together, but maybe in the future, who knows? Dear Gerard, thank you for all your help on the review and for always trying to explain the complicated world of statistics to me. Dear Mentje, Ria, Bote and Helene, thank you for all your help during these years in taking care of documents, sending packages and arranging every form needed. You make the life of a PhD student so much easier. To all my friends and colleagues in the department, thank you for making the lab such an enjoyable place to work! It was great to know you and to party and hang around with you. We sure had a great time! Dear Yunia, Christine, Cleo, Eva and Noortje, thank you for being such nice roommates and for creating a great atmosphere in our office. Dear Jihane, Gerben, Gosia, Agata, Asia, Justyna, Annemieke, Tjakko, Mats, Jianghua, Renee, Jan J, Mahdi, Anna, Paul and Karin, thank your for all the work discussions and suggestions; thank you for always cheering up the lab, even at nights and in the weekends; and thank you for all the great fun we had in all the (crazy) parties! © In the diagnostics section, thank you Jos, Bea, Yvonne, Krista B, Krista K and Chantal for all your help in the lab. Dear Marina and Edwin, thank you for taking care of the finances and for solving all the computer-related problems. Dear Bahram and Monique, thank you for always being so nice and always willing to help. Bahram, I really enjoyed our lunch breaks where we talked about so many different things and where you taught me a little about your culture. It was great to get to know you! Dear Joke, Conny, Peter and Jorieke, thank you for your comments and suggestions during our Friday morning meetings.

185 Acknowledgements

Dear Joke, thank you also for all your help in sorting out patient data and for arranging everything so quickly. To my dear friends in Groningen, Ana, Pedro, Mateusz, Marysia, Susana, Carlos, Roberta, Bispo, Ana Duarte, Bram, Karla, Arjan, Gilda, Mirjam, Daiana, Adriana, Jihane, Bahram, Susanne, Silvia, Raquel, Rajesh, Gwenny and Danny, thank you for all your support and friendship. You were my family here in Groningen and I really enjoyed our time together. You made everything easier and I am truly grateful to have met you all! To my dear friends in Portugal, thank you for all the support and friendship. It is very good to know that, no matter the distances, you are always there for me. Thank you for cheering me up whenever i needed it and thank you for bringing a little of Portugal with you in every visit. Dear Ana Gracinda, thank you also for being my paranimf! It is truly amazing to have you next to me on such an important day! Dear Ricardo, thank you for your support and for always making me believe that I could do this. It was great to share a part of this journey with you. To my dear family, Mae, Pai, Daniel (the King!), Guida, Toninha and Mimi. Thank you for all your support and for always wanting the best for me, for trying to understand my decisions, even if sometimes they didn't make much sense to you. You always kept me motivated and focussed and showed me how important and satisfying it is to achieve the goals we have planned. It is a pity that Pai is no longer here with us ... I wish he could see the end result of my hard work ... But I'm sure that he is looking out for us and I hope I have made him proud with everything I've achieved! And finally, my dear Kaushal, you are the best thing from these years! Thank you for always being there for me since almost the beginning of this journey, for always supporting me and cheering me up whenever I didn't get the results I wanted (lots of times)! © You are my sunshine in this rainy Holland and I am looking forward to our life together.

Maria

186 187