The Bone Marrow Niche in the Pathogenesis of Multiple Myeloma a Role for Wnt Signaling and Adrenomedullin

A role for Wnt signaling and Adrenomedullin boneThe marrowpathogenesis niche in the of Multiple Myeloma The bone marrow niche in the pathogenesis of Multiple Myeloma A role for Wnt signaling and Adrenomedullin

pracownia DTP i grafiki www.przygotowalniaDTP.pl ISBN 978-83-63861-16-2 Kinga Anna Kocemba

okladka_nowa.indd 1-3 2014-01-20 11:57:46 The bone marrow niche in the pathogenesis of Multiple Myeloma A role for Wnt signaling and Adrenomedullin The bone marrow niche in the pathogenesis of Multiple Myeloma A role for Wnt signaling and Adrenomedullin Kinga Anna Kocemba

The research described in this thesis was funded by the Dutch Cancer Society

Editing Przygotowalnia Pracownia DTP i Grafiki Graphic design and typesetting Przygotowalnia Pracownia DTP i Grafiki Cover: Antelope Canyon, Arizona, USA by Kinga Anna Kocemba © Copyright by Kinga Anna Kocemba

ISBN 978-83-63861-16-2 Amsterdam–Kraków 2014

Pracownia DTP i Grafiki ul. Łużycka 71c/9 30–693 Kraków [email protected] www.przygotowalniaDTP.pl Moim rodzicom To my parents

The bone marrow niche in the pathogenesis of Multiple Myeloma A role for Wnt signaling and Adrenomedullin

academisch proefschrift ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op donderdag 6 maart 2014, te 14:00 uur door Kinga Anna Kocemba geboren te Oświęcim, Polen Promotiecommissie promotor Prof. dr. S.T. Pals co-promotor Dr. M. Spaargaren overige leden Dr. M.M. Maurice Prof. dr. C.J.M. van Noesel Prof. dr. M.H.J. van Oers Prof. dr. P. Sonneveld Prof. dr. K. Vanderkerken

Faculteit der Geneeskunde Contents

9 chapter 1 General introduction

37 chapter 2 Wnt signaling reaching gale force during multiple myeloma progression

59 chapter 3 Transcriptional silencing of the Wnt-antagonist DKK1 by promoter methylation is associated with enhanced Wnt signaling in advanced multiple myeloma

87 chapter 4 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma and is associated with aggressive disease

111 chapter 5 N-cadherin-mediated interaction with multiple myeloma cells inhibits osteoblast differentiation 143 chapter 6 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma and promotes angiogenesis

201 chapter 7 Summary and discussion

213 chapter 8 Nederlandse samenvatting Acknowledgments Curriculum vitae Chapter 11

General introduction

B cell development and B-lineage malignancies

The development of B cells is characterized by sequential molecular events regulated by B-lineage specific transcription factors. During early development in the 1 bone marrow, B cells acquire the ability to express cell surface-bound immuno- globulins, which constitutie the antigen-binding part of the B cell receptor (BCR) multiprotein complex.1 BCR diversity, required for antigen recognition, is generated by somatic recombination, combining various gene segments within the 3 immunoglobuline (Ig) heavy and light chain loci. The heavy chain is assembled from Variable (V), Diversity (D) and Joining (J) gene segments that somatically 4 recombine with the Constant (C) region exons, generating unique immunoglobu- lins of diverse antigenic specificity. The light chain variable region (which can be of 5 kappa or lambda type) is composed of only a V and J gene segment. The diversity of the IgV regions is further enhanced by addition of nucleotides at the junctions 6 of the segments. Upon successful rearrangement of Ig light and heavy chain genes, a complete IgM molecule is expressed at the cell surface, which is required for sur- 7 vival. Non self-reactive immature B cells exit the bone marrow, whereas self-reactive immature B cells undergo apoptosis (clonal deletion), or generate a new B cell receptor by receptor editing.2,3,4,5 The immature B cell then migrate to the secondary lymphoid organs, i.e. lymph nodes, spleen and mucosa-associated lymphoid tissue (MALT), to complete the process of maturation. Upon antigen encounter, these naive B cells may undergo antigen-specific B cell differentiation. Initially, antigen is internalized by the BCR and processed peptides are presented by MHC class II molecules to Th cells. This interaction with Th cells provides the B cells with co-stimulatory signals, crucial for further differentiation. B cells may now differentiate into short-lived plasmablast outside of the germinal center (GC), or migrate into a primary follicle to initiate the formation of a GC.6,7 Germinal centers can be

11 Chapter 1

subdivided in two distinct zones, called the dark and light zones. Rapidly prolifer- ating B cells (centroblasts) form the dark zone of the GC. In this zone, somatic hypermutation (SHM) of the immunoglobulin genes of the B cells takes place.8–10 This involves introduction of point mutations by the enzyme called activation-induced cytidine deaminase (AID), which deaminates cytidines to uracil. This results in C-T and G-A transitions in the V region of immunoglobulin (Ig) genes, creating Ig variants with altered affinity for a particular antigen.11 The mutated B cells (centrocytes) migrate to the light zone of the GC. Here, they re-encounter antigen presented by the follicular dendritic cells (FDC), and a selection process, based on the affinity of the BCR for antigen, is initiated. Depending on the strength of BCR signal, B cells will receive a survival and proliferation signal from the FDCs and Th cells, or will die by apoptosis.12–14 High affinity B cells will process antigen and present this to antigen-specific T cells,15 which provide stimulatory signals (e.g. T cell receptor (TCR)/CD3-MHC class II, CD40-CD40L, CD80/86-CD28 and several cytokines),6,13,16–18 resulting in class switch recombination (CSR) and further differentiation into either memory B cells or plasma cells.9,17,18 VDJ recombination (VDJR), as well as (CSR) and SHM, involve the generation of DNA breaks, potentially dangerous events that predispose to chromosomal translocations Indeed, the DNA breaks that are induced by VDJR, CSR and SHM coincide with the sites of chromosomal translocations that involve the IgH or IgL loci in many lymphoid malignancies.19–21 For instance, chromosomal translocation that involve the IgH switch region and the partner oncogene (e.g. BCL-2, BCL6, c-MYC, CCND1, CCND3, FGFR3-MMSET and MAFC), have been identified in follicular lymphoma (FL), diffuse large B cell lymphoma (DLBCL), Burkitt’s lymphoma and multiple myeloma (MM).22–25 Moreover, during the SHM process in the GC, AID can introduce mutations in non-immunoglobulin genes, e.g. the oncogenes BCL6, MYC, PIM1, PAX5.19,26 In general, each lymphoma subtype resembles a B cell trapped at the particular stage of B cell development (Figure 1), as determined by the presence or absence of somatic hypermutation and the gene expression profile.27,28 For instance, for Burkitt lymphoma, follicular lymphoma, and the germinal center B cell-like (GCB) subtype of diffuse large B cell lymphoma (DLBCL), the normal cellular counterpart is the germinal center B cell,29,30 whereas the activated B cell-like (ABC) subtype of DLBCL resembles post-germinal center B cells/plasmablasts.29 Mucosa-associated lymphoid tissue (MALT) lymphomas are extra nodal in origin and phenotypically related to post-germinal center marginal zone B cells.31 Hairy cell leukemia has mutated Ig genes and class switched Ig heavy chains, along with a gene expression profile pointing to post-germinal center memory B cell as a cell of origin.32

12 General introduction

Figure 1. Normal B cell development and the related stages of B cell malignancy Schematic representation of B cell differentiation. The malignant counterparts are indicated by italic font. BM-bone marrow, ALL-acute lymphoblastic leukemia, MALT-mucosa associated lymphoid tissue, MCL-mucosa associated lymphoid tissue, DLBCL-diffuse large B cell lymphoma, CLL-chronic 3 lymphocytic leukemia, FL-follicular lymphoma, MM-multiple myeloma. 4 Multiple Myeloma 5 Multiple myeloma (MM) is characterized by a clonal proliferation of plasma cells 6 (PCs) in the bone marrow (BM), often associated with pancytopenia and osteolytic bone disease. It is one of the most frequent hematological cancers and remains 7 largely incurable, despite high-dose chemotherapy with additional stem cells support. The presence of somatic hypermutation of the immunoglobulin variable region genes in MM cells indicates that the cell of origin in myeloma is a post-germinal center B cell.33,34 In the majority of cases MM arises from a pre-malignant expansion of plasma cells called monoclonal gammopathy of undetermined significance (MGUS), with a prevalence of approximately 3% and 5% in people older than 50 and 70 years, respectively.35,36 Based on chromosomal studies, two main cytogenetic subgroups can be discriminated: a group characterized by a high in- cidence of five recurrent IgH translocations and loss of chromosome 13/13q14; and a hyperdiploid group, characterized by multiple trisomies.24 The recurrent translocations in the non-hyperdiploid group involve the IgH switch region 14q32

13 Chapter 1

and the translocation partners: cyclin D1 (CCND1) (11q13), cyclin D3 (CCND3) (6p21), MAFC (16q23), FGFR3 and MMSET (4p16), and MAFB (20q11).37 Hyperdiploid myeloma is characterized by trisomies of multiple odd chromosomes (3, 5, 7, 9, 11, 15, 19, and 21). Together with t(11;14), hyperdiploidy confers a relatively favorable prognosis, whereas MAFC, MAFB, or FGFR3/MMSET activation and deletion of chromosome 13 and/or 17 are associated with a poor prognosis.38,39 Interestingly, there is no single genetic event that distinguishes MGUS from MM. However, several genetic aberrations have been associated with disease progression. These include translocations and other genetic aberrations involving cMYC, activating mutations in KRAS, NRAS and FGFR3, and inactivating mutations or deletions of TRAF3, TP53, RB1 and PTEN.40–42 Moreover, aberrant activation of two oncogenic pathways, i.e. the NF-ĸB and Wnt pathways, is considered as a driving force for disease progression (Figure 2).43–48

Figure 2. Aberrant genetic events during progression of multiple myeloma Genetic aberration occurring during the progression of MM. For more details see the text.

Molecular classification of multiple myeloma

Multiple myeloma is a heterogeneous disease, with genetic diversity as well as significant variation in clinical features and prognosis. Recently, several groups have employed mRNA microarray technology to molecularly sub-classify newly diagnosed MM patients. Two major gene-expression profile-based classification systems have thus far been reported, the translocation and cyclin D classification (TC)49

14 General introduction and the University of Arkansas for Medical Sciences (UAMS) molecular classification of myeloma.50 The TC classification system comprises 8 distinct groups, based on overexpression of genes deregulated by primary IgH translocations and transcriptional activation of cyclin D,49 whereas the UAMS contains 7 distinct subgroups including; translocation clusters MS [t(4;14)], MF [t(14;16)/t(14;20)], CD-1/2 [t(11;14)] and [t(6;14)], a hyperdiploid cluster (HY), a cluster with proliferation-associated genes (PR), and a cluster mainly characterized by a low inci- dence of osteolytic bone disease (LB). Furthermore, this study reported a group of cases with a myeloid signature that was excluded from the analyses.50 Recently, a study by Broyl et al. revealed 3 novel clusters.51 These are the NF-κB, cancer- testis antigen (CTA), and PRL3 clusters. The NF-κB cluster was characterized by hyperdiploidy in 66% of cases, demonstrating clear differential expression of NF-κB pathway genes, while the CTA cluster was characterized by overexpression of cancer-testis antigens. The PRL3 cluster consisted of 9 cases only, with overexpression of a limited number of genes including tyrosine phosphatase PTP4A3 1 (PRL3), protein tyrosine phosphatase, receptor-type, Z polypeptide 1 (PTPRZ1), and suppressor of cytokine signaling 3 (SOCS3).51 3 The bone marrow environment in multiple myeloma 4 Since MM mainly progresses in the bone marrow, signals from this microenvi- 5 ronment play a crucial role in plasma cell growth, migration and survival. The MM microenvironment is formed by clonal plasma cells, extracellular matrix 6 proteins (ECM) proteins and bone marrow stromal cells (BMSC). Of particular interest is the observation that factors normally expressed within the bone mar- 7 row to specifically regulate and maintain the HSC niche, are also involved in enhancing myeloma cell growth, supporting the notion that that myeloma cells “hijack” the normal hematopoietic niche, thereby contributing to pancytopenia.52,53 Moreover, the MM BM microenvironment was shown to confer survival and chemoresistance of MM cells to current therapies. Of note, the interaction between MM cells and local BM environment is bi-directional. MM destroys the bone marrow homeostasis, resulting in anemia, aberrant angiogenesis and osteolytic bone disease, whereas the BM environment supports the growth and survival of malignant plasma cells, via adhesion molecules, cytokines and growth factors (Figure 3).54 Specifically, BMSC secrete of a variety of cytokines and growth factors including: IL-6, IGF-1, VEGF, TGF-b, tumor necrosis

15 Chapter 1 factor alpha (TNF-alpha), stromal cell-derived factor 1a (SDF-1a), basic fibroblast growth factor (bFGF), macrophage inflammatory protein 1a (MIP-1a), stem-cell factor (SCF), hepatocyte growth factor (HGF), B-cell-activating factor (BAFF), IL-1b, IL-3, IL-10, IL-15, and IL-21, as well as Ang-1 and matrix metalloproteinases (e.g. MMP-2 and MMP-9).46,55–64,65 These cytokines, secreted by BMSC, activate major signaling pathways contributing to MM pathogenesis directly, by triggering the key tumor-cell responses growth, survival, migration and drug resistance, and indirectly, by modifying the tumor microenvironment, increasing tumor angiogenesis and bone resorption. In addition, adhesive interaction between malignant plasma cells and the cells in the BM environment seems to becrucial in survival and drugs resistance of malignant plasma cells. In particular, MM cells express beta1 (CD29) integrins, including very-late- activating antigens 4 and 5 (VLA-4, VLA-5),66,67 as well as integrin avb3.68,69 Integrin-mediated adhesion critically enhances MM-cell growth and survival, and most importantly confers protection against drug-induced apoptosis. This is a consequence of the direct cell-cell contact, as well as NF-κB-dependent transcription and secretion of IL-6, a major MM growth, survival, and drug resistance factor.67,70

Figure 3. The interaction between MM and the BM niche A schematic presentation of the interactions between malignant plasma cells and the BM environment. Myeloma cells produce factors that directly or indirectly modify the BM environment by induction of aberrant angiogenesis and osteolysis. In particular osteolysis is triggered by inhibition of osteoblast differentiation and stimulation of osteoclast activity. In parallel, the BM derived growth and survival factors together with the adhesive interaction between MM cells and the bone marrow stromal cells further support the progression of MM cells in the bone marrow niche.

16 General introduction

The role of MM-induced angiogenesis during the progression of the disease bone marrow hypoxia The presence of hypoxia in the normal BM was demonstrated in several independent studies. In studies in mice, Parmar et al. showed that a fraction of hematopoietic stem cells are exposed to hypoxic oxygen levels in normal bone marrow, whereas most of the committed and differentiated progenitors appear to be relatively well oxygenated.71 This observation was confirmed by Levesque et al., who clearly demonstrated the presence of hypoxia in the endosteal niche, with an oxygen tension below 10 mmHg, whereas the central BM was reported to be normoxic in steady state conditions, probably because of the proximity to endothelial sinuses.72 Of note, it has been postulated that these endosteal and vascular niches have a key role in determining the function and the faith of hematopoietic stem cells: whereas the hypoxic endosteal niche maintains he- 1 matopoietic stem cells in a quiescent, immature state, the normoxic vascular niche promotes the proliferation and maturation of hematopoietic stem cells.71,72 The intriguing question whether or not MM cell behavior is also dependent on hypoxic niches awaits further investigation. So far, a few studies have demon- 3 strated the presence of low oxygen tension in MM-infiltrated BM in the human and in murine models. Using the 5T2MM mouse model of MM, Asosingh 4 et al., reported hypoxia in the whole BM environment, in control as well as in the 5T2MM-diseased mice,73 by using pimonidazole as a probe and analyz- 5 ing HIF1α protein. Although both control and MM-infiltrated bone marrow were hypoxic, hypoxia in myelomatous bone was significantly decreased.73 6 These data suggest that expansion of MM starts in the hypoxic BM environment and that MM-induced angiogenesis leads to an increased oxygen pressure 7 during disease progression. However, contrary to these findings, a recent study using the 5T33MM demonstrated that the myelomatous BM environment is actually more hypoxic than the normal BM microenvironment and promotes metastasis of MM cells.74 This observation is in accordance with findings in solid tumors, where poorly organized tumor-induced vasculature is unable to provide a blood supply sufficient to restore a normoxic environment. In human bone marrow, Colla et al., reported the presence of hypoxia in the aspirates collected from healthy donors, MGUS and MM patients.75 Moreover, studies by these authors described nuclear accumulation of HIF1α in all MM patients analyzed. In addition, recent studies by Martin et al. and Giatromanolaki et al. also confirmed HIF1α accumulation in malignant plasma cells in the BM.76–78

17 Chapter 1

Studies from the Anderson laboratory further extended these findings by demonstrating HIF1α protein stabilization and activity in some MM patients under normoxic conditions. Specifically, it was demonstrated that c-Myc collaborates with HIF1α to induce expression of VEGFA, an important growth factor for MM and endothelial cells.79 aberrant angiogenesis in multiple myeloma MM was the first hematological malignancy in which a prognostic value of aberrant angiogenesis was described.80 In the MM infiltrated BM microenvironment aberrant neovascularization (“angiogenesis”) is present. This is reflected by endothelial cells activation and increased capillary permeability, as well as by increased microvessel density (MVD).81,82 Of note, BM angiogenesis parallels disease progression and correlates with event-free survival (EFS) and overall survival (OS).83–86 The cause(s) of the myeloma angiogenic switch are currently the subject of investigation. The increased BM angiogenesis in MM is considered to be related to an imbalance between the production of pro- and anti-angiogenic factors by myeloma cells and the microenvironment. Indeed, over-production of VEGFA, bFGF, IGF-1, ANG-1, OPN and HGF is reflected by their elevated levels in the BM plasma and peripheral serum of MM.87,88 Interestingly, however, there is no clear difference in the expression of VEGF and bFGF, the primary angiogenic inducers secreted by myeloma cells, between MGUS, smoldering MM, and active MM. This suggests that the observed angiogenic switch between MGUS and active MM is, at least in part, the consequence of increasing number of malignant plasma cells in the bone marrow enviroment.89 Indeed, recently, a model of MM-induced angiogenesis was proposed in which normal bone marrow plasma cells (BMPCs) induce controlled angiogenesis by secreting of a slight surplus of pro-angiogenic over anti-angiogenic factors. In accordance with this scenario, the increase in plasma cell number during MM progression leads to an increase of the absolute amount of pro-angiogenic factors produced in the BM.90 According to this model, it is not necessary that MM cells show a differential expression of pro-angiogenic and anti-angiogenic genes compared with BMPCs, although aberrant expression of pro-angiogenic genes and downregulation of anti-angiogenic genes may further enhance angiogenesis in individual MM patients. Moreover, the bone marrow environment presumably is chronically hypoxic and the influence of hypoxia and HIF1α on the production of pro-angiogenic molecules is well-documented. Indeed, studies have revealed a positive correlation between HIF1α expression and the level of BM angiogenesis in MM biopsy specimens.76 Furthermore, forced overexpression of HIF1α in

18 General introduction

MM cells significantly enhances MM-induced angiogenesis.78 Gene expression profiling of MM plasma cells, cultured under normoxic or hypoxic conditions, has shown a significant upregulation in the expression of angiogenic genes such as interleukin-8 (IL-8) and VEGFA by hypoxia. Conversely, small interfering RNA-mediated knockdown of HIF1α in MM cell lines, in hypoxic conditions, significantly reduces MM-induced angiogenic response in vitro, as determined by the downregulation of pro-angiogenic molecules.75 Thus, given that hypoxia is constantly present in the BM environment, it is likely, that pro-angiogenic properties of normal and malignant plasma cells are, at least partially, determined by HIF1α accumulation in hypoxic BM environment. adrenomedullin One of the factors involved in angiogenesis is Adrenomedullin (AM). AM was initially identified as a vasodilator peptide isolated from a human pheochromocytoma.91 AM acts through the G-protein-coupled receptor calcitonin-receptor 1 like receptor (CLR), which form a complex with the receptor activity modifying proteins 2 (RAMP2) or 3 (RAMP3), leading to the formation of AM1 and AM2 receptors, respectively.92 AM and its receptors are expressed in several tissues, including the heart and blood vessels, kidneys, lungs, atrium, gastrointestinal 3 tract, spleen and thymus, endocrine glands and brain.91 AM is strongly induced by hypoxia, through transcription factor HIF1α,93,94 and is aberrantly expressed in 4 a variety of malignant tissues, where it mediates cell growth and survival in an autocrine fashion.95–97 Moreover, via paracrine action, adrenomedullin is involved 5 in blood vessel morphogenesis, vasculogenesis, and tumor angiogenesis.98,99 The importance of AM for blood vessel morphogenesis was highlighted by the finding 6 that homozygous AM-/- mice die in utero due to the aberrant angiogenesis in the placenta.100 Moreover, AM overexpression, by tumors cells, contributes to angio- 7 genesis.96,101,102 Interestingly, AM acts synergistically with VEGFA, since binding of AM to the CRLR/RAMP2 transactivates the VEGFR-2, which is responsible for most pro-angiogenic effects of VEGFA, including the stimulation of endothelial cell differentiation, proliferation, migration and morphogenesis. Importantly, however, blocking of VEGFA alone does not inhibit the AM-induced endothelial cells (EC) proliferation, showing that VEGFA is not required for AM-induced angiogenic response.103 Thus, currently available evidence points to AM as a potential target in cancer therapy, interfering with its role as autocrine growth factor and as paracrine stimulator of angiogenesis.

19 Chapter 1

The significance of aberrantly activated Wnt and NF-ĸB pathways in multiple myeloma progression wnt signaling in multiple myeloma The Wnt pathway plays an important role during vertebrate and invertebrate development. Wnt proteins are crucial signaling molecules, which regulate diverse cellular processes, such as differentiation, cell migration, survival, cell movement and proliferation. Recently, the importance of activated Wnt signaling and production of Wnt antagonists (e.g. DKK1) in multiple myeloma (MM) has been extensively discussed because of the crucial role of this cascades not only in myeloma progression 43,46,104–108 but also in the pathogenesis of osteolytic bone disease.109–111 In MM, functional signaling through two distinct Wnt pathways (Wnt/ β-catenin and Wnt/RhoA) has been identified. Activation of Wnt/RhoA signaling is associated with myeloma cell adhesion and drug resistance,105,106 whereas increased Wnt/β-catenin signaling by Wnt ligands, leads to enhancement in proliferation, survival, dissemination and also therapy resistance to lenalidomide.43,46,104,107,108 Importantly, the Wnt/β-catenin signaling plays also a significant direct role in the regulation of osteoblast function, responsible for bone formation, and indirectly controls of osteoclast function, leading to bone resorption. Disruption of Wnt signaling by myeloma plasma cell-derived Wnt inhibitors, mainly DKK1, results in suppression of osteoblast differentiation and enhanced osteoclast function.109 Conversely, however, direct stimulation of the Wnt pathway and blockage of DKK1 prevents MM-induced bone disease and may also affect myeloma cell growth in vivo.112–116 In chapter 2, we review the current understanding of the role of Wnt signaling in the pathogenesis of MM and of MM bone disease, and discuss the pros and cons of targeting this pathway in the treatment of MM. n-cadherin The critical molecule in canonical Wnt-signaling, β-catenin, has a dual function: by binding the transcription factor TCF it can control gene transcription, and by binding the adhesion molecules of the cadherin family it can control cell adhesion. Cadherins comprise a family of calcium-dependent cell adhesion molecules.117–119 Homophylic interaction between cadherins molecules regulate many basic developmental processes and are fundamental for the maintenance of both, embryonic and adult tissue structure.118,119 During gastrulation N-cadherin

20 General introduction is expressed by cells of the primitive streak undergoing epithelial-mesenchymal transition (EMT), a process characterized by down-regulation of E-cadherin and up-regulation of N-cadherin.119–121 Moreover, during osteogenesis N-cadherin is expressed at all stages of bone formation and its expression increases at the stages of nodule formation and mineralization.121–125 Several independent studies showed an important role of N-cadherin in osteoblasts differentiation and function. Moreover, N-cadherin expressed by osteoblasts has been suggested to play a role in creating the niche for the hematopoietic stem cells (HSC) maint- ance,126 although it is controversial whether or not HSC themselves express N-cadherin.127–131 In adult individuals, N-cadherin is expressed mainly by neural tissue, myocytes, and endothelial cells, contributing to stability of the vessels.132 In cancer, expression of N-cadherin is related to EMT, which is associated with an invasive phenotype and enhanced metastatic potential.120 A possible explanation for the connection between N-cadherin expression and invasiveness is its interaction with FGFR-1. By preventing internalization of FGFR-1, N-cadherin may 1 cause sustained MAPK activation and overexpression of MMP-9.120,133 In lymphoid malignancies, expression of N-cadherin was reported in T cell leukemia cell lines134–138 and t(1;19) translocated B-LBL/ALL.135 These studies also revealed that N-cadherin can mediate adhesion of lymphoma cells to stroma. Altogether, 3 these finding suggest the potential role of N-cadherin in lymphocyte development as well as in the process of malignant transformation. 4 nf-ĸb signaling in multiple myeloma progression 5 It is well established that activation of the NF-κB pathway is important for the survival of healthy and malignant plasma cells.139,140 Of note, around 50% of MM cell 6 lines (MMCLs) and most primary MM samples have an increased level of NF-κB activity, based on expression of a transcription signature of 11 MM-NF-κB target 7 genes.48 The high level of NF-κB activity in most MM tumors results from extrin- sic signaling by APRIL and BAFF ligands, which are produced by bone marrow stromal cells.60,139 These two proteins are the main NF-κB inducing ligands, and are crucial survival factors for normal plasma cells and MM cells in the BM. In addition, mutations leading to constitutive NF-κB signaling have been observed in MM. These involve the TRAF2, TRAF3, CYLD, cIAP1/cIAP2, NFKB1, NFKB2, CD40, LTBR, TACI, and NIK leading to constitutive activation of NF-κB signaling.44 According to an integrated analysis of high-density oligonucleotide array (CGH) and gene expression profiling data, NF-κB mutations are present in at least 17% of primary MMs and 40% of MM cell lines.44 These activating mutations in NF-κB components, which occur during disease progression, render

21 Chapter 1

MM tumors less dependent upon paracrine signals from the bone marrow. This increased autonomy of the malignant plasma cells allows MM dissemination and progression to plasma cell leukemia. tumor-suppressing function of cyld One of the many molecules involved in controlling NFĸ-B activity is CYLD. CYLD was initially identified as a tumor suppressor for familial cylindromatosis, an autosomal dominant predisposing to multiple benign neoplasms of skin appendages. Cylindromatosis patients carry heterozygous germ-line mutations in carboxyl terminal end of the CYLD gene, whereas the wild-type CYLD allele undergoes loss of heterozygosity, thus emphazing the tumor suppressor role of CYLD.141,142 CYLD encodes a 107 kDa protein that is ubiquitously expressed and contains a deubiqutinating domain (UCH) at the C-terminus, which removes lysine 63 (K63) linked polyubiquitination from several target distinct proteins.143 The initial clue for a signaling function of CYLD came from a RNAi-based functional screening study, which identified it as a DUB that negatively regulates NF-κB pathway activation,144 by removing Lys-63-linked polyubiquitin chains from TNF receptor associated factor 2 and 6 (TRAF2 and TRAF6) and NF-ĸB essential modulator (NEMO).144–146 More recent studies identified additional molecular targets for CYLD, but further validation is required to determine the functional consequences of these interactions. Dysregulation of NF-κB activity, thereby increasing apoptosis resistance, has been proposed as mechanism of transformation upon loss of CYLD. However, inhibition of NF-κB pathway is not the only function of CYLD. For instance, Stegmeier et al. reported a role of CYLD in the regulation of the cell cycle progression, controlling mitotic entry.147 Interestingly, this novel function required the DUB catalytic activity of CYLD, but was independent of the NF-κB pathway. It was proposed that CYLD positively regulates the function of polo-like kinase 1 (PLK-1), probably by removing K-63-linked ubiquitin chains from PLK-1 or its upstream regulators.147 Thus, according to this study, CYLD may positively regulate the cellular proliferation via its interaction with PLK-1 kinase. Indeed, as most of cylindromas are benign, it can be hypothesized that the proliferative disadvantage, caused by downregulation of CYLD function, helps to restrain tumor growth and inhibits tumor progression. However, in contradiction with this study, Wickstrom et al. revealed that CYLD negatively regulates proliferation of keratinocytes and melanoma cells, which involves association with alpha-tubulin and microtubules via its CAP-Gly domain. Translocation of CYLD to the perinuclear region is achieved by an inhibitory interaction of CYLD with histone deactylase-6 (HDAC6), resulting

22 General introduction in elevated perinuclear levels of acetylated alpha tubulin. This facilitates the interaction between CYLD and BCL-3, leading to deubiquitination of BCL-3 and preventing its nuclear translocation. As a consequence, the NF-κB p50/BCL-3 and NF-κB p52/BCL-3 transcription is inhibited, reducing the expression of cyclin D1. As expression of cyclin D1 is crucial for G1/S transition, it is likely that CYLD delays G1/S by inhibiting the expression of cyclin D1 transition.148 An- other important function of CYLD is related to the presence of three cytoskeleton-associated protein glycine-rich (CAP-Gly) domains, which are responsible for CYLD association with microtubules. CYLD has been shown to associate with microtubules by binding to tubulin and to promote microtubule assembly and stability. In accordance with this function, CYLD is important for cell migration, as was revealed in vitro in a wound healing assay.149 Interestingly, recent studies by Tauriello et al. uncovered a novel interesting function of CYLD, i.e. a negative regulatory role in proximal events in Wnt/β-catenin signaling.150 It was shown that silencing of CYLD markedly enhances Wnt-induced accumulation 1 of β-catenin and target gene activation. At the molecular level, CYLD removes K-63-linked ubiquitin from Dvl protein. The Dvl protein contains a conserved DIX domain which mediates interaction between Dvl and Axin, playing an important role in Wnt signal transduction. Thus, in CYLD knockdown cells, en- 3 hanced ubiquitination of Dvl, in particular of its polymerization-prone DIX domain, results in enhanced Wnt responses.150 4 cyld in multiple myeloma 5 Importantly, recent studies reported that CYLD is a target for genetic aberration in malignant plasma cells. For instance, Keats et al., identified biallelic deletion 6 and inactivating mutations in the CYLD gene.44 In addition, Annuziata et al. reported bi-allelic loss of the CYLD in primary MM cases with high NF-κB signa- 7 ture expression.48 Moreover, LOH of chromosome 16q, in which the CYLD gene resides, is observed in a large subset of MM patients. Although multiple genes are affected in the MM cases with 16q LOH, CYLD was reported as one of the two most likely genes responsible for the poor prognosis of patients with 16q LOH.151

Aims and outline of this thesis

The studies described in this thesis investigated the communication between malignant plasma cells and the bone marrow environment. During disease progression this interaction is bidirectional; MM cells destroy the normal BM

23 Chapter 1

homeostasis, to create the optimal tumor niche for their progression, whereas the BM environment, via adhesive interaction or secreted cytokines, chemokines and growth factors, support the proliferation and survival of the MM plasma cells. Chapter 2 discusses the role of aberrantly activated Wnt pathway in MM, specifically in tumor progression and the development of osteolytic bone disease. Al- though this subject is extensively studied, the role of Wnt still is debated since dis- parate results have been reported. This chapter extensively discuss the importance of Wnt and the current controversies related to the issue of Wnt in the progression of MM. In chapter 3, we investigated the presence of a Wnt-pathway negative feedback loop in malignant plasma cells. We revealed that the loss of DKK1 in malignant plasma cells is associated with hyperactivation of the Wnt pathway and tumor progression. Our data revealed that aberrant promoter methylation is a major mechanism causing DKK1 downregulation during MM progression. In chapter 4, we show that CYLD expression is frequently lost in MM, what strongly correlates with a proliferative gene-expression profile in MM patients. Our in vitro results demonstrated that the loss of CYLD hyperactivates Wnt and NF-ĸB signaling and significantly enhances proliferation of malignant plasma cells. In line with this observations, we found that low expression of CYLD is correlated with inferior progression free survival and overall survival in MM patients, pointing to the loss of CYLD expression as a new prognostic factor in MM. In chapter 5, we identified N-cadherin, as being aberrantly expressed by the malignant plasma cells of a subset of MM patients and analyzed its role in the progression of MM. We determined that N-cadherin, which mediates homophilic adhesion, is involved in the BM localization of MM cells, contributing to inhibition of osteoblasts differentiation, thereby revealing a role of this protein in MM-induced osteolytic bone disease. In chapter 6, we identified adrenomedullin as an angiogenic molecule aberrantly expressed by the malignant plasma cells of a subset of MM patients. Our study revealed that AM is the most highly upregulated gene upon exposure of MM cells to hypoxia. Moreover, we demonstrated that MM-driven promotion of endothelial cell proliferation and tube formation is augmented by expression of AM and strongly repressed by inhibition of endogenous AM activity. Finally, chapter 7 discusses and summarizes the results presented in this thesis.

24 General introduction

References

1. Hardy RR, Hayakawa K. B Cell development pathways. Annu Rev Immunol 2001; 19: 595–621. 2. Tonegawa S. Somatic generation of antibody diversity. Nature 1983; 302(5909): 575–581. 3. Rajewsky K. Clonal selection and learning in the antibody system. Nature 1996; 381(6585): 751–758. 4. Nussenzweig MC. Immune receptor editing: revise and select. Cell 1998; 95(7): 875–878. 5. Fugmann SD, Lee AI, Shockett PE, Villey IJ, Schatz DG. The RAG proteins and V(D) J recombination: complexes, ends, and transposition. Annu Rev Immunol 2000; 18: 495–527. 6. Liu YJ, Zhang J, Lane PJ, Chan EY, MacLennan IC. Sites of specific B cell activation in primary and secondary responses to T cell-dependent and T cell-independent antigens. Eur J Immunol 1991; 21(12): 2951–2962. 1 7. Klein U, Dalla-Favera R. Germinal centres: role in B-cell physiology and malignancy. Nat Rev Immunol 2008; 8(1): 22–33. 8. Kuppers R, Zhao M, Hansmann ML, Rajewsky K. Tracing B cell development in human germinal centres by molecular analysis of single cells picked from histological sections. Embo J 1993; 12(13): 4955–4967. 3 9. Jacob J, Kelsoe G, Rajewsky K, Weiss U. Intraclonal generation of antibody mutants in germinal centres. Nature 1991; 354(6352): 389–392. 4 10. Berek C, Berger A, Apel M. Maturation of the immune response in germinal centers. Cell 1991; 67(6): 1121–1129. 5 11. Di Noia JM, Neuberger MS. Molecular mechanisms of antibody somatic hypermutation. Annu Rev Biochem 2007; 76: 1–22. 6 12. MacLennan IC. Germinal centers. Annu Rev Immunol 1994; 12: 117–139. 13. Liu YJ, Arpin C. Germinal center development. Immunol Rev 1997; 156: 111–126. 7 14. Dubois B, Barthelemy C, Durand I, Liu YJ, Caux C, et al. Toward a role of dendritic cells in the germinal center reaction: triggering of B cell proliferation and isotype switching. J Immunol 1999; 162(6): 3428–3436. 15. Fuller KA, Kanagawa O, Nahm MH. T cells within germinal centers are specific for the immunizing antigen. J Immunol 1993; 151(9): 4505–4512. 16. Kawabe T, Naka T, Yoshida K, Tanaka T, Fujiwara H, et al. The immune responses in CD40-deficient mice: impaired immunoglobulin class switching and germinal center formation. Immunity 1994; 1(3): 167–178. 17. Cosgrove D, Gray D, Dierich A, Kaufman J, Lemeur M, et al. Mice lacking MHC class II molecules. Cell 1991; 66(5): 1051–1066. 18. Arpin C, Dechanet J, Van Kooten C, Merville P, Grouard G, et al. Generation of memory B cells and plasma cells in vitro. Science 1995; 268(5211): 720–722.

25 Chapter 1

19. Lenz G, Staudt LM. Aggressive lymphomas. N Engl J Med; 362(15): 1417–1429. 20. Allen CD, Okada T, Cyster JG. Germinal-center organization and cellular dynamics. Immunity 2007; 27(2): 190–202. 21. Kuppers R, Klein U, Hansmann ML, Rajewsky K. Cellular origin of human B-cell lymphomas. N Engl J Med 1999; 341(20): 1520–1529. 22. Lo Coco F, Ye BH, Lista F, Corradini P, Offit K, et al. Rearrangements of the BCL6 gene in diffuse large cell non-Hodgkin’s lymphoma. Blood 1994; 83(7): 1757–1759. 23. Kuppers R, Dalla-Favera R. Mechanisms of chromosomal translocations in B cell lymphomas. Oncogene 2001; 20(40): 5580–5594. 24. Bergsagel PL, Kuehl WM. Chromosome translocations in multiple myeloma. Onco- gene 2001; 20(40): 5611–5622. 25. Bastard C, Deweindt C, Kerckaert JP, Lenormand B, Rossi A, et al. LAZ3 rearrangements in non-Hodgkin’s lymphoma: correlation with histology, immunophenotype, karyotype, and clinical outcome in 217 patients. Blood 1994; 83(9): 2423–2427. 26. Liu M, Duke JL, Richter DJ, Vinuesa CG, Goodnow CC, et al. Two levels of protection for the B cell genome during somatic hypermutation. Nature 2008; 451(7180): 841–845. 27. Stevenson F, Sahota S, Zhu D, Ottensmeier C, Chapman C, et al. Insight into the origin and clonal history of B-cell tumors as revealed by analysis of immunoglobulin variable region genes. Immunol Rev 1998; 162: 247–259. 28. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000; 403(6769): 503–511. 29. Wright G, Tan B, Rosenwald A, Hurt EH, Wiestner A, et al. A gene expression-based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proc Natl Acad Sci U S A 2003; 100(17): 9991–9996. 30. Dave SS, Fu K, Wright GW, Lam LT, Kluin P, et al. Molecular diagnosis of Burkitt’s lymphoma. N Engl J Med 2006; 354(23): 2431–2442. 31. Isaacson PG, Du MQ. MALT lymphoma: from morphology to molecules. Nat Rev Cancer 2004; 4(8): 644–653. 32. Basso K, Liso A, Tiacci E, Benedetti R, Pulsoni A, et al. Gene expression profiling of hairy cell leukemia reveals a phenotype related to memory B cells with altered expression of chemokine and adhesion receptors. J Exp Med 2004; 199(1): 59–68. 33. Hallek M, Bergsagel PL, Anderson KC. Multiple myeloma: increasing evidence for a multistep transformation process. Blood 1998; 91(1): 3–21. 34. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2(3): 175–187. 35. Munshi NC. Plasma cell disorders: an historical perspective. Hematology Am Soc Hematol Educ Program 2008: 297. 36. Kyle RA, Therneau TM, Rajkumar SV, Offord JR, Larson DR, et al. A long-term study of prognosis in monoclonal gammopathy of undetermined significance. N Engl J Med 2002; 346(8): 564–569.

26 General introduction

37. Fonseca R, Debes-Marun CS, Picken EB, Dewald GW, Bryant SC, et al. The recurrent IgH translocations are highly associated with nonhyperdiploid variant multiple myeloma. Blood 2003; 102(7): 2562–2567. 38. Fonseca R, Blood E, Rue M, Harrington D, Oken MM, et al. Clinical and biologic implications of recurrent genomic aberrations in myeloma. Blood 2003; 101(11): 4569–4575. 39. Smadja NV, Leroux D, Soulier J, Dumont S, Arnould C, et al. Further cytogenetic characterization of multiple myeloma confirms that 14q32 translocations are a very rare event in hyperdiploid cases. Genes Chromosomes Cancer 2003; 38(3): 234–239. 40. Seidl S, Kaufmann H, Drach J. New insights into the pathophysiology of multiple myeloma. Lancet Oncol 2003; 4(9): 557–564. 41. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009; 374(9686): 324–339. 42. Mitsiades CS, Mitsiades N, Munshi NC, Anderson KC. Focus on multiple myeloma. Cancer Cell 2004; 6(5): 439–444. 1 43. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/TCF transcriptional complex in the treatment of multiple myeloma. Proc Natl Acad Sci U S A 2007; 104(18): 7516–7521. 44. Keats JJ, Fonseca R, Chesi M, Schop R, Baker A, et al. Promiscuous mutations acti- 3 vate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell 2007; 12(2): 131–44. 45. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is 4 a target of Wnt/{beta}-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 5 46. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci 6 U S A 2004; 101(16): 6122–6127. 47. Bjorklund CC, Ma W, Wang ZQ, Davis RE, Kuhn DJ, et al. Evidence of a role for acti- 7 vation of Wnt/beta-catenin signaling in the resistance of plasma cells to lenalidomide. J Biol Chem 2011; 286(13): 11009–11020. 48. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12(2): 115–130. 49. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, et al. Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood 2005; 106(1): 296–303. 50. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108(6): 2020–2028. 51. Broyl A, Hose D, Lokhorst H, de Knegt Y, Peeters J, et al. Gene expression profiling for molecular classification of multiple myeloma in newly diagnosed patients. Blood 2010; 116(14): 2543–2553.

27 Chapter 1

52. Burger JA, Ghia P, Rosenwald A, Caligaris-Cappio F. The microenvironment in mature B-cell malignancies: a target for new treatment strategies. Blood 2009; 114(16): 3367–3375. 53. Noll JE, Williams SA, Purton LE, Zannettino AC. Tug of war in the haematopoietic stem cell niche: do myeloma plasma cells compete for the HSC niche? Blood Can- cer J 2012; 2: e91. 54. Podar K, Richardson PG, Hideshima T, Chauhan D, Anderson KC. The malignant clone and the bone-marrow environment. Best Pract Res Clin Haematol 2007; 20(4): 597–612. 55. Vacca A, Ribatti D, Presta M, Minischetti M, Iurlaro M, et al. Bone marrow neovascularization, plasma cell angiogenic potential, and matrix metalloproteinase-2 secretion parallel progression of human multiple myeloma. Blood 1999; 93(9): 3064–3073. 56. Barille S, Akhoundi C, Collette M, Mellerin MP, Rapp MJ, et al. Metalloproteinases in multiple myeloma: production of matrix metalloproteinase-9 (MMP-9), activation of proMMP-2, and induction of MMP-1 by myeloma cells. Blood 1997; 90(4): 1649–1655. 57. Kawano M, Hirano T, Matsuda T, Taga T, Horii Y, et al. Autocrine generation and requirement of BSF-2/IL-6 for human multiple myelomas. Nature 1988; 332(6159): 83–85. 58. Ehrlich LA, Chung HY, Ghobrial I, Choi SJ, Morandi F, et al. IL-3 is a potential inhibitor of osteoblast differentiation in multiple myeloma. Blood 2005; 106(4): 1407–1414. 59. Cozzolino F, Torcia M, Aldinucci D, Rubartelli A, Miliani A, et al. Production of interleukin-1 by bone marrow myeloma cells. Blood 1989; 74(1): 380–387. 60. Tai YT, Li XF, Breitkreutz I, Song W, Neri P, et al. Role of B-cell-activating factor in adhesion and growth of human multiple myeloma cells in the bone marrow microenvironment. Cancer Res 2006; 66(13): 6675–6682. 61. Podar K, Anderson KC. The pathophysiologic role of VEGF in hematologic malignancies: therapeutic implications. Blood 2005; 105(4): 1383–1395. 62. Borset M, Seidel C, Hjorth-Hansen H, Waage A, Sundan A. The role of hepatocyte growth factor and its receptor c-Met in multiple myeloma and other blood malignancies. Leuk Lymphoma 1999; 32(3–4): 249–256. 63. Urashima M, Ogata A, Chauhan D, Hatziyanni M, Vidriales MB, et al. Transforming growth factor-beta1: differential effects on multiple myeloma versus normal B cells. Blood 1996; 87(5): 1928–1938. 64. Hideshima T, Chauhan D, Hayashi T, Podar K, Akiyama M, et al. The biological sequelae of stromal cell-derived factor-1alpha in multiple myeloma. Mol Cancer Ther 2002; 1(7): 539–544. 65. Podar K, Hideshima T, Chauhan D, Anderson KC. Targeting signalling pathways for the treatment of multiple myeloma. Expert Opin Ther Targets 2005; 9(2): 359–381.

28 General introduction

66. Jensen GS, Belch AR, Mant MJ, Ruether BA, Yacyshyn BR, et al. Expression of multiple beta 1 integrins on circulating monoclonal B cells in patients with multiple myeloma. Am J Hematol 1993; 43(1): 29–36. 67. Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell adhesion mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in human myeloma cell lines. Blood 1999; 93(5): 1658–1667. 68. Vacca A, Ria R, Presta M, Ribatti D, Iurlaro M, et al. alpha(v)beta(3) integrin engagement modulates cell adhesion, proliferation, and protease secretion in human lymphoid tumor cells. Exp Hematol 2001; 29(8): 993–1003. 69. Ria R, Vacca A, Ribatti D, Di Raimondo F, Merchionne F, et al. Alpha(v)beta(3) integrin engagement enhances cell invasiveness in human multiple myeloma. Haemato- logica 2002; 87(8): 836–845. 70. Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D, Anderson KC. Adhesion of human myeloma-derived cell lines to bone marrow stromal cells stimulates interleukin-6 secretion. Blood 1993; 82(12): 3712–3720. 71. Parmar K, Mauch P, Vergilio JA, Sackstein R, Down JD. Distribution of hematopoi- 1 etic stem cells in the bone marrow according to regional hypoxia. Proc Natl Acad Sci U S A 2007; 104(13): 5431–5436. 72. Levesque JP, Winkler IG, Hendy J, Williams B, Helwani F, et al. Hematopoietic pro- genitor cell mobilization results in hypoxia with increased hypoxia-inducible tran- 3 scription factor-1 alpha and vascular endothelial growth factor A in bone marrow. Stem Cells 2007; 25(8): 1954–1965. 4 73. Asosingh K, De Raeve H, de Ridder M, Storme GA, Willems A, et al. Role of the hypoxic bone marrow microenvironment in 5T2MM murine myeloma tumor progression. Haematologica 2005; 90(6): 810–817. 5 74. Azab AK, Hu J, Quang P, Azab F, Pitsillides C, et al. Hypoxia promotes dissemination of multiple myeloma through acquisition of epithelial to mesenchymal transition-like 6 features. Blood 2012; 119(24): 5782–5794. 75. Colla S, Storti P, Donofrio G, Todoerti K, Bolzoni M, et al. Low bone marrow oxygen 7 tension and hypoxia-inducible factor-1alpha overexpression characterize patients with multiple myeloma: role on the transcriptional and proangiogenic profiles of CD138(+) cells. Leukemia 2010; 24(11): 1967–1970. 76. Giatromanolaki A, Bai M, Margaritis D, Bourantas KL, Koukourakis MI, et al. Hypox- ia and activated VEGF/receptor pathway in multiple myeloma. Anticancer Res 2010; 30(7): 2831–2836. 77. Martin SK, Diamond P, Gronthos S, Peet DJ, Zannettino AC. The emerging role of hypoxia, HIF1 and HIF-2 in multiple myeloma. Leukemia 2011; 25(10): 1533–1542. 78. Martin SK, Diamond P, Williams SA, To LB, Peet DJ, et al. Hypoxia-inducible factor-2 is a novel regulator of aberrant CXCL12 expression in multiple myeloma plasma cells. Haematologica 2010; 95(5): 776–784.

29 Chapter 1

79. Zhang J, Sattler M, Tonon G, Grabher C, Lababidi S, et al. Targeting angiogenesis via a c-Myc/hypoxia-inducible factor-1alpha-dependent pathway in multiple myeloma. Cancer Res 2009; 69(12): 5082–5090. 80. Jakob C, Sterz J, Zavrski I, Heider U, Kleeberg L, et al. Angiogenesis in multiple myeloma. Eur J Cancer 2006; 42(11): 1581–1590. 81. Vacca A, Ria R, Semeraro F, Merchionne F, Coluccia M, et al. Endothelial cells in the bone marrow of patients with multiple myeloma. Blood 2003; 102(9): 3340–3348. 82. Vacca A, Ribatti D, Roncali L, Ranieri G, Serio G, et al. Bone marrow angiogenesis and progression in multiple myeloma. Br J Haematol 1994; 87(3): 503–508. 83. Hillengass J, Wasser K, Delorme S, Kiessling F, Zechmann C, et al. Lumbar bone marrow microcirculation measurements from dynamic contrast-enhanced magnetic resonance imaging is a predictor of event-free survival in progressive multiple myeloma. Clin Cancer Res 2007; 13(2 Pt 1): 475–481. 84. Pour L, Svachova H, Adam Z, Almasi M, Buresova L, et al. Levels of angiogenic factors in patients with multiple myeloma correlate with treatment response. Ann Hematol 2010; 89(4): 385–389. 85. Rajkumar SV, Leong T, Roche PC, Fonseca R, Dispenzieri A, et al. Prognostic value of bone marrow angiogenesis in multiple myeloma. Clin Cancer Res 2000; 6(8): 3111–3116. 86. Rana C, Sharma S, Agrawal V, Singh U. Bone marrow angiogenesis in multiple myeloma and its correlation with clinicopathological factors. Ann Hematol 2010; 89(8): 789–794. 87. Alexandrakis MG, Passam FJ, Ganotakis E, Dafnis E, Dambaki C, et al. Bone marrow microvascular density and angiogenic growth factors in multiple myeloma. Clin Chem Lab Med 2004; 42(10): 1122–1126. 88. Cibeira MT, Rozman M, Segarra M, Lozano E, Rosinol L, et al. Bone marrow angiogenesis and angiogenic factors in multiple myeloma treated with novel agents. Cytokine 2008; 41(3): 244–253. 89. Kumar S, Witzig TE, Timm M, Haug J, Wellik L, et al. Bone marrow angiogenic ability and expression of angiogenic cytokines in myeloma: evidence favoring loss of marrow angiogenesis inhibitory activity with disease progression. Blood 2004; 104(4): 1159–1165. 90. Hose D, Moreaux J, Meissner T, Seckinger A, Goldschmidt H, et al. Induction of angiogenesis by normal and malignant plasma cells. Blood 2009; 114(1): 128–143. 91. Kitamura K, Kangawa K, Kawamoto M, Ichiki Y, Nakamura S, et al. Adrenomedul- lin: a novel hypotensive peptide isolated from human pheochromocytoma. Biochem Biophys Res Commun 1993; 192(2): 553–560. 92. Fernandez-Sauze S, Delfino C, Mabrouk K, Dussert C, Chinot O, et al. Effects of adrenomedullin on endothelial cells in the multistep process of angiogenesis: involvement of CRLR/RAMP2 and CRLR/RAMP3 receptors. Int J Cancer 2004; 108(6): 797–804.

30 General introduction

93. Garayoa M, Martinez A, Lee S, Pio R, An WG, et al. Hypoxia-inducible factor-1 (HIF1) up-regulates adrenomedullin expression in human tumor cell lines during oxygen deprivation: a possible promotion mechanism of carcinogenesis. Mol Endocrinol 2000; 14(6): 848–862. 94. Nguyen SV, Claycomb WC. Hypoxia regulates the expression of the adrenomedullin and HIF1 genes in cultured HL-1 cardiomyocytes. Biochem Biophys Res Commun 1999; 265(2): 382–386. 95. Miller MJ, Martinez A, Unsworth EJ, Thiele CJ, Moody TW, et al. Adrenomedullin expression in human tumortumor cell lines. Its potential role as an autocrine growth factor. J Biol Chem 1996; 271(38): 23345–23351. 96. Oehler MK, Hague S, Rees MC, Bicknell R. Adrenomedullin promotes formation of xenografted endometrial tumors by stimulation of autocrine growth and angiogenesis. Oncogene 2002; 21(18): 2815–2821. 97. Ramachandran V, Arumugam T, Hwang RF, Greenson JK, Simeone DM, et al. Adre- nomedullin is expressed in pancreatic cancer and stimulates cell proliferation and invasion in an autocrine manner via the adrenomedullin receptor, ADMR. Cancer 1 Res 2007; 67(6): 2666–2675. 98. Deville JL, Salas S, Figarella-Branger D, Ouafik L, Daniel L. Adrenomedullin as a therapeutic target in angiogenesis. Expert Opin Ther Targets 2010; 14(10): 1059–1072. 99. Kim W, Moon SO, Sung MJ, Kim SH, Lee S, et al. Angiogenic role of adrenomedul- 3 lin through activation of Akt, mitogen-activated protein kinase, and focal adhesion kinase in endothelial cells. Faseb J 2003; 17(13): 1937–1939. 4 100. Caron KM, Smithies O. Extreme hydrops fetalis and cardiovascular abnormalities in mice lacking a functional Adrenomedullin gene. Proc Natl Acad Sci U S A 2001; 98(2): 615–619. 5 101. Ishikawa T, Chen J, Wang J, Okada F, Sugiyama T, et al. Adrenomedullin antagonist suppresses in vivo growth of human pancreatic cancer cells in SCID mice by sup- 6 pressing angiogenesis. Oncogene 2003; 22(8): 1238–1242. 102. Martinez A, Julian M, Bregonzio C, Notari L, Moody TW, et al. Identification of vaso- 7 active nonpeptidic positive and negative modulators of adrenomedullin using a neutralizing antibody-based screening strategy. Endocrinology 2004; 145(8): 3858–65. 103. Guidolin D, Albertin G, Spinazzi R, Sorato E, Mascarin A, et al. Adrenomedullin stimulates angiogenic response in cultured human vascular endothelial cells: involvement of the vascular endothelial growth factor receptor 2. Peptides 2008; 29(11): 2013–2023. 104. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/beta-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 105. Qiang YW, Walsh K, Yao L, Kedei N, Blumberg PM, et al. Wnts induce migration and invasion of myeloma plasma cells. Blood 2005; 106(5): 1786–1793.

31 Chapter 1

106. Kobune M, Chiba H, Kato J, Kato K, Nakamura K, et al. Wnt3/RhoA/ROCK signaling pathway is involved in adhesion-mediated drug resistance of multiple myeloma in an autocrine mechanism. Mol Cancer Ther 2007; 6(6): 1774–1784. 107. Ashihara E, Kawata E, Nakagawa Y, Shimazaski C, Kuroda J, et al. beta-catenin small interfering RNA successfully suppressed progression of multiple myeloma in a mouse model. Clin Cancer Res 2009; 15(8): 2731–2738. 108. Bjorklund CC, Ma W, Wang ZQ, Davis RE, Kuhn DJ, et al. Evidence of a role for activation of Wnt/beta-catenin signaling in the resistance of plasma cells to lenalidomide. J Biol Chem 2011; 286(13): 11009–11020. 109. Qiang YW, Chen Y, Stephens O, Brown N, Chen B, et al. Myeloma-derived Dick- kopf-1 disrupts Wnt-regulated osteoprotegerin and RANKL production by osteoblasts: a potential mechanism underlying osteolytic bone lesions in multiple myeloma. Blood 2008; 112(1): 196–207. 110. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med 2003; 349(26): 2483–2494. 111. Haaber J, Abildgaard N, Knudsen LM, Dahl IM, Lodahl M, et al. Myeloma cell expression of 10 candidate genes for osteolytic bone disease. Only overexpression of DKK1 correlates with clinical bone involvement at diagnosis. Br J Haematol 2008; 140(1): 25–35. 112. Edwards CM, Edwards JR, Lwin ST, Esparza J, Oyajobi BO, et al. Increasing Wnt signaling in the bone marrow microenvironment inhibits the development of myeloma bone disease and reduces tumor burden in bone in vivo. Blood 2008; 111(5): 2833–2842. 113. Fulciniti M, Tassone P, Hideshima T, Vallet S, Nanjappa P, et al. Anti-DKK1 mAb (BHQ880) as a potential therapeutic agent for multiple myeloma. Blood 2009; 114(2): 371–379. 114. Gunn WG, Krause U, Lee N, Gregory CA. Pharmaceutical inhibition of glycogen synthetase kinase-3beta reduces multiple myeloma-induced bone disease in a novel murine plasmacytoma xenograft model. Blood 2011; 117(5): 1641–1651. 115. Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy JD, Jr., et al. Inhibit- ing Dickkopf-1 (Dkk1) removes suppression of bone formation and prevents the development of osteolytic bone disease in multiple myeloma. J Bone Miner Res 2009; 24(3): 425–436. 116. Yaccoby S, Ling W, Zhan F, Walker R, Barlogie B, et al. Antibody-based inhibition of DKK1 suppresses tumor-induced bone resorption and multiple myeloma growth in vivo. Blood 2007; 109(5): 2106–2111. 117. Shapiro L, Weis WI. Structure and biochemistry of cadherins and catenins. Cold Spring Harb Perspect Biol 2009; 1(3): a003053. 118. Gumbiner BM. Regulation of cadherin-mediated adhesion in morphogenesis. Nat Rev Mol Cell Biol 2005; 6(8): 622–634.

32 General introduction

119. Derycke LD, Bracke ME. N-cadherin in the spotlight of cell-cell adhesion, differentiation, embryogenesis, invasion and signalling. Int J Dev Biol 2004; 48(5–6): 463–76. 120. Mariotti A, Perotti A, Sessa C, Ruegg C. N-cadherin as a therapeutic target in cancer. Expert Opin Investig Drugs 2007; 16(4): 451–465. 121. Hatta K, Takeichi M. Expression of N-cadherin adhesion molecules associated with early morphogenetic events in chick development. Nature 1986; 320(6061): 447–449. 122. Tuan RS. Cellular signaling in developmental chondrogenesis: N-cadherin, Wnts, and BMP-2. J Bone Joint Surg Am 2003; 85-A Suppl 2: 137–141. 123. Stains JP, Civitelli R. Cell-cell interactions in regulating osteogenesis and osteoblast function. Birth Defects Res C Embryo Today 2005; 75(1): 72–80. 124. Marie PJ. Role of N-cadherin in bone formation. J Cell Physiol 2002; 190(3): 297–305. 125. Ferrari SL, Traianedes K, Thorne M, Lafage-Proust MH, Genever P, et al. A role for N-cadherin in the development of the differentiated osteoblastic phenotype. J Bone Miner Res 2000; 15(2): 198–208. 126. Zhang J, Niu C, Ye L, Huang H, He X, et al. Identification of the haematopoietic stem cell niche and control of the niche size. Nature 2003; 425(6960): 836–841. 1 127. Li P, Zon LI. Resolving the controversy about N-cadherin and hematopoietic stem cells. Cell Stem Cell 2010; 6(3): 199–202. 128. Kiel MJ, Radice GL, Morrison SJ. Lack of evidence that hematopoietic stem cells depend on N-cadherin-mediated adhesion to osteoblasts for their maintenance. Cell 3 Stem Cell 2007; 1(2): 204–217. 129. Hosokawa K, Arai F, Yoshihara H, Nakamura Y, Gomei Y, et al. Function of oxida- 4 tive stress in the regulation of hematopoietic stem cell-niche interaction. Biochem Biophys Res Commun 2007; 363(3): 578–583. 130. Hosokawa K, Arai F, Yoshihara H, Iwasaki H, Hembree M, et al. Cadherin-based 5 adhesion is a potential target for niche manipulation to protect hematopoietic stem cells in adult bone marrow. Cell Stem Cell 2010; 6(3): 194–198. 6 131. Hooper AT, Butler J, Petit I, Rafii S. Does N-cadherin regulate interaction of hematopoietic stem cells with their niches? Cell Stem Cell 2007; 1(2): 127–129. 7 132. Luo Y, Radice GL. N-cadherin acts upstream of VE-cadherin in controlling vascular morphogenesis. J Cell Biol 2005; 169(1): 29–34. 133. Suyama K, Shapiro I, Guttman M, Hazan RB. A signaling pathway leading to metastasis is controlled by N-cadherin and the FGF receptor. Cancer Cell 2002; 2(4): 301–314. 134. Tsutsui J, Moriyama M, Arima N, Ohtsubo H, Tanaka H, et al. Expression of cadherin- catenin complexes in human leukemia cell lines. J Biochem 1996; 120(5): 1034–1039. 135. Nygren MK, Dosen-Dahl G, Stubberud H, Walchli S, Munthe E, et al. beta-catenin is involved in N-cadherin-dependent adhesion, but not in canonical Wnt signaling in E2A-PBX1-positive B acute lymphoblastic leukemia cells. Exp Hematol 2009; 37(2): 225–233.

33 Chapter 1

136. Matsuyoshi N, Toda K, Imamura S. N-cadherin expression in human adult T-cell leukemia cell line. Arch Dermatol Res 1998; 290(4): 223–225. 137. Makagiansar IT, Yusuf-Makagiansar H, Ikesue A, Calcagno AM, Murray JS, et al. N-cadherin involvement in the heterotypic adherence of malignant T-cells to epi- thelia. Mol Cell Biochem 2002; 233(1–2): 1–8. 138. Kawamura-Kodama K, Tsutsui J, Suzuki ST, Kanzaki T, Ozawa M. N-cadherin expressed on malignant T cell lymphoma cells is functional, and promotes heterotypic adhesion between the lymphoma cells and mesenchymal cells expressing N-cadherin. J Invest Dermatol 1999; 112(1): 62–66. 139. Moreaux J, Veyrune JL, De Vos J, Klein B. APRIL is overexpressed in cancer: link with tumor progression. BMC Cancer 2009; 9: 83. 140. O’Connor BP, Raman VS, Erickson LD, Cook WJ, Weaver LK, et al. BCMA is essential for the survival of long-lived bone marrow plasma cells. J Exp Med 2004; 199(1): 91–98. 141. Bignell GR, Warren W, Seal S, Takahashi M, Rapley E, et al. Identification of the familial cylindromatosis tumor-suppressor gene. Nat Genet 2000; 25(2): 160–165. 142. Massoumi R, Paus R. Cylindromatosis and the CYLD gene: new lessons on the molecular principles of epithelial growth control. Bioessays 2007; 29(12): 1203–1214. 143. Komander D, Lord CJ, Scheel H, Swift S, Hofmann K, et al. The structure of the CYLD USP domain explains its specificity for Lys63-linked polyubiquitin and reveals a B box module. Mol Cell 2008; 29(4): 451–64. 144. Brummelkamp TR, Nijman SM, Dirac AM, Bernards R. Loss of the cylindromatosis tumor suppressor inhibits apoptosis by activating NF-kappaB. Nature 2003; 424(6950): 797–801. 145. Kovalenko A, Chable-Bessia C, Cantarella G, Israel A, Wallach D, et al. The tumor suppressor CYLD negatively regulates NF-kappaB signalling by deubiquitination. Nature 2003; 424(6950): 801–805. 146. Trompouki E, Hatzivassiliou E, Tsichritzis T, Farmer H, Ashworth A, et al. CYLD is a deubiquitinating enzyme that negatively regulates NF-kappaB activation by TNFR family members. Nature 2003; 424(6950): 793–796. 147. Stegmeier F, Sowa ME, Nalepa G, Gygi SP, Harper JW, et al. The tumor suppressor CYLD regulates entry into mitosis. Proc Natl Acad Sci U S A 2007; 104(21): 8869–8874. 148. Wickstrom SA, Masoumi KC, Khochbin S, Fassler R, Massoumi R. CYLD negatively regulates cell-cycle progression by inactivating HDAC6 and increasing the levels of acetylated tubulin. Embo J 2010; 29(1): 131–144. 149. Gao J, Huo L, Sun X, Liu M, Li D, et al. The tumor suppressor CYLD regulates microtubule dynamics and plays a role in cell migration. J Biol Chem 2008; 283(14): 8802–8809.

34 General introduction

150. Tauriello DV, Haegebarth A, Kuper I, Edelmann MJ, Henraat M, et al. Loss of the tumor suppressor CYLD enhances Wnt/beta-catenin signaling through K63-linked ubiquitination of Dvl. Mol Cell 2010; 37(5): 607–619. 151. Jenner MW, Leone PE, Walker BA, Ross FM, Johnson DC, et al. Gene mapping and expression analysis of 16q loss of heterozygosity identifies WWOX and CYLD as being important in determining clinical outcome in multiple myeloma. Blood 2007; 110(9): 3291–3300.

3 4 5 6 7

Chapter2

Wnt signaling reaching gale force during multiple myeloma progression

Kinga A. Kocemba,1 Harmen van Andel,1 Martin F. M. de Rooij,1 Marie José Kersten,2 Marcel Spaargaren1,3 and Steven T. Pals1,3

1Departments of Pathology and 2Hematology, Academic Medical Center, University of Amsterdam, and Lymphoma and Myeloma Center Amsterdam (LYMMCARE), The Netherlands,3 These authors share last authorship

Manuscript in preparation Abstract

Aberrant activation of Wnt signaling plays a crucial role in the pathogenesis of various human cancers and is typically caused by mutations in the adenomatous polyposis gene (APC) or in the β-catenin gene (CTNNB1). As discussed in this review, multiple myelomas (MMs) often display evidence of Wnt pathway activation but do not contain these pathway intrinsic mutations. Instead, the Wnt pathway in MM is intact and activation results from auto- and/or paracrine stimulation by Wnt ligands. During tumor progression, this activation is promoted by epigenetic silencing of (soluble) negative feed-back regulators, like secreted frizzled-related proteins (sFRPs) and Dickkopf-1 (DKK1). Moreover, signaling may also be enhanced by genetic events affecting several recently identified positive- and negative Wnt-pathway regulators. Functional evidence indicates that deregulated Wnt signaling in MM plays two distinct pathogenic roles: i) aberrant activation of Wnt canonical and non-canonical Wnt pathways promotes tumor dissemination, proliferation, and drug resistance; ii) overexpression of soluble Wnt inhibitors like sFRPs and DKK1 by MMs contributes to osteolytic bone disease by inhibiting osteoblast differentiation. The ligand dependence of the aberrant Wnt signaling activity in MM implies that targeting Wnt secretion with small molecule inhibitors presents an interesting option for the treatment of MM.

38 Introduction multiple myeloma Multiple myeloma (MM) is a malignant plasma cell disorder that accounts for approximately 10% of all hematological cancers. It develops from a pre-malig- 1 nant condition termed monoclonal gammopathy of undetermined significance (MGUS). In about 50% of MGUS and MM patients, the clonal plasma cells harbor translocations involving the immunoglobulin heavy chain (IgH) locus on chromosome 14q32 and one of the following partners: 11q13 (CCND1, cyclin 3 D1 gene), 4p16.3 (FGFR-3 and MMSET gene), 6p21 (CCND3, cyclin D3 gene), 16q23 (MAFC gene) and 20q11 (MAFB gene). Most of the remaining cases 4 (IgH non-translocated MM) are associated with hyperdiploidy characterized by trisomies of chromosomes 3, 5, 7, 9, 11, 15, 19 and 21.1 Genetic abnormalities 5 including Ras mutations, p16 (CDKN2A) promotor methylation, changes involving the MYC family of oncogenes, secondary chromosomal translocations and 6 deletions, and p53 mutations have all been identified in clonal plasma cells in association with progression to the advanced stage MM.2–5 In addition, the bone 7 marrow (BM) microenvironment becomes heavily modified during disease progression and plays a crucial role in the biology of MM. This interaction between MM cells and the microenvironment is bi-directional: MM cells disrupt the homeostasis of the BM, resulting in anemia, aberrant angiogenesis, and osteolytic bone disease, while the BM microenvironment supports the growth and survival of the malignant plasma cells through signals mediated by adhesion molecules, cytokines, and growth factors.6 Major signaling routes deregulated in MM include the extracellular signal- regulated kinase (ERK), phosphatidylinositol 3-kinase (PI3K)/AKT, Janus kinase (JNK), signal transducer and activator of transcription (JAK/STAT), and NF-κB pathways. In addition, as discussed in this review, over the last decade

39 Chapter 2

evidence has accumulated indicating that deregulated Wnt signaling plays an important role in the pathogenesis of MM. This deregulation leads to aberrant expression of multiple Wnt pathway components, including agonists and antagonists, as well as illegitimate activation of canonical7 and non-canonical Wnt signaling.8–10 Although the full impact of this deregulation on the biology of MM is not yet defined, effects on tumor cell migration, proliferation and drug resistance have been reported.7,8,10,11 In addition to these direct consequences for the tumor, aberrantly expressed Wnt pathway components also strongly affect the tumor microenvironment, inhibiting osteoblast differentiation and promoting angiogenesis.6,12–14 Although loss of APC or gain of function mutations in CTNNB1 have not been found in MM, recent data suggests that genetic and epigenetic alterations affecting Wnt pathway regulators play a critical role in the aberrant activation of the pathway. the wnt pathway The term “Wnt” is derived from a combination of the names for the Drosophila melanogaster segment polarity gene Wingless and its vertebrate homolog-Inte- grase-1 (Int-1), a mouse proto-oncogene that was discovered as an integration site for mouse mammary tumor virus.15,16 There are 19 Wnt genes in the human genome, all encoding lipid-modified secreted glycoproteins, which act as ligands for cell surface receptor-mediated signal transduction pathways regulating a variety of cellular activities, including cell fate determination, proliferation, migration, and cell polarity. The lipid modification of Wnt proteins involves covalent attach- ment of palmitic acid on the first cysteine residue and palmitoleic acid on a highly conserved serine residue. Whereas the palmitoylation of Wnts is required for binding to their cognate frizzled receptors, initiating signaling, glycosylation of Wnts is required for their secretion. Studies in Drosophila and vertebrates have shown that Wnt signals can be transduced in distinct ways; by a well-defined “canonical” Wnt/β-catenin pathway, or by either of two “non-canonical” β-catenin independent pathways.17–19 canonical wnt signaling Wnt/β-catenin signaling is the best characterized Wnt signaling pathway. A major effector of the canonical Wnt signaling pathway is the transcription factor β-catenin. In the absence of Wnt proteins, β-catenin interacts with APC and Axin scaffold proteins in the cytoplasm and is a phosphorylation substrate for the kinases Casein Kinase 1 (CK1) and glycogen synthase kinase (GSK)3β. Phos- phorylated β-catenin is ubiquitinated and subsequently destructed by the pro-

40 Wnt signaling reaching gale force during multiple myeloma progression teosome. Wnt ligand binding to Frizzled (Fz) family receptors, in complex with the coreceptor LRP5/6 and Dishevelled (DVL), promotes the phosphorylation of LRP5/6 by CK1 and GSK3β (the same kinases that are involved in β-catenin phosphorylation in the destruction complex). Phosphorylation of LRP5/6 creates a docking site for AXIN1, resulting in its sequestration from the destruction complex and β-catenin stabilization. Active, non–phosphorylated β-catenin translocates to the nucleus where it binds TCF/LEF transcription factors and mediates expression of Wnt responsive genes (Figure 1).20–22

3 4 5 6 Figure 1. The canonical Wnt signaling pathway 7 Off state: In the absence of Wnt signaling free cytoplasmic β-catenin is kept at very low levels by proteo- somal degradation. β-catenin degradation is accomplished through active phosphorylation at conserved regions by glycogen synthase kinase 3β (GSK-3β) and Casein Kinase 1 (CK1). In addition, without the presence of R-spondin, ZNRF3 promotes turnover of Fz and LRP6 by acting as an ubiquitin ligase for these proteins. DKK1, sFRP, WIF1, and sclerostin are soluble Wnt cascade inhibitors, preventing Wnt signaling, even in the presence of Wnt ligands. CYLD acts as negative Wnt pathway regulator by deubiquitinating Dishevelled (DVL). In nucleus, in the absence of β-catenin, TCF/LEF occupies and represses Wnt target genes, assisted by transcriptional co-repressors such as Groucho. On state: Upon binding of a Wnt to Fz-family receptors in complex with the co-receptor LRP5/6 and DVL, phosphorylation of LRP5/6 by CK1 and GSK3β is initiated. Phosphorylation of LRP5/6 creates a docking site for AXIN1, resulting in its sequestration from the destruction complex and β-catenin stabilization. Active, non- phosphorylated β-catenin translocates to the nucleus, replaces Groucho from TCF/LEF, and recruits transcriptional co-activators to drive target gene expression. Moreover, if present, R-spondin binds to both, ZNRF3 and LGR4, promoting ZNRF3 turnover, likely through autoubiquitination, thereby stabi- lizing Fz and LRP5/6 for a greater Wnt response.

41 Chapter 2

As most other signaling cascades, the Wnt pathway contains multiple negative and positive regulatory elements, which either limit or enhance the strength and duration of the signal triggered by the initial stimulus. Thus, paracrine and/or autocrine stimulation of Wnt signaling induces expression of several negative intracellular and secreted feedback regulators, including AXIN2, DKK1, and sFRPs, resulting in signal attenuation once a certain threshold has been reached.23,24 Recent studies have revealed a crucial role of protein ubiquitination in the regulation of Wnt signaling. It was shown that the deubiquitinase (DUB) CYLD negatively regulates Wnt signaling by removing K63-linked polyubiquitin chains from Dishevelled (Dvl).25 In addition, ubiquitination and deubiquitination of Fz was found to modulate the cellular responsiveness to Wnts.26 This ubiquitination is controlled by transmembrane E3 ubiquitin ligase zinc and ring finger 3 (ZNRF3) or its homologue RNF43. ZNRF3/RNF43 ubiquitinates Fz and promotes the degradation of frizzled and LRP6, leading to attenuated canonical and non-canonical Wnt signaling. R-spondin a positive modulator of Wnt signaling was shown to bind to ZNRF3, in addition to transmembrane LGR4/5 receptors, resulting in membrane clearance of ZNRF3 thereby increasing Fz and LRP6 membrane expression levels and enhancing responses to Wnts.27,28 non-canonical wnt signaling The non-canonical Wnt signaling pathways in Drosophila and vertebrates are less well characterized. They also involve Wnt-Fz binding, but are independent of LRPs and β-catenin. Based on the major intracellular mediators used, they are designated the Wnt/Jun N-terminal kinase (Wnt/JNK) or Wnt/Ca2+ pathway. The Wnt/JNK pathway largely overlaps with the planar cell polarity pathway, originally described in Drosophila. In this pathway, receptor triggering leads to Dvl mediated activation of the small GTPase RhoA and downstream protein kinases, including JNK and Rho kinase, which affects cytoskeletal dynamics.29–32 In the Wnt/Ca2+ pathway, G-proteins, phospholipase C (PLC), and phosphodiesterase (PDE) are activated. Elevation of intracellular calcium levels activates enzymes such as Ca2+/calmodulin dependent kinase II and protein kinase C, resulting in altered cell motility. Importantly, the Wnt/Ca2+ pathway has been linked to the activation of Nemo-Like Kinase (NLK), which is able to phosphorylate TCF transcription factors and thereby inhibits canonical Wnt signalling.29,30 soluble wnt pathway inhibitors Wnt signaling is negatively regulated by a multitude of secreted proteins. These include Wnt inhibitory factor 1 (Wif1), secreted frizzled related proteins (sFRPs),

42 Wnt signaling reaching gale force during multiple myeloma progression

Dickkopfs (DKKs), and sclerostin. The sFRP family consists of five members, each containing a cysteine rich domain (CRD) with 30–50% sequence homology with the CRD of Fz receptors. Both, WIF1 and sFRPs bind to Wnt ligands thereby inhibiting Wnt pathway activity. In addition, sFRPs may also inhibit Wnt signaling by forming an inhibitory complex with Fz receptors. DKKs have a different mode of action: They act by forming a ternary complex with LRP5/6, resulting in endocytosis and thus depletion of LRP receptors from the cell surface. In this way they prevent Fz activation, even in the presence of Wnt ligands. Sclerostin, which was recently added to the list of extracellular Wnt antagonists, also binds to LRP5 and LRP6 co-receptor but with a lower affinity than DKK1.33–35

The role of Wnt signaling in multiple myeloma 1 the role of non-canonical wnt signaling in multiple myeloma The first evidence for a role of (non-canonical) Wnt signaling in MM came from a study by Qiang et al.9 These authors demonstrated that Wnt3a induces rear- 3 rangement of the actin cytoskeleton as well as striking morphological changes in myeloma cells. These responses were completely inhibited by sFRP-1, indicating 4 the involvement of Fz receptors, but the DKK1 and DKK2 proteins had no effect, implying that LRP co-receptors were not involved. The morphological changes 5 were associated with Rho activation and could be completely blocked by a Rho- associated kinase inhibitor.10 Subsequent studies from the same laboratory 6 showed that multiple members of the Wnt family (Wnt3a, Wnt1, Wnt4) promote myeloma cell migration and invasion. This Wnt-mediated migration was 7 associated with activation of RhoA and members of the protein kinase C (PKC) family, including PKCα, PKCβ, and PKCμ and involved induction of macro- molecular signaling complexes containing Dvl, RhoA, and PKCs.10 Thus, by stimulating the migration of malignant plasma cells, non-canonical Wnt signaling may promote MM cell dissemination within the bone marrow.10 An additional functional consequence of non-canonical Wnt signaling in MM involves its regulatory effect on integrin-mediated cell adhesion. As shown by Kobune et al.8 Wnt stimulation promotes myeloma cell adhesion to extracellular matrix and stromal cells, through activation of the RhoA/ROCK kinase pathway. This integrin-mediated adhesion was shown to regulate cell-adhesion mediated drug resistance (CAM-DR) to doxorubicine.

43 Chapter 2 activation and function of canonical wnt signaling in multiple myeloma Initial evidence for a role of canonical Wnt pathway activation in the pathogenesis of MM came from a study by Derksen et al.7, demonstrating that MM plasma cells, unlike normal BM plasma cells, express nuclear and non-phosphorylated β-catenin, suggesting active β-catenin/TCF mediated transcription. Stimulation of Wnt signaling with exogenous Wnt ligand (Wnt3a), LiCL, or a constitutively active mutant of β-catenin (S33Y), enhanced accumulation and nuclear localization of β-catenin and promoted proliferation. In contrast, disruption of β-catenin/ TCF activity by dominant negative TCF led to inhibition of MM cells proliferation. Importantly, no mutations in APC or β-catenin (CTNNB1) were found in MMs. These data indicate that MM cells are dependent on active Wnt signaling, involving autocrine Wnts, which was further stimulated by exogenous (paracrine) Wnt ligands.7 Consistent with these observations, analysis of gene-expression profiling data of primary MMs36 revealed co-expression of various Fzs (e.g. Fz 1, 3, 6, 7 and 8) and the co-receptor LRP6, as well as various Wnts (including 4, 5A, 5B, 6, 10A and 16). Approximately half of the primary MMs co-expressed LRP6 with at least one Fz genes and one Wnt gene, indicating that these MMs are well equipped to evoke autocrine Wnt pathway activation. Furthermore, Wnt ligands are produced by bone marrow stromal cells,7,37 providing a source of paracrine Wnt pathway activation. These findings were supported by Sukhdeo et al.,38 who also observed massive upregulation of multiple Wnt signaling pathway genes in primary MM cells. In line with the study of Derksen,7 active, non-phosphorylated β-catenin was found in the nucleus of the cells of the majority of MM cell lines. Blockage of β-catenin/TCF mediated transcription, by using the β-catenin/TCF complex inhibitor PKF115-584, resulted in downregulation of Wnt target genes as well as cell cycle arrest, apoptosis, and activation of apoptotic regulators.38 These findings were supported by in vivo data, showing that MM growth in SCID mice was effectively inhibited by the compound.38 Subsequent studies from the same laboratory demonstrated that Wnt/β-catenin pathway does not only affect the G1 phase of the cell cycle but also G2/M transition.11 By targeting β-catenin with shR- NAs, new potential Wnt target genes involved in cell cycle progression and check- point regulation were identified. Among these genes, AURKA/B was shown to act as a key regulator of β-catenin-mediated effects on cell cycle and MM growth, suggesting an important role for this protein in the Wnt-mediated pathogenic effects. Importantly, targeting β-catenin protein expression caused significant tumor reduction and increased survival in a xenograft mouse model of MM.11 The finding that silencing of β-catenin by siRNAs inhibits of MM tumor growth in vivo

44 Wnt signaling reaching gale force during multiple myeloma progression was confirmed by Ashihara et al.39 In the 5TGM1 mouse myeloma model, Ed- wards et al.40 found that Wnt pathway activation by LiCl induces accumulation of β-catenin and enhanced Wnt target gene expression, but does not alter the in vitro proliferation rate or the Wnt-pathway dependent tumor expansion in the BM microenvironment. However, LiCl enhanced the growth of subcutaneously inoculat- ed 5TGM1 cells, which was prevented by overexpression of a dominant-negative TCF4, confirming the Wnt signaling dependency.40 A recent study by Bjorklund et al.41 indicates a role for Wnt/β-catenin signaling in drug resistance. Exposure of malignant plasma cells to lenalidomide enhanced β-catenin/TCF mediated transcription and expression of Wnt target genes. This lenalidomide-mediated Wnt pathway activation, or activation by Wnt3a or β-catenin, reduced the anti-proliferative effect of lenalidomide. These effects were reversed by shRNA-mediated down-regulation of β-catenin, suggesting that targeting Wnt/β-catenin signaling may help to overcome lenalidomide resistance in MM. In line with the results discussed above Qiang et al.42 reported that Wnt3a, either alone or in combination 1 with IL-6 and insulin-like growth factor (IGF)-1 induces β-catenin stabilization and Wnt reporter activity. However, these authors did not find an enhancing effect on MM proliferation but reported that Wnt3a overexpression attenuates bone disease and tumor growth of a human MM cell line transplanted in human bone 3 implants in SCID mice.42 In another study,43 GSK3β inhibition by 6-bromoin- dirubin-3-oxime (BIO) led to improved bone quality at the bone-tumor interface 4 as well as to increased tumor necrosis in a murine model for MM-bone disease, while a pro-apoptotic effect of BIO was found in human MM cell lines in vitro. 5 Taken together, although a few discrepant reports exist, which might be explained by the difference in experimental models or distinct levels of Wnt path- 6 way activation,44 most of the above studies support a scenario in which aberrant canonical and non-canonical Wnt pathway activation acts as an important factor 7 enhancing MM aggressiveness by promoting cell motility, cell cycle progression, and drug-resistance. This Wnt pathway activation depends on autocrine and/or paracrine Wnts and not on mutations in the APC or CTNNB1, causing ligand- independent activation. However, as will be discussed below, loss of negative regulation of Wnt signaling is common in MMs and appears to be an important factor in Wnt pathway activation during disease progression. epigenetic and genetic events affecting wnt signaling in multiple myeloma During MM progression, negative regulators of Wnt signaling are common target of epigenetic silencing. Thus, Chim et al.45 reported silencing by promoter

45 Chapter 2

hypermethylation of the Wnt antagonists WIF1, DKK3, APC as well as sFRP1, -2, -4 and -5 in MM cell lines45 These cell lines displayed active Wnt signaling, which was suppressed by treatment with the demethylating agent by 5-azadC, re- expression of these antagonists. Of primary MM samples, 40% showed methylation of one or more of these seven genes, indicating that methylation of soluble Wnt antagonists is common in MM.45 These data suggest that Wnt inhibitors may act as tumor suppressor genes that need to be inactivated in order to reach optimal levels of Wnt pathway activation during MM progression. Indeed, a study by Jost et al.46 revealed that whereas hypermethylation of sFRP1 and -2 genes is already present in MGUS and remained present at all subsequent MM stages, sFRP5 methylation was restricted to advanced stages of MM and plasma cell leukaemia (PCL).46 Intriguingly, these studies did not explore the methylation status of DKK1, a major Wnt antagonist secreted by MM cells, which contributes to MM bone disease by inhibiting osteoblast differentiation (discussed below). We recently demonstrated36 that DKK1 is also a target of methylation and that DKK1 methylation is largely limited to advanced stage MM. Hence, during the initial stages of MM evolution overexpression and secretion of DKK1 by malignant plasma cells modifies the bone marrow niche, inhibiting Wnt-signaling dependent osteoblast differentiation and creating an optimal environment for MM growth and progression. In advanced stage MM, however, silencing of DKK1 unleashes the Wnt pathway in MM cells, promoting proliferation of MM and drug resistance. Taken together, these data suggest that aberrant Wnt pathway activation in MM is the consequence of “releasing the brake” rather than of “hitting the gas”. This hypothesis is corroborated by recent studies from our laboratory, identifying CYLD as an important negative regulator of Wnt signaling in MM. (manuscript in preparation). CYLD was originally identified as a tumor suppressor gene mutated in familial cylindromatosis (Brooke-Spiegler syndrome), an autosomal dominant disorder predisposing to benign tumors of skin appendages.47 Subsequent studies have linked loss of the tumor suppressor function of CYLD to the pathogenesis of several other tumors including melanoma, T-cell acute lymphoblastic lymphoma (T-ALL), and colon and hepatocellular carcinoma.48–51 In MM, loss of CYLD, resulting from biallelic deletion or inactivating mutations, is among the most common genomic aberrations.52,53 The CYLD protein is a member of the USP family of deubiquitinating enzymes (DUBs), which act by specifically removing lysine(K)-63-linked polyubiquitin chains from substrate proteins.54 In contrast to lysine-48-linked polyubiquitination marking proteins for proteasomal degradation, K63-linked ubiquitination enhances protein stability and facilitates

46 Wnt signaling reaching gale force during multiple myeloma progression protein-protein interaction. A number of studies have shown that CYLD acts as a negative regulator of nuclear factor-ĸB (NF-ĸB) signaling removing K63- linked polyubiquitin chains from TRAF2, TRAF6, and NEMO.55,56 Other CYLD substrates important for NF-ĸB signaling include RIPK1, BCL3 and TAK1.57–59 In addition, it was recently reported that CYLD may also acts as a negative regulator of proximal events in Wnt/β-catenin signaling. Loss of CYLD causes hyperubiquitination of the DIX domain of the adapter protein Dishevelled (Dvl), leading to enhanced Dvl polymerization and Wnt signaling.25 Indeed, human cylindroma skin tumors that arise from mutations in CYLD were found to display hyperactive Wnt signaling, suggesting that the tumor growth instigated by loss of CYLD involves enhancement of Wnt responses.25 We recently observed that CYLD expression in human myeloma cell lines as well as primary MMs is highly variable and that the protein is functionally involved in the regulation of MM cell growth and survival. In MM patients, low CYLD expression is associated with poor progression free survival and overall survival. Functional assays employ- 1 ing inducible CYLD silencing or overexpression revealed that CYLD represses autocrine as well as ligand-induced Wnt/β-catenin signaling, and that low CYLD expression in primary MMs is strongly associated with the presence of a Wnt signaling gene-expression signature. These findings identify CYLD as a regulator 3 of Wnt/ β-catenin signaling in MM and suggest that loss of CYLD enhances MM aggressiveness through a mechanism involving Wnt pathway activation resulting 4 in enhanced proliferation. (manuscript in preparation) In conclusion, loss of negative regulators of Wnt signaling by epigenetic and 5 genetic events causes hyperactivation of the Wnt pathway in advanced MM, contributing to enhanced cell proliferation and drug resistance. 6 7 The role of Wnt pathway in the pathogenesis of multiple myeloma bone disease wnt inhibitors and multiple myeloma bone disease During the last decade, it has become clear that osteoblast differentiation is critically dependent on Wnt signaling.60 Interestingly, Tian et al.61 showed that most primary MMs overexpress and secrete the Wnt inhibitor DKK1 and that this overexpression was strongly correlated to the presence of osteolytic bone disease. Similar data were reported by Oshima et al.62 for the secreted Wnt signaling inhibitor sFRP2: MM cells from patients with advanced bone lesions secreted

47 Chapter 2

sFRP2 and conditioned media of these cells suppressed osteoblast mineralization and alkaline phosphatase activity. In addition, MM cells can also secrete the FRZB/sFRP3, which was shown to be upregulated at the transition of MGUS to multiple myeloma,63,64 and correlated with the presence of osteolytic lesions.12 More recently, studies by Brunetti et al.,65 and Colluci et al.66 reported overexpression of sclerostin, an osteocyte-specific negative regulator of Wnt signaling, by human myeloma cell lines and primary MM cells. Taken together, the data indicate an important role for secreted Wnt inhibitors in MM bone disease, inhibiting Wnt signaling in osteoblasts and thereby interfering with osteoblast differentiation and inducing apoptosis. Importantly, in addition to causing bone disease, inhibition of osteoblast differentiation may also promote MM growth since immature osteoblasts express high levels of IL-6, a central growth and survival factor for myeloma plasma cells (Figure 2).68 In parallel, inhibition of Wnt signaling in osteoblasts enhances the expression of receptor activator of NF-ĸB ligand (RANKL), downregulating the expression of osteoprotegerin (OPG) in immature osteoblast.67 The resulting increased RANKL/OPG ratio leads to osteoclast activation, hence promoting osteolytic bone disease.

Figure 2. Model of MM-bone disease Myeloma cells produce factors that directly or indirectly activate osteoclasts, such as macrophage inflammatory protein-1a (MIP-1a), IL-3 and HGF. The bone destructive process releases factors that increase MM growth, thereby aggravating the osteolytic process. In addition, osteoclasts also support the growth of myeloma cells through secretion of IL-6, APRIL and osteopontin, and by adhesive interactions, stimulating the proliferation of malignant plasma cells. Moreover, myeloma cells produce DKK1, sFRP2, FRZB, HGF, IL-7 and sclerostin (SOST), which suppress differentiation of mesenchymal cells into osteoblast and formation of new bone. In turn, the mesenchymal cells are the source of IL-6, VEGF, IGF-1, HGF, TNF-a, bFGF, IL-6 and Wnts, which support the growth of malignant plasma cells in the bone marrow. The Wnt inhibitors secreted by malignant plasma cells also influence the autocrine Wnt pathway in MM cells, inhibiting the progression of the disease.

48 Wnt signaling reaching gale force during multiple myeloma progression dkk1 inhibition in multiple myeloma in vivo models Based on the early focus on DKK1 as mediator of osteolytic bone disease,61 a number of studies have explored the effect of targeting DKK1. In a study by Yaccoby et al.70 SCID-rab mice were engrafted with BM from MM patients expressing different levels of DKK1 and treated with control or DKK1-neutralizing antibodies. Compared to the controls, mice treated with anti-DKK1 antibody showed increased bone mineral density (BMD), which was associated with reduced tumor burden. The bones of anti-DKK1-treated animals contained increased numbers of osteocalcin-expressing osteoblasts and diminished number of multinucleated tartrate-resistant acidic phosphatase (TRAP)-expressing osteoclasts. Of note, anti-DKK1 treatment also significantly increased bone-mineral density of im- planted normal bone, suggesting that DKK1 is an important physiological regulator of bone remodeling. In the 5T2MM murine MM model, Heath et al.71 also found similar stimulatory effects of anti-DKK1 antibody treatment on osteoblasts and bone mineralization. However, in contrast to the study by Yaccoby et al.,70 1 no effect on osteoclast numbers or tumor burden were found. An anabolic effect of DKK1 inhibition on the bone was also reported by Fulciniti et al.,72 employing a SCID-hu MM model. In this model, anti-DKK1 treatment significantly increased osteoblast numbers, serum osteocalcin level, and trabecular bone, and 3 inhibited MM cell growth. This growth inhibition was only observed for MM growth in the BM microenvironment but not for subcutaneous tumor growth, in- 4 dicating that anti-DKK1 treatment does not directly target MM cells but depends on the stromal component. Possible mechanisms include reduced IL-6 expression 5 by BM stromal cells and/or inhibition of adhesion between stromal and MM cells. Taken together these studies indicate that anti-DKK1 antibody treatment can 6 inhibit MM-bone disease in animal models, suggesting that it presents a valuable therapeutic asset for patients suffering from MM bone disease. In addition, in 7 several models, anti-DKK1 treatment attenuates tumor growth within the bone marrow environment, presumably by indirect effects on BMSCs. How DKK1 expression is activated in MM is currently incompletely understood. Gene-expression profiling-defined classification of MMs has revealed that DKK1 expression and osteolysis is significantly elevated in hyperdiploid disease with char- acteristics of microenvironment dependence.73 These data suggest that DKK1 activation is an essential event in disease progression, especially in forms of the disease lacking oncogene-activating translocations. In a study by Colla et al.74 evidence was provided that oxidative stress- mediated JNK activation may partly drive this DKK1 overexpression in MM. It should be noted, however, that DKK1 itself presents a prominent transcriptional target of Wnt signaling, that acts as

49 Chapter 2 negative feedback-regulator in many different cell types.24,75,76 Recent studies from our laboratory indicate that DKK1 can also be a transcriptional target of Wnt signaling in MM cells (unpublished observation). Hence, inhibition of DKK1 activity in the BM environment could potentially lead to hyperactivation of Wnt signaling in MM plasma cells, stimulating MM proliferation and dissemination. Although this scenario is not supported by a study of Dun et al.,77 which suggests that targeting DKK1 exclusively influences the Wnt pathway in BM stromal cells but not in the MM cells themselves, two independent studies, by Kocemba at al.36 and Kobune et al.,8 revealed that DKK1 can instead directly inhibit Wnt signaling activity in MM plasma cells. Furthermore, DKK1 expression is often lost in advanced stage, clinically aggressive MMs and in MM cell lines.61 This is caused by promotor methylation and correlates with enhanced Wnt signaling in the MM plasma cells, suggesting that DKK1 loss may unleash Wnt signaling.36 As discussed in a previous paragraph this may not only promote proliferation of MM plasma cells but also enhance their metastatic potential.11 Furthermore, an effect on lenalidomide resistance was noted.41 Thus, in MM, DKK1 does not only promote bone disease but also appears to act as a tumor suppressor. Therefore, blocking of DKK1 activity, although inhibiting its osteolytic activity disease, may lead to MM cells dissemination, confer drug resistance and promote the tumor growth, especially in the extramedulary location.40 Consistent with this notion, MM cell lines generally express very low levels of DKK1.36 These cell lines are derived from MM subgroups with oncogenic translocations involving the immunoglobulin heavy chain (IgH) locus on chromosome 14q32 and a partner oncogene, but never from hyperdiploid MMs as these remain fully dependent on the bone marrow environment.78 The low level of DKK1 in BM-independent MM cells further suggests that downregulation of DKK1, by unleashing the Wnt signaling, can contribute to the aggressiveness and progression of MM disease.

Conclusions and therapeutic perspectives

Aberrant activation of Wnt signaling plays a crucial role in the pathogenesis of various human cancers. As discussed in this review, multiple myelomas (MMs) often display evidence of Wnt pathway activation but do not contain pathway intrinsic mutations in APC or CTNNB1. Instead, Wnt pathway activation results from auto- and/or paracrine stimulation by Wnt ligands and is promoted by epigenetic silencing of (soluble) negative feed-back regulators and

50 Wnt signaling reaching gale force during multiple myeloma progression by genetic events affecting positive and negative Wnt-pathway regulators. De- regulated Wnt signaling in MMs promotes tumor growth, dissemination, and drug resistance. On the other hand, overexpression of soluble Wnt (feed-back) inhibitors by MMs contributes to osteolytic bone disease by inhibiting osteoblast differentiation. The complex structure of the Wnt pathway makes it, in prin- ciple, amenable to therapeutic intervention at multiple distinct levels, ranging from inhibition of receptor-ligand interaction at the cell surface to interference with TCF/β-catenin complex formation and blocking of target gene expression and/or function (Figure 3).

3 4 5 6 7

Figure 3. Wnt pathway inhibitors in I/II clinical trials for cancer OMP-18R5 – a monoclonal antibody which targets the Fz receptors to block association with Wnt ligands (NCT01345201).

51 Chapter 2

OMP-54F28 – a fusion protein of the Fz8 ligand-binding domain with the Fc region of a human immunoglobulin. It binds and sequesters soluble Wnt ligands, impairing their interaction with receptors in tissues. (NCT01608867). PRI-724 – a small-molecule inhibitor of the interaction between β-catenin and creb-binding protein (CBP). Disrupting this interaction prevents activation of transcription by Wnt signaling. (NCT01606579, NCT01302405). LGK974 – a small-molecule inhibitor of the acyltransferase porcupine. Enzymatic activity of acyltransferase porcupine is crucial in the secretion of Wnt ligands. (NCT01351103).

From a therapeutic perspective the fact that the Wnt signaling pathway in MM is essentially intact and that it activation is strictly ligand depend is of great interest. It implies that targeting Wnt secretion with small molecule inhibitors that interfere with the function of the acyltransferase porcupine presents an interesting option for the treatment of MM. Interestingly, LGK974 a small molecule porcupine inhibitor is currently in clinical trials for cystic pancreatic cancer; like MM this tumor depends on activation by Wnt ligands. Furthermore, modulat- ing negative regulators of Wnt signaling, e.g. by employing demethylating agents, presents a therapeutic option. In employing Wnt-targeted treatments it will be important to define the therapeutic window to achieve a maximal anti-MM effects with a minimal impact on osteoblasts differentiation. If Wnt-targeting compounds prove safe and effective, their therapeutic application in MM may represent a milestone in the treatment of the disease.

References

1. Rajkumar SV. Multiple myeloma. Current Problems in Cancer 2009; 33(1): 7–64. 2. Chng WJ, Glebov O, Bergsagel PL, Kuehl WM. Genetic events in the pathogenesis of multiple myeloma. Best Practice & Research 2007; 20(4): 571–596. 3. Kyle RA, Rajkumar SV. Monoclonal gammopathy of undetermined significance and smoldering multiple myeloma. Hematology/Oncology Clinics of North America 2007; 21(6): 1093–1113, ix. 4. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009; 374(9686): 324–339. 5. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nature Reviews 2012; 12(5): 335–348. 6. Edwards CM, Zhuang J, Mundy GR. The pathogenesis of the bone disease of multiple myeloma. Bone 2008; 42(6): 1007–1013. 7. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proceedings of the Na- tional Academy of Sciences of the United States of America 2004; 101(16): 6122–6127.

52 Wnt signaling reaching gale force during multiple myeloma progression

8. Kobune M, Chiba H, Kato J, Kato K, Nakamura K, et al. Wnt3/RhoA/ROCK signaling pathway is involved in adhesion-mediated drug resistance of multiple myeloma in an autocrine mechanism. Molecular Cancer Therapeutics 2007; 6(6): 1774–1784. 9. Qiang YW, Endo Y, Rubin JS, Rudikoff S. Wnt signaling in B-cell neoplasia. Oncogene 2003; 22(10): 1536–1545. 10. Qiang YW, Walsh K, Yao L, Kedei N, Blumberg PM, et al. Wnts induce migration and invasion of myeloma plasma cells. Blood 2005; 106(5): 1786–1793. 11. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/beta-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 12. Stewart JP, Shaughnessy JD, Jr. Role of osteoblast suppression in multiple myeloma. Journal of Cellular Biochemistry 2006; 98(1): 1–13. 13. Goodwin AM, D’Amore PA. Wnt signaling in the vasculature. Angiogenesis 2002; 5(1–2): 1–9. 14. Zhang B, Ma JX. Wnt pathway antagonists and angiogenesis. Protein & Cell 2010; 1(10): 898–906. 1 15. Nusse R, Varmus HE. Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Cell 1982; 31(1): 99–109. 16. Sharma RP, Chopra VL. Effect of the Wingless (wg1) mutation on wing and haltere de- 3 velopment in Drosophila melanogaster. Developmental Biology 1976; 48(2): 461–465. 17. Moon RT, Brown JD, Torres M. WNTs modulate cell fate and behavior during verte- 4 brate development. Trends Genet 1997; 13(4): 157–162. 18. Reichsman F, Smith L, Cumberledge S. Glycosaminoglycans can modulate extracellular localization of the wingless protein and promote signal transduction. The Journal 5 of Cell Biology 1996; 135(3): 819–827. 19. Willert K, Brown JD, Danenberg E, Duncan AW, Weissman IL, et al. Wnt proteins 6 are lipid-modified and can act as stem cell growth factors. Nature 2003; 423(6938): 448–452. 7 20. Macdonald BT, Semenov MV, He X. SnapShot: Wnt/beta-catenin signaling. Cell 2007; 131(6): 1204. 21. van de Wetering M, Cavallo R, Dooijes D, van Beest M, van Es J, et al. Armadillo coactivates transcription driven by the product of the Drosophila segment polarity gene dTCF. Cell 1997; 88(6): 789–799. 22. van Noort M, Meeldijk J, van der Zee R, Destree O, Clevers H. Wnt signaling controls the phosphorylation status of beta-catenin. The Journal of Biological Chemistry 2002; 277(20): 17901–17905. 23. Jho EH, Zhang T, Domon C, Joo CK, Freund JN, et al. Wnt/beta-catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of the signaling pathway. Molecular and Cellular Biology 2002; 22(4): 1172–1183.

53 Chapter 2

24. Niida A, Hiroko T, Kasai M, Furukawa Y, Nakamura Y, et al. DKK1, a negative regulator of Wnt signaling, is a target of the beta-catenin/TCF pathway. Oncogene 2004; 23(52): 8520–8526. 25. Tauriello DV, Haegebarth A, Kuper I, Edelmann MJ, Henraat M, et al. Loss of the tumor suppressor CYLD enhances Wnt/beta-catenin signaling through K63-linked ubiquitination of Dvl. Molecular Cell 2010; 37(5): 607–619. 26. Mukai A, Yamamoto-Hino M, Awano W, Watanabe W, Komada M, et al. Balanced ubiquitylation and deubiquitylation of Frizzled regulate cellular responsiveness to Wg/Wnt. The EMBO Journal 2010; 29(13): 2114–2125. 27. Hao HX, Xie Y, Zhang Y, Charlat O, Oster E, et al. ZNRF3 promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 2012; 485(7397): 195–200. 28. Koo BK, Spit M, Jordens I, Low TY, Stange DE, et al. Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 2012; 488(7413): 665–669. 29. Kuhl M, Sheldahl LC, Malbon CC, Moon RT. Ca(2+)/calmodulin-dependent protein kinase II is stimulated by Wnt and Frizzled homologs and promotes ventral cell fates in Xenopus. The Journal of Biological Chemistry 2000; 275(17): 12701–12711. 30. Kuhl M, Sheldahl LC, Park M, Miller JR, Moon RT. The Wnt/Ca2+ pathway: a new vertebrate Wnt signaling pathway takes shape. Trends Genet 2000; 16(7): 279–283. 31. Segalen M, Bellaiche Y. Cell division orientation and planar cell polarity pathways. Seminars in Cell & Developmental Biology 2009; 20(8): 972–977. 32. Wansleeben C, Meijlink F. The planar cell polarity pathway in vertebrate development. Dev Dyn 2011; 240(3): 616–626. 33. Kawano Y, Kypta R. Secreted antagonists of the Wnt signalling pathway. Journal of Cell Science 2003; 116(Pt 13): 2627–2634. 34. Brunkow ME, Gardner JC, Van Ness J, Paeper BW, Kovacevich BR, et al. Bone dys- plasia sclerosteosis results from loss of the SOST gene product, a novel cystine knot- containing protein. American Journal of Human Genetics 2001; 68(3): 577–589. 35. van Bezooijen RL, ten Dijke P, Papapoulos SE, Lowik CW. SOST/sclerostin, an osteocyte-derived negative regulator of bone formation. Cytokine & Growth Factor Reviews 2005; 16(3): 319–327. 36. Kocemba KA, Groen RW, van Andel H, Kersten MJ, Mahtouk K, et al. Transcriptional silencing of the Wnt-antagonist DKK1 by promoter methylation is associated with enhanced Wnt signaling in advanced multiple myeloma. PloS One 2012; 7(2): e30359. 37. Van Den Berg DJ, Sharma AK, Bruno E, Hoffman R. Role of members of the Wnt gene family in human hematopoiesis. Blood 1998; 92(9): 3189–3202. 38. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/ TCF transcriptional complex in the treatment of multiple myeloma. Proceedings of the National Academy of Sciences of the United States of America 2007; 104(18): 7516–7521.

54 Wnt signaling reaching gale force during multiple myeloma progression

39. Ashihara E, Kawata E, Nakagawa Y, Shimazaski C, Kuroda J, et al. beta-catenin small interfering RNA successfully suppressed progression of multiple myeloma in a mouse model. Clin Cancer Res 2009; 15(8): 2731–2738. 40. Edwards CM, Edwards JR, Lwin ST, Esparza J, Oyajobi BO, et al. Increasing Wnt signaling in the bone marrow microenvironment inhibits the development of myeloma bone disease and reduces tumor burden in bone in vivo. Blood 2008; 111(5): 2833–2842. 41. Bjorklund CC, Ma W, Wang ZQ, Davis RE, Kuhn DJ, et al. Evidence of a role for activation of Wnt/beta-catenin signaling in the resistance of plasma cells to lenalidomide. The Journal of Biological Chemistry 2011; 286(13): 11009–11020. 42. Qiang YW, Shaughnessy JD, Jr., Yaccoby S. Wnt3a signaling within bone inhibits multiple myeloma bone disease and tumor growth. Blood 2008; 112(2): 374–382. 43. Gunn WG, Krause U, Lee N, Gregory CA. Pharmaceutical inhibition of glycogen synthetase kinase-3beta reduces multiple myeloma-induced bone disease in a novel murine plasmacytoma xenograft model. Blood 2011; 117(5): 1641–1651. 44. Luis TC, Naber BA, Roozen PP, Brugman MH, de Haas EF, et al. Canonical wnt sign- 1 aling regulates hematopoiesis in a dosage-dependent fashion. Cell Stem Cell 2011; 9(4): 345–356. 45. Chim CS, Pang R, Fung TK, Choi CL, Liang R. Epigenetic dysregulation of Wnt signaling pathway in multiple myeloma. Leukemia 2007; 21(12): 2527–2536. 3 46. Jost E, Gezer D, Wilop S, Suzuki H, Herman JG, et al. Epigenetic dysregulation of secreted Frizzled-related proteins in multiple myeloma. Cancer Letters 2009; 281(1): 4 24–31. 47. Bignell GR, Warren W, Seal S, Takahashi M, Rapley E, et al. Identification of the familial cylindromatosis tumour-suppressor gene. Nature Genetics 2000; 25(2): 160–165. 5 48. Hellerbrand C, Bumes E, Bataille F, Dietmaier W, Massoumi R, et al. Reduced expression of CYLD in human colon and hepatocellular carcinomas. Carcinogenesis 2007; 6 28(1): 21–27. 49. Ke H, Augustine CK, Gandham VD, Jin JY, Tyler DS, et al. CYLD Inhibits Melanoma 7 Growth and Progression through Suppression of the JNK/AP-1 and beta1-Integrin Signaling Pathways. The Journal of Investigative Dermatology 2013; 133(1): 221–229. 50. D’Altri T, Gonzalez J, Aifantis I, Espinosa L, Bigas A. Hes1 expression and CYLD repression are essential events downstream of Notch1 in T-cell leukemia. Cell Cycle 2011; 10(7): 1031–1036. 51. Espinosa L, Cathelin S, D’Altri T, Trimarchi T, Statnikov A, et al. The Notch/Hes1 pathway sustains NF-kappaB activation through CYLD repression in T cell leukemia. Cancer Cell 2010; 18(3): 268–281. 52. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12(2): 115–130.

55 Chapter 2

53. Keats JJ, Fonseca R, Chesi M, Schop R, Baker A, et al. Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell 2007; 12(2): 131–144. 54. Sun SC. CYLD: a tumor suppressor deubiquitinase regulating NF-kappaB activation and diverse biological processes. Cell Death and Differentiation 2010; 17(1): 25–34. 55. Brummelkamp TR, Nijman SM, Dirac AM, Bernards R. Loss of the cylindromatosis tumour suppressor inhibits apoptosis by activating NF-kappaB. Nature 2003; 424(6950): 797–801. 56. Kovalenko A, Chable-Bessia C, Cantarella G, Israel A, Wallach D, et al. The tumour suppressor CYLD negatively regulates NF-kappaB signalling by deubiquitination. Nature 2003; 424(6950): 801–805. 57. Massoumi R, Chmielarska K, Hennecke K, Pfeifer A, Fassler R. Cyld inhibits tumor cell proliferation by blocking Bcl-3-dependent NF-kappaB signaling. Cell 2006; 125(4): 665–677. 58. Reiley WW, Jin W, Lee AJ, Wright A, Wu X, et al. Deubiquitinating enzyme CYLD negatively regulates the ubiquitin-dependent kinase Tak1 and prevents abnormal T cell responses. The Journal of Experimental Medicine 2007; 204(6): 1475–1485. 59. Wright A, Reiley WW, Chang M, Jin W, Lee AJ, et al. Regulation of early wave of germ cell apoptosis and spermatogenesis by deubiquitinating enzyme CYLD. Developmen- tal Cell 2007; 13(5): 705–716. 60. Monroe DG, McGee-Lawrence ME, Oursler MJ, Westendorf JJ. Update on Wnt signaling in bone cell biology and bone disease. Gene 2012; 492(1): 1–18. 61. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. The New England Journal of Medicine 2003; 349(26): 2483–2494. 62. Oshima T, Abe M, Asano J, Hara T, Kitazoe K, et al. Myeloma cells suppress bone formation by secreting a soluble Wnt inhibitor, sFRP-2. Blood 2005; 106(9): 3160–3165. 63. Davies FE, Dring AM, Li C, Rawstron AC, Shammas MA, et al. Insights into the multistep transformation of MGUS to myeloma using microarray expression analysis. Blood 2003; 102(13): 4504–4511. 64. De Vos J, Couderc G, Tarte K, Jourdan M, Requirand G, et al. Identifying intercellular signaling genes expressed in malignant plasma cells by using complementary DNA arrays. Blood 2001; 98(3): 771–780. 65. Brunetti G, Oranger A, Mori G, Specchia G, Rinaldi E, et al. Sclerostin is overexpressed by plasma cells from multiple myeloma patients. Annals of the New York Academy of Sciences 2011; 1237: 19–23. 66. Colucci S, Brunetti G, Oranger A, Mori G, Sardone F, et al. Myeloma cells suppress osteoblasts through sclerostin secretion. Blood Cancer Journal 2011; 1(6): e27.

56 Wnt signaling reaching gale force during multiple myeloma progression

67. Qiang YW, Chen Y, Stephens O, Brown N, Chen B, et al. Myeloma-derived Dickkopf-1 disrupts Wnt-regulated osteoprotegerin and RANKL production by osteoblasts: a potential mechanism underlying osteolytic bone lesions in multiple myeloma. Blood 2008; 112(1): 196–207. 68. Klein B, Zhang XG, Jourdan M, Content J, Houssiau F, et al. Paracrine rather than autocrine regulation of myeloma-cell growth and differentiation by interleukin-6. Blood 1989; 73(2): 517–526. 69. Abe M, Hiura K, Wilde J, Shioyasono A, Moriyama K, et al. Osteoclasts enhance myeloma cell growth and survival via cell-cell contact: a vicious cycle between bone destruction and myeloma expansion. Blood 2004; 104(8): 2484–2491. 70. Yaccoby S, Ling W, Zhan F, Walker R, Barlogie B, et al. Antibody-based inhibition of DKK1 suppresses tumor-induced bone resorption and multiple myeloma growth in vivo. Blood 2007; 109(5): 2106–2111. 71. Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy JD, Jr., et al. Inhibiting Dickkopf-1 (Dkk1) removes suppression of bone formation and prevents the development of osteolytic bone disease in multiple myeloma. J Bone Miner Res 2009; 1 24(3): 425–436. 72. Fulciniti M, Tassone P, Hideshima T, Vallet S, Nanjappa P, et al. Anti-DKK1 mAb (BHQ880) as a potential therapeutic agent for multiple myeloma. Blood 2009; 114(2): 371–379. 3 73. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108(6): 2020–2028. 4 74. Colla S, Tagliaferri S, Morandi F, Lunghi P, Donofrio G, et al. The new tumor-suppressor gene inhibitor of growth family member 4 (ING4) regulates the production of proangiogenic molecules by myeloma cells and suppresses hypoxia-inducible factor-1 5 alpha (HIF-1alpha) activity: involvement in myeloma-induced angiogenesis. Blood 2007; 110(13): 4464–4475. 6 75. Gonzalez-Sancho JM, Aguilera O, Garcia JM, Pendas-Franco N, Pena C, et al. The Wnt antagonist DICKKOPF-1 gene is a downstream target of beta-catenin/TCF and 7 is downregulated in human colon cancer. Oncogene 2005; 24(6): 1098–1103. 76. Chamorro MN, Schwartz DR, Vonica A, Brivanlou AH, Cho KR, et al. FGF-20 and DKK1 are transcriptional targets of beta-catenin and FGF-20 is implicated in cancer and development. The EMBO Journal 2005; 24(1): 73–84. 77. Dun X, Jiang H, Zou J, Shi J, Zhou L, et al. Differential expression of DKK-1 binding receptors on stromal cells and myeloma cells results in their distinct response to secreted DKK-1 in myeloma. Molecular Cancer 2010; 9: 247. 78. Robbiani DF, Chesi M, Bergsagel PL. Bone lesions in molecular subtypes of multiple myeloma. The New England Journal of Medicine 2004; 351(2): 197–198.

Chapter3

Transcriptional silencing of the Wnt-antagonist DKK1 by promoter methylation is associated with enhanced Wnt signaling in advanced multiple myeloma

Kinga A. Kocemba1,3, Richard W. J. Groen1,3. Harmen van Andel1, Marie Jose Kersten2, Karène Mahtouk1, Marcel Spaargaren1,4, Steven T. Pals1,4

1Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands,2 Department of Hematology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands, 3These authors contributed equally to this work, 4These authors also contributed equally to this work

PLoS One 2012; 7(2): e30359 Abstract

The Wnt/β-catenin pathway plays a crucial role in the pathogenesis of various human cancers. In multiple myeloma (MM), aberrant auto-and/or paracrine activation of canonical Wnt signaling promotes proliferation and dissemination, while overexpression of the Wnt inhibitor Dickkopf1 (DKK1) by MM cells contributes to osteolytic bone disease by inhibiting osteoblast differentiation. Since DKK1 itself is a target of TCF/β-catenin mediated transcription, these findings suggest that DKK1 is part of a negative feedback loop in MM and may act as a tumor suppressor. In line with this hypothesis, we show here that DKK1 expression is low or undetectable in a subset of patients with advanced MM as well as in MM cell lines. This absence of DKK1 is correlated with enhanced Wnt pathway activation, evidenced by nuclear accumulation of β-catenin, which in turn can be antagonized by restoring DKK1 expression. Analysis of the DKK1 promoter revealed CpG island methylation in several MM cell lines as well as in MM cells from patients with advanced MM. Moreover, demethylation of the DKK1promoter restores DKK1 expression, which results in inhibition of β-catenin/TCF-mediated gene transcription in MM lines. Taken together, our data identify aberrant methylation of the DKK1 promoter as a cause of DKK1 silencing in advanced stage MM, which may play an important role in the progression of MM by unleashing Wnt signaling.

60 Introduction

Multiple myeloma (MM), one of the most common hematological malignancies in adults, is characterized by a clonal expansion of malignant plasma cells in the bone marrow, associated with suppression of normal hematopoiesis, renal failure, 1 and osteolytic bone lesions.1,2 These bone lesions have been shown to be the result of uncoupled or imbalanced bone remodeling with decreased formation and increased resorption of bone tissue, due to impaired osteoblast differentiation and aberrant osteoclast activation.3 Recent studies have identified canonical Wnt 3 signaling as a key signal pathway in both normal bone homeostasis and in the pathogenesis of MM bone disease.4–6 4 The canonical Wnt/β-catenin signaling pathway plays a central role in modu- lating the delicate balance between stemness and differentiation in several adult 5 stem cell niches, including the intestinal crypt and the hematopoietic stem cell niche in the bone marrow.7–9 Wnt genes encode a family of 19 secreted glyco- 6 proteins, which promiscuously interact with several Frizzled (FRZ) receptors and the low-density lipoprotein receptor-related protein 5/6 (LRP5/6). The key 7 event in the Wnt signaling pathway is the stabilization of β-catenin. Signaling by Wnt proteins results in inhibition of glycogen synthase kinase-3β (GSK3β) activity and dissociation of the adenomatous polyposis coli (APC)/axin complex, resulting in accumulation of β-catenin, which translocates to the nucleus. Here, β-catenin interacts with T cell factor (TCF) transcription factors to drive transcription of target genes.10 The Wnt pathway is regulated by a large number of antagonists, including the secreted frizzled-related proteins (sFRPs) and the Dickkopf (DKK) family proteins. These two classes of antagonists either act by direct binding to the Wnt ligands (the sFRPs) or by interacting with the LPR5/6 coreceptors, preventing binding of the Wnt proteins to the FRZ/LRP receptor complex (the DKKs).11

61 Chapter 3

Recent studies indicate that the Wnt signaling plays at least two distinct roles in the pathogenesis of MM. On the one hand, studies by our own laboratory12 as well as by the Anderson laboratory13 have demonstrated that MMs can display aberrant activation of the canonical Wnt signaling pathway. This Wnt pathway activation presumably results from auto- and/or paracrine stimulation by Wnts, and promotes tumor proliferation and dissemination.12,13 On the other hand, as first shown by Tian and colleagues,6 MMs overexpress and secrete the Wnt signaling inhibitor Dickkopf-1 (DKK1), which contributes to osteolytic bone disease by inhibiting osteoblast differentiation. 14–17 Similar to DKK1, secretion of the Wnt inhibitor sFRP2 by MM cells may also promote myeloma bone disease.5 Since both DKK1 and sFRP2 are established targets of TCF/β-catenin-mediated transcription,18–20 these findings suggests the presence of a negative feedback loop in MM in which DKK1 and sFRP2 act as potential tumor suppressors. In line with this hypothesis, we show here that DKK1 expression is often low or undetectable in advanced myeloma and is absent in MM cell lines, which are generally derived from advanced extramedullary myeloma. This silencing of DKK1 is caused by methylation of the DKK1 promoter and unleashes β-catenin/TCF mediated transcription

Materials and Methods ethics statements The study involving human biopsy samples was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee of The University of Amsterdam, AIEC (Algemene Instellingsgebonden Ethische Com- missie). Patients gave written informed consent for the sample collection. case selection and classification A panel of BM biopsy specimen from 41 MM and 7 MGUS patients, obtained at clinical diagnosis, was selected from the files of the Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Nether- lands. All patients were staged according to the Salmon–Durie system. For statistical analysis patients at stage I and II disease were grouped together (n = 16) and classified as early MM, whereas patients with stage III disease (n = 25) were classified as having advanced MM.

62 Transcriptional silencing of the Wnt-antagonist DKK1… microarray analysis For the analysis of expression of Wnt family members in MM patients, expression data publically available and deposited in the NIH Gene Expression Omni- bus (GEO National Center for Biotechnology Information [NCBI], http://www. ncbi.nlm.nih.gov/geo/ were used. These concerned the U133 Plus2.0 affymetrix oligonucleotide microarray data from 559 newly diagnosed MM patients included in total therapy 2/3 (TT2, TT3), provided by the University of Arkansas for Medical Sciences, GSE2658.21 immunohistochemistry/immunocytochemistry Immunohistochemical staining was performed on formalin-fixed, plastic-embedded bone marrow sections as described previously.22 Sections were deplastified in acetone, after which endogenous peroxidase was blocked with a 0.3% solution of H2O2 in methanol and followed by antigen retrieval for 10 minutes in TRIS/EDTA buffer (respectively 10 mM/1 mM) pH 9.0 at 100°C. After blocking 1 with serum free blocker (DAKO, Carpinteria, CA), the slides were either incubated for one hour at room temperature with anti-CD138 (IQP-153 IQ Prod- ucts, Groningen, The Netherlands) or overnight at 4°C with anti-DKK1 (Abcam, Cambridge, MA), or anti-β-catenin (clone 14 BD Biosciences, Erembodegem, 3 Belgium). For CD138 and β-catenin, binding of the antibody was visualized using the PowerVision plus detection system (Immunovision Technologies, Duiven, 4 The Netherlands) and 3, 3-diaminobenzidine (Sigma-Aldrich, St Louis, MO). Whereas binding of the DKK1-antibody was visualized with a biotinylated rab- 5 bit anti-goat antibody, followed by horseradish peroxidase (HRP)-conjugated streptavidin (DAKO) and DAB+(DAKO). The sections were counterstained 6 with hematoxylin (Merck, Darmstadt, Germany), washed and subsequently de- hydrated through graded alcohol, cleared in xylene, and coverslipped. Immuno- 7 cytochemical staining was performed on formalin-fixed cells. Briefly, slides were incubated for 30 minutes with PBS 0.3% Triton X-100 followed by blocking with serum free blocker (DAKO, Carpinteria, CA), the slides were then incubated overnight at 4°C with anti-DKK1. Binding of the DKK1-antibody was visualized with a biotinylated rabbit anti-goat antibody, followed by horseradish peroxidase (HRP)-conjugated streptavidin (DAKO) and 3-amino-9-ethyl carbazole (AEC). The biopsies were analyzed for DKK1 and β-catenin expression by two independent observers (RWJG and STP). DKK1 expression was scored in three semi-quantitative categories, i.e., low (0–25%), intermediate (25–75%) and high (75–100%), with the percentages indicating the number of DKK1 positive MM plasma cells. For β-catenin expression the intensity of nuclear staining was scored

63 Chapter 3

as negative/low, intermediate or high. For statistical analysis, the cases scored as negative/low and intermediate were grouped together as low β-catenin. methylation analysis Genomic DNA was extracted and purified from MM bone marrow suspensions and cell lines using DNAzol (Invitrogen Life technologies, Breda, The Nether- lands) according to the manufacturer’s protocol. The extracted DNA was modified by treatment with sodium bisulfite using the EpiTect Bisulfite Kit (Qiagen, Hilden, Germany). The methylation status of the DKK1gene was assessed using both methylation specific PCR (MSP) and bisulfate sequencing, as described by Aquilera et al.23 cell culture and demethylation treatment MM cell lines L363, UM-1, OPM-1, RPMI 8226, and PC-3 were cultured in RPMI medium 1640 (Invitrogen Life technologies) containing 10% clone I serum (HyClone), 100 units/ml penicillin, and 100 µg/ml streptomycin. XG-1 and LME-1 cell lines were cultured in IMDM medium (Invitrogen Life technologies) supplemented with transferrin (20 µg/ml) and β-mercaptoethanol (50 µM). For XG-1, medium was additionally supplemented with IL-6 (500 pg/m). 293T cells were cultured in DMEM medium containing 10% clone I serum (HyClone), 100 units/ml penicillin, and 100 µg/ml streptomycin. The PC-3 cell line was kindly provided by Christopher Hall. L-cells and Wnt3a-producing L-cells (L-306.72 Mouse L fibroblasts, stably transfected with Wnt3a) were cultured in DMEM medium (Invitrogen Life technologies) containing 10% clone I serum (HyClone), 100 units/ml penicillin, and 100 µg/ml streptomycin. Conditioned medium was harvested from 95% confluent flasks every 72 h and stored at 4°C. 5-aza-2′-deoxycytidine (5-aza-CdR Sigma-Aldrich) treatment was performed for 72 h at a final concentration of 5 µM, after which they were harvested and lysed in Trizol (Invitrogen Life technologies). rt-pcr Total RNA was isolated using Trizol according to the manufacturer’s protocol (In- vitrogen Life technologies). The RNA was further purified using iso-propanol precipitation and was concentrated using the RNeasy MinElute Cleanup kit (Qiagen). The quantity of total RNA was measured using a NanoDrop ND-1000 Spectro- photometer (NanoDrop Technologies, Wilmington, DE). 5 µg of total RNA was used for cDNA synthesis as described previously. The PCR mixture contained: 100 ng of cDNA, 1× PCR Rxn buffer (Invitrogen Life technologies), 0.2 mmol/L

64 Transcriptional silencing of the Wnt-antagonist DKK1…

dNTP, 2 mmol/L MgCl2, 0.2 µmol/L of each primer, and 1 U platinum Taq polymerase (Invitrogen Life technologies). PCR conditions were: denaturing at 95°C for 5 minutes, followed by 30 cycles of 30 s at 95°C, 30 s at 65°C (DKK1), 55°C (β-actin) and 30 s at 72°C. The reaction was completed for 10 minutes at 72°C. Primers used were: DKK1 forward (5′-GATCATAGCACCTTGGATGGG-3′) DKK1 reverse (5′-CAGTCTGATGACCGG-3′) β-actin forward (5′-GGATGCAGAAGGAG ATCACTG-3′) β-actin reverse (5′-TCCACACGGAGTACTTG-3′). All primers were manufactured by Sigma-Aldrich (Haverhill, UK). western blot analysis Conditioned medium was directly lysed in sample buffer, separated by SDS/10% PAGE, and blotted. Primary goat anti-DKK1 antibody was detected by a horseradish peroxidase-conjugated swine anti-goat antibody, followed by detection using Lumi-Light PLUS western blotting substrate (Roche). Protein was harvested from MM cell lines, fractionated using a nuclear/cytosol fractionation kit 1 (BioVision), separated by 10% SDS-polyacrylamide gel electrophoresis and subsequently blotted. The following antibodies were used: β-catenin (clone 14 BD Biosciences, Erembodegem, Belgium), β-tubulin (SantaCruz Biotechnologies), and histone H2B (Imgenex). Primary antibodies were detected by HRP-con- 3 jugated secondary antibodies, followed by detection using Lumi-Light PLUS western blotting substrate (Roche). 4 retroviral vector production and transduction 5 of multiple myeloma cell lines The LZRS-pBMN-DKK1-IRES-eGFP retroviral vector encoding DKK1 and 6 eGFP gene was generated by inserting the DKK1 gene from pcDNA3.1 into the BamHI and XhoI sites of LZRS-pBMN-IRES-eGFP vector (S-001-AB provided 7 by Dr. G. Nolan). The amphotropic Phoenix packaging cell line was transfected either with LZRS-pBMN-IRES-eGFP or LZRS-pBMN-DKK1-IRES-eGFP by the use of Fugene (Roche). The selection of transfected cells was performed with puromycine. Virus was harvested from the packaging cell lines when 95% of the cells were eGFP-positive as analyzed by fluorescence activated cell sorting (FACS). Transduction of MM cell lines was done by 16 h of incubation with virus supernatant in the presence of 10 µg/ml retronectin (Takara Biomedicals, Shiga, Japan). 2 days after the transduction eGFP positive cells were analyzed and sorted by FACS. The transduced MM cells were expanded and examined regularly for eGFP and DKK1 expression. The expression of DKK1 protein was determined in conditioned medium by western blot analysis.

65 Chapter 3 luciferase assay MM cells (10 mln) were transfected with TOPFLASH reporter construct (5 ug) and pRL-TK (1 ug) in 500 ul serum-free medium. Treatment with Wnt3a conditioned medium or L-cell conditioned medium was initiated 24 h upon transfection and subsequently maintained for 24 h. Next, MM cells were lysed in 100 ul passive lysis buffer (Promega) and firefly luciferase activity and Renilla luciferase activity were measured using the dual luciferase assay kit (Promega) following the manufacturers instruction. Renilla luciferase activity served as an internal control for transfection efficiency. statistical analysis The chi-square (χ2) test was used to test the correlation between the DKK1 expression and the stage of disease, the relation between expression of nuclear β-catenin and the stage of disease and the correlation between expression of DKK1 and nuclear β-catenin.

Results wnt signaling activation is associated with advanced stage multiple myeloma Previous studies have demonstrated that the malignant plasma cells in MM can display Wnt pathway activation. Since no mutations in Wnt pathway components were detected, this activation is presumably caused by auto- and/or paracrine stimulation by Wnt ligands secreted in the bone marrow microenvironment.12,13 In support of a ligand-dependent activation scenario, analysis of gene expression profiling data of the malignant plasma cells of 345 MM patients (http://www.ncbi. nlm.nih.gov/geo/ accession number GSE2658) revealed frequent (co-)expression of various Frizzleds (e.g. Fzd 1, 3, 6, 7 and 8) and the co-receptor LRP6 (LRP5 could not be studied because of defective probe sets), as well as various Wnts (e.g. 4, 5A, 5B 6, 10A and 16). Moreover, 51% of the primary MMs co-express the co-receptor LRP6 with at least one of the Frizzled genes and one of the Wnt genes (Table S1). Hence, in the majority of MM patients the malignant plasma cells appear to be well equipped to evoke autocrine activation of the Wnt pathway. Furthermore, Wnt ligands are produced by bone marrow stromal cells,12,24,25 which may cause paracrine Wnt pathway activation. To explore whether MM cells localized within the bone marrow microenvironment indeed display evidence of active Wnt signaling, we analyzed a panel of MM

66 Transcriptional silencing of the Wnt-antagonist DKK1… bone marrow biopsies for nuclear expression of β-catenin, a key feature of active canonical Wnt signaling.22,26 The panel consisted of 41 MM and 7 MGUS biopsies patients all biopsies were obtained at first diagnosis. Immunohistochemical studies revealed strong β-catenin staining in the malignant plasma cells of 34% of the MM patients (Figure 1 A–B). Interestingly, as shown in Figure 1B, β-catenin expression was significantly correlated with disease stage: nuclear β-catenin was present in none of the MGUS patients, 19% of the stage I/II patients, and 44% of the patients with advanced MM (p < 0.05). These findings imply the presence of active Wnt signaling in BM-cocalized MM cells and show that this is associated with advanced stage MM.

3 4 5 6 7

Figure 1. The relation between the Wnt pathway activation and the loss of DKK1 expression during MM progression (A) Representative pictures of immunohistochemical staining of a multiple myeloma patient displaying either high DKK1 and low β-catenin expression (upper panel) or with DKK1 loss and increased nuclear β-catenin localization (lower panel). Immunohistochemical stainings are shown for CD138

67 Chapter 3

(left column), β-catenin (middle column) and DKK1 (right column). (B) Nuclear β-catenin expression in relation to multiple myeloma progression (n = 48, p < 0.05). (C) DKK1 expression in relation to multiple myeloma progression (p < 0.05). (D) Representative pictures of immunocytochemical staining of multiple myeloma cell lines with goat polyclonal anti-DKK1 antibody (magnification: 400×). Prostate cancer cell line (PC-3) was used as positive control (PC) for the DKK1 staining. (E) Relation between the loss of DKK1 expression and nuclear localization of β-catenin (p > 0.05). A significant correlation (p < 0.05) between expression of nuclear β-catenin and DKK1 was observed in the two extreme groups identified based on β-catenin expression. * indicates p value < 0.05. absence of dkk1 expression unleashes wnt pathway activation in multiple myeloma Since stimulation by (autocrine and/or paracrine) Wnt ligands appears to underly Wnt pathway activation in MM, loss of secreted Wnt pathway antagonists like DKKs and sFRPs could have a major impact on the pathogenesis of MM. In particular DKK1, which binds to LRP5/6 thereby preventing activation of the pathway by Wnt ligands, might act as a potent negative regulator as it has been shown to be overexpressed by MM cells.18–20 To gain insight into the expression of the DKK1 protein in primary MM samples, we studied DKK1 expression in the above described panel of BM biopsies. DKK1 was detected in most (81%) of the biopsies. However, in a significant proportion expression was either restricted to a subfrac- tion of the malignant plasma cells or was entirely lost (Figure 1 A, C). In contrast to expression of β-catenin, DKK1 expression was negatively correlated with disease stage: whereas all MGUS patients and the majority of early (stage I/II) MM patients demonstrated high DKK1 expression, DKK1 was either partially or completely absent in 68% of the patients with advanced MM (p < 0.05) (Figure 1C). Furthermore, DKK1 protein expression was not detected in any of the studied MM cell lines (Figure 1D). Direct comparison between the level of DKK1 and nuclear β-catenin expression in individual patients indeed revealed a significant inverse correlation between these parameters (p < 0.05) (Figure 1E). Our results demonstrate a correlation between DKK1 expression and MM stage with a low or undetectable levels of DKK1 in a subset of patients with advanced stage MM (p < 0.05), and an inverse relation to nuclear β-catenin expression. The above findings suggest that silencing of DKK1 may unleash activation of the canonical Wnt pathway by Wnt ligands, leading to increased levels of nuclear β-catenin in MM cells of patients with advanced disease. To explore whether DKK1 expression by MM cells can indeed control the response of MM cells to ligand-induced Wnt pathway activation, we restored the DKK1 expression in the MM cell lines OPM-1 and UM-1 by retroviral transduction (Figure 2A). OPM-1 and UM-1 lack mutations in Wnt pathway components and co-express Frizzleds, the co-receptor LRP6, and Wnts, and have been reported to display constitu-

68 Transcriptional silencing of the Wnt-antagonist DKK1… tively active β-catenin/TCF-dependent transcription.12,27,28 As shown in Figure 2B, introduction of DKK1 expression in OPM-1 and UM-1 cells leads to reduced Wnt3a induced and baseline β-catenin levels. Furthermore, it results in a significant inhibition of ligand-induced TCF-mediated transcription (Figure 2C). These findings suggest thatDKK1 expression by MM cells could indeed restrain the response of these cells to Wnt ligands expressed in the BM microenvironment.

3 4 5 6 7

Figure 2. DKK1 expression represses Wnt pathway activation in MM (A) MM cell lines OPM-1 and UM-1 were transduced with either the LZRS-pBMN-IRES-eGFP (control) or the LZRS-pBMN-DKK1-IRES-eGFP (DKK1) virus. Conditioned medium of sorted, transduced cells was harvested and immunoblotted using a goat polyclonal antibody against DKK1. Representative immunoblot confirms the expression of DKK1 in the conditioned medium of LZRS- pBMN-DKK1-IRES-eGFP transduced cells. β-actin is shown as internal control for equal cell number. (B) Cytoplasmic and nuclear proteins were prepared from the LZRS-pBMN-IRES-eGFP (control) or

69 Chapter 3

the LZRS-pBMN-DKK1-IRES-eGFP (DKK1) MM cells, stimulated for 24h with Wnt3a conditioned medium (+). As a control, L-cells conditioned medium was applied (-).To assess β-catenin accumulation, nuclear and cytoplasmic cells lysate was immunoblotted by using a monoclonal anti-β-catenin antibody. The bottom part of the blot was stained with β-tubulin and Histone H2B as controls for cytoplasmic and nuclear proteins, respectively. (C) LZRS-pBMN-IRES-eGFP (control) or the LZRS-pBMN-DKK1- IRES-eGFP (DKK1) cells were transfected with TOPFLASH reporter and renilla contruct. 24 hours upon transfection cells were treated with L-cells conditioned medium (-) or Wnt3a conditioned medium (+).The relative light units value of LZRS-pBMN-IRES-eGFP cells treated with L-cells conditioned medium was normalized to 1. The mean ± SD of representative experiment performed in triplicate is shown. * indicates p value < 0.05 *** indicates p value < 0.001. by student’s t-test. dkk1 promoter hypermethylation in multiple myeloma cell lines and primary tumors To further analyze the absence of DKK1 expression in MM, we examined MM cell lines forDKK1 mRNA expression. In MM cell lines, which all lack detectable levels of DKK1 protein (Figure 1D and Figure S1), the DKK1 transcript was either undetectable (UM-1, RPMI 8226 and OPM-1), or weakly expressed (XG-1, L363 and LME-1) in comparison to the DKK1 expression in the (positive control) prostate cancer cell line PC-3 (Figure 3A).29 Since it has been reported that DKK1 is a target for epigenetic inactivation by CpG island promoter hypermethylation in several forms of cancer,23,30–33 we hypothesized that epigenetic silencing could also be responsible for the lack of DKK1 in primary MM cells and cell lines. Hence, we studied the DKK1 promoter region, including the CpG island encompassing the transcriptional start site for methylation using a methylation specific PCR (MSP) and a bisulphate- sequencing PCR (BSP) (Figure 3B). DNA isolated from the colon cancer cell lines HT-29 and DLD-1, previously reported to be unmethylated and methylated, respectively,23 was used to validate our experimental set-up (Figure 3C). Combined MSP and BSP analysis revealed that four of the 6 MM cell lines tested, i.e., L363, LME-1, UM-1, and OPM-1, showed hypermethylation of the DKK1 promoter, while XG-1 and RMPI8226 were unmethylated (Figure 3 D–E). Notably, the BSP analysis of OPM-1 revealed a methylation pattern that was not detected by the MSP primers, thus showing the necessity of using both techniques to adequately perform methylation studies. Interestingly, in the L363 cell line four out of ten sequenced clones displayed extensive CpG methylation, suggesting that either a subset of the cells show methylation or that there is partial monoallelic methylation within individual cells. The lack of DKK1 expression despite the absence of promoter methylation in RPMI8226 cells indicates either the mere lack of transcriptional activation or an alternative mechanism of DKK1 silencing.

70 Transcriptional silencing of the Wnt-antagonist DKK1…

3 4 5 6 7

Figure 3. DKK1 promoter methylation in MM cell lines (A) Total RNA was isolated and RT–PCR analyses were performed with the specific primers indicated. Complementary DNA from prostate cancer cell line (PC-3) was used as positive control (PC) for DKK1 expression. The β-actin expression was used as a loading control. (B) Schematic representation of the promoter area analyzed for DKK1, containing a CpG island. White arrows indicate the positions of primers used for bisulfite sequencing, and black arrows indicate the positions of primers used

71 Chapter 3

for methylation specific PCR. Each of the CpG dinucleotides is presented as open circle. (C) Upper panel. Representation of bisulfite genomic sequencing results of 5 clones of the DKK1promoter region in HT-29 and DLD-1 colon cell lines used as unmethylated (U_DNA) and methylated (M_DNA) control, respectively. The amplified 326 bp product corresponds to the DKK1 promoter region from -193 to +122. In total, 18 CpG dinucleotides (CpGs) within the CpG island were analyzed and are represented as open and closed circles, which indicate unmethylated and methylated CpG sites, respectively. Lower panel. Electropherograms of bisulfite modified DNA from DKK1 CpG island in HT-29 (U_DNA) and DLD-1 (M_DNA) cells. (D) Methylation specific PCR of the CpG island of the DKK1promoter region in MM cell lines. DNA bands in lanes labeled with U indicate PCR products amplified with primers recognizing unmethylated promoter sequences, whereas DNA bands in lanes labeled with M represent amplified products with primers designed for the methylated template. (E) Bisulfite sequencing analysis of the the DKK1 promoter region in multiple myeloma cell lines, open circles indicating unmethylated CpG sites, and closed circles representing methylated CpG sites. Percentages indicate the fraction of methylated CpG dinucleotides of the total CpG sites analyzed.

Since DKK1 expression appears to be decreased in MM cells isolated from patients with advanced stage MM, we next investigated the DKK1 promoter associated CpG island in the bone marrow mononuclear cells (BM-MNC) of patients with MM and compared this to BM-MNC from healthy donors. All 12 patients had stage III disease. The DKK1 promoter was hypermethylated in 33% of the primary MM specimens studied, as determined by BSP (Figure S2) and MSP analysis (Figure 4). In accordance with a previous study,32 none of the healthy donor BM samples tested displayed methylation of the DKK1 promoter (data not shown).

Figure 4. Analysis of DKK1 promoter methylation in MM bone marrow samples Methylation specific PCR of the CpG island of the DKK1 promoter region in the bone marrow samples of twelve patients with multiple myeloma (P1–P12), HT-29 and DLD-1 colon cell lines were used as unmethylated (U_DNA) and methylated (M_DNA) control respectively. DNA bands in lanes labeled with U and M indicate PCR products amplified with primers recognizing unmethylated and methylated promoter sequences respectively. promoter methylation silences dkk1 expression To assess whether DKK1 promoter methylation indeed plays a causative role in the transcriptional silencing of DKK1 in MM, we investigated whether DKK1 expression could be restored or enhanced by treatment with the demethylating agent 5-aza-2-deoxycytidine. Analysis by BSP of the genomic DNA isolated from OPM-1 and UM-1 cells, which combine a lack of DKK1 expression with hy-

72 Transcriptional silencing of the Wnt-antagonist DKK1… permethylation of its promoter (Figures 1D and Figure 3), confirmed that treatment results in a decrease in methylation of the DKK1 promoter (Figure 5A). Importantly, treatment of these cell lines results in restoration of DKK1 expression (Figure 5B). In addition, in LME-1, a cell line with a hypermethylated DKK1 promoter and low DKK1expression, DKK1 expression was increased upon 5-aza- 2-deoxycytidine treatment. As expected, treatment of XG-1 and RPMI8226, the two cell lines that lack methylation of theDKK1 promoter, did not affect the expression (Figure 5B). Taken together, these data establish a direct role for CpG island methylation in epigenetic silencing of DKK1 expression in MM.

3 4 5 Figure 5. Restoration of DKK1 expression in MM cell lines by 6 5-aza-2-deoxycytidine treatment (A) DKK1 promoter methylation analyzed by bisulfite genomic sequencing of 10 clones, on DNA isolated from 5-aza-2-deoxycytidine treated (5-aza-CdR) and untreated (PBS) MM cell lines UM-1 and 7 OPM-1. Frequency of methylation was calculated by dividing the number of methylated CpG sites by the total number of analyzed CpG sites. (B) Reverse transcriptase-PCR analysis for DKK1 gene expression in multiple myeloma cell lines in the absence and presence of the demethylating agent 5-aza- 2-deoxycytidine. β-actin expression is shown as an input control.

Discussion

We have previously demonstrated that the Wnt pathway, which is essential for normal T and B cell development8,34 and plays a key role in the development of several types of cancer,35,36 is aberrantly activated in MM and can promote MM tumor growth.12 Subsequent studies have confirmed the oncogenic potential

73 Chapter 3

of the Wnt pathway in MM by demonstrating that targeting the Wnt pathway by drugs or siRNAs leads to inhibition of MM growth.12,13,37–39 In addition, by knock- ing down β-catenin, Dutta-Simmons et al. revealed new potential Wnt/β-catenin transcriptional targets involved in various aspects of cell-cycle progression, such as CDC25 A/B, cyclins, cyclin-dependent kinases, and AurKA/B. Among the genes significantly downregulated upon β-catenin knock down, a significant proportion had LEF/TCF4 (GCTTTGT/A) binding sites in their promoters, identifying them as putative direct Wnt target genes.13 Although the exact cause(s) of aberrant Wnt pathway activation in MM has not yet been established, the absence of detectable Wnt pathway mutations12 as well as the (over)expression of Wnt ligands in the BM microenvironment by both stromal cells and by the MM cells themselves,12,28 (and table S1), suggests a key role for autocrine and/or paracrine stimulation. As a consequence, loss of secreted Wnt pathway antagonists like DKKs and sFRPs could have a major impact on the pathogenesis of MM. Indeed, we observed that whereas the DKK1 protein is strongly expressed in most primary MMs, the expression of this Wnt antagonist is down-regulated or even completely absent in a subgroup of advanced (stage III) MMs (Figure 1A, C). In addition, the DKK1 protein was undetectable in MM cell lines, which represent the ultimate, microenvironment independent, phase of MM tumor progression and are almost invariably derived from extramedullary MMs (Figure 1D and Figure S1). Interestingly, low or undetectable of DKK1 protein expression in BM samples of MM patients was correlated with an increased nuclear expression of β-catenin, a hallmark of canonical Wnt signaling (Figure 1E). DKK1 is a major Wnt pathway antagonist which acts by interfering with the binding of Wnt ligands to the LRP5/6 coreceptor.40 Importantly, the DKK1 gene itself is a direct target of β-catenin/TCF-mediated transcription,18–20 and DKK1 has been implicated in the feed-back regulation of Wnt signaling in several biological systems.41–43 Consist- ent with a tumor suppressor function, DKK1 silencing during tumor progression has been reported in several types of cancer.23,30–33 Our observation, that DKK1 levels can be low or undetectable in advanced MM and that restoration of its expression inhibits β-catenin/TCF transcriptional activity (Figure 2C), suggests that silencing of DKK1 may contribute to activation of the canonical Wnt pathway during MM progression. Like loss of function mutations, aberrant methylation of the promoter of tumor suppressor genes can provide a selective advantage to neoplastic cells.44–48 We identified DKK1promoter hypermethylation as a mechanism underlying the absence of DKK1 expression in MM (Figure 3, 4, 5). In four of the 6 MM cell

74 Transcriptional silencing of the Wnt-antagonist DKK1… lines tested, i.e., L363, LME-1, UM-1, and OPM-1, we showed hypermethylation of the DKK1 promoter (Figure 3). The CpG island analyzed encompasses the first exon of the DKK1 gene, which encodes the transcriptional and translational start sites as well as a significant part of the region upstream of the coding sequence, an organization characteristic of genes targeted by epigenetic silencing.46 Indeed, this CpG island has previously been implicated in DKK1 silencing in several types of cancer, including colorectal cancer, gastric cancer, breast cancer, medulloblastoma and leukemia.23,30–33 Importantly, the promoter methylation was reduced and DKK1 expression was either restored and/or markedly increased by the DNA demethylating agent 5-aza-2-deoxycytidine (Figure 5A–B), confirming that the observed aberrant methylation indeed was instrumental in the silencing of DKK1 expression (Figure 5B). Interestingly, 5-azacytidine has been reported to have significant cytotoxic activity against MM cell lines as well as patient-derived malignant plasma cells, but not against peripheral blood mononuclear cells.49 Of the cell lines used in our study, OPM-1 and UM-1 display a very high sensitiv- 1 ity to 5-azacytidine and treatment of these cells with this compound result not only in promoter demethylation but also in rapid and extensive cell death, which explains the rather modest induction of DKK1 expression. Indeed, in the LME-1 cell line, which shows a lower sensitivity to 5-azacytidine-induced cell death, 3 treatment results in a much stronger upregulation of DKK1 mRNA expression. In primary MM, we also observed dense methylation of the DKK1-associated 4 CpG island (Figure 4 and Figure S2). Since methylation of the DKK1 promoter was not observed in normal bone marrow samples,32 this methylation can be con- 5 sidered aberrant and disease-related. In addition to silencing of DKK1, silencing of other Wnt antagonists could also contribute to the enhanced Wnt signaling in 6 advanced MM. Consistent with this notion, Chim et al., have reported that constitutive Wnt signaling in MM cell lines is associated with methylation dependent 7 silencing of several Wnt inhibitors, including the sFRP1, 2, 4 and 5. Methylation of at least one of these soluble Wnt inhibitors was observed in most primary MM bone marrow samples.50 Our finding that the DKK1 promoter is methylated and Wnt pathway is hyperactivated in advanced multiple myeloma, strongly suggests the presence of autocrine Wnt signaling in malignant plasma cells. In accordance with this hypothesis we observed inhibition of nuclear β-catenin levels and of Wnt reporter activity upon restoration of DKK1 in MM cells (Figure 2 A–C). Importantly, MM cell lines used in this experiment have dense methylation of the DKK1promoter around the transcription start region and lack detectable DKK1 transcript (Figure 1D, Figure 3A and Figure S1). Taken together, these data suggest that activation of

75 Chapter 3

Wnt signaling in these cell lines is the consequence of DKK1 silencing and could reflect the progression-dependent Wnt pathway activation in patients with advanced MM. Our current study, in conjunction with work of others, points to a multi- facetted role of DKK1 in the pathogenesis of MM. Studies by Shaughnessy and collegues have previously also reported that DKK1 is strongly expressed by the malignant plasma cells of most MM patients.6 It was shown that secretion of the DKK1 can contribute to MM bone disease by inhibiting Wnt signaling in osteoblasts, thereby interfering with their differentiation.14–17 Furthermore, in line with our current findings, which suggests that DKK1 may act as a feed-back tumor suppressor, these authors also reported loss of DKK1 protein expression in a subgroup of patients with advanced MM.6 In addition to causing bone disease, inhibition of osteoblast differentiation by DKK1 may also promote MM growth, since mature osteoblasts can suppress myeloma growth, whereas immature osteoblasts express high levels of IL-6, a central growth and survival factor for myeloma plasma cells.51 Furthermore, DKK1 enhances the expression of receptor activator of NF-kappa B ligand (RANKL) and downregulates the expression of osteoprotegerin (OPG) in immature osteoblast.17 The resulting increased RANKL/OPG ratio leads to osteoclast activation promoting osteolytic bone disease. Osteoclasts may also support the growth of myeloma cells through secretion of IL-6 and osteopontin, and by adhesive interactions, stimulating the proliferation of malignant plasma cells.52 Thus, like several other soluble factors expressed by MM cells, for example vascular endothelial growth factor (VEGF) and hepatocyte growth factor (HGF),53–55 DKK1 can both, exercise paracrine effects on the BM microenvironment, and affect the MM cells in an autocrine fashion the net effect could be either enhanced or reduced tumor growth. Consistent with this hypothesis, by employing a MM SCID/rab mouse model, Yaccoby et al. showed that treatment with anti-human DKK1-neutralizing antibody stimulates osteoblast activity, reduces osteoclastogenesis, and promotes bone formation in myelomatous and nonmyelomatous bones.56 MM burden was also reduced, but notably, not in all mice bearing human myeloma cells.57 Simi- lar results were also obtained in a SCID/hu mouse model by Fulciniti et al.56,57 Together, these studies suggest that MM bone disease and tumor growth are in- terdependent, at least at the intramedullary stage, and that increased bone formation as a consequence of neutralization of DKK1, may also control MM growth. However, in a 5T2MM murine myeloma model treatment with the anti-DKK1 antibody BHQ880 also caused a reduction of osteolytic bone lesions but did not have any effect on tumor burden.58 Give our current finding that DKK1 inhibits

76 Transcriptional silencing of the Wnt-antagonist DKK1… autocrine canonical Wnt signaling in MM cells, inhibition of DKK1 could hyper- activate the Wnt pathway and thereby promote tumor growth, especially at extramedullary sites. Indeed, stimulation of the Wnt signaling pathway in a 5TGM1 mouse myeloma model significantly increased subcutaneous tumor growth.59 In patients, extramedullary growth is associated with aggressive disease, occurring subsequent to the osteolytic bone disease, often resulting in plasma cell leukemia. Importantly, in a human-mouse xenograft MM model, Dutta-Simmons et al. demonstrated that the Wnt pathway not only controls the proliferation of MM plasma cells but also their metastatic potential.13 Taken together, these studies suggest a scenario in which DKK1 has a dual, stage depend, role: whereas high DKK1 expression in early MM contributes to a tumor permissive micronenviro- ment within the BM, advanced MMs that have acquired BM independence may benefit from DKK1 loss, which enhanced Wnt signaling and thereby promotes MM growth and dissemination. Although blocking DKK1 inhibits osteolytic bone disease in vivo, targeting DKK1 in MM patients could enhance Wnt path- 1 way activity in MM plasma cells, which might increase the metastatic potential and extramedullary growth of the tumor. In conclusion, our study establishes for the first time a relation between low or absence of DKK1 expression and the presence Wnt pathway activation dur- 3 ing MM progression. Moreover, we demonstrate the presence of a functional ligand-dependent Wnt signaling in MM cells and identify methylation of the 4 DKK1 promoter as a mechanism underlying the absence of DKK1 expression in advanced stage MM. These data strongly suggest that epigenetic silencing of 5 DKK1 unleashes Wnt signaling in a subset of advanced myelomas, promoting disease progression. 6 7 References

1. Podar K, Chauhan D, Anderson KC. Bone marrow microenvironment and the identification of new targets for myeloma therapy. Leukemia 2009; 23: 10–24. 2. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009; 374: 324–329. 3. Edwards CM, Zhuang J, Mundy GR. The pathogenesis of the bone disease of multiple myeloma. Bone 2008; 42: 1007–1013. 4. Giuliani N, Morandi F, Tagliaferri S, Lazzaretti M, Donofrio G, et al. Production of Wnt inhibitors by myeloma cells: potential effects on canonical Wnt pathway in the bone microenvironment. Cancer 2007; Res 67: 7665–7674.

77 Chapter 3

5. Oshima T, Abe M, Asano J, Hara T, Kitazoe K, et al. Myeloma cells suppress bone formation by secreting a soluble Wnt inhibitor, sFRP-2. Blood 2005; 106: 3160–3165. 6. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med 2003; 349: 2483–2494. 7. Malhotra S, Kincade PW. Wnt-related molecules and signaling pathway equilibrium in hematopoiesis. Cell Stem Cell 2009; 4: 27–36. 8. Staal FJ, Clevers HC. WNT signalling and haematopoiesis: a WNT-WNT situation. Nat Rev Immunol 2005; 5: 21–30. 9. Staal FJ, Sen JM. The canonical Wnt signaling pathway plays an important role in lymphopoiesis and hematopoiesis. Eur J Immunol 2008; 38: 1788–1794. 10. van de Wetering M, Cavallo R, Dooijes D, van Beest M, van Es J, et al. Armadillo coactivates transcription driven by the product of the Drosophila segment polarity gene dTCF. Cell 1997; 88: 789–799. 11. Kawano Y, Kypta R. Secreted antagonists of the Wnt signalling pathway. J Cell Sci 2003; 116: 2627–2634. 12. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci U S A 2004; 101: 6122–6127. 13. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/{beta}-catenin involved in multiple myeloma disease progression. Blood 2009; 114: 2699–2708. 14. Haaber J, Abildgaard N, Knudsen LM, Dahl IM, Lodahl M, et al. Myeloma cell expression of 10 candidate genes for osteolytic bone disease. Only overexpression of DKK1 correlates with clinical bone involvement at diagnosis. Br J Haematol 2008; 140: 25–35. 15. Pinzone JJ, Hall BM, Thudi NK, Vonau M, Qiang YW, et al. The role of Dickkopf-1 in bone development, homeostasis, and disease. Blood 2009; 113: 517–525. 16. Qiang YW, Barlogie B, Rudikoff S, Shaughnessy JD Jr. Dkk1-induced inhibition of Wnt signaling in osteoblast differentiation is an underlying mechanism of bone loss in multiple myeloma. Bone 2008; 42: 669–680. 17. Qiang YW, Chen Y, Stephens O, Brown N, Chen B, et al. Myeloma-derived Dick- kopf-1 disrupts Wnt-regulated osteoprotegerin and RANKL production by osteoblasts: a potential mechanism underlying osteolytic bone lesions in multiple myeloma. Blood 2008; 112: 196–207. 18. Chamorro MN, Schwartz DR, Vonica A, Brivanlou AH, Cho KR, et al. FGF-20 and DKK1 are transcriptional targets of beta-catenin and FGF-20 is implicated in cancer and development. Embo J 2005; 24: 73–84. 19. Lescher B, Haenig B, Kispert A. sFRP-2 is a target of the Wnt-4 signaling pathway in the developing metanephric kidney. Dev Dyn 1998; 213: 440–451.

78 Transcriptional silencing of the Wnt-antagonist DKK1…

20. Niida A, Hiroko T, Kasai M, Furukawa Y, Nakamura Y, et al. DKK1, a negative regulator of Wnt signaling, is a target of the beta-catenin/TCF pathway. Oncogene 2004; 23: 8520–8526. 21. Zhan F, Barlogie B, Arzoumanian V, Huang Y, Williams DR, et al. Gene-expression signature of benign monoclonal gammopathy evident in multiple myeloma is linked to good prognosis. Blood 2007; 109: 1692–1700. 22. Groen RW, Oud ME, Schilder-Tol EJ, Overdijk MB, ten Berge D, et al. Illegitimate WNT pathway activation by beta-catenin mutation or autocrine stimulation in T-cell malignancies. Cancer Res 2008; 68: 6969–6977. 23. Aguilera O, Fraga MF, Ballestar E, Paz MF, Herranz M, et al. Epigenetic inactivation of the Wnt antagonist DICKKOPF-1 (DKK1) gene in human colorectal cancer. Oncogene 2006; 25: 4116–4121. 24. Dosen G, Tenstad E, Nygren MK, Stubberud H, Funderud S, et al. Wnt expression and canonical Wnt signaling in human bone marrow B lymphopoiesis. BMC Im- munology 2006; 7: 13–30. 25. Mahtouk K, Moreaux J, Hose D, Reme T, Meissner T, et al. Growth factors in multiple 1 myeloma: a comprehensive analysis of their expression in tumor cells and bone marrow environment using Affymetrix microarrays. BMC Cancer 2010; 10: 198–216. 26. Batlle E, Bacani J, Begthel H, Jonkheer S, Gregorieff A, et al. EphB receptor activity suppresses colorectal cancer progression. Nature 2005; 435: 1126–1130. 27. Qiang YW, Endo Y, Rubin JS, Rudikoff S Wnt signaling in B-cell neoplasia. Oncogene 3 2003; 22: 1536–1545. 4 28. Qiang YW, Walsh K, Yao L, Kedei N, Blumberg PM, et al. Wnts induce migration and invasion of myeloma plasma cells. Blood 2005; 106: 1786–1793. 29. Hall CL, Bafico A, Dai J, Aaronson SA, Keller ET. Prostate cancer cells promote os- 5 teoblastic bone metastases through Wnts. Cancer Res 2005; 65: 7554–7560. 30. Sato H, Suzuki H, Toyota M, Nojima M, Maruyama R, et al. Frequent epigenetic 6 inactivation of DICKKOPF family genes in human gastrointestinal tumors. Carcino- genesis 2007; 28: 2459–2466. 7 31. Suzuki H, Toyota M, Carraway H, Gabrielson E, Ohmura T, et al. Frequent epigenetic inactivation of Wnt antagonist genes in breast cancer. Br J Cancer 2008; 98: 1147–1156. 32. Suzuki R, Onizuka M, Kojima M, Shimada M, Fukagawa S, et al. Preferential hypermethylation of the Dickkopf-1 promoter in core-binding factor leukaemia. Br J Haematol 2007; 138: 624–631. 33. Vibhakar R, Foltz G, Yoon JG, Field L, Lee H, et al. Dickkopf-1 is an epigenetically silenced candidate tumor suppressor gene in medulloblastoma. Neuro Oncol 2007; 9: 135–144. 34. Timm A, Grosschedl R. Wnt signaling in lymphopoiesis. Curr Top Microbiol Im- munol 2005; 290: 225–252.

79 Chapter 3

35. Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 2006; 127: 469–480. 36. Polakis P. Wnt signaling and cancer. Genes Dev 2000; 14: 1837–1851. 37. Ashihara E, Kawata E, Nakagawa Y, Shimazaski C, Kuroda J, et al. Beta-catenin small interfering RNA successfully suppressed progression of multiple myeloma in a mouse model. Clin Cancer Res 2009; 15: 2731–2738. 38. Schmidt M, Sievers E, Endo T, Lu D, Carson D, et al. Targeting Wnt pathway in lymphoma and myeloma cells. Br J Haematol 2009; 144: 796–798. 39. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/TCF transcriptional complex in the treatment of multiple myeloma. Proc Natl Acad Sci USA 2007; 104: 7516–7521. 40. Semenov MV, Tamai K, Brott BK, Kuhl M, Sokol S, et al. Head inducer Dickkopf-1 is a ligand for Wnt coreceptor LRP6. Curr Biol 2001; 11: 951–961. 41. Lewis SL, Khoo PL, De Young RA, Steiner K, Wilcock C, et al. Dkk1 and Wnt3 interact to control head morphogenesis in the mouse. Development 2008; 135: 1791–1801. 42. Sick S, Reinker S, Timmer J, Schlake T. WNT and DKK determine hair follicle spac- ing through a reaction-diffusion mechanism. Science 2006; 314: 1447–1450. 43. Yamaguchi Y, Passeron T, Hoashi T, Watabe H, Rouzaud F, et al. Dickkopf 1 (DKK1) regulates skin pigmentation and thickness by affecting Wnt/ beta-catenin signaling in keratinocytes. Faseb J 2008; 22: 1009–1020. 44. Egger G, Liang G, Aparicio A, Jones PA. Epigenetics in human disease and prospects for epigenetic therapy. Nature 2004; 429: 457–463. 45. Esteller M, Fraga MF, Paz MF, Campo E, Colomer D, et al. Cancer epigenetics and methylation. Science 2002; 297: 1807–1808. 46. Feltus FA, Lee EK, Costello JF, Plass C, Vertino PM. Predicting aberrant CpG island methylation. Proc Natl Acad Sci U S A 2003; 100: 12253–12258. 47. Herman JG. Epigenetic changes in cancer and preneoplasia. Cold Spring Harb Symp Quant Biol 2005; 70: 329–333. 48. Klein G. Epigenetics: surveillance team against cancer. Nature 2005; 434: 150. 49. Kiziltepe T, Hideshima T, Catley L, Raje N, Yasui H, et al. 5-Azacytidine, a DNA methyltransferase inhibitor, induces ATR-mediated DNA double-strand break responses, apoptosis, and synergistic cytotoxicity with doxorubicin and bortezomib against multiple myeloma cells. Mol Cancer Ther 2007; 6: 1718–1727. 50. Chim CS, Pang R, Fung TK, Choi CL, Liang R. Epigenetic dysregulation of Wnt signaling pathway in multiple myeloma. Leukemia 2007; 21: 2527–2536. 51. Klein B, Zhang XG, Jourdan M, Content J, Houssiau F, et al. Paracrine rather than autocrine regulation of myeloma-cell growth and differentiation by interleukin-6. Blood 1989; 73: 517–526.

80 Transcriptional silencing of the Wnt-antagonist DKK1…

52. Abe M, Hiura K, Wilde J, Shioyasono A, Moriyama K, et al. Osteoclasts enhance myeloma cell growth and survival via cell-cell contact: a vicious cycle between bone destruction and myeloma expansion. Blood 2004; 104: 2484–2491. 53. Derksen PW, Keehnen RM, Evers LM, van Oers MH, Spaargaren M, et al. Cell surface proteoglycan syndecan-1 mediates hepatocyte growth factor binding and promotes Met signaling in multiple myeloma. Blood 2002; 99: 1405–1410. 54. Le Gouill S, Podar K, Amiot M, Hideshima T, Chauhan D, et al. VEGF induces Mcl-1 up-regulation and protects multiple myeloma cells against apoptosis. Blood 2004; 104: 2886–2892. 55. Standal T, Abildgaard N, Fagerli UM, Stordal B, Hjertner O, et al. HGF inhibits BMP- induced osteoblastogenesis: possible implications for the bone disease of multiple myeloma. Blood 2007; 109: 3024–3030. 56. Yaccoby S, Ling W, Zhan F, Walker R, Barlogie B, et al. Antibody-based inhibition of DKK1 suppresses tumor-induced bone resorption and multiple myeloma growth in vivo. Blood 2007; 109: 2106–2111. 57. Fulciniti M, Tassone P, Hideshima T, Vallet S, Nanjappa P, et al. Anti-DKK1 mAb 1 (BHQ880) as a potential therapeutic agent for multiple myeloma. Blood 2009; 114: 371–379. 58. Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy JD Jr., et al. Inhibiting Dickkopf-1 (Dkk1) removes suppression of bone formation and prevents the development of osteolytic bone disease in multiple myeloma. J Bone Miner Res 2009; 24: 3 425–436. 4 59. Edwards CM, Edwards JR, Lwin ST, Esparza J, Oyajobi BO, et al. Increasing Wnt signaling in the bone marrow microenvironment inhibits the development of myeloma bone disease and reduces tumor burden in bone in vivo. Blood 2008; 111: 5 2833–2842. 6 7

81 Chapter 3 supplementary material

Figure S1 Bisulfite genomic sequencing of the DKK1 promoter in MM cell line. Representative pictures of immunocytochemical staining of multiple myeloma cell lines with goat polyclonal anti-DKK1 antibody (magnification: 400×). Prostate cancer cell line (PC-3) was used as positive control for the DKK1 staining.

82 Transcriptional silencing of the Wnt-antagonist DKK1…

3 Figure S2 4 Bisulfite genomic sequencing of the DKK1 promoter in MM bone marrow samples.Bisulfite sequencing analysis was performed on DNA isolated from total bone marrow samples of twelve MM patients (P1–P12). For individual patient 5 clones of the DKK1 promoter region are presented. Open 5 circles indicate unmethylated CpG sites; closed circles represent methylated CpG sites. 6 7

83 Chapter 3

Table S1 Expression of Wnt family members in 345 MM patients from total therapy 2 (TT2) patients set.

84 Transcriptional silencing of the Wnt-antagonist DKK1…

3 4 5 6 7

Chapter4

Loss of CYLD expression unleashes Wnt signaling in multiple myeloma and is associated with aggressive disease

Kinga A. Kocemba1,5, Harmen van Andel1,5, Anneke de Haan-Kramer1, Annemiek Broyl3, Mark van Duin3, Pieter Sonneveld3, Madelon M. Maurice4, Marie José Kersten2, Marcel Spaargaren1,6, Steven T. Pals1,6,7

1Department of Pathology, and 2Hematology, Lymphoma and Myeloma Center Amsterdam (LYMMCARE), Academic Medical Center, University of Amsterdam, 1105 AZ Amsterdam, The Netherlands,3 Department of Hematology, Erasmus MC, Rotterdam, The Netherlands, 4Department of Cell Biology, University MC Utrecht, the Netherlands, 5These authors contributed equally to this work,6 These authors share last authorship

Manuscript in preparation Abstract

In multiple myeloma (MM), loss of the gene encoding the deubiquitinating enzyme CYLD is a common recurrent genomic aberration. Here, we investigated the expression of CYLD in human myeloma cell lines (HMCLs) and in a large panel of primary MMs and explored its relation to disease progression and prognosis, as well as its functional impact on intracellular signaling and MM cell survival. We show that CYLD expression in HMCLs and primary MMs is highly variable and that it is involved in the regulation of MM cell growth and survival. In primary MMs, CYLD loss was associated with disease progression from MGUS to MM and correlated with reduced progression free- and overall-survival. Func- tional assays employing inducible CYLD silencing or overexpression revealed that CYLD acts as a negative regulator of both NF-ĸB and Wnt/β-catenin signaling and sensitizes MM cells to NF-ĸB stimuli and Wnt ligands. Interestingly, in primary MMs, low CYLD expression was strongly correlated to the presence of a proliferative- and Wnt signaling gene-expression signature. Our findings identify CYLD as a regulator of both NF-ĸB and Wnt/β-catenin signaling in MM and indicate that loss of CYLD enhances MM aggressiveness through a mechanism involving Wnt pathway activation.

88 Introduction

Multiple Myeloma (MM) is a still incurable B-lineage malignancy, characterized by uncontrolled expansion of malignant plasma cells in the bone marrow (BM). The transition of a normal plasma cell to a fully transformed, aggressive myeloma 1 cell is a multistep process, which requires the acquisition of multiple structural and numerical genomic abnormalities. Most of this evolution takes place in the BM, indicating that the interaction with the BM microenvironment plays a critical role in the pathogenesis of MM.1,2 The most common recurrent genomic aber- 3 rations in MM are found in both the pre-malignant monoclonal gammopathy of undetermined significance (MGUS) and MM, and involve translocation of the 4 immunoglobulin heavy chain (IgH) locus and one of the following partner genes: CCND1 (Cyclin D1), MMSET, CCND3 (Cyclin D3), c-MAF or MAFB. Most of the 5 remaining cases of MM are associated with trisomies of chromosomes 3, 5, 7, 9, 11, 15, 19 and 21.3 Other recurrent genomic abnormalities in MM include activating 6 RAS mutations, deletions of 6q, 13q, 8p, 16q, or 17q, or changes involving MYC family genes, and are mostly associated with progression to advanced stage MM.4–6 7 Deletion of all or part of chromosome 16q has been reported in approximately 15–20% of MM cases and is associated with an unfavorable prognosis.7 Gene expression profiling and gene mapping using single-nucleotide polymorphism (SNP) arrays identified the CYLD gene on 16q12.1 as a potential determinant of poor prognosis in a subset of these cases.7 CYLD was originally identified as a tumor suppressor gene mutated in familial cylindromatosis (Brooke-Spiegler syndrome), an autosomal dominant disorder predisposing to benign tumors of skin appendages.8 Subsequent studies have linked loss of the tumor suppressor function of CYLD to the pathogenesis of several other tumors including melanoma, T-cell acute lymphoblastic leukemia (T-ALL), and colon and hepatocellular carcinoma.9–11 In MM, loss of CYLD, resulting from biallelic deletion or inactivating

89 Chapter 4 mutations, is among the most common genomic abnormalities.12,13 CYLD is a member of the USP family of deubiquitinating enzymes (DUBs). A number of studies have shown that CYLD acts as a negative regulator of nuclear factor-ĸB (NF-ĸB) signaling by removing K63-linked polyubiquitin chains from TRAF2, TRAF6, and NEMO.14,15 Other CYLD substrates important for NF-ĸB signaling include RIPK1, BCL3 and TAK1.16–18 More recently, it was reported that CYLD can also acts as a negative regulator of proximal events in Wnt/β-catenin signaling. It was shown that loss of CYLD causes hyperubiquitination of the DIX domain of the adapter protein Dishevelled (Dvl), leading to enhanced Dvl polymerization and Wnt signaling.19 Importantly, human cylindroma skin tumors that arise from mutations in CYLD were indeed found to display hyperactive Wnt signaling, suggesting that the tumor growth instigated by loss of CYLD involves enhancement of Wnt responses.19 In the present study, we further investigated the expression of CYLD in MM and explored its relation to disease progression and prognosis as well as its functional impact on intracellular signaling and on MM growth and survival. Our results identify CYLD as a regulator of both NF-ĸB and Wnt/β-catenin signaling in MM and suggest that loss of CYLD enhances MM aggressiveness through a mechanism involving aberrant Wnt pathway activation.

Material and methods cell culture UM-1, OPM-1, UM-3, U266, FRAVEL, ARH-77, L363, NCI-H929, RPMI 8226 were cultured in RPMI medium 1640 (Invitrogen Life Technologies, Carsbad, CA) containing 10% clone I serum (HyClone, Logan, UT), 100 units/ml penicillin (Sigma Aldrich, St. Louis, MO), and 100 µg/ml streptomycin (Sigma Aldrich). XG-1, XG-3, ANBL-6 and LME-1 were cultured in IMDM medium (Invitrogen Life technologies) supplemented with 20 µg/ml human apo-transferrin (Sigma Aldrich) and 50 µM β-mercaptoethanol (Sigma Aldrich). For XG-1, XG-3 and ANBL-6 medium was additionally supplemented with 500 pg/mL IL-6 (Prospec, East Brunswick, NJ). L-cells (CCL1.3, ATCC, Manassas, VA) and Wnt3a-producing L-cells (CRL-2467, ATCC) were cultured in DMEM medium (Invitro- gen Life technologies) containing 10% clone I serum (HyClone), 100 units/ml penicillin, and 100 µg/ml streptomycin. Conditioned medium was harvested confluent cultures every 72 h and stored at 4°C. The doxycycline-inducible UM- 1shCYLD/a/b cell lines were generated as described previously, using the T-REx

90 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

System (Invitrogen Life technologies).20 shRNA sequences targeting CYLD were previously described by Tauriello et al.19 CYLD knock down was obtained by incubating UM-1shCYLDa/b cells for 72 h with 0.2 µg/ml doxycyclin (Sigma Aldrich). LZRS-pBMN-CYLD-IRES-YFP was generated by inserting CYLD between the BamHI and NotI restriction site of LZRS-pBMN-IRES-YFP using standard cloning techniques. Phoenix-GALV packaging cells were transfected with either LZRS-pBMN-IRES-YFP or LZRS-pBMN-CYLD-IRES-YFP by the use of Fugene 6 according to the manufacturer’s instructions (Roche, Basel, Switzerland). 48 hours after transfection, medium containing 2 µg/ml puromycin (Sigma Aldrich) was added to enrich virus producing cells to > 95% YFP+ over a 7 day course. Transduction of MM cell lines was performed on 10µg/mL retronectin (Takara Biomedicals, Japan) coated plates by centrifuging for 90 minutes on 35°C. Subsequently, cells were incubated for 16h on 37°C after which the virus supernatant was replaced with culture medium. 48 hours after transduction the % of YFP positive cells was analyzed on a FACSCantoII flow cytometer (BD Biosci- 1 ences, Franklin Lakes, NJ) and subsequently sorted. q-pcr Total RNA was isolated using TRIzol (Invitrogen Life Technologies) according to 3 the manufacturer’s instructions. The quantity of total RNA was measured using a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, 4 DE). cDNA synthesis was performed as previously described.21 The PCR mixture contained: 1 µl of cDNA (0.1µg/mL), 0,2 ul of each 10µM primers, 2,5 µl of Sensi- 5 Mix (Bioline, London, UK) and 2,5 µl of dH2O. PCRs were conducted in triplo on a Lightcycler 480 (Roche). Primers used were: CYLD forward (5′‑TGCAGGCT 6 GTACGGATGGAACCT‑3′); CYLD reverse (5′‑TCCTGATGCAGCCTCCA CCT ‑3′); CD74 forward (5′‑TGACCAGCGCGA CCTTATCT‑3′); CD74 re- 7 verse (5′‑GAGCAGGTGCATCACATGGT‑3′), RELB forward (5′‑CATCGA GCTCC GGGATTGT‑3′); RELB reverse (5′‑TCTTCCTGGTTCTTCCAT TGGGCA‑3′) TBP forward (5′‑CACCTCGCTCGCAGACACCAC‑3′); TBP reverse (5′‑GAGAGCAGCAGAGATGGAAGGAAAAC‑3′),CCND1 forward (5′‑CCTGTCCTACTACCGCCTCA‑3′), CCND1 reverse (5’‑CTTGACTCC AGCAGGGCTTC‑3′). All primers were manufactured by Biolegio (Nijmegen, The Netherlands) reporter assay 1 × 106 cells were transfected by electroporation with the TOPflash/FOPflash Wnt reporter construct (10 μg) or NF-ĸB reporter construct (10 μg) in combination

91 Chapter 4 with pRL-TK-Renilla (2 μg) in 500ul serum-free medium. Cells were allowed to recover and express reporter plasmids for 24h at 37°C. After 24h, cells were stimulated over night with 1:1 (v/v) diluted control or Wnt conditioned medium in case of Wnt assays or 50 ng/mL TNFα (Prospec) or vehicle in case of NF-ĸB assays. Firefly luciferase and renilla luciferase activity were measured using the dual luciferase assay kit (Promega, Madison, WI) on a GloMax-Multi+ (Promega) according to the manufacturer’s instructions. Renilla luciferase activity served as a control for transfection efficiency. immunoblotting and facs analysis Protein for immunoblotting was harvested from MM cell lines, separated by 10% SDS-polyacrylamide gel electrophoresis and subsequently blotted. Nuclear and cytosolic fractions were prepared using the Nuclear/Cytosolic fractionation kit (Biovision, Milpitas, CA) according to the manufacturer’s instructions. The following antibodies were used for immunoblotting: rabbit anti-histone 2B (Abcam, Cambridge, UK), mouse anti-tubulin (Sigma Aldrich), mouse anti-P65 (Santa Cruz Biotechnology, Dallas, TX), mouse anti CYLD (clone E-10, Santa Cruz Biotechnology), mouse anti-lamin A/C (clone 4C11, Cell Signaling, Beverly MA) and mouse anti-β-catenin (clone 14, BD Biosciences). Primary antibodies were detected by HRP-conjugated secondary antibodies (DAKO, Glostrup, Denmark), followed by detection using Amersham ECL Westernblot Detection Reagent (GE Healthcare, Little Chalfont, UK). multiple myeloma proliferation and drug toxicity assay FACS-sorted YFP positive MM cells transduced with either LZRS-pBMN-IRES- YFP or LZRS-pBMN-CYLD-IRES-YFP were plated at a density of 15000 cells/ well on flat bottom 96-well plates in 200 μl of RPMI 1640 medium containing 5% FCS. After 3 and 5 days, the number of living cells was determined based on a FSC/SSC dot plot analyzed on a FACSCantoII flow cytometer (BD Biosciences, Franklin Lakes, NJ). The results are presented as mean +/- SD of samples assayed in triplicate. All experiments were performed at least three times. Student’s t-test was used for statistical data comparison. cell cycle analysis Cells were fixed in ice-cold 70% ethanol for 24 hours. Cells were washed twice with PBS/0.1%BSA and treated for 30 minutes with 250 µg/mL RNAse A (Bioke, Leiden, The Netherlands) at 37°C after which cells were stained with 10µg/mL

92 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma… propidium iodide (Invitrogen Life Technologies). Cell cycle distribution was ana- lysed by flowcytometry and quantified by Flowjo software (Flowjo, Ashland, OR). patients and statistical analysis Gene expression data publically available and deposited in the NIH Gene Expres- sion Omnibus (GEO) National Center for Biotechnology Information [NCBI]. These concerned the U133 Plus 2.0 affymetrix oligonucleotide microarray data from; 44 MGUS patients, 22 healthy donors and 559 newly diagnosed MM patients included in total therapy 2/3 (TT2, TT3), provided by the Donna D. and Donald M. Lambert Laboratory of Myeloma Genetics, University of Arkansas for Medical Sciences, Little Rock, AR, USA22 and the gene expression and survival data of MM patients included in randomized clinical trials, i.e. the Institu- tional Review Board-approved HOVON-65/GMMG-HD4 (ISRCTN64455289) trial for newly diagnosed patients with MM.23 The MM NF-кB profile was determined by Annuziata et al.12 Genes comprising 1 the NF-кB profile in MM were those that were decreased in expression by > 40% in at least six of eight time points following treatment of L363 cells with IKKβ inhibitor (MLN120b) (Millennium Pharmaceuticals, Cambridge, MA) for 8–24h in three separate experiments (accession number GSE8487). Genes were chosen if 3 they correlated in expression across the MM cell lines (r > 0.5). NF-кB profile genes, using Affymetrix U133plus 2.0 data, relied on the following probe sets: 4 210538_s_at (BIRC3), 202644_s_at (TNFAIP3), 207535_s_at (NFKB2), 204116_ at (IL2RG), 203927_at (NFKBIE), 205205_at (RELB), 201502_s_at (NFKBIA), 5 209619_at (CD74), 203471_s_at (PLEK), 210018_x_at (MALT1), 223709_s_at (WNT10A). 6 The MM Wnt profile was determined based on study reported by Dutta-Sim- mons et al.24 Genes comprising the Wnt profile in MM were those that most down- 7 regulated by β-catenin knockdown compared with control shRNA in MM1.S cells and contained LEF1/TCF4 binding sites (GCTTTGT/A) in their promoters. Genes were chosen if they correlated in expression across the MM patients. Wnt profile genes, using Affymetrix U133plus 2.0 data, relied on the following probe sets: 207828_s_at (CENPF), 208079_s_at (AURKA), 225655_at (UHRF1), 218755_at (KIF20A), 223307_at (CDCA3), 202095_s_at (BIRC5), 218039_at (NUSAP1), 1555772_a_at (CDC25A), 203276_at (LMNB1), 210559_s_at (CDC2). The probe set for CYLD expression was chosen based on the highest frequency of P calls (present calls) in MM patients data sets. The call (“present” or “absent”) is determined by Affymetrix GCOS-software and indicates whether a gene is reliably expressed or not.

93 Chapter 4

The Kruskal-Wallis test with Dunn correction for multiple comparisons was used to compare CYLD gene expression in different molecular subgroups in MM patients and to analyze the difference in CYLD expression during disease progression. Two-tailed t-tests were employed to statistically analyze the in vitro data using Prism 5.0 software (Graphpad Software, San Diego, CA). p < 0.05 was defined as statistically significant. Cox proportional hazard regression was used to assess the influence of CYLD expression on survival outcome. Survival curves were estimated by the product-limit methods of Kaplan and Meier and compared using the log-rank test using Prism 5.0 software. Spearman rank correlation coefficients were used to determine correlations between CYLD expression and NF-ĸB and Wnt profile and to analyze the correlation between Wnt and NF-ĸB profile and the proliferation index (PI). Array comparative genomic hybredization (aCGH) on the UM-3 cell line was performed using the Agilent 180K (Amadid 023363) oligo-array platform (Agilent Inc, Santa Clara, CA). Data was analyzed to the NCBI Build 36.1/hg18 browser.

Results human myeloma cell lines and primary mms show highly variable levels of cyld expression Loss of the tumor suppressor function of CYLD in cancer can be caused by mu- tational inactivation of CYLD25–27, but may also involve alternative mechanisms, including transcriptional and epigenetic suppression of CYLD function by phosphorylation or microRNAs.9,11,28,29 To investigate CYLD expression and function in MM, we initially assessed CYLD mRNA and protein expression in a panel of human myeloma cell lines (HMCLs), using qPCR and immunoblotting. As is shown in Figure 1, HMCLs displayed a broad range of CYLD expression levels: Whereas moderate to high levels of CYLD mRNA were found in ANBL-6, L363, RPMI8226, U266, XG-1, and XG-3, CYLD expression in the other cell lines was either low (Fravel, NCI-H929, LME-1, OPM-1, and UM-1) or undetectable (UM-3) (Figure 1A). CYLD protein levels largely, but incompletely, paralleled these mRNA expression data (Figure 1B): Immunoblotting showed high expression levels in ANBL-6, L363, RPMI8226, U266, and XG-1, low levels in Fravel, NCI-H929, LME-1, OPM-1, and XG-3, and a complete absence of CYLD protein in UM-3. However, for XG-3, protein levels were relatively low compared to mRNA expression. Similar to HMCLs, primary MMs also displayed a broad range of CYLD mRNA and protein expression levels (Figure 1C and D). Notably,

94 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

CYLD protein was not expressed in 2 of 5 samples tested (Figure 1D), but in one of these samples (pMM5) CYLD mRNA was readily detectable. Hence, CYLD expression in both HMCLs and primary MM cells is highly variable and involves regulation at the transcriptional as well as post-transcriptional level.

3 4 5 6 7

Figure 1. CYLD is variably expressed in malignant plasma cells (A) Analysis of CYLD mRNA expression in MM cell lines as determined by quantitative RT-PCR. TBP was used as an input control. (B) Western blot analyses of CYLD protein expression. β-actin was used as a loading control. (C) CYLD mRNA in 5 primary MM. TBP was used as an input control. (D) Western blot analyses of CYLD protein expression in 5 primary MM. β-actin expression was used as a loading control.

95 Chapter 4 loss of cyld expression in multiple myeloma is associated with disease progression and with a proliferation gene-expression signature To obtain a more global view of CYLD mRNA expression in primary MMs and explore its relation to disease progression and MM molecular subgroup, we next analyzed a publically available microarray dataset containing 414 newly diagnosed MM patients22 for CYLD expression. Interestingly, this analysis revealed that CYLD expression is markedly reduced in a subset of patients with overt MM, compared to normal bone marrow plasma cells (BMPC) (Figure 2A). By contrast, in MGUS and smoldering myeloma patients, CYLD levels were comparable to normal BPMCs (Figure 2A). These results indicate that disease progression from MGUS and smoldering MM to full blown MM is accompanied by loss of CYLD expression in a subset of MM patients. Analysis of CYLD expression in distinct gene-expression profiling defined MM molecular subgroups22, revealed that individual MM cases with reduced CYLD expression can be found in each subgroups. Importantly, however, CYLD expression in the proliferation (PR) subgroup, characterized by over expression of numerous cell cycle- and proliferation-related genes and associated with an adverse prognosis, was strongly decreased in comparison with all other molecular subgroups (p < 0.001) (Figure 2B). Hence, loss of CYLD expression in primary MMs is associated with disease progression and with a proliferative phenotype.

Figure 2. CYLD expression in relation to MM progression and molecular subgroup classification (A) CYLD mRNA expression analyses of publically available micro-array data.22 22 healthy donors bone marrow plasma cells (BMPCs), 44 monoclonal gammopathy of undetermined significance (MGUS) patients, 12 smoldering MM (sMM) patients and 414 MM patients classified in 7 molecular subgroups. A subset of MMs shows low/absent CYLD expression. (B) CYLD mRNA expression in different molecular MM subgroups.22 CYLD expression in the proliferation (PR) subgroup is significantly lower compared to other molecular subgroups (PR vs LB, HY, CD1, CD2, MS, p < 0.001, PR vs MF p < 0.05, by Kruskal-Wallis test with Dunn correction for multiple comparisons).

96 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

3 4 5 6 7

Figure 3. Restoration of CYLD expression represses growth and survival of CYLD-deficient MM cells (A) Array comparative genomic hybridization (aCGH) showed a homozygous deletion of the CYLD locus (16q21) in UM-3, resulting in the absence of CYLD mRNA and protein (Figure 1). (B) UM-3 cells were transduced with CYLD (LZRS-pBMN-CYLD-IRES-YFP) or control vector (LZRS-pBMN-IRES-YFP). Expression of CYLD was analyzed by immunoblot. (C) left panel: UM-3 cells were transduced with CYLD (LZRS-pBMN-CYLD-IRES-YFP) or a control vector (LZRS-pBMN-IRES-YFP) containing YFP, allowing analyses of the percentage of transduced cells over time. The proportion of YFP positive cells declined in CYLD transduced cells population, but was stable in control vector transduced cells. The percentage of YFP positive cells was normalized to 100% at time point 0. Results are presented as mean +/- SD of samples assayed

97 Chapter 4

in triplicate. All experiments were performed at least three times. Student’s t-test was used for statistical data comparison. *** indicates p value < 0.001, by student’s t-test. right panel: sorted CYLD or control vector transduced UM-3 cells were cultured in low serum conditions for 5 days. The number of viable cells was determined on indicated time points. The number of living cells at day 0 was normalized to 1. The mean ± SD of a representative experiment performed in triplicate is shown. *** indicates p value < 0.001, by student’s t-test. (D) sorted CYLD or control vector transduced UM-3 cells were analyzed for cell cycle distribution by propidium iodide staining. The mean ± SD of three independent tranductions is shown (lower panel). *** indicates p value < 0.001, ** indicates p value < 0.01, *indicates p value < 0.05, by student’s t-test. cyld regulates growth and survival of multiple myeloma cells In MM, biallelic deletion or inactivating mutations of CYLD are among the most common genomic aberrations, suggesting a tumor suppressor function in these tumors.7,12,13 As is shown in Figure 3A, the HMCL UM-3 models these patients since it contains a homozygous deletion of CYLD, explaining the complete absence of CYLD mRNA and protein expression (Figure 1A and B). To provide direct functional evidence for a tumor suppressor role of CYLD in MM, we explored the impact of re-introducing CYLD on the growth and survival of UM-3 cells (Figure 3B). Upon retroviral transduction of UM-3 cells with a vector expressing CYLD and YFP, CYLD expressing cells were outcompeted by CYLD negative cells, resulting in a progressive loss of YFP+ cells (Figure 3C, left panel). This growth suppressive effect of CYLD was confirmed in experiments comparing growth of FACS-sorted UM-3 cells transduced with either a CYLD containing or a control vector (Figure 3C, right panel). Further analysis revealed an increase in cell death and inhibition of proliferation after ectopic expression of CYLD, as shown by an increased subG1 population and decreased fraction of cells in S phase, respectively (Figure 3D). Hence, CYLD acts as a tumor suppressor in MM by repressing proliferation and survival. loss of cyld sensitizes multiple myeloma cells to both nf-ĸb-stimuli and wnt ligands CYLD is an established negative regulator of NF-ĸB signaling14–18, and has recently been reported to also negatively regulate Wnt/β-catenin signaling.19 Loss of CYLD expression might thus increase the sensitivity of MM cells to NF-ĸB-stimuli and Wnt ligands expressed in BM microenvironment. To explore this notion, we stably transfected UM-1 MM cells with doxycyclin-inducible shRNAs to silence CYLD expression (Figure 4A). Of note, UM-1 does not carry any known genetic abnormalities in components of the NF-ĸB or Wnt/β-catenin pathways.21 To rule out off-target effects of the shRNAs, two different UM-1-shCYLD lines, each

98 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma… containing a different non-overlapping targeting sequence for CYLD, were generated (shCYLDa and shCYLDb). As shown in Figure 4A, doxycyclin treatment of UM1-shCYLDa or UM1-shCYLDb cells led to a strong reduction in CYLD expression but did not affect CYLD expression in control (TR) cells, containing the Tet-repressor without shCYLD.

3 4 5 6 7

Figure 4. CYLD inhibits NF-ĸB signaling in malignant plasma cells (A) UM-1 cells transfected with a doxycyclin inducible vector containing a shRNA targeting CYLD (UM1-shCYLDa/b) or a TET repressor only (UM-1 TR) were treated -/+ doxycyclin for 72h. CYLD expression was analyzed by immunoblot. β-actin expression was used as a loading

99 Chapter 4

control. (B) UM-1shCYLD cells were treated for 72h -/+ doxycyclin and subsequently stimulated for 24h -/+ TNF alpha (50ng/mL). Subcellular localization of the p65 NF-ĸB subunit was analyzed by immunoblotting the nuclear and cytoplasmic cell fractions. β-tubulin (cytoplasm) and histone 2B (nucleus) were used as loading controls. (C) left panel: UM-1shCYLDa cells were treated for 72h -/+ doxycyclin and subsequently transfected with a NF-ĸB luciferase reporter and renilla luciferase. 24 hours after transfection cells were treated -/+ TNF alpha (50ng/ mL) for 16 hours. Luciferase activity was corrected for renilla luciferase activity (Rela- tive light units, RLU) and values were normalized to untreated cells. The mean ± SD of a representative experiment performed in triplicate is shown. *** indicates p value < 0.001, by student’s t-test. right panel: UM-1shCYLDa cells were treated for 72h -/+ doxycyclin and PCR analysis of NF-ĸB target gene (CD74 and RELB) mRNA was performed. HPRT1 was used as an input control. (E) Sorted CYLD or control vector transduced UM-3 cells were stimulated for 24h -/+ TNF alpha (50ng/mL). Subcellular distribution of the p65 NF-ĸB subunit was analyzed by immunoblotting nuclear and cytosolic cell fractions. β-tubulin (cytoplasm) and lamin A/C (nucleus) were used as loading control. (F) Sorted CYLD or control vector transduced UM-3 cells were transfected with a NF-ĸB luciferase reporter and renilla luciferase. 24 hours after transfection cells were treated -/+ TNF alpha (50ng/mL) for 16 hours. Luciferase activity was corrected for renilla luciferase activity (Relative light units, RLU) and values were normalized to untreated cells. The mean ± SD of a representative experiment performed in triplicate is shown. *** indicates p value < 0.001, ns-not significant, by student’s t-test.

Silencing of CYLD in UM-1 resulted in a strongly enhanced NF-ĸB signaling: CYLD knockdown increased nuclear localization of the NF-ĸB subunit p65/RelA, which was further enhanced by activating NF-ĸB signaling with TNFα, mimicking paracrine signaling. (Figure 4B). Furthermore, CYLD silencing strongly enhanced NF-ĸB luciferase reporter activity in both basal and TNFα-stimulated conditions (Figure 4C), and resulted in upregulation of the NF-ĸB target genes CD74 and RelB (Figure 4D).12 Conversely, introduction of CYLD in UM-3 cells inhibited NF-ĸB signaling as evidenced by decreased basal and TNFα-induced levels of nuclear p65/RelA as well as by strongly reduced NF-ĸB reporter activity (Figure 4E and F). Hence, our data show that CYLD regulates NF-ĸB signaling in MM. Interestingly, in addition to affecting NF-ĸB signaling, silencing of CYLD also profoundly affected Wnt/β-catenin signaling: CYLD knockdown resulted in increased basal levels of nuclear β-catenin, a hallmark of active Wnt/β-catenin signaling (Figure 5A). Furthermore, stimulation with the canonical Wnt ligand Wnt3a further boosted levels of β-catenin (Figure 5A). In addition, CYLD knockdown led to increased basal and Wnt3a- stimulated TCF4/β-catenin-mediated luciferase reporter activity and strong upregulation of the Wnt target gene cyclin D1 (Figure 5C). Conversely, introduction of CYLD in UM-3 cells resulted in a suppression of Wnt/β-catenin signaling, evidenced by decreased levels of nuclear β-catenin (Figure 5D) and decreased Wnt reporter activity (Figure 5E). Taken together, these these data demonstrate that loss of CYLD has profound effects on both NF-ĸB and Wnt/β-catenin signaling in MM cells and support

100 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma… the notion that CYLD loss during disease progression can enhance sensitivity of MM cells to NF-ĸB-stimuli and Wnt ligands provided by the bone marrow microenvironment.

3 4 5 6 7

Figure 5. Silencing of CYLD expression promotes Wnt signaling in malignant plasma cells (A) UM-1shCYLDa cells were treated for 72h -/+ doxycyclin and subsequently stimulated for 24h with Wnt3a or control conditioned medium. Subcellular localization of β-catenin was analyzed by immunoblotting nuclear and cytoplasmic cell fractions. β-tubulin (cytoplasm) and histone 2B (nucleus) were used as loading control. (B) UM-1shCYLDa cells were treated for 72h with -/+doxycyclin and subsequently transfected with a β-catenin driven Wnt luciferase reporter (TOPFlash) or a control reporter containing mutated β-catenin binding sites (FOPFlash) together with renilla luciferase. 24 hours after transfection, cells were treated

101 Chapter 4

with Wnt3a or control conditioned medium for 24 hours. Luciferase values were corrected for renilla luciferase (relative light units, RLU). Wnt activity is presented as TOPFlash values divided by FOPFlash values. The mean ± SD of a representative experiment performed in triplicate is shown. *indicates p value < 0.05 ***indicates p value < 0.001. by student’s t-test. (C) UM-1shCYLDa cells were treated for 72h -/+ doxycycline and subsequently stimulated for 24h with Wnt3a conditioned medium or control conditioned medium. mRNA expression of the Wnt target gene CCND1 was analyzed by qPCR. TBP was used as an input control. (D) Sorted CYLD or control vector transduced UM-3 cells were stimulated for 24h with Wnt3a or control conditioned medium. Subcellular distribution of β-catenin was analyzed by immunoblotting nuclear and cytosolic cell fractions. β-tubulin (cytoplasm) and lamin A/C (nucleus) were used as loading control. (E) Sorted CYLD or control vector transduced UM-3 cells were transfected with a β-catenin driven Wnt luciferase reporter (TOPFlash) or a control reporter containing mutated β-catenin binding sites (FOPFlash) together with renilla luciferase. 24 hours after transfection, cells were treated with Wnt3a or control conditioned medium for 24 hours. Luciferase values were corrected for renilla luciferase (relative light units, RLU). Wnt activity is presented as TOPFlash values divided by FOPFlash values. The mean ± SD of a representative experiment performed in triplicate is shown. *indicates p value < 0.05 ***indicates p value < 0.001 by student’s t-test. loss of cyld in primary mms is associated with a wnt gene-expression signature and with an unfavorable prognosis The above findings suggest that loss of CYLD can unleash NF-ĸB and/or Wnt signaling during MM progression. To explore this notion, we analyzed the relation between CYLD mRNA expression and the strength of Wnt and NF-ĸB gene-expression signatures in MM patients from the TT2/TT3 study groups (Figure 6).22 Interestingly, as shown in Figure 6A (left panel), CYLD expression showed a significant negative correlation to the Wnt profile, whereas it was positively correlated to the expression NF-ĸB profile genes. Furthermore, analysis of the individual molecular MM subgroups revealed that the proliferation (PR) subgroup, characterized by low CYLD expression (Figure 2B), displayed a significantly stronger Wnt signature than the other molecular subgroups (Figure 6B, left panel), while the NF-ĸB signature in the PR subgroup was weak (Figure 6B, right panel). Thus, the PR group is characterized by active Wnt signaling, but shows reduced NF-ĸB signaling, suggesting that Wnt signaling might drive proliferation. Consistent with this notion, we found a strong positive correlation (R = 0.9, P < 0.001) between the GEP-defined proliferation index (PI) and the Wnt expression signature (Figure 6C, left panel). By contrast, the PI showed a weak but significant negative correlation to the NF-ĸB profile (Figure 6C, right panel). Together, these findings suggest a scenario in which loss of CYLD enhances autocrine and/or paracrine Wnt signaling, leading to enhanced proliferation and tumor progression.

102 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

3 4 5 6 Figure 6. Low CYLD expression is associated with a Wnt gene-expression signature and a proliferation phenotype 7 (A) left panel: Correlation between Wnt profile and CYLD expression in the TT2/TT3 MM patient data set. The proliferation subgroup (PR) is marked in red. Spearman’s rank correlation coefficient = -0.3, p < 0.001, right panel: Correlation between NF-ĸB profile and CYLD expression in the TT2/TT3 MM patient data set. The proliferation subgroup (PR) is marked in red. Spearman’s rank correlation coefficient = 0.2, p < 0.001, (B) Wnt (left panel) and NF-ĸB (right panel) expression profile in 7 different genetic MM subgroups of the TT2/TT3 patient data set. The Wnt expression profile was significantly higher in the proliferation subgroup (PR) in comparison to other subgroups, whereas the NF-ĸB expression profile was significantly lower in the PR subgroup in comparison to others subgroups ***indicates p value < 0.001, by Kruskal Walis test with Dunn correction for multiple comparison. (C) left panel: Correlation between Wnt expression profile and proliferation index (left panel) in the TT2/TT3 MM patient data set. The proliferation subgroup (PR) is marked in red. Spearman’s rank correlation coefficient = 0.9, p < 0.001, right panel: Correlation between NF-ĸB profile and proliferation index in the TT2/TT3 MM patient data set. The proliferation subgroup (PR) is marked in red. Spearman’s rank correlation coefficient = -0.2, p < 0.05.

103 Chapter 4

Our finding that CYLD is a negative regulator of Wnt signaling in MM and that loss of CYLD expression is associated with a Wnt signature and enhanced proliferation, suggests a relation to disease outcome. To address this hypothesis, we studied the relation between CYLD expression and disease outcome in a large group of newly diagnosed MM patients (HOVON65 MM dataset, n = 328). For analysis, the cohort was divided into a CYLD high and a CYLD low group determined by median CYLD expression as cut-off. As shown in Figure 7, there was a significant correlation between low expression of CYLD and inferior overall survival (P < 0.01) and progression-free survival (p < 0.05). Hence, loss of CYLD expression in newly diagnosed MM patients is associated with an unfavorable prognosis.

Figure 7. Low CYLD expression is associated with poor prognosis of MM patients The HOVON-65 micro-array dataset containing newly diagnosed MM patients was used to correlate CYLD expression (median highest and median lowest) to disease outcome. Kaplan- Meier curves are shown for overall survival (OS, upper panel) and progression free survival (PFS, lower panel). Low expression of CYLD is associated with inferior OS and PFS in MM patients, p < 0.05 and p < 0.01 by the log-rank test.

104 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

Discussion

CYLD was originally identified as a tumor suppressor gene mutated in familial cylindromatosis8 and has subsequently been implicated in the pathogenesis of several cancers.9–11 In MM, loss of CYLD is among the most common genomic aberrations, suggesting an important role in the pathogenesis of MM.12,13 In this study, we confirm previous reports showing that CYLD is lost in a proportion of MM patients.7,12,13 In addition, we show that CYLD expression is highly variable in both primary MMs and HMCLs, with evidence of transcriptional as well as post-transcriptional regulation (Figure 1). In malignant melanoma and T-ALL, down-regulation of CYLD has been shown to involve transcriptional repression mediated by SNAIL and the Notch/Hes1 pathway, respectively.9,11 Furthermore, CYLD expression and function can also be subject of epigenetic regulation, involving phosphorylation or regulation by microRNAs.28,29 Further studies are needed to define the exact mechanisms of CYLD (de)regulation in MM. 1 Analysis of CYLD expression in healthy donor plasma cells, MGUS, smoldering MM, and full blown MMs revealed that CYLD is lost in a subset of MM patients, but not MGUS patients, suggesting an instrumental role of CYLD loss in disease progression. In line with this notion, CYLD expression was signifi- 3 cantly reduced in the proliferation (PR) molecular subgroup (Figure 2B), which is characterized by a high proliferation rate and an unfavorable prognosis, pre- 4 sumably reflecting disease progression.22 Furthermore, low expression of CYLD was correlated with inferior progression-free and overall survival in a large co- 5 hort of primary MMs (Figure 7). Conceivably, down-regulation or complete loss of CYLD plays an instrumental role in this aggressive behavior. In support of this 6 notion, introduction of CYLD into UM-3 cells, which lack CYLD expression due to a homozygous deletion (Figure 3A, B) and thereby model a recurrent genomic 7 aberration in primary MMs12,13, resulted in growth inhibition (Figure 3C), increased cell death, and a decreased proportion of cycling cells (Figure 3D).These findings provide prove of concept for a functional role of CYLD in the control of MM growth and survival. We observed that loss of CYLD has profound effects on both NF-ĸB and Wnt/β-catenin signaling in MM cells by sensitizing MM cells to NF-ĸB as well as to Wnt ligands (Figures 4 and 5). CYLD is an established regulator of NF-ĸB signaling that removes K63-linked polyubiquitin chains from proteins involved in the NF-ĸB signaling cascade.14–18 In this way, it acts as a negative regulator of NF-ĸB signaling in various biological processes including lymphoid development, inflammation, and cancer.27 Our current finding that CYLD regulates basal

105 Chapter 4

as well as autocrine/paracrine NF-ĸB activity represents the first functional evidence that CYLD is also involved in the regulation of NF-ĸB signaling in MM (Figure 4). Interestingly, however, we observed that CYLD also functions as a potent negative regulator of Wnt signaling in MM cells, controlling both basal and ligand-induced Wnt/β-catenin signaling (Figure 5). This finding cor- roborates a recent study, showing that CYLD can control Wnt signaling through deubiquitination of Dishevelled (Dvl), an adapter protein which transmits proximal Wnt signals and is dependent on K63-linked ubiquitination for its function.19 These findings imply that down-regulation or loss of CYLD expression potentially increases the sensitivity of MM cells to both NF-ĸB-stimuli and Wnt ligands expressed in the BM microenvironment, thus unleashing NF-ĸB as well as Wnt signaling. However, although we indeed found a significant negative correlation between CYLD expression and expression of Wnt target genes in primary MM samples, we found a (weak) positive correlation to the NF-ĸB signature (Fig- ure 6A). This suggests that loss of CYLD predominantly activates Wnt signaling in MM in vivo. Consistent with this scenario the PR subgroup, which is characterized by low CYLD expression (Figure 2B), displayed a significantly stronger Wnt but weaker NF-ĸB signature as compared to all other molecular subgroups (Fig- ure 6B). A possible explanation for the positive correlation between CYLD expression and NF-ĸB activation is that CYLD acts as a classical negative feedback target, however, we did not observe upregulation of CYLD after NF-ĸB activation in vitro (data not shown). Taken together, these findings support a scenario in which loss of CYLD enhances autocrine and/or paracrine Wnt signaling, promoting proliferation and mediating tumor progression. Indeed, the strong positive correlation (R = 0.9, P < 0.001) between the GEP-defined proliferation index (PI) and the Wnt expression signature (Figure 6C) suggests that Wnt signaling is instrumental in driving MM proliferation in vivo. Previous studies from our own and other laboratories have demonstrated that aberrant activation of the canonical Wnt pathway is common in advanced MM and plays a potentially important role in the pathogenesis of MM by promoting cell growth, survival, and drug-resistance.21,24,30,31 Although activation of Wnt signaling in human cancer is typically caused by mutations in the adenomatous polyposis coli gene (APC) or in β-catenin (CTNNB1), resulting in ligand-independent Wnt pathway activation, these mutations are absent in MM. Instead, Wnt pathway activation in MM involves stimulation of an intact Wnt signaling pathway by autocrine and/or paracrine Wnt ligands, which are (over)expressed by by stromal cells in the MM microenvironment.21,30–32 Interestingly, in addition to CYLD loss or down regulation, silencing of other negative regulators of Wnt

106 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma… signaling also appears to play a role in MM progression. Promoter methylation of Wnt antagonists, including WIF1, DKK3, SFRP1, -2, -4, and -5 has been reported in MM and was associated with disease progression.33,34 Furthermore, in a recent study, we demonstrated that DKK1, a major negative feed-back regulator of Wnt-signaling, is also a target of promoter methylation, especially in advanced stage MM.31 Our current findings that CYLD acts as a negative regulator of MM growth and Wnt/β-catenin signaling and that low CYLD expression in primary MMs is associated with a Wnt signaling and proliferative gene-expression signature, strongly suggest that CYLD down-regulation enhances MM aggressiveness through a mechanism involving Wnt pathway activation. Presumably, Wnt/β-catenin signaling-mediated effects on the cell cycle via AurKA/B may play a role in these pathogenic effects.24 The fact that Wnt signaling in MM is mainly ligand-driven highlights a possible therapeutic role for drugs interfering with proximal Wnt signaling. Recent studies showed potent anti-tumor activity of LGK974 in pancreatic cancer35, which targets the acyltransferase Porcupine 1 (PORCN) that is essential for Wnt secretion and may therefore be an interesting new option in the treatment of MM. 3 References 4 1. Hideshima T, Mitsiades C, Tonon G, Richardson PG, Anderson KC. Understand- ing multiple myeloma pathogenesis in the bone marrow to identify new therapeutic 5 targets. Nat Rev Cancer 2007; 7(8): 585–598. 2. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host inter- 6 actions. Nat Rev Cancer 2002; 2(3): 175–187. 3. Kumar S, Fonseca R, Ketterling RP, Dispenzieri A, Lacy MQ, et al. Trisomies in multiple myeloma: impact on survival in patients with high-risk cytogenetics. Blood 2013; 7 119(9): 2100–2105. 4. Chng WJ, Glebov O, Bergsagel PL, Kuehl WM. Genetic events in the pathogenesis of multiple myeloma. Best Pract Res Clin Haematol 2007; 20(4): 571–596. 5. Kyle RA, Rajkumar SV. Criteria for diagnosis, staging, risk stratification and response assessment of multiple myeloma. Leukemia 2009; 23(1): 3–9. 6. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009; 374(9686): 324–339. 7. Jenner MW, Leone PE, Walker BA, Ross FM, Johnson DC, et al. Gene mapping and expression analysis of 16q loss of heterozygosity identifies WWOX and CYLD as being important in determining clinical outcome in multiple myeloma. Blood 2007; 110(9): 3291–3300.

107 Chapter 4

8. Bignell GR, Warren W, Seal S, Takahashi M, Rapley E, et al. Identification of the familial cylindromatosis tumour-suppressor gene. Nat Genet 2000; 25(2): 160–165. 9. Espinosa L, Cathelin S, D’Altri T, Trimarchi T, Statnikov A, et al. The Notch/Hes1 pathway sustains NF-kappaB activation through CYLD repression in T cell leukemia. Cancer Cell 2010; 18(3): 268–281. 10. Hellerbrand C, Bumes E, Bataille F, Dietmaier W, Massoumi R, et al. Reduced expression of CYLD in human colon and hepatocellular carcinomas. Carcinogenesis 2007; 28(1): 21–27. 11. Massoumi R, Kuphal S, Hellerbrand C, Haas B, Wild P, et al. Down-regulation of CYLD expression by Snail promotes tumor progression in malignant melanoma. J Exp Med 2009; 206(1): 221–232. 12. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12(2): 115–130. 13. Keats JJ, Fonseca R, Chesi M, Schop R, Baker A, et al. Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell 2007; 12(2): 131–144. 14. Brummelkamp TR, Nijman SM, Dirac AM, Bernards R. Loss of the cylindromatosis tumour suppressor inhibits apoptosis by activating NF-kappaB. Nature 2003; 424(6950): 797–801. 15. Kovalenko A, Chable-Bessia C, Cantarella G, Israël A, Wallach D, et al. The tumour suppressor CYLD negatively regulates NF-kappaB signalling by deubiquitination. Nature 2003; 424(6950): 801–805. 16. Massoumi R, Chmielarska K, Hennecke K, Pfeifer A, Fässler R. Cyld inhibits tumor cell proliferation by blocking Bcl-3-dependent NF-kappaB signaling. Cell 2006; 125(4): 665–677. 17. Reiley WW, Jin W, Lee AJ, Wright A, Wu X, et al. Deubiquitinating enzyme CYLD negatively regulates the ubiquitin-dependent kinase Tak1 and prevents abnormal T cell responses. J Exp Med 2007; 204(6): 1475–1485. 18. Wright A, Reiley WW, Chang M, Jin W, Lee AJ, et al. Regulation of early wave of germ cell apoptosis and spermatogenesis by deubiquitinating enzyme CYLD. Dev Cell 2007; 13(5): 705–716. 19. Tauriello DV, Haegebarth A, Kuper I, Edelmann MJ, Henraat M, et al. Loss of the tumor suppressor CYLD enhances Wnt/beta-catenin signaling through K63-linked ubiquitination of Dvl. Mol Cell 2010; 37(5): 607–619. 20. Reijmers RM, Groen RW, Rozemuller H, Kuil A, de Haan-Kramer A, et al. Targeting EXT1 reveals a crucial role for heparan sulfate in the growth of multiple myeloma. Blood 2010; 115(3): 601–604. 21. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci U S A 2004; 101(16): 6122–6127.

108 Loss of CYLD expression unleashes Wnt signaling in multiple myeloma…

22. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108(6): 2020–2028. 23. Sonneveld P, Schmidt-Wolf IG, van der Holt B, El Jarari L, Bertsch U, et al. Bort- ezomib Induction and Maintenance Treatment in Patients With Newly Diagnosed Multiple Myeloma: Results of the Randomized Phase III HOVON-65/ GMMG-HD4 Trial. J Clin Oncol 2012; 30(24): 2946–2955. 24. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/beta-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 25. Alameda JP, Moreno-Maldonado R, Navarro M, Bravo A, Ramírez A, et al. An inactivating CYLD mutation promotes skin tumor progression by conferring enhanced proliferative, survival and angiogenic properties to epidermal cancer cells. Oncogene 2010; 29(50): 6522–6532. 26. An CH, Kim SS, Kang MR, Kim YR, Kim HS, et al. Frameshift mutations of ATBF1, WNT9A, CYLD and PARK2 in gastric and colorectal carcinomas with high micro- satellite instability. Pathology 2010; 42(6): 583–585. 1 27. Massoumi R. CYLD: a deubiquitination enzyme with multiple roles in cancer. Future Oncol 2011; 7(2): 285–297. 28. Hutti JE, Shen RR, Abbott DW, Zhou AY, Sprott KM, et al. Phosphorylation of the tumor suppressor CYLD by the breast cancer oncogene IKKepsilon promotes cell 3 transformation. Mol Cell 2009; 34(4): 461–472. 29. Iliopoulos D, Jaeger SA, Hirsch HA, Bulyk ML, Struhl K. STAT3 activation of miR- 21 and miR-181b-1 via PTEN and CYLD are part of the epigenetic switch linking 4 inflammation to cancer. Mol Cell 2010; 39(4): 493–506. 30. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/TCF 5 transcriptional complex in the treatment of multiple myeloma. Proc Natl Acad Sci U S A 2007; 104(18): 7516–7521. 6 31. Kocemba KA, Groen RW, van Andel H, Kersten MJ, Mahtouk K, et al. Transcriptional silencing of the Wnt-antagonist DKK1 by promoter methylation is associated with 7 enhanced Wnt signaling in advanced multiple myeloma. PLoS One 2012; 7(2): e30359. 32. Van Den Berg DJ, Sharma AK, Bruno E, Hoffman R. Role of members of the Wnt gene family in human hematopoiesis. Blood 1998; 92(9): 3189–3202. 33. Chim CS, Pang R, Fung TK, Choi CL, Liang R. Epigenetic dysregulation of Wnt signaling pathway in multiple myeloma. Leukemia 2007; 21(12): 2527–2536. 34. Jost E, Gezer D, Wilop S, Suzuki H, Herman JG, et al. Epigenetic dysregulation of secreted Frizzled-related proteins in multiple myeloma. Cancer Lett 2009; 281(1): 24–31. 35. Jiang X, Hao HX, Growney JD, Woolfenden S, Bottiglio C, et al. Inactivating mutations of RNF43 confer Wnt dependency in pancreatic ductal adenocarcinoma. Proc Natl Acad Sci U S A 2013; 110(31): 12649–12654.

109

Chapter5

N-cadherin-mediated interaction with multiple myeloma cells inhibits osteoblast differentiation

Richard W.J. Groen1, Martin F.M. de Rooij1, Kinga A. Kocemba1, Rogier M. Reijmers1, Anneke de Haan-Kramer1, Marije B. Overdijk1, Linda Aalders2, Henk Rozemuller3, Anton C.M. Martens2,3, P. Leif Bergsagel4, Marie José Kersten5, Steven T. Pals1 and Marcel Spaargaren1

1Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands,2 Department of Cell Biology, University Medical Center Utrecht, Utrecht, The Netherlands,3 Department of Immunology, University Medical Center Utrecht, Utrecht, The Netherlands,4 Comprehensive Cancer Center, Mayo Clinic Arizona, Scottsdale, Arizona, USA, 5Department of Hematology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Haematologica 2011; 96(11): 1653–1661 abstract background Multiple myeloma is a hematologic malignancy characterized by a clonal expansion of malignant plasma cells in the bone marrow, which is accompanied by the development of osteolytic lesions and/or diffuse osteopenia. The intricate bi- directional interaction with the bone marrow microenvironment plays a critical role in sustaining the growth and survival of myeloma cells during tumor progression. Identification and functional analysis of the (adhesion) molecules involved in this interaction will provide important insights into the pathogenesis of multiple myeloma. design and methods Multiple myeloma cell lines and patients’ samples were analyzed for expression of the adhesion molecule N-cadherin by immunoblotting, flow cytometry, immunofluorescence microscopy, immunohistochemistry and expression microarray. In addition, by means of blocking antibodies and inducible RNA interference we studied the functional consequence of N-cadherin expression for the myeloma cells, by analysis of adhesion, migration and growth, and for the bone marrow microenvironment, by analysis of osteogenic differentiation. results The malignant plasma cells in approximately half of the multiple myeloma patients, belonging to specific genetic subgroups, aberrantly expressed the homophilic adhesion molecule N-cad-herin. N-cadherin-mediated cell-substrate or homotypic cell-cell adhesion did not contribute to myeloma cell growth in vitro. However, N-cadherin directly mediated the bone marrow localization/retention

112 N-cadherin-mediated interaction with multiple myeloma cells… of myeloma cells in vivo, and facilitated a close interaction between myeloma cells and N-cadherin-positive osteoblasts. Furthermore, this N-cadherin-mediated interaction contributed to the ability of myeloma cells to inhibit osteoblastogenesis. conclusions Taken together, our data show that myeloma cells frequently display aberrant expression of N-cadherin and that N-cadherin mediates the interaction of myeloma cells with the bone marrow microenvironment, in particular the osteoblasts. This N-cadherin-mediated interaction inhibits osteoblast differentiation and may play an important role in the pathogenesis of myeloma bone disease.

3 4 5 6 7

113

Introduction

Multiple myeloma (MM) is a neoplasm characterized by clonal expansion of malignant plasma cells. The transition of a normal plasma cell to a fully transformed, aggressive myeloma is considered to be a multistep process, which re- 1 quires the acquisition of chromosomal translocations and mutations in multiple genes.1,2 Most of this evolution takes place in the bone marrow (BM), indicating that the interaction with the BM microenvironment plays a critical role in the pathogenesis of MM. 3 As in normal plasma cell homing, a major player in the recruitment to and retention of MM plasma cells in the BM is the CXCL12/CXCR4 axis.3,4 The 4 chemokine CXCL12 promotes transendothelial migration and induces α4β1 mediated adhesion to VCAM-1 and fibronectin.5 Malignant plasma cells express 5 several cell-surface molecules which mediate either homotypic cell adhesion, e.g. N-CAM, or adhesion with the extracellular matrix or other cells in the BM micro- 6 environment, e.g. the integrins α4β1 and α5β1 and mucin1.1,6 These interactions control long-term survival and proliferation,7 and may also render the MM cells 7 resistant to the pro-apoptotic effects of conventional chemotherapies,6 a process known as cell adhesion-mediated drug resistance (CAM-DR). In addition, these interactions may also affect the BM microenvironment. Under physiological conditions bone remodeling is a continuous process in which osteoclasts mediate resorption of “old” bone tissue, followed by new bone formation by the osteoblasts. These two processes are tightly regulated by signals involving the RANKL-RANK axis, the Wnt signaling pathway, and N-cadherin engagement,8–10 with the last interaction serving an important role in osteoblast function and differentiation. One of the characteristic features of MM is osteolytic bone destruction, resulting from the activation of osteoclasts and inhibition of osteoblast function. Notably, the actions of several factors produced by

115 Chapter 5

MM cells and involved in the deregulation of osteoblast differentiation, such as DKK1,11 sFRP-2,12 and H G F, 13 were shown to be partially dependent on adhesion of myeloma cells to the osteoblasts via integrin α4β1.14 We have previously shown that malignant plasma cells from MM patients overexpress β-catenin, including the non-phosphorylated form, leading to active β-catenin/TCF-mediated transcription and MM cell proliferation.15 Subsequent studies confirmed the presence of Wnt pathway activation in MM and its importance for MM cell proliferation and survival.16,17 Interestingly, our immunohistochemical studies revealed that β-catenin is localized in the nucleus as well as at the plasma membrane, at the site of cell-cell contact. Indeed, in addition to its role as a transcriptional regulator in the Wnt signaling pathway, β-catenin is also a key regulator of cadherin-mediated adhesion.18 Cadherins comprise a family of transmembrane adhesion molecules that mediate calcium-dependent cell-cell adhesion through homophilic interactions. In the “classical” cad-herins, the conserved cytoplasmic domain forms a complex with the catenins p120catenin, β-catenin and α-catenin, which are possible regulators of cadherin function and link it to the cytoskeleton.19 In this study, we show that malignant plasma cells of a subset of MM patients express N-cadherin. We explored the function of N-cadherin expression in MM pathogenesis, focusing on its potential role in bone marrow localization of malignant cells and myeloma bone disease.

Design and Methods cell lines and antibodies MM cell lines were cultured as described previously.15 Further details on the cell lines, culture conditions, and antibodies are provided in the Supplementary De- sign and Methods. A doxycycline-inducible NCI-H929 MM cell line was generated as described previously,20 using the T-REx™ System (Invitrogen Life Technologies), containing the TET repressor only (H929 TR), or combined with a specific shRNA against CDH2, 5′-GAGCCT-GAAGCCAACCTTA-3′ (H929 shCDH2). N-cadherin knockdown was obtained by incubating NCI-H929 shCDH2 cells for 5 days with 0.2 μg/mL doxycycline. cell growth assessment Cells were plated (5 × 103) in a 96-well plate coated with recombinant N-cadherin/ Fc chimera (1 μg/mL; R&D Systems, Abingdon, UK), or BSA as control. Further

116 N-cadherin-mediated interaction with multiple myeloma cells… details on the cell growth assessment assays are provided in the Supplementary Design and Methods. immunofluorescent microscopy, immunohistochemistry and flow cytometry N-cadherin (mouse mAb clone 32) and β-catenin (rabbit pAb H-102) expression in MM cell lines was analyzed by immunofluorescence microscopy on paraform- aldehyde-fixed cytospins. Expression was detected, using Alexa488-conjugated goat anti-mouse and Alexa568-conjugated goat anti-rabbit as secondary antibodies, and analyzed by confocal laser scan microscopy (CLSM). Immunohistochemical staining was performed on formalin-fixed, plastic-embedded tissue sections as described elsewhere.15 Further details are provided in the Supplementary Design and Methods. Flow cytometry analysis of N-cadherin expression was determined by incubation of cells with an anti-N-cadherin monoclonal antibody (clone GC-4, Sigma- 1 Aldrich) followed by biotinylated goat anti-mouse IgG1 (Southern Biotechnology, Birmingham, AL, USA) and subsequently RPE-conjugated streptavidin (DAKO). Analysis was carried out on a FACScalibur flow cytometer (BD Biosciences) with CellQuestTM software (BD Biosciences). 3 immunoprecipitation, immunoblot analysis, 4 cell adhesion and migration assays Immunoprecipitation, western blotting, adhesion and migration assays were per- 5 formed essentially as described previously.21,22 Further details on these assays are provided in the Supplementary Design and Methods. 6 sample preparation and microarray 7 hybridization and analysis Isolation of plasma cells and profiling of RNA were performed as described previously.23 Further details are provided in the Supplementary Design and Methods. homing assay H929 shCDH2 cells were incubated for 5 days with (KD) or without (WT) 0.2 μg/mL doxycycline. Cells were labeled with either 0.25 μM Cell tracker Green (CMFDA; Invitrogen) or 2 μM Cell tracker Orange (CMTMR; Invitrogen), mixed (1:1), and 15 × 106 cells were injected intravenously into RAG2−/− γc−/− mice. Each WT/KD combination was analyzed by adoptive transfer of eight recipient mice, and included a dye-swap. After 24 h, blood and bone marrow were collected and

117 Chapter 5

analyzed by FACS to quantify dye-labeled cells. The percentage of homing cells was corrected for the input ratio. All animal experiments were conducted according to the Institutional Guidelines of the University Medical Center Utrecht, after acquiring permission from the local Ethical Committee for Animal Experimenta- tion and in accordance with current Dutch laws on animal experiments. osteoblast differentiation KS483 cells and primary human mesenchymal stromal cells were cultured and differentiated essentially as described elsewhere,24,25 in the absence or presence of MM cells. Further details are provided in the Supplementary Design and Methods. real-time reverse transcription polymerase chain reaction RNA isolation and cDNA synthesis were performed as described previously.15 The quantitative reverse transcription-polymerase chain reaction (qRT-PCR) runs were performed on a Roche LightCycler 1.5 using FastStart DNA Mas- ter SYBR Green I kit (Roche, Basel, Switzerland). Results were analyzed using LinReg PCR analysis software (version 7.526). Expression was normalized over β-2-microglobulin expression. Further details are provided in the Supplementary Design and Methods and Supplementary Tables 1 and 2. statistical analysis The unpaired two-tailed Student’s t-test was used to determine the statistical significance of differences between means, unless otherwise stated. *P< 0.05; **P < 0.01; ***P < 0.001; ns: not significant.

Results multiple myeloma cells express a functional β-catenin/n-cadherin complex at the plasma membrane We previously reported that MM plasma cells over-express β-catenin. Stimulation of Wnt signaling with either the glycogen synthase kinase-3β (GSK3β) inhibitor LiCl or Wnt3a led to further accumulation and nuclear localization of β-catenin, resulting in enhanced TCF-mediated transcription and increased cell proliferation.15 Interestingly, however, β-catenin was not only localized in the nucleus of the MM cells, but was also frequently observed at the plasma membrane, at the

118 N-cadherin-mediated interaction with multiple myeloma cells… cell-cell contact sites between adjacent MM cells (Figure 1A). This observation suggested that β-catenin could also be involved in intercellular adhesion, presumably through interaction with a classical cadherin.18 To explore this hypothesis, we screened a panel of MM cell lines for expression of cadherins. Western blot analysis revealed that, while none of the myeloma cell lines tested expressed E-cadherin, four of seven were N-cadherin positive (Figure 1B). This expression was confirmed by FACS analysis (Supplementary Figure S1). Notably, the cell lines with most prevalent expression of N-cadherin also displayed the highest levels of β-catenin (Figure 1B). Furthermore, confocal laser scan microscopy revealed co-localization of N-cadherin and β-catenin at the cell-cell junctions between adjacent MM cells, suggesting a physical interaction between the two proteins (Figure 1C). Indeed, co-precipitation of N-cadherin and β-catenin confirmed the existence of an association between these molecules at the MM plasma membrane, with the majority of N-cadherin attached to β-catenin (Figure 1D). 1

3 4 5 6 7

Figure 1. Multiple myeloma cells express N-cadherin (A) β-catenin is expressed in MM cells at the cell-cell junctions. The MM cell line OPM-1 is stained with a pAb H-102 against β-catenin and detected with Alexa568-conjugated goat anti-rabbit by confocal laser scan microscopy (CLSM). (B) Cadherin expression in MM cell lines. Cell lysates were immunoblotted using monoclonal antibodies against E-cadherin (HECD-1), N-cadherin (clone 32) and β-catenin (clone 14). The breast carcinoma cell line T47D and the malignant glioma cell line U251 MG were used as positive controls for E-cadherin and N-cadherin, respectively. β-actin was used as the loading control. (C) N-cadherin co-localizes with β-catenin at the cell-cell contacts. OPM-1 cells were double- stained with an antibody against N-cadherin (clone 32) and β-catenin (H-102), followed by secondary antibodies Alexa488-conjugated goat anti-mouse and Alexa568-conjugated goat anti-rabbit. Expres- sion of N-cadherin (green; left panel) and β-catenin (red; middle panel) and co-localization (orange; right panel) was detected by CLSM. (D) Co-immunoprecipitation of N-cadherin with β-catenin. Cell

119 Chapter 5

lysates of the MM cell line OPM-1 were immunoprecipitated (IP) with monoclonal antibody clone 32 (N-cadherin), clone 14 (β-catenin), or control IgG1 antibody, pre-coupled to Protein G-Sepharose beads. Western blots (WB) were stained with anti-N-cadherin (clone 32) and anti-β-catenin (clone 14). As a specificity control, mock incubations without lysates, were immunoblotted and stained. Total lysates (TL) were immunoblotted and stained with anti-β-catenin (clone 14), as an input control.

Classical cadherins, including N-cadherin, mediate cell-cell adhesion via homophilic interaction. To assess the functional integrity of the N-cadherin/ β-catenin complexes on the MM cells we tested the ability of N-cadherin-expressing MM cells to adhere to an N-cadherin-coated surface. As shown in Figure 2A, the N-cadherin-positive MM cell lines OPM-1 and NCI-H929 specifically adhered to recombinant N-cadherin, an effect which could be prevented completely by the N-cadherin blocking antibody (GC-4). Notably, the N-cadherin-negative MM cell lines L363 and UM3 did not show specific binding to N-cadherin (data not shown). CD138-positive MACS-purified primary MM cells, expressing N-cadherin, also showed specific adhesion to coated N-cadherin (Figure 2B), whereas primary MM cells without N-cadherin did not adhere (data not shown).

Figure 2. N-cadherin-mediated adhesion of MM cells (A) Adhesion of MM cells to recombinant N-cadherin. OPM-1 (left panel) and NCI-H929 (right panel) cells adhere to a surface coated with recombinant N-cadherin (1 μg/mL), compared to BSA-coating as a negative control. This adhesion could be prevented by an N-cadherin blocking antibody, GC-4 (10 μg/mL). (B) Adhesion of primary patients’ material to recombinant N-cadherin. Primary myeloma cells (pMM) were MACS purified using CD138-coated beads (> 95% pure) and were allowed to adhere to a surface coated with recombinant N-cadherin (1 μg/mL), compared to BSA-coating as a negative control. expression of n-cadherin in primary multiple myelomas To assess the prevalence of N-cadherin expression in primary MM, in purified plasma cells from a panel of 559 MM patients, using Affymetrix oligonucleotide microarrays we analyzed N-cadherin expression in relation to characteristic recurrent chromosomal translocations and expression of cyclin D1 and D2 (TC groups).23 As depicted, CDH2, the gene encoding N-cadherin, was highly expressed (> 2x) in 83% of the MM samples bearing the t(4;14)(p16;q32) translo-

120 N-cadherin-mediated interaction with multiple myeloma cells… cation involving MMSET. Furthermore, the subgroups characterized by high expression of cyclin D1, either alone (D1) or together with high expression of cyclin D2 (D1 + D2), consisted of two distinct populations: one with high and one with low expression of CDH2. Expression of CDH2 was less prevalent in samples with the translocations involving 11q13, 6p21, or MAF, and in the subgroup characterized by low cyclin D1 but high cyclin D2 expression (D2) (Figure 3A, left panel). Although the 4p16 translocation is known to correlate with poor prognosis, no independent prognostic value could be found for N-cad-herin expression (data not shown). Importantly, analysis of the freely accessible microarray data of the Lambert Laboratory revealed that expression of N-cadherin is absent/low in normal bone marrow plasma cells from healthy donors, whereas expression is weakly elevated in MGUS (Figure 3A, right panel). In agreement with our data, analysis of CDH2 expression in this data set, in which the major myeloma subtypes are defined by gene expression profile-derived classification,27 revealed high expression of CDH2 in over 90% of the MMSET expression subgroup (MS) characterized 1 by the 4p16 MMSET translocation, and low expression in the MAF expression subgroup (MF) characterized by MAF/MAFB translocations. Furthermore, the hyperdiploid subgroup (HY) revealed distinct populations with either high or low CDH2 expression, which is in line with the high concordance of the HY subgroup 3 with our D1 subgroup.28 Consistent with the mRNA expression data, immunohistochemical study of 4 bone marrow biopsies of MM patients (n = 43) demonstrated N-cadherin protein expression in the malignant cells of approximately 50% of the patients (Figure 3B). 5 Besides membrane expression, several of these tumors displayed strong cytoplasmic N-cadherin staining. As in the MM cell lines (Figure 1C), N-cadherin and 6 β-catenin in the primary MM often localized at the cell-cell junctions between adjacent MM cells (Figure 3B), and between MM cells and the bone-lining cells 7 (Figure 3C). Our observations identify N-cadherin as a myeloma-associated protein displaying deregulated expression in a subset of MM. n-cadherin-mediated adhesion does not affect multiple myeloma growth Since N-cadherin expression has been described to promote survival19 and to suppress cell proliferation in other cell types,29 we examined the role of both heterotypic as well as homotypic N-cadherin-mediated adhesion in MM growth. The direct effect of heterotypic adhesion was mimicked by seeding MM cell lines, with different levels of N-cadherin expression (Figure 1B), on recombinant N-cadherin and monitoring the growth for 4 days. Although the cells of the N-cadherin-

121 Chapter 5 expressing cell lines essentially grew as single cells on the N-cadherin coating as compared to the formation of cell aggregates on the BSA coating (Supplementary Figure S2A), no differences in growth rate were observed (Supplementary Figure S2B). Similarly, no difference in cell proliferation was observed, as determined by 3H-thymidine incorporation (data not shown). In addition, the survival of MM cells in a single cell growth assay was not altered by culturing on a recombinant N-cadherin-coated surface (data not shown).

Figure 3. Expression of N-cadherin in primary MM (A) Affymetrix expression profiles of N-cadherin in MM. Gene expression of 559 newly diagnosed MM patients was measured by U133 Plus2.0 Affymetrix oligonucleotide microarray probeset 203440_at, summarized with MAS5, median normalized and plotted against chromosomal translocations/cyclin D expression (left panel); and publicly available genetic MM data of 345 MM patients from the total therapy 2 (TT2) patient set were plotted against disease progression and genetic profiles (right panel). The expression of N-cadherin by plasma cells of patients in the MMSET expression (MS) and the hyperdiploid (HY) subgroups was statistically higher than that by normal bone marrow plasma cells (BMPC) from healthy donors (P < 0.001 by the Kruskal-Wallis test). (B) Expression of N-cadherin in primary MM. Consecutive sections of plastic-embedded BM-biopsies of MM patients were immunohistochemi- cally stained with antibodies against CD138 (B-B4; left panels), N-cadherin (clone 32; middle panels), and β-catenin (clone 14; right panels). Original magnifications x64. (C) Immunohistochemical stainings of N-cadherin and β-catenin expression on MM cells adjacent to N-cadherin positive bone lining osteoblasts. N-cadherin is often localized between MM cells and the bone-lining cells (arrowhead). Plastic- embedded sections from a MM patient stained with antibodies against β-catenin (clone 14; upper panel) and N-cadherin (clone 32; lower panel) are shown. Original magnification x 100.

122 N-cadherin-mediated interaction with multiple myeloma cells…

To examine the contribution of the homotypic adhesion to the growth of MM cells, we cultured MM cells in the presence of an N-cadherin blocking antibody. However, addition of antibodies to the culture did not result in altered proliferation (data not shown). To determine the direct effect of N-cadherin expression on MM growth, we generated NCI-H929 cells stably transfected with a doxycycline- inducible shRNA against CDH2 (H929 shCDH2). As shown in Supplementary Figure S2C, doxycycline treatment of the cells for 5 days resulted in a 70% reduction of N-cadherin expression. Assessment of the growth of the H929 shCDH2 cells in comparison to the negative control H929 TR cells, containing the TET repressor but not the CDH2 shRNA, showed that the doxycycline-induced knockdown of N-cadherin did not result in an aberrant growth pattern (Supple- mentary Figure S2D). Performing similar experiments on a N-cadherin coating also did not result in a difference in the growth rate of the cells (Supplementary Figure S2E). Collectively, these data show that neither N-cadherin expression as such, nor N-cadherin-mediated cell-substrate or homotypic cell-cell adhesion, 1 directly affects MM cell growth in vitro. n-cadherin plays a role in the retention of multiple myeloma cells in the bone marrow 3 Since N-cadherin has been implicated in migration and tumor metastasis,19,30,31 we investigated whether N-cadherin plays a role in MM cell homing. H929 4 shCDH2 cells were incubated with or without doxycycline for 5 days, and subsequently fluorescently labeled with either CMFDA or CMTMR. Cells were mixed 5 (1:1), and injected intravenously into Rag-2−/−γc−/− mice. Analysis of blood and bone marrow of these mice (n = 4) revealed that reduced N-cadherin expres- 6 sion resulted in higher levels of circulating cells and reduced homing to the bone marrow (Figure 4A), with a dye swap (n = 4) resulting in similar results (data 7 not shown). In support of these results, both basal motility (Figure 4B) and SDF- 1-induced migration (Figure 4C) of the N-cad-herin-expressing OPM-1 cells was potentiated in transwell migration assays when the transwells were coated with recombinant N-cadherin (from 9% to 15%, and from 45% to 80%, respectively), whereas migration of the N-cadherin-negative L363 cells was not affected. Since endothelial cells express N-cadherin, we further explored a possible role for N-cadherin in MM cell homing by transendothelial migration assays using HUVEC cells. However, in contrast to the strong inhibition observed with an α4-integrin blocking antibody, neither blocking nor knockdown of N-cadherin affected transendothelial migration towards SDF-1 (Figure 4D). Taken together, our results establish the importance of N-cadherin in localization of MM cells

123 Chapter 5 in the bone marrow, most likely reflecting a role for N-cadherin in bone marrow retention rather than in active homing of MM cells.

Figure 4. N-cadherin-mediated retention of MM cells in the bone marrow (A) H929 shCDH2 cells were incubated with or without doxycycline for 5 days, and subsequently labeled with either CMFDA or CMTMR. Cells were mixed (1:1), and injected intravenously into Rag-2−/− γc−/− mice. Each CMFDA/CMTMR combination was analyzed by adoptive transfer of eight recipient mice, which included a dye-swap. After 24 h, blood and bone marrow were collected and FACS analyzed to quantify dye-labeled cells. The percentage of homing cells was corrected for the input ratio. The bars represent the means ± SD for four mice. (B) MM cells were allowed to migrate for 4 h in the absence of SDF-1 in transwells coated with recombinant N-cadherin (1 μg/mL), compared to BSA-coating as a negative control. (C) MM cells were allowed to migrate for 4 h towards 100 ng/mL SDF-1 in transwells coated with recombinant N-cadherin (1 μg/mL), or sVCAM-1 (1 μg/mL) compared to BSA-coating as a negative control. (D) H929 shCDH2 cells were allowed to migrate for 5 h towards 100 ng/mL SDF-1 over an endothelial monolayer. H929 shCDH2 cells were cultured in the presence (+dox) or absence (-dox) of doxycycline for 5 days before migration (left panel). In addition to the doxycycline-treatment, MM cells were coated with blocking antibodies against N-cadherin (GC-4) or α4-integrin (HP2/1) before migration (right panel). n-cadherin-mediated interaction with multiple myeloma cells suppresses osteoblast differentiation MM-related osteolytic bone disease is caused by an imbalance between osteoblast and osteoclast activity. MM cells have been shown to suppress osteoblast differentiation and activity via at least two different mechanisms, i.e., by secreting soluble factors11–13 and by direct cell-cell contact with osteoblasts.14 This,

124 N-cadherin-mediated interaction with multiple myeloma cells… combined with the observations that N-cadherin-positive MM cells reside in close proximity to the osteoblasts in the bone marrow of MM patients (Fig- ure 3C) and that N-cadherin-mediated adhesion plays an important role in osteoblast maturation,9,32,33 prompted us to explore the possible contribution of N-cadherin-mediated adhesion to the contact-dependent suppression of osteoblast differentiation by MM cells. After confirming the high N-cadherin expression by pre-osteoblastic cells (Figure 5A), we determined whether MM cells could adhere to these cells in an N-cadherin-dependent manner. To avoid integrin-mediated adhesion, experiments were performed in the presence of calcium as the only divalent cation. Indeed, more than 60% of the N-cadherin-positive MM cells adhered to the osteoblasts (Supplementary Figure S3A, left panels), and, moreover, this adhesion could be blocked by pre-incubation of the MM cells with an antibody that blocked N-cadherin function (Supplementary Figure S3A, right panels and Figure 5B). This heterotypic cell-cell interaction was further investigated using 1 doxycycline-inducible H929 shCDH2 cells, which upon doxycycline-treatment displayed an approximately 70% reduction of N-cadherin expression (Supple- mentary Figure S2C). In line with the blocking antibody results (Supplementary Figure S3A and Figure 5B), these cells showed diminished adhesion to osteo- 3 blasts upon silencing of N-cadherin expression, whereas no difference in adhesion was observed with the control H929 TR cells (Supplementary Figure S3B 4 and Figure 5C). To investigate the effect of N-cadherin-mediated MM adhesion on osteoblast 5 differentiation, the doxycycline-inducible cells were co-cultured with murine KS483 pre-osteoblastic cells which, upon reaching confluence and the addition 6 of ascorbic acid, differentiate into mature osteoblasts expressing alkaline phosphatase (ALP). Co-cultures of KS483 cells with either H929 shCDH2 cells or 7 H929 TR cells resulted in a strong inhibition of ALP activity (Figure 5D). In- terestingly, doxycycline-induced knockdown of N-cadherin markedly attenuated the ability of H929 shCDH2 cells to inhibit osteoblast differentiation, whereas doxycycline treatment of the control H929 TR cells had no effect (Figure 5D). The inhibitory effect of this N-cadherin-mediated interaction on osteoblast differentiation was further substantiated by measuring the mRNA levels of the early osteogenic markers Akp2 and Col1a1 and the late marker Bglap (Figure 5E), encoding alkaline phosphatase, collagen type I, alpha1 and osteocalcin, respectively. As for ALP activity, the ability of MM cells to inhibit the expression of Akp2, Col1a1 and Bglap was significantly diminished upon N-cadherin knockdown (Figure 5E), whereas no significant change was observed in the expression of

125 Chapter 5 the (pre-)osteogenic transcription factors Osx and Runx2 (Supplementary Fig- ure S4A). Notably, co-culture of the H929 MM cells did not affect the expression of the osteoclastogenic factors Cdh2, Vcam1, Tnfsf11/Rankl, or Il6 by the KS483 cells (Supplementary Figure S4A-B). In addition, co-cultures of primary human mesenchymal stromal cells with H929 shCDH2 cells or H929 TR cells also revealed a diminished inhibition of the late osteogenic marker BGLAP upon N-cadherin knockdown, while there was no change in expression of the early markers ALPL or COL1A1 (Figure 5F). Furthermore, in the co-cultures with either KS483 or these primary mesenchymal stromal cells, N-cadherin silencing did not reduce MM growth. Thus, N-cadherin-mediated adhesion does not control production of MM-growth supportive cytokines by the osteoblasts, and the observed impaired inhibition of osteoblast differentiation upon N-cadherin silencing is not due to reduced MM cell numbers (Supplementary Figure S4C). Taken together, these observations show that N-cadherin plays an important role in the interaction of MM cells with osteoblasts, and establish an important role for this N-cadherin- mediated interaction in the inhibition of osteoblastogenesis.

Figure 5. N-cadherin mediates inhibition of osteoblast differentiation by MM cells (A) N-cadherin expression in osteoblastic cell lines. Cell lysates were immunoblotted using a monoclonal antibody against N-cadherin (clone 32), and β-actin was used as a loading control. (B) N-cadherin- mediated adhesion of MM cells to osteoblasts. MM cell lines were allowed to adhere to C3H10T1/2

126 N-cadherin-mediated interaction with multiple myeloma cells…

cells in the presence of an N-cadherin blocking antibody (GC-4) or isotype control antibodies. (C) N-cadherin knockdown abolishes N-cadherin-mediated adhesion of MM cells to osteoblasts. H929 TR and H929 shCDH2 cells were incubated with or without doxycycline for 5 days, and subsequently allowed to adhere to C3H10T1/2 cells. (D) N-cadherin mediates MM cell-controlled inhibition of osteoblast differentiation. Murine (pre-)osteoblastic KS483 cells were, upon confluence, further differentiated in the presence or absence of MM cells, either treated with (black bars) or without doxycycline (white bars), and subsequently stained for alkaline phosphatase (ALP) expression and quantified. (E) N-cadherin represses alkaline phosphatase (Akp2), collagen type I, alpha1 (Col1a1), and osteocalcin (Bglap) gene expression. Murine KS483 cells were differentiated in the presence or absence of H929 TR or H929 shCDH2 cells with (black bars) or without doxycycline (white bars), and subsequently analyzed by qRT-PCR. (F) N-cadherin represses the late human osteogenic differentiation. Human primary MSC were differentiated in the presence or absence of H929-TR or H929-shCDH2 cells with (black bars) or without doxycycline (white bars), and subsequently analyzed by qRT-PCR for the expression of the early osteogenic markers alkaline phosphatase (ALPL), and collagen type I, alpha1 (COL1A1), and the late marker osteocalcin (BGLAP).

Discussion 1 Although MM cell growth is driven by genetic alterations like translocation and mutations, these cells still remain dependent upon the BM microenvironment. The interactions of MM cells with the BM microenvironment, either directly via adhesion molecules or indirectly via the ensuing stimulation of autocrine/par- 3 acrine production of cytokines, activate a broad range of proliferative and anti- apoptotic signaling pathways.6,7 Here, we identified a new interactant of MM cells 4 with the BM microenvironment, i.e. N-cadherin. We observed high expression of N-cadherin on the malignant plasma cells in 5 a subset of approximately 50% of primary MM as well as MM cell lines, but not in normal BM plasma cells (Figures 1 and 3). The functionality of the expressed 6 N-cadherin was initially demonstrated by means of in vitro homophilic adhesion assays (Figure 2). As expected, N-cadherin co-localizes and physically interacts 7 with β-catenin (Figures 1 and 3). Furthermore, high levels of N-cadherin seem to correlate with high levels of β-catenin (Figure 1B). This increase in β-catenin protein levels might be explained by the “protective” effect of N-cadherin binding to β-catenin, preventing degradation of the latter.34,35 Our gene expression profiling analysis of a large group of MM patients revealed that CDH2, the gene encoding N-cadherin, is highly, but not exclusively, expressed in MM cells bearing a t(4;14)(p16;q32) translocation. Although this MM subtype is associated with poor prognosis,36,37 expression of CDH2 by itself does not predict prognosis. Notably, in a previous independent gene expression profiling study of 29 primary MM samples, CDH2 was among the genes up-regulated in five patients carrying the t(4;14) MMSET translocation.38 Moreover,CDH2 was among the

127 Chapter 5

genes down-regulated upon silencing of the MMSET gene in the KMS-11 cell- line.39 Also, in our panel of MM cell-lines, the two cell-lines that carry the t(4;14) translocation, i.e., OPM-1 and NCI-H929, displayed prominent N-cadherin expression (Figure 1). Taken together, these data indicate that the aberrant expression of N-cadherin in a subgroup of MM patients may, at least in part, be the indirect consequence of overexpression of the transcriptional repressor MMSET. Increasing evidence indicates that the gain of N-cadherin expression in solid tumor cells is associated with enhanced invasive potential and metastasis.30,31,40 In line with these reports, we here show that N-cadherin potentiates basal MM cell motility as well as SDF-1-induced migration (Figure 4B-C), and contributes to the homing to and/or retention in the BM of MM cells (Figure 4A). In order to separate these two processes, we performed a transendothelial migration assay, showing that knockdown of N-cadherin and/or treatment with an N-cadherin blocking antibody, did not affect migration of MM cells across an endothelial barrier (Figure 4D), whereas a blocking antibody targeting α4-integrin did suppress migration of the cells. In this context it is important to note that endothelial cells mainly display extrajunctional localization of N-cadherin,41,42 which might facilitate adhesion rather than migration in these static migration assays. Never- theless, our results favor the hypothesis that N-cadherin is involved in the retention of MM cells in the BM, which may be the consequence of homotypic and/or heterotypic cell-cell adhesion within the BM microenvironment. Indeed, N-cadherin expression by MM cells participates in homophilic, heterotypic adhesion of MM cells with their environment, e.g., with osteoblasts (Figures 2 and 5). It has become apparent that the interplay between MM cells and bone is bi-directional. Important clinical sequelae of this bidirectional interaction is increased osteoclast activity and suppression of osteoblast function, resulting in osteolytic bone lesions.6,7 Besides inhibition by soluble factors, such as Wnt antagonists, activin A, interleukin-3 and interleukin-7, osteoblast function and maturation can also be suppressed through direct contact between MM cells and pre-osteoblastic cells mediated via α4β1-VCAM-1 and/or N-CAM-N-CAM interactions.43 The latter findings, in addition to the reported key role of cadherin-based interactions in osteoblast function and differentiation,9,32,33prompted us to investigate whether N-cadherin-mediated adhesion of MM cells to osteoblasts (Figure 5B–C and Sup- plementary Figure S3) can contribute to inhibition of osteoblast differentiation. Indeed, silencing of N-cadherin in myeloma cells impaired these cells’ capacity to suppress osteoblast differentiation (Figure 5D-F). It should, however, be noted that although these data clearly show that N-cadherin-mediated adhesion of MM cells to osteoblasts contributes to inhibition of osteoblast differentiation, no cor-

128 N-cadherin-mediated interaction with multiple myeloma cells… relation could be found between CDH2 expression and the presence of osteolytic lesions (data not shown), and N-cadherin expression was also (weakly) enhanced in some cases of MGUS (Figure 3A, right panel). This pre-malignant stage of disease is characterized by a small expansion of clonal plasma cells, diffusely located throughout the bone marrow, and by the absence of osteolytic lesions. Presumably, inhibition of osteoblast differentiation may require local high amounts of plasma cells. Notably, expression of another molecule associated with bone disease, DKK1, has also been observed in MGUS.11 Furthermore, since N-cadherin expression not only contributes to the inhibition of osteoblast differentiation, but also to migration and BM homing/retention (Figure 4), it may well be that in this pre-malignant stage of disease N-cadherin is mainly involved in determining the localization of the plasma cells. Together our results indicate that N-cadherin expression contributes to osteolysis, but is neither critical nor sufficient for this process. In addition to the indirect role of N-cadherin discussed above, N-cadherin- mediated adhesion may also directly inhibit osteoblast differentiation. Since both 1 loss of function and over-expression of N-cadherin have been shown to result in osteoporosis,32,33,44 two distinct mechanisms may account for this. The adhesion may induce relocation of N-cadherin from the intercellular osteoblast junctions to the site of MM contact, resulting in loss of osteoblast-osteoblast contact and 3 osteoblast function. Alternatively, adhesion may enhance N-cadherin stability and expression, resulting in inhibition of osteoblast differentiation by interfer- 4 ing with Wnt signaling,33 a pathway playing a central role in the osteolytic bone disease observed in MM.11,12,43 5 Apart from contributing to osteolytic bone disease, the N-cadherin-mediated interaction of MM with osteoblasts may also contribute to another sequela of 6 MM, i.e. the development of pancytopenia. Since osteoblasts have a central role in the organization of the endosteal stem cell niche,45 it is tempting to speculate 7 that N-cadherin expression might enable MM cells to access this niche, leading to niche dysregulation and thereby contributing to the pancytopenia that is typically observed in MM patients. In conclusion, our data indicate that N-cadherin-mediated interaction of MM cells with osteoblasts may be involved in bone marrow retention and results in reduced osteoblast differentiation, which may play a crucial role in the pathogenesis of myeloma bone disease. Targeting N-cadherin, e.g., with a recently developed small cyclic peptide ADH-1,46 may prove to be a successful novel means of therapeutic intervention in a subgroup of MM patients.

129 Chapter 5

References

1. Hallek M, Bergsagel PL, Anderson KC. Multiple myeloma: increasing evidence for a multistep transformation process. Blood 1998; 91(1): 3–21. 2. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2(3): 175–187. 3. Alsayed Y, Ngo H, Runnels J, Leleu X, Singha UK, et al. Mechanisms of regulation of CXCR4/SDF-1 (CXCL12)-dependent migration and homing in multiple myeloma. Blood 2007; 109(7): 2708–2717. 4. Pals ST, de Gorter DJ, Spaargaren M. Lymphoma dissemination: the other face of lymphocyte homing. Blood 2007; 110(9): 3102–3111. 5. Sanz-Rodriguez F, Hidalgo A, Teixido J. Chemokine stromal cell-derived factor- 1alpha modulates VLA-4 integrin-mediated multiple myeloma cell adhesion to CS-1/fibronectin and VCAM-1. Blood 2001; 97(2): 346–351. 6. Hideshima T, Mitsiades C, Tonon G, Richardson PG, Anderson KC. Understanding multiple myeloma pathogenesis in the bone marrow to identify new therapeutic targets. Nat Rev Cancer 2007; 7(8): 585–598. 7. Podar K, Chauhan D, Anderson KC. Bone marrow microenvironment and the identification of new targets for myeloma therapy. Leukemia 2009; 23(1): 10–24. 8. Clohisy D. Cellular mechanisms of osteolysis. J Bone Joint Surg 2003; Am85-A (Suppl 1): S4–6. 9. Mbalaviele G, Shin CS, Civitelli R. Cell-cell adhesion and signaling through cadherins: connecting bone cells in their microenvironment. J Bone Miner Res 2006; 21(12): 1821–1827. 10. Chen Y, Alman BA. Wnt pathway, an essential role in bone regeneration. J Cell Bio- chem 2009; 106(3): 353–362. 11. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med 2003; 349(26): 2483–2494. 12. Oshima T, Abe M, Asano J, Hara T, Kitazoe K, et al. Myeloma cells suppress bone formation by secreting a soluble Wnt inhibitor, sFRP-2. Blood 2005; 106(9): 3160–3165. 13. Standal T, Abildgaard N, Fagerli UM, Stordal B, Hjertner O, et al.(2007) HGF inhibits BMP-induced osteoblastogenesis: possible implications for the bone disease of multiple myeloma. Blood 2005; 109(7): 3024–3030. 14. Giuliani N, Colla S, Morandi F, Lazzaretti M, Sala R, et al. Myeloma cells block RUNX2/CBFA1 activity in human bone marrow osteoblast progenitors and inhibit osteoblast formation and differentiation. Blood 2005; 106(7): 2472–2483. 15. Derksen PWB, Tjin E, Meijer HP, Klok MD, Mac Gillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci USA 2004; 101(16): 6122–6127.

130 N-cadherin-mediated interaction with multiple myeloma cells…

16. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/TCF transcriptional complex in the treatment of multiple myeloma. Proc Natl Acad Sci USA 2007; 104(18): 7516–7521. 17. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/beta-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 18. Harris TJ, Peifer M. Decisions, decisions: beta-catenin chooses between adhesion and transcription. Trends Cell Biol 2005; 15(5): 234–237. 19. Derycke LD, Bracke ME. N-cadherin in the spotlight of cell-cell adhesion, differentiation, embryogenesis, invasion and signalling. Int J Dev Biol 2004; 48(5–6): 463–476. 20. Reijmers RM, Groen RW, Rozemuller H, Kuil A, de Haan-Kramer A, et al. Targeting EXT1 reveals a crucial role for heparan sulfate in the growth of multiple myeloma. Blood 2010; 115(3): 601–604. 21. de Gorter DJ, Beuling EA, Kersseboom R, Middendorp S, van Gils JM, et al. Bruton’s tyrosine kinase and phospholipase Cgamma2 mediate chemokine-controlled B cell migration and homing. Immunity 2007; 26(1): 93–104. 1 22. Spaargaren M, Beuling EA, Rurup ML, Meijer HP, Klok MD, et al. The B cell antigen receptor controls integrin activity through Btk and PLCgamma2. J Exp Med 2003; 198(10): 1539–1550. 23. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, et al. Cyclin D dysregulation: an 3 early and unifying pathogenic event in multiple myeloma. Blood 2005; 106(1): 296–303. 24. van der Horst G, van Bezooijen RL, Deckers MM, Hoogendam J, Visser A, et al. Dif- 4 ferentiation of murine preosteoblastic KS483 cells depends on autocrine bone morphogenetic protein signaling during all phases of osteoblast formation. Bone 2002; 31(6): 661–669. 5 25. Prins HJ, Rozemuller H, Vonk-Griffioen S, Verweij VG, Dhert WJ, et al. Bone-forming capacity of mesenchymal stromal cells when cultured in the presence of human platelet 6 lysate as substitute for fetal bovine serum. Tissue Eng Part A 2009; 15(12): 3741–3751. 26. Ramakers C, Ruijter JM, Deprez RH, Moorman AF. Assumption-free analysis of 7 quantitative real-time polymerase chain reaction (PCR) data. Neurosci Lett 2003; 339(1): 62–66. 27. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108(6): 2020–2028. 28. Fonseca R, Bergsagel PL, Drach J, Shaughnessy J, Gutierrez N, et al. International Myeloma Working Group molecular classification of multiple myeloma: spotlight review. Leukemia 2009; 23(12): 2210–2221. 29. Kamei J, Toyofuku T, Hori M. Negative regulation of p21 by beta-catenin/TCF signaling: a novel mechanism by which cell adhesion molecules regulate cell proliferation. Biochem Biophys Res Commun 2003; 312(2): 380–387. 30. Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. Cadherin switching. J Cell Sci 2008; 121(Pt 6): 727–735.

131 Chapter 5

31. Blaschuk OW, Devemy E. Cadherins as novel targets for anti-cancer therapy. Eur J Pharmacol 2009; 625(1–3): 195–198. 32. Cheng SL, Shin CS, Towler DA, Civitelli R. A dominant negative cadherin inhibits osteoblast differentiation. J Bone Miner Res 2000; 15(12): 2362–2370. 33. Hay E, Laplantine E, Geoffroy V, Frain M, Kohler T, et al. N-cadherin interacts with axin and LRP5 to negatively regulate Wnt/beta-catenin signaling, osteoblast function, and bone formation. Mol Cell Biol 2009; 29(4): 953–964. 34. Sadot E, Simcha I, Shtutman M, Ben-Ze’ev A, Geiger B. Inhibition of beta-catenin-mediated transactivation by cadherin derivatives. Proc Natl Acad Sci USA 1998; 95(26): 15339–15344. 35. Heuberger J, Birchmeier W. Interplay of cadherin-mediated cell adhesion and canonical wnt signaling. Cold Spring Harb Perspect Biol 2010; 2(2): a002915. 36. Keats JJ, Reiman T, Maxwell CA, Taylor BJ, Larratt LM, et al. In multiple myeloma, t(4;14)(p16;q32) is an adverse prognostic factor irrespective of FGFR3 expression. Blood 2003; 101(4): 1520–1529. 37. Chng WJ, Kuehl WM, Bergsagel PL, Fonseca R. Translocation t(4;14) retains prognostic significance even in the setting of high-risk molecular signature. Leukemia 2008; 22(2): 459–461. 38. Dring AM, Davies FE, Fenton JA, Roddam PL, Scott K, et al. A global expression- based analysis of the consequences of the t(4;14) translocation in myeloma. Clin Cancer Res 2004; 10(17): 5692–5701. 39. Lauring J, Abukhdeir AM, Konishi H, Garay JP, Gustin JP, et al. The multiple myeloma associated MMSET gene contributes to cellular adhesion, clonogenic growth, and tumorigenicity. Blood 2008; 111(2): 856–864. 40. Hazan RB, Phillips GR, Qiao RF, Norton L, Aaronson SA. Exogenous expression of N-cadherin in breast cancer cells induces cell migration, invasion, and metastasis. J Cell Biol 2000; 148(4): 779–790. 41. Salomon D, Ayalon O, Patel-King R, Hynes RO, Geiger B. Extrajunctional distribution of N-cadherin in cultured human endothelial cells. J Cell Sci 1992; 102(Pt 1): 7–17. 42. Navarro P, Ruco L, Dejana E Differential localization of VE- and N-cadherins in human endothelial cells: VE-cadherin competes with N-cadherin for junctional localization. J Cell Biol 1998; 140(6): 1475–1484. 43. Roodman GD. Osteoblast function in myeloma. Bone 2011; 48(1): 135–140. 44. Castro CH, Shin CS, Stains JP, Cheng SL, Sheikh S, et al. Targeted expression of a dominant-negative N-cadherin in vivo delays peak bone mass and increases adi- pogenesis. J Cell Sci 2004; 117(Pt 13): 2853–2864. 45. Li P, Zon LI. Resolving the controversy about N-cadherin and hematopoietic stem cells. Cell Stem Cell 2010; 6(3): 199–202. 46. Williams E, Williams G, Gour BJ, Blaschuk OW, Doherty P. A novel family of cyclic peptide antagonists suggests that N-cadherin specificity is determined by amino acids that flank the HAV motif. J Biol Chem 2000; 275(6): 4007–4012.

132 N-cadherin-mediated interaction with multiple myeloma cells… supplementary material Supplementary Design and Methods antibodies Monoclonal antibodies were: anti-N-cadherin, clone 32 (IgG1); anti-β-catenin, clone 14 (IgG1) (both from BD Biosciences, Erembodegem, Belgium); anti- β-actin, clone AC-15 (IgG1); anti-N-cadherin, clone GC-4 (IgG1) (both from Sigma-Aldrich, St Louis, MO, USA); anti-E-cadherin, clone HECD-1 (IgG1) (Ta- kara Bio, Shiga, Japan); anti-CD138, clone B-B4 (IgG1) (IQ Products, Groningen, The Netherlands); and IgG1 control antibody (DAKO, Carpinteria, CA, USA). Polyclonal antibodies used were: rabbit anti-human β-catenin, H-102 (Santa Cruz Biotechnology, Santa Cruz, CA; USA); horseradish peroxidase (HRP)- conjugated rabbit anti-mouse; R-phycoerythrin (RPE)-conjugated streptavidin (both from DAKO); biotinylated goat anti-mouse IgG1 (Southern Biotechnology, Birmingham, AL, USA); Alexa488-conjugated goat anti-mouse and Alexa568- 1 conjugated goat anti-rabbit (both from Invitrogen Life technologies, Breda, The Netherlands). cell culture 3 Multiple myeloma (MM) cell lines, UM-1, UM-3, L363, OPM-1, NCI-H929, XG-1 and LME-1 were cultured as described previously (Derksen et al., 2004). 4 The murine cell lines C2C12 and C3H10T1/2 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen Life Technologies) and murine 5 pre-osteoblastic KS483 cells were grown in minimum essential medium (MEM) alpha (Invitrogen Life Technologies), both supplemented with 10% fetal calf se- 6 rum (FCS), penicillin (50 U/mL) and streptomycin (50 μg/mL) (both from Inv- itrogen Life Technologies). 7 cell growth assessment Cells were plated (5 × 103) in a 96-well plate coated with recombinant N-cadherin/Fc chimera (1 μg/mL; R&D Systems, Abingdon, UK), or BSA as control. When assessing the growth of the doxycycline-inducible cell line, knockdown was obtained by incubating NCI-H929 shCDH2 cells for 5 days with 0.2 μg/mL doxycycline prior to the assay, and maintained by addition of 0.2 μg/mL doxycycline to the culture medium. Cell numbers and viability were determined by means of trypan blue staining (Sigma-Aldrich). The effect of co-culture with KS483 or human mesenchymal stem cells on MM growth was analyzed using luciferase-marked myeloma cells, which were generated as described previously

133 Chapter 5

(Rozemuller et al., 2008). Cell viability was determined with a lumi- nometer as described by McMillin et al. (McMillin et al., 2010). immunoprecipitation and immunoblot analysis Immunoprecipitation and western blot analysis were performed as described previously (de Gorter et al., 2007). The immunoblots were stained with anti-N- cadherin (clone 32), anti-E-cadherin (HECD-1), or anti-β-catenin (clone 14). Equal loading was confirmed with anti-β-actin. Primary antibodies were detected by HRP-conjugated rabbit anti-mouse, followed by detection using Lumi-Light- PLUS western blotting substrate (Roche, Basel, Switzerland). cell adhesion assays Cell-substrate adhesion assays were done as described previously (Spaarga- ren et al., 2003), in triplicate on flat-bottom 96-well plates (Costar, Cambridge, MA, USA) coated overnight at 4°C with PBS containing 1 μg/mL recombinant N-cadherin/Fc chimera (R&D Systems), 4% BSA, or for 15 min at 37°C with 1 mg/mL poly-l-lysine (PLL), and blocked for 2 h at 37°C with 4% BSA in RPMI 1640. MM cells (105 cells/100 μL) were plated and incubated at 37°C for 20 min. To block N-cadherin-mediated adhesion, cells were incubated with the monoclonal antibody GC-4 (10 μg/mL) for 30 min at 4°C, prior to the adhesion assay. Results are presented as percentages of maximum adhesion, as measured by adhesion to the PLL-coated surface, and the bars represent the means ±SD of a triplicate experiment of at least three independent experiments. Malignant plasma cells were isolated from bone marrow aspirates from MM patients using magnetic activated cell sorting (MACS) as described elsewhere (Derksen et al., 2003). For cell-cell adhesion assays, C3H10T1/2 osteoblastic cells were seeded at a density of 7500 cells/200 μL, in a 96-well flat-bottom tissue culture plates (Cos- tar). Twenty-four hours after plating an adhesion assay was performed by adding MM cells (105 cells/100 μL) and incubating for 20 min either in culture medium as a control, or in Hanks’ balanced salt solution (HBSS) in the presence or absence of 2 mM calcium chloride, and in the presence of an N-cadherin blocking antibody (10 μg/mL), or an isotype antibody as a control. Images were captured using an EVOS original camera (AMG, Mill Creek, WA, USA) and processed with Adobe Photoshop. The results are expressed as relative adhesion with the adhesion of the non-pretreated MM cells to C3H10T1/2 cells in HBSS supplemented with calcium and the isotype control normalized to 100. The bars represent the means ± SD of four measurements, representative of at least three independent experiments.

134 N-cadherin-mediated interaction with multiple myeloma cells… migration assays Migration assays were performed in triplicate as described previously (de Gorter et al., 2007), with transwells (8μm pore size; Costar) coated with 1 μg/mL recombinant N-cadherin/Fc chimera (R&D Systems), sVCAM-1 (R&D Systems), or BSA (fraction V; Sigma-Aldrich) coating as a control. Transendothelial migration was performed by growing a confluent layer of HUVEC cells on a fibronectin-coated transwell insert. Subsequently, H929 shCDH2 cells (5 × 105), either induced with or without doxycycline, were added and allowed to migrate for 5 h towards 100 ng/mL SDF-1, in the presence or absence of blocking antibodies against N-cadherin (GC-4), or α4-integrin (HP2/1). The amount of viable migrating cells was determined by fluorescence-activated cell sorting (FACS) and expressed as a percentage of the input. The percentage of non-pretreated cells was normalized to 100%. The bars represent the means ± SD of three measurements, representative of at least three independent experiments. 1 sample preparation and microarray hybridization and analysis Isolation of plasma cells and RNA profiling were performed as described previously (Bergsagel et al., 2005). Gene expression was measured by U133 Plus2.0 3 Affymetrix oligonucleotide microarray probeset 203440_at of 559 newly diagnosed MM patients. Expression data were summarized with MAS5 using default 4 parameters in Affymetrix GeneChip operating software, median normalized and plotted against genomic aberrations. This research study was performed with 5 the approval of the institutional review board, and all subjects provided written informed consent in accordance with the Declaration of Helsinki. In addition, 6 publicly available U133 Plus2.0 Affymetrix oligonucleotide microarray data, provided by the Donna D. and Donald M. Lambert Laboratory of Myeloma Genetics, 7 were used to analyze the expression of N-cadherin on the plasma cells of 345 MM patients from the total therapy 2 (TT2) patient set. MAS5 summarized data have been deposited in the NIH Gene Expression Omnibus (GEO; National Center for Biotechnology Information [NCBI], http://www.ncbi.nlm.nih.gov/geo/ under accession number GSE2658. immunohistochemistry Immunohistochemical staining was performed on formalin-fixed, plastic-embedded tissue sections. Endogenous peroxidase activity was blocked with 0.3% H2O2 in methanol. For antigen retrieval sections were boiled for 10 min in a Tris/EDTA buffer (respectively 10 mM/1 mM) pH9, after which they were blocked with 10%

135 Chapter 5 normal goat serum. Followed by incubation for either 1 h at room temperature with mouse monoclonal antibody against CD138 or overnight at 4°C with mouse monoclonal antibodies against N-cadherin (clone 32) or β-catenin (clone 14). Binding of the antibody was visualized using the PowerVision plus detection system (Immunovision Technologies, Duiven, The Netherlands) and 3,3-diaminobenzidine (Sigma). The sections were counterstained with hematoxylin (Merck, Darmstadt, Germany), washed and protected with a cover slip. osteoblast differentiation KS483 cells were seeded at a density of 12000 cells/cm2 and cultured until confluence. After confluence ascorbic acid (50 μg/mL) was added to the medium, and the medium was changed every 3 days. Co-cultures were initiated from the day of confluence, day 4, by addition of MM cells (25 × 103/well of a 24-well plate for alkaline phosphatase expression; 106/well of a 6-well plate for RNA samples) and maintained for 1 week, in the presence or absence of doxycycline (0.2 μg/mL; Sigma-Aldrich). Subsequently, cultures were stained for alkaline phosphatase expression as described by van der Horst et al. (van der Horst et al., 2002), or cells were lysed in Tri Reagent (Sigma-Aldrich). Co-cultures with primary human mesenchymal stromal cells (MSC), obtained and expanded as described previously (Prins et al., 2009), were initiated by plating 8 × 104 MSC in a 6-well plate in a platelet-lysate supplemented medium (Prins et al., 2009). After 24 h, osteogenic differentiation was started by changing the medium with NH OsteoDiff human medium (Miltenyi Biotec, Bergisch Gladbach, Germany), in the presence or absence of 106 MM cells. Cultures were maintained for 1 week and medium was changed every 3 days. MSC were positively selected from the MM cells by MACS as described previously (Derksen et al., 2003), using anti-CD73 (clone AD2, BD Biosciences). real-time reverse transcription polymerase chain reaction RNA isolation and cDNA synthesis were performed as described previously (Derksen et al., 2004). The quantitative reverse transcription polymerase chain reaction (qRT-PCR) runs were performed on a Roche LightCycler 1.5 using Fast- Start DNA Master SYBR Green I kit (Roche, Basel, Switzerland). Results were analyzed using LinReg PCR analysis software (version 7.5; Ramakers et al., 2003). Expression was normalized over beta-2-microglobulin expression. Mouse-specific primers were designed recognizing alkaline phosphatase (Akp2), osteocalcin (Bglap), collagen, type I, alpha1 (Col1A1), sp7 transcription

136 N-cadherin-mediated interaction with multiple myeloma cells… factor 7/osterix (Osx), runt-related transcription factor 2 (Runx2), vascular cell adhesion molecule 1 (Vcam1), tumor necrosis factor (ligand) superfamily member 11 aka Rankl (Tnfsf1), interleukin 6 (Il6), N-cadherin (Cdh2), and beta- 2-microglobulin (B2m). Human primers were designed recognizing alkaline phosphatase (ALPL), collagen type I, alpha1 (COL1A1), osteocalcin (BGLAP), and beta-2-microglobulin (B2M). The sequences are shown in Supplementary Tables S1 and S2.

Supplementary Table 1. Mouse-specific primer sets for quantitative PCR

Forward (5′–3′) Reverse (5′–3′) 1 Alkaline phosphatase Akp2 GGATAACGAGATGC CATCCAGTTCGTATTC CACC CAC Osteocalcin Bglap CAATAAGGTAGTGA CTGGTCTGATAGCTCG ACAGACTCC TCAC Collagen, type I, alphal Col1A1 CAAAGGAGAACCCG GGTCCAGGCAGTCCG 1 GTGCTAC GAAG Sp7 transcription factor 7/ Sp7/Osx TCCCATTCTCCCTCC GGGACTGGAGCCATA osterix CTCT GTGAG Runt-related transcription Runx2 GATCTGAGATTTGTG CCACTGTCACTTTAAT factor 2 GGC AGC 3 Interleukin 6 Il6 TGATGGATGCTACCA TTCATGTACTCCAGGT AACTGG AGCTATGG 4 Tumor necrosis factor Tnfsfll AAGACACACTACCTG CCACAATGTGTTGCA (ligand) superfamily, primerset 1 ACTCCTGC GTTCC 5 member 11 Tnfsfll ACTCTGGAGAGTGA CCATGAGCCTTCCATC primerset 2 AGACACACTAC ATAG 6 Tnfsfll CCAGCCATTTGCAC AGCAGGGAAGGGTTG primerset 3 ACCTC GACA 7 Vascular cell adhesion Vcaml TGGTGAAATGGAAT GACCCAGATGGTGGT molecule 1 CTGAACC TTCC Cadherin 2 Cdh2 CCTCCATGTGCCGG CAATTTCACCAGAAG ATAG CCTCC Beta-2-microglobulin B2m CTGGTGCTTGTCTC GGTGGAACTGTGTTA ACTGACC CGTAGC

All primers were manufactured by Sigma-Aldrich (Haverhill, UK). The results are expressed as relative inhibition, with the co-culture of KS483 (or primary MSCs) and H929 in the absence of doxycycline normalized to 100. The bars represent the means ± SD of three measurements, representative of at least three independent experiments.

137 Chapter 5

Supplementary Table 2. Human primer sets for quantitative PCR

Forward (5′–3′) Reverse (5′–3′) 1 Alkaline phosphatase ALPL ACAAGCACTCCCAC TCACGTTGTTCCTG TTCATCTGGA TTCAGCTCGT Osteocalcin BGLAP GGCAGCGAGGTAGT GAT GT GGTCAGCC GAAGAG AACTCGT Collagen, type I, alpha1 COLlAl AGGGCCAAGACGAA AGATCACGTCATCG GACATC CACAACA Beta-2-microglobulin B2M GTCTTTCAGCAAGG CTTCAAACCTCCAT ACTGGTC GATGC

Supplementary Figure 1. Multiple myeloma cells express N-cadherin Fluorescence-activated cell sorting (FACS) analysis for N-cadherin protein expression in MM cell lines. Cells were stained with anti-N-cadherin monoclonal antibody GC-4 (open histogram) or isotype control (filled histogram).

138 N-cadherin-mediated interaction with multiple myeloma cells…

3 4 5 6 Supplementary Figure 2. N-cadherin does not affect MM cell growth (A) Aggregation of MM cells blocked by N-cadherin coating. MM cell lines OPM-1 (top panel) and 7 NCI-H929 (bottom panel) were plated on N-cadherin (1 μg/mL; right column) or on BSA as a control (left column). Representative pictures of the aggregation of the MM cell lines are shown. (B) MM cells (5 × 103) were plated on BSA (·) or N-cadherin (■: 1 μg/mL) coated surfaces and cultured for 4 days. Cell viability was determined using FACS. The growth curves represent the means ± SD of three measurements representative of at least three independent experiments. (C) Knockdown of N-cadherin in the MM cell line NCI-H929, containing a doxycycline-inducible shRNA targeting N-cadherin (H929 shCDH2). Cells were incubated with (black line) or without doxycycline (gray line) for 5 days, and subsequently analyzed by FACS (left panel) using an N-cadherin monoclonal antibody (clone GC-4) and an isotype control (filled gray histogram). By western blot analysis (right panel) knockdown was confirmed (H929 shCDH2) and compared to the control cell line containing the TET repressor only (H929 TR), using a monoclonal antibody N-cadherin (clone 32). β-actin was used as a loading control. (D, E) The growth rate of H929 shCDH2 (right panels) was compared with that of H929 TR (left panels), in the presence (▲) and absence (■) of doxycycline, with (E) or without (D) N-cadherin coating (1 μg/mL), over a 4-day culture period. The growth curves represent the means ± SD of three measurements representative of at least three independent experiments.

139 Chapter 5

Supplementary Figure 3. N-cadherin-mediated adhesion to osteoblasts (A) MM cell lines were allowed to adhere to C3H10T1/2 cells in Hanks’ balanced salt solution (HBSS) in the presence of calcium in combination with an N-cadherin blocking antibody (GC-4) or isotype control antibodies. Representative pictures of the adhesion in HBSS supplemented with calcium and the isotype as a control (left column), and HBSs supplemented with calcium and the blocking antibody GC-4 (right column). (B) N-cadherin knockdown abolishes N-cadherin-mediated adhesion of MM cells to osteoblasts. H929 TR (top panel) and H929 shCDH2 cells (bottom panel) were incubated with (right column) or without (left column) doxycycline for 5 days, and subsequently allowed to adhere to C3H10T1/2 cells in HBSS in the presence of calcium. Representative pictures of the adhesion are shown.

140 N-cadherin-mediated interaction with multiple myeloma cells…

3 4 5 6 Supplementary Figure 4. N-cadherin-mediated inhibition of osteoblast differentiation does not involve inhibition of osteogenic transcription factors, or a change in MM cell 7 viability (A) Murine KS483 cells were differentiated in the presence or absence of H929 TR or H929 shCDH2 cells with (black bars) or without doxycycline (white bars), and subsequently analyzed by qRT-PCR for the expression of sp7 transcription factor 7/ osterix (Sp7/Osx), runt-related transcription factor 2 (Runx2), cadherin 2 (Cdh2), and vascular cell adhesion molecule 1 (Vcaml). The results are expressed as relative inhibition, with the culture of KS483 without H929 cells (or doxycycline) set to 100. ns: not significant, by Student’s t-test. (B) Co-incubation of murine KS483 cells with MM cells does not result in an induction of interleukin 6 (Il6), or tumor necrosis factor (ligand) superfamily, member 11, (Tnfsfll), also known as Rank ligand, using murine inflamed kidney and lymph node respectively as positive controls. Representative agarose gels are shown. (C) Doxycycline-induced knockdown of N-cadherin does not affect MM viability in co-cultures with KS483 or primary human mesenchymal stromal cells (MSC). The growth rate of H929 shCDH2 was compared with that of H929 TR, in the presence (black bars) and absence (white bars) of doxycycline, in co-culture with murine KS483 or human MSC, over a 48 h culture period. The results are expressed as relative growth induction as compared to the input of MM cells at day 0.

141

Chapter6

The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma and promotes angiogenesis

Kinga A. Kocemba1,4, Harmen van Andel1,4, Anneke de Haan-Kramer1, Karène Mahtouk1, Rogier Versteeg2, Marie José Kersten3, Marcel Spaargaren1,5 and Steven T. Pals1,5

1Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands,2 Department of Human Genetics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands and3 Department of Hematology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands,4 These authors contributed equally to this work,5 These authors share last authorship

Leukemia 2013; 27(8): 1729–1737 Abstract

In multiple myeloma (MM), angiogenesis is strongly correlated to disease progression and unfavorable outcome, and may be promoted by bone marrow hypoxia. Employing gene-expression profiling, we here identified the pro-angiogenic factor adrenomedullin (AM) as the most highly upregulated gene in MM cells exposed to hypoxia. Malignant plasma cells from the majority of MM patients, belonging to distinct genetic subgroups, aberrantly express AM. Already under normoxic conditions, a subset of MM highly expressed and secreted AM, which could not be further enhanced by hypoxia or cobalt chloride-induced stabilization of hypoxia-inducible factor (HIF)1α. In line with this, expression of AM did not correlate with expression of a panel of established hypoxia-/HIF1α-target genes in MM patients. We demonstrate that MM-driven promotion of endothelial cell proliferation and tube formation is augmented by inducible expression of AM and strongly repressed by inhibition of endogenous and hypoxia-induced AM activity. Together, our results demonstrate that MM cells, both in a hypoxia- dependent and independent fashion, aberrantly express and secrete AM, which can mediate MM-induced angiogenesis. Thus, AM secretion can be a major driving force for the angiogenic switch observed during MM evolution, which renders AM a putative target for MM therapy.

144 Introduction

Multiple myeloma (MM) is a neoplasm characterized by expansion of malignant plasma cells in the bone marrow. The transition of a normal plasma cell to a fully transformed, aggressive myeloma cell is a multistep process, which requires the 1 acquisition of chromosomal translocations and mutations in multiple genes. Most of this evolution takes place in the bone marrow (BM), indicating that the interaction with the BM microenvironment has a critical role in the pathogenesis of MM.1,2 In the MM-infiltrated BM, aberrant neovascularization (angiogenesis) is 3 almost invariably present, and is associated with endothelial activation, increased capillary permeability and hyperperfusion.3,4 Importantly, BM angiogenesis in 4 MM parallels disease progression and is correlated with poor event-free and overall survival,5,6 while after successful treatment, microvessel density returns 5 to normal.7,8 The pathogenesis of MM-induced angiogenesis has not yet been fully eluci- 6 dated. It is driven by genetic alterations in the MM cells resulting in an imbalance between the production of pro- and anti-angiogenic factors by myeloma cells and 7 the microenvironment. This is reflected by elevated levels of pro-angiogenic factors, including VEGFA, bFGF, and HGF, in the BM plasma and peripheral blood of MM patients.9 This imbalance results in an “angiogenic switch”, which takes place on the verge of progression of monoclonal gammopathy of undetermined significance (MGUS) to active MM. As some of the important angiogenic factors secreted by myeloma cells, including VEGFA and bFGF, are equally expressed by tumor cells isolated from MGUS, smoldering MM and active MM,10 it has been suggested that the angiogenic switch could also be the consequence of increasing tumor burden, rather than that of aberrant expression of pro-angiogenic factors alone. In line with this notion, a study by Hose et al.11 revealed that even normal bone marrow plasma cells (BMPC) have significant pro-angiogenic properties.

145 Chapter 6

On the other hand, chronic hypoxia may also have an important role in BM angiogenesis in MM.12–14 This is suggested by studies demonstrating stabilization and nuclear localization of the HIF1α protein in malignant plasma cells.12,13,15,16 Indeed, by employing gene-expression profiling, Colla et al.13 demonstrated that hypoxia affects the transcriptional and angiogenic profiles of myeloma cells, leading to increased expression of VEGFA and IL-8 among other pro-angiogenic factors. Interestingly, HIF1α protein stabilization and activity in MM cells may also occur under normoxic conditions and can, in collaboration with c-MYC, induce VEGFA-mediated angiogenesis.15 Together, these findings suggest a central role for the hypoxia-HIF1α axis in MM-related BM angiogenesis. In the present study, we further explored this possibility by studying the transcriptional response of MM cells to hypoxia, using gene-expression microarrays. Interestingly, we identified the pro-angiogenic factor adrenomedullin (AM) as the most highly hypoxia-induced gene in MM cells. In primary myelomas, AM expression was found to be increased during disease progression from MGUS to MM. In addition, endogenous, ectopically expressed and hypoxia-triggered AM secretion by primary MMs and MM cell lines enhanced angiogenesis. Of note, several MM cell lines and primary MMs expressed high levels of AM under normoxic conditions, suggesting regulation independent of the hypoxia- HIF1α axis. Taken together, our results identify AM as a potential driver of the angiogenic switch and promising therapeutic target in MM.

Materials and methods preparation of complementary rna, microarray hybridization and gene-expression profiling analysis RNA was extracted with the RNeasy Kit (Qiagen, Hilden, Germany) or the SV-total RNA extraction kit (Promega, Fitchburg, WI, USA) and Trizol (Invitrogen Life Technologies, Carslbad, CA, USA), in accordance with the manufacturer’s instructions. Biotinylated complementary RNA was amplified with a double in vitro transcription, according to the Affymetrix small sample labeling protocol vII (Affymetrix, Santa Clara, CA, USA). The biotinylated complementary RNA was fragmented and hybridized to the HG-U133 Plus 2.0 GeneChip oligonucleotide arrays according to the manufacturer’s instructions (Affymetrix). Fluorescence intensities were quantified and analyzed using the GCOS software (Affymetrix). Arrays were scaled to an average intensity of 100. Differentially expressed genes

146 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma… were identified by a Student’s t-test, and P-values were adjusted for multiple comparisons using the Benjamini and Hochberg correction. The threshold for significance was set to a P-value of ≤ 0.05. Among those genes, those with a fold change of ≥ 2 were retained. Among the genes with a fold change ≥ 2 in a given population, those with 100% absent call in this population were considered not to be biologi- cally relevant and were removed. The call (“present” or “absent”) is determined by Affymetrix GCOS software and indicates whether a gene is reliably expressed or not. For a global analysis of AM expression, gene-expression data publically available and deposited in the NIH Gene Expression Omnibus (GEO) National Center for Biotechnology Information (NCBI), http://www.ncbi.nlm.nih.gov/geo/ under accession number GSE2658, were used. These concerned the U133 Plus 2.0 Affymetrix oligonucleotide microarray data from 559 newly diagnosed MM patients included in total therapy 2/3(TT2, TT3), provided by the Donna D and Donald M Lambert Laboratory of Myeloma Genetics, University of Arkansas for Medical Sciences, Little Rock, AR, USA.17 1 Primary MM samples used for additional studies were obtained during routine diagnostic procedures at the Academic Medical Center, Amsterdam, the Nether- lands. Mononuclear cells were harvested by standard Ficoll/Paque density gradi- ent centrifugation (Amersham Pharmacia Biosciences, Roozendaal, the Neth- 3 erlands) and CD138+ cells were sorted by positive selection using anti-CD138 antibody (clone BB4, Instruchemie, Delfzijl, the Netherlands) and Dynabead- 4 conjugated goat anti-mouse IgG (Dynal, Oslo, Norway). 5 nf-κb profile The MM NF-κB profile was determined by Annuziata et al.18 Genes comprising 6 the NF-κB profile in MM were those that were decreased in expression by > 40% in at least six of eight time points following treatment of L363 cells with IKKβ inhibi- 7 tor (MLN120b) (Millennium Pharmaceuticals, Cambridge, MA, USA) for 8–24 h in three separate experiments (accession number GSE8487). Genes were chosen if they correlated in expression across the MM cell lines (r > 0.5). NF-κB profile genes, using Affymetrix U133 Plus 2.0 data, relied on the following probe sets: 210538_s_ at (BIRC3), 202644_s_at (TNFAIP3), 207535_s_at (NFKB2), 204116_at (IL2RG), 203927_at (NFKBIE), 205205_at (RELB), 201502_s_at (NFKBIA), 209619_at (CD74), 203471_s_at (PLEK), 210018_x_at (MALT1), 223709_s_at (WNT10A). cell culture Human myeloma cell lines (HMCLs) L363, UM-1, OPM-1, NCI-H929 and RPMI8226 were cultured in RPMI medium 1640 (Invitrogen Life Technologies)

147 Chapter 6

containing 10% clone I serum (HyClone, Waltham, MA, USA), 100 units per ml of penicillin, and 100 μg per ml of streptomycin. LME-1 cell line was cultured in IMDM medium (Invitrogen Life Technologies) supplemented with transferrin (20 μg/ml) and β-mercaptoethanol (50μM). An AM cDNA doxycycline- inducible NCI-H929 cell line (NCI-H929/TR/AM) was generated as described previously, using the T-REx System (Invitrogen Life technologies).19 AM overexpression was obtained by incubating NCI-H929/TR/AM cells for 24 h with 0.2 μg/ml doxycycline. Conditioned medium was harvested from 48 h cultures initiated at a concentration of 106 cells/ml in RPMI 10% FCS medium. Control medium (RPMI 10% FCS) was pretreated for 48h in 37°C next to conditioned medium. NF-κB stimulation was performed by culturing MM cells (at a concentration of 106 cells/ml in RPMI medium containing 10% FCS) for 48h with PMA

or TNFα (100ng/ml). The treatment with cobalt chloride (CoCl2) was performed by culturing MM cells (at a concentration of 106 cells/ml in RPMI medium containing 10% FCS) for 24 h with the final concentration of CoCl2 (100 uM). HUVECs were prepared from human umbilical cord veins as described previously.20 The adherent endothelial cells (culture flasks were coated with 1% gelatin) were maintained in RPMI medium (Invitrogen Life Technologies) containing 10% clone I serum (HyClone), 10% normal human serum, 3 μg/ml of basic fibroblast growth factor (bFGF), 100 units/ml of penicillin/streptomycin and 2mmol/l l-glutamine (complete medium), and incubated at 37°C in 5% CO2. At confluence, the cells were detached by trypsin and used in experiments before the sixth passage. rt-pcr Total RNA was isolated using Trizol according to the manufacturer’s protocol (Invitrogen Life Technologies). The RNA was further purified using isopro- panol precipitation and was concentrated using the RNeasy MinElute Clean- up kit (Qiagen). The quantity of total RNA was measured using a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Five microgram of total RNA was used for cDNA synthesis as described previously.21 The PCR mixture contained: 2 μl of cDNA, 1 x PCR Rxn buffer (In- vitrogen Life Technologies), 0.2 mmol/l dNTP, 2 mmol/l MgCl2, 0.2 μmol/l of each primer, and 1 U platinum Taq polymerase (Invitrogen Life Technologies). PCR conditions were: denaturing at 95°C for 5 min, followed by 30 cycles of 30 s at 95°C, 30s at 58°C and 30 s at 72°C. The reaction was completed for 10 min at 72°C. Primers used were: AM forward (5′-CTCTGAGTCGTGGGAAGAG G-3′); AM reverse (5′-CGTGTGCTTGTGGCTTAGAA-3′); PFKFB4 forward (5′-GGGATGGC GTCCCCACGGG-3′); PFKFB4 reverse (5′-CGCTCTCC

148 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

GTTCTCGGGTG-3′), AK3L forward (5′-TGCTGCCAGGCTAAGACAGTA CAA-3′); AK3L reverse (5′-TCTTCC TGGTTCTTCCATTGGGCA-3′) BNIP3 forward (5′-CACCTCGCTCGCAGACAC CAC-3′); BNIP3 reverse(5′-GAG AGCAGCAGAGATGGAAGGAAAAC-3′), VEGF forward (5′-CTCTACCT CCACCATGCCAAGT-3′); VEGF reverse (5′-ATCTGGTTC CCGAAAC CCTGAG-3′), CD74 forward (5′-TGACCAGCGCGACCTTATC-3′);CD74 reverse (5′-GAGCAGGTGCATCACATGGT-3′), RELB forward (5′-CATCG AGCTCCGGGATTGT-3′); RELB reverse (5′-CTTCAGGGACCCAGCGTT GTA-3′), calcitonin-receptor-like receptor (CRLR) forward (5′-TGCTCTGTG AAGGCATT TAC-3′); CRLR reverse (5′-CAGAATTGCTTGAACCTCTC-3′), RAMP2 forward (5′-GGACGGTGAAGAACTATGAG-3′); RAMP2 reverse (5′-ATCATGGCCAGGA GTACATC-3′), HPRT1 forward (5′-TGGCGTCG TGATTAGTGATG-3′); HPRT1 reverse (5′-TATCCAACACTTCGTGGGG T-3′). All primers were manufactured by Sigma-Aldrich (Haverhill, UK). 1 huvec proliferation and tube formation assay HUVECs were plated at a density of 5000 cells/well on pre-coated 96-well plates in 100 μl of RPMI 1640 medium containing 10% FCS and 10% NHS. The next day, the medium was removed and HUVECs were stimulated with MM condi- 3 tioned medium (CM), harvested from 48 h cultures initiated at a concentration of 106cells/ml in RPMI with 10% FCS. The AM inhibitors (37133 and 16311) 4 were obtained from the laboratory of Frank Cutitta and used at a final concentration of 1 μM.22 bFGF was used at a final concentration of 3 μg/ml. After 72h, the 5 number of living cells was determined based on a FSC/SSC dot plot. The results are presented as mean + / - s.d. of samples assayed in triplicate. All experiments 6 were performed at least three times. Student’s t-test was used for statistical data comparison. For the tube formation assay, HUVECs were seeded on matrigel in 7 15-well μ-slide angiogenesis plates, and were stimulated with MM conditioned medium. Cultures were then incubated for 6h at 37°C. At the end of the incubation period, from each well, three fields of view were photographed and the number of meshes per field was quantified. The data were expressed as mean ± s.d. of a representative experiment in triplicate. elisa and western blot The concentration of AM in MM supernatants was determined by ELISA (Phoe- nix Pharmaceutical, Burlingame, CA, USA). MM cell line conditioned media were obtained after 48 h culture at a concentration of 106cells/ml in RPMI 1% FCS medium. Control medium (RPMI 1% FCS) was pre-treated for 48 h at 37°C.

149 Chapter 6

Protein for immunoblotting was harvested from MM cell lines, separated by 10% SDS-polyacrylamide gel electrophoresis and subsequently blotted. The following antibodies were used for immunoblotting: anti-HIF1α polyclonal antibody (NB100-449, Novus Biologicals, Littleton, CO, USA) and anti-β-actin monoclonal antibody (clone AC-15, Sigma-Aldrich, St. Louis, MO, USA). Primary antibodies were detected by HRP-conjugated secondary antibodies, followed by detection using Lumi-Light PLUS western blotting substrate (Roche, Penzberg, Germany). statistical analysis Nonparametric statistics were used with Prism 5.0 software (Graphpad Software, San Diego, CA, USA). Spearman rank correlation coefficients were used to determine correlations (in 559 newly diagnosed MM patients included in the Total Therapy 2/3(TT2, TT3) trials between (i) expression of the AM gene and the MM hypoxia target genes, (ii) expression of the AM gene and the NF-κB profile genes, (iii) AM gene and the HOXB7 gene and (iv) expression of the AM gene and the ING4 gene. The Kruskal-Wallis test was used to compare AM gene expression in plasma cells derived from 559 newly diagnosed MM patients treated in the TT2 and TT3 trials, 4 MGUS patients and 22 healthy donors BMPC. Two-tailed t-tests were employed to analyze the in vitro data using Prism 5.0 software (Graph- pad Software, San Diego, CA, USA). A P-value of < 0.05 was considered to be statistically significant. The gene ontology (GO) biological pathways regulated by hypoxia were identified using DAVID software (Database for Annotations, Visualization and Integrated Discovery).23

Results transcriptional response of multiple myeloma cells to hypoxia To gain a global view of the transcriptional response of multiple myeloma cells to hypoxia, we compared the gene-expression profile of the HMCLs UM-1 and OPM-1 cultured under normoxic and hypoxic conditions for 16 h, using Affym- etrix human genome U133 plus 2.0 arrays. HIF1α stabilization, a characteristic response to hypoxic stimulation, was confirmed by immunoblotting (Figure 1A). Hypoxia resulted in a consistent > 2-fold increase in gene expression of 311 genes in UM-1 cells and 290 genes in OPM-1 cells (Supplementary Table 1 and 2). Im- portantly, the hypoxic gene-expression signature (Table 1) contained many genes

150 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma… that have previously been associated with hypoxia responses and/or represent established HIF1α target genes (underlined in Table 1). To identify if any GO classes were enriched in these two differentially expressed gene sets, GO analysis was performed using DAVID bioinformatics resource. The enriched GO “biological process” categories were found to be enriched with a P-value cutoff of P < 0.05 and a fold enrichment ≥ 2, and are shown in Supplementary Figures 1 and 2. Importantly, the “hypoxia” pathway, as well as pathways known to be regulated in response to hypoxia were found for both cell lines.

3 4 5 6 7 Figure 1. Transcriptional response of multiple myeloma cells to hypoxia (A) HIF1α accumulation in the HMCLs OPM-1 and UM-1 in response to hypoxia. HMCLs were cultured under normoxic (-) or hypoxic (+) (1% O2) conditions for 16 h. HIF1α accumulation was detected using western blot. β-actin expression is shown as an input control. (B) Expression of hypoxia target genes in the HMCLs, OPM-1 and UM-1 MM cell lines, as determined by RT-PCR following 16 h exposure to normoxia (-) or hypoxia (+) (1% O2). HPRT1 expression is shown as an input control.

Those pathways contained many genes that have previously been associated with hypoxia responses and/or represent established HIF1α target genes (underlined in Table 1). It comprised genes involved in the metabolic response to hypoxia, including glucose import (SLC2A6 and SLC2A1), glycolysis (PFKFB4, PFKFB3, ALDOC, ENO1, HK2 and HK1), downregulation of oxidative phosphorylation

151 Chapter 6

(PDK1 and MXI1); genes involved in cell proliferation and apoptosis (BNIP3, BNIP3L, PRF1, PLEKHF1 and APLP1), genes encoding transcription factors and signaling molecules (SREBF1, JUN, MAF, WDR54, WDR5B, RAB20, RAP1 and GAP) and several pro-angiogenic genes (VEGFA, AM and ANGPTL4). For AM, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase-4 (PFKFB4), BCL2/adenovirus E1B 19kDa interacting protein 3 (BNIP3), vascular endothelial growth factor (VEGF), and adenylate-kinase-3-like-1 (AK3L), hypoxia-mediated induction was confirmed by RT-PCR (Figure 1B). Among the hypoxia-repressed genes were a large group of genes linked to cell proliferation/DNA replication and protein synthesis, energy-costly processes that are typically repressed in response to hypoxic stress (Table 1).

Table 1. Common main genes induced by hypoxia in OPM-1 and UM-1 cell line

Main genes induced by hypoxia Main genes downregulated by hypoxia Extracellular glucose import Cell cycle and proliferation SLC2A6, SLC2A1 CDC27, CDC23, CDC20, MYC, PLK1, CDK6, HDAC9

Glycolytic breakdown of glucose Regulation of DNA replication PFKFB4, PFKFB3, ALDOC, ENO1, HK2, HK1 RFC3, TOP1, CCDC88A Inhibition of mitochondrial activity Regulation of translation PDK1, MXI1 EPRS, EIF2B3, FARSB, YARS2, EIF4G3 Angiogenesis — VEGFA, AM, ANGPTL4 — Apoptosis — BNIP3, BNIP3L, PRF1, PLEKHF1, APLP1 — Transcription factors — SREBF1, JUN, MAF — Signal transduction — WDR54, WDR5B, RAB20,RAP1,GAP AK3L — Response to oxidative stress — IDH1 — Chemokines and chemokine receptor — CXCR4, CCL28 — Main genes induced by hypoxia, as revealed using DAVID Gene Functional Classification Tool. Confirmed HIF-1 alpha target genes are underlined.

Interestingly, in both OPM-1 and UM-1 cells, AM was identified as the top- regulated hypoxia target gene, displaying a 48- and 57-fold induction, respectively. The fact that AM has been reported to be a transcriptional target gene of HIF1α24 and acts as a potent angiogenic factor in a number of solid tumors,25–29 prompted

152 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma… us to further investigate its expression, regulation and role in myeloma-induced angiogenesis. expression and regulation of am in hmcls and primary multiple myeloma cells To explore the expression and regulation of AM in MM, we initially assessed the AM mRNA and protein expression in a panel of HMCLs cultured under normoxic conditions (Figures 2A–C). While high or moderate AM mRNA levels were found in L363, LME-1 and RPMI8226, AM expression in the other HMCLs was either low or undetectable (OPM-1, NCI-H929 and UM-1) (Figure 2A).

3 4 5 6 7

Figure 2. Expression and regulation of AM in MM cells (A) Expression of AM mRNA in a panel of HMCLs. AM mRNA expression was analyzed by RT-PCR. HPRT1 was used as a loading control. (B) AM protein expression by HMCLs. Immunocytochemical staining of HMCLs with anti-AM antibody, showing intracytoplasmic expression (magnification: x 400). (C) AM protein secretion by HMCLs. AM concentration in the CM of HMCLs cells cultured for 48 h

153 Chapter 6

was measured by ELISA (n = 3). (D) Expression of Am and selected hypoxia target genes in the panel of HMCLs. Hypoxia target genes expression was analyzed by RT-PCR. HPRT1 expression is shown as input control. (E) HIF1α accumulation in HMCLs in response to hypoxia. HMCLs were cultured under normoxic (-) or hypoxic (+) (1% O2) conditions for 16 h. HIF1α accumulation was determined by immunoblotting. β-actin expression is shown as an input control. (F) Upper panel: AM mRNA expression in HMCLs exposed to hypoxia. Cells were cultured for 16 h under normoxic (-) or hypoxic (+) conditions. HprT1 expression is shown as an input control. Lower panel: AM mRNA expression in

HMCLs exposed to CoCl2. AM mRNA expression in HMCLs cultured in the absence (-) or presence

(+) of 100 μM CoCl2 for 24 h. Immunocytochemical AM protein detection matched these mRNA expression data (Figure 2B), showing strong cytoplasmic staining of L363, LME-1 and RPMI 8226, but no or very weak AM expression in the other cell lines. In addition, significant quantities of secreted AM protein could also be detected in the culture supernatants of L363 and LME-1 and, at lower levels, of RPMI8226 (Figure 2C). Hence, under normoxic culture conditions, HMCLs show highly variable levels of AM mRNA and protein expression. As AM is a transcriptional target of HIF1α,24 high expression of AM in HMCL under normoxic conditions could be a consequence of aberrant HIF1α pathway activation. If so, the AM-positive HMCLs L363, LME-1 and RPMI8226 would be predicted to be positive for an expression-signature comprising multiple HIF1α targets genes.30 However, AM expression in the tested HMCLs did not correlate with that of other HIF1α/hypoxia target genes (Figure 2D). This suggests that the normoxic AM expression in L363, LME-1 and RPMI8226 cells is regulated by a HIF1α-independent mechanism. Next, we subjected all HMCLs to hypoxia, which resulted in effective HIF1α stabilization (Figure 2E). Interestingly, AM expression was not only strongly induced in OPM-1 and UM-1 cells, but also in NCI-H929 and in RPMI8226 cells, with low and moderate normoxic AM gene expression, respectively. The already high normoxic AM expression in LME-1 and L363 cells was not further increased by hypoxia (Figure 2F upper panel). A similar induction of AM

expression in HMCLs was obtained with CoCl2 (Figure 2F lower panel). CoCl2 stabilizes HIF1α by inhibiting prolyl hydroxylase domain-containing proteins (PHDs), which, in normoxic conditions, hydroxylate HIF1α. This hydroxylation is required for interaction of HIF1α with the VHL protein, which targets HIF1α for degradation by the 26S proteasome.31 Thus, induction of AM expression by treatment with CoCl2 confirms the role of HIF1α in hypoxia-driven upregulation of AM in MM cells (Figure 2F).

154 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma… expression of am is related to multiple myeloma disease progression and molecular subgroup Consistent with the results obtained in HMCLs, primary MMs isolated and processed under standard, that is, non-hypoxic, conditions, also showed highly variable levels of AM mRNA expression (Figure 3A). As shown in Figure 3B, hypoxic stimulation of these primary MM cells resulted in an increased AM mRNA levels in cells with low baseline AM expression.

3 4 5 6 7

Figure 3. Expression of AM in primary MM samples (A) Expression of AM mRNA in purified primary MM cells. Total RNA was isolated from primary MM cells and RT-PCR was performed. HPRT1 was used as a loading control. (B) Regulation of AM expression in primary MM cells by hypoxia. Primary MM cells were cultured for 48 h in normoxic (-) or hypoxic (+) (1% O2) conditions and RT-PCR was performed. HPRT1 expression is shown as an input control. (C) AM expression in MM is related to the disease progression and molecular subgroup. AM mRNA expression in 559 newly diagnosed MM patients from TT2/3, 44 MGUS patients and 22 healthy donors BMPC samples assessed by U133plus 2.0 Affymetrix oligonucleotide. Left panel: AM expression in BMPCs, MGUs, and MM. AM expression in MM cells was increased compared with BMPC (P < 0.001) and MGUS cells (P < 0.001). Right panel: AM expression in different genetic MM subgroups. Compared with BMPCs, AM expression was significantly increased in PR, LB, HY, CD2 and MY subgroups (P < 0.001).

155 Chapter 6

To obtain a global view of AM expression in relation to disease progression and MM molecular subgroup, we analyzed Affymetrix oligonucleotide microarray data from a panel of 559 newly diagnosed MM patients included in total therapy 2/3 (TT2, TT3).17 Interestingly, this analysis revealed that, compared with expression in normal BMPC, AM expression is already elevated in a small subset of MGUS patients and is markedly increased in the majority of primary MMs (P < 0.001) (Figure 3C, left panel). Further analysis of AM expression in specific MM molecular subgroups, as previously identified by gene-expression profiling (PR, LB, MF, HY, CD1, CD2, MS amd MY),17 revealed significantly increased AM expression (P < 0.05) in the PR, LB, HY, CD2 and MY subgroups compared with BMPC and MGUS patients (Figure 3C, right panel). Hence, expression of AM by MM cells is related to disease progression and molecular subgroup. adrenomedullin expression in primary mms is not related to expression of other hypoxia target genes or nf-κb profile genes The data presented above suggest that aberrant expression of AM by MM cells can be driven by hypoxia-dependent as well as hypoxia-independent mechanisms. To further strengthen this notion, we assessed the relation between expression of AM and a panel of 11 “MM hypoxia signature genes” (Supplementary Table 3). These genes were selected because they were (i) highly induced by hypoxia in both OPM-1 and UM-1 cells (Supplementary Table 1 and 2), and significantly expressed in at least 10% of primary MM patients; (ii) confirmed to be hypoxia target genes in multiple independent studies; (iii) contained functional hypoxia response elements (HREs) in their promoter.30 Consistent with our observations in HMCLs, analysis of the TT2/TT3 data set17 revealed that AM expression in primary MM is not correlated to most of the MM hypoxia signature genes. The only significant correlations found were between AM and PDK1 and HK2 (Supplementary Ta- ble 4); however, the correlation coefficients were low (< 0.2). As the NF-κB pathway can drive angiogenesis32–34 and expression of pro-inflammatory cytokines, including cytokines that can regulate AM,35–37 we also assessed the relation between the NF-κB pathway and AM. As shown in Supplementary Table 4, analysis of the TT2/TT3 data set did not reveal an association between AM expression and expression of NF-κB profile genes, as defined by previous studies in MM.18 Consistent with this finding, stimulation of the NF-κB pathway in MM cell lines (UM-1 and OPM-1) with PMA or TNFα, as confirmed by the upregulation of established NF-κB target genes, including CD74 and RELB, did not enhance the (low) AM gene expression in these cells (Supplementary Figure 3).

156 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Other candidate regulators of AM expression and MM-induced angiogenesis are the homeobox gene HOXB7 and the tumor-suppressor gene inhibitor of growth family member 4 (ING4). HOXB7 can mediate tumor-induced angiogenesis,38 and its expression in MM cells is correlated with that of several pro-angiogenic factors.39 However, analysis of the TT2/3 primary MM patients’ data set, did not reveal a correlation between HOXB7 and AM gene expression (Supplementary Table 4). Similarly, expression of ING4, which represses angiogenesis in solid tumors40 and is downregulated in MM cells compared with that in normal plasma cells,41 was not correlated with AM expression (Supplemen- tary Table 4). adrenomedullin contributes to mm-induced angiogenesis MM cells generally express multiple angiogenic factors. Consequently, MM conditioned media (CM) show significant pro-angiogenic activity.11 To assess wheth- 1 er AM can contribute to this activity, we generated NCI-H929 cells with inducible overexpression of AM, by stably transfecting a doxycycline-inducible AM cDNA (NCI-H929/TR/AM). Doxycycline treatment of these cells resulted in a clear induction of AM mRNA, resulting in a seven-fold increase in AM protein level, 3 while doxycyclin treatment of control (NCI-H929/TR) cells did not affect AM expression levels (Figure 4A). Importantly, as shown in Figure 4B, CM derived 4 from doxycycline-treated NCI-H929/TR/AM cells enhanced proliferation of human umbilical vein endothelial cells (HUVECs), which express the AM recep- 5 tors CRLR and RAMP2-modifying protein (Supplementary Figure 4) by almost two-fold. Moreover, AM-enriched CM stimulated endothelial mesh formation 6 by almost five-fold (Figure 4C). Both HUVEC proliferation and angiogenesis were AM-specific, as both were completely abrogated by 37133 and 16311, two 7 highly specific small-molecule AM inhibitors that antagonize signaling by binding AM directly, thereby preventing receptor binding.22 Of note, by the use of the CM from doxycycline-treated control NCI-H929/TR cells, non-specific effects of doxycycline were excluded and the AM-specificity of the small molecule inhibitors 37133 and 16311 was confirmed (Figure 4B right panel). To address whether endogenously produced AM can also contribute to MM- induced angiogenesis, we next assessed the angiogenic properties of CM from L363 MM cells, which produce high levels of AM under normoxic conditions (Figure 2C). As shown in Figure 5, CM from L363 cells displayed potent angiogenic activity, promoting HUVEC proliferation, which was decreased two- fold by treatment with 16311 and 37133, respectively (Figure 5A) Furthermore,

157 Chapter 6 blocking AM by 16311 and 37133 decreased L363 CM induced mesh formation approximately two-fold (Figure 5B). Taken together, these data demonstrate that overexpression of AM by malignant plasma cells can significantly contribute to their pro-angiogenic properties.

Figure 4. Overexpression of AM enhances the pro-angiogenic activity of MM cells (A) Inducible expression of AM. NCI-H929 MM cells, containing either a doxycycline-inducible AM cDNA plus a TET repressor (NCI-H929/TR/AM) or a TET repressor only (NCI-H929/TR) were cultured without (-) or with (+) doxycycline for 24 h. AM expression was measured by RT-PCR (Left panel) and ELISA (Right panel). (B) AM overexpression enhances the pro-angiogenic activity of MM

158 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

supernatants Left panel. AM induction enhances HUVEC proliferation. HUVECs cells were cultured in supernatant of NCI-H929/TR/AM cells treated for 48 h without (- dox) or with (+ dox) doxycycline. This stimulating effect was fully blocked by 37133 and 16311, two small molecule inhibitors of AM (*P-value 0.05). The number of living cells in the control (supernatant from NCI-H929/TR/AM cells cultured for 48 h without doxycycline) was normalized to 100%. Right panel: specificity control: HUVECs cells were cultured in CM of NCI-H929/TR cells that had been treated for 48 h without (- dox) or with (+ dox) doxycycline. The number of living cells in the control (supernatant from NCI-H929/TR cells cultured for 48 h without doxycycline) was normalized to 100%. (C) AM promotes MM-induced angiogenesis. AM induction by doxycyclin strongly enhanced the formation of three-dimensional capillary- like tubular structures by HUVEC cells. HUVEC were cultured with CM from NCI-H929/TR/AM cell that had been cultured for 48 h without (- dox) or with (+ dox) doxycycline. The stimulating effect of the + dox CM was fully blocked by 37133 and 16311, two small molecule inhibitors of AM. Left panel: HUVEC cultures in matrigel showing endothelial mesh formation in cells cultured with CM from NCI-H929/TR/AM (- dox) or (+ dox) cells, in the presence or absence of the small molecule AM inhibitors 37133 and 16311. Right panel: quantification of endothelial mesh formation (*P-value < 0.05).

To determine whether hypoxia-driven AM secretion by primary MM cells can contribute directly to MM-induced angiogenesis, two independent primary MM cell samples were subjected to hypoxia and AM expression and angiogenic 1 potential were investigated. Exposure to hypoxia resulted in a clear increase in AM mRNA levels in both primary MM cell samples (Figure 5C), as well as in an increased AM protein secretion (1.6- and 3.2-fold in MM1 and 2, respectively) (Figure 5D). Functionally, this hypoxia-driven AM secretion resulted in 3 an approximately two-fold increase in HUVEC numbers, that was inhibited to normoxic or subnormoxic levels by blocking AM signaling with 16311. Impor- 4 tantly, 16311 did not influence bFGF-induced HUVEC proliferation, confirming specificity of the inhibition (Figure 5E). Taken together, these results demonstrate 5 that AM is a key mediator of angiogenesis in MM, strongly upregulated under hypoxic conditions. 6 7 Discussion

In this study, we explored the transcriptional response of MM cells to hypoxia, by comparing the gene expression profile of the HMCLs UM-1 and OPM-1 under normoxic and hypoxic conditions. The hypoxia-induced gene expression signature of these MM cells (Table 1, Supplementary Table 1 and 2, Figure 1) contained multiple genes that have previously been associated with hypoxia responses and/or represent established HIF1α target genes, including genes involved in the metabolic response to hypoxia, in cell proliferation and apoptosis. Furthermore, it comprised several pro-angiogenic genes, including VEGFA, ANGPTL4 and AM, of which AM was identified as the top-regulated gene (40-50-fold induction).

159 Chapter 6

Figure 5. Endogenous AM expression contibutes to MM induced angiogenesis (A) Contribution of endogenous AM to the mitogenic effect of CM from L363 cells. HUVEC proliferation was measured in CM in the presence or absence of the small molecule AM inhibitors 37133 and 16311. The mean ±s.d. of a representative experiment performed in triplicate is shown. (B) Contribu- tion of endogenous AM to the angiogenic effect of CM from L363 cells. Endothelial mesh formation

160 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

was measured in L363 CM in the presence or absence of the small molecule AM inhibitors 37133 and 16311. The mean ±s.d. of a representative experiment performed in triplicate is shown. (*P-value < 0.05; ***P-value < 0.001). (C) Relative expression of AM mRNA in purified primary MM cells. Total RNA was isolated from primary MM (pMM) cells and RT-PCR was performed. AM expression was normalized to HPRT1 gene expression. (D) AM protein secretion by primary MM cells. AM concentration in the CM of primary MM cells cultured for 48 h in normoxic and/or hypoxic conditions was measured by ELISA. (E) Contribution of pMM-derived endogenous and hypoxia-induced AM to HUVECs proliferation, in the presence or absence of the small-molecule AM inhibitor 16311. bFGF was used as a specificity control for 16311. (***P-value < 0.001). Interestingly, however, although AM is a well-established HIF1α target gene containing HRE sites in its promoter, as major regulatory sequences,42 we observed that several HMCLs and primary MMs also expressed high levels of AM under normoxic conditions (Figures 2A–C). Importantly, MM cells with normoxic AM expression did not show aberrant normoxic HIF1α stabilization and displayed no overexpression of other HIF1α/hypoxia target genes (Figures 2D–F), suggesting normoxic regulation of AM expression by HIF1α-independent mechanisms. In line with this notion, our analysis of a large MM gene-expression data set17 1 revealed no consistent correlation between AM expression and expression of other hypoxia/HIF1α target genes. These findings imply that mechanisms other than hypoxia can contribute to AM expression in malignant plasma cells, and are consistent with a scenario in which both HIF1α-dependent and independent 3 mechanisms contribute to the “angiogenic switch” in MM. We observed that AM expression in malignant plasma cells of MM patients is 4 significantly higher than in normal BMPC or plasma cells patients with MGUS (Figure 3C). This elevated AM expression was present in most molecular MM 5 subgroups (PR, LB, HY, CD2 and MY) (Figure 3D). Hence, AM overexpression is related to disease progression, suggesting that AM may have an important 6 functional contribution in angiogenesis, which is associated with progression of MGUS to clinically overt MM.43 Our functional studies strongly support this sce- 7 nario as they demonstrate that (i) endogenous, hypoxia-induced and ectopic expression of AM in MM cells strongly promotes angiogenic activity of MM cells, as shown by enhanced endothelial cell proliferation and mesh formation (Figures 4 and 5); (ii) blockage of AM strongly reduces the pro-angiogenic activity of MM cells (Figure 5). Adrenomedullin is involved in blood vessel morphogenesis, vasculogenesis and tumor angiogenesis.29,44 AM-null mice die in utero as a result of defective vasculogenesis,45 while AM overexpression by tumors mediates tumor angiogenesis.26,28,46 In addition, AM can directly promote tumor growth.25,47 AM stimulates angiogenesis by binding the CRLR, which is widely expressed on normal and hypoxic endothelial cells.48 Interestingly, binding of AM to the CRLR/RAMP2

161 Chapter 6

can transactivate the VEGFR-2, which is responsible for most pro-angiogenic effects of VEGFA49, including the stimulation of endothelial cell differentiation, proliferation, migration and morphogenesis. This AM-induced VEGFR-2 transactivation does not require VEGFA, suggesting that AM can functionally mimic VEGFA, and thereby contribute to MM-induced angiogenesis. Although it has recently been suggested that HIF1α is a major regulator of the pro-angiogenic profile of myeloma cells and of MM-induced angiogenesis, deregulation of other transcriptional pathways may potentially also contribute to the angiogenic switch and overexpression of AM in MM. Since pro-inflammatory cytokines have been shown to induce increased AM secretion,35–37 we explored the role of NF-κB signaling in the regulation of AM. However, expression of NF-κB profile genes and AM in primary MM patients (Supplementary Table 4) were not correlated. Moreover, stimulation of the NF-κB pathway in MM cells in vitro did not influence expression of AM (Supplementary Figure 3), suggesting that the NF-κB pathway is not involved in the regulation of AM in malignant plasma cells. Other candidate regulators of AM expression and MM-induced angiogenesis are the homeobox gene HOXB7 and the tumor-suppressor gene inhibitor of growth family member 4 (ING4). HOXB7 has been shown to mediate tumor-induced angiogenesis and tumor progression in solid cancers by regulating VEGF, IL-8, bFGF2 and Ang-2.38 In MM cells, HOXB7 expression is correlated with and can control overexpression of several pro-angiogenic factors.39 However, our analysis of the TT2/3 primary MM patients’ data set, did not reveal a correlation between HOXB7 and AM gene expression (Supplementary Table 4). Similarly, expression of ING4, which acts as a repressor of angiogenesis in solid tumors40 and shown reduced expression in MM cells compared with normal plasma cells,41 was not correlated with AM expression (Supplementary Table 4). Hence, neither NF-κB, HOXB7 nor ING4, appear to have a role in AM deregulation in MM plasma cells, and further studies are needed to unravel the molecular mechanism(s) involved. In conclusion, our results demonstrate that MM cells, both in a hypoxia-dependent and -independent fashion, aberrantly express and secrete AM, which can mediate MM-induced angiogenesis. This aberrant AM expression could be a major driving force for the angiogenic switch observed during MM progression, which renders AM a putative target for anti-angiogenic therapy in MM.

162 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

References

1. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2: 175–187. 2. Hideshima T, Mitsiades C, Tonon G, Richardson PG, Anderson KC. Understanding multiple myeloma pathogenesis in the bone marrow to identify new therapeutic targets. Nat Rev Cancer 2007; 7: 585–598. 3. Vacca A, Ribatti D. Bone marrow angiogenesis in multiple myeloma. Leukemia 2006; 20: 193–199. 4. Giuliani N, Storti P, Bolzoni M, Palma BD, Bonomini S. Angiogenesis and multiple myeloma. Cancer Microenviron 2011; 4: 325–337. 5. Rana C, Sharma S, Agrawal V, Singh U. Bone marrow angiogenesis in multiple myeloma and its correlation with clinicopathological factors. Ann Hematol 2010; 89: 789–794. 6. Bhatti SS, Kumar L, Dinda AK, Dawar R. Prognostic value of bone marrow angiogenesis in multiple myeloma: use of light microscopy as well as computerized image 1 analyzer in the assessment of microvessel density and total vascular area in multiple myeloma and its correlation with various clinical, histological, and laboratory parameters. Am J Hematol 2006; 81: 649–656. 7. Sezer O, Niemoller K, Kaufmann O, Eucker J, Jakob C, et al. Decrease of bone mar- 3 row angiogenesis in myeloma patients achieving a remission after chemotherapy. Eur J Haematol 2001; 66: 238–244. 4 8. Kumar S, Fonseca R, Dispenzieri A, Lacy MQ, Lust JA, et al. Bone marrow angiogenesis in multiple myeloma: effect of therapy. Br J Haematol 2002; 119: 665–671. 9. Ribatti D, Nico B, Vacca A. Importance of the bone marrow microenvironment in in- 5 ducing the angiogenic response in multiple myeloma. Oncogene 2006; 25: 4257–4266. 10. Kumar S, Witzig TE, Timm M, Haug J, Wellik L, et al. Bone marrow angiogenic ability 6 and expression of angiogenic cytokines in myeloma: evidence favoring loss of marrow angiogenesis inhibitory activity with disease progression. Blood 2004; 104: 1159–1165. 7 11. Hose D, Moreaux J, Meissner T, Seckinger A, Goldschmidt H, et al. Induction of angiogenesis by normal and malignant plasma cells. Blood 2009; 114: 128–143. 12. Martin SK, Diamond P, Gronthos S, Peet DJ, Zannettino AC. The emerging role of hypoxia, HIF-1 and HIF-2 in multiple myeloma. Leukemia 2011; 25: 1533–1542. 13. Colla S, Storti P, Donofrio G, Todoerti K, Bolzoni M, et al. Low bone marrow oxygen tension and hypoxia-inducible factor-1alpha overexpression characterize patients with multiple myeloma: role on the transcriptional and proangiogenic profiles of CD138 (+) cells. Leukemia 2010; 24: 1967–1970. 14. Asosingh K, De Raeve H, de Ridder M, Storme GA, Willems A, et al. Role of the hypoxic bone marrow microenvironment in 5T2MM murine myeloma tumor progression. Haematologica 2005; 90: 810–817.

163 Chapter 6

15. Zhang J, Sattler M, Tonon G, Grabher C, Lababidi S, et al. Targeting angiogenesis via a c-Myc/hypoxia-inducible factor-1alpha-dependent pathway in multiple myeloma. Cancer Res 2009; 69: 5082–5090. 16. Giatromanolaki A, Bai M, Margaritis D, Bourantas KL, Koukourakis MI, et al. Hy- poxia and activated VEGF/receptor pathway in multiple myeloma. Anticancer Res 2010; 30: 2831–2836. 17. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108: 2020–2028. 18. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12:115–130. 19. Reijmers RM, Groen RW, Rozemuller H, Kuil A, de Haan-Kramer A, et al. Targeting EXT1 reveals a crucial role for heparan sulfate in the growth of multiple myeloma. Blood 2010; 115: 601–604. 20. Baudin B, Bruneel A, Bosselut N, Vaubourdolle M. A protocol for isolation and culture of human umbilical vein endothelial cells. Nat Protoc 2007; 2: 481–485. 21. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci U S A 2004; 101: 6122–6127. 22. Martínez A, Julian M, Bregonzio C, Notari L, Moody TW, et al. Identification of vaso- active nonpeptidic positive and negative modulators of adrenomedullin using a neutralizing antibody-based screening strategy. Endocrinology 2004; 145: 3858–3865. 23. Huang da W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 2009; 37: 1–13. 24. Takenaga K. Angiogenic signaling aberrantly induced by tumor hypoxia. Front Bi- osci 2011; 16: 31–48. 25. Ouafik L, Sauze S, Boudouresque F, Chinot O, Delfino C, et al. Neutralization of adrenomedullin inhibits the growth of human glioblastoma cell lines in vitro and suppresses tumor xenograft growth in vivo. Am J Pathol 2002; 160: 1279–1792. 26. Oehler MK, Hague S, Rees MC, Bicknell R. Adrenomedullin promotes formation of xenografted endometrial tumors by stimulation of autocrine growth and angiogenesis. Oncogene 2002; 21: 2815–2821. 27. Kaafarani I, Fernandez-Sauze S, Berenguer C, Chinot O, Delfino C, et al. Targeting adrenomedullin receptors with systemic delivery of neutralizing antibodies inhibits tumor angiogenesis and suppresses growth of human tumor xenografts in mice. FASEB J 2009; 23: 3424–3435. 28. Ishikawa T, Chen J, Wang J, Okada F, Sugiyama T, et al. Adrenomedullin antagonist suppresses in vivo growth of human pancreatic cancer cells in SCID mice by suppressing angiogenesis. Oncogene 2003; 22: 1238–1242.

164 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

29. Deville JL, Salas S, Figarella-Branger D, Ouafik L, Daniel L. Adrenomedullin as a therapeutic target in angiogenesis. Expert Opin Ther Targets 2010; 14:1059–1072. 30. Benita Y, Kikuchi H, Smith AD, Zhang MQ, Chung DC, et al. An integrative genom- ics approach identifies Hypoxia Inducible Factor-1 (HIF-1)-target genes that form the core response to hypoxia. Nucleic Acids Res 2009; 37: 4587–4602. 31. Yu F, White SB, Zhao Q, Lee FS. HIF-1alpha binding to VHL is regulated by stimulus- sensitive proline hydroxylation. Proc Natl Acad Sci U S A 2001; 98: 9630–9635. 32. Tabruyn SP, Griffioen AW, NF-kappa B. A new player in angiostatic therapy. Angio- genesis 2008; 11: 101–106. 33. Schmidt D, Textor B, Pein OT, Licht AH, Andrecht S, et al. Critical role for NF- kappaB-induced JunB in VEGF regulation and tumor angiogenesis. EMBOJ 2007; 26: 710–719. 34. Huang S, Pettaway CA, Uehara H, Bucana CD, Fidler IJ. Blockade of NF-kappaB activity in human prostate cancer cells is associated with suppression of angiogenesis, invasion, and metastasis. Oncogene 2001; 20: 4188–4197. 35. Takahashi K, Nakayama M, Totsune K, Murakami O, Sone M, et al. Increased secre- 1 tion of adrenomedullin from cultured human astrocytes by cytokines. J Neurochem 2000; 74: 99–103. 36. Sugo S, Minamino N, Shoji H, Kangawa K, Kitamura K, et al. Production and secretion of adrenomedullin from vascular smooth muscle cells: augmented production 3 by tumor necrosis factor-alpha. Biochem Biophys Res Commun 1994; 203: 719–726. 37. Horio T, Nishikimi T, Yoshihara F, Nagaya N, Matsuo H, et al. Production and secre- 4 tion of adrenomedullin in cultured rat cardiac myocytes and non-myocytes: stimulation by interleukin-1beta and tumor necrosis factor-alpha. Endocrinology 1998; 139: 4576–4580. 5 38. Care A, Felicetti F, Meccia E, Bottero L, Parenza M, et al. HOXB7: a key factor for tumor-associated angiogenic switch. Cancer Res 2001; 61: 6532–6539. 6 39. Storti P, Donofrio G, Colla S, Airoldi I, Bolzoni M, et al. HOXB7 expression by myeloma cells regulates their pro-angiogenic properties in multiple myeloma patients. 7 Leukemia 2011; 25: 527–537. 40. Garkavtsev I, Kozin SV, Chernova O, Xu L, Winkler F, et al. The candidate tumour suppressor protein ING4 regulates brain tumour growth and angiogenesis. Nature 2004; 428: 328–332. 41. Colla S, Tagliaferri S, Morandi F, Lunghi P, Donofrio G, et al. The new tumor-suppressor gene inhibitor of growth family member 4 (ING4) regulates the production of proangiogenic molecules by myeloma cells and suppresses hypoxia-inducible factor-1 alpha (HIF-1alpha) activity: involvement in myeloma-induced angiogenesis. Blood 2007; 110: 4464–4475. 42. Garayoa M, Martinez A, Lee S, Pio R, An WG, et al. Hypoxia-inducible factor-1 (HIF-1) up-regulates adrenomedullin expression in human tumor cell lines during

165 Chapter 6

oxygen deprivation: a possible promotion mechanism of carcinogenesis. Mol Endo- crinol 2000; 14: 848–862. 43. Jakob C, Sterz J, Zavrski I, Heider U, Kleeberg L, et al. Angiogenesis in multiple myeloma. Eur J Cancer 2006; 42: 1581–1590. 44. Kim W, Moon SO, Sung MJ, Kim SH, Lee S, et al. Angiogenic role of adrenomedullin through activation of Akt, mitogen-activated protein kinase, and focal adhesion kinase in endothelial cells. FASEB J 2003; 17: 1937–1939. 45. Caron KM, Smithies O. Extreme hydrops fetalis and cardiovascular abnormalities in mice lacking a functional Adrenomedullin gene. Proc Natl Acad Sci U S A 2001; 98: 615–619. 46. Martinez A, Vos M, Guedez L, Kaur G, Chen Z, et al. The effects of adrenomedullin overexpression in breast tumor cells. J Natl Cancer Inst 2002; 94: 1226–1237. 47. Ramachandran V, Arumugam T, Hwang RF, Greenson JK, Simeone DM, et al. Adre- nomedullin is expressed in pancreatic cancer and stimulates cell proliferation and invasion in an autocrine manner via the adrenomedullin receptor, ADMR. Cancer Res 2007; 67: 2666–2675. 48. Fernandez-Sauze S, Delfino C, Mabrouk K, Dussert C, Chinot O, et al. Effects of adrenomedullin on endothelial cells in the multistep process of angiogenesis: involvement of CRLR/RAMP2 and CRLR/RAMP3 receptors. Int J Cancer 2004; 108: 797–804. 49. Guidolin D, Albertin G, Spinazzi R, Sorato E, Mascarin A, et al. Adrenomedullin stimulates angiogenic response in cultured human vascular endothelial cells: involvement of the vascular endothelial growth factor receptor 2. Peptides 2008; 29: 2013–2023.

166 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma… supplementary material

3 4 5 6 7

Supplementary Figure 1 Gene ontology groups with significant over/under-representation among genes differentially expressed in hypoxic conditions (fold change ≥ 2 or ≤ 0,5) in OPM-1 cells. Only the pathways having significant alteration (p < 0.05) are presented. For each GO pathway, the bar shows the x-fold enrichment of the pathway in the dataset.

167 Chapter 6

Supplementary Figure 2 Gene ontology groups with significant over/under-representation among genes differentially expressed in hypoxic conditions (fold change ≥ 2 or ≤ 0,5) in UM-1 cells. Only the pathways having significant alteration (p < 0.05) are presented. For each GO pathway, the bar shows the x-fold enrichment of the pathway in the dataset.

Supplementary Figure 3. Response of MM cell lines to NF-кB stimulation RT-PCR analysis of NF-кB target genes and AM expression in OPM-1 and UM-1 MM cell lines following 48h exposure to PMA and/or TNF alpha (100ng/ml). HPRT1 expression is shown as an input control.

168 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Supplementary Figure 4. Expression of AM receptors in HUVECs cells Reverse transcriptase-PCR analysis for AM receptors: CRLR and RAMP2 modifying protein. HPRT1 expression is shown as an input control. As a positive control (positive) for RT-PCR the cDNA from kidney was used, as negative control (negative) water was used.

Supplementary Table 1. OPM-1 cell line Probe Set ID Score(d) Gene Symbol Gene Title 1 202912_at 48,48 ADM adrenomedullin 203828_s_at 40,08 IL32 interleukin 32 223836_at 36,93 FGFBP2 fibroblast growth factor binding protein 2 228499_at 23,46 PFKFB4 6-phosphofructo-2-kinase/fructose-2,6-biphos- 3 phatase 4 214978_s_at 22,91 PPFIA4 protein tyrosine phosphatase, receptor type, f polypeptide (PTPRF), interacting protein (liprin), 4 alpha 4 202887_s_at 10,39 DDIT4 DNA-damage-inducible transcript 4 5 204347_at 10,17 LOC645619 /// similar to Adenylate kinase isoenzyme 4, mito- LOC731007 chondrial (ATP-AMP transphosphorylase) /// similar to Adenylate kinase isoenzyme 4, mitochondrial 6 (Adenylate kinase 3-like 1) (ATP-AMP transphosphorylase) 7 202022_at 10,10 ALDOC aldolase C, fructose-bisphosphate 1560112_at 9,29 WDFY2 WD repeat and FYVE domain containing 2 201170_s_at 9,04 BHLHB2 basic helix-loop-helix domain containing, class B, 2 225342_at 8,38 AK3L1 adenylate kinase 3-like 1 201212_at 7,73 LGMN legumain 203438_at 6,93 STC2 stanniocalcin 2 207543_s_at 6,30 P4HA1 procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase), alpha polypeptide I 226985_at 6,23 FGD5 FYVE, RhoGEF and PH domain containing 5 201849_at 6,06 BNIP3 BCL2/adenovirus E1B 19kDa interacting protein 3 202718_at 5,51 IGFBP2 insulin-like growth factor binding protein 2, 36kDa

169 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 203027_s_at 4,98 MVD mevalonate (diphospho) decarboxylase 221214_s_at 4,84 NELF nasal embryonic LHRH factor 226682_at 4,82 LOC283666 hypothetical protein LOC283666 210512_s_at 4,75 VEGFA vascular endothelial growth factor A 226549_at 4,68 SBK1 SH3-binding domain kinase 1 215616_s_at 4,62 JMJD2B jumonji domain containing 2B 202934_at 4,62 HK2 hexokinase 2 202481_at 4,60 DHRS3 dehydrogenase/reductase (SDR family) member 3 226452_at 4,55 PDK1 pyruvate dehydrogenase kinase, isozyme 1 202497_x_at 4,42 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 227337_at 4,35 ANKRD37 ankyrin repeat domain 37 203282_at 4,22 GBE1 glucan (1,4-alpha-), branching enzyme 1 (glycogen branching enzyme, Andersen disease, glycogen storage disease type IV) 202464_s_at 4,16 PFKFB3 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 235850_at 4,13 WDR5B WD repeat domain 5B 205349_at 3,92 GNA15 guanine nucleotide binding protein (G protein), alpha 15 (Gq class) 212689_s_at 3,82 JMJD1A jumonji domain containing 1A 212813_at 3,76 JAM3 junctional adhesion molecule 3 219622_at 3,68 RAB20 RAB20, member RAS oncogene family 218205_s_at 3,67 MKNK2 MAP kinase interacting serine/threonine kinase 2 200697_at 3,67 HK1 hexokinase 1 223046_at 3,64 EGLN1 egl nine homolog 1 (C. elegans) 234312_s_at 3,57 ACSS2 acyl-CoA synthetase short-chain family member 2 221478_at 3,52 BNIP3L BCL2/adenovirus E1B 19kDa interacting protein 3-like 41386_i_at 3,47 JMJD3 jumonji domain containing 3 218274_s_at 3,36 ANKZF1 ankyrin repeat and zinc finger domain containing 1 225533_at 3,33 PHF19 PHD finger protein 19 222856_at 3,33 APLN apelin, AGTRL1 ligand 226160_at 3,32 H6PD hexose-6-phosphate dehydrogenase (glucose 1-dehydrogenase) 225249_at 3,30 SPPL2B signal peptide peptidase-like 2B 225530_at 3,30 MOBKL2A MOB1, Mps One Binder kinase activator-like 2A (yeast) 222446_s_at 3,26 BACE2 beta-site APP-cleaving enzyme 2

170 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 200831_s_at 3,24 SCD stearoyl-CoA desaturase (delta-9-desaturase) 211902_x_at 3,23 TRA@ T cell receptor alpha locus 227799_at 3,23 MYO1G myosin IG 210130_s_at 3,14 TM7SF2 transmembrane 7 superfamily member 2 201389_at 3,09 ITGA5 integrin, alpha 5 (fibronectin receptor, alpha polypeptide) 226971_at 3,09 CCDC136 coiled-coil domain containing 136 203823_at 3,08 RGS3 regulator of G-protein signalling 3 1569496_s_at 3,07 SPON2 Spondin 2, extracellular matrix protein 239004_at 3,07 SQSTM1 Sequestosome 1 240258_at 2,96 ENO1 enolase 1, (alpha) 202245_at 2,95 LSS lanosterol synthase (2,3-oxidosqualene-lanosterol cyclase) 203911_at 2,94 RAP1GAP RAP1 GTPase activating protein 219014_at 2,93 PLAC8 placenta-specific 8 1 235231_at 2,92 ZNF789 zinc finger protein 789 207030_s_at 2,91 CSRP2 cysteine and glycine-rich protein 2 200965_s_at 2,89 ABLIM1 actin binding LIM protein 1 235900_at 2,88 SPNS3 spinster homolog 3 (Drosophila) 3 209034_at 2,87 PNRC1 proline-rich nuclear receptor coactivator 1 207351_s_at 2,85 SH2D2A SH2 domain protein 2A 4 215898_at 2,85 TTLL5 tubulin tyrosine ligase-like family, member 5 206655_s_at 2,83 GP1BB /// glycoprotein Ib (platelet), beta polypeptide /// 5 SEPT5 septin 5 212518_at 2,83 PIP5K1C phosphatidylinositol-4-phosphate 5-kinase, type I, gamma 6 221964_at 2,83 TULP3 tubby like protein 3 208308_s_at 2,80 GPI glucose phosphate isomerase 7 228741_s_at 2,80 HCN3 hyperpolarization activated cyclic nucleotide-gated potassium channel 3 208154_at 2,79 LOC51336 mesenchymal stem cell protein DSCD28 202308_at 2,79 SREBF1 sterol regulatory element binding transcription factor 1 219862_s_at 2,79 NARF nuclear prelamin A recognition factor 210132_at 2,78 EFNA3 ephrin-A3 218507_at 2,77 HIG2 hypoxia-inducible protein 2 226446_at 2,76 HES6 hairy and enhancer of split 6 (Drosophila) 208322_s_at 2,75 ST3GAL1 ST3 beta-galactoside alpha-2,3-sialyltransferase 1 200696_s_at 2,73 GSN gelsolin (amyloidosis, Finnish type)

171 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 200632_s_at 2,72 NDRG1 N-myc downstream regulated gene 1 224473_x_at 2,70 LZTS2 leucine zipper, putative tumor suppressor 2 242361_at 2,70 IMMT Inner membrane protein, mitochondrial (mitofilin) 220091_at 2,69 SLC2A6 solute carrier family 2 (facilitated glucose transporter), member 6 209462_at 2,68 APLP1 amyloid beta (A4) precursor-like protein 1 212218_s_at 2,67 FASN fatty acid synthase 219752_at 2,66 RASAL1 RAS protein activator like 1 (GAP1 like) 221011_s_at 2,66 LBH limb bud and heart development homolog (mouse) 223326_s_at 2,66 FLJ22795 /// hypothetical protein FLJ22795 /// similar to cis- LOC727751 Golgi matrix protein GM130 229902_at 2,65 FLT4 fms-related tyrosine kinase 4 1555842_at 2,65 LOC284356 hypothetical protein LOC284356 218625_at 2,64 NRN1 neuritin 1 203946_s_at 2,64 ARG2 arginase, type II 223621_at 2,63 PNMA3 paraneoplastic antigen MA3 225301_s_at 2,62 MYO5B myosin VB 229055_at 2,62 GPR68 G protein-coupled receptor 68 221757_at 2,61 PIK3IP1 phosphoinositide-3-kinase interacting protein 1 219815_at 2,61 GAL3ST4 galactose-3-O-sulfotransferase 4 204981_at 2,60 SLC22A18 solute carrier family 22 (organic cation transporter), member 18 211037_s_at 2,60 LENG4 leukocyte receptor cluster (LRC) member 4 243363_at 2,60 LOC641518 hypothetical protein LOC641518 201313_at 2,58 ENO2 enolase 2 (gamma, neuronal) 204994_at 2,58 MX2 myxovirus (influenza virus) resistance 2 (mouse) 223640_at 2,56 HCST hematopoietic cell signal transducer 1556599_s_at 2,55 ARPP-21 cyclic AMP-regulated phosphoprotein, 21 kD 225542_at 2,55 CENTB5 centaurin, beta 5 218697_at 2,55 NCKIPSD NCK interacting protein with SH3 domain 201791_s_at 2,54 DHCR7 7-dehydrocholesterol reductase 210220_at 2,54 FZD2 frizzled homolog 2 (Drosophila) 204164_at 2,54 SIPA1 signal-induced proliferation-associated gene 1 204365_s_at 2,53 REEP1 receptor accessory protein 1 212496_s_at 2,51 JMJD2B jumonji domain containing 2B 225262_at 2,49 FOSL2 FOS-like antigen 2 212430_at 2,48 RBM38 RNA binding motif protein 38 212045_at 2,48 GLG1 golgi apparatus protein 1

172 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 217838_s_at 2,48 EVL Enah/Vasp-like 231720_s_at 2,48 JAM3 junctional adhesion molecule 3 241954_at 2,47 FDFT1 Farnesyl-diphosphate farnesyltransferase 1 214177_s_at 2,46 PBXIP1 pre-B-cell leukemia homeobox interacting protein 1 235226_at 2,46 CDC2L6 cell division cycle 2-like 6 (CDK8-like) 231124_x_at 2,45 LY9 lymphocyte antigen 9 236275_at 2,45 KRBA1 KRAB-A domain containing 1 227353_at 2,45 TMC8 Transmembrane channel-like 8 212561_at 2,44 RAB6IP1 RAB6 interacting protein 1 200737_at 2,44 PGK1 phosphoglycerate kinase 1 203894_at 2,43 TUBG2 tubulin, gamma 2 211924_s_at 2,43 PLAUR plasminogen activator, urokinase receptor 1553611_s_at 2,42 FLJ33790 hypothetical protein FLJ33790 1562271_x_at 2,42 ARHGEF7 Rho guanine nucleotide exchange factor (GEF) 7 1 221778_at 2,42 JHDM1D jumonji C domain-containing histone demethylase 1 homolog D (S. cerevisiae) 1565935_at 2,42 LOC91431 prematurely terminated mRNA decay factor-like 232946_s_at 2,42 NADSYN1 NAD synthetase 1 3 227383_at 2,42 LOC727820 hypothetical protein LOC727820 210711_at 2,40 MGC5457 hypothetical protein MGC5457 4 238740_at 2,39 AARSD1 alanyl-tRNA synthetase domain containing 1 208926_at 2,39 NEU1 sialidase 1 (lysosomal sialidase) 1554977_at 2,38 LOC198437 bA299N6.3 5 223216_x_at 2,37 FBXO16 /// zinc finger protein 395 /// F-box protein 16 ZNF395 6 203937_s_at 2,35 TAF1C TATA box binding protein (TBP)-associated factor, RNA polymerase I, C, 110kDa 7 209889_at 2,35 SEC31B SEC31 homolog B (S. cerevisiae) 49679_s_at 2,35 MMP24 Matrix metallopeptidase 24 (membrane-inserted) 223727_at 2,34 KCNIP2 Kv channel interacting protein 2 224565_at 2,34 TncRNA trophoblast-derived noncoding RNA 230619_at 2,34 ARNT aryl hydrocarbon receptor nuclear translocator 36554_at 2,34 ASMTL acetylserotonin O-methyltransferase-like 213270_at 2,34 MPP2 membrane protein, palmitoylated 2 (MAGUK p55 subfamily member 2) 226728_at 2,33 SLC27A1 solute carrier family 27 (fatty acid transporter), member 1 244385_at 2,32 JMJD2C Jumonji domain containing 2C

173 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 239137_x_at 2,31 MGC45491 hypothetical protein MGC45491 226390_at 2,31 STARD4 START domain containing 4, sterol regulated 238551_at 2,30 FUT11 fucosyltransferase 11 (alpha (1,3) fucosyltransferase) 219236_at 2,30 PAQR6 progestin and adipoQ receptor family member VI 218019_s_at 2,30 PDXK pyridoxal (pyridoxine, vitamin B6) kinase 214667_s_at 2,30 TP53I11 tumor protein p53 inducible protein 11 208964_s_at 2,30 FADS1 fatty acid desaturase 1 227281_at 2,30 SLC29A4 solute carrier family 29 (nucleoside transporters), member 4 229872_s_at 2,30 LOC642441 /// hypothetical LOC642441 /// hypothetical protein LOC730256 /// LOC730256 /// hypothetical protein LOC730257 LOC730257 217783_s_at 2,29 YPEL5 yippee-like 5 (Drosophila) 202769_at 2,29 CCNG2 cyclin G2 218756_s_at 2,28 MGC4172 short-chain dehydrogenase/reductase 228906_at 2,28 CXXC6 CXXC finger 6 225191_at 2,28 CIRBP cold inducible RNA binding protein 222175_s_at 2,27 PCQAP PC2 (positive cofactor 2, multiprotein complex) glutamine/Q-rich-associated protein 213060_s_at 2,26 CHI3L2 chitinase 3-like 2 209566_at 2,26 INSIG2 insulin induced gene 2 213011_s_at 2,26 TPI1 triosephosphate isomerase 1 201673_s_at 2,26 GYS1 glycogen synthase 1 (muscle) 203238_s_at 2,25 NOTCH3 Notch homolog 3 (Drosophila) 1566720_at 2,25 LOC376693 hypothetical LOC376693 209218_at 2,25 SQLE squalene epoxidase 209348_s_at 2,25 MAF v-maf musculoaponeurotic fibrosarcoma oncogene homolog (avian) 201251_at 2,25 PKM2 pyruvate kinase, muscle 219499_at 2,25 SEC61A2 Sec61 alpha 2 subunit (S. cerevisiae) 228730_s_at 2,25 SCRN2 secernin 2 50965_at 2,24 RAB26 RAB26, member RAS oncogene family 201556_s_at 2,24 VAMP2 vesicle-associated membrane protein 2 (synapto- brevin 2) 1552455_at 2,23 PRUNE2 prune homolog 2 (Drosophila) 226068_at 2,23 SYK Spleen tyrosine kinase 209640_at 2,23 PML promyelocytic leukemia 209962_at 2,22 EPOR erythropoietin receptor

174 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 203521_s_at 2,22 ZNF318 zinc finger protein 318 204118_at 2,22 CD48 CD48 molecule 218030_at 2,22 GIT1 G protein-coupled receptor kinase interactor 1 210010_s_at 2,22 SLC25A1 solute carrier family 25 (mitochondrial carrier; citrate transporter), member 1 1556058_s_at 2,22 SPEN spen homolog, transcriptional regulator (Dros- ophila) 200808_s_at 2,21 ZYX zyxin 222395_s_at 2,21 UBE2Z ubiquitin-conjugating enzyme E2Z (putative) 228296_at 2,21 YPEL1 yippee-like 1 (Drosophila) 219566_at 2,20 PLEKHF1 pleckstrin homology domain containing, family F (with FYVE domain) member 1 206467_x_at 2,20 RTEL1 /// tumor necrosis factor receptor superfamily, mem- TNFRSF6B ber 6b, decoy /// regulator of telomere elongation helicase 1 1 203317_at 2,20 PSD4 pleckstrin and Sec7 domain containing 4 203394_s_at 2,20 HES1 hairy and enhancer of split 1, (Drosophila) 36907_at 2,20 MVK mevalonate kinase (mevalonic aciduria) 1553101_a_at 2,19 ALKBH5 alkB, alkylation repair homolog 5 (E. coli) 204379_s_at 2,19 FGFR3 fibroblast growth factor receptor 3 (achondroplasia, 3 thanatophoric dwarfism) 210854_x_at 2,19 SLC6A8 solute carrier family 6 (neurotransmitter transporter, 4 creatine), member 8 235729_at 2,19 ZNF514 zinc finger protein 514 5 203950_s_at 2,19 CLCN6 chloride channel 6 1554077_a_at 2,19 TMEM53 transmembrane protein 53 6 233358_at 2,18 FLJ14311 hypothetical gene FLJ14311 1555843_at 2,18 HNRPM Heterogeneous nuclear ribonucleoprotein M 7 214755_at 2,18 UAP1L1 UDP-N-acteylglucosamine pyrophosphorylase 1-like 1 223378_at 2,18 GLIS2 GLIS family zinc finger 2 208807_s_at 2,18 CHD3 chromodomain helicase DNA binding protein 3 212567_s_at 2,18 MAP4 microtubule-associated protein 4 218149_s_at 2,17 ZNF395 zinc finger protein 395 231823_s_at 2,16 SH3PXD2B SH3 and PX domains 2B 224602_at 2,16 LOC401152 HCV F-transactivated protein 1 204044_at 2,16 QPRT quinolinate phosphoribosyltransferase (nicotinate- nucleotide pyrophosphorylase (carboxylating)) 222746_s_at 2,15 BSPRY B-box and SPRY domain containing 41657_at 2,15 STK11 serine/threonine kinase 11

175 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 219188_s_at 2,15 LRP16 LRP16 protein 203366_at 2,14 POLG polymerase (DNA directed), gamma 200827_at 2,14 PLOD1 procollagen-lysine 1, 2-oxoglutarate 5-dioxygenase 1 201749_at 2,14 ECE1 Endothelin converting enzyme 1 202328_s_at 2,13 PKD1 polycystic kidney disease 1 (autosomal dominant) 202068_s_at 2,13 LDLR low density lipoprotein receptor (familial hypercholesterolemia) 200872_at 2,13 S100A10 S100 calcium binding protein A10 223666_at 2,12 SNX5 Sorting nexin 5 202624_s_at 2,12 CABIN1 calcineurin binding protein 1 35617_at 2,12 MAPK7 mitogen-activated protein kinase 7 229001_at 2,11 PPP1R3E Protein phosphatase 1, regulatory (inhibitor) subunit 3E 225718_at 2,11 KIAA1715 KIAA1715 219825_at 2,11 CYP26B1 cytochrome P450, family 26, subfamily B, polypeptide 1 207196_s_at 2,11 TNIP1 TNFAIP3 interacting protein 1 238996_x_at 2,11 ALDOA aldolase A, fructose-bisphosphate 206039_at 2,11 RAB33A RAB33A, member RAS oncogene family 242956_at 2,11 IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble 205633_s_at 2,11 ALAS1 aminolevulinate, delta-, synthase 1 202962_at 2,10 KIF13B kinesin family member 13B 57082_at 2,10 LDLRAP1 low density lipoprotein receptor adaptor protein 1 219020_at 2,10 HS1BP3 HCLS1 binding protein 3 218759_at 2,10 DVL2 dishevelled, dsh homolog 2 (Drosophila) 218543_s_at 2,10 PARP12 poly (ADP-ribose) polymerase family, member 12 218068_s_at 2,09 ZNF672 zinc finger protein 672 203043_at 2,09 ZBED1 zinc finger, BED-type containing 1 48825_at 2,09 ING4 inhibitor of growth family, member 4 211529_x_at 2,09 HLA-G HLA-G histocompatibility antigen, class I, G 201059_at 2,09 CTTN cortactin 201537_s_at 2,08 DUSP3 dual specificity phosphatase 3 (vaccinia virus phosphatase VH1-related) 209051_s_at 2,08 RALGDS ral guanine nucleotide dissociation stimulator 207100_s_at 2,08 VAMP1 vesicle-associated membrane protein 1 (synapto- brevin 1) 224824_at 2,08 FAM36A family with sequence similarity 36, member A 242621_at 2,08 ZNF498 zinc finger protein 498

176 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 212669_at 2,08 CAMK2G calcium/calmodulin-dependent protein kinase (CaM kinase) II gamma 208881_x_at 2,07 IDI1 isopentenyl-diphosphate delta isomerase 1 213787_s_at 2,07 EBP emopamil binding protein (sterol isomerase) 211065_x_at 2,07 PFKL phosphofructokinase, liver 228028_at 2,07 FAM59B family with sequence similarity 59, member B 212793_at 2,06 DAAM2 dishevelled associated activator of morphogenesis 2 224821_at 2,06 ABHD14B abhydrolase domain containing 14B 232463_at 2,06 CXYorf10 chromosome X and Y open reading frame 10 210070_s_at 2,06 CHKB /// choline kinase beta /// carnitine palmitoyltrans- CPT1B ferase 1B (muscle) 223392_s_at 2,06 TSHZ3 teashirt family zinc finger 3 204899_s_at 2,05 SAP30 Sin3A-associated protein, 30kDa 206724_at 2,05 CBX4 chromobox homolog 4 (Pc class homolog, Dros- ophila) 1 227392_at 2,05 NISCH nischarin 37996_s_at 2,05 DMPK dystrophia myotonica-protein kinase 209795_at 2,05 CD69 CD69 molecule 222795_s_at 2,04 PLCXD1 phosphatidylinositol-specific phospholipase C, X 3 domain containing 1 208998_at 2,04 UCP2 uncoupling protein 2 (mitochondrial, proton 4 carrier) 220310_at 2,03 TUBAL3 tubulin, alpha-like 3 32137_at 2,03 JAG2 jagged 2 5 224929_at 2,03 TMEM173 transmembrane protein 173 217584_at 2,03 NPC1 Niemann-Pick disease, type C1 6 37005_at 2,03 NBL1 neuroblastoma, suppression of tumorigenicity 1 221812_at 2,03 FBXO42 F-box protein 42 7 203402_at 2,02 KCNAB2 potassium voltage-gated channel, shaker-related subfamily, beta member 2 219894_at 2,02 MAGEL2 MAGE-like 2 207595_s_at 2,02 BMP1 bone morphogenetic protein 1 201625_s_at 2,02 INSIG1 insulin induced gene 1 212558_at 2,02 SPRY1 sprouty homolog 1, antagonist of FGF signaling (Drosophila) 205662_at 2,02 EPPB9 B9 protein 216953_s_at 2,02 WT1 Wilms tumor 1 218284_at 2,01 SMAD3 SMAD family member 3 214617_at 2,01 PRF1 perforin 1 (pore forming protein)

177 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 228000_at 2,01 ADC arginine decarboxylase 231024_at 2,01 LOC572558 hypothetical locus LOC572558 202973_x_at 2,01 FAM13A1 family with sequence similarity 13, member A1 207163_s_at 2,01 AKT1 v-akt murine thymoma viral oncogene homolog 1 213827_at 2,01 SNX26 sorting nexin 26 214950_at 2,01 IL9R /// interleukin 9 receptor /// similar to Interleukin-9 LOC729486 receptor precursor (IL-9R) (CD129 antigen) 238590_x_at 2,01 TMEM107 transmembrane protein 107 202039_at 2,01 MYO18A /// TGFB1-induced anti-apoptotic factor 1 /// myosin TIAF1 XVIIIA 205536_at 2,01 VAV2 vav 2 oncogene 1567032_s_at 2,01 ZNF160 zinc finger protein 160 228149_at 2,00 FLJ31818 hypothetical protein FLJ31818 217937_s_at 2,00 HDAC7A histone deacetylase 7A 1554240_a_at 2,00 ITGAL integrin, alpha L (antigen CD11A (p180), lymphocyte function-associated antigen 1; alpha polypeptide)

1556348_at 0,19 HEATR1 HEAT repeat containing 1 1556144_at 0,22 DHX30 DEAH (Asp-Glu-Ala-His) box polypeptide 30 201796_s_at 0,24 VARS valyl-tRNA synthetase 242442_x_at 0,24 RG9MTD2 RNA (guanine-9-) methyltransferase domain containing 2 207793_s_at 0,25 EPB41 erythrocyte membrane protein band 4.1 (elliptocy- tosis 1, RH-linked) 206591_at 0,25 RAG1 recombination activating gene 1 218470_at 0,25 YARS2 tyrosyl-tRNA synthetase 2, mitochondrial 207057_at 0,27 SLC16A7 solute carrier family 16, member 7 (monocarboxylic acid transporter 2) 223527_s_at 0,28 CDADC1 cytidine and dCMP deaminase domain containing 1 1553749_at 0,29 FAM76B family with sequence similarity 76, member B 201702_s_at 0,29 PPP1R10 protein phosphatase 1, regulatory (inhibitor) subunit 10 227361_at 0,30 HS3ST3B1 heparan sulfate (glucosamine) 3-O-sulfotransferase 3B1 238960_s_at 0,31 LARP4 La ribonucleoprotein domain family, member 4 204686_at 0,32 IRS1 insulin receptor substrate 1 201843_s_at 0,33 EFEMP1 EGF-containing fibulin-like extracellular matrix protein 1

178 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 224188_s_at 0,33 XPNPEP3 X-prolyl aminopeptidase (aminopeptidase P) 3, putative 222808_at 0,33 ALG13 asparagine-linked glycosylation 13 homolog (S. cerevisiae) 201936_s_at 0,33 EIF4G3 eukaryotic translation initiation factor 4 gamma, 3 205072_s_at 0,33 XRCC4 X-ray repair complementing defective repair in Chinese hamster cells 4 239439_at 0,34 AFF4 AF4/FMR2 family, member 4 1554067_at 0,34 FLJ32549 hypothetical protein FLJ32549 235006_at 0,35 MGC13017 similar to RIKEN cDNA A430101B06 gene 244680_at 0,35 GLRB glycine receptor, beta 240546_at 0,35 LOC389043 hypothetical gene supported by AK125982; BC042817 216550_x_at 0,35 ANKRD12 ankyrin repeat domain 12 221606_s_at 0,35 NSBP1 nucleosomal binding protein 1 1 222486_s_at 0,36 ADAMTS1 ADAM metallopeptidase with thrombospondin type 1 motif, 1 220764_at 0,37 PPP4R2 protein phosphatase 4, regulatory subunit 2 205659_at 0,37 HDAC9 histone deacetylase 9 3 222849_s_at 0,37 SCRN3 secernin 3 1555829_at 0,38 FAM62B family with sequence similarity 62 (C2 domain containing) member B 4 222011_s_at 0,38 TCP1 t-complex 1 1553810_a_at 0,38 KIAA1524 KIAA1524 5 235287_at 0,38 CDK6 cyclin-dependent kinase 6 219927_at 0,38 FCF1 FCF1 small subunit (SSU) processome component 6 homolog (S. cerevisiae) 227338_at 0,38 LOC440983 hypothetical gene supported by BC066916 7 225028_at 0,39 LOC550643 hypothetical protein LOC550643 235338_s_at 0,39 SETDB2 SET domain, bifurcated 2 214056_at 0,39 MCL1 Myeloid cell leukemia sequence 1 (BCL2-related) 225158_at 0,39 GFM1 G elongation factor, mitochondrial 1 223465_at 0,40 COL4A3BP collagen, type IV, alpha 3 (Goodpasture antigen) binding protein 219074_at 0,40 TMEM34 transmembrane protein 34 218859_s_at 0,40 ESF1 ESF1, nucleolar pre-rRNA processing protein, homolog (S. cerevisiae) 223758_s_at 0,40 GTF2H2 general transcription factor IIH, polypeptide 2, 44kDa 230352_at 0,40 PRPS2 Phosphoribosyl pyrophosphate synthetase 2

179 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 202362_at 0,40 RAP1A RAP1A, member of RAS oncogene family 212824_at 0,40 FUBP3 far upstream element (FUSE) binding protein 3 1553645_at 0,40 FLJ39502 hypothetical protein FLJ39502 225484_at 0,40 TSGA14 testis specific, 14 213704_at 0,41 RABGGTB Rab geranylgeranyltransferase, beta subunit 209388_at 0,41 PAPOLA poly(A) polymerase alpha 238506_at 0,41 LRRC58 leucine rich repeat containing 58 214710_s_at 0,41 CCNB1 cyclin B1 231534_at 0,41 CDC2 Cell division cycle 2, G1 to S and G2 to M 214277_at 0,41 COX11 /// COX11 homolog, cytochrome c oxidase assembly COX11P protein (yeast) /// COX11 homolog, cytochrome c oxidase assembly protein (yeast) pseudogene 215442_s_at 0,41 TSHR thyroid stimulating hormone receptor 226019_at 0,42 OMA1 OMA1 homolog, zinc metallopeptidase (S. cerevisiae) 46947_at 0,42 GNL3L guanine nucleotide binding protein-like 3 (nucleolar)-like 215081_at 0,42 KIAA1024 KIAA1024 protein 202124_s_at 0,42 TRAK2 trafficking protein, kinesin binding 2 1559343_at 0,42 SNRPN Small nuclear ribonucleoprotein polypeptide N 1553801_a_at 0,42 C14orf126 chromosome 14 open reading frame 126 204120_s_at 0,42 ADK adenosine kinase 214953_s_at 0,42 APP amyloid beta (A4) precursor protein (peptidase nexin-II, Alzheimer disease) 235744_at 0,42 PPTC7 PTC7 protein phosphatase homolog (S. cerevisiae) 212249_at 0,42 PIK3R1 phosphoinositide-3-kinase, regulatory subunit 1 (p85 alpha) 236917_at 0,43 LRRC34 leucine rich repeat containing 34 205046_at 0,43 CENPE centromere protein E, 312kDa 235388_at 0,43 CHD9 chromodomain helicase DNA binding protein 9 212570_at 0,43 ENDOD1 endonuclease domain containing 1 221079_s_at 0,43 METTL2A /// methyltransferase like 2B /// methyltransferase METTL2B like 2A 235597_s_at 0,43 RGPD1 /// RANBP2-like and GRIP domain containing 1 /// RGPD2 /// RANBP2-like and GRIP domain containing 3 /// RGPD3 RANBP2-like and GRIP domain containing 2 220797_at 0,43 METT10D methyltransferase 10 domain containing 210733_at 0,43 TRAM1 Translocation associated membrane protein 1 235003_at 0,43 UHMK1 U2AF homology motif (UHM) kinase 1 228530_at 0,43 RP11-11C5.2 Similar to RIKEN cDNA 2410129H14

180 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 225686_at 0,44 FAM33A family with sequence similarity 33, member A 220060_s_at 0,44 C12orf48 chromosome 12 open reading frame 48 235737_at 0,44 TSLP thymic stromal lymphopoietin 213320_at 0,44 PRMT3 protein arginine methyltransferase 3 215223_s_at 0,44 SOD2 superoxide dismutase 2, mitochondrial 213907_at 0,44 EEF1E1 Eukaryotic translation elongation factor 1 epsilon 1 235644_at 0,44 CCDC138 coiled-coil domain containing 138 221213_s_at 0,45 SUHW4 suppressor of hairy wing homolog 4 (Drosophila) 218647_s_at 0,45 YRDC yrdC domain containing (E. coli) 208900_s_at 0,45 TOP1 topoisomerase (DNA) I 212615_at 0,45 CHD9 chromodomain helicase DNA binding protein 9 202240_at 0,45 PLK1 polo-like kinase 1 (Drosophila) 205369_x_at 0,45 DBT dihydrolipoamide branched chain transacylase E2 205129_at 0,45 NPM3 nucleophosmin/nucleoplasmin, 3 1 202146_at 0,46 IFRD1 interferon-related developmental regulator 1 218535_s_at 0,46 RIOK2 RIO kinase 2 (yeast) 1569366_a_at 0,46 ZNF569 zinc finger protein 569 211406_at 0,46 IER3IP1 immediate early response 3 interacting protein 1 3 219036_at 0,46 CEP70 centrosomal protein 70kDa 216493_s_at 0,46 IGF2BP3 /// insulin-like growth factor 2 mRNA binding protein LOC645468 /// 3 /// similar to insulin-like growth factor 2 mRNA 4 LOC651107 binding protein 3 /// similar to IGF-II mRNA- binding protein 3 5 221561_at 0,46 SOAT1 sterol O-acyltransferase (acyl-Coenzyme A: choles- terol acyltransferase) 1 203276_at 0,46 LMNB1 lamin B1 6 212105_s_at 0,46 DHX9 DEAH (Asp-Glu-Ala-His) box polypeptide 9 202872_at 0,46 ATP6V1C1 ATPase, H+ transporting, lysosomal 42kDa, V1 7 subunit C1 231820_x_at 0,46 ZNF587 zinc finger protein 587 239233_at 0,47 CCDC88A coiled-coil domain containing 88A 211697_x_at 0,47 PNO1 partner of NOB1 homolog (S. cerevisiae) 225114_at 0,47 AGPS alkylglycerone phosphate synthase 201151_s_at 0,47 MBNL1 muscleblind-like (Drosophila) 220346_at 0,47 MTHFD2L methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2-like 214583_at 0,47 RSC1A1 regulatory solute carrier protein, family 1, member 1 219031_s_at 0,47 NIP7 nuclear import 7 homolog (S. cerevisiae) 221705_s_at 0,47 SIKE suppressor of IKK epsilon

181 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 219493_at 0,47 SHCBP1 SHC SH2-domain binding protein 1 212575_at 0,47 C19orf6 chromosome 19 open reading frame 6 200841_s_at 0,47 EPRS glutamyl-prolyl-tRNA synthetase 226794_at 0,47 STXBP5 syntaxin binding protein 5 (tomosyn) 223254_s_at 0,47 KIAA1333 KIAA1333 208042_at 0,47 AGGF1 angiogenic factor with G patch and FHA domains 1 217878_s_at 0,47 CDC27 cell division cycle 27 homolog (S. cerevisiae) 222850_s_at 0,47 DNAJB14 DnaJ (Hsp40) homolog, subfamily B, member 14 204807_at 0,47 TMEM5 transmembrane protein 5 213655_at 0,48 YWHAE Tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, epsilon polypeptide 205097_at 0,48 SLC26A2 solute carrier family 26 (sulfate transporter), member 2 227161_at 0,48 NOM1 nucleolar protein with MIF4G domain 1 223651_x_at 0,48 CDC23 cell division cycle 23 homolog (S. cerevisiae) 231784_s_at 0,48 WDSOF1 WD repeats and SOF1 domain containing 209257_s_at 0,48 SMC3 structural maintenance of chromosomes 3 225834_at 0,48 FAM72A /// family with sequence similarity 72, member A /// LOC653820 /// similar to family with sequence similarity 72, LOC729533 member A 209754_s_at 0,48 TMPO thymopoietin 228204_at 0,48 PSMB4 Proteasome (prosome, macropain) subunit, beta type, 4 218712_at 0,48 C1orf109 chromosome 1 open reading frame 109 226762_at 0,48 PURB purine-rich element binding protein B 231175_at 0,48 C6orf65 chromosome 6 open reading frame 65 230618_s_at 0,48 BAT2D1 BAT2 domain containing 1 210802_s_at 0,48 DIMT1L DIM1 dimethyladenosine transferase 1-like (S. cerevisiae) 1557984_s_at 0,48 FLJ21908 hypothetical protein FLJ21908 205362_s_at 0,48 PFDN4 prefoldin subunit 4 235463_s_at 0,48 LASS6 LAG1 homolog, ceramide synthase 6 (S. cerevisiae) 222642_s_at 0,48 TMEM33 transmembrane protein 33 229128_s_at 0,48 ANP32E Acidic (leucine-rich) nuclear phosphoprotein 32 family, member E 209865_at 0,48 SLC35A3 solute carrier family 35 (UDP-N-acetylglucosamine (UDP-GlcNAc) transporter), member A3 1555638_a_at 0,48 SAMSN1 SAM domain, SH3 domain and nuclear localization signals 1 212952_at 0,48 CALR Calreticulin

182 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 217759_at 0,48 TRIM44 tripartite motif-containing 44 226765_at 0,49 SPTBN1 Spectrin, beta, non-erythrocytic 1 213982_s_at 0,49 RABGAP1L RAB GTPase activating protein 1-like 217547_x_at 0,49 ZNF675 zinc finger protein 675 236140_at 0,49 GCLM glutamate-cysteine ligase, modifier subunit 220330_s_at 0,49 SAMSN1 SAM domain, SH3 domain and nuclear localization signals 1 229194_at 0,49 PCGF5 polycomb group ring finger 5 222360_at 0,49 DPH5 DPH5 homolog (S. cerevisiae) 229125_at 0,49 ANKRD38 ankyrin repeat domain 38 203465_at 0,49 MRPL19 mitochondrial ribosomal protein L19 219918_s_at 0,49 ASPM asp (abnormal spindle) homolog, microcephaly associated (Drosophila) 204962_s_at 0,49 CENPA centromere protein A 221683_s_at 0,49 CEP290 centrosomal protein 290kDa 1 236834_at 0,49 SCFD2 sec1 family domain containing 2 219546_at 0,49 BMP2K BMP2 inducible kinase 213926_s_at 0,49 HRB HIV-1 Rev binding protein 225300_at 0,49 C15orf23 chromosome 15 open reading frame 23 3 217620_s_at 0,49 PIK3CB phosphoinositide-3-kinase, catalytic, beta polypeptide 4 203790_s_at 0,49 HRSP12 heat-responsive protein 12 206989_s_at 0,49 SFRS2IP splicing factor, arginine/serine-rich 2, interacting 5 protein 1557132_at 0,50 WDR17 WD repeat domain 17 208653_s_at 0,50 CD164 CD164 molecule, sialomucin 6 237466_s_at 0,50 HHIP hedgehog interacting protein 217299_s_at 0,50 NBN nibrin 7 201996_s_at 0,50 SPEN spen homolog, transcriptional regulator (Dros- ophila) 1554873_at 0,50 CSPP1 centrosome and spindle pole associated protein 1 224309_s_at 0,50 SUGT1 SGT1, suppressor of G2 allele of SKP1 (S. cerevisiae) 1552263_at 0,50 MAPK1 mitogen-activated protein kinase 1 209451_at 0,50 TANK TRAF family member-associated NFKB activator 242648_at 0,50 KLHL8 kelch-like 8 (Drosophila) 227510_x_at 0,50 PRO1073 PRO1073 protein

183 Chapter 6

Supplementary Table 2. UM-1 cell line

Probe Set ID Score(d) Gene Symbol Gene Title 202912_at 57,63 ADM adrenomedullin 228499_at 36,58 PFKFB4 6-phosphofructo-2-kinase/ fructose-2,6-biphosphatase 4 242064_at 24,83 SDK2 sidekick homolog 2 (chicken) 221009_s_at 13,21 ANGPTL4 angiopoietin-like 4 203438_at 12,97 STC2 stanniocalcin 2 214978_s_at 12,97 PPFIA4 protein tyrosine phosphatase, receptor type, f polypeptide (PTPRF), interacting protein (liprin), alpha 4 227337_at 12,34 ANKRD37 ankyrin repeat domain 37 238551_at 11,21 FUT11 fucosyltransferase 11 (alpha (1,3) fucosyltransferase) 235850_at 10,99 WDR5B WD repeat domain 5B 234312_s_at 10,88 ACSS2 acyl-CoA synthetase short-chain family member 2 204348_s_at 10,63 AK3L1 adenylate kinase 3-like 1 243435_at 10,59 KCNQ1OT1 KCNQ1 overlapping transcript 1 226682_at 10,47 LOC283666 hypothetical protein LOC283666 204347_at 9,57 LOC645619 /// similar to Adenylate kinase isoenzyme 4, LOC731007 mitochondrial (ATP-AMP transphosphorylase) /// similar to Adenylate kinase isoenzyme 4, mitochondrial (Adenylate kinase 3-like 1) (ATP-AMP transphosphorylase) 202022_at 9,51 ALDOC aldolase C, fructose-bisphosphate 202497_x_at 9,47 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 227404_s_at 8,55 EGR1 Early growth response 1 202718_at 8,45 IGFBP2 insulin-like growth factor binding protein 2, 36kDa 201849_at 8,19 BNIP3 BCL2/adenovirus E1B 19kDa interacting protein 3 207543_s_at 8,08 P4HA1 procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase), alpha polypeptide I 215925_s_at 7,08 CD72 CD72 molecule 202934_at 6,91 HK2 hexokinase 2 202464_s_at 6,70 PFKFB3 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 221478_at 6,63 BNIP3L BCL2/adenovirus E1B 19kDa interacting protein 3-like 226452_at 6,62 PDK1 pyruvate dehydrogenase kinase, isozyme 1 210512_s_at 6,58 VEGFA vascular endothelial growth factor A

184 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 203282_at 6,00 GBE1 glucan (1,4-alpha-), branching enzyme 1 (glycogen branching enzyme, Andersen disease, glycogen storage disease type IV) 224797_at 5,93 ARRDC3 arrestin domain containing 3 210426_x_at 5,92 RORA RAR-related orphan receptor A 203946_s_at 5,77 ARG2 arginase, type II 1558770_a_at 5,60 C17orf38 chromosome 17 open reading frame 38 206991_s_at 5,53 CCR5 /// chemokine (C-C motif) receptor 5 /// similar to LOC727797 C-C chemokine receptor type 5 (C-C CKR-5) (CC-CKR-5) (CCR-5) (CCR5) (HIV-1 fusion coreceptor) (CHEMR13) (CD195 antigen) 228483_s_at 5,46 TAF9B TAF9B RNA polymerase II, TATA box binding protein (TBP)-associated factor, 31kDa 209566_at 5,39 INSIG2 insulin induced gene 2 219670_at 5,34 C1orf165 chromosome 1 open reading frame 165 223216_x_at 5,33 FBXO16 /// zinc finger protein 395 /// F-box protein 16 1 ZNF395 223046_at 5,04 EGLN1 egl nine homolog 1 (C. elegans) 218149_s_at 5,04 ZNF395 zinc finger protein 395 203027_s_at 5,01 MVD mevalonate (diphospho) decarboxylase 3 206580_s_at 4,97 EFEMP2 EGF-containing fibulin-like extracellular matrix protein 2 4 225557_at 4,95 AXUD1 AXIN1 up-regulated 1 202218_s_at 4,92 FADS2 fatty acid desaturase 2 5 206686_at 4,89 PDK1 pyruvate dehydrogenase kinase, isozyme 1 207425_s_at 4,89 9-Sep septin 9 218274_s_at 4,86 ANKZF1 ankyrin repeat and zinc finger domain containing 1 6 208513_at 4,83 FOXB1 forkhead box B1 205599_at 4,72 TRAF1 TNF receptor-associated factor 1 7 202149_at 4,69 NEDD9 neural precursor cell expressed, developmentally down-regulated 9 221757_at 4,68 PIK3IP1 phosphoinositide-3-kinase interacting protein 1 219622_at 4,55 RAB20 RAB20, member RAS oncogene family 202393_s_at 4,55 KLF10 Kruppel-like factor 10 209446_s_at 4,44 C7orf44 chromosome 7 open reading frame 44 201389_at 4,39 ITGA5 integrin, alpha 5 (fibronectin receptor, alpha polypeptide) 212689_s_at 4,36 JMJD1A jumonji domain containing 1A 201693_s_at 4,36 EGR1 early growth response 1

185 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 226985_at 4,33 FGD5 FYVE, RhoGEF and PH domain containing 5 217028_at 4,28 CXCR4 chemokine (C-X-C motif) receptor 4 224345_x_at 4,12 C3orf28 chromosome 3 open reading frame 28 209795_at 4,07 CD69 CD69 molecule 212281_s_at 4,07 TMEM97 transmembrane protein 97 244178_at 4,06 COMMD7 COMM domain containing 7 1569453_a_at 4,06 LOC692247 hypothetical locus LOC692247 201368_at 4,04 ZFP36L2 zinc finger protein 36, C3H type-like 2 202426_s_at 4,04 RXRA retinoid X receptor, alpha 201790_s_at 3,93 DHCR7 7-dehydrocholesterol reductase 205484_at 3,93 SIT1 signaling threshold regulating transmembrane adaptor 1 226390_at 3,81 STARD4 START domain containing 4, sterol regulated 228891_at 3,72 C9orf164 chromosome 9 open reading frame 164 227868_at 3,68 LOC154761 hypothetical protein LOC154761 230398_at 3,63 TNS4 tensin 4 202973_x_at 3,58 FAM13A1 family with sequence similarity 13, member A1 211434_s_at 3,57 CCRL2 /// chemokine (C-C motif) receptor-like 2 /// similar LOC727811 to chemokine (C-C motif) receptor-like 2 240258_at 3,55 ENO1 enolase 1, (alpha) 201625_s_at 3,53 INSIG1 insulin induced gene 1 219014_at 3,53 PLAC8 placenta-specific 8 214593_at 3,53 PIAS2 protein inhibitor of activated STAT, 2 202887_s_at 3,51 DDIT4 DNA-damage-inducible transcript 4 212496_s_at 3,47 JMJD2B jumonji domain containing 2B 202562_s_at 3,47 C14orf1 chromosome 14 open reading frame 1 210132_at 3,46 EFNA3 ephrin-A3 206039_at 3,44 RAB33A RAB33A, member RAS oncogene family 218012_at 3,44 TSPYL2 TSPY-like 2 228749_at 3,43 ZDBF2 zinc finger, DBF-type containing 2 215649_s_at 3,43 MVK mevalonate kinase (mevalonic aciduria) 208657_s_at 3,40 9-Sep septin 9 226114_at 3,40 ZNF436 zinc finger protein 436 223553_s_at 3,35 DOK3 docking protein 3 238996_x_at 3,33 ALDOA aldolase A, fructose-bisphosphate 202449_s_at 3,32 RXRA retinoid X receptor, alpha 209034_at 3,28 PNRC1 proline-rich nuclear receptor coactivator 1

186 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 202245_at 3,27 LSS lanosterol synthase (2,3-oxidosqualene-lanosterol cyclase) 224602_at 3,25 LOC401152 HCV F-transactivated protein 1 215734_at 3,24 C19orf36 chromosome 19 open reading frame 36 219236_at 3,21 PAQR6 progestin and adipoQ receptor family member VI 202996_at 3,19 POLD4 polymerase (DNA-directed), delta 4 209859_at 3,19 TRIM9 tripartite motif-containing 9 238965_at 3,16 C21orf2 Chromosome 21 open reading frame 2 209279_s_at 3,15 NSDHL NAD(P) dependent steroid dehydrogenase-like 201673_s_at 3,13 GYS1 glycogen synthase 1 (muscle) 1567032_s_at 3,13 ZNF160 zinc finger protein 160 202644_s_at 3,12 TNFAIP3 tumor necrosis factor, alpha-induced protein 3 222870_s_at 3,10 B3GNT2 UDP-GlcNAc:betaGal beta-1,3-N-acetylglucosaminyltransferase 2 209822_s_at 3,09 VLDLR very low density lipoprotein receptor 1 1555349_a_at 3,06 ITGB2 integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) 242621_at 3,06 ZNF498 zinc finger protein 498 218638_s_at 3,05 SPON2 spondin 2, extracellular matrix protein 3 201464_x_at 3,04 JUN jun oncogene 210130_s_at 3,03 TM7SF2 transmembrane 7 superfamily member 2 4 223522_at 2,99 C9orf45 chromosome 9 open reading frame 45 228098_s_at 2,96 MYLIP myosin regulatory light chain interacting protein 5 237120_at 2,96 KRT77 keratin 77 205497_at 2,95 ZNF175 zinc finger protein 175 6 218368_s_at 2,90 TNFRSF12A tumor necrosis factor receptor superfamily, member 12A 233358_at 2,89 FLJ14311 hypothetical gene FLJ14311 7 207339_s_at 2,88 LTB lymphotoxin beta (TNF superfamily, member 3) 201170_s_at 2,87 BHLHB2 basic helix-loop-helix domain containing, class B, 2 36711_at 2,85 MAFF v-maf musculoaponeurotic fibrosarcoma oncogene homolog F (avian) 209608_s_at 2,84 ACAT2 acetyl-Coenzyme A acetyltransferase 2 (acetoacetyl Coenzyme A thiolase) 208763_s_at 2,84 TSC22D3 TSC22 domain family, member 3 204243_at 2,84 RLF rearranged L-myc fusion 202472_at 2,83 MPI mannose phosphate isomerase 219183_s_at 2,83 PSCD4 pleckstrin homology, Sec7 and coiled-coil domains 4

187 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 1569835_at 2,81 LOC339352 Similar to ATP binding domain 3 202331_at 2,80 BCKDHA branched chain keto acid dehydrogenase E1, alpha polypeptide 203950_s_at 2,79 CLCN6 chloride channel 6 218205_s_at 2,77 MKNK2 MAP kinase interacting serine/threonine kinase 2 1569909_at 2,77 KRT6L keratin 6L 228738_at 2,76 D2HGDH D-2-hydroxyglutarate dehydrogenase 224772_at 2,76 NAV1 neuron navigator 1 206907_at 2,75 TNFSF9 tumor necrosis factor (ligand) superfamily, member 9 219417_s_at 2,74 C17orf59 chromosome 17 open reading frame 59 209975_at 2,73 CYP2E1 cytochrome P450, family 2, subfamily E, polypeptide 1 206828_at 2,72 TXK TXK tyrosine kinase 210105_s_at 2,72 FYN FYN oncogene related to SRC, FGR, YES 206478_at 2,71 KIAA0125 KIAA0125 206348_s_at 2,70 PDK3 pyruvate dehydrogenase kinase, isozyme 3 1562403_a_at 2,69 SLC8A3 solute carrier family 8 (sodium-calcium exchanger), member 3 235948_at 2,69 FAM80A family with sequence similarity 80, member A 241954_at 2,68 FDFT1 Farnesyl-diphosphate farnesyltransferase 1 205790_at 2,66 SKAP1 src kinase associated phosphoprotein 1 203317_at 2,65 PSD4 pleckstrin and Sec7 domain containing 4 201537_s_at 2,64 DUSP3 dual specificity phosphatase 3 (vaccinia virus phosphatase VH1-related) 1552343_s_at 2,64 PDE7A phosphodiesterase 7A 226471_at 2,64 GGTL3 gamma-glutamyltransferase-like 3 219862_s_at 2,63 NARF nuclear prelamin A recognition factor 1565611_at 2,62 DYNLRB1 Dynein, light chain, roadblock-type 1 206583_at 2,61 ZNF673 zinc finger protein 673 214847_s_at 2,60 GPSM3 G-protein signalling modulator 3 (AGS3-like, C. elegans) 201313_at 2,60 ENO2 enolase 2 (gamma, neuronal) 211423_s_at 2,59 SC5DL sterol-C5-desaturase (ERG3 delta-5-desaturase homolog, S. cerevisiae)-like 204484_at 2,59 PIK3C2B phosphoinositide-3-kinase, class 2, beta polypeptide 227134_at 2,57 SYTL1 synaptotagmin-like 1 219563_at 2,57 C14orf139 chromosome 14 open reading frame 139

188 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 200827_at 2,56 PLOD1 procollagen-lysine 1, 2-oxoglutarate 5-dioxygenase 1 212368_at 2,56 ZNF292 zinc finger protein 292 212190_at 2,56 SERPINE2 serpin peptidase inhibitor, clade E (nexin, plasminogen activator inhibitor type 1), member 2 212958_x_at 2,55 PAM peptidylglycine alpha-amidating monooxygenase 37796_at 2,53 LRCH4 leucine-rich repeats and calponin homology (CH) domain containing 4 224925_at 2,53 PREX1 phosphatidylinositol 3,4,5-trisphosphate-dependent RAC exchanger 1 207196_s_at 2,52 TNIP1 TNFAIP3 interacting protein 1 209146_at 2,49 SC4MOL sterol-C4-methyl oxidase-like 202364_at 2,48 MXI1 MAX interactor 1 200697_at 2,47 HK1 hexokinase 1 213693_s_at 2,46 MUC1 mucin 1, cell surface associated 200832_s_at 2,46 SCD stearoyl-CoA desaturase (delta-9-desaturase) 1 205141_at 2,45 ANG angiogenin, ribonuclease, RNase A family, 5 216086_at 2,44 SV2C synaptic vesicle glycoprotein 2C 229063_s_at 2,44 CCDC107 coiled-coil domain containing 107 225530_at 2,43 MOBKL2A MOB1, Mps One Binder kinase activator-like 2A 3 (yeast) 241348_at 2,42 ZNF654 zinc finger protein 654 4 231056_at 2,42 LOC339352 similar to ATP binding domain 3 243296_at 2,42 PBEF1 Pre-B-cell colony enhancing factor 1 5 201251_at 2,41 PKM2 pyruvate kinase, muscle 222451_s_at 2,41 ZDHHC9 zinc finger, DHHC-type containing 9 225383_at 2,40 ZNF275 zinc finger protein 275 6 215189_at 2,40 KRT86 keratin 86 226470_at 2,39 GGTL3 gamma-glutamyltransferase-like 3 7 41386_i_at 2,39 JMJD3 jumonji domain containing 3 235231_at 2,39 ZNF789 zinc finger protein 789 34210_at 2,38 CD52 CD52 molecule 201309_x_at 2,36 C5orf13 chromosome 5 open reading frame 13 222906_at 2,36 FLVCR1 feline leukemia virus subgroup C cellular receptor 1 230086_at 2,35 FNBP1 formin binding protein 1 220091_at 2,34 SLC2A6 solute carrier family 2 (facilitated glucose transporter), member 6 218145_at 2,34 TRIB3 tribbles homolog 3 (Drosophila) 210561_s_at 2,33 WSB1 WD repeat and SOCS box-containing 1

189 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 205904_at 2,33 MICA MHC class I polypeptide-related sequence A 223179_at 2,32 YPEL3 yippee-like 3 (Drosophila) 204899_s_at 2,32 SAP30 Sin3A-associated protein, 30kDa 202219_at 2,32 SLC6A8 solute carrier family 6 (neurotransmitter transporter, creatine), member 8 225262_at 2,31 FOSL2 FOS-like antigen 2 210337_s_at 2,31 ACLY ATP citrate lyase 202803_s_at 2,31 ITGB2 integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) 1553537_at 2,30 KRT73 keratin 73 221750_at 2,29 HMGCS1 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1 (soluble) 220310_at 2,29 TUBAL3 tubulin, alpha-like 3 232809_s_at 2,28 FLT1 Fms-related tyrosine kinase 1 (vascular endothelial growth factor/vascular permeability factor receptor) 212813_at 2,28 JAM3 junctional adhesion molecule 3 228149_at 2,27 FLJ31818 hypothetical protein FLJ31818 212552_at 2,27 HPCAL1 hippocalcin-like 1 242592_at 2,27 GPR137C G protein-coupled receptor 137C 204669_s_at 2,26 RNF24 ring finger protein 24 222856_at 2,26 APLN apelin, AGTRL1 ligand 1554860_at 2,26 PTPN7 protein tyrosine phosphatase, non-receptor type 7 218625_at 2,26 NRN1 neuritin 1 229086_at 2,25 C1orf213 chromosome 1 open reading frame 213 226442_at 2,24 ABTB1 ankyrin repeat and BTB (POZ) domain containing 1 201508_at 2,24 IGFBP4 insulin-like growth factor binding protein 4 225918_at 2,24 LOC146346 hypothetical protein LOC146346 214861_at 2,24 JMJD2C jumonji domain containing 2C 212745_s_at 2,24 BBS4 Bardet-Biedl syndrome 4 225718_at 2,23 KIAA1715 KIAA1715 201275_at 2,23 FDPS farnesyl diphosphate synthase (farnesyl pyrophosphate synthetase, dimethylallyltranstransferase, geranyltranstransferase) 1569212_at 2,22 LOC619207 scavenger receptor protein family member 202068_s_at 2,22 LDLR low density lipoprotein receptor (familial hypercholesterolemia) 210362_x_at 2,22 PML promyelocytic leukemia

190 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 201626_at 2,22 INSIG1 insulin induced gene 1 210203_at 2,21 CNOT4 CCR4-NOT transcription complex, subunit 4 205267_at 2,21 POU2AF1 POU domain, class 2, associating factor 1 204205_at 2,21 APOBEC3G apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like 3G 235213_at 2,21 ITPKB Inositol 1,4,5-trisphosphate 3-kinase B 226275_at 2,20 MXD1 MAX dimerization protein 1 202492_at 2,19 ATG9A ATG9 autophagy related 9 homolog A (S. cerevisiae) 243264_s_at 2,18 C8orf44 /// serum/glucocorticoid regulated kinase family, SGK3 member 3 /// chromosome 8 open reading frame 44 217356_s_at 2,18 PGK1 phosphoglycerate kinase 1 226811_at 2,18 FAM46C family with sequence similarity 46, member C 201810_s_at 2,17 SH3BP5 SH3-domain binding protein 5 (BTK-associated) 221011_s_at 2,16 LBH limb bud and heart development homolog (mouse) 1 207590_s_at 2,16 CENPI centromere protein I 209827_s_at 2,16 IL16 interleukin 16 (lymphocyte chemoattractant factor) 228906_at 2,16 CXXC6 CXXC finger 6 205839_s_at 2,16 BZRAP1 benzodiazapine receptor (peripheral) associated 3 protein 1 238784_at 2,16 DPY19L2 dpy-19-like 2 (C. elegans) 4 220081_x_at 2,15 HSD17B7 /// hydroxysteroid (17-beta) dehydrogenase 7 /// HSD17B7P2 /// hydroxysteroid (17-beta) dehydrogenase 7 pseu- LOC730412 dogene 2 /// similar to hydroxysteroid (17-beta) 5 dehydrogenase 7 200808_s_at 2,15 ZYX zyxin 6 224027_at 2,15 CCL28 chemokine (C-C motif) ligand 28 215785_s_at 2,15 CYFIP2 cytoplasmic FMR1 interacting protein 2 7 210534_s_at 2,14 EPPB9 B9 protein 201102_s_at 2,14 PFKL phosphofructokinase, liver 208881_x_at 2,14 IDI1 isopentenyl-diphosphate delta isomerase 1 217937_s_at 2,14 HDAC7A histone deacetylase 7A 238540_at 2,14 LOC401320 hypothetical LOC401320 40020_at 2,13 CELSR3 cadherin, EGF LAG seven-pass G-type receptor 3 (flamingo homolog, Drosophila) 200872_at 2,13 S100A10 S100 calcium binding protein A10 213787_s_at 2,13 EBP emopamil binding protein (sterol isomerase) 1554452_a_at 2,13 HIG2 hypoxia-inducible protein 2 203037_s_at 2,13 MTSS1 metastasis suppressor 1

191 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 221572_s_at 2,12 SLC26A6 solute carrier family 26, member 6 207908_at 2,12 KRT2 keratin 2 (epidermal ichthyosis bullosa of Siemens) 202260_s_at 2,11 STXBP1 syntaxin binding protein 1 207686_s_at 2,11 CASP8 caspase 8, apoptosis-related cysteine peptidase 227645_at 2,10 PIK3R5 phosphoinositide-3-kinase, regulatory subunit 5, p101 228667_at 2,10 AGPAT4 1-acylglycerol-3-phosphate O-acyltransferase 4 (lysophosphatidic acid acyltransferase, delta) 230012_at 2,10 C17orf44 chromosome 17 open reading frame 44 201250_s_at 2,10 SLC2A1 solute carrier family 2 (facilitated glucose transporter), member 1 204225_at 2,10 HDAC4 histone deacetylase 4 203208_s_at 2,10 MTFR1 mitochondrial fission regulator 1 205250_s_at 2,09 CEP290 centrosomal protein 290kDa 239137_x_at 2,09 MGC45491 hypothetical protein MGC45491 214730_s_at 2,09 GLG1 golgi apparatus protein 1 39966_at 2,09 CSPG5 chondroitin sulfate proteoglycan 5 (neuroglycan C) 202081_at 2,09 IER2 immediate early response 2 218498_s_at 2,09 ERO1L ERO1-like (S. cerevisiae) 209882_at 2,09 RIT1 Ras-like without CAAX 1 212641_at 2,08 HIVEP2 human immunodeficiency virus type I enhancer binding protein 2 228194_s_at 2,08 SORCS1 sortilin-related VPS10 domain containing receptor 1 219125_s_at 2,08 RAG1AP1 recombination activating gene 1 activating protein 1 200921_s_at 2,08 BTG1 B-cell translocation gene 1, anti-proliferative 241905_at 2,07 PIK3C2A Phosphoinositide-3-kinase, class 2, alpha polypeptide 213397_x_at 2,07 RNASE4 ribonuclease, RNase A family, 4 204106_at 2,07 TESK1 testis-specific kinase 1 218840_s_at 2,07 NADSYN1 NAD synthetase 1 203521_s_at 2,06 ZNF318 zinc finger protein 318 1558525_at 2,06 LOC283901 hypothetical protein LOC283901 221813_at 2,06 FBXO42 F-box protein 42 221828_s_at 2,06 FAM125B family with sequence similarity 125, member B 210556_at 2,06 NFATC3 nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 3 202539_s_at 2,06 HMGCR 3-hydroxy-3-methylglutaryl-Coenzyme A reductase 202769_at 2,05 CCNG2 cyclin G2

192 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 239761_at 2,05 GCNT1 glucosaminyl (N-acetyl) transferase 1, core 2 (beta-1,6-N-acetylglucosaminyltransferase) 205045_at 2,05 AKAP10 A kinase (PRKA) anchor protein 10 228550_at 2,05 RTN4R reticulon 4 receptor 210031_at 2,05 CD247 CD247 molecule 223498_at 2,05 SPECC1 sperm antigen with calponin homology and coiled-coil domains 1 208964_s_at 2,05 FADS1 fatty acid desaturase 1 214500_at 2,04 H2AFY H2A histone family, member Y 223392_s_at 2,04 TSHZ3 teashirt family zinc finger 3 203349_s_at 2,04 ETV5 ets variant gene 5 (ets-related molecule) 219629_at 2,04 FAM118A family with sequence similarity 118, member A 233813_at 2,04 PPP1R16B protein phosphatase 1, regulatory (inhibitor) subunit 16B 205107_s_at 2,03 EFNA4 ephrin-A4 1 228739_at 2,03 CYS1 cystin 1 225869_s_at 2,03 UNC93B1 unc-93 homolog B1 (C. elegans) 201637_s_at 2,03 FXR1 fragile X mental retardation, autosomal homolog 1 222790_s_at 2,03 RSBN1 round spermatid basic protein 1 3 225715_at 2,03 KIAA1303 raptor 212770_at 2,03 TLE3 transducin-like enhancer of split 3 (E(sp1) 4 homolog, Drosophila) 244556_at 2,02 LCP2 Lymphocyte cytosolic protein 2 (SH2 domain containing leukocyte protein of 76kDa) 5 203929_s_at 2,02 MAPT microtubule-associated protein tau 221214_s_at 2,02 NELF nasal embryonic LHRH factor 6 203767_s_at 2,01 STS steroid sulfatase (microsomal), arylsulfatase C, isozyme S 7 218630_at 2,01 MKS1 Meckel syndrome, type 1 59375_at 2,01 MYO15B myosin XVB pseudogene 1554079_at 2,01 GALNTL4 UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase-like 4 227412_at 2,01 PPP1R3E protein phosphatase 1, regulatory (inhibitor) subunit 3E 234302_s_at 2,01 ALKBH5 alkB, alkylation repair homolog 5 (E. coli) 226925_at 2,01 ACPL2 acid phosphatase-like 2 217677_at 2,00 PLEKHA2 pleckstrin homology domain containing, family A (phosphoinositide binding specific) member 2 35617_at 2,00 MAPK7 mitogen-activated protein kinase 7

193 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 229437_at 2,00 BIC BIC transcript 227383_at 2,00 LOC727820 hypothetical protein LOC727820 226037_s_at 0,13 LOC728198 /// TAF9B RNA polymerase II, TATA box binding TAF9B protein (TBP)-associated factor, 31kDa /// similar to transcription associated factor 9B 225779_at 0,22 SLC27A4 solute carrier family 27 (fatty acid transporter), member 4 219522_at 0,26 FJX1 four jointed box 1 (Drosophila) 225484_at 0,29 TSGA14 testis specific, 14 1557309_at 0,32 DENND1B DENN/MADD domain containing 1B 224870_at 0,33 KIAA0114 KIAA0114 237466_s_at 0,34 HHIP hedgehog interacting protein 226884_at 0,36 LRRN1 leucine rich repeat neuronal 1 235117_at 0,36 CHAC2 ChaC, cation transport regulator homolog 2 (E. coli) 219676_at 0,36 ZSCAN16 zinc finger and SCAN domain containing 16 209567_at 0,36 RRS1 RRS1 ribosome biogenesis regulator homolog (S. cerevisiae) 222714_s_at 0,36 LACTB2 lactamase, beta 2 232155_at 0,38 KIAA1618 KIAA1618 220311_at 0,38 N6AMT1 N-6 adenine-specific DNA methyltransferase 1 (putative) 224632_at 0,38 GPATCH4 G patch domain containing 4 1554331_a_at 0,39 LRRC18 leucine rich repeat containing 18 1556348_at 0,39 HEATR1 HEAT repeat containing 1 212333_at 0,40 FAM98A family with sequence similarity 98, member A 222162_s_at 0,40 ADAMTS1 ADAM metallopeptidase with thrombospondin type 1 motif, 1 209771_x_at 0,40 CD24 CD24 molecule 1553111_a_at 0,40 KBTBD6 kelch repeat and BTB (POZ) domain containing 6 219050_s_at 0,41 ZNHIT2 zinc finger, HIT type 2 202870_s_at 0,41 CDC20 cell division cycle 20 homolog (S. cerevisiae) 225200_at 0,41 DPH3 DPH3, KTI11 homolog (S. cerevisiae) 205992_s_at 0,41 IL15 interleukin 15 228050_at 0,42 UTP15 UTP15, U3 small nucleolar ribonucleoprotein, homolog (S. cerevisiae) 232291_at 0,42 MIRH1 microRNA host gene (non-protein coding) 1 225748_at 0,42 LTV1 LTV1 homolog (S. cerevisiae) 204127_at 0,42 RFC3 replication factor C (activator 1) 3, 38kDa

194 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 221168_at 0,42 PRDM13 PR domain containing 13 1555950_a_at 0,43 CD55 CD55 molecule, decay accelerating factor for complement (Cromer blood group) 228355_s_at 0,43 NDUFA12L NDUFA12-like 222958_s_at 0,43 DEPDC1 DEP domain containing 1 223651_x_at 0,43 CDC23 cell division cycle 23 homolog (S. cerevisiae) 236302_at 0,43 PPM1E protein phosphatase 1E (PP2C domain containing) 209595_at 0,43 GTF2F2 general transcription factor IIF, polypeptide 2, 30kDa 201656_at 0,43 ITGA6 integrin, alpha 6 207057_at 0,43 SLC16A7 solute carrier family 16, member 7 (monocarboxylic acid transporter 2) 231784_s_at 0,43 WDSOF1 WD repeats and SOF1 domain containing 228645_at 0,43 SNHG9 small nucleolar RNA host gene (non-protein coding) 9 1 231798_at 0,43 NOG Noggin 218647_s_at 0,43 YRDC yrdC domain containing (E. coli) 202715_at 0,44 CAD carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase 3 219031_s_at 0,44 NIP7 nuclear import 7 homolog (S. cerevisiae) 242260_at 0,44 MATR3 Matrin 3 4 224450_s_at 0,44 RIOK1 RIO kinase 1 (yeast) 223113_at 0,44 TMEM138 transmembrane protein 138 241933_at 0,44 QRSL1 Glutaminyl-tRNA synthase 5 (glutamine-hydrolyzing)-like 1 205215_at 0,44 RNF2 ring finger protein 2 6 213427_at 0,44 RPP40 ribonuclease P 40kDa subunit 230170_at 0,44 OSM oncostatin M 7 203321_s_at 0,44 ZNF508 zinc finger protein 508 202345_s_at 0,44 FABP5 /// fatty acid binding protein 5 (psoriasis-associated) LOC728641 /// /// similar to Fatty acid-binding protein, epidermal LOC729163 /// (E-FABP) (Psoriasis-associated fatty acid-binding LOC731043 /// protein homolog) (PA-FABP) LOC732031 200848_at 0,44 AHCYL1 S-adenosylhomocysteine hydrolase-like 1 218590_at 0,44 PEO1 progressive external ophthalmoplegia 1 220083_x_at 0,44 UCHL5 ubiquitin carboxyl-terminal hydrolase L5 212510_at 0,45 GPD1L glycerol-3-phosphate dehydrogenase 1-like 220295_x_at 0,45 DEPDC1 /// DEP domain containing 1 /// similar to DEP LOC730888 domain containing 1

195 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 215388_s_at 0,45 CFH /// complement factor H /// complement factor CFHR1 H-related 1 231102_at 0,45 CROT carnitine O-octanoyltransferase 213320_at 0,45 PRMT3 protein arginine methyltransferase 3 1569190_at 0,45 SCLT1 sodium channel and clathrin linker 1 201023_at 0,45 TAF7 TAF7 RNA polymerase II, TATA box binding protein (TBP)-associated factor, 55kDa 203956_at 0,45 MORC2 MORC family CW-type zinc finger 2 225291_at 0,46 PNPT1 polyribonucleotide nucleotidyltransferase 1 231240_at 0,46 DIO2 deiodinase, iodothyronine, type II 204905_s_at 0,46 EEF1E1 eukaryotic translation elongation factor 1 epsilon 1 214277_at 0,46 COX11 /// COX11 homolog, cytochrome c oxidase assembly COX11P protein (yeast) /// COX11 homolog, cytochrome c oxidase assembly protein (yeast) pseudogene 225682_s_at 0,46 POLR3H polymerase (RNA) III (DNA directed) polypeptide H (22.9kD) 204033_at 0,46 TRIP13 thyroid hormone receptor interactor 13 203119_at 0,46 CCDC86 coiled-coil domain containing 86 226059_at 0,46 TOMM40L translocase of outer mitochondrial membrane 40 homolog (yeast)-like 202903_at 0,46 LSM5 LSM5 homolog, U6 small nuclear RNA associated (S. cerevisiae) 221586_s_at 0,46 E2F5 E2F transcription factor 5, p130-binding 225768_at 0,47 NR1D2 nuclear receptor subfamily 1, group D, member 2 231863_at 0,47 ING3 inhibitor of growth family, member 3 204114_at 0,47 NID2 nidogen 2 (osteonidogen) 205929_at 0,47 GPA33 glycoprotein A33 (transmembrane) 222962_s_at 0,47 MCM10 minichromosome maintenance complex component 10 225158_at 0,47 GFM1 G elongation factor, mitochondrial 1 201516_at 0,47 SRM spermidine synthase 205129_at 0,47 NPM3 nucleophosmin/nucleoplasmin, 3 203178_at 0,47 GATM glycine amidinotransferase (L-arginine:glycine amidinotransferase) 224441_s_at 0,47 USP45 ubiquitin specific peptidase 45 203196_at 0,47 ABCC4 ATP-binding cassette, sub-family C (CFTR/MRP), member 4 222771_s_at 0,47 MYEF2 myelin expression factor 2 205518_s_at 0,47 CMAH cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMP-N-acetylneuraminate monooxygenase)

196 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Probe Set ID Score(d) Gene Symbol Gene Title 235795_at 0,47 PAX6 paired box gene 6 (aniridia, keratitis) 219068_x_at 0,47 ATAD3A ATPase family, AAA domain containing 3A 218670_at 0,47 PUS1 pseudouridylate synthase 1 209100_at 0,47 IFRD2 interferon-related developmental regulator 2 218488_at 0,48 EIF2B3 eukaryotic translation initiation factor 2B, subunit 3 gamma, 58kDa 221549_at 0,48 GRWD1 glutamate-rich WD repeat containing 1 203737_s_at 0,48 PPRC1 peroxisome proliferator-activated receptor gamma, coactivator-related 1 209205_s_at 0,48 LMO4 LIM domain only 4 230243_at 0,48 RG9MTD2 RNA (guanine-9-) methyltransferase domain containing 2 226882_x_at 0,48 WDR4 WD repeat domain 4 205046_at 0,48 CENPE centromere protein E, 312kDa 228802_at 0,48 RBPMS2 RNA binding protein with multiple splicing 2 1 228381_at 0,48 ATF7IP2 Activating transcription factor 7 interacting protein 2 1554557_at 0,48 ATP11B ATPase, Class VI, type 11B 235289_at 0,48 EIF5A2 eukaryotic translation initiation factor 5A2 3 227594_at 0,48 ZMYM6 zinc finger, MYM-type 6 219546_at 0,48 BMP2K BMP2 inducible kinase 4 204769_s_at 0,48 TAP2 transporter 2, ATP-binding cassette, sub-family B (MDR/TAP) 218719_s_at 0,48 GINS3 GINS complex subunit 3 (Psf3 homolog) 5 223035_s_at 0,48 FARSB phenylalanyl-tRNA synthetase, beta subunit 224204_x_at 0,48 ARNTL2 aryl hydrocarbon receptor nuclear 6 translocator-like 2 223403_s_at 0,48 POLR1B polymerase (RNA) I polypeptide B, 128kDa 7 238762_at 0,49 MTHFD2L methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2-like 220688_s_at 0,49 MRTO4 mRNA turnover 4 homolog (S. cerevisiae) 218219_s_at 0,49 LANCL2 LanC lantibiotic synthetase component C-like 2 (bacterial) 226284_at 0,49 ZBTB2 zinc finger and BTB domain containing 2 204175_at 0,49 ZNF593 zinc finger protein 593 218680_x_at 0,49 HYPK Huntingtin interacting protein K 228397_at 0,49 TUG1 taurine upregulated gene 1 219037_at 0,49 RRP15 ribosomal RNA processing 15 homolog (S. cerevisiae) 206235_at 0,49 LIG4 ligase IV, DNA, ATP-dependent

197 Chapter 6

Probe Set ID Score(d) Gene Symbol Gene Title 203708_at 0,49 PDE4B phosphodiesterase 4B, cAMP-specific (phosphodiesterase E4 dunce homolog, Drosophila) 224721_at 0,49 WDR75 WD repeat domain 75 224315_at 0,49 DDX20 DEAD (Asp-Glu-Ala-Asp) box polypeptide 20 203465_at 0,49 MRPL19 mitochondrial ribosomal protein L19 218695_at 0,49 EXOSC4 exosome component 4 201644_at 0,50 TSTA3 tissue specific transplantation antigen P35B 202431_s_at 0,50 MYC v-myc myelocytomatosis viral oncogene homolog (avian) 202069_s_at 0,50 IDH3A isocitrate dehydrogenase 3 (NAD+) alpha 222765_x_at 0,50 ESF1 ESF1, nucleolar pre-rRNA processing protein, homolog (S. cerevisiae) 213189_at 0,50 MINA MYC induced nuclear antigen 219131_at 0,50 UBIAD1 UbiA prenyltransferase domain containing 1 228043_at 0,50 UTP15 UTP15, U3 small nucleolar ribonucleoprotein, homolog (S. cerevisiae) 220643_s_at 0,50 FAIM Fas apoptotic inhibitory molecule 225834_at 0,50 FAM72A /// family with sequence similarity 72, member A /// LOC653820 /// similar to family with sequence similarity 72, LOC729533 member A 1553535_a_at 0,50 RANGAP1 Ran GTPase activating protein 1 215011_at 0,50 SNHG3 small nucleolar RNA host gene (non-protein coding) 3 222500_at 0,50 PPIL1 peptidylprolyl isomerase (cyclophilin)-like 1 218470_at 0,50 YARS2 tyrosyl-tRNA synthetase 2, mitochondrial 218981_at 0,50 ACN9 ACN9 homolog (S. cerevisiae) 224654_at 0,50 DDX21 DEAD (Asp-Glu-Ala-Asp) box polypeptide 21

198 The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma…

Supplementary Table 3. MM hypoxia-target genes

Gene ID Probe set ID ADM 202912_at ALDOC 202022_at AK3L1 225342_at SLC2A3 202497_x_at VEGFA 210512_s_at BNIP3 201848_s_at PDK1 226452_at DDIT4 202887_s_at P4HA1 207543_s_at HK2 202934_at ANKRD37 227337_at 1 Supplementary Table 4 Spearman’s rank correlation coefficient matrix for AM in relation to HOXB7, ING4, NF-ĸB profile genes and hypoxia target genes in 559 primary MM patients, and in MM subgroups (PR, LB, HY, CD2, MY) with significantly elevated AM expression. 3 MM PR LB HY CD2 MY patients subgroup subgroup subgroup subgroup subgroup (n = 559) (n = 47) (n = 58) (n = 116) (n = 60) (n = 145) 4 AM AM AM AM AM AM ALDOC ns ns ns ns ns ns 5 AK3L1 ns ns ns *-0.2 *-0.3 ns SLC2A3 ns ns ns ns ns ns 6 VEGFA ns ns ns *-0.2 ns ns BNIP3 ns ns ns ns ns ns 7 PDK1 ***0.1 ns **0.3 *0.2 ns ns DDIT4 ns ns ns ns ns ns P4HA1 ns ns ns ns ns ns HK2 ***0.2 ns ns ***0.3 ns **0.3 ANKRD37 ns ns ns ns ns ns HOXB7 ns ns ns ns ns ns ING4 ns ns ns ns ns ns NF-ĸB profile genes ns ns ns ns ns ns ns-not significant, *** indicates p value < 0.001, **indicates p value < 0.01, *indicates p value < 0.05, NF-ĸB profile comprised 11 NF-κB MM signature genes determined by Annunziata et. al [19]

199

Chapter7

Summary and discussion

Summary and Discussion

Multiple myeloma (MM) is an incurable plasma cell malignancy characterized by a monoclonal proliferation of plasma cells in the bone marrow (BM), associated with osteolytic bone lesions, aberrant angiogenesis and pancytopenia.1–8 In 1 the BM, MM cells receive signals to survive and proliferate due to the existence of functional, bi-directional interactions between the MM cells and the other cells in the BM microenvironment.7,9,10 In this thesis, the interaction between the BM environment and MM cells was extensively studied. In particular, we identi- 3 fied novel proteins, N-cadherin and adrenomedullin, contributing to osteolytic bone disease and aberrant MM-induced angiogenesis, respectively. Furthermore, 4 the research presented in this thesis sheds a new light on the role of DKK1 in the progression of MM. Up to now, it was well established that elevated level of DKK1 5 inhibit osteoblast differentiation and is an important trigger of osteolytic bone disease, contributing to the creation of a niche for MM progression.11–13 Interest- 6 ingly, however, our data revealed that DKK1 has a dual pathogenic role in MM since it can also inhibit tumor growth, by suppressing Wnt signaling in malignant 7 plasma cells. During the progression of MM this inhibitory role of DKK1 is lost as a result of aberrant promoter methylation. In search for additional genetic and epigenetic events involved in aberrant Wnt pathway activation in MM, we uncovered downregulation of CYLD expression as a novel prognostic factor in MM. In particular, we showed that the loss of CYLD enhances the Wnt and NF-ĸB signaling, sensitizing MM cells to the Wnt and NF-ĸB ligands. Of note, our data identified downregulation of CYLD as a trigger of malignant plasma cells proliferation and strongly indicates that loss of CYLD enhances MM aggressiveness, through a mechanism involving Wnt pathway hyperactivation.

203 Chapter 7

The Wnt pathway and its negative-feedback regulator DKK1 in malignant plasma cells

Aberrant activation of Wnt signaling plays a crucial role in the pathogenesis of several types of cancer and is most often caused by mutations in Wnt pathway components such as adenomatous polyposis coli (APC) or β-catenin (CTNNB1).14 However, although the Wnt pathway is frequently hyperactivated in MMs, pathway intrinsic mutations have not (yet) been found,15 and the aberrant Wnt activity is the consequence of auto- and/or paracrine stimulation by Wnt ligands. This activation of canonical and non-canonical Wnt signaling promotes dissemination, proliferation, and drug resistance of malignant plasma cells.15–20 Interest- ingly, the malignant plasma cells also express high level of the DKK1, a feed-back inhibitor of Wnt signalling. DKK1 can promote bone disease by inhibiting Wnt signaling in osteoblasts thereby preventing their differentiation.11–13 This may indirectly promote MM growth, since immature osteoblasts express high levels of IL-6, a central growth and survival factor for myeloma plasma cells.21 Further- more, DKK1 enhances the expression of receptor activator of NF-kappa B ligand (RANKL) and downregulates the expression of osteoprotegerin (OPG) in immature osteoblast. The increased RANKL/OPG ratio leads to osteoclast activation promoting osteolytic bone disease.12 Osteoclasts may also support the growth of myeloma cells through secretion of IL-6 and osteopontin, and by adhesive interactions, stimulating the proliferation of malignant plasma cells.22,23 Thus, DKK1 can both exercise paracrine effects on the BM microenvironment, and affect the MM growth by creating an optimal niche for tumor progression. In chapter 2 we review the literature on the role of the Wnt signaling and DKK1 secretion in MM. In particular, we carefully assess the literature on the potential value of Dickkopf-1 (DKK1) as a new therapeutic target for MM. As shown by several studies, anti-DKK1 antibody treatment can inhibit MM-bone disease in various animal models,24–26 strongly suggesting that it might present a valuable therapeutic asset for patients suffering from MM bone disease. Impor- tantly, the data presented in chapter 3 suggest an alternative scenario for the role of DKK1 in the MM. We demonstrate that DKK1 is expressed from the initial phase of the disease onwards and that MM cells are fully responsive to DKK1, confirming that Dickkopf-1 can downmodulates the activity of Wnt pathway in malignant plasma cells. Interestingly, however, in advanced stage MMs and MM cell lines, when the bone marrow independence is acquired, DKK1 expression is lost as a result of aberrant promoter methylation. This observation clearly points to DKK1 as a tumor suppressor for malignant plasma cells. Thus, the data pre-

204 Summary and discussion sented in chapter 3, suggest that therapy based on DKK1 inhibition needs to be carefully monitored since it may not only suppresses the osteolytic bone disease but also promotes the MM tumor growth, especially at extramedullary locations. Indeed, using the 5TGM1 mouse model, it was shown that stimulation of Wnt pathway by lithium chloride significantly increases subcutaneous MM growth.27

Loss of CYLD expression unleashes Wnt signaling in MM

The BM environment supports the growth and survival of malignant plasma cells, via adhesion molecules, cytokines and growth factors. Among the different signaling pathways activated in MM cells, by external stimulation, the NF-κB and Wnt pathway presumably are of major importance.15–17,19,20,28–30 Interestingly, the prevalence of aberrant genetic and epigenetic events in NF-κB and Wnt pathway seems to 1 be substantially higher in MM cell lines (MMCLs) compared to primary MMs,28,30–32 which suggests that constitutively active NF-κB and Wnt signaling renders the tumor cells independent of the BM microenvironment. In chapter 4, we focused on cylindromatosis (CYLD) gene; loss and inactivating mutations of this gene were 3 recently described in MM.28,30,33,34 The CYLD gene was initially identified as a tumor suppressor, mutated in familial cylindromatosis and multiple familial trichoepithe- 4 lioma patients.35 CYLD is a deubiquitinating enzyme acting as a negative regulator of NF-κB and Wnt signaling, by removing lysine-63-linked polyubiquitin chains 5 from NF-κB and Wnt activating proteins.36–38 We established that CYLD acts as an important negative regulator of Wnt and NF-κB signaling in malignant plasma cells. 6 Furthermore, by studying a gene expression data set containing mRNA samples of MMs, we found a significantly lower expression of CYLD in the proliferation sub- 7 group (PR) of MM patients, characterized by poor prognosis and high expression of genes involved in cell growth.39 In support of this observation, introduction of CYLD into UM-3 cells, which lack CYLD expression due to a homozygous deletion, resulted in growth inhibition and increased cell death. Moreover, the proliferation subgroup is characterized by the high expression of Wnt target genes, suggesting that genetic alteration of CYLD expression, together with epigenetic silencing of Wnt inhibitors, contribute to aberrant activation of the Wnt pathway in malignant plasma cells. Importantly, in line with this observation, low expression of CYLD was correlated with inferior progression-free and overall survival in a large cohort of primary MMs confirming that the downregulation or complete loss of this gene plays an instrumental role in the aggressive behavior of malignant plasma cells.

205 Chapter 7

A role of N-cadherin in MM interaction with the BM microenvironment

In chapter 5 we studied the role of N-cadherin in MM. We observed high N-cadherin expression in 50% of MM patients and MM cell lines. Analysis of N-cadherin in the MM subgroups revealed that it is highly, but not exclusively, expressed in MMs bearing a t(4;14) translocation. This subgroup is characterized by a poor prognosis, but high N-cadherin expression is not an independent prognostic factor. The data point to N-cadherin as an important motility protein in malignant plasma cells. Indeed, in solid tumors overexpression of N-cadherin is part of the switch that occurs during epithelial-mesenchymal transition (EMT).40 The process of EMT is directly related to cancer invasiveness, reflected by enhanced cell migration and invasion, resulting in metastatic dissemination. Non-motile polarized epithelial cells, embedded via cell-cell junctions, convert into individual, non-polarized motile and invasive mesenchymal cells.41 However, in MM overexpression of N-cadherin did not affect the (trans-endothelial) migration, and our data suggest that N-cadherin may be involved in the BM retention rather than in dissemination of malignant plasma cells. N-cadherin mediated interaction between MM cells and osteoblasts partially inhibited osteoblast differentiation, suggesting a contribution to osteolytic bone disease. Of note, N-cadherin is expressed by osteoblasts during all stages of the bone formation and N-cadherin-mediated interactions between osteoblasts have shown to be crucial for their differentiation and function.42 In addition, the N-cadherin mediated adhesion between osteoblast and malignant plasma cells might contribute to another feature of MM, i.e. induction of pancytopenia. Since osteoblasts have a central role in the organization of the endosteal hematopoietic stem cell niche,43 it is conceivable that overexpression of N-cadherin by malignant plasma cells, facilitates their access to this niche, leading to dysregulation of hematopoiesis and pancytopenia.

The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma and promotes angiogenesis

The pathogenesis of MM-induced BM angiogenesis is not yet fully understood. At the verge of progression of MGUS to active MM, elevated levels of pro-angiogenic factors in BM plasma and blood result in an “angiogenic switch”.44–46 Since several of the crucial angiogenic factors secreted by MM cells, including

206 Summary and discussion

VEGFA and bFGF, are equally expressed by tumor cells from MGUS, smoldering MM and active MM,47 it has been suggested that the angiogenic switch could be the consequence of increasing tumor burden, rather than aberrant expression of pro-angiogenic factors per se.48 On the other hand, chronic hypoxia may also play an important role in BM angiogenesis in MM.49–51 This is suggested by studies demonstrating that stabilization and nuclear localization of HIF1α affects the transcriptional and angiogenic profiles of myeloma cells, leading to increased expression of VEGFA and IL-8 among other pro-angiogenic factors.50 In chapter 6, we further explored the role of hypoxia in MM, by studying the transcriptional response of MM cells to low oxygen. Interestingly, by the use of gene expression microarray, we identified the pro-angiogenic factor adrenomedullin as the most highly hypoxia-induced gene in MM cells. Adrenomedullin is involved in blood vessel morphogenesis, vasculogenesis, and tumor angiogenesis.52,53 AM stimulates angiogenesis by binding the calcitonin-receptor-like receptor (CRLR), which is widely expressed on normal and hypoxic endothelial cells.54 Importantly, bind- 1 ing of AM to the CRLR/RAMP2 can also transactivate the VEGFR-2, which is responsible for most pro-angiogenic effects of VEGFA, including the stimulation of endothelial cell differentiation, proliferation, migration and morphogenesis. This AM-induced VEGFR-2 transactivation does not require VEGFA, suggest- 3 ing that AM can functionally mimic VEGFA, and thereby contribute to MM- induced angiogenesis.55 Interestingly, however, although AM is a well established 4 HIF1α target gene, containing HRE sites in its promoter as major regulatory sequences,56,57 we observed that several HMCLs and primary MMs also expressed 5 high levels of AM under normoxic conditions. Importantly, these MM cells with normoxic AM expression did not show aberrant basal HIF1α stabilization and 6 displayed no overexpression of other HIF1α/hypoxia target genes, suggesting normoxic regulation of AM expression by HIF1α-independent mechanisms. 7 In line with this notion, our analysis of a large MM gene-expression data set revealed no consistent correlation between AM expression and expression of other hypoxia/HIF1α target genes. These findings imply that mechanisms other than hypoxia can contribute to AM expression in malignant plasma cells, and are consistent with a scenario in which both HIF1α-dependent and independent mechanisms contribute to the “angiogenic switch” in MM. The functional studies presented in chapter 6 strongly support the angiogenic role of AM in MM progression since they demonstrate that forced overexpression and hypoxia-induced AM in MM cells strongly promotes the pro-angiogenic activity of MM cells, as revealed by enhanced endothelial cell proliferation and mesh formation, whereas blockage of endogenously produced as well as hypoxia-triggered AM strongly

207 Chapter 7 reduces the pro-angiogenic activity of MM cells. Thus, the data, demonstrate that MM cells, both in a hypoxia-dependent and independent fashion, aberrantly express and secrete AM, which can mediates MM-induced angiogenesis. This aberrant AM expression could be a major driving force for the angiogenic switch observed during MM progression, which renders AM a novel target for anti- angiogenic therapy in MM.

References

1. Anderson KC, Carrasco RD. Pathogenesis of myeloma. Annu Rev Pathol 2011; 6: 249–274. 2. Bergsagel PL, Kuehl WM. Molecular pathogenesis and a consequent classification of multiple myeloma. J Clin Oncol 2005; 23(26): 6333–6338. 3. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2(3): 175–187. 4. Kyle RA, Rajkumar SV. Multiple myeloma. Blood 2008; 111(6): 2962–2972. 5. Palumbo A, Anderson K. Multiple myeloma. N Engl J Med 2011; 364(11): 1046–1060. 6. Podar K, Chauhan D, Anderson KC. Bone marrow microenvironment and the identification of new targets for myeloma therapy. Leukemia 2009; 23(1): 10–24. 7. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009; 374(9686): 324–339. 8. Rajkumar SV. Multiple myeloma. Curr Probl Cancer 2009; 33(1): 7–64. 9. Giuliani N, Rizzoli V. Myeloma cells and bone marrow osteoblast interactions: role in the development of osteolytic lesions in multiple myeloma. Leuk Lymphoma 2007; 48(12): 2323–2329. 10. Seidl S, Kaufmann H, Drach J. New insights into the pathophysiology of multiple myeloma. Lancet Oncol 2003; 4(9): 557–564. 11. Qiang YW, Barlogie B, Rudikoff S, Shaughnessy JD, Jr. Dkk1-induced inhibition of Wnt signaling in osteoblast differentiation is an underlying mechanism of bone loss in multiple myeloma. Bone 2008; 42(4): 669–80. 12. Qiang YW, Chen Y, Stephens O, Brown N, Chen B, et al. Myeloma-derived Dick- kopf-1 disrupts Wnt-regulated osteoprotegerin and RANKL production by osteoblasts: a potential mechanism underlying osteolytic bone lesions in multiple myeloma. Blood 2008; 112(1): 196–207. 13. Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med 2003; 349(26): 2483–2494. 14. Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 2006; 127(3): 469–80.

208 Summary and discussion

15. Derksen PW, Tjin E, Meijer HP, Klok MD, MacGillavry HD, et al. Illegitimate WNT signaling promotes proliferation of multiple myeloma cells. Proc Natl Acad Sci U S A 2004; 101(16): 6122–6127. 16. Bjorklund CC, Ma W, Wang ZQ, Davis RE, Kuhn DJ, et al. Evidence of a role for activation of Wnt/beta-catenin signaling in the resistance of plasma cells to lenalidomide. J Biol Chem 2011; 286(13): 11009–11020. 17. Dutta-Simmons J, Zhang Y, Gorgun G, Gatt M, Mani M, et al. Aurora kinase A is a target of Wnt/beta-catenin involved in multiple myeloma disease progression. Blood 2009; 114(13): 2699–2708. 18. Qiang YW, Endo Y, Rubin JS, Rudikoff S. Wnt signaling in B-cell neoplasia. Onco- gene 2003; 22(10): 1536–1545. 19. Qiang YW, Walsh K, Yao L, Kedei N, Blumberg PM, et al. Wnts induce migration and invasion of myeloma plasma cells. Blood 2005; 106(5): 1786–1793. 20. Sukhdeo K, Mani M, Zhang Y, Dutta J, Yasui H, et al. Targeting the beta-catenin/TCF transcriptional complex in the treatment of multiple myeloma. Proc Natl Acad Sci U S A 2007; 104(18): 7516–7521. 1 21. Klein B, Zhang XG, Jourdan M, Content J, Houssiau F, et al. Paracrine rather than autocrine regulation of myeloma-cell growth and differentiation by interleukin-6. Blood 1989; 73(2): 517–26. 22. Abe M, Hiura K, Wilde J, Shioyasono A, Moriyama K, et al. Osteoclasts enhance 3 myeloma cell growth and survival via cell-cell contact: a vicious cycle between bone destruction and myeloma expansion. Blood 2004; 104(8): 2484–2491. 4 23. Matsumoto T, Abe M. Bone destruction in multiple myeloma. Ann N Y Acad Sci 2006; 1068: 319–326. 24. Fulciniti M, Tassone P, Hideshima T, Vallet S, Nanjappa P, et al. Anti-DKK1 mAb 5 (BHQ880) as a potential therapeutic agent for multiple myeloma. Blood 2009; 114(2): 371–379. 6 25. Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy JD, Jr., et al. Inhibiting Dickkopf-1 (Dkk1) removes suppression of bone formation and prevents the devel- 7 opment of osteolytic bone disease in multiple myeloma. J Bone Miner Res 2009; 24(3): 425–436. 26. Yaccoby S, Ling W, Zhan F, Walker R, Barlogie B, et al. Antibody-based inhibition of DKK1 suppresses tumor-induced bone resorption and multiple myeloma growth in vivo. Blood 2007; 109(5): 2106–2111. 27. Edwards CM, Edwards JR, Lwin ST, Esparza J, Oyajobi BO, et al. Increasing Wnt signaling in the bone marrow microenvironment inhibits the development of myeloma bone disease and reduces tumor burden in bone in vivo. Blood 2008; 111(5): 2833–2842. 28. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12(2): 115–130.

209 Chapter 7

29. Ashihara E, Kawata E, Nakagawa Y, Shimazaski C, Kuroda J, et al. beta-catenin small interfering RNA successfully suppressed progression of multiple myeloma in a mouse model. Clin Cancer Res 2009; 15(8): 2731–2738. 30. Demchenko YN, Glebov OK, Zingone A, Keats JJ, Bergsagel PL, et al. Classical and/or alternative NF-kappaB pathway activation in multiple myeloma. Blood 2010; 115(17): 3541–3552. 31. Chim CS, Pang R, Fung TK, Choi CL, Liang R. Epigenetic dysregulation of Wnt signaling pathway in multiple myeloma. Leukemia 2007; 21(12): 2527–2536. 32. Demchenko YN, Kuehl WM. A critical role for the NFkB pathway in multiple myeloma. Oncotarget 2010; 1(1): 59–68. 33. Keats JJ, Fonseca R, Chesi M, Schop R, Baker A, et al. Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell 2007; 12(2): 131–144. 34. Jenner MW, Leone PE, Walker BA, Ross FM, Johnson DC, et al. Gene mapping and expression analysis of 16q loss of heterozygosity identifies WWOX and CYLD as being important in determining clinical outcome in multiple myeloma. Blood 2007; 110(9): 3291–3300. 35. Bignell GR, Warren W, Seal S, Takahashi M, Rapley E, et al. Identification of the familial cylindromatosis tumour-suppressor gene. Nat Genet 2000; 25(2): 160–165. 36. Brummelkamp TR, Nijman SM, Dirac AM, Bernards R. Loss of the cylindromatosis tumour suppressor inhibits apoptosis by activating NF-kappaB. Nature 2003; 424(6950): 797–801. 37. Kovalenko A, Chable-Bessia C, Cantarella G, Israel A, Wallach D, et al. The tumour suppressor CYLD negatively regulates NF-kappaB signalling by deubiquitination. Nature 2003; 424(6950): 801–805. 38. Tauriello DV, Haegebarth A, Kuper I, Edelmann MJ, Henraat M, et al. Loss of the tumor suppressor CYLD enhances Wnt/beta-catenin signaling through K63-linked ubiquitination of Dvl. Mol Cell 2010; 37(5): 607–619. 39. Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, et al. The molecular classification of multiple myeloma. Blood 2006; 108(6): 2020–2028. 40. Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. Cadherin switching. J Cell Sci 2008; 121(Pt 6): 727–735. 41. Kang Y, Massague J. Epithelial-mesenchymal transitions: twist in development and metastasis. Cell 2004; 118(3): 277–279. 42. Marie PJ. Role of N-cadherin in bone formation. J Cell Physiol 2002; 190(3): 297–305. 43. Calvi LM, Adams GB, Weibrecht KW, Weber JM, Olson DP, et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature 2003; 425(6960): 841–846. 44. Alexandrakis MG, Passam FJ, Ganotakis E, Dafnis E, Dambaki C, et al. Bone marrow microvascular density and angiogenic growth factors in multiple myeloma. Clin Chem Lab Med 2004; 42(10): 1122–1126.

210 Summary and discussion

45. Di Raimondo F, Azzaro MP, Palumbo G, Bagnato S, Giustolisi G, et al. Angiogenic factors in multiple myeloma: higher levels in bone marrow than in peripheral blood. Haematologica 2000; 85(8): 800–805. 46. Jakob C, Sterz J, Zavrski I, Heider U, Kleeberg L, et al. Angiogenesis in multiple myeloma. Eur J Cancer 2006; 42(11): 1581–1590. 47. Kumar S, Witzig TE, Timm M, Haug J, Wellik L, et al. Bone marrow angiogenic ability and expression of angiogenic cytokines in myeloma: evidence favoring loss of marrow angiogenesis inhibitory activity with disease progression. Blood 2004; 104(4): 1159–1165. 48. Hose D, Moreaux J, Meissner T, Seckinger A, Goldschmidt H, et al. Induction of angiogenesis by normal and malignant plasma cells. Blood 2009; 114(1): 128–143. 49. Asosingh K, De Raeve H, de Ridder M, Storme GA, Willems A, et al. Role of the hypoxic bone marrow microenvironment in 5T2MM murine myeloma tumor progression. Haematologica 2005; 90(6): 810–817. 50. Colla S, Storti P, Donofrio G, Todoerti K, Bolzoni M, et al. Low bone marrow oxygen tension and hypoxia-inducible factor-1alpha overexpression characterize patients 1 with multiple myeloma: role on the transcriptional and proangiogenic profiles of CD138(+) cells. Leukemia 2010; 24(11): 1967–1970. 51. Martin SK, Diamond P, Gronthos S, Peet DJ, Zannettino AC. The emerging role of hypoxia, HIF1 and HIF-2 in multiple myeloma. Leukemia 2011; 25(10): 1533–1542. 3 52. Deville JL, Salas S, Figarella-Branger D, Ouafik L, Daniel L. Adrenomedullin as a therapeutic target in angiogenesis. Expert Opin Ther Targets 2010; 14(10): 1059–1072. 4 53. Kim W, Moon SO, Sung MJ, Kim SH, Lee S, et al. Angiogenic role of adrenomedullin through activation of Akt, mitogen-activated protein kinase, and focal adhesion kinase in endothelial cells. Faseb J 2003; 17(13): 1937–1939. 5 54. Fernandez-Sauze S, Delfino C, Mabrouk K, Dussert C, Chinot O, et al. Effects of adrenomedullin on endothelial cells in the multistep process of angiogenesis: involvement 6 of CRLR/RAMP2 and CRLR/RAMP3 receptors. Int J Cancer 2004; 108(6): 797–804. 55. Guidolin D, Albertin G, Spinazzi R, Sorato E, Mascarin A, et al. Adrenomedullin 7 stimulates angiogenic response in cultured human vascular endothelial cells: involvement of the vascular endothelial growth factor receptor 2. Peptides 2008; 29(11): 2013–2023. 56. Garayoa M, Martinez A, Lee S, Pio R, An WG, et al. Hypoxia-inducible factor-1 (HIF1) up-regulates adrenomedullin expression in human tumor cell lines during oxygen deprivation: a possible promotion mechanism of carcinogenesis. Mol Endocrinol 2000; 14(6): 848–862. 57. Nguyen SV, Claycomb WC. Hypoxia regulates the expression of the adrenomedullin and HIF1 genes in cultured HL-1 cardiomyocytes. Biochem Biophys Res Commun 1999; 265(2): 382–386.

211

Chapter8

Nederlandse samenvatting Acknowledgments Curriculum vitae

Nederlandse samenvatting

Multipel myeloom (MM) is een ongeneeslijke plasmacel maligniteit die wordt gekenmerkt door klonale uitgroei van plasmacellen in het beenmerg (BM), met 1 als gevolg het ontstaan van osteolytische botlaesies, aberrante angiogenese en pancytopenie. In het BM zorgen andere celtypen en extracellulaire matrix mo- leculen (ECM) voor signalen die essentieel zijn voor de groei en overleving van MM cellen. In dit proefschrift is de wisselwerking tussen deze BM niche en MM 3 cellen uitvoerig onderzocht. In hoofdstuk 2 wordt de huidige literatuur over de rol van Wnt signalering 4 en de door tumorcellen uitgescheiden Wnt remmer Dickkopf-1 (DKK1) in de pathogenese van MM bediscussieerd. In het bijzonder wordt de literatuur be- 5 sproken waarin DKK1 als therapeutisch aangrijpingspunt wordt gebruikt in de behandeling van MM. In verschillende studies met verschillende proefdier- 6 modellen is aangetoond dat het ontstaan van MM geïnduceerde botlaesies ge- remd kan worden met DKK1 neutraliserende antilichamen. Door het wegnemen 7 van de remmende werking van DKK1 op Wnt signalering in osteoblasten wordt botaanmaak bevorderd wat mogelijk een belangrijke nieuwe stap kan zijn in de 8 behandeling van MM. In hoofdstuk 3 is te zien dat DKK1 ook in staat is om Wnt signalering in de maligne plasmacellen zelf te remmen. Tevens verliezen MM cellen tijdens ziekteprogressie juist de mogelijkheid om DKK1 uit te scheiden, dit als gevolg van methylering van de DKK1 promotor, wat correleert met verhoogde Wnt signalering in MM patiënten. Deze data duiden erop dat, hoewel het wegnemen van de remmende functie van DKK1 op botaanmaak het ontstaan van botlaesies remt, DKK1 mogelijk juist een tumorsuppressief effect heeft op de MM cellen zelf via remming van Wnt signalering. Het wegnemen van deze rem op Wnt signalering in MM cellen kan mogelijk leiden tot tumor dissemina- tie, drugresistentie en tumor uitgroei als gevolg van verhoogde Wnt signalering.

215 Chapter 8

Deze data suggereren dat toepassing van DKK1-neutraliserende therapie kri- tisch bekeken en geëvalueerd moet worden aangezien niet alleen de osteolytische botlaesies onderdrukt worden maar ook, vooral extramedullair gelokaliseerde, MM tumorgroei kan worden gestimuleerd. In hoofdstuk 4 ligt de focus op de gevolgen van genetische defecten in het cylindromatosis (CYLD) gen in de pathogenese van MM. Verlies van CYLD functie als gevolg van deleties en/of inactiverende mutaties zijn recent in MM beschreven. Het CYLD gen werd oorspronkelijk beschreven als een tumorsuppressor waarin inactiverende kiembaan mutaties aanwezig zijn in patiënten die lijden aan familiare cylindromatosis (FC) en multiple familiare trichoepithelioma (MFT). CYLD is een deubiquitinase dat een remmende werking heeft op NF-κB en Wnt signalering. CYLD verwijdert polyubiquitine ketens die via lysine-63 zijn gekoppeld en een stabiliserende werking hebben op eiwitten die betrokken zijn bij NF-κB en Wnt signalering. Ook in MM cellen is CYLD een belangrijke negatieve regulator van NF-κB en Wnt signalering. Zo resulteert kunstmatige remming van CYLD in MM cellen tot een sterke toename van NF-κB en Wnt signalering en leidt ectopische expressie van CYLD in een deficiënte cellijn juist tot remming van beide signaleringscascades. Bovendien is er een significant lagere expressie van CYLD mRNA in de proliferatie (PR) genetische subgroep die wordt gekenmerkt door hoge expressie van Wnt targetgenen. Ook is er een negatieve correlatie tussen CYLD expressie en progressievrije en totale overleving in een groot cohort van primaire MM patiënten. Deze bevindingen suggereren dat verlies van CYLD functie, als gevolg van genetische defecten, Wnt/β-catenine en NF-κB signalering faciliteert en zo bijdraagt aan ziekteprogressie. In hoofdstuk 5 wordt de rol van het adhesiemolekuul N-cadherine in MM onderzocht. Ongeveer 50% van de onderzochte primaire MM en MM cellijnen brengt N-cadherine hoog tot expressie. Verdere analyse van N-cadherine expressie in verschillende MM genetische subgroepen laat zien dat N-cadherine hoog, maar niet exclusief, tot expressie komt in MM patiënten met een t(4;14)/MM- SET translocatie. Deze subgroep wordt gekenmerkt door een slechte prognose, hoewel hoge N-cadherine expressie geen onafhankelijke prognostische factor is. Functioneel is N-cadherine betrokken bij migratie en homotypische adhesie aan BM niche cellen zoals osteoblasten. Blokkering van N-cadherine met neutraliserende antilichamen of remming van N-cadherine expressie met shRNAs leidt tot verminderde adhesie en toegenomen motiliteit/migratie met als gevolg een verhoogd aantal circulerende MM cellen in vivo. Echter, remming van N- cadherine heeft geen effect op transendotheliale migratie, wat mogelijk wijst op een rol van N-cadherine in retentie van MM cellen in het BM. Interessant genoeg

216 Nederlandse samenvatting is de N-cadherine gemedieerde MM-osteoblast interactie niet eenzijdig. Zo leidt deze wisselwerking ook tot remming van osteoblastdifferentiatie, wat het ontstaan van osteolytische botlaesies kan bevorderen. Mogelijke verklaringen zijn verlies van onderlinge N-cadherine interacties tussen osteoblasten of remming van Wnt signalering, hetgeen beide een remmende werking heeft op osteoblastdifferentiatie en botaanmaak. Bovendien hebben osteoblasten een centrale rol in de stamcelniche en leidt N-cadherine gemedieerde binding van MM cellen aan osteoblasten mogelijk tot verdringing van stamcellen uit deze niche die essentieel is voor het goed functioneren van stamcellen. Dit draagt mogelijk bij aan het ontstaan van pancytopenie, een veel voorkomend gevolg van MM. In Hoofdstuk 6 wordt de transcriptionele respons van MM cellen op hypoxie en het effect hiervan op de beenmergomgeving onderzocht. Na blootstelling van MM cellijnen aan hypoxie werd de pro-angiogenetische factor adrenomedullin (AM) als het sterkst opgereguleerde gen geïdentificeerd. AM stimuleert 1 angiogenese via binding aan de calcitonin-receptor-like receptor (CRLR), die tot expressie komt op normale en hypoxische epitheliale cellen. De AM promotor bevat verscheidene bindingsplaatsen voor de hypoxie effector HIF-1α en hypoxie vormt dus een belangrijke trigger voor AM expressie. Echter, ook 3 onder normoxische condities worden door verscheidene MM cellijnen en primaire MM al grote hoeveelheiden AM uitgescheiden. Interessant genoeg is er 4 in deze MM geen stabilisatie van HIF1α detecteerbaar en komen andere HIF1 α/hypoxie target genen niet tot expressie. Deze bevindingen suggereren dat zo- 5 wel hypoxie-afhankelijke als hypoxie-onafhankelijke mechanismen bijdragen aan AM secretie door maligne plasmacellen. Uit functionele proeven blijkt dat 6 zowel hypoxie-geïnduceerde als ectopische AM expressie een sterk pro-angioge- netisch effect heeft op endotheelcellen, wat kan worden geblokkeerd door speci- 7 fieke AM remmers. Samengevat laten de data zien dat aberrante AM expressie een belangrijke drijvende kracht is van de ‘angiogenetische switch’ die wordt 8 geobserveerd tijdens MM progressie, hetgeen AM een potentieel target maakt voor anti-angiogenetische therapie in MM.

217 Acknowledgments

Having lived 7 years in the Netherlands, I had never imagined the moment I am writing the last part of my thesis, acknowledgments, but this moment has come so let me start, of course from the top… My Dear Bosses, Steven and Marcel… Steven, I would like to thank you very much for giving me opportunity to get the PhD in your group. I highly appreciate all your support and help. Although, my ‘PhD journey’ was sometimes difficult, I see now how much I am getting out of everything I learnt in your lab. Thank you very much Steven for everything… When I close my eyes and I recall all the different meetings I was giving the talks, I immediately see my second boss, Marcel, sitting in the front row and mov- ing one hand up and the second hand down. Marcel was continuing this original pantomime, which only us both could understand, until the presentation was over☺. In this way, Marcel was at every single meeting helping me to keep the proper loudness of my voice and the speed of my talking. Marcel, you taught me a lot of things. I am very very grateful for all your help and support during my PhD. At that point, I would like to thank Richard Groen (for whom I used to work as a student), who first came up with the idea of me doing the PhD. Now, after all these years I can certainly say, Richard had an excellent idea. Richard thank you very much for everything…, I wish you all the best in your life and the research career. Thinking about my lab-life I can definitely divide it on two parts, the lab-life without Anneke and the lab-life with Anneke. I do not remember exactly when Anneke joined me (probably around 3rd year of my PhD study) but I remember very well that from this moment I was never ever alone in the lab, Anneke shared with me the joy of the good results but even more importantly she supported me

218 Acknowledgments in disappointing moments so I did not have to go through them alone. Anneke, thank you SO much for all your help and support. It was hyper-great to work with you. I am so much grateful for all you did… In the end of my PhD when I was running out of my energy sources, I was super lucky to meet Harmen, who joined me as a student and immediately it turned out we enjoy working, chatting, spinning, life discussing and beer drink- ing together. Harmen, it was brilliant to work with you. I am very grateful for all your help. Thank you vvvvvvvvvv much… At that point, I would like to mention Zaira who was helping me with ADM project. Zaira, I was very happy to work with you☺. All the best to you. The cooperation and ideas exchange are very important during the PhD and I was very lucky to meet great cooperators. I would like to thank Madelon Mau- rice and Danielle Tauriello from the University of Utrecht. In fact, the idea of CYLD project was born during the Wnt meeting in Washington and with the 1 support and help of Madelon and Danielle I could reveal the role of CYLD in multiple myeloma. I would like to thank Mark van Duin and Annemiek Broyl from Professor Pieter Sonneveld group for their cooperation in CYLD project. In a special way I would like to thank Mark, who has spent a lot of time with me 3 helping to understand and break through the difficult array data. I would like to thank very much Marie Jose Kersten, who was always very 4 supportive for me and provided a lot of help with the study of patients materials. Many thanks I direct to Peter T. and Ono who shared their experience with me 5 in relation to HUVECs culture and isolation. I would like to thank Elco for help with the immunohistochemistry and Berend for help with FACS sorting. 6 I also would like to mention Chris van der Loos† (2013), a colleague who I remember for his help in difficult immunohistochemistry. 7 I was not always lucky with my experiments in the lab but I had amazing luck to be always surrounded by the great people in the lab. I would like to thank 8 Annemieke, Leonie and Loes for being always willing to help whenever I asked. Monique and Esther, I thank very much for all the help with PCR related issues. In fact, Monique was the first person in the lab who taught me in practice how PCR works, and from that day, I am a fan of PCR and a fan of Monique☺. In general, I would like to thank all people from diagnostic team for being always helpful. In the lab-life there are always some difficult problems, but in the pathology department, we have the lab-fairy Nike, who has a magic skill to make all problems disappear. Nike, thank you very much for everything, I felt your positive energy and support during my whole PhD.

219 Chapter 8

I also direct my thanks to computer guys team, Patric, Remco and Ronald, who were always willing to solve my computer problems and to Frank, who was able to arrange all administration issues in a funny way. And now time for my AIOs’ room roommates☺. I would like to thank the ‘old experienced AIOs’, (this is how I saw Rogier R. and Febe when I started my PhD) for always being willing to share their experience with me. I wish you all the best in your lives and careers. When old AIOs left, the new ones came. I had the great time with Martine, Martin (because of Martin the Epstein-Bar was re-discovered), Linda and Katka. All the best to you. With Richard B. and Robbert H. I sit in the same part of the room (behind the wall) to the last day of my PhD. They were great colleagues, and I really enjoyed sharing the room with them. All the best guys. At that point I would like to mention Jeroen G., (who joined the department unfortunately in the end of my PhD) and Carel van Noesel,who gave me a lot of good remarks related to my projects. Thank you very much. Jurrit and Sander, two brilliant guys in the myeloma group, who were working on colon cancer☺. Although our research subjects were far away from each other’s, I always felt very strongly the positive and super highly supportive signals from you both. Chinese say: ‘A good neighbor-a found treasure’, I was also very lucky to have the best neighbor ever at the Anatomy Department, Aho. So now, time for Tamas, the super original, controversial, mysterious indi- vidualist. Tamas, I really enjoyed a lot our controversial life discussions☺. All the best to you, it was a great experience to meet you during my ‘PhD journey’. In this point, I would also like to thank Monika, who was always very supportive for me. Monika, it was great to have small polish-girls- chat with you. In general, there were many great people in the pathology department, especially kidney group, who I did not mention personally, but I would like to thank all of them for the great, unique and unforgettable atmosphere of this place. Foreigners are always attracted to other foreigners and it was always the case with me. Elena, Emanuele, Maria, Anna S., Marco, Gleb, Karene and Anand. It was hyper fantastic to meet you and to spend time with you. Thank you very very much guys for everything. In research life people often come and go what gives the impression that noth- ing is stable. But fortunately, I was very lucky to have Anand around, who was in the pathology department when I came and who is still there. Anand, many things we did together, a lot of discussions, and a great time in general. Thank

220 Acknowledgments you very much Anand for being stably around during my PhD time. It was really great and very important for me. Here, I would like to direct my thanks also to Pran, who, although left to UK, was also long time stably around in the pathology department and came with many interesting ideas, especially in hypoxia project. Thank you very much Pran. As for Karene, I can certainly say she was my best and the most supportive friend during the time she was around (and later as well). I remember our lunches; small and big talks about really everything. I also remember very exactly when(too often probably☺) I was asking Karene, ‘Do you think the day of my defense will ever come?’ and Karene was always answering the same ‘sure, I have no doubts’. Karene, I am so grateful for your friendship. It was a big luck for me to meet you. When in spring 2005 I met the Socrates student from Lisbon, Emilio, for the first time, I did not expect at all than 9 years later, we will be still friends, having fun and still discovering the new places in the Netherlands. Emilio, although we 1 see each other not more often than once in 3 years, there is always something to tell and to discuss, this happens only between the real friends… Many thanks I direct to Ania B. and Zofia M., it is quite amazing to keep in touch constantly from years although being thousands of kilometers from each 3 other. I cannot imagine to survive all these years without our skype conferences, organized every time if ‘emergency problem solving’ was needed. 4 I would like to thank my polish friends, who stayed in Netherlands, which is not so bad because I can always visit them☺. Agata M., Agnieszka S., Monika, 5 Robert, Justyna and Marcin, thank you all for your support and great time we had together during my ‘dutch times’. 6 And of course my ‘very old’ and best of the best friends, the one and only Asia D., Anna S., Aga B., Gosia H. and Agatka T., who were always hyper-sup- 7 portive and constantly from the 4th year of my PhD has been asking the same question ‘is the date of the defense maybe already set?’ so finally, it is set and we 8 can finally celebrate… I was finishing my PhD thesis in Poland, working in the Chair of Medical Bio- chemistry at Jagiellonian University Medical College. At that point I would like to thank professor Piotr Laidler and all the staff from the Chair of Medical Biochem- istry whose support in the last part of my PhD I could feel very well. Especially, I thank my roommates (rather roomfriends) and the friends of the room 102D☺. Serdeczne dziękuję mojej najbliższej rodzinie, Kasi, Piotrkowi, Oli, Kubusiowi i cioci Helenie, których wsparcie przez wszystkie te lata było dla mnie ogromnie cenne.

221 Chapter 8

Wyjątkowo szczególne podziękowania kieruję do moich kochanych rodziców, za to, że zawsze byli przy mnie i zawsze bardzo silnie mnie wspierali. Nigdy nie osiągnęłabym tego co mam gdyby nie ich wiara we mnie…

Kinga K.

222 Curriculum vitae

Kinga Kocemba was born in 1980 in Oświęcim in Poland. In 2005, she gradu- ated from Jagiellonian University in Krakow where she was awarded the title of 1 Master Degree in the Department of Biochemistry at the faculty of Biotechnol- ogy, Biochemistry and Biophysics. As an undergraduate student she participated in the Socrates Erasmus program broadening her experience in the Department of Pathology, at the Academic Medical Center of the University of Amsterdam. 3 The main subject Kinga Kocemba was working on for six months was related to the role of DKK1 in the progression of multiple myeloma. In December 2005, 4 she moved to the Netherlands, where she continued her science career as a PhD researcher in the Department of Pathology, at the Academic Medical Center of 5 the University of Amsterdam, under supervision of Professor Steven Pals. The focus of her research was the determination of molecular alterations in malig- 6 nant plasma cells facilitating their interaction with the bone marrow environment, with emphasis on Wnt signaling. 7 She is the first author and co-author of several articles in prestigious journals as: Leukemia, Haematologica, PLoS ONE. The data obtained during her PhD 8 study was presented on several well-known conferences in Greece, U.S.A, Sweden, France and the Netherlands. In October 2012, she took up a position in the Chair of Medical Biochemistry, Jagiellonian University-Collegium Medicum in Poland, where she continues her research on multiple myeloma, with main focus on the role of hypoxia in the progression of the disease.

223