ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2013

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 928

Transcriptional and EpigeneticRegulation of Epithelial-Mesenchymal Transition

E-JEAN TAN

ISSN 1651-6206ISBN 978-91-554-8734-8urn:nbn:se:uu:diva-206120

Dissertation presented at Uppsala University to be publicly examined in Room C8:301,Biomedical Centre, Uppsala University, Husargatan 3, Uppsala, Friday, October 18, 2013 at13:15 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will beconducted in English.

AbstractTan, E.-J. 2013. Transcriptional and Epigenetic Regulation of Epithelial-MesenchymalTransition. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine 928. 51 pp. Uppsala. ISBN 978-91-554-8734-8.

The transforming growth factor beta (TGFβ) is a cytokine that regulates a plethora of cellularprocesses such as cell proliferation, differentiation, migration and apoptosis. TGFβ signalsvia serine/threonine kinase receptors and activates the Smads to regulate gene expression.Enigmatically, TGFβ has a dichotomous role as a tumor suppressor and a tumor promoter incancer. At early stages of tumorigenesis, TGFβ acts as a tumor suppressor by exerting growthinhibitory effects and inducing apoptosis. However, at advanced stages, TGFβ contributes totumor malignancy by promoting invasion and metastasis.

The pro-tumorigenic TGFβ potently triggers an embryonic program known as epithelial-mesenchymal transition (EMT). EMT is a dynamic process whereby polarized epithelial cellsadapt a mesenchymal morphology, thereby facilitating migration and invasion. Downregulationof cell-cell adhesion molecules, such as E-cadherin and ZO-1, is an eminent feature of EMT.TGFβ induces EMT by upregulating a non-histone chromatin factor, high mobility group A2(HMGA2). This thesis focuses on elucidating the molecular mechanisms by which HMGA2elicits EMT.

We found that HMGA2 regulates a network of EMT transcription factors (EMT-TFs), suchas members of the Snail, ZEB and Twist families, during TGFβ-induced EMT. HMGA2 caninteract with Smad complexes to synergistically induce Snail expression. HMGA2 also directlybinds and activates the Twist promoter. We used mouse mammary epithelial cells overexpressingHMGA2, which are mesenchymal in morphology and highly invasive, as a constitutive EMTmodel. Snail and Twist have complementary roles in HMGA2-mesenchymal cells during EMT,and tight junctions were restored upon silencing of both Snail and Twist in these cells. Finally,we also demonstrate that HMGA2 can epigenetically silence the E-cadherin gene. In summary,HMGA2 modulates multiple reprogramming events to promote EMT and invasion.

Keywords: TGFβ, EMT, HMGA2, Snail, Twist, E-cadherin, epigenetics, invasion

E-Jean Tan, Uppsala University, Ludwig Institute for Cancer Research, Box 595,SE-751 24 Uppsala, Sweden. Science for Life Laboratory, SciLifeLab, Box 256,SE-751 05 Uppsala, Sweden.

© E-Jean Tan 2013

ISSN 1651-6206ISBN 978-91-554-8734-8urn:nbn:se:uu:diva-206120 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-206120)

Till min älskade Daniel

”Att arbeta hårt och mycket räcker inte. Man måste arbeta hårt och mycket tillsammans.”

- Mattias Dalhgren

(To work hard and a lot is not enough,you have to work hard and a lot together.)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Thuault S, Tan E-J, Peinado H, Cano A, Heldin C-H, and

Moustakas A. (2008). HMGA2 and Smads co-regulate SNAIL1 expression during induction of epithelial-to-mesenchymal tran-sition. The Journal of Biological Chemistry, 283: 33437–46.

II Tan E-J, Thuault S, Caja L, Carletti T, Heldin C-H, and Mous-takas A. (2012). Regulation of transcription factor Twist ex-pression by the DNA architectural protein high mobility group A2 during epithelial-to-mesenchymal transition. The Journal of Biological Chemistry, 287: 7134–45.

III Tan E-J, Kahata K, Thuault S, Heldin C-H, and Moustakas A.

(2013). The high mobility group A2 protein epigenetically si-lences the Cdh1 gene during epithelial-to-mesenchymal transi-tion. Manuscript.

Reprints were made with permission from the respective publishers.

“Do or do not. There is no try.”- Yoda

Contents

Introduction ................................................................................................... 11

Transcriptional control and epigenetics ........................................................ 12 Transcription ............................................................................................ 12 Epigenetics ............................................................................................... 13

Histone modification ........................................................................... 13 DNA methylation ................................................................................ 14 MicroRNA ........................................................................................... 15

TGFβ signaling ............................................................................................. 16 TGFβ ligands and receptors ..................................................................... 16 Smad transducers ...................................................................................... 17 Non-Smad signaling ................................................................................. 20 TGFβ in cancer ......................................................................................... 21

High mobility group A proteins .................................................................... 23 HMGA2 expression and its regulation ..................................................... 23 HMGA2 mode of action ........................................................................... 24 HMGA2 in cancers ................................................................................... 26

Epithelial-mesenchymal transition ................................................................ 27 EMT in embryogenesis ............................................................................ 28 TGFβ induction of EMT .......................................................................... 28 Transcriptional inducers of EMT ............................................................. 29 Epigenetic regulation in EMT .................................................................. 31 Attributes of EMT .................................................................................... 32

Present Investigation ..................................................................................... 33 Paper I: HMGA2 and Smads co-regulate Snail1 ...................................... 33 Paper II: HMGA2 regulates Twist1 during EMT .................................... 34 Paper III: HMGA2 epigenetic regulation of E-cadherin .......................... 34

Concluding remarks ...................................................................................... 36

Acknowledgements ....................................................................................... 38

References ..................................................................................................... 40

Abbreviations

AP1 Activator protein 1 AZA 5-aza-2’-deoxycytidine bHLH Basic helix-loop-helix CDK Cyclin-dependent kinase ChIP Chromatin immunoprecipitation CSC Cancer stem cell DNMT DNA methyltransferase HAT Histone acetyltransferase HDAC Histone deacetylase HMGA2 High mobility group A2 ECM Extracellular matrix EMT Epithelial-mesenchymal transition EMT-TF EMT-inducing transcription factor GSK3 Glycogen synthase kinase 3 HSC Hematopoietic stem cell KDM Histone lysine demethylase KMT Histone lysine methyltransferase MAPK Mitogen-activated protein kinase MH1/2 Mad homology 1/2 miRNA MicroRNA MET Mesenchymal-epithelial transition mTORC1/2 Mammalian target of rapamycin complex 1/2 NES Nuclear export signal NLS Nuclear localization signal NSC Neuronal stem cell PDGF Platelet-derived growth factor PDGFR Platelet-derived growth factor receptor PI3K Phosphotidylinositol-3’ kinase PRC1/2 Polycomb repressive complex 1/2 PTM Post-translational modification SBE Smad-binding element TβRI TGFβ type I receptor TβRII TGFβ type II receptor TF Transcription factor TGFβ Transforming growth factor beta TSS Transcription start site

11

Introduction

The cell is the fundamental unit of life and approximately 100 trillion (1014) cells make up the human body. Although every human cell has the same genetic information, what dictates a cell fate and function lies in its gene expression programs. Cells receive and respond to cues from the environ-ment, other cells and itself. This cellular communication is vital for biologi-cal processes during development, immunity, tissue repair and homeostasis. From the signals received, cells relay the 'messages' to their nuclei which function like air traffic control towers; thus cells know which genes to turn on or off, and when to press the switch.

There are multiple layers of control that orchestrate such gene activities in a stringent and sophisticated manner. Gene expression can be regulated from the transcriptional level (which is the most utilized mode) to post-translational level after the protein has been made. Transcriptional regulation is controlled by transcription factors and cofactors, and is dependent on the chromatin landscape. When gene expression programs go awry, the conse-quences often lead to the development and progression of diseases, such as cancer.

Epithelial-mesenchymal transition (EMT) is an important differentiation event during embryogenesis and cancer progression, which involves several gene expression programs. During EMT, polarized epithelial cells dissolve their cell-cell and cell-matrix contacts and change their morphology into motile, mesenchymal cells. This thesis investigates how the cytokine trans-forming growth factor beta (TGFβ), through high mobility group A2 (HMGA2) protein, regulates a complex network of transcription factors to reprogram and establish EMT.

12

Transcriptional control and epigenetics

The human genome project has mapped out and catalogued about 20,000 protein-coding genes, which is less than 2% of the genome (1). The rest of the DNA, previously thought to be 'junk', has recently been annotated under the Encyclopedia of DNA Elements (ENCODE) initiative and found to har-bor functional and regulatory elements that modulate gene expression (2). The transcription and expression of a gene has to be tightly regulated and coordinated in a spatial-temporal manner. This provides the basis for cellular processes such as proliferation, differentiation, metabolism and apoptosis. Deregulation of gene expression programs can have pathological conse-quences (3).

Transcription The transcription process is regulated by transcription factors (TFs) and reg-ulatory elements such as core and proximal promoters, enhancers, silencers, insulators or locus control regions (4). TFs are proteins that recognize and bind to specific sequences on the DNA and they can regulate recruitment and activities of RNA polymerase II and cofactors to target genes. General TFs form a pre-initiation complex to direct RNA polymerase II to the core promoter to carry out basal transcriptions (4, 5). Activating TFs (activators) promote and enhance gene transcription together with their cofactors, termed as coactivators. Conversely, transcription is blocked by repressive TFs (re-pressors) and their cofactors, or corepressors. Repressors function by pre-venting or competing with activators to bind specific DNA sites; or sterically hinder the pre-initiation complex assembly (4).

Soon after transcription is initiated, the RNA polymerase complexes are stalled by pause control factors. In order to escape from the promoter, addi-tional TFs and elongation factors are recruited to release RNA polymerase II from the pause control factors and allow transcription to continue (3). The multi-protein Mediator complex is another crucial cofactor which partici-pates in almost all steps of the transcription process, from the assembly of pre-initiation complex to the elongation of mRNA transcript (6).

13

Epigenetics Gene regulation is also dependent on the chromatin conformation, which is influenced by chemical modifications and chromatin-binding factors. The chromatin is a highly ordered structure of packed DNA and histones, which form nucleosomes. Nucleosomal packing helps to compact the whole ge-nome to fit into the nucleus. A network of mobile proteins surrounds the chromatin and dynamically modulates the chromatin structure in order to control precise gene activities (5). There are two general states of the chro-matin: heterochromatin and euchromatin. The highly condensed heteroch-romatin poses a barrier to TFs and RNA polymerase complexes, and thus, is associated with transcriptional repression. In contrast, the loosely packed euchromatin is in an ‘open’ conformation which allows easy access to the transcriptional machinery and therefore, is transcriptionally active (5).

Epigenetics is the heredity of a phenotype that is the result of alterations on the chromatin and not due to genetic changes in the DNA nucleotide se-quence (7). There are two main epigenetic mechanisms, histone modification and DNA methylation, which are intrinsically linked and have a great impact on chromatin function (8). For example, the transcription start sites (TSS) of active genes are often associated with active histone marks and void of DNA methylation (5, 9). Although epigenetic modifications can be stably inhe-rited, they could also dynamically change depending on developmental or environmental cues. Epigenetic regulation is an important program used in cell fate specification and lineage commitment. It is also used as one of the reprogramming strategies of cell identity during cancer development (10–12).

Histone modification The single unit of the chromatin is called a nucleosome that is wrapped around by 147 bp DNA. A nucleosome is an octamer consisting of 2 mole-cules of each histones H2A, H2B, H3 and H4. The histone N-terminal tails protrude out of the nucleosome and are readily prone to post-translational modifications (PTMs) such as acetylation, methylation, phosphorylation and ubiquitination, and thereby affecting chromatin structure. Based on the his-tone PTM marks, also known as the histone code, the chromatin decon-denses and extends to facilitate DNA processes like transcription, replication or repair; alternatively, the chromatin condenses into a compact form to si-lence genes (13). These histone PTM marks can also serve as docking sites for other chromatin-binding proteins and H3 acetylation and methylation are some of the PTMs most extensively studied, which will be briefly described here.

Histones are positively charged and form very tight interactions with the negatively charged DNA. Acetylation of lysine residues on the histone N-

14

terminal tails by histone acetyltransferases (HATs) neutralizes the positive charges and potentiates the chromatin into an ‘open’ conformation. In con-trast, histone deacetylases (HDACs) deacetylate lysine residues, and thus, are associated with repression (13).

Histone methylation on lysine residues does not impose an overall net charge on the nucleosome as acetylation does. Lysines can be mono-, di-, or trimethylated by histone lysine methyltransferases (KMTs), whereas methyl-groups are removed (demethylation) by histone lysine demethylases (KDMs) (14). Methylation of histone H3 on lysine 4 (H3K4me) is found enriched at active promoters, while methylation of lysine 9 or 27 (H3K9me or H3K27me) is associated with inactive promoters and heterochromatin re-gions (13, 14). Both KMTs and KDMs are highly specific enzymes in target-ing particular lysine residues, especially in regard to their methylation status (14). For example, G9a (KMT1C) mono- or dimethylates H3 lysine 9 (H3K9me1/me2) from the unmethylated state, whereas SUV39H1 (KMT1A) trimethylates H3K9 (H3K9me3); and LSD1 (KDM1A) can demethylate H3K4me1/me2 but is unable to catalyze H3K4me3 to H3K4me2 (14). The repressive histone H3K9me3 and H3K27me3 moieties can serve as target sites for heterochromatin protein 1 (HP1) and Polycomb repressive com-plexes (PRC), respectively, both of which are linked to DNA methylation and involved in the long-term silencing of the chromatin (15).

DNA methylation DNA methylation is an important gene regulatory mechanism that controls many biological processes, such as chromosome X-inactivation, genomic imprinting and maintenance of genomic stability (8, 9, 16). Defects and ab-errations in this epigenetic event lead to embryonic lethality and are linked to cancer progression (10, 16).

In mammals, DNA can be methylated on the carbon-5 position of the cy-tosine ring by DNA methyltransferases (DNMTs), generating 5-methylcytosine. Such methylated cytosines are usually located in CpG di-nucleotides (i.e., a cytosine bridged to a guanine by a phosphodiester bond). Most CpG sites in the genome are methylated except for those found clus-tered together in regions known as CpG islands, some of which are found near active gene promoters. When the CpG islands near the TSS become methylated, then such genes are often stably silenced (9).

DNMT1 is known as the maintenance DNMT, which preserves the me-thylation pattern of genes after every replication. It binds hemimethylated DNA and methylates the newly synthesized, unmethylated strand (Figure 1a). De novo DNMTs, DNMT3A and DNMT3B, are able to recognize both unmethylated and hemimethylated DNA so as to establish de novo methyla-tion (Figure1b) (17). DNMT3L is also related to the DNMT3 family but lacks catalytic activity. However, it can interact with DNMT3A and

15

DNMT3B to enhance their DNA methylating activities (17, 18). Further-more, DNMT3L forms a complex with DNMT3A and guides it to the tar-geted loci (19).

Figure 1. DNA methylation. (a) During DNA replication, maintenance methylation is carried out by DNMT1, which recognizes and binds hemimethylated DNA. (b) De novo methylation by DNMT3 family members. White circles, unmethylated CpG; black circles, methylated CpG.

DNA methylation and histone modification are interrelated and contribute to the chromatin architecture by establishing and maintaining long-term re-pression (8, 15). Methylated cytosines provide binding sites for the methyl-cytosine-binding domain (MBD) family of proteins, which recruit corepres-sors and HDACs (20). Alternatively, the DNMTs are directed to the chroma-tin by histone-modifying enzymes or chromatin-binding proteins. DNMTs can interact with HDAC (21); KMTs such as G9a, SUV39H1, EZH2 and SETDB1 (22–25); and HP1 (23). Hence, this leads to heterochromatinization and gene silencing (15).

MicroRNA MicroRNA (miRNA) regulates gene expression post-transcriptionally. It is a non-coding single stranded RNA that averages 22 nucleotides in length. miRNA is first transcribed by RNA polymerase II into long primary miRNA (pri-miRNA), which is cleaved by the RNase Drosha to yield precursor miRNA (pre-miRNA) with a characteristic stem-loop secondary structure. Pre-miRNA is exported by Exportin-5 from the nucleus into the cytoplasm, where it is further processed into mature miRNA by the RNase Dicer. miR-NAs silence genes by base-pairing their so-called ‘seed’ sequences (2 – 7 nucleotides) to their mRNA targets, usually located at the 3’ untranslated regions. This leads to degradation of the targeted mRNA and/or inhibition of its translation (26).

16

TGFβ signaling

TGFβ is a cytokine and the prototype of the TGFβ family whose 33 mem-bers include bone morphogenetic proteins (BMP), activins, growth and diffe-rentiation factors (GDF) and Müllerian inhibiting substance (MIS). The TGFβ family signaling pathways are involved in embryonic development, tissue homeostasis and repair, and modulation of the immune system (27, 28). TGFβ propagates its signals in a cascading fashion: ligand binds to re-ceptors, receptors activate their substrates to mediate downstream events, such as gene transcription. Dichotomously, TGFβ is a tumor suppressor at early stages of tumorigenesis, but at advanced stages, it contributes towards tumor malignancy by promoting EMT, invasion and metastasis (29–32).

TGFβ ligands and receptors There are 3 isoforms of TGFβ (TGFβ1, TGFβ2 and TGFβ3) in mammals which are secreted in latent forms and stored as inactive complexes in the extracellular matrix (ECM). Latent TGFβ is activated and liberated from the ECM by proteases and matrix metalloproteinases (33).

TGFβ ligands elicit their cellular effects by binding to type I and type II serine/threonine kinase receptors. Seven type I receptors (activin receptor-like kinases, ALK1-7) and five type II receptors (ActRII, ActRIIB, TβRII, BMPRII and MISRII) are found in vertebrates (27, 28). They are structurally related and contain an N-terminal extracellular domain, a transmembrane domain and a C-terminal kinase domain. The type I receptor has an addition-al GS domain adjacent to its kinase domain (27). TGFβ also binds to acces-sory receptors called betaglycan, or endoglin in endothelial cells, which ei-ther enhance or limit TGFβ signaling (34).

The dimeric TGFβ molecules bind two type I receptors, TβRI (ALK5) and two constitutively active type II receptors (TβRII) as a heterotetrameric complex (Figure 2). TGFβ has a higher affinity for TβRII and does not inte-ract with TβRI on its own (27, 35). Thus, TGFβ first binds to TβRII and then recruits TβRI to the complex. The GS domain of TβRI is phosphorylated by TβRII on serine and threonine residues and activated TβRI phosphorylates the Smad proteins that are the intracellular mediators in the TGFβ signaling cascade (27).

17

Figure 2. The TGFβ/Smad pathway. TGFβ ligand binding leads to formation of a TβRII-TβRI heterotetrameric receptor complex. TβRI is activated and phosphory-lates the R-Smads, which in turn translocate with co-Smad to the nucleus. In the nucleus, the Smad complexes associate with other cofactors to regulate expression of target genes.

The kinase domains of TβRI and TβRII conform to the typical se-rine/threonine protein kinase structure, but they have been shown to phos-phorylate also tyrosine residues, and therefore, are dual-specificity kinases (36, 37). According to this specificity, TβRI can be autophosphorylated on its tyrosine residues, to which adapter protein ShcA recognizes and binds. ShcA gets phosphorylated by TβRI and recruits the Sos1/Grb2 complex to activate the mitogen-activated protein kinase (MAPK) kinase pathway (37).

Smad transducers The TGFβ family members transduce their signals through the Smad pro-teins which are divided into three groups: receptor-activated Smad (R-Smad), common Smad (co-Smad) and inhibitory Smad (I-Smad). TGFβ signaling leads to accumulation of activated Smad complexes in the nucleus,

18

where they regulate genes involved in a plethora of cellular responses such as cell growth, differentiation, migration and apoptosis, in a context- and cell type-dependent manner (38–40).

The R-Smads include Smad1, Smad2, Smad3, Smad5 and Smad8. They share a common partner or the co-Smad, Smad4. The I-Smads, Smad6 and Smad7, are themselves induced by TGFβ/BMP to inhibit TGFβ/BMP signal-ing pathways as a negative feedback mechanism. The TGFβ and activin pathways signal through Smad2 and Smad3, whereas BMP and GDF signal via Smad1, Smad5 and Smad8 (28). The activated type I receptor phospho-rylates the R-Smads at their C-terminal di-serine motifs and the activated R-Smads associate with Smad4; together, they translocate to the nucleus to regulate gene expression (Figure 2).

Smad structure The Smad structure consists of an N-terminal Mad homology 1 (MH1) do-main that is bridged to the C-terminal Mad homology 2 (MH2) domain by a linker region (Figure 3) (40).

Figure 3. Illustration of Smad protein structure with MH1 (purple), linker (beige) and MH2 (blue) domains. NLS, nuclear localization signal; β-hairpin, DNA-binding loop; NES, nuclear export signal; L3, L3 loop; SXS, motif phosphorylated by type I receptors.

The MH1 domain is highly conserved in all R-Smad and co-Smad proteins, but not in the I-Smads. I-Smad lacks the MH1 domain. A nuclear localiza-tion signal (NLS) and a β-hairpin loop that binds DNA are found within the MH1 domain (40). A common splice variant of Smad2 is unable to bind DNA due to a unique exon insertion in the DNA-binding domain, whereas a Smad2 variant lacking this insertion is able to bind DNA efficiently. Al-though R-Smads and co-Smad bind DNA via their MH1 domains, they do so with low affinities and therefore, cooperate with other TFs to strengthen the binding of the Smad complexes onto the DNA (41).

19

The MH2 domain is the most conserved domain among all the Smads and mediates many protein-protein interactions. Nuclear export signal (NES) motifs are found in the MH2 domains of R-Smad and I-Smad. In the case of Smad4, the NES motif is found in the linker domain (40). The specificity of the R-Smads for their respective receptors is defined by the L3 loops and basic pockets in the MH2 domains of the R-Smads and the L45 loops in the kinase domain of type I receptors (41). The type I receptors phosphorylate the C-terminal SXS motif of the R-Smads. The linker regions of the Smad proteins contain multiple serine and threonine residues that are phosphory-lated by MAPK, glycogen synthase kinase 3 (GSK3) and cyclin-dependent kinases (CDKs) (40), and proline-tyrosine motifs that are recognized by WW-domain-containing proteins, e.g. Smurf or NEDD4 (38, 42).

Gene regulation by Smad proteins The TGFβ family signaling pathways can have differential cellular effects depending on the cell type and cell context. TGFβ can inhibit cell growth yet also promote cell proliferation; it can promote differentiation or maintain stem cell pluripotency (38, 39). How the Smad proteins achieve specificity in mediating these versatile cellular responses is dependent on: (i) the type of ligand stimulus; (ii) organization of Smad complexes in association with specific DNA response elements; and (iii) interactions with TFs and cofac-tors that affect the chromatin structure or transcriptional machinery (39). Moreover, the desired signal intensity and duration is controlled by the activ-ity and turnover of Smad proteins through post-translational modifications, such as phosphorylation, ubiquitination, sumoylation and ADP-ribosylation (38, 42).

R-Smad and co-Smad complexes recognize and bind to specific DNA se-quences in the promoters or regulatory elements of their target genes (40), which has been further affirmed in recent genome-wide chromatin immuno-precipitation analyses (43). The Smad3-Smad4 complex binds selectively to the Smad-binding element (SBE) which has the consensus sequence 5’-GTCT-3’, or its complementary sequence 5’-AGAC-3’ (44–46). Some pro-moters lack the canonical SBE but contain GC-rich sequences which are preferentially recognized by Smad1 or Smad5 (47, 48).

Smad proteins cooperate with a wide repertoire of TFs, e.g. forkhead, homeobox and activator protein 1 (AP1) families, to selectively bind specific DNA sequences and regulate different sets of genes (40). Master transcrip-tional regulators (Oct4, MyoD and PU.1) recruit and direct Smad complexes to bind specific genomic sites to regulate gene programs during cell-lineage specification of embryonic stem cells and progenitor cells (49). During TGFβ-induced EMT, Smad complexes with c-Jun, JunB and Fra-1, members of the AP1 family, to induce genes related to invasiveness (50).

TGFβ initiates cell cycle arrest by first repressing c-Myc and then induc-ing the CDK inhibitors, p21Cip1 and p15Ink4B (51–54). Smad3-Smad4 com-

20

plexes cooperate with E2F4/5 to repress the c-myc promoter at the TGFβ-inhibitory element (TIE) site (55). In addition, Smad3-Smad4 complexes interact with Sp1 and FoxO to bind the proximal and distal regions of the p21Cip1 promoter, respectively (51, 52). This suggests that multiple, differen-tial Smad-partner complexes could occur on a single gene. Furthermore, Smad-FoxO complexes, as a common set of TF partners, could coordinate a synexpression group of 11 genes that partake in the regulation of growth inhibitory, stress- and adaptive-responses to TGFβ signals (54). Activated Smads can induce one gene and later cooperate with their gene product to regulate another gene, which is described as a ‘self-enabling’ mechanism (40). For example, ATF3 is a repressor induced upon TGFβ signaling and then interacts with Smad3 to repress Id1 transcription (56).

In addition to TFs, Smad proteins interact with cofactors that are part of the transcriptional machinery (40). Coactivators like p300, CBP and the Me-diator complex bring the Smad proteins into closer proximity to the RNA polymerase II complex. As p300 and CBP also possess HAT activity, they acetylate Smad2 and Smad3 and enhance the transcriptional activities of the Smad proteins (57). Transcriptional repression by Smad proteins is mediated through recruitment of HDAC to the repressor complex. For example, TGFβ-activated Smad3 binds HDAC4/5 to inhibit Runx2-dependent tran-scription of genes involved in osteoblast differentiation (58). Smad proteins could also interact with various corepressors, c-Ski and SnoN being the best studied, which repress Smad-induced transcription as a negative feedback mechanism (40).

Non-Smad signaling TGFβ also activates other pathways apart from Smads, e.g. MAPK, Rho GTPases and phosphotidylinositol-3’ kinase (PI3K)/Akt (59). MAPKs relay TGFβ signals independently of Smad activation and they modulate Smad activities by phosphorylating the linker region of Smad proteins (42). TβRI with a L45 loop mutation renders it unable to bind and activate Smad pro-teins but its kinase activity is retained, enabling the activation of JNK and p38 MAPK pathways in response to TGFβ stimuli (60, 61). TβRI induces apoptosis by recruiting tumor necrosis factor α receptor-associated factor 6 (TRAF6) to ubiquitinate and activate TGFβ-activated kinase 1 (TAK1) to activate the p38 MAPK pathway (62). Furthermore, TRAF6 poly-ubiquitination of TβRI can lead to cleavage of the TβRI intracellular domain (ICD), which translocates into the nucleus to activate, e.g. Snail and MMP2 genes (63). Whether the TβRI ICD phosphorylates Smads in the nucleus remains unknown. As mentioned earlier, TβRI can phosphorylate ShcA and activate the ERK MAPK pathway, which regulates cell proliferation and migration (37). In a similar fashion, TβRII is phosphorylated by Src tyrosine

21

kinase and engages Grb2 and ShcA to promote activation of the p38 MAPK pathway, thereby promoting proliferation and invasion of breast cancer cells (64).

TGFβ activates small Rho GTPases which regulate actin cytoskeletal dy-namics (65). RhoA is required for stress fiber formation and acquisition of mesenchymal properties during TGFβ-induced EMT (66). Conversely, TβRII phosphorylates the polarity protein Par6 and recruits Smurf1 to ubi-quitinate and degrade RhoA, thus leading to the dissolution of tight junctions (67). Activation of mammalian target of rapamycin complexes (mTORC), via the PI3K/Akt pathway, by TGFβ leads to increased protein synthesis, increased cell size, EMT and invasiveness (68, 69).

TGFβ in cancer TGFβ signaling is important in metazoan biology and any aberrations in the pathways often have deleterious consequences, such as cancer development (31). TGFβ functions as a tumor suppressor because it is able to inhibit cell proliferation during early tumorigenesis (38, 39). Some cancers counteract this by inactivating the TGFβ receptors and Smad proteins. Smad4 is often deleted in pancreatic cancers (31).

However, at later stages of tumor progression, TGFβ plays a turnabout role as a tumor promoter. Tumor cells become resistant to the cytostatic and apoptotic effects by TGFβ, but instead hijack TGFβ signaling pathways and use them to their advantage to evade immune surveillance, promote produc-tion of growth factors and undergo EMT, which is associated to invasion and metastasis (30, 31). EMT is linked to endowing tumor cells with ‘stem cell’-like properties, i.e. cancer stem cells (CSCs) (70, 71). CSCs are a small pop-ulation of cells that are capable of self-renewal, resistant to conventional therapeutics, and potentially seeding new tumors (72). TGFβ enriches the formation of CSCs in mammary epithelial cells undergoing EMT and in so-called claudinlow breast cancer cells (70, 71, 73). EMT induction by TGFβ will be discussed later.

Finally, TGFβ is implicated in metastasis, a complex, multistep progres-sion of malignant cancer cells entering the body circulatory system and forming tumors at secondary sites, which is often the main cause of death (32). Several pro-metastatic genes have been identified that predispose tu-mor cells to metastasize towards specific organs (74). TGFβ upregulates parathyroid hormone-related protein (PTHrP), interleukin 11 (IL-11) and connective tissue growth factor (CTGF) in breast cancer cells to promote bone metastases (75, 76). Angiopoeitin-like 4 (ANGTL4), another direct target of TGFβ, is upregulated in breast cancer cells and favors lung metasta-sis. ANGTL4 allows cancer cells to extravasate into the lung parenchyma by disrupting vascular endothelial cell-cell junctions (77). In cooperation with

22

Ras and mutant p53, TGFβ antagonizes the ability of p63 to protect against metastasis (78). TGFβ upregulates miRNAs that promote metastatic poten-tial, or metastamirs, e.g. miR-181a. miR-181a targets and degrades the pro-apoptotic protein Bim, thereby resisting anoikis; and high miR-181a expres-sion correlates with poor overall survival in cancer patients (79).

23

High mobility group A proteins

The chromatin is a highly ordered structure that controls the accessibility of DNA to nuclear proteins regulating replication, transcription, recombination and repair. The non-histone high mobility group (HMG) chromatin proteins are one group of architects of chromatin structuring that were identified due to their high electrophoretic mobility in polyacrylamide gels. The HMGA subfamily consists of two members encoded by different genes, Hmga1 and Hmga2 (80). HMGA2 will be the main focus of this thesis.

The structure of HMGA protein is characterized by three DNA-binding domains called ‘AT-hooks’ because they recognize and bind to AT-rich se-quences in the minor groove of DNA. The C-terminal is rich in acidic amino acids and mediates protein-protein interactions (Figure 4) (80). An NLS is found within the second AT-hook of the HMGA2 protein (81).

Figure 4. Illustration of a HMGA protein structure with three ‘AT’-hooks which bind DNA, and an acidic domain at the C-terminal.

HMGA2 expression and its regulation HMGA proteins are predominantly expressed during embryogenesis and hardly expressed in differentiated cells or normal adult tissues (82, 83). HMGA2 is also expressed in human embryonic stem cells (84). Overexpres-sion of HMGA2 by amplification or translocation is associated with benign and malignant neoplasias (85).

HMGA2 plays a role in cell proliferation, differentiation and self-renewal. It affects the cell cycle by activating the transcription of Cyclin A and Cyclin B2 (86, 87). HMGA2 relieves E2F1 repression by retinoblastoma protein (pRb) and causes cells to progress into the S phase of the cell cycle (88). The tumor suppressors p16INK4a and p19ARF, which cause cell cycle arrest, have been reported to be inhibited by HMGA2 (89). Hmga2 knockout mice dis-played a pygmy phenotype with reduced birth weight, craniofacial defects

24

and the adults were 40% of the size of wildtype littermates (82). Hmga2-/- mice were shown to have lesser fat cells than wildtype mice due to defective adipocyte proliferation, implying a role for HMGA2 in adipogenesis (90). In contrast, transgenic mice overexpressing Hmga2 exhibit gigantism (91). Coincidently, HMGA2 has been identified in genome-wide association stu-dies as a gene associated to human height stature (92).

During the early stage of heart development, HMGA2 cooperates with Smad proteins to positively regulate the Nkx2.5 gene, which encodes a tran-scription factor that regulates cardiomyocyte differentiation (93). HMGA2 has been implicated in neuronal stem cell (NSC) development. Hmga2 tran-scripts were found to be highly expressed in NSCs from fetal and young mice but its expression decreased with aging. While HMGA2 is not required for NSC lineage differentiation, it is required for maintenance of self-renewal ability in young mice (89). Early-stage NSCs have higher potential for proliferation and self-renewal and this neurogenic potential decreases over the course of brain development. HMGA2 expression is important for neurogenic potential in early-stage NSCs and forced expression of HMGA2 in late-stage NSCs can block astrocyte differentiation (94).

HMGA2 was found as a TGFβ target in a transcriptomic screen; it is upregulated in a Smad-dependent manner and promotes induction of EMT (95). Wnt10b, which is involved in mammary gland development, regulates HMGA2 expression via the Wnt/β-catenin signaling pathway. HMGA2 le-vels were diminished in embryonic mammary cells from Wnt10b knockout mice compared to the wildtype mice (96). HMGA2 is also regulated post-transcriptionally by the let-7 miRNA family. Let-7 miRNA represses HMGA2 expression by targeting the 3’UTR of Hmga2 mRNA for degrada-tion (97, 98). The reciprocal expression of let-7 and HMGA2 is also ob-served during NSC development; let-7 expression pattern is low in fetal NSC but high in adult NSC (89). Similarly, the inverse Let-7 and HMGA2 rela-tionship, that is regulated by Lin28b, is also observed in fetal and adult he-matopoietic stem cells (HSCs) (99).

HMGA2 mode of action HMGA proteins do not have transcriptional activity per se. They bind and cause conformational changes to DNA by bending, straightening, and un-winding mechanisms, and this facilitates the binding of TFs to DNA (Figure 5). HMGA proteins interact with their protein partners which result in in-creased or decreased binding affinities for DNA (80). HMGA2 binds directly to the Snail1 and Twist1 promoters, and possibly induces conformation changes for the basic transcriptional machinery to bind favorably and acti-vate transcription (100, 101). On the other hand, HMGA2-induced DNA bending prevents transcriptional complexes from binding to the ERCC1

25

gene, and hence, inhibits the expression of ERCC1, which is involved in the DNA repair process (102). Additionally, HMGA proteins recognize and bind specific chromatin structures in a sequence-independent manner (103). They bind to nucleosomes (84, 103) and can also displace histone H1 proteins (94, 104, 105), thereby affecting chromatin conformation (Figure 5a).

HMGA proteins are involved in many protein-protein interactions (106). HMGA2 interacts with Smad proteins to activate transcription of Nkx2.5 and Snail1 genes (93, 100). The assembly of interferon-β enhanceosome by HMGA proteins is a good example of how HMGA binds and bends the DNA sharply so that TFs (ATF2, c-Jun, IRF and NF-κB) are brought into good proximity of each other for protein-protein interactions and transcrip-tional initiation (Figure 5b) (107, 108).

Figure 5. Modulation of DNA by HMGA proteins. (a) HMGA extends the chroma-tin from a ‘closed’ to ‘open’ conformation. (b) HMGA bends DNA to allow tran-scription factors to bind. HMGA proteins can also interact and affect binding affini-ties of transcription factors for DNA. TF, transcription factor.

As mentioned earlier, HMGA2 promotes cell cycle progression into the S phase by alleviating E2F1 activity. The pRb protein is an important regulator of the cell cycle, as it binds and inactivates E2F1 to prevent inappropriate cell division. HMGA2 binds to pRb without disrupting the pRb-E2F1 com-plex but instead displaces HDAC from the pRb-E2F1 complex to enhance E2F1 transcriptional activity (88). Similarly, HMGA2 displaces HDAC from the hTERT promoter and interacts with Sp1 to drive hTERT expression (109). In addition to directly binding the Cyclin A promoter, HMGA2 also interferes with the binding of the repressor p120E4F to the promoter (86). Through these protein-DNA and protein-protein interactions, HMGA pro-teins participate in a plethora of biological processes (106).

26

HMGA2 in cancers HMGA2 overexpression has been found in several cancers and is often asso-ciated with poor prognosis and metastasis (85). Its role in gene regulation, i.e. cell cycle genes, further exemplifies its ability to promote cell prolifera-tion and tumor growth. HMGA2 regulates Cyclin A and Cyclin B2 expres-sion and knockdown of HMGA2 in breast cancer cells reduced proliferation rates (86, 87, 96). In a transgenic mouse model, HMGA2 induces pituitary adenomas by enhancing E2F1 activity (88). HMGA2 indirectly inhibits p16INK4a and p19ARF in NSCs, which cause cell cycle arrest (89). HMGA2 also promotes neoplastic transformation of rat thyroid cells by activating JunB and Fra1 of the AP1 family of transcription factors (110). High expres-sion of HMGA2 in stroma cells modulates the tumor microenvironment by stimulating paracrine Wnt signaling in neighboring prostate epithelia, thus promoting prostate adenoma formation (111).

EMT reprograms epithelial cells to become mesenchymal-like with gain of invasive properties (29, 112). HMGA2 plays a central role in TGFβ-induced EMT by regulating the expression of EMT-TFs, such as members of Snail, ZEB and Twist families (95, 100, 101). Furthermore, HMGA2 pro-motes metastasis of breast cancer cells in mouse xenograft models (113, 114).

The tumorigenicity effects by HMGA2 can be suppressed by the let-7 miRNA (97, 98). Downregulation of let-7 is often correlated with HMGA2 overexpression, invasion and metastasis (115, 116). Forced expression of let-7 reduced self-renewal ability and inhibited tumorigenesis in breast CSCs (115). However, HMGA2 silencing in the same breast CSCs had no adverse effects on mammosphere formation but enhanced differentiation of CSCs (115). The striking similarities between the roles of HMGA2 in NSC and CSC, suggest that HMGA2 is not required to confer self-renewal properties, but instead, is crucial for the maintenance of self-renewing potential and the retention of cells in the undifferentiated state (89, 94, 115).

27

Epithelial-mesenchymal transition

During morphogenesis, epithelial-mesenchymal transition (EMT) is vital for the formation of mesenchymal cells, tissues or organs. It is a highly dynamic process whereby polarized epithelial cells change their morphology into motile mesenchymal cells. EMT involves the disassembly of cell-cell adhe-sion molecules, upregulation of mesenchymal proteins and remodeling of the cytoskeleton and ECM (Figure 6). The reverse process, mesenchymal-epithelial transition (MET), occurs after cells migrate and home into new sites for tissue formation. An eminent EMT feature is the loss of E-cadherin, a cell-cell adhesion molecule that is important for the maintenance of epi-thelial architecture. Concurrently, cells upregulate mesenchymal markers such as N-cadherin and vimentin to acquire migratory and invasive proper-ties (29).

Figure 6. Epithelial-mesenchymal transition (EMT). An epithelial cell undergoing EMT (outlined in red), dissolves cell-cell junctions and acquires mesenchymal prop-erties for migration and invasion (mesenchymal cell in red). EMT-TFs (purple text) induce EMT by downregulating epithelial (green text) and upregulating mesen-chymal genes (red text).

E-cadherin is a transmembrane glycoprotein with a Ca2+-binding extracellu-lar domain that forms homophilic interactions with adjacent cells. The intra-cellular domain complexes with catenins, linking it to the actin cytoskeleton network and further strengthening adhesions with neighboring cells (117).

28

Downregulation of E-cadherin during EMT represents a determining step to destabilize the epithelial architecture by reducing cell-cell contacts and reor-ganizing the cytoskeleton, so a single cell may readily break away from its neighbor.

EMT in cancer progression can confer tumor cells with motility, survival, drug resistance and stem cell-like properties. Such EMT features provide advantages for invasion, dissemination and metastasis. These tumor cells dissociate from neighboring cells, invade the basement membrane and the walls of neighboring blood vessels into the blood stream where they have to survive and disseminate. Upon reaching the secondary site, they extravasate into the tissues and undergo MET to establish secondary metastases (29, 30).

EMT in embryogenesis Gastrulation occurs to form the ectoderm, mesoderm and endoderm. Epi-thelial cells undergo EMT in order to ingress the primitive streak and form the mesoderm (29). After gastrulation, neural crest cells between the neural plate and non-neural ectoderm, via the EMT program, delaminate from the neural epithelia. The newly formed mesenchymal cells can migrate over long distances to new tissue areas to differentiate into various cell types (118). Cardiac valve development in the chick involves TGFβ signaling during EMT. TGFβ2 is required to initiate EMT, but it is TGFβ3 that promotes invasion of endocardial cells into the cardiac cushion to establish the septa and valves (119). Moreover, TGFβ3 is required for EMT-formation of the palate in mice, as TGFβ knockout mice display a cleft palate phenotype (120).

TGFβ induction of EMT EMT is promoted by many signaling pathways, e.g. TGFβ, Wnt, Notch, receptor tyrosine kinases (RTK) and integrins, that often crosstalk with each other to orchestrate elaborate gene programs and protein networks to remo-del the cell architecture and function. These signaling pathways, including microenvironmental factors like hypoxia or inflammation, activate effector proteins such as MAPK, PI3K, NFκB, integrin-linked kinase (ILK), Src and Rho GTPases, to regulate EMT (121).

TGFβ is a major inducer of EMT in embryogenesis and cancer progres-sion (30). Transgenic mice with TGFβ1 overexpression in keratinocytes exposed to a chemical carcinogen led to higher incidences of developing spindle cell carcinomas, which were highly invasive and produced endogen-ous TGFβ3. Control mice developed benign skin tumors when exposed to the same carcinogen. This suggests that TGFβ1 can drive EMT in vivo (122).

29

In vitro studies of EMT showed that TGFβ downregulates epithelial markers, like E-cadherin and cytokeratins, and upregulates mesenchymal markers, including N-cadherin, fibronectin, vimentin and α-smooth muscle actin (30). TGFβ mediates EMT through activation of Smad2 and Smad3 (123, 124). Furthermore, TGFβ regulates HMGA2 via the Smad pathway to elicit EMT (95). HMGA2, in turn, cooperates with Smads to upregulate TFs involved in inducing EMT, such as Snai1l and Twist1 (100, 101). Activated Smads can also interact with Snail1 to repress E-cadherin expression (125). Platelet-derived growth factor (PDGF) ligands and receptors are induced by TGFβ to promote EMT, invasion and metastasis in hepatocellular carcinoma (126). In addition, TGFβ integrates with the Notch signaling pathway by inducing Notch ligand Jagged1 and target gene Hey1 in a Smad3-dependent manner, at the onset of EMT (127). TGFβ also represses the inhibitor of differentia-tion proteins, Id2 and Id3, during EMT (95). Id proteins antagonize basic helix-loop-helix (bHLH) proteins, such as Twist1 which is a TF that induces EMT. Id2 and Id3 are also direct gene targets of BMP7 signaling pathway, which could explain how BMP7 can inhibit TGFβ-induced EMT and cause MET (128).

TGFβ plays a role to disrupt cell polarity during EMT. TβRI is found in complex with Par6 and occludin at epithelial tight junctions. Phosphoryla-tion of Par6 by TβRII leads to RhoA degradation and dissolution of tight junctions in a TGFβ-dependent manner (67). In addition, TGFβ negatively regulates RhoA through another mechanism by inducing miR-155 expres-sion to disrupt tight junction assembly (129). Par3, which is part of the Par6/aPKC complex that maintains epithelial apico-basal polarity, was shown to be downregulated by TGFβ in kidney epithelial cells (130).

RhoA activation by TGFβ is required for stress fiber formation during EMT (66). Recently, mTORC2 was shown to be activated by TGFβ to in-duce EMT and contribute to cell migration and invasion. The mTORC2 is also required for TGFβ-activation of RhoA (69). The disparity between RhoA activation (66) and degradation (67) during TGFβ-induced EMT re-mains unclear. Although TGFβ activates mTORC1 as well, this does not trigger EMT but instead leads to increase in cell size and enhances invasion (68). The mechanisms of TGFβ signaling that drive EMT, whether through Smad or in coordination with other pathways, are diverse and not yet fully understood.

Transcriptional inducers of EMT The signaling pathways that trigger EMT often cascade down to a common plane, that is, regulation of transcription factors which induce the EMT pro-gram. These EMT-inducing transcription factors (EMT-TF) coordinately exert transcriptional control over EMT by regulating epithelial and mesen-

30

chymal genes (112). The most heavily studied EMT-TFs are members of the Snail, ZEB and Twist families, which converge to repress the expression of prototype epithelial marker, E-cadherin.

Snail1 (Snail) and Snail2 (Slug) are zinc finger TFs, composed of a highly conserved C-terminal region which contains four to six C2H2-type zinc fin-gers and a more divergent N-terminal region. The zinc fingers bind to DNA by recognizing E-box elements with the consensus CANNTG. The N-terminal contains a SNAG domain that mediates repressor function. Snail1 mutants in mice are embryonic lethal because they are unable to complete gastrulation (131). Snail1 binds directly to E-boxes of the E-cadherin pro-moter and recruits the transcriptional corepressor Sin3A and HDAC1/2 complexes (132). Snail2, ZEB1 and ZEB2 use similar mechanisms to repress the E-cadherin gene, whereas Twist1 uses an indirect mechanism (133).

The ZEB family consists of ZEB1 (δEF1) and ZEB2 (SIP1). They have two zinc finger clusters at each end, a middle homeodomain that does not bind DNA, a Smad binding domain and a CtBP interaction domain. ZEB1- or ZEB2-knockouts in mice are not viable, which indicate that the two ZEBs are not able to compensate for each other’s function despite being structural-ly similar (134). In addition to repressing E-cadherin, ZEB2 represses other epithelial molecules: tight junction, adherens junction, desmosome and gap junction genes to promote invasion (135, 136). ZEB1 represses the Lgl2 gene, a component of the SCRIB basolateral polarity complex, and its loss is associated with metastasis (137).

Twist1 belongs to the bHLH family of transcription factors. Its structure contains a basic region and a helix-loop-helix domain for dimerization. Twist1-deficient mice die shortly after gastrulation due to defects in neural tube closure and malformation of the cranium and limbs (138). Twist1 does not repress the E-cadherin promoter directly but cooperates with other TFs or repressors, such as Snail2 (139), Bmi (140) and SET8 (KDM5A) (141). Twist1 positively regulates the expression of N-cadherin and Akt2 genes, promoting migration and invasion (142, 143). The requirement of Twist1 for invasion and lung metastasis has been shown by knocking down Twist1 in a mouse mammary tumor model. Silencing of Twist1 had no effect on forma-tion of primary tumors compared with control mice, but lung metastases were drastically reduced (144). Twist1 upregulates PDGF receptor α (PDGFRα), which participates in invadopodia formation that facilitates me-tastasis. High expression of Twist1 and PDGFRα correlated with poor sur-vival outcome (145).

EMT-TFs could positively regulate and cooperate with each other. Snail1 and Twist1 co-regulate Snail2 and ZEBs (101, 146). Several studies have suggested an interplay of EMT-TFs during tumor progression. Snail1 is upregulated at the onset of EMT and leads to the initiation of the invasive process, whereas the other EMT-TFs are subsequently induced to promote and maintain the migratory and invasive phenotype (133).

31

Epigenetic regulation in EMT EMT-TFs drive many extensive and distinct gene programs during EMT: gene programs to downregulate epithelial genes, upregulate mesenchymal genes, induce anti-apoptotic genes, enrich for CSCs, etc. Such reprogram-ming events involve epigenetic rewiring for the alteration of the chromatin landscape to impose transcriptional changes. Genome-wide histone modifi-cation changes during TGFβ-induced EMT were observed in mouse hepato-cytes, whereby H3K9me2 (repressive mark) was reduced and H3K4me3 (active mark) was increased within the large, organized heterochromatin K9 modifications (LOCK) regions, while DNA methylation remained unaf-fected (147). Similarly, mouse mammary cells undergoing EMT exhibit a genome-wide redistribution of H3K27me3 repressive marks, which is me-diated by histone methyltransferase EZH2. EZH2 was shown to be regulated by another TF, Sox4, and knockdown of EZH2 affected EMT and reduced metastasis in mouse models (148)

However, de novo DNA methylation of several promoters, including E-cadherin, were found to occur in immortalized human mammary epithelial cells undergoing EMT (149). Furthermore, E-cadherin repression during EMT in breast cancer cell lines were associated to promoter hypermethyla-tion, rather than mutations (150). TGFβ/Smad signaling is required to main-tain DNA methylation patterns of epithelial genes as knockdown of Smad2 reverted the methylated status and blocked EMT (151).

EMT-TFs themselves interact with many epigenetic regulators and some examples have been mentioned in the previous section. Snail1 forms a com-plex with LSD1 to demethylate H3K4me3 and thereby removes the chroma-tin ‘active’ status (152). Snail1 recruits histone modifying enzymes to induce a repressive state at the E-cadherin gene, e.g. HDAC complexes (132). G9a and SUV39H1, which di- and tri-methylate H3K9, respectively, were shown to be required for Snail-mediated repression of E-cadherin (153, 154). In addition, Snail1 forms a complex between G9a and DNMT1 to modulate DNA methylation of E-cadherin promoter (153).

The EMT program is also regulated by miRNA networks that target EMT-TFs, as well as epithelial and mesenchymal markers (155). The miR-200 family (miR-200a/b/c, miR-141 and miR-429) and miR-205, which negatively regulates ZEB1 and ZEB2, are downregulated during TGFβ-induced EMT. Forced expression of miR-200s in mesenchymal cells led to high E-cadherin levels and MET (156, 157). In a feedback loop, ZEB1 binds and represses the promoters of miR-200c and miR-141 to reinforce EMT (158). Snail1 and miR-34 are another example of a double feedback loop: Snail1 is repressed by miR-34 and vice versa (159). Twist1 can induce tran-scription of miR-10b which inhibits tumor suppressor HoxD10, and as a result, small GTPase RhoC is no longer sequestered by HoxD10 and able to promote metastasis (160).

32

Attributes of EMT EMT contributes to tumor progression by facilitating invasion, intravasation and metastasis. Epithelial cells that have undergone EMT detach from the epithelium and acquire properties associated with cell survival, i.e. resistance to apoptosis, anoikis, senescence, chemotherapeutics and immune surveil-lance; and gain of stem cell-like features (29).

Snail1 decreases cell proliferation to favor invasion over growth and also promotes cell survival (161). ZEB1 and Twist1 abrogate oncogene-induced premature senescence (162, 163). Twist1 also inhibits apoptosis by downre-gulating the ARF-p53 pathway, as well as confers resistance to chemothera-peutic agents (138).

CSCs are a small subset of cells found within the heterogeneities of tu-mors, which have self-renewing ability and are highly tumorigenic (72). Ectopic expression of Snail1 or Twist1 in human mammary epithelial cells caused cells to undergo EMT and acquire CSC characteristics, which are defined by the expression of CD44high/CD24low surface markers and in-creased ability to generate mammospheres (70, 71). In addition, Twist1 can induce the transcription of Bmi, which promotes stem cell self-renewal (140).

33

Present Investigation

TGFβ is a potent inducer of EMT and Thuault et al. have shown for the first time that HMGA2 is required in this process (95). Our aim was to delve into the molecular mechanisms by which HMGA2 elicits EMT.

Paper I: HMGA2 and Smads co-regulate Snail1 HMGA2 and Smads co-regulate Snail1 expression during induction of epi-thelial-to-mesenchymal transition. TGFβ is a major inducer of EMT and requires HMGA2 to drive the EMT process (95). TGFβ/Smad signaling activates the transcription of the Hmga2 gene which subsequently leads to the induction of other EMT-TFs, such as members of the Snail, ZEB and Twist families, to regulate EMT.

Here, we showed that Snail1 is robustly upregulated in response to TGFβ, and HMGA2 is required for full induction of Snail1 expression. We then analyzed the regulation of Snai1l gene expression by TGFβ and HMGA2 and found that heteromeric Smad3/Smad4 and HMGA2 can synergistically activate the Snail1 promoter. In-depth analysis of the Snail1 promoter dem-onstrated that HMGA2 and Smads bind to the proximal region on their own and their binding to the promoter is enhanced when both factors are present. Furthermore, HMGA2 can interact with Smad2, Smad3 and Smad4, which explains the cooperation between HMGA2 and Smads in regulating Snail1 expression.

We also found that Snail1 acts as a major effector downstream of HMGA2 for the induction of EMT, as Snail1 depletion partially reversed the mesenchymal phenotype associated with ectopic HMGA2 expression. In addition, we found that the effects of Snail1 depletion are extended to the concomitant decreased expression of other specific E-cadherin repressors like Snail2, ZEB1 and ZEB2. This suggests that TGFβ signaling induces and uses HMGA2 as an integrator to coordinate a hierarchical network of tran-scription factors to orchestrate the EMT program.

34

Paper II: HMGA2 regulates Twist1 during EMT Regulation of transcription factor Twist expression by the DNA architectural protein high mobility group A2 during epithelial-to-mesenchymal transition. In Paper I, we have shown that ectopic expression of HMGA2 in NMuMG epithelial cells leads to a constitutive EMT phenotype with high mRNA ex-pression levels of Snail1, Snail2, ZEB1, ZEB2 and Twist1. Knockdown of Snail1 in HMGA2-overexpressing mesenchymal cells led to an incomplete epithelial morphology, with a concomitant decrease in the levels of Snail2, ZEB1 and ZEB2. However, Twist1 levels remained unaffected which proba-bly explains the partial epithelial reversion (100).

We were interested to know if depletion of both Snail1 and Twist1 could reverse HMGA2-mesenchymal cells back to an epithelial phenotype. RNAi experiments of Snail1 and Twist1 double-knockdowns in HMGA2-mesenchymal cells had an epithelial-like morphology as re-organization of the epithelial marker ZO-1 at the tight junctions was more distinctive, com-pared to knockdown of Snail1 alone in the same mesenchymal cells. Despite Snail1 and Twist1 being depleted, E-cadherin continued to be suppressed by HMGA2 via unknown mechanisms.

As HMGA2 induces Twist1 expression, we examined if HMGA2 can ac-tivate the Twist1 promoter. Indeed, HMGA2 directly bound and robustly activated the Twist1 promoter. Mapping studies of the Twist1 promoter showed HMGA2 binding near the transcription start site which contains a short T-rich motif. Mutations in this T-rich motif abolished the binding of HMGA2. In summary, HMGA2 can induce multiple transcriptional regula-tors of the EMT program and acts as a master integrator of mesenchymal differentiation.

Paper III: HMGA2 epigenetic regulation of E-cadherin The high mobility group A2 protein epigenetically silences the Cdh1 gene during epithelial-to-mesenchymal transition. The loss of the tumor suppressor E-cadherin (Cdh1) is a key event during tumorigenesis and EMT (29, 112). We showed in Paper II that knockdown of both Snail1 and Twist1 in HMGA2-overexpressing mesenchymal cells led to reassembly of the tight junctions, but E-cadherin continued to be re-pressed (101).

We hypothesized that HMGA2, as a chromatin-binding protein, has a role in epigenetically silencing the Cdh1 gene. We found that, in HMGA2-overexpressing cells, the Cdh1 promoter was DNA-methylated and con-tained repressive histone modification marks, e.g. tri-methylation of lysine 9

35

and 27 residues on histone H3. Chromatin immunoprecipitation (ChIP) ana-lyses showed that both HMGA2 and DNMT3A, a DNA-methylating en-zyme, occupied the proximal region of the Cdh1 promoter. Treatment of HMGA2-overexpressing cells with a DNA demethylating drug, 5-aza-2’-deoxycytidine (AZA), restored the expression of E-cadherin by blocking the binding of HMGA2 and DNMT3A to its promoter.

Using an online gene expression analysis tool, we found an inverse corre-lation between HMGA2 and E-cadherin expression in breast cancer cells. Migration and invasion rates were greatly reduced when HMGA2 was knocked down in the invasive, MDA-MB-231 breast cancer cells. Here, we described a new epigenetic role for HMGA2 in the long-term silencing of the E-cadherin gene, as part of the EMT program and consequently facilitat-ing migration and invasion.

36

Concluding remarks

EMT is a complex process involving multiple gene expression programs that advocate tumor progression, invasion and metastasis. Many growth factors, TGFβ in particular, and the microenvironment promote EMT by transcrip-tionally and epigenetically regulating the epithelial and mesenchymal genes involved. There is an abundance of literature characterizing EMT-TFs in orchestrating EMT, and in recent years, there has been a shift in focus to-wards the epigenetic control of EMT. This is unsurprising as gene regulation by TFs and epigenetic modifiers are tightly inter-linked, after all, TFs are dependent on the chromatin milieu to be able to access and bind DNA (5).

The non-histone chromatin protein HMGA2 is the heavy weight of this thesis - the regulation of EMT-TFs (Paper I and II) and alteration of the chromatin landscape for long-term suppression of the E-cadherin gene (Pa-per III) to promote EMT. HMGA2 regulates a hierarchical network of EMT-TFs, namely Snail1, Snail2, ZEB1, ZEB2 and Twist1. Knockdown of Snail1 and/or Twist1 in HMGA2-overexpressing cells was insufficient to fully re-vert to MET as E-cadherin remained repressed. One reason is that HMGA2 was still highly expressed and may regulate another EMT-TF as a compensa-tory mechanism. Secondly, the knockdown efficiency was not 100%. On the other hand, how desirable is it to achieve a ‘MET’ state? There has been growing evidence that MET promotes metastatic colonization and formation of secondary tumors (164).

EMT-TFs like Snail1 and Twist1 endow tumor cells with self-renewal properties, thereby giving rise to CSCs (70, 71). Snail1 and Twist1 are regu-lated by HMGA2 (Paper I and II), and HMGA2 is required for the self-renewal potential in fetal NSC and HSC (89, 94, 99). Overexpression of the HMGA2 antagonist, let-7 miRNA, reduced self-renewal ability in NSC and CSC; while downregulation of let-7 stabilized HMGA2 expression and en-hanced self-renewal ability (89, 115). Recently, this self-renewal ability of CSC, NSC or HSC, points to a central role of Lin28, which inhibits let-7 miRNA and de-represses HMGA2, creating a Lin28/let-7/HMGA2 axis (12, 94, 99, 115). Thus, whether HMGA2 is able to generate CSCs, directly or through EMT-TFs, remains to be elucidated.

Overexpression of HMGA2 in NMuMG epithelial cells led to increased migration and invasion rates; while knockdown of endogenous HMGA2 in metastatic MDA-MB-231 breast cancer cells blocked migration and inva-sion. E-cadherin is a well-known suppressor of invasion (117) and its ex-

37

pression is often inversely correlated with HMGA2 (113, 114), which is also reflected in our studies here. This suggests that HMGA2 is involved in mi-gration and invasion, which explains the metastatic ability of HMGA2 in breast cancer cells (113, 114).

In Paper III, we showed that HMGA2 has an epigenetic role in the long-term silencing of the E-cadherin gene and this involves DNMT3A which is upregulated by HMGA2. We have yet to elucidate how DNMT3A is re-cruited, and if DNMT1 and DNMT3B contribute to the hypermethylation of the E-cadherin promoter. CTCF is a chromatin insulator protein that partici-pates in many DNA processes and is unable to bind methylated DNA re-gions (9). Witcher and Emerson have shown that CTCF creates protective boundaries around tumor suppressor genes, including E-cadherin. The loss of these boundaries is associated with the repression of tumor suppressor genes in cancer cells. (165). We showed that high HMGA2 expression deters CTCF binding to the E-cadherin promoter by unknown mechanism, and disrupts this protective boundary. Does HMGA2 modulate the chromatin architecture such that CTCF is unable to bind and DNA methylation by DNMT ensues, or does HMGA2 recruit the epigenetic machinery to first silence the gene and the resultant structure is unfavorable for CTCF binding? This chicken-and-egg question on what happens first, is difficult to resolve. It would be interesting to know if HMGA2 epigenetically regulates other epithelial genes, e.g. tight junction genes, in a similar fashion.

Recently, it has been shown that HMGA2 regulates another epigenetic factor during breast cancer tumorigenesis. HMGA2 downregulates TET1, a methylcytosine dioxygenase which catalyzes the demethylation of DNA. Depletion of HMGA2 led to TET1 upregulation and TET1 demethylates its target genes as well its own promoter for further robust induction to suppress tumor growth and metastasis (113). This suggests a link between HMGA2 and DNA methylation events, i.e. HMGA2 could induce DNMT3A and re-press TET1 to prevent demethylation.

It would be interesting to know which other EMT-TF or epigenetic-regulator genes are under the regulation of HMGA2. Transcriptomic micro-arrays have been performed to identify HMGA2 target genes in colon and gastric cancer cells (166, 167). However, this does not provide information on chromatin regulation. A ChIP-sequencing profile would provide a better understanding of genome-wide regulation of genes and chromatin by HMGA2. Such a data set can be used in compar-ative analyses with other epigenetic marks, such as histone modifications, DNMTs and CTCF binding patterns, and nucleosome positioning, some of which are already annotated in the EN-CODE repository.

38

Acknowledgements

The work in this thesis has been undertaken at the Ludwig Institute for Can-cer Research Ltd. Chasing after the Philosophiae Doctor Degree has gone beyond the scope of science; the road has been fun and I have met many good people from all over the world. I draw my lessons and strength from you and your rich life experiences. I like to express my heart-felt gratitude here. The list is long and I apologize if I’d neglected to mention someone here. Nevertheless, I thank you all for your support and encouragement. Boss, Aristeidis Moustakas. You are one of the coolest superiors I have ever come across. You’ve been extremely supportive in my scientific develop-ment and also of my family life. I have learnt and gained a lot under your ever-optimistic leadership. Thank you for giving me so many opportunities and introducing many proteins to me.

Big Boss and co-supervisor, Carl-Henrik (Calle) Heldin. Despite your hectic schedule, you always take a keen interest and time in all our projects and progress. I appreciate that a lot. Thank you for your kind support.

Collaborators: Apple guru, Kaoru Kahata. You have been a tremendous help, especially in the last 2 years, otherwise I won’t have gotten so far. It was a privilege to work with you. Sylvie Thuault, thanks for getting me started and your continuous participation in our works. Laia Caja Puigsu-birà, thanks for your help. You also often perform a nonsense-knock-out and then a sense-knock-in to my head. Tea Carletti, you were a terrific student to work with. Stefan Termén, thanks for letting me help out. Additional thanks to external collaborators, Amparo Cano and Hector Peinado.

Green-fingers, caring Anita Morén: you’re a pillar in the lab and your presence never fails to give me assurance. Coffee junkies, Katia Savary and gorilLars van der Heide, caffeine OD makes us super giggly-silly; TGFβ enhances balls formation and caffeine sustains Smad signal hyperactivity. I like to thank the rest of the group for their help and discussions: Peter Lönn (Pete Mapletree), Michael Vanlandewijck, Markus Dahl, Erna Raja, Ulla Engström, Yukihide Wantanabe, Jon Carthy, Varun Maturi, Masoud Raz-mara, Panagiotis (Panos) Papoutsoglou, Constantine (Costas) Kolliopoulos, Andries Blokzijl; and Aris’ Angels: Mahsa Dadras, Kalliopi (Kallia) Tzav-laki and Claudia Bellomo.

I like to thank the ever-helpful Monday J.Clubbers, and group leaders and their lab members: Maréne Landström, Paraskevi (Evi) Heldin, Johan

39

Lennartson, Ingvar Ferby, Carina Hellberg, Johan Ericsson and Pontus Aspenström.

Doc. Bike, Ulf Hellman: changing bicycle tires for a person, it’s for one season; teach her to change tires, it’ll be for all seasons. You and Christian Schmees, thanks for speaking the same language - food. We had many nice lunches together and were the unofficial food critics of Uppsala restaurants. We could have written a “Ludwigen Guide vs. Michelin Guide”. Verena von Bülow, be it science, fashion or motherhood, I love hanging out with you. Elsa Papadimitriou and Jessica Rössler, even though you were here for a short time, I miss your friendship a lot. Eftheria Vasilaki (pronounced Ria vad-so-lucky), lively ChIP master and my runway model for Paul Frank, Desigual and Camper. My role model Anders Sundqvist, you showed me it’s all possible with family, sports and science. PCR Pro Masato Morikawa, your tips did the trick! Planning Pro Berit Olofsson and super happy Helena (Posh) Porsch, whether it’s the day or night shift, I enjoyed our blah-blahs. Merima Mehic saved my life with a coffee machine. Sushikompisar Car-men Herrera Hidalgo and Ana Rosa Saez Ibanez: Eat your sushi! In sync with Giulia Pilia, don’t we love our lucky culture hood! Linda Sooman and Takashi Terabayashi, thanks for the office banter. My favorite blonds, Inna Kozlova and Mariya Yakymovych, I love your good hearty laughs and op-timisms in life.

Charlotte (Lotti) Rosman, Aive Åhgren and Aino Ruusala, always having old-school lab tricks up their sleeves. Ulf (Uffe) Hedberg, computer and gadgets, dance away. Lars-Erik Hermansson, Lasse kan allt! Eva Hallin, Run on. Ingegärd Schiller, Christer Wernstedt, Gulbritt och Inger. You all have such a zest for life. Tack allihop för ert stöd och lunchsällskap.

Rachel Nong Yuan (农园), we came to Winterland about the same time and grew together. You’ve got humor and have been a tremendous support. Thanks for sticking around. Folks at home: Dad, E-Wen, E-Lis, Wai Leng and families@316, I am grateful to you for always keeping me in your thoughts. Elsa Ng, your handwritten letters and gifts always bring me com-fort. I also like to thank the Högberg families for their 3Cs: care, concern and cooking. My brats, Caelan and Travis, you’ve taught me one of the many important life lessons - the essence of time.

Lastly, my wonderful husband Daniel Högberg, I won’t have made it without you. Thanks for arranging your busy schedule to accommodate my fantastiskt cells’ ¤#(&*/ needs. You’ve always been so patient, encouraging and thoughtful so this wild horse can enjoy her food and Korean dramas.

~ May the SMADs be with you ~

40

References

1. International Human Genome Sequencing Consortium (2004) Finishing the euchromatic sequence of the human genome. Nature, 431, 931–45.

2. Dunham,I., Kundaje,A., Aldred,S.F., Collins,P.J., Davis,C.A., Doyle,F., Epstein,C.B., Frietze,S., Harrow,J., Kaul,R., et al. (2012) An integrated encyclopedia of DNA elements in the human genome. Nature, 489, 57–74.

3. Lee,T.I. and Young,R.A. (2013) Transcriptional regulation and its misregulation in disease. Cell, 152, 1237–51.

4. Maston,G.A., Evans,S.K. and Green,M.R. (2006) Transcriptional regulatory elements in the human genome. Annu. Rev. Genomics Hum. Genet., 7, 29–59.

5. Li,B., Carey,M. and Workman,J.L. (2007) The role of chromatin during transcription. Cell, 128, 707–19.

6. Malik,S. and Roeder,R.G. (2010) The metazoan Mediator co-activator complex as an integrative hub for transcriptional regulation. Nat. Rev. Genet., 11, 761–72.

7. Berger,S.L., Kouzarides,T., Shiekhattar,R. and Shilatifard,A. (2009) An operational definition of epigenetics. Genes Dev., 23, 781–3.

8. Cedar,H. and Bergman,Y. (2009) Linking DNA methylation and histone modification: patterns and paradigms. Nat. Rev. Genet., 10, 295–304.

9. Jones,P.A. (2012) Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat. Rev. Genet., 13, 484–92.

10. Baylin,S.B. and Jones,P.A. (2011) A decade of exploring the cancer epigenome - biological and translational implications. Nat. Rev. Cancer, 11, 726–34.

11. Dawson,M.A. and Kouzarides,T. (2012) Cancer epigenetics: from mechanism to therapy. Cell, 150, 12–27.

12. Suva,M.L., Riggi,N. and Bernstein,B.E. (2013) Epigenetic Reprogramming in Cancer. Science, 339, 1567–1570.

13. Kouzarides,T. (2007) Chromatin modifications and their function. Cell, 128, 693–705.

14. Black,J.C., Van Rechem,C. and Whetstine,J.R. (2012) Histone lysine methylation dynamics: establishment, regulation, and biological impact. Mol. Cell, 48, 491–507.

15. Beisel,C. and Paro,R. (2011) Silencing chromatin: comparing modes and mechanisms. Nat. Rev. Genet., 12, 123–35.

16. Smith,Z.D. and Meissner,A. (2013) DNA methylation: roles in mammalian development. Nat. Rev. Genet., 14, 204–20.

17. Jurkowska,R.Z., Jurkowski,T.P. and Jeltsch,A. (2011) Structure and function of mammalian DNA methyltransferases. Chembiochem., 12, 206–22.

18. Suetake,I., Shinozaki,F., Miyagawa,J., Takeshima,H. and Tajima,S. (2004) DNMT3L stimulates the DNA methylation activity of Dnmt3a and Dnmt3b through a direct interaction. J. Biol. Chem., 279, 27816–23.

41

19. Ooi,S.K.T., Qiu,C., Bernstein,E., Li,K., Jia,D., Yang,Z., Erdjument-Bromage,H., Tempst,P., Lin,S.-P., Allis,C.D., et al. (2007) DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature, 448, 714–7.

20. Sasai,N. and Defossez,P.-A. (2009) Many paths to one goal? The proteins that recognize methylated DNA in eukaryotes. Int. J. Dev. Biol., 53, 323–34.

21. Fuks,F., Burgers,W.A., Godin,N., Kasai,M. and Kouzarides,T. (2001) Dnmt3a binds deacetylases and is recruited by a sequence-specific repressor to silence transcription. EMBO J., 20, 2536–44.

22. Epsztejn-Litman,S., Feldman,N., Abu-Remaileh,M., Shufaro,Y., Gerson,A., Ueda,J., Deplus,R., Fuks,F., Shinkai,Y., Cedar,H., et al. (2008) De novo DNA methylation promoted by G9a prevents reprogramming of embryonically silenced genes. Nat. Struct. Mol. Biol., 15, 1176–83.

23. Fuks,F., Hurd,P.J., Deplus,R. and Kouzarides,T. (2003) The DNA methyltransferases associate with HP1 and the SUV39H1 histone methyltransferase. Nucleic Acids Res., 31, 2305–12.

24. Viré,E., Brenner,C., Deplus,R., Blanchon,L., Fraga,M., Didelot,C., Morey,L., Van Eynde,A., Bernard,D., Vanderwinden,J.-M., et al. (2006) The Polycomb group protein EZH2 directly controls DNA methylation. Nature, 439, 871–4.

25. Li,H., Rauch,T., Chen,Z.-X., Szabó,P.E., Riggs,A.D. and Pfeifer,G.P. (2006) The histone methyltransferase SETDB1 and the DNA methyltransferase DNMT3A interact directly and localize to promoters silenced in cancer cells. J. Biol. Chem., 281, 19489–500.

26. Winter,J., Jung,S., Keller,S., Gregory,R.I. and Diederichs,S. (2009) Many roads to maturity: microRNA biogenesis pathways and their regulation. Nat. Cell Biol., 11, 228–34.

27. Massagué,J. (1998) TGF-beta signal transduction. Annu. Rev. Biochem., 67, 753–91.

28. Moustakas,A. and Heldin,C.-H. (2009) The regulation of TGFbeta signal transduction. Development, 136, 3699–714.

29. Thiery,J.P., Acloque,H., Huang,R.Y.J. and Nieto,M.A. (2009) Epithelial-mesenchymal transitions in development and disease. Cell, 139, 871–90.

30. Moustakas,A. and Heldin,C.-H. (2012) Induction of epithelial-mesenchymal transition by transforming growth factor β. Semin. Cancer Biol., 22, 446–54.

31. Massagué,J. (2008) TGFbeta in Cancer. Cell, 134, 215–30. 32. Padua,D. and Massagué,J. (2009) Roles of TGFbeta in metastasis. Cell Res.,

19, 89–102. 33. Ten Dijke,P. and Arthur,H.M. (2007) Extracellular control of TGFbeta

signalling in vascular development and disease. Nat. Rev. Mol. Cell Biol., 8, 857–69.

34. Bernabeu,C., Lopez-Novoa,J.M. and Quintanilla,M. (2009) The emerging role of TGF-beta superfamily coreceptors in cancer. Biochim. Biophys. Acta, 1792, 954–73.

35. Groppe,J., Hinck,C.S., Samavarchi-Tehrani,P., Zubieta,C., Schuermann,J.P., Taylor,A.B., Schwarz,P.M., Wrana,J.L. and Hinck,A.P. (2008) Cooperative assembly of TGF-beta superfamily signaling complexes is mediated by two disparate mechanisms and distinct modes of receptor binding. Mol. Cell, 29, 157–68.

42

36. Lawler,S., Feng,X.H., Chen,R.H., Maruoka,E.M., Turck,C.W., Griswold-Prenner,I. and Derynck,R. (1997) The type II transforming growth factor-beta receptor autophosphorylates not only on serine and threonine but also on tyrosine residues. J. Biol. Chem., 272, 14850–9.

37. Lee,M.K., Pardoux,C., Hall,M.C., Lee,P.S., Warburton,D., Qing,J., Smith,S.M. and Derynck,R. (2007) TGF-beta activates Erk MAP kinase signalling through direct phosphorylation of ShcA. EMBO J., 26, 3957–67.

38. Heldin,C.-H. and Moustakas,A. (2012) Role of Smads in TGFβ signaling. Cell Tissue Res., 347, 21–36.

39. Massagué,J. (2012) TGFβ signalling in context. Nat. Rev. Mol. Cell Biol., 13, 616–30.

40. Massagué,J., Seoane,J. and Wotton,D. (2005) Smad transcription factors. Genes Dev., 19, 2783–810.

41. Shi,Y. (2006) Structural insights into Smad function and specificity. In ten Dijke,P., Heldin,C.-H. (eds), Smad Signal Transduction. Springer, Dordrecht, pp. 215–233.

42. Sundqvist,A., Ten Dijke,P. and van Dam,H. (2012) Key signaling nodes in mammary gland development and cancer: Smad signal integration in epithelial cell plasticity. Breast Cancer Res., 14, 204.

43. Morikawa,M., Koinuma,D., Miyazono,K. and Heldin,C.-H. (2013) Genome-wide mechanisms of Smad binding. Oncogene, 32, 1609–15.

44. Zawel,L., Dai,J.L., Buckhaults,P., Zhou,S., Kinzler,K.W., Vogelstein,B. and Kern,S.E. (1998) Human Smad3 and Smad4 are sequence-specific transcription activators. Mol. Cell, 1, 611–7.

45. Koinuma,D., Tsutsumi,S., Kamimura,N., Taniguchi,H., Miyazawa,K., Sunamura,M., Imamura,T., Miyazono,K. and Aburatani,H. (2009) Chromatin immunoprecipitation on microarray analysis of Smad2/3 binding sites reveals roles of ETS1 and TFAP2A in transforming growth factor beta signaling. Mol. Cell. Biol., 29, 172–86.

46. Koinuma,D., Tsutsumi,S., Kamimura,N., Imamura,T., Aburatani,H. and Miyazono,K. (2009) Promoter-wide analysis of Smad4 binding sites in human epithelial cells. Cancer Sci., 100, 2133–42.

47. Kim,J., Johnson,K., Chen,H.J., Carroll,S. and Laughon,A. (1997) Drosophila Mad binds to DNA and directly mediates activation of vestigial by Decapentaplegic. Nature, 388, 304–8.

48. Morikawa,M., Koinuma,D., Tsutsumi,S., Vasilaki,E., Kanki,Y., Heldin,C.-H., Aburatani,H. and Miyazono,K. (2011) ChIP-seq reveals cell type-specific binding patterns of BMP-specific Smads and a novel binding motif. Nucleic Acids Res., 39, 8712–27.

49. Mullen,A.C., Orlando,D.A., Newman,J.J., Lovén,J., Kumar,R.M., Bilodeau,S., Reddy,J., Guenther,M.G., DeKoter,R.P. and Young,R.A. (2011) Master transcription factors determine cell-type-specific responses to TGF-β signaling. Cell, 147, 565–76.

50. Sundqvist,A., Zieba,A., Vasilaki,E., Herrera Hidalgo,C., Söderberg,O., Koinuma,D., Miyazono,K., Heldin,C.-H., Landegren,U., Ten Dijke,P., et al. (2012) Specific interactions between Smad proteins and AP-1 components determine TGFβ-induced breast cancer cell invasion. Oncogene, 32, 3606-15.

43

51. Pardali,K., Kurisaki,A., Morén,A., ten Dijke,P., Kardassis,D. and Moustakas,A. (2000) Role of Smad proteins and transcription factor Sp1 in p21(Waf1/Cip1) regulation by transforming growth factor-beta. J. Biol. Chem., 275, 29244–56.

52. Seoane,J., Le,H.-V., Shen,L., Anderson,S.A. and Massagué,J. (2004) Integration of Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation. Cell, 117, 211–23.

53. Seoane,J., Pouponnot,C., Staller,P., Schader,M., Eilers,M. and Massagué,J. (2001) TGFbeta influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b. Nat. Cell Biol., 3, 400–8.

54. Gomis,R.R., Alarcón,C., He,W., Wang,Q., Seoane,J., Lash,A. and Massagué,J. (2006) A FoxO-Smad synexpression group in human keratinocytes. Proc. Natl. Acad. Sci. U. S. A., 103, 12747–52.

55. Frederick,J.P., Liberati,N.T., Waddell,D.S., Shi,Y. and Wang,X.-F. (2004) Transforming growth factor beta-mediated transcriptional repression of c-myc is dependent on direct binding of Smad3 to a novel repressive Smad binding element. Mol. Cell. Biol., 24, 2546–59.

56. Kang,Y., Chen,C.-R. and Massagué,J. (2003) A self-enabling TGFbeta response coupled to stress signaling: Smad engages stress response factor ATF3 for Id1 repression in epithelial cells. Mol. Cell, 11, 915–26.

57. Simonsson,M., Kanduri,M., Grönroos,E., Heldin,C.-H. and Ericsson,J. (2006) The DNA binding activities of Smad2 and Smad3 are regulated by coactivator-mediated acetylation. J. Biol. Chem., 281, 39870–80.

58. Kang,J.S., Alliston,T., Delston,R. and Derynck,R. (2005) Repression of Runx2 function by TGF-beta through recruitment of class II histone deacetylases by Smad3. EMBO J., 24, 2543–55.

59. Mu,Y., Gudey,S.K. and Landström,M. (2012) Non-Smad signaling pathways. Cell Tissue Res., 347, 11–20.

60. Yu,L., Hébert,M.C. and Zhang,Y.E. (2002) TGF-beta receptor-activated p38 MAP kinase mediates Smad-independent TGF-beta responses. EMBO J., 21, 3749–59.

61. Itoh,S., Thorikay,M., Kowanetz,M., Moustakas,A., Itoh,F., Heldin,C.-H. and ten Dijke,P. (2003) Elucidation of Smad requirement in transforming growth factor-beta type I receptor-induced responses. J. Biol. Chem., 278, 3751–61.

62. Sorrentino,A., Thakur,N., Grimsby,S., Marcusson,A., von Bulow,V., Schuster,N., Zhang,S., Heldin,C.-H. and Landström,M. (2008) The type I TGF-beta receptor engages TRAF6 to activate TAK1 in a receptor kinase-independent manner. Nat. Cell Biol., 10, 1199–207.

63. Mu,Y., Sundar,R., Thakur,N., Ekman,M., Gudey,S.K., Yakymovych,M., Hermansson,A., Dimitriou,H., Bengoechea-Alonso,M.T., Ericsson,J., et al. (2011) TRAF6 ubiquitinates TGFβ type I receptor to promote its cleavage and nuclear translocation in cancer. Nat. Commun, 2, 330.

64. Galliher,A.J. and Schiemann,W.P. (2007) Src phosphorylates Tyr284 in TGF-beta type II receptor and regulates TGF-beta stimulation of p38 MAPK during breast cancer cell proliferation and invasion. Cancer Res., 67, 3752–8.

65. Kardassis,D., Murphy,C., Fotsis,T., Moustakas,A. and Stournaras,C. (2009) Control of transforming growth factor beta signal transduction by small GTPases. FEBS J., 276, 2947–65.

44

66. Bhowmick,N.A., Ghiassi,M., Bakin,A., Aakre,M., Lundquist,C.A., Engel,M.E., Arteaga,C.L. and Moses,H.L. (2001) Transforming growth factor-beta1 mediates epithelial to mesenchymal transdifferentiation through a RhoA-dependent mechanism. Mol. Biol. Cell, 12, 27–36.

67. Ozdamar,B., Bose,R., Barrios-Rodiles,M., Wang,H.-R., Zhang,Y. and Wrana,J.L. (2005) Regulation of the polarity protein Par6 by TGFbeta receptors controls epithelial cell plasticity. Science, 307, 1603–9.

68. Lamouille,S. and Derynck,R. (2007) Cell size and invasion in TGF-beta-induced epithelial to mesenchymal transition is regulated by activation of the mTOR pathway. J. Cell Biol., 178, 437–51.

69. Lamouille,S., Connolly,E., Smyth,J.W., Akhurst,R.J. and Derynck,R. (2012) TGF-β-induced activation of mTOR complex 2 drives epithelial-mesenchymal transition and cell invasion. J. Cell Sci., 125, 1259–73.

70. Mani,S.A., Guo,W., Liao,M.-J., Eaton,E.N., Ayyanan,A., Zhou,A.Y., Brooks,M., Reinhard,F., Zhang,C.C., Shipitsin,M., et al. (2008) The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell, 133, 704–15.

71. Morel,A.-P., Lièvre,M., Thomas,C., Hinkal,G., Ansieau,S. and Puisieux,A. (2008) Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS One, 3, e2888.

72. Baccelli,I. and Trumpp,A. (2012) The evolving concept of cancer and metastasis stem cells. J. Cell Biol., 198, 281–93.

73. Bruna,A., Greenwood,W., Le Quesne,J., Teschendorff,A., Miranda-Saavedra,D., Rueda,O.M., Sandoval,J.L., Vidakovic,A.T., Saadi,A., Pharoah,P., et al. (2012) TGFβ induces the formation of tumour-initiating cells in claudinlow breast cancer. Nat. Commun., 3, 1055.

74. Nguyen,D.X., Bos,P.D. and Massagué,J. (2009) Metastasis: from dissemination to organ-specific colonization. Nat. Rev. Cancer, 9, 274–84.

75. Yin,J.J., Selander,K., Chirgwin,J.M., Dallas,M., Grubbs,B.G., Wieser,R., Massagué,J., Mundy,G.R. and Guise,T.A. (1999) TGF-beta signaling blockade inhibits PTHrP secretion by breast cancer cells and bone metastases development. J. Clin. Invest., 103, 197–206.

76. Kang,Y., Siegel,P.M., Shu,W., Drobnjak,M., Kakonen,S.M., Cordón-Cardo,C., Guise,T.A. and Massagué,J. (2003) A multigenic program mediating breast cancer metastasis to bone. Cancer cell, 3, 537–49.

77. Padua,D., Zhang,X.H.-F., Wang,Q., Nadal,C., Gerald,W.L., Gomis,R.R. and Massagué,J. (2008) TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell, 133, 66–77.

78. Adorno,M., Cordenonsi,M., Montagner,M., Dupont,S., Wong,C., Hann,B., Solari,A., Bobisse,S., Rondina,M.B., Guzzardo,V., et al. (2009) A Mutant-p53/Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell, 137, 87–98.

79. Taylor,M.A., Sossey-Alaoui,K., Thompson,C.L., Danielpour,D. and Schiemann,W.P. (2013) TGF-β upregulates miR-181a expression to promote breast cancer metastasis. J. Clin. Invest., 123, 150–63.

80. Cleynen,I. and Van de Ven,W.J.M. (2008) The HMGA proteins: a myriad of functions (Review). Int. J. Oncol., 32, 289–305.

45

81. Cattaruzzi,G., Altamura,S., Tessari,M.A., Rustighi,A., Giancotti,V., Pucillo,C. and Manfioletti,G. (2007) The second AT-hook of the architectural transcription factor HMGA2 is determinant for nuclear localization and function. Nucleic Acids Res., 35, 1751–60.

82. Zhou,X., Benson,K.F., Ashar,H.R. and Chada,K. (1995) Mutation responsible for the mouse pygmy phenotype in the developmentally regulated factor HMGI-C. Nature, 376, 771–4.

83. Rogalla,P., Drechsler,K., Frey,G., Hennig,Y., Helmke,B., Bonk,U. and Bullerdiek,J. (1996) HMGI-C expression patterns in human tissues. Implications for the genesis of frequent mesenchymal tumors. Am. J. Pathol., 149, 775–9.

84. Li,O., Vasudevan,D., Davey,C.A. and Dröge,P. (2006) High-level expression of DNA architectural factor HMGA2 and its association with nucleosomes in human embryonic stem cells. Genesis, 44, 523–9.

85. Fedele,M. and Fusco,A. (2010) HMGA and cancer. Biochim. Biophys. Acta, 1799, 48–54.

86. Tessari,M.A., Gostissa,M., Altamura,S., Sgarra,R., Rustighi,A., Salvagno,C., Caretti,G., Imbriano,C., Mantovani,R., Del Sal,G., et al. (2003) Transcriptional activation of the cyclin A gene by the architectural transcription factor HMGA2. Mol. Cell. Biol., 23, 9104–16.

87. De Martino,I., Visone,R., Wierinckx,A., Palmieri,D., Ferraro,A., Cappabianca,P., Chiappetta,G., Forzati,F., Lombardi,G., Colao,A., et al. (2009) HMGA proteins up-regulate CCNB2 gene in mouse and human pituitary adenomas. Cancer Res., 69, 1844–50.

88. Fedele,M., Visone,R., De Martino,I., Troncone,G., Palmieri,D., Battista,S., Ciarmiello,A., Pallante,P., Arra,C., Melillo,R.M., et al. (2006) HMGA2 induces pituitary tumorigenesis by enhancing E2F1 activity. Cancer cell, 9, 459–71.

89. Nishino,J., Kim,I., Chada,K. and Morrison,S.J. (2008) Hmga2 promotes neural stem cell self-renewal in young but not old mice by reducing p16Ink4a and p19Arf Expression. Cell, 135, 227–39.

90. Anand,A. and Chada,K. (2000) In vivo modulation of Hmgic reduces obesity. Nat. Genet., 24, 377–80.

91. Battista,S., Fidanza,V., Fedele,M., Klein-Szanto,A.J., Outwater,E., Brunner,H., Santoro,M., Croce,C.M. and Fusco,A. (1999) The expression of a truncated HMGI-C gene induces gigantism associated with lipomatosis. Cancer Res., 59, 4793–7.

92. Weedon,M.N., Lettre,G., Freathy,R.M., Lindgren,C.M., Voight,B.F., Perry,J.R.B., Elliott,K.S., Hackett,R., Guiducci,C., Shields,B., et al. (2007) A common variant of HMGA2 is associated with adult and childhood height in the general population. Nat. Genet., 39, 1245–50.

93. Monzen,K., Ito,Y., Naito,A.T., Kasai,H., Hiroi,Y., Hayashi,D., Shiojima,I., Yamazaki,T., Miyazono,K., Asashima,M., et al. (2008) A crucial role of a high mobility group protein HMGA2 in cardiogenesis. Nat. Cell Biol., 10, 567–74.

94. Kishi,Y., Fujii,Y., Hirabayashi,Y. and Gotoh,Y. (2012) HMGA regulates the global chromatin state and neurogenic potential in neocortical precursor cells. Nat. Neurosci., 15, 1127–33.

46

95. Thuault,S., Valcourt,U., Petersen,M., Manfioletti,G., Heldin,C.-H. and Moustakas,A. (2006) Transforming growth factor-beta employs HMGA2 to elicit epithelial-mesenchymal transition. J. Cell Biol., 174, 175–83.

96. Wend,P., Runke,S., Wend,K., Anchondo,B., Yesayan,M., Jardon,M., Hardie,N., Loddenkemper,C., Ulasov,I., Lesniak,M.S., et al. (2013) WNT10B/β-catenin signalling induces HMGA2 and proliferation in metastatic triple-negative breast cancer. EMBO Mol. Med., 5, 264–79.

97. Lee,Y.S. and Dutta,A. (2007) The tumor suppressor microRNA let-7 represses the HMGA2 oncogene. Genes Dev., 21, 1025–30.

98. Mayr,C., Hemann,M.T. and Bartel,D.P. (2007) Disrupting the Pairing Between let-7 and Hmga2 Enhances Oncogenic Transformation. Science, 315, 1576–1579.

99. Copley,M.R., Babovic,S., Benz,C., Knapp,D.J.H.F., Beer,P.A., Kent,D.G., Wohrer,S., Treloar,D.Q., Day,C., Rowe,K., et al. (2013) The Lin28b-let-7-Hmga2 axis determines the higher self-renewal potential of fetal haematopoietic stem cells. Nat. Cell Biol., 15, 916-25.

100. Thuault,S., Tan,E.-J., Peinado,H., Cano,A., Heldin,C.-H. and Moustakas,A. (2008) HMGA2 and Smads co-regulate SNAIL1 expression during induction of epithelial-to-mesenchymal transition. J. Biol. Chem., 283, 33437–46.

101. Tan,E.-J., Thuault,S., Caja,L., Carletti,T., Heldin,C.-H. and Moustakas,A. (2012) Regulation of transcription factor Twist expression by the DNA architectural protein high mobility group A2 during epithelial-to-mesenchymal transition. J. Biol. Chem., 287, 7134–45.

102. Borrmann,L., Schwanbeck,R., Heyduk,T., Seebeck,B., Rogalla,P., Bullerdiek,J. and Wisniewski,J.R. (2003) High mobility group A2 protein and its derivatives bind a specific region of the promoter of DNA repair gene ERCC1 and modulate its activity. Nucleic Acids Res., 31, 6841–51.

103. Catez,F. and Hock,R. (2010) Binding and interplay of HMG proteins on chromatin: lessons from live cell imaging. Biochim Biophys.Acta,1799,15-27.

104. Zhao,K., Käs,E., Gonzalez,E. and Laemmli,U.K. (1993) SAR-dependent mobilization of histone H1 by HMG-I/Y in vitro: HMG-I/Y is enriched in H1-depleted chromatin. EMBO J., 12, 3237–47.

105. Funayama,R., Saito,M., Tanobe,H. and Ishikawa,F. (2006) Loss of linker histone H1 in cellular senescence. J. Cell Biol., 175, 869–80.

106. Sgarra,R., Zammitti,S., Lo Sardo,A., Maurizio,E., Arnoldo,L., Pegoraro,S., Giancotti,V. and Manfioletti,G. HMGA molecular network: From transcriptional regulation to chromatin remodeling. Biochim. Biophys. Acta, 1799, 37–47.

107. Yie,J., Merika,M., Munshi,N., Chen,G. and Thanos,D. (1999) The role of HMG I(Y) in the assembly and function of the IFN-beta enhanceosome. EMBO J., 18, 3074–89.

108. Mantovani,F., Covaceuszach,S., Rustighi,A., Sgarra,R., Heath,C., Goodwin,G.H. and Manfioletti,G. (1998) NF-kappaB mediated transcriptional activation is enhanced by the architectural factor HMGI-C. Nucleic Acids Res., 26, 1433–9.

109. Li,A.Y.-J., Lin,H.H., Kuo,C.-Y., Shih,H.-M., Wang,C.C.C., Yen,Y. and Ann,D.K. (2011) High-mobility group A2 protein modulates hTERT transcription to promote tumorigenesis. Mol. Cell. Biol., 31, 2605–17.

47

110. Vallone,D., Battista,S., Pierantoni,G.M., Fedele,M., Casalino,L., Santoro,M., Viglietto,G., Fusco,A. and Verde,P. (1997) Neoplastic transformation of rat thyroid cells requires the junB and fra-1 gene induction which is dependent on the HMGI-C gene product. EMBO J., 16, 5310–21.

111. Zong,Y., Huang,J., Sankarasharma,D., Morikawa,T., Fukayama,M., Epstein,J.I., Chada,K.K. and Witte,O.N. (2012) Stromal epigenetic dysregulation is sufficient to initiate mouse prostate cancer via paracrine Wnt signaling. Proc. Natl. Acad. Sci. U. S. A., 109, E3395–404.

112. Craene,B. De and Berx,G. (2012) Regulatory networks defining EMT during cancer initiation and progression. Nat. Rev. Cancer, 13, 97–110.

113. Sun,M., Song,C.-X., Huang,H., Frankenberger,C.A., Sankarasharma,D., Gomes,S., Chen,P., Chen,J., Chada,K.K., He,C., et al. (2013) HMGA2/TET1/HOXA9 signaling pathway regulates breast cancer growth and metastasis. Proc. Natl. Acad. Sci. U. S. A., 10.1073/pnas.1305172110.

114. Morishita,A., Zaidi,M.R., Mitoro,A., Sankarasharma,D., Szabolcs,M., Okada,Y., D’Armiento,J. and Chada,K. (2013) HMGA2 is a Driver of Tumor Metastasis. Cancer Res., 73, 4289-99.

115. Yu,F., Yao,H., Zhu,P., Zhang,X., Pan,Q., Gong,C., Huang,Y., Hu,X., Su,F., Lieberman,J., et al. (2007) let-7 regulates self renewal and tumorigenicity of breast cancer cells. Cell, 131, 1109–23.

116. Shell,S., Park,S.-M., Radjabi,A.R., Schickel,R., Kistner,E.O., Jewell,D.A., Feig,C., Lengyel,E. and Peter,M.E. (2007) Let-7 expression defines two differentiation stages of cancer. Proc. Natl. Acad. Sci. U. S. A., 104, 11400–5.

117. Gheldof,A. and Berx,G. (2013) Cadherins and epithelial-to-mesenchymal transition. Prog. Mol. Biol. Thransl. Sci., 116, 317–36.

118. Nieto,M.A. (2001) The early steps of neural crest development. Mech. Dev., 105, 27–35.

119. Boyer,A.S., Ayerinskas,I.I., Vincent,E.B., McKinney,L.A., Weeks,D.L. and Runyan,R.B. (1999) TGFbeta2 and TGFbeta3 have separate and sequential activities during epithelial-mesenchymal cell transformation in the embryonic heart. Dev. Biol., 208, 530–45.

120. Nawshad,A. and Hay,E.D. (2003) TGFbeta3 signaling activates transcription of the LEF1 gene to induce epithelial mesenchymal transformation during mouse palate development. J. Cell Biol., 163, 1291–301.

121. Moustakas,A. and Heldin,C.-H. (2007) Signaling networks guiding epithelial-mesenchymal transitions during embryogenesis and cancer progression. Cancer Sci., 98, 1512–20.

122. Cui,W., Fowlis,D.J., Bryson,S., Duffie,E., Ireland,H., Balmain,A. and Akhurst,R.J. (1996) TGFbeta1 inhibits the formation of benign skin tumors, but enhances progression to invasive spindle carcinomas in transgenic mice. Cell, 86, 531–42.

123. Piek,E., Moustakas,A., Kurisaki,A., Heldin,C.H. and ten Dijke,P. (1999) TGF-(beta) type I receptor/ALK-5 and Smad proteins mediate epithelial to mesenchymal transdifferentiation in NMuMG breast epithelial cells. J. Cell Sci., 112 ( Pt 2, 4557–68.

124. Valcourt,U., Kowanetz,M., Niimi,H., Heldin,C.-H. and Moustakas,A. (2005) TGF-beta and the Smad signaling pathway support transcriptomic reprogramming during epithelial-mesenchymal cell transition. Mol. Biol. Cell, 16, 1987–2002.

48

125. Vincent,T., Neve,E.P.A., Johnson,J.R., Kukalev,A., Rojo,F., Albanell,J., Pietras,K., Virtanen,I., Philipson,L., Leopold,P.L., et al. (2009) A SNAIL1-SMAD3/4 transcriptional repressor complex promotes TGF-beta mediated epithelial-mesenchymal transition. Nat. Cell Biol., 11, 943–50.

126. Fischer,A.N.M., Fuchs,E., Mikula,M., Huber,H., Beug,H. and Mikulits,W. (2007) PDGF essentially links TGF-beta signaling to nuclear beta-catenin accumulation in hepatocellular carcinoma progression. Oncogene, 26, 3395–405.

127. Zavadil,J., Cermak,L., Soto-Nieves,N. and Böttinger,E.P. (2004) Integration of TGF-beta/Smad and Jagged1/Notch signalling in epithelial-to-mesenchymal transition. EMBO J., 23, 1155–65.

128. Kowanetz,M., Valcourt,U., Bergström,R., Heldin,C.-H. and Moustakas,A. (2004) Id2 and Id3 define the potency of cell proliferation and differentiation responses to transforming growth factor beta and bone morphogenetic protein. Mol. Cell. Biol., 24, 4241–54.

129. Kong,W., Yang,H., He,L., Zhao,J., Coppola,D., Dalton,W.S. and Cheng,J.Q. (2008) MicroRNA-155 is regulated by the transforming growth factor beta/Smad pathway and contributes to epithelial cell plasticity by targeting RhoA. Mol. Cell. Biol., 28, 6773–84.

130. Wong,E.W.P., Mruk,D.D., Lee,W.M. and Cheng,C.Y. (2008) Par3/Par6 polarity complex coordinates apical ectoplasmic specialization and blood-testis barrier restructuring during spermatogenesis. Proc. Natl. Acad. Sci. U. S. A., 105, 9657–62.

131. Nieto,M.A. (2002) The snail superfamily of zinc-finger transcription factors. Nat. Rev. Mol. Cell Biol., 3, 155–66.

132. Peinado,H., Ballestar,E., Esteller,M. and Cano,A. (2004) Snail mediates E-cadherin repression by the recruitment of the Sin3A/histone deacetylase 1 (HDAC1)/HDAC2 complex. Mol. Cell. Biol., 24, 306–19.

133. Peinado,H., Olmeda,D. and Cano,A. (2007) Snail, Zeb and bHLH factors in tumour progression: an alliance against the epithelial phenotype? Nat. Rev. Cancer, 7, 415–28.

134. Vandewalle,C., Van Roy,F. and Berx,G. (2009) The role of the ZEB family of transcription factors in development and disease. Cell Mol. Life Sci., 66, 773–87.

135. Comijn,J., Berx,G., Vermassen,P., Verschueren,K., van Grunsven,L., Bruyneel,E., Mareel,M., Huylebroeck,D. and van Roy,F. (2001) The two-handed E box binding zinc finger protein SIP1 downregulates E-cadherin and induces invasion. Mol. Cell, 7, 1267–78.

136. Vandewalle,C., Comijn,J., De Craene,B., Vermassen,P., Bruyneel,E., Andersen,H., Tulchinsky,E., Van Roy,F. and Berx,G. (2005) SIP1/ZEB2 induces EMT by repressing genes of different epithelial cell-cell junctions. Nucleic Acids Res., 33, 6566–78.

137. Spaderna,S., Schmalhofer,O., Wahlbuhl,M., Dimmler,A., Bauer,K., Sultan,A., Hlubek,F., Jung,A., Strand,D., Eger,A., et al. (2008) The transcriptional repressor ZEB1 promotes metastasis and loss of cell polarity in cancer. Cancer Res., 68, 537–44.

138. Ansieau,S., Morel,A.-P., Hinkal,G., Bastid,J. and Puisieux,A. (2010) TWISTing an embryonic transcription factor into an oncoprotein. Oncogene, 29, 3173–84.

49

139. Casas,E., Kim,J., Bendesky,A., Ohno-Machado,L., Wolfe,C.J. and Yang,J. (2011) Snail2 is an essential mediator of Twist1-induced epithelial mesenchymal transition and metastasis. Cancer Res., 71, 245–54.

140. Yang,M.-H., Hsu,D.S.-S., Wang,H.-J.H.-W., Lan,H.-Y., Yang,W.-H., Huang,C.-H., Kao,S.-Y., Tzeng,C.-H., Tai,S.-K., Chang,S.-Y., et al. (2010) Bmi1 is essential in Twist1-induced epithelial-mesenchymal transition. Nat. Cell Biol., 12, 982–92.

141. Yang,F., Sun,L., Li,Q., Han,X., Lei,L., Zhang,H. and Shang,Y. (2012) SET8 promotes epithelial-mesenchymal transition and confers TWIST dual transcriptional activities. EMBO J., 31, 110–23.

142. Alexander,N.R., Tran,N.L., Rekapally,H., Summers,C.E., Glackin,C. and Heimark,R.L. (2006) N-cadherin gene expression in prostate carcinoma is modulated by integrin-dependent nuclear translocation of Twist1. Cancer Res., 66, 3365–9.

143. Cheng,G.Z., Zhang,W.Z., Sun,M., Wang,Q., Coppola,D., Mansour,M., Xu,L.M., Costanzo,C., Cheng,J.Q. and Wang,L.-H. (2008) Twist is transcriptionally induced by activation of STAT3 and mediates STAT3 oncogenic function. J. Biol. Chem., 283, 14665–73.

144. Yang,J., Mani,S.A., Donaher,J.L., Ramaswamy,S., Itzykson,R.A., Come,C., Savagner,P., Gitelman,I., Richardson,A. and Weinberg,R.A. (2004) Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell, 117, 927–39.

145. Eckert,M.A., Lwin,T.M., Chang,A.T., Kim,J., Danis,E., Ohno-Machado,L. and Yang,J. (2011) Twist1-induced invadopodia formation promotes tumor metastasis. Cancer cell, 19, 372–86.

146. Dave,N., Guaita-Esteruelas,S., Gutarra,S., Frias,À., Beltran,M., Peiró,S. and de Herreros,A.G. (2011) Functional cooperation between Snail1 and twist in the regulation of ZEB1 expression during epithelial to mesenchymal transition. J. Biol. Chem., 286, 12024–32.

147. McDonald,O.G., Wu,H., Timp,W., Doi,A. and Feinberg,A.P. (2011) Genome-scale epigenetic reprogramming during epithelial-to-mesenchymal transition. Nat. Struct. Mol. Biol., 18, 867–74.

148. Tiwari,N., Tiwari,V.K., Waldmeier,L., Balwierz,P.J., Arnold,P., Pachkov,M., Meyer-Schaller,N., Schübeler,D., van Nimwegen,E. and Christofori,G. (2013) Sox4 Is a Master Regulator of Epithelial-Mesenchymal Transition by Controlling Ezh2 Expression and Epigenetic Reprogramming. Cancer Cell, 23, 768–783.

149. Dumont,N., Wilson,M.B., Crawford,Y.G., Reynolds,P.A., Sigaroudinia,M. and Tlsty,T.D. (2008) Sustained induction of epithelial to mesenchymal transition activates DNA methylation of genes silenced in basal-like breast cancers. Proc. Natl. Acad. Sci. U. S. A., 105, 14867–72.

150. Lombaerts,M., van Wezel,T., Philippo,K., Dierssen,J.W.F., Zimmerman,R.M.E., Oosting,J., van Eijk,R., Eilers,P.H., van de Water,B., Cornelisse,C.J., et al. (2006) E-cadherin transcriptional downregulation by promoter methylation but not mutation is related to epithelial-to-mesenchymal transition in breast cancer cell lines. Br. J. Cancer, 94, 661–71.

50

151. Papageorgis,P., Lambert,A.W., Ozturk,S., Gao,F., Pan,H., Manne,U., Alekseyev,Y.O., Thiagalingam,A., Abdolmaleky,H.M., Lenburg,M., et al. (2010) Smad signaling is required to maintain epigenetic silencing during breast cancer progression. Cancer Res., 70, 968–78.

152. Lin,T., Ponn,A., Hu,X., Law,B.K. and Lu,J. (2010) Requirement of the histone demethylase LSD1 in Snai1-mediated transcriptional repression during epithelial-mesenchymal transition. Oncogene, 29, 4896–904.

153. Dong,C., Wu,Y., Yao,J., Wang,Y., Yu,Y., Rychahou,P.G., Evers,B.M. and Zhou,B.P. (2012) G9a interacts with Snail and is critical for Snail-mediated E-cadherin repression in human breast cancer. J. Clin. Invest., 122, 1469–86.

154. Dong,C., Wu,Y., Wang,Y., Wang,C., Kang,T., Rychahou,P.G., Chi,Y.-I., Evers,B.M. and Zhou,B.P. (2013) Interaction with Suv39H1 is critical for Snail-mediated E-cadherin repression in breast cancer. Oncogene, 32,1351-62.

155. Lamouille,S., Subramanyam,D., Blelloch,R. and Derynck,R. (2013) Regulation of epithelial-mesenchymal and mesenchymal-epithelial transitions by microRNAs. Curr. Opin. Cell Biol., 25, 200–7.

156. Gregory,P.A., Bert,A.G., Paterson,E.L., Barry,S.C., Tsykin,A., Farshid,G., Vadas,M.A., Khew-Goodall,Y. and Goodall,G.J. (2008) The miR-200 family and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 and SIP1. Nat. Cell Biol., 10, 593–601.

157. Korpal,M., Lee,E.S., Hu,G. and Kang,Y. (2008) The miR-200 family inhibits epithelial-mesenchymal transition and cancer cell migration by direct targeting of E-cadherin transcriptional repressors ZEB1 and ZEB2. J. Biol. Chem., 283, 14910–4.

158. Burk,U., Schubert,J., Wellner,U., Schmalhofer,O., Vincan,E., Spaderna,S. and Brabletz,T. (2008) A reciprocal repression between ZEB1 and members of the miR-200 family promotes EMT and invasion in cancer cells. EMBO Rep., 9, 582–9.

159. Siemens,H., Jackstadt,R., Hünten,S., Kaller,M., Menssen,A., Götz,U. and Hermeking,H. (2011) miR-34 and SNAIL form a double-negative feedback loop to regulate epithelial-mesenchymal transitions. Cell cycle, 10, 4256–71.

160. Ma,L., Teruya-Feldstein,J. and Weinberg,R.A. (2007) Tumour invasion and metastasis initiated by microRNA-10b in breast cancer. Nature, 449, 682–8.

161. Vega,S., Morales,A. V, Ocaña,O.H., Valdés,F., Fabregat,I. and Nieto,M.A. (2004) Snail blocks the cell cycle and confers resistance to cell death. Genes Dev., 18, 1131–43.

162. Liu,Y., El-Naggar,S., Darling,D.S., Higashi,Y. and Dean,D.C. (2008) Zeb1 links epithelial-mesenchymal transition and cellular senescence. Development, 135, 579–88.

163. Ansieau,S., Bastid,J., Doreau,A., Morel,A.-P., Bouchet,B.P., Thomas,C., Fauvet,F., Puisieux,I., Doglioni,C., Piccinin,S., et al. (2008) Induction of EMT by twist proteins as a collateral effect of tumor-promoting inactivation of premature senescence. Cancer cell, 14, 79–89.

164. Gunasinghe,N.P.A.D., Wells,A., Thompson,E.W. and Hugo,H.J. (2012) Mesenchymal-epithelial transition (MET) as a mechanism for metastatic colonisation in breast cancer. Cancer Metastasis Rev., 31, 469–78.

51

165. Witcher,M. and Emerson,B.M. (2009) Epigenetic silencing of the p16(INK4a) tumor suppressor is associated with loss of CTCF binding and a chromatin boundary. Mol. Cell, 34, 271–84.

166. Winter,N., Nimzyk,R., Bösche,C., Meyer,A. and Bullerdiek,J. (2011) Chromatin immunoprecipitation to analyze DNA binding sites of HMGA2. PLoS One, 6, e18837.

167. Zha,L., Wang,Z., Tang,W., Zhang,N., Liao,G. and Huang,Z. (2012) Genome-wide analysis of HMGA2 transcription factor binding sites by ChIP on chip in gastric carcinoma cells. Mol. Cell. Biochem., 364, 243–51.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 928

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine.

Distribution: publications.uu.seurn:nbn:se:uu:diva-206120

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2013