Cross-Regulation Between Tgfβ/BMP Signalling and the Metabolic LKB1

To my family and friends along the way

For your love and support

List of Papers

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

I Morén, A. #, Raja, E. #, Heldin, C-H., Moustakas, A. Negative regulation of TGFβ signalling by the kinase LKB1 and the scaffolding protein LIP1. Journal of Biological Chemistry (2011), 286(1):341-53. II Raja, E. Zieba, A., Morén, A., Voytyuk, I., Söderberg, O., Heldin, C-H., Moustakas, A. The tumour suppressor kinase LKB1 negatively regulates bone morphogenetic receptor signalling. Manuscript submitted (2012). III Lönn, P. #, Vanlandewijck, M. #, Raja, E., Kowanetz, M., Wa- tanabe, Y., Kowanetz, K., Vasilaki, E., Heldin, C-H., Mousta- kas, A. Transcriptional induction of salt-inducible kinase 1 by transforming growth factor β leads to negative regulation of type I receptor signalling in cooperation with the Smurf2 ubiquitin ligase. Journal of Biological Chemistry (2012), 287(16):12867-78. IV Van der Heide, L., Raja, E., Heldin, C-H., Moustakas, A. Tran- scriptional induction of Nuak2 by TGFβ leads to phosphorylation and inhibition of the translational regulator Raptor. Manu- script (2012).

# Indicates that these authors contributed equally to the work

Reprints were made with permission from the respective publishers.

Contents

Introduction ...... 9 TGFβ/BMP Signalling ...... 10 A brief history ...... 10 The Smad-dependent signalling pathway ...... 10 TGFβ Signalling Regulators ...... 14 Receptor regulation ...... 14 Smads regulation ...... 17 Signalling cross-talk through the non-Smad signalling pathways ...... 20 TGFβ signalling and the MAPK pathways ...... 20 TGFβ signalling and the PI3K/Akt – mTOR pathway ...... 21 Physiological responses to TGFβ signalling ...... 22 Aberrant TGFβ signalling in diseases ...... 26 Physiological responses to BMP signalling ...... 27 Aberrant BMP signalling in diseases ...... 28 LKB1 signalling ...... 30 LKB1 tumour suppressor ...... 30 LKB1 structure and function ...... 31 LKB1 signalling pathways ...... 34 LKB1-AMPK pathways ...... 34 LKB1-SIK1 pathways ...... 36 LKB1-NUAK2 pathways ...... 37 The physiological roles of LKB1 ...... 38 LKB1 signalling cross-talk ...... 39 Aim of study ...... 41 Present Investigation ...... 42 Future perspectives ...... 45 References ...... 49 Acknowledgements ...... 64

Abbreviations

ACVR Activin receptor AMPK Adenosine monophosphate-activated kinase BAMBI BMP and activin membrane bound inhibitor bHLH Basic helix loop helix BMP Bone morphogenetic protein BMPR Bone morphogenetic protein receptor CBP CREB binding protein CDK Cyclin dependent kinase EGF Epidermal growth factor EMT Epithelial to mesenchymal transition FGF Fibroblast growth factor GSK Glycogen synthase kinase HDAC Histone deacetylase LIP1 LKB1-interacting protein 1 LKB1 Liver kinase B1 MAPK Mitogen activated protein kinase MMP Matrix metalloproteinase mTOR Mammalian target of rapamycin NUAK2 Sucrose non-fermenting (SNF1)-like kinase 2 PGC1α Peroxisome proliferator-activated receptor-gamma coactivator-1 alpha PI3K Phosphoinositide-3’-kinase Raptor Regulatory-associated protein of mTOR SARA Smad anchor for receptor activation SIK1 Salt inducible kinase 1 Smad Mother against decapentaplegic (Dpp) homolog Smurf Smad ubiquitination-related factor STRAD STE20-related kinase adaptor TAK1 TGFβ-activated kinase 1 TGFβ Transforming growth factor-β TGFβR Transforming growth factor-β receptor TNF Tumour necrosis factor TRAF6 TNF receptor-associated factor 6

Introduction

Communication is an absolute necessity for the formation and maintenance of an organised, dynamic and functional community that works towards a certain goal. Likewise, our body is built from different types of cells that communicate with each other to ensure proper function during foetal development and homeostasis of the adult organism. Cells secrete many different chemical signals or cytokines to communicate with each other. For example, the TGFβ/BMP family of ligands are crucial cytokines secreted by cells to regulate many cellular processes such as proliferation, death, differentiation and migration in both the embryo and adult tissue.

Cell communication or signalling pathways cross-regulate each other at many levels to ensure a very tight control of the cellular outcomes in response to cellular changes. When communication breaks down, cells can no longer achieve the desired responses to changes in the cellular environment. Accumulations of aberrant cell signalling have been shown to lead to diseases, with cancer being a prominent example.

This thesis work focuses on the cross regulation of the TGFβ/BMP pathways and LKB1 pathway. The LKB1 pathway is a tumour suppressor pathway that plays a major role in the regulation of energy metabolism to control cell proliferation during energetic stress. The TGFβ/BMP pathways are also involved in tumourigenesis and here we elucidate the novel mechanism on how these two signalling pathways cross-talk with each other and the resulting consequences on cellular physiology.

9 TGFβ/BMP Signalling

A brief history Transforming growth factor-β (TGFβ) was first identified in 1978 as a growth factor secreted by fibroblasts transformed by Moloney sarcoma virus, leading to their anchorage-independent growth in soft agar (de Larco and Todaro 1978). The TGFβ family of growth factors is now known to consist of nearly 40 related cytokines, subdivided into the TGFβ/Αctivin/Nodal subfamily (for example TGFβ1, TGFβ2, TGFβ3) and the bone morphogenetic protein (BMP)/growth and differentiation factor (GDF)/Mullerian inhibiting substance (MIS) subfamily (for example BMP2, BMP4, BMP7) (Massague, Seoane et al. 2005).

BMP was first discovered by Marshall R Urist in 1965 as a bone inducing substance contained in demineralised bone matrix. He found that when he implanted a demineralised bone into muscular tissues, a few weeks later cartilage and bone tissues with bone marrow were formed in the muscular tissue (Urist 1965). In 1971 he was able to isolate this bone inducing substance from the bone matrix with protein denaturants and called it ‘bone morphogenetic protein’ (Urist and Strates 1971).

The Smad-dependent signalling pathway The TGFβ/BMP signalling cascade is initiated by binding of ligand to the heterodimeric type I and type II serine/threonine kinase receptors (Figure 1). The TGFβ receptors consist of an N-terminal extracellular ligand binding domain, a transmembrane domain and a C-terminal kinase domain. Conven- tionally, the receptors for the TGFβ ligand subfamily are the type II receptors TβR-II, ActR-IIA and ActR-IIB in combination with the type I receptors TβR-I Activin receptor-like kinase 4 (ALK4), ALK5 and ALK7, whereas the BMP type II receptors are BMPRII, ActRIIA, ActRIIB in combination with the type I receptors ALK2, ALK3 and ALK6 (Shi and Massague 2003). Nonetheless, there are exceptions to this convention. Receptor combinations from both the TGFβ and BMP subfamilies can be used by cells to respond to certain ligands of the family (Daly, Randall et al. 2008).

10 TGFβ ligands bind strongly to the constitutively active type II receptor, which then recruits and phosphorylates the type I receptor at the Gly-Ser (GS) domain, which is the SGSGSG sequence N-terminal to the cytoplasmic kinase domain of the receptor (Shi and Massague 2003). BMP ligands, however, bind strongly to the type I receptor and have a low affinity for the type II receptor. Ligand bound type I BMP receptor then binds to the type II BMP receptor to be phosphorylated and activated (Kirsch, Sebald et al. 2000).

Activated type I receptors activate the receptor regulated-Smads (R-Smads) by phosphorylation on the Ser-X-Ser motif within the C-terminal MH2 domain. Phosphorylation of R-Smads enables their interaction with the common mediator –Smad (co-Smad), Smad4, and together they act as transcription factors. While there is only one common-Smad, there are five R-Smads, Smad2 and Smad3 for TGFβ signalling and Smad1, Smad5, and Smad8 for BMP signalling (Massague, Seoane et al. 2005).

The cytoplasmic Smad2 and Smad3 can bind to Smad anchor for receptor activation (SARA), a phospholipid binding protein that facilitates R-Smad recruitment and phosphorylation by the TGFβ type I receptor at the plasma membrane or in early endosomes (Tsukazaki, Chiang et al. 1998; Di Gug- lielmo, Le Roy et al. 2003). Similarly, Smad1 is bound to endofin as a Smad anchor for activation by the BMP receptors. Smad1 phosphorylation can occur on the early endosomes, once the ligand-bound receptors have been internalised (Shi, Chang et al. 2007), or on the plasma membrane (Hartung, Bitton-Worms et al. 2006).

TGFβ receptors activate predominantly Smad2 and Smad3 whereas the BMP receptors activate Smad1, Smad5 or Smad8. The specific choice of Smads to be phosphorylated by the type I receptors is mediated by the receptor’s L45 loop sequence that interacts with the L3 loop of the R-Smad (Feng and Derynck 1997; Lo, Chen et al. 1998). Nevertheless, there are reports which describe that TGF-β, through ALK5 and together with BMP type I receptors, activates Smad1/5/8 to mediate some of the TGF-β regulated cell processes such as cell migration and anchorage-independent growth of tumour cells (Daly, Randall et al. 2008; Liu, Schilling et al. 2009).

The signal transducing Smads have two domains, the N-terminal MH1 domain and the C-terminal MH2 domain connected by a linker region. The MH1 domain is the DNA-binding domain whereas the MH2 domain is the transactivation domain that interacts with the transcriptional co-activators p300 and CREBbinding protein (CBP). The C-terminal end of the MH2 domain of R-Smads contains the Ser-X-Ser motif where the phosphorylation by the type I receptors occurs. The MH2 domain of Smad4 lacks the Ser-X- Ser motif, thus it cannot be phosphorylated by the type I receptors. The MH2

11 domain is also responsible for the formation of homomeric or heteromeric Smad complexes and facilitates Smad nucleocytoplasmic shuttling. The R- Smads have a PPXY motif in their linker region which binds ubiquitin ligases of the Smurf and NEDD4 families, which regulate their degradation (Shi and Massagué 2003; Lin, Chen et al. 2008).

Receptor phosphorylated Smads are released from SARA protein complex and subsequently form a homomeric complex with each other or a heteromeric complex with Smad4. These complexes then translocate into the nucleus to regulate gene transcription in cooperation with other DNA binding proteins (Hill 2009).

Different mechanisms have been proposed on how the Smads are imported into the nucleus. One view suggests that the Smad complexes enter the nucleus via the nuclear transport receptors, importinα and importinβ. Direct binding of Smads to these importins are possible through the conserved lysine rich nuclear localisation signal (NLS)-like sequence in their MH1 domain. Another view suggests that the MH2 domain of the Smads interact directly with the nucleoporins CAN/Nup214 for nuclear import (Hill 2009).

In the nucleus, TGFβ-responsive Smad complexes bind to the minimal Smad binding element (SBE) DNA sequence (5’-AGAC- 3’) through the β-hairpin structure on their MH1 domain (Dennler, Itoh et al. 1998; Zawel, Dai et al. 1998). The BMP responsive Smads (Smad1/5/8) bind weakly to SBE and stronger to GC-rich boxes such as the GCCGnCGC sequence on some target gene promoters, such as Id1 (Kusanagi, Inoue et al. 2000; Korchynskyi and ten Dijke 2002; Lopez-Rovira, Chalaux et al. 2002). In endothelial cells, Smad1/5 also binds to target genes related to the Notch pathway such as the Notch ligand Jag1 to activate the neighbouring Notch signalling pathway (Morikawa, Koinuma et al. 2011).

In addition to Smad4, the R-Smads bind to many other transcription co- activators and co-repressors. A specific combination of these factors with the Smads and their expression and availability in different cell types determines the specific cellular response to TGFβ or BMP signalling (Massague, Seoane et al. 2005).

Smads interact with the general transcriptional co-activator p300/CBP leading to chromatin modification by histone acetylation, priming for gene transcriptional activation (Bannister and Kouzarides 1996; Ogryzko, Schiltz et al. 1996; Janknecht, Wells et al. 1998; Pouponnot, Jayaraman et al. 1998).

More specifically, Smads also bind to DNA binding proteins such as the forkhead family of transcription factors (for example FoxH1, FoxO1) and

12 Sp1 to activate the p21Cip1 gene expression that is required for the cytostatic effect of TGFβ (Chen, Rubock et al. 1996; Pardali, Kurisaki et al. 2000; Seoane, Le et al. 2004). The BMP responsive Smad1/5/8 interacts with transcription factors such as Runx2 and Menin to regulate gene expression that directs mesenchymal progenitor cells to osteoblastic differentiation (Zhang, Yasui et al. 2000; Sowa, Kaji et al. 2004).

On the contrary, Smads, in complex with transcriptional co-repressors, deac- tivate many genes in response to TGFβ or BMP stimulation. To regulate deacetylation, Smad3 binds to the histone deacetylase HDAC4 (Kang, Allis- ton et al. 2005). TGFβ signalling also results in a complex formation of the Smad3-E2F4/5-p107 co-repressor to repress c-Myc transcription. C-Myc repression is necessary for TGFβ-induced cell cycle arrest (Chen, Kang et al. 2002). An example of BMP-mediated gene repression is through the complex formed by the transcriptional repressor Nkx3.2 with Smad1/Smad4 and HDAC1, which regulates chondrocyte differentiation (Kim and Lassar 2003). TGFβ BMP BMP TGFβ

Plasma membrame β RI RII RI BMP β TGF BMPRII

TGF P P

Smad4

SARA

Smad2/3 P Endofin P

P R-Smad Smad4

P TF P TF Smad2/3 Smad4 Smad1/5/8 Smad4 CAGA BRE

Figure 1. TGFβ/BMP signalling – Smad dependent pathway. TGFβ/BMP ligands dimerise and bind to their corresponding receptors. Upon phopsphorylation (red circle), the receptor activates the R-Smads from the plasma membrane or endosomes. Phosphorylated R-Smads form complexes with Smad4 and translocate to the nucleus to regulate gene transcription, together with other transcription factors (TF) on the CAGA element for TGFβ signalling or BMP Response Element (BRE) for BMP signalling.

13 TGFβ Signalling Regulators The dynamic fine tuning of the intensity, spatial and timely regulations of TGFβ signalling are responsible for the broad yet targeted cellular processes during embryogenesis and adult tissue homeostasis. This is achieved through the work of many signalling regulators (Figure 2 and 3) and cross-talk with other signalling pathways (Figure 4).

Receptor regulation Regulators of ligand availability Extracellularly, there are ligand binding proteins which sequester away the ligands from receptor binding, such as α2 macroglobulin which binds to TGFβ, follistatin binds to activins and BMPs, whereas noggin, chordin and sclerostin bind to the BMPs as well (Katagiri, Suda et al. 2008).

Besides these extracellular signalling inhibitors, extracellular signalling en- hancers have been described too. This includes, for example, BMP1 and kielin/chordin-like protein (KCP). BMP1 is a metalloproteinase that cleaves chordin to release bound BMPs for binding to their receptors. KCP is a protein with 18 cysteine-rich domains that binds to BMP7 (similar to chordin) but, instead of inhibiting BMP7 as chordin does, KCP promotes BMP7 binding to the type I receptors (Katagiri, Suda et al. 2008).

Co-receptors Co-receptors are localised on the membrane and regulate ligand binding or activity of signalling receptors. For instance, β-glycan or the TGFβ type III receptor is important especially for the binding of TGFβ2 to the TGFβ type II receptor and signalling (Brown, Boyer et al. 1999). β-Glycan has been reported to affect BMP signalling as well by forming a complex with ALK3 and ALK6 which results in differential trafficking of these receptors and signalling responses (Lee, Kirkbride et al. 2009).

Glycosylphosphatidylinositol (GPI) membrane anchored proteins such as repulsive guidance molecule (RGMa) and RGMb act as co-receptors for BMP2 and BMP4 but not BMP7. RGMa and RGMb interact with BMP type I receptors and enhance BMP signalling (Halbrooks, Ding et al. 2007; Xia, Yu et al. 2007). However, this role is not yet clear as another report de- scribes RGMb as an inhibitor of BMP signalling in C2C12 myoblasts (Ka- nomata, Kokabu et al. 2009).

Additionally, BMP and activin membrane-bound inhibitor (BAMBI), a pseudoreceptor (a receptor without an intracellular kinase domain), attenuates BMP signalling by competing with the type I receptor and thus prevents

14 the proper assembly of the type I-type II receptor complex upon activation with a ligand (Onichtchouk, Chen et al. 1999).

Receptor phosphatases Although SARA facilitates the recruitment of Smads to the TGFβ receptors, SARA also plays a negative role in TGFβ signalling by presenting the protein phosphatase PP1c to dephosphorylate the TGFβ type I receptor (Shi, Sun et al. 2004). Similarly, on the BMP signalling side, endofin has been reported to play a similar role as SARA in that it recruits PP1c, leading to BMP receptor type I dephosphorylation (Zhang, Qiu et al. 2009).

Dullard is another phosphatase that was shown to dephosphorylate the BMP type I receptor ALK3 (Satow, Kurisaki et al. 2006). On top of that, Dullard promotes BMPRII ubiquitination and proteasomal degradation via the caveolar pathway and in a phosphatase activity-dependent manner during Xenopus neural development (Satow, Kurisaki et al. 2006).

Receptor post-translational regulation There is another class of Smads whose expression are induced by TGFβ/BMP signalling as part of a negative feedback response. These Smads are called the inhibitory Smads (I-Smads), and include Smad6 and Smad7. Smad7 has a wider scope in that it inhibits both TGFβ and BMP signalling, whereas Smad6 functions more specifically towards inhibition of BMP signalling (Goto, Kamiya et al. 2007).

The I-Smads support receptor down-regulation by recruiting E3 ubiquitin ligases to the receptors. Smad7 associates with HECT type E3 ubiquitin ligases such as Smurf1, Smurf2, WWP1 and NEDD4-2. Receptors bound to I- Smads and Smurf are internalised via the lipid raft-caveolae pathway to be targeted for polyubiquitination. At the receptor complex, the E3 ligases at- tach ubiquitin moieties to receptor lysine residues and tag the receptor complex for proteasomal degradation (Itoh and ten Dijke 2007).

TGFβ/BMP signalling also provides another level of regulation on receptor degradation by inducing SIK1 which cooperates with Smad7 and Smurf2 to enhance TGFβ type I receptor ALK5 degradation (Kowanetz, Lonn et al. 2008; Lonn, Vanlandewijck et al. 2012).

Ubiquitination is a reversible process. Ubiquitin-specific protease (USP) 4 interacts with TGFβ type I receptor and deubiquitylates the receptor to stabilise the receptor and induce TGFβ signalling. Akt or protein kinase B (PKB) promotes USP4 activity by inducing its cytoplasmic and membrane localisation and protein stability (Zhang, Zhou et al. 2012).

15 Another important post-translational regulation of TGFβ receptor comes from sumoylation by Ubc9. SUMO, a ubiquitin like molecule, is attached to its target proteins via lysine modification by sumo-conjugating enzymes. Sumoylated TGFβ receptor facilitates Smad3 recruitment and phosphorylation, thus positively regulating TGFβ signalling (Kang, Saunier et al. 2008).

Finally, post translational modification by phosphorylation has been reported to occur on the BMP type II receptor (BMPRII) tail by the cyclic guanosine 3’, 5’-monophosphate (cGMP)-dependent kinase I (cGKI) in the absence of BMP2 signalling. With BMP2 stimulation, cGKI interacts with active R- Smad complexes and promotes the translocation of such complexes to the nucleus for further gene regulation (Schwappacher, Weiske et al. 2009).

Another protein that also binds to the tail domain of BMPRII is the tribbles- like protein 3 (Trb3). Trb3 interaction with the BMPRII is disrupted by BMP4 stimulation and ‘released’ Trb3 then interacts with Smurf1 to promote Smurf1 degradation via the ubiquitin-proteasome pathway. Degrada- tion of Smurf1 leads to increased BMP signalling (Chan, Nguyen et al. 2007).

Regulation of Smad activation by the receptor Besides playing a role in promoting receptor degradation, the direct interaction of the inhibitory Smads with the type I receptors competes with R- Smads binding to the receptor and therefore hinders the R-Smads from being activated (Hayashi, Abdollah et al. 1997; Imamura, Takase et al. 1997; Na- kao, Afrakhte et al. 1997).

Two proteins that support this feature of Smad7 are the serine/threonine kinase receptor-associated protein (STRAP) and BAMBI. STRAP recruits Smad7 and stabilises Smad7 interaction with the TGFβ type I receptor, thus promoting the Smad7 function of blocking Smad2/3 binding to the receptor and suppressing TGFβ signalling (Datta and Moses 2000). Similarly, BAMBI forms a complex with Smad7 and the TGFβ type I receptor ALK5 to impair Smad3-ALK5 interaction and prevent Smad3 phosphorylation (Yan, Lin et al. 2009).

Moreover, FKBP12 binds to the TGFβ type I receptor on its GS domain when the receptor is inactive in the absence of ligand binding. This prevents ligand independent R-Smads binding and activation by the receptor (Huse, Chen et al. 1999).

Furthermore, the transmembrane prostate androgen-induced RNA (TME- PAI) is a direct target gene of TGFβ signalling that provides negative feedback to the signalling pathway. TMEPAI contains a smad-interacting motif

16 (SIM) that binds Smad2/3 and prevents them from SARA association and phosphorylation by the type I receptors (Watanabe, Itoh et al. 2010).

Ligand availability: BMP1 KCP TGFβ α-macroglobulin Plasma membrane co-receptors: chordin TGFβ Β-Glycan noggin RGMa/RGMb sclerostin Pseudoreceptor BAMBI RI

Receptor degradation: RII I-Smads P Receptor phosphatases: SIK1 PP1C Ubiquitin ligases Dullard USP4 deubiquitinase

Other Receptor interacting proteins : Ubc9 sumo ligase cGKI Trb3 FKBP12 STRAP

Figure 2. Receptor regulators in TGFβ/BMP signaling.

Smads regulation Regulation of Smad transcriptional activity Ski and SnoN are nuclear proteins that are negative regulators of TGFβ signalling. Ski and SnoN interact with Smad2/3 and Smad4 leading to disrup- tion of this active heteromeric Smad complex and inhibition of Smad transcriptional activity (Stroschein, Wang et al. 1999; Deheuninck and Luo 2009). In addition, Ski and SnoN further impair activation of TGFβ target genes by inhibition of R-Smad binding to the transcriptional co-activators p300/CBP and recruitment of transcriptional co-repressors such as HDAC and n-CORs to the Smad-regulated promoters (Deheuninck and Luo 2009).

Interestingly, in BMP signalling, only Ski could bind to Smad1/5 but not SnoN. As in the case of TGFβ signalling, the interaction of Ski with Smad1/5 results in negative regulation of BMP-induced gene transcription and regulation of neural cell fate in Xenopus development (Wang, Mariani et al. 2000). Similarly, Nanog competes with p300/CBP for binding to Smad1, thus suppressing BMP signalling and BMP-induced differentiation in mouse embryonic stem cells (Suzuki, Raya et al. 2006).

17 Smad post-translational regulation In addition to the activating phosphorylation of the C-terminal MH2 domain by the type I receptors, Smad activity is regulated by a number of other post- translational modifications such as phosphorylation, poly-ADP-ribosylation (PARylation), ubiquitination, sumoylation and acetylation.

Smad4 is phosphorylated on its Thr77 residue within the MH1 domain by LKB1 (Moren, Raja et al. 2010). This phosphorylation occurs very close to the β-hairpin structure that is responsible for Smad4 binding to DNA, thus disrupting Smad4 binding to DNA and its transcriptional activity upon TGFβ and BMP stimulation (Moren, Raja et al. 2010). On the other hand, Smad4 phosphorylation by Erk MAPK on its linker region may enhance Smad4 transcriptional activity by stabilising the heteromeric Smad complex with p300/CBP (Roelen, Cohen et al. 2003).

The transcriptional activity of Smad3 and Smad4 are inhibited by the interaction with poly-ADP ribose polymerase 1 (PARP1). Smad3 and Smad4 are poly-ADP ribosylated and this modification results in dissociation of Smad complexes from target gene promoters and suppression of TGFβ-induced EMT (Lonn, van der Heide et al. 2010).

Smad ubiquitination is mediated by E3 ligases such as Smurf1, Smurf2, WWP1, Nedd4-2, Jab1, CHIP, SCF, AIP4/Itch, Arkadia and Ectodermin. While all these ubiquitin ligases are associated with polyubiquitination and targeting the Smads for proteasomal degradation, several of them have been reported to promote Smad function before termination of signalling by degradation. For example, ubiquitination by AIP4/Itch enhances Smad2 C- terminal phosphorylation by association with type I TGFβ receptor and ubiquitination by Arkadia is required for R-Smads to activate transcription (Lonn, Moren et al. 2009).

Phosphorylation of the linker domain in R-Smads is related to its ubiquitination. In BMP signalling, Smad1 linker phosphorylation is mediated by CDK8/9 which is followed by YAP association and increased Smad1 transcriptional activity while recruiting Smurf1 to promote Smad1 ubiquitination and protein turnover as well (Alarcon, Zaromytidou et al. 2009).

Moreover, CDK8/9 phosphorylates Smad3 on its linker domain leading to recruitment of NEDD4-2 and Smad3 polyubiquitination and degradation. Besides CDK8/9, ERK2, CDK2/4 and GSK3β were also described to phosphorylate R-Smads on the linker domain for recruitment of E3 ligases and polyubiquitination leading to proteasomal degradation (Lonn, Moren et al. 2009).

Post-translational modifications are reversible events. Dephosphorylation of the Smad2/3 linker domain by small C-terminal domain phosphatases (SCP) leads to enhanced TGFβ signalling (Sapkota, Knockaert et al. 2006). On the other hand, SCP attenuates BMP signalling by dephosphorylating Smad1 C- terminal phosphorylation (Knockaert, Sapkota et al. 2006).

Likewise, PPM1A/PP2Cα dephosphorylates the C-terminal phosphorylation sites of Smad2 and Smad3 in the nucleus and stimulates the export of Smads back to the cytoplasm (Lin, Duan et al. 2006). Recently, myotubularin- related protein-4 (MTMR4) was also reported to function as a phosphatase for activated Smad2/3 in the endosomes (Yu, Pan et al. 2010).

More negative regulation comes from sumoylation of Smad3 and Smad4 by the sumo ligases protein inhibitor of activated STAT (PIASy). Sumoylation of Smad3 and Smad4 suppresses their transcriptional activity (Long, Matsu- ura et al. 2003; Long, Wang et al. 2004). However, there are reports describing the roles of Smad4 sumoylation for Smad4 stability and nuclear localisation (Lin, Liang et al. 2003a; Lin, Liang et al. 2003b).

While most of the post-translational modifications described so far result in negative regulation of Smads in TGFβ/BMP signalling, acetylation potentiates Smad function. In the nucleus, p300/CBP acetylates the R-Smads. Smad2 is acetylated on its MH1 domain whereas Smad3 on its MH2 domain and both acetylations promote the transcriptional activity of Smads in response to TGFβ signalling (Simonsson, Kanduri et al. 2006; Inoue, Itoh et al. 2007; Tu and Luo 2007).

Smad7 is ubiquitinated by Smurf1/2 and Arkadia leading to its proteasomal degradation (Kavsak, Rasmussen et al. 2000; Ebisawa, Fukuchi et al. 2001; Koinuma, Shinozaki et al. 2003). However, acetylation of Smad7 by p300 at the same residues that could be ubiquitinated leads to inhibition of Smad7 ubiquitination and degradation (Gronroos, Hellman et al. 2002). This process is reversible by HDAC which removes the acetyl moiety from Smad7 so it can be ubiquitinated and degraded again (Simonsson, Heldin et al. 2005).

Inhibit R-Smad activation by the receptor: I-Smads P TMEPAI Co- R-Smad Smad Inhibit Smad binding to transcription co-activators: Ski/SnoN Nanog

P Co- TF R-Smad Smad DNA binding element

Post translational regulators affecting Smads transcriptional activity: LKB1 ERK MAPK phosphorylation CDK8/9 GSK3β SCP PPM1A/PP2Cα de-phosphorylation MTMR4 PARP1 parylation PIASy sumoylation P300/CBP acetylation

Figure 3. Smad regulators in TGFβ/BMP signaling.

Signalling cross-talk through the non-Smad signalling pathways Although the canonical TGFβ/BMP signalling cascade involves Smad activation, the non-Smad signalling pathway is equally important in exerting TGFβ/BMP regulated cell processes. TGFβ signalling cross-talks with many signalling pathways, however, here an emphasis will be placed on the cross- talk mechanism between TGFβ signalling with the mitogen-activated protein kinase (MAPK) and the phosphoinositide-3-kinase (PI3K) - mTOR pathways (Figure 4).

TGFβ signalling and the MAPK pathways The ERK, JNK and p38 MAPK pathways are activated by growth factors such as epidermal growth factor (EGF), fibroblast growth factor (FGF), insu- lin, platelet-derived growth factor (PDGF) and TGFβ/BMP. TGFβ/BMP- activated MAPK mediates linker phosphorylation of R-Smads and Smad4 as

20 mentioned previously and additionally, MAPKs also phosphorylate transcription factors that cooperate with Smads such as the AP-1 family, including Jun, Fos, Maf and ATF, to regulate gene transcriptions (Guo and Wang 2009).

How does TGFβ or BMP signalling activate MAPKs? The TGFβ type I receptor binds Smad7 whereas the BMP type I receptor binds an X-linked inhibitor of apoptosis/TAK1 binding protein (XIAP/TAB) complex. Both Smad7 and XIAP/TAB complex recruit TGFβ-activated kinase (TAK)1 to the receptors (Moustakas and Heldin 2005).

A later finding indicates that TNF receptor-associated factor (TRAF)6 ubiquitin ligase constitutively interacts with the TGFβ type I receptor and upon ligand binding and receptor dimerisation, receptor bound TRAF6 dimerises and becomes activated by auto-ubiquitination. Activated TRAF6 mediates TAK1 Lys-63 ubiquitination and activation (Sorrentino, Thakur et al. 2008). Activated TAK1 phosphorylates MKK3/4/7 which in turn phosphorylates p38 and JNK leading to apoptosis (Sorrentino, Thakur et al. 2008).

TRAF6-mediated Lys-63 ubiquitination of the TGFβ type I receptor causes the receptor cleavage by tumour necrosis factor (TNF) alpha-converting enzyme (TACE), in a protein kinase C (PKC)ζ - dependent manner. Finally, cleaved receptor translocates to the nucleus to cooperate with p300 for activation of cell invasion-related genes such as Snail and matrix metalloproteinase (MMP) 2 (Mu, Sundar et al. 2011).

TGFβ signalling and the PI3K/Akt – mTOR pathway The branch of PI3K/Akt pathway that is activated by TGFβ does not support TGFβ-induced apoptosis but instead promotes cell survival towards epithelial to mesenchymal transition and cell invasion (Conery, Cao et al. 2004; Lamouille and Derynck 2007; Lamouille, Connolly et al. 2012).

TGFβ signalling activates the mTOR complex 1 pathway via activation of PI3K and Akt. Activated mTOR kinase results in phosphorylation of the S6 kinase and eukaryotic initiation factor 4E-binding protein 1, both of which promote protein synthesis leading to an increase in cell size, EMT and cell invasion (Lamouille and Derynck 2007).

Recently, TGFβ signalling was reported to induce mTOR complex 2 kinase activity to positively regulate actin reorganisation and RhoA activation during TGFβ-induced EMT. mTORC2 activity is also required for expression of MMP9 and TGFβ-dependent cell invasion (Lamouille, Connolly et al. 2012).

BMP TGFβ

TGFβ BMP Plasma membrame β RI RII RI TRAF6 BMP β TGF P BMPRII TGF P Smad7 P

P P PI3K/Akt MKK 3/4/7

P P P P mTORC mTORC 1 2 p38 JNK

S6K & 4E-BP1 activation Actin organisation Protein synthesis RhoA activation Apoptosis MMP9 expression

EMT Invasion

Figure 4. TGFβ/BMP signalling – non-Smad signalling pathways. TRAF6 and TAK1 are both activated by ubiquitination (orange circle chain). Activated TAK1 phosphorylates MKKs and MAPKs to induce apoptosis. TGFβ signalling induces PI3K/Akt pathway to activate the mTORC1 and mTORC2 pathways, both leading to EMT and cell invasion.

Physiological responses to TGFβ signalling TGFβ signalling, through the Smad or non-Smad pathway, is well known to regulate cellular processes such as cell proliferation, cell differentiation, epithelial to mesenchymal transition, as well as cell migration and invasion. The effects of TGFβ signalling are very complex and context-dependent, whether it is the intracellular or extracellular environmental factors.

Epithelial cell growth The cytostatic effect of TGFβ signalling in epithelial cells is very well studied. TGFβ signalling mediated by Smad2/3/4 represses the expression of c- Myc, Id1, Id2 and Id3 and activates the expression of cyclin dependent

22 kinase inhibitor (CDK) inhibitors such as p15, p21, and p57 to induce cell cycle arrest (Siegel and Massague 2003)

The activity of CDKs is crucial for cell cycle progression from the G1 phase to the S phase or G2 phase to M phase. Inhibitors of CDKs therefore provide checkpoints for proper balance during the process of cell division. C-Myc is a repressor of p15 and p21 transcription and thus the TGFβ-induced c-Myc repression further enhances the expression of p15/p21 required for cytostasis (Claassen and Hann 2000; Seoane, Pouponnot et al. 2001; Staller, Peukert et al. 2001).

The FoxO family of transcription factors cooperate with TGFβ-induced Smad3/4 to activate p21 transcription. However, activated PI3K signalling leads to Akt-mediated phosphorylation of FoxO transcription factors and their translocation from the nucleus to the cytoplasm. Therefore, excessive PI3K-Akt signalling suppresses TGFβ-induced p21 transcription (Seoane, Le et al. 2004). Smad proteins also cooperate with both Sp1 and p53 transcription factors to up-regulate p21 levels for TGFβ-mediated growth inhibition (Moustakas and Kardassis 1998; Pardali, Kurisaki et al. 2000; Corde- nonsi, Dupont et al. 2003)

The inhibitors of DNA binding/differentiation (Id) proteins are suppressed by TGFβ signalling as they promote cell proliferation through negative regulation of the gene transcription that is required for epithelial differentiation. Id proteins bind and inhibit the function of basic-helix-loop-helix transcription factors that induce differentiation. This indirectly links cell differentiation to inhibition of cell proliferation (Ruzinova and Benezra 2003).

Importantly, epithelial cell growth is also influenced by TGFβ signalling in its neighbouring stromal cells. This was demonstrated in a couple of studies in murine mammary and prostate glands. In the mammary gland, dominant negative expression of type II TGFβ receptor in the stroma results in hyperplasia of the adjacent mammary epithelium (Joseph, Gorska et al. 1999). In the prostate gland, targeted loss of type II TGFβ receptor expression in stromal fibroblasts leads to its over-proliferation as well as hyperplasia of the neighbouring prostate epithelial cells (Bhowmick, Chytil et al. 2004).

Epithelial growth inhibition by the surrounding stromal cells may be mediated by TGFβ signalling that suppresses the production of growth factors such as TGFα or hepatocyte growth factor (HGF), which promote epithelial cell growth (Joseph, Gorska et al. 1999; Bhowmick, Chytil et al. 2004).

Another aspect of cell growth regulation is the process of programmed cell death or apoptosis. TGFβ signalling via activation of the Smad pathway

23 induces the expression of the transcription factor TIEG1 (TGFβ−inducible early response genes), DAPK (death-associated protein kinase) and SHIP (SH2 domain-containing inositol-5-phosphatase) to promote apoptosis (Ta- chibana, Imoto et al. 1997; Jang, Chen et al. 2002; Valderrama-Carvajal, Cocolakis et al. 2002).

A transcriptional-independent mechanism involves the protein ARTS (apoptosis-related protein in TGFβ signalling) and Daxx (death-associated protein) (Larisch, Yi et al. 2000; Perlman, Schiemann et al. 2001). ARTS is a mitochondrial protein which activates caspase 3 to induce apoptosis in response to TGFβ (Larisch, Yi et al. 2000). Daxx, on the other hand, is an adapter protein that activates JNK (Jun amino-terminal kinase) to induce apoptosis (Perlman, Schiemann et al. 2001). In addition, as mentioned in the previous section, TGFβ−activated TRAF6-TAK1-p38 pathway also activates JNK to promote apoptosis (Sorrentino, Thakur et al. 2008).

Lung development Lung morphogenesis is one example of TGFβ control in epithelial cell proliferation and differentiation during development. In vitro experiments suggest that TGFβ1 and TGFβ2 inhibit differentiation of mouse embryonic lung cells and branching morphogenesis (Zhou, Dey et al. 1996; Beers, Solarin et al. 1998). In the same direction, mice with suppressed TGFβ signalling because of expression of a dominant negative form of TGFβ type II receptor in the lung epithelium showed a more pronounced branching morphogenesis due to increased cell proliferation and differentiation (Zhao, Sime et al. 1998).

In addition to TGFβ, a concentration gradient of BMP4 during lung morphogenesis controls the differentiation of lung epithelium to create the proper proximal-distal axis of the airways (Bellusci, Henderson et al. 1996). Exposure to high BMP4 signalling inhibits proximal cell differentiation in the terminal buds of the lungs (Weaver, Yingling et al. 1999).

Bone development Bone development is a dynamic process that requires a fine balance between bone deposition and bone resorption. In this setting, TGFβ plays a critical role by regulating osteoclast differentiation and the cross-talk between osteoblasts and osteoclasts during bone development.

TGFβ positively regulates osteoclast differentiation and activity in a time and dose dependent manner. Osteoclasts produce latent TGFβ and are them- selves responsive to TGFβ stimulation. TGFβ signalling potentiates osteoclast differentiation, and this may involve TGFβ−dependent induction of suppressor of cytokine signalling (SOCS) expression and tumour necrosis

24 factor receptor (TNFR). Osteoclast activity involving MMP2 and MMP9 in the bone matrix activates latent TGFβ and further promotes TGFβ signalling which eventually results in inhibition of bone resorption as a negative feedback loop. Osteoblasts express factors that are required for osteoclastogene- sis, such as tumour necrosis factor (ligand for TNFR), macrophage-colony stimulating factor (M-CSF) and osteoprotegerin (OPG), which are induced by TGFβ to provide cross regulation of osteoclasts by osteoblasts. However, high levels of TGFβ result in decreased expression of these factors as a negative feedback to the system (Alliston, Piek et al. 2008).

Blood vessels and heart development Blood vessels are generated through vasculogenesis and angiogenesis. Vas- culogenesis refers to the formation of new vessels from proliferating progenitor cells, whereas angiogenesis is the sprouting of new vessels from the pre-existing blood vessels. Nevertheless, both processes are dependent on endothelial cell function and involve vascular smooth muscle cell recruitment which supports the integrity of the vessel wall (Goumans, Liu et al. 2009).

Interestingly, TGFβ signalling through the type I receptors ALK5 and ALK1 gives opposite responses in endothelial cells. ALK5 signalling activates Smad2/3 to inhibit endothelial cell proliferation, migration and vessel formation. In contrast, ALK1 signalling activates Smad1/5 to promote endothelial cell proliferation, migration and vessel formation (Goumans, Valdimarsdottir et al. 2002; Goumans, Valdimarsdottir et al. 2003). Nevertheless, ALK5 and ALK1 signalling are not independent of each other and the balance between these two pathways is needed for proper blood vessel formation or homeostasis (Oh, Seki et al. 2000; Goumans, Valdimarsdottir et al. 2002; Gou- mans, Valdimarsdottir et al. 2003).

Endothelial cells cooperate with vascular smooth muscle cells to form functional vessels. TGFβ signalling is an important mediator that provides ade- quate communication between these two cell types during vessel formation. Endothelial cells produce TGFβ ligand to induce smooth muscle cell differentiation as well as to stimulate VEGF production, which regulates the growth and differentiation of both endothelial and smooth muscle cells (Ya- mamoto, Kozawa et al. 2001; Ding, Darland et al. 2004).

Moreover, TGFβ signalling promotes endothelial to mesenchymal transition (Endo-MT) which is a process whereby endothelial cells acquire a mesenchymal, smooth muscle cell-like phenotype (Kokudo, Suzuki et al. 2008). In the same line, TGFβ induces the process of epithelial to mesenchymal transition (EMT) during heart development. EMT enables epithelial cells to trans- form by losing their epithelial cell-cell junctions and becoming mesenchy-

25 mal-like, acquiring the migratory and invasive behaviour that is essential during embryogenesis (Nakajima, Yamagishi et al. 2000).

Aberrant TGFβ signalling in diseases Due to the widespread involvement of TGFβ signalling in many cellular processes in many tissues, aberrant TGFβ signalling has been implicated in the development of many diseases. Examples of vascular disease and cancer are given in this chapter.

Genetic vascular disease Inactivating mutations of ALK1 have been associated with an autosomal dominant disease called hereditary hemorrhagic telangiectasia (HHT). Lack of TGFβ signalling via ALK1 affects ALK5 signalling as well, which overall causes impaired vascular network formation. HHT is characterised by haemorrhages from vascular lesions (obvious in the skin and mucosa) and abnormal lung, brain and liver formation (Fernandez, Sanz-Rodriguez et al. 2006).

Cancer and metastasis It is generally accepted that TGFβ signalling plays a dual role in cancer development, as a tumour suppressor in the early stages and as a tumour promoter as the cancer progresses and metastasises.

Inactivating mutations or allelic loss of the TGFβ receptor type I or II and Smad4 have been observed in many tumour types confirming the tumour suppressive role of TGFβ signalling. Conversely, in the progression of many types of tumours, such as breast cancer, prostate cancer and colon cancer, TGFβ signalling remains active and is positively correlated with metastasis (Padua and Massague 2009).

As cells acquire many genetic changes while becoming malignant, cancer cells could by-pass the TGFβ−induced cytostatic or apoptotic responses by, for example, loss-of-function mutations in TGFβ-Smad cooperating transcription factors that are essential for induction of genes required for cytostasis or apoptosis. These cells could then go further into cancer development, taking advantage of other TGFβ-induced responses that are pro- tumourigenic (Padua and Massague 2009).

To promote metastasis, TGFβ signalling suppresses the immune response by inhibition of T lymphocytes which are able to recognise cancer cells and eliminate them (Gorelik and Flavell 2001). TGFβ signalling also promotes angiogenesis which facilitates oxygen and nutrient delivery to cancer cells.

26 This is achieved through induction of vascular endothelial growth factor (VEGF) and connective tissue growth factor (CTGF) (Sanchez-Elsner, Bo- tella et al. 2001; Shimo, Nakanishi et al. 2001). Moreover, TGFβ is a potent inducer of EMT to support cancer cell migration and invasion into new tissue niches. This process is often characterised by a loss of E-cadherin protein that marks cell-cell adhesions and a gain of mesenchymal proteins such as fibronectin and actin stress fiber formation that promote cell migration (Heldin, Vanlandewijck et al. 2012).

TGFβ plays a prominent role in the metastases of breast or prostate cancers to the bone and lung. In bone metastasis, as mentioned in the previous section, TGFβ signalling promotes osteoclast differentiation and their function to degrade bone matrix and create metastatic bone lesions. In lung metastasis, TGFβ signalling induces angiopoeitin-like 4 (ANGPTL4) expression that impairs vascular cell junctions and enhances vascular permeability so that breast cancer cells can disseminate into the lung (Padua, Zhang et al. 2008).

Physiological responses to BMP signalling Like TGFβ, BMPs play important roles in embryogenesis and tissue maintenance through controlling various cellular processes. However, here an emphasis is placed on the role of BMPs in regulating cell differentiation.

Bone differentiation The most well known function of BMP is to induce differentiation of mesenchymal progenitor cells into mature osteoblasts or chondrocytes, thus con- tributing to the formation of bone and cartilage. The bone-inducing ability of BMP is dependent on the activation of the type I receptors and also activation of Smad1/5/8 (Akiyama, Katagiri et al. 1997; Chen, Ji et al. 1998; Fujii, Takeda et al. 1999). However, non-Smad pathways activated by the BMP receptors, such as the MAP kinase pathways, have also been shown to lead to osteoblast differentiation in vitro (Vinals, Lopez-Rovira et al. 2002; Guicheux, Lemonnier et al. 2003).

BMP signalling increases the expression of transcription factors such as Runx2, Dlx2, Dlx5 and osterix to promote osteoblast or chondrocyte differentiation (Ducy, Zhang et al. 1997; Miyama, Yamada et al. 1999; Naka- shima, Zhou et al. 2002). Runx2 is a critical transcription factor required for bone formation, as suggested by Runx2 knock-out mice which do not have any bone at all (Otto, Thornell et al. 1997). BMP activates Smad signalling to induce the expression of Dlx5 which in turn induces Runx2 expression.

27 Runx2 interacts with Smad1/5 to regulate transcription of genes required for osteoblast differentiation (Zhang, Yasui et al. 2000).

Muscle differentiation In muscle tissue, BMPs play an antagonistic role by blocking myogenesis through inhibition of myoblast differentiation. Myoblast differentiation is a process activated by bHLH transcription factors such as MyoD and myogenin and this process is inhibited by BMPs partly because of the induction of Id1 protein, which represses MyoD/myogenin-induced gene transcription (Katagiri, Akiyama et al. 1997; Katagiri, Imada et al. 2002).

Neural cell differentiation BMPs such as BMP2 also inhibit neurogenesis of neuroepithelial cells by inducing Id1 and Id3, leading to suppression of transcription activated by neurogenic bHLH transcription factors such as neurogenin, NeuroD and Mash1 (Nakashima, Takizawa et al. 2001). Nonetheless, BMP2, together with leukemia inhibitory factor (LIF) and p300, promotes astrocytic differentiation of neural stem cells and therefore suggests that BMPs regulate the fate of neural stem cells (Nakashima, Yanagisawa et al. 1999).

Adipocyte differentiation BMP7 promotes the differentiation of brown pre-adipocytes and mesenchymal progenitor cells towards brown adipocyte cell fate. This is achieved through induction of brown adipogenic factors such as PR-domain containing 16 (PRDM16), peroxisome proliferator-activated receptor-γ (PPARγ), PPARγ coactivator-1α (PGC1-α), uncoupling protein 1 (UCP1), CCAAT enhancer binding proteins (CEBPs) and of mitochondrial biogenesis. Mice overexpressing BMP7 have increased brown fat tissue, increased energy expenditure and a lower weight gain, suggesting a possible use of BMP7 for the treatment of obesity (Tseng, Kokkotou et al. 2008).

Aberrant BMP signalling in diseases Some genetic mutations in the BMP signalling pathway have been described including mutations in the BMPR2 gene, BMPRIa/ALK3, ALK2/ACVR1 and also SMAD4. Mutations in the BMPR2 gene have been linked to the familial or sporadic cases of pulmonary arterial hypertension (Deng, Morse et al. 2000).

BMP signalling promotes the survival of pulmonary arterial endothelial cells and suppresses proliferation of pulmonary artery smooth muscle cells. BMP signalling defects caused by the BMPR2 mutations may have caused the opposite events, a decrease in the survival of the endothelial cells accompa-

28 nied by an increase in smooth muscle cells, which leads to obstruction in the small pulmonary arteries (Yang, Long et al. 2005; Teichert-Kuliszewska, Kutryk et al. 2006).

Loss-of-function mutations in ALK3 and Smad4 are linked with the juvenile polyposis syndrome (JPS), an autosomal dominant genetic disorder that is characterised by gastrointestinal hamartomatous polyps and a significantly elevated risk for developing cancer. About 50-60% cases of JPS are associated with ALK3/Smad4 mutations (Brosens, Langeveld et al. 2011). This may mean that the BMP signalling pathway plays a role in regulating proper cell growth in the gastrointestinal tract (Howe, Bair et al. 2001). It has also been reported that the BMP pathway is inactivated in many sporadic colorectal cancers (Kodach, Wiercinska et al. 2008).

Gain of function mutations in the ALK2 gene result in a syndrome called fibrodysplasia ossificans progressiva (FOP). These mutations occur in the glycine-serine rich domain of the BMP type I receptor ALK2, rendering it constitutively active (Shore, Xu et al. 2006). FOP patients suffer from pro- gressive postnatal ossification of soft tissues. A small molecule inhibitor for the BMP type I receptor (LDN-193189) has been developed that may be useful for the treatment of FOP patients (Yu, Deng et al. 2008).

Similar to TGFβ, BMPs seem to play dual roles in different cancers as well. In colorectal cancer and glioblastoma, BMPs may act as tumour suppressors (Piccirillo, Reynolds et al. 2006; Kodach, Wiercinska et al. 2008), whereas in breast cancer, pancreatic cancer, prostate cancer and bone cancer, BMPs may have pro-tumourigenic and pro-invasive/metastatic functions (Yoshi- kawa, Nakase et al. 2004; Feeley, Gamradt et al. 2005; Katsuno, Hanyu et al. 2008; Gordon, Kirkbride et al. 2009).

29 LKB1 signalling

LKB1 tumour suppressor Liver kinase B1 (LKB1) or serine/threonine kinase 11 (STK11) is a tumour suppressor kinase that regulates cellular metabolism, cell death, cell proliferation and cell polarity (Figure 5) (Alessi, Sakamoto et al. 2006).

Inactivating mutations of the LKB1 gene have been linked to the Peutz Jegh- ers syndrome (PJS). PJS was first described by Dr Johannes Peutz in 1922 and further by Dr. Harold Jeghers in 1949, but the genetic link to LKB1 was made in 1998 by Akseli Hemminki, a graduate student from Helsinki Uni- versity (Hemminki, Markie et al. 1998). PJS is an autosomal dominant disease marked by gastrointestinal hamartomatous polyps, mucocutaneous pigmentation and a higher risk of developing malignancies. About 93% of PJS patients get cancer by the age of 43, both gastrointestinal related or un- related cancers (Alessi, Sakamoto et al. 2006).

Somatic mutations leading to the loss of function of LKB1 are rarely de- tected; however, they have been reported in some sporadic cancers (Kata- jisto, Vallenius et al. 2007). A prominent example is from non-small cell lung carcinoma where 15-35% of cases harbour LKB1 mutations (Ji, Ramsey et al. 2007) with the addition of 20% of cervical carcinoma cases (Wingo, Gallardo et al. 2009) and a small fraction in malignant melanoma (Guldberg, thor Straten et al. 1999; Rowan, Bataille et al. 1999).

Intriguingly, although the loss of LKB1 clearly leads to a loss of control over cell growth and tumour development, the loss of LKB1 has also been shown to give rise to resistance from oncogenic transformations such as Ras. This finding places LKB1 as a context-dependent tumour suppressor, suggesting the possibility that loss of function of LKB1 may happen early during tumourigenesis but LKB1 may be beneficial at later stages of malignant progression (Bardeesy, Sinha et al. 2002).

30 LKB1 structure and function

LKB1 post-translational modifications Human LKB1 consists of 433 residues. The N-terminal domain of LKB1 contains a nuclear localisation signal; the kinase domain contains two auto- phosphorylation sites and the C-terminal domain has several phosphorylation sites by upstream kinases. LKB1 is also farnesylated on the C-terminal domain at C433 (Alessi, Sakamoto et al. 2006). Very little is known about the function of this modification but it has been suggested that it is important for anchoring LKB1 to the plasma membrane where it can then regulate cell polarity (Martin and St Johnston 2003).

In addition, LKB1 is acetylated on Lys48 and this acetylation inhibits its function. A protein deacetylase SIRT1 removes this acetylation and promotes LKB1 function by increasing its cytoplasmic localisation and binding to its activating partner, STE20-related kinase adaptor (STRAD). Deacetyla- tion of LKB1 was shown to increase LKB1 activity, indicated by an increase in MARK1 (Microtubule affinity-regulating kinase) and AMPK (AMP- regulated kinase) phosphorylation levels. The AMPK and MARK kinases are known to act downstream of LKB1 (Lan, Cacicedo et al. 2008).

The C-terminal domain of LKB1 is very important because naturally occur- ring mutations in this domain lead to a decrease in activation of AMPK downstream of LKB1, although these mutations do not affect LKB1 kinase activity per se. C-terminal mutations also impair the function of LKB1 to establish and maintain intestinal cell polarity (Forcet, Etienne-Manneville et al. 2005).

A number of upstream kinases phosphorylate LKB1 at its C-terminal domain. P90 RSK and cAMP-dependent protein kinase A phosphorylate LKB1 at Ser431 and this phosphorylation is important for LKB1 to suppress cell growth (Sapkota, Kieloch et al. 2001). In response to ONOO- (peroxynitrite), protein kinase C ζ also phosphorylates human LKB1 at Ser428 (Ser431 in mice) to promote its nuclear export and inhibit signalling by the pro-survival kinase Akt, leading to induction of apoptosis in endothelial cells (Song, Xie et al. 2008). PKCζ was also reported to phosphorylate LKB1 at Ser307 to promote LKB1 binding to STRADα and their nuclear to cytoplasmic export. This phosphorylation site is important for LKB1 to suppress angiogenesis (Xie, Dong et al. 2009).

LKB1-STRADα-Mo25 complex LKB1 functions in a trimeric complex, consisting of LKB1-STRAD- Mo25 in a 1:1:1 ratio. STRAD is a pseudokinase and is therefore inactive. Mo25

31 (mouse protein of 40 kDa) is a scaffolding protein and together with ATP, it binds to STRAD and keeps STRAD in a closed conformation that is typical of active kinases. STRAD in a closed/active conformation binds LKB1 as its pseudosubstrate and allows LKB1 to change to its active conformation, which is further stabilised by Mo25 that interacts with the LKB1 activation loop (Zeqiraj, Filippi et al. 2009).

LKB1, in complex with STRAD and Mo25, represents the active kinase which is able to phosphorylate its downstream substrates much more effi- ciently. For example, LKB1, in a complex with STRAD and Mo25, results in a 100-fold increase in downstream AMPK phosphorylation (Hawley, Boudeau et al. 2003).

STRAD also facilitates LKB1 nuclear export by forming a complex with exportins CRM1/exportin1 and exportin7 but inhibits LKB1 nuclear import by competitive binding to importinα (Dorfman and Macara 2008). Further- more, the interaction of LKB1 with LKB1 interacting protein 1 (LIP1) induces LKB1 cytoplasmic localisation to support LKB1 activation (Smith, Rayter et al. 2001).

The complex of LKB1-STRAD-Mo25 does not form spontaneously when the three components are mixed in vitro. LKB1 interacts with heat shock protein 90 (Hsp90) and Cdc37 (kinase-specific targeting subunit for Hsp90) when it is not in complex with STRAD or Mo25 and there is a possibility that Hsp90 and Cdc37 may facilitate LKB1-STRAD-Mo25 complex formation. Hsp90 stabilises LKB1 protein by inhibiting its proteasomal degradation induced by the E3 ubiquitin ligase CHIP (Alessi, Sakamoto et al. 2006).

On the contrary, LKB1 complex activity was shown to be negatively regulated by microRNA-451 in gliomas. miR-451 was able to repress the expression of Mo25 leading to inhibition LKB1/AMPK signalling, which is required for the control of cell migration and cell cycle arrest during glucose deprivation. The level of miR-451 is influenced by the level of glucose, high glucose induces miR-451 expression to support cell growth whereas low glucose reduces miR-451 level for cell survival and migration (Godlewski, Nowicki et al. 2010).

Most reports have described the function of STRADα and Mo25 as co- activators of LKB1, but recently their LKB1-independent functions have- been uncovered. New targets of Mo25 include the SPAK/OSR1 kinases that regulate ion homeostasis and blood pressure and the MST3/MST4/YSK1 kinases that are important during development (Filippi, de los Heros et al. 2011). In LKB1 null cell lines, STRADα activates the Rac1-PAK1 pathway to drive cell motility and invasion (Eggers, Kline et al. 2012).

32 Regulators of LKB1 expression Sex hormones have been reported to regulate LKB1 expression in various cell lines. 17β estradiol (E2) negatively regulates LKB1 transcription and protein levels in MCF-7 cells. ERα binds to the LKB1 promoter constitutively and E2 stimulation reduces ERα binding to the LKB1 promoter, thus lowering LKB1 promoter activity and expression (Brown, McInnes et al. 2011)

However, another independent report showed the opposite, that 17β estradiol treatment enhanced LKB1 mRNA expression (Linher-Melville, Zantinge et al. 2012). In adipocytes, it was similarly shown that testosterone suppressed LKB1 mRNA expression and this negative effect was reversed by estrogen treatment. LKB1 mRNA level was correlated with AMPK phosphorylation (McInnes, Brown et al. 2011).

Recently, microRNA-199a3p was found to repress LKB1 transcription during liver fibrosis. The farnesoid X receptor (FXR) and its ligand, however, represses miRNA-199a3p expression and up-regulates LKB1 level (Lee, Kim et al. 2012).

Kinases regulating LKB1 activity: LKB1 complex P90RSK LKB1 localisation: formation : PKA SIRT1 Hsp90 PKCζ PKCζ Cdc37 B-RAF LIP1 miR-451 Akt STRAD TAK1

LKB1 expression: Other regulators 17β-Estradiol Mo STRAD of LKB1 activity: miR-19913p 25 Bile acid signalling Testosterone LKB1 P FGF21 signalling

P AMPK1/2

NUAK1/2 BRSK1/2 MARK 1-4

QIK SIK QSK

Cell growth Cell polarity Energy metabolism Cell survival Cell death & proliferation

Figure 5. Regulators of LKB1 activity and LKB1 signalling via the AMPK family.

33 LKB1 signalling pathways LKB1, together with STRAD and Mo25, activates 14 downstream kinases (AMPK1, AMPK2, NUAK1, NUAK2, BRSK1, BRSK2, QIK, QSK, SIK, MARK1, MARK2, MARK3, MARK4, MELK) to regulate many cellular processes such as cell growth, cell polarity and cellular energy metabolism (Lizcano, Göransson et al. 2004).

LKB1-AMPK pathways Cellular energy regulation The best-studied LKB1 signalling pathway is probably the pathway involving AMPK in the regulation of cell survival during energetic stress (Figure 6). During energy depletion, when the level of cellular AMP is high, AMP induces the formation of AMPK complexes (catalytic α subunit and regulatory β and γ subunits) to facilitate its phosphorylation at the T loop (Thr172) by LKB1 (Shaw, Kosmatka et al. 2004). Besides AMP, there is evidence that adenosine diphosphate (ADP) binds to the γ subunit of AMPK, leading to its Thr172 phosphorylation by calcium/calmodium-dependent protein kinase kinase (CaMKK) β (Oakhill, Steel et al. 2011).

Activation of AMPK activates the tuberous sclerosis complex 1 and 2 (TSC1 – TSC2) which further inactivates the Rheb GTPase and thus results in inhibition of the mTOR pathway, a pathway that positively regulates protein synthesis and cell growth (Alessi, Sakamoto et al. 2006). AMPK activation also inhibits the mTOR signalling by inhibitory phosphorylation of Raptor, a protein that is a binding partner for mTOR in the mTOR complex 1 (mTORC1) (Gwinn, Shackelford et al. 2008). Therefore, AMPK activation by LKB1 during energy depletion ensures that cells minimise their ATP- consuming processes.

AMPK activation by LKB1 also functions to upregulate ATP producing processes. One example is the stimulation of glucose transport into cells by AMPK during exercise. AMPK activation inhibits Rab-GTPase-activating proteins (AS160 and TBC1D1), resulting in activation of Rab8A which promotes the translocation of GLUT4 (glucose transporter 4) from intracellular vesicles to the plasma membrane (Sakamoto and Holman 2008). Addi- tionally, LKB1 has been shown to facilitate the trafficking of monocarboxylate transporter (Sln/MCT1) to the plasma membrane, enabling the cells to receive input from extra energy sources, such as lactate and butyrate (Jang, Lee et al. 2008).

Independent of LKB1, AMPK is phosphorylated and inactivated by protein kinase A (PKA) in response to β-adrenergic signalling in primary mouse

34 adipocytes. This results in AMPK Ser173 phosphorylation, which impedes the Thr172 phosphorylation and activation by LKB1. Ser173 phosphorylated AMPK functions to promote lipolysis and the release of free fatty acids as an energy substrate during energy demand (Djouder, Tuerk et al. 2010).

Although LKB1 is well known as a tumour suppressor protein, it is also a protein that promotes cell survival under low energy conditions. Tumour cells, especially those that are close to the core of the tumour mass are de- prived of energy due to a lack of oxygen and nutrients and thus this activation of the LKB1-AMPK pathway may give them an advantage for survival.

The LKB1 – AMPK pathway phosphorylates p27 (a cyclin-dependent kinase inhibitor) at Thr198 to stabilise it and allowing MCF-7 cell survival during growth arrest by entering autophagy (Liang, Shao et al. 2007). Unc-51-like kinase (Ulk1) phosphorylates AMPK providing a negative feedback loop to this autophagy inducing pathway (Loffler, Alers et al. 2011).

Mo STRAD 25 NADPH level sustained LKB1 P & cell survival P ACC 1/2

P AMPK

P mTORC1 Raptor Inhibition of protein synthesis & cell growth

Rheb

Figure 6. The LKB1-AMPK pathways regulate energy metabolism. In low ATP conditions, LKB1 activates AMPK by phosphorylation (red P circle) and AMPK in turn phosphorylates both TSC2 and Raptor (inhibitory phosphorylation represented by blue P circle) to suppress mTORC1 activity and protein synthesis. AMPK also inhibits ACC1/2 by phosphorylation leading to sustained NADPH level and cell survival.

An independent study further emphasised this concept by describing a role for LKB1-AMPK positive regulation on NADPH levels during energetic stress to increase cancer cell survival. Activated AMPK inhibits the acetyl carboxylases ACC1 and ACC2 during metabolic stress, which enables maintenance of NADPH level generated from the pentose phosphate pathway.

35 This is achieved through decreasing NADPH consumption in fatty acid synthesis and increasing NADPH production in fatty acid oxidation. The inhibition of ACC1/ACC2 by AMPK promotes lung cancer cell survival and anchorage independent growth (Jeon, Chandel et al. 2012)

Cellular polarity regulation The MARKs, BRSKs and AMPKs are regulators of cellular polarity. The microtubule affinity-regulating kinases (MARKs) regulate cell polarity as suggested by their homolog in Drosophila, Par1 kinase (Martin and St Johnston 2003). The brain-specific kinases (BRSK) are expressed mainly in the nervous system and play an important role in polarising neurons to form axons and dendrites (Kishi, Pan et al. 2005).

The AMPKs are involved in the assembly of epithelial tight junctions in a calcium-rich environment and their activation promotes the maintenance of tight junctions during calcium depletion (Zheng and Cantley 2007). Since AMPK is a protein kinase with a crucial role in regulating cellular responses to energetic stress, AMPK allows for the preservation of cell polarity even during energy depletion. To promote epithelial cell polarity, AMPK phosphorylates myosin regulatory light chain (MRLC) and this is facilitated by dystroglycan (the receptor for perlecan, an extracellular matrix proteoglycan) which possibly helps MRLC to be appropriately localised for AMPK phosphorylation (Mirouse, Christoforou et al. 2009).

The maintenance of epithelial polarity and structure by the LKB1 pathway strongly supports resistance of mammary cells towards oncogenic activation by c-Myc (Partanen, Nieminen et al. 2007). Loss of LKB1 leads to decreased epithelial integrity and oncogenic synergy with c-Myc towards cell proliferation. Thus, it is clear that the tumour suppressor function of LKB1 is tightly linked to its positive regulation on cell polarity (Partanen, Tervonen et al. 2012).

Moreover, LKB1-AMPK signalling contributes to the formation of a proper mitotic spindle. AMPK phosphorylates myosin regulatory light chain (MRLC) at the spindle pole to control microtubule organisation for mitotic progression (Thaiparambil, Eggers et al. 2012).

LKB1-SIK1 pathways Salt inducible kinase 1 (SIK1) was discovered in the adrenal glands of rats fed with a high salt diet. SIK1 plays a central role as a sodium sensor and modulates cellular sodium transport in response to hormones or changes in water/sodium levels in our body. An increase in intracellular sodium increases the Na+ and K+ ATPase activity to pump sodium out of cells. In this

36 process, SIK1 cooperates with Na+ and K+ ATPase to enforce their maxi- mum sodium pumping activity (Jaitovich and Bertorello 2010).

In another metabolic pathway, SIK1 is required as a class II HDAC kinase that phosphorylates HDACs to activate the myocyte enhancer factor 2 (MEF2) transcription factor that induces the expression of muscle specific genes and enhance the survival of skeletal muscle cells during physical exercise (Berdeaux, Goebel et al. 2007).

To maintain epithelial cell polarity and integrity, LKB1-SIK1 signalling was shown to mediate E-cadherin expression and cell junction formation by down-regulating EMT transcription factors which repress E-cadherin, such as Snail2, Zeb1, Zeb2 (Eneling, Brion et al. 2012).

Similarly, it was reported that the LKB1-SIK1 pathway serves as a tumour suppressor pathway by positively regulating p53-dependent anoikis (apoptotic cell death induced by lack of adhesion) and suppressing anchorage independent growth and lung metastasis from breast cancer. SIK1 catalytic activity was shown to be crucial for p53 phosphorylation during loss of cell adhesion (Cheng, Liu et al. 2009).

LKB1-NUAK2 pathways Not much is known about the function of NUAK2 or sucrose non-fermenting AMPK-related kinase (SNARK), however, some reports have described its roles in regulation of metabolism, cell migration, and cell growth/survival during tumourigenesis.

In skeletal muscle metabolism, LKB1-dependent NUAK2 activity is induced by muscle contraction and glucose transport is increased as a compensation for the loss of energy during exercise (Koh, Toyoda et al. 2010).

The architecture of actin stress fibers is crucial for cellular processes such as cell shape, cell adhesion, cell division and cell motility. NUAK2 modulates the function of myosin regulatory light chain (MLC) phosphatase by interacting with the myosin phosphatase Rho-interacting protein (MRIP). The NUAK2-MRIP interaction inhibits the function of MYPT1, the regulatory subunit of MLC phosphatase. As a result, dephosphorylation of MLC is reduced and formation of actin stress fibers is elevated (Vallenius, Vaahtomeri et al. 2011).

In human melanoma, NUAK2 has tumour promoting properties as it induces cell cycle progression and increased immortality, cell migration and up- regulation of the mTOR pathway. Nevertheless, the mechanism of NUAK2-

37 mediated oncogenic effects has not been reported (Namiki, Tanemura et al. 2011).

The physiological roles of LKB1 To summarise, LKB1 is expressed ubiquitously in many tissues in our body and is important for the maintenance of tissue integrity, energy regulation and sensor for tissue survival, ensuring proper cell polarity for mitosis and cell migration but also providing checkpoint to control cell proliferation by working together with the downstream AMPKs and p53 tumour suppressor.

Many studies have been carried out on the role of LKB1 pathway in epithelial and muscle cells. In addition, emerging evidence from 3 independent groups demonstrated a role for LKB1 in maintaining the balance in hematopoietic stem cell (HSC) homeostasis (Gan, Hu et al. 2010; Gurumurthy, Xie et al. 2010; Nakada, Saunders et al. 2010).

When LKB1 was conditionally deleted in hematopoietic tissues, there was an initial increase in HSCs cell proliferation that was followed by elevated apoptosis and autophagy due to metabolic and genotoxic stress. LKB1 defi- ciency is also associated with decreased levels of PPARγ co-activator 1 (PGC1), a transcription factor that is important for mitochondrial biogenesis. Therefore, the loss of LKB1 in HSCs leads to ATP reduction and disrupted mitochondrial potential. Nevertheless, it is not yet known which downstream AMPK is involved in this process as AMPK and mTORC1 are not involved (Gan, Hu et al. 2010; Gurumurthy, Xie et al. 2010; Nakada, Saunders et al. 2010).

Likewise, LKB1 plays a role in the immune system. LKB1 is a regulator of T-cell development, viability, metabolism and activation. Unlike its role in hematopoietic tissues, LKB1 acts through activation of the AMPK and mTOR pathways in T-cells. Loss of LKB1 leads to reduced T-cell proliferation, defective development and impaired glucose/lipid metabolism which influences cell survival. Interestingly, it was also noted that loss of LKB1 promotes T-cell function and inflammatory cytokine production (MacIver, Blagih et al. 2011).

38 LKB1 signalling cross-talk

Signalling pathways affecting LKB1 activity Bile acid (taurocholate) and fibroblast growth factor 21 (FGF21) positively regulate LKB1 activity. Bile acid stimulates cAMP-Epac-MEK-Rap1 pathway which activates LKB1-AMPK signalling to induce hepatocyte polarisa- tion during canalicular network formation (Fu, Wakabayashi et al. 2011).

On the other hand, FGF21 is a metabolic regulator that works via the LKB1- AMPK pathway to activate sirtuin1 (SIRT1), which deacetylates PGC-1α. Effectively, this leads to elevated mitochondrial oxidative function and energy production in adipocytes (Chau, Gao et al. 2010).

Additionally, the activity of TAK1 is required for LKB1 activity and its ability to phosphorylate downstream AMPK in mouse embryonic fibroblasts. Activators of AMPK (for example AICAR and metformin) could activate TAK1 as a positive feedback (Xie, Zhang et al. 2006).

Negative regulators of LKB1 activity have been identified as well, including tumour promoters such as B-RAF and Akt. The oncogenic B-RAF V600E mutation enhances LKB1 phosphorylation by ERK and Rsk kinases. This phosphorylation inhibits LKB1 ability to activate AMPK and finally contributes to melanoma cell proliferation (Zheng, Jeong et al. 2009).

Similarly, Akt also acts upstream of LKB1 and phosphorylates LKB1 at Ser334, leading to a decreased binding to STRADα but an increased binding to the 14-3-3 nuclear proteins. As a consequence, LKB1 cannot translocate to the cytoplasm and its tumour suppressor function of inhibiting breast cancer proliferation is compromised (Liu, Siu et al. 2012).

LKB1 affects other signalling pathways Since LKB1 is a tumour suppressor, it is important to know the mechanism of how LKB1 controls cell proliferation. When LKB1 was expressed in cancer cell lines, they were arrested at the G1 phase and this process was due to induced p21 levels which were in turn mediated by p53 transcriptional activity on the p21 promoter (Tiainen, Vaahtomeri et al. 2002). Another report suggests that LKB1 could also work via AMPK to activate p53 and inhibit cell proliferation (Jones, Plas et al. 2005).

Conversely, as was mentioned, the LKB1 tumour suppressor effect could be context-dependent, as LKB1 has been shown to be required for Akt mediated phosphorylation of downstream pro-apoptotic proteins, thus suppressing apoptosis. It was further explained that LKB1 supports Akt function in the

39 context of cells expressing constitutively active Akt but not in cells without Akt activation (Zhong, Liu et al. 2008).

There are more reports describing the cross-talk of LKB1 signalling with other pathways. Despite being a tumour suppressor, LKB1 has been shown to positively regulate estrogen receptor-α (ER-α) signalling (Nath-Sain and Marignani 2009) which gives an advantage to cancer cell growth.

LKB1 may play a role in Wnt signalling, as reported by elevated WNT5A level in LKB1 (-/+) mice and human PJS polyps (Lai, Robinson et al. 2011). In another study, LKB1 was suggested to inhibit Wnt signalling by phos- phorylating and re-directing PAR1A activity towards other unknown pathways (Spicer, Rayter et al. 2003).

Most relevant to this thesis work is the study of LKB1 cross-talk with TGFβ signalling. Three studies have suggested that LKB1 is required for TGFβ signalling and TGFβ mediated cellular processes such as suppression of cell growth to prevent gastrointestinal polyposis, induction of myofibroblast differentiation and vascular smooth muscle recruitment during angiogenesis (Katajisto, Vaahtomeri et al. 2008; Londesborough, Vaahtomeri et al. 2008; Vaahtomeri, Ventela et al. 2008).

It was described that a loss of LKB1 in stromal cells, by targeted deletion, resulted in decreased TGFβ production and secretion, leading to less exposure of epithelial cells to TGFβ and promotion of epithelial cell proliferation due to a loss of TGFβ signalling that mediates cell cycle arrest (Katajisto, Vaahtomeri et al. 2008). Furthermore, restoration of LKB1 in LKB1 null mouse embryonic fibroblasts promotes TGFβ signalling and induces differentiation (Vaahtomeri, Ventela et al. 2008).

Our findings suggest a complementary scenario. LKB1 negatively regulates TGFβ/BMP signalling by inhibitory phosphorylation of Smad4 and down- regulation of the BMP type I receptor (Moren, Raja et al. 2011). This context dependent role of LKB1 in TGFβ signalling is intriguing and one possibility could point to the unknown factors that work together with LKB1 that are differentially expressed in different cell types. Possibly, co-expression of STRADα and Mo25α as LKB1 co-activators in our study may also activate certain factors that direct LKB1 towards negative regulation of TGFβ signalling.

Alternatively, the role of downstream AMPK family kinases needs to be addressed. Based on the multi-functionality of the AMPKs, we could pro- pose alternative cell type-dependent mechanisms whereby LKB1 either enhances or suppresses TGFβ signalling.

40 Aim of study

Our interest in the study of LKB1 cross-talk with TGFβ signalling was based on the observation that Smad4 associates with the LKB1 interacting protein 1 (LIP1) (Smith, Rayter et al. 2001). We decided to further investigate the role of LKB1 in TGFβ signalling.

I. In Paper I, we evaluated the function of the LIP-LKB1 complex with Smad4. Since LKB1 is a kinase, it was rational to test Smad4 phosphorylation and furthermore we examined the effect of this phosphorylation on the function of Smad4 and finally the overall influence on TGFβ and BMP signalling and TGFβ- mediated cellular responses such as EMT.

II. Paper II aimed to study more the role of LKB1 in BMP signalling and investigate its effects on components of BMP signalling other than Smad4. Our analysis focused on the receptor level and we identified a new mechanism of action for LKB1.

III. Paper III was originated from the microarray screen that identified SIK1 as a TGFβ target gene (Kowanetz, Lonn et al. 2008). Here we aimed to further demonstrate the negative role of SIK1 in TGFβRI signalling via Smurf1 ubiquitin ligase and to study the mechanism of SIK1 promoter induction by TGFβ.

IV. Lastly, Paper IV was inspired by Paper III and here we tried to screen for other AMPKs that could be transcriptionally regulated by TGFβ signalling. We identified NUAK2 as such a kinase and further analysed the function of NUAK2 in TGFβ and LKB1 signalling.

41 Present Investigation

Paper I. Negative regulation of TGFβ signalling by the kinase LKB1 and the scaffolding protein LIP1

In this study, we discovered the biological function of the LKB1-LIP1- Smad4 complex in relation to TGFβ signalling. We found that the association of LKB1 with Smad4 leads to Smad4 phosphorylation on Thr 77. This phosphorylation may provide for the negative regulation on Smad4 binding to DNA in response to TGFβ stimulation (Figure 7).

Indeed, co-incubation of LKB1 and Smad4 in vitro led to suppressed Smad4 binding to small fragments of BMP or TGFβ promoters. This was further confirmed by chromatin immuno-precipitation assay. Gain of function of LKB1 led to decreased Smad4 binding to the TGFβ-induced promoter, PAI1. Moreover, LKB1-induced inhibition on Smad4 correlated with reduced BMP- or TGFβ-dependent promoter activities as reflected by luciferase signals and endogenous BMP- or TGFβ-dependent gene expres- sions.

Finally, just as LKB1 suppressed TGFβ signalling, it also suppressed TGFβ- induced EMT as seen morphologically and via its effects on EMT markers such as Snail, Slug, fibronectin and E-cadherin. In the same line, LKB1 blocked TGFβ-stimulated breast cancer cell invasion to emphasise its role as a negative regulator of TGFβ signalling and as a tumour suppressor.

Paper II. The tumour suppressor kinase LKB1 negatively regulates bone morphogenetic protein receptor signalling

The clear evidence that LKB1 modulates TGFβ signalling led us to investigate more thoroughly the role of LKB1 on BMP signalling. From paper I we had observed that LKB1 negatively affects BMP-induced gene responses. In this paper we describe that in addition to Smad4, LKB1 negatively regulates BMP signalling through inhibition of BMP type I receptor activity. LKB1 interacts with both inhibitory Smad7 and ALK2 receptor to promote ALK2 receptor degradation via the proteasome (Figure 7).

42 LKB1-induced inhibition of BMP type I receptor results in suppression of Smad1/5/8 phosphorylation, decreased BMP-dependent promoter activity and gene transcription and reduced BMP-stimulated osteoblast differentiation.

Paper III. Transcriptional induction of salt-inducible kinase 1 by transforming growth factor β leads to negative regulation of type I receptor signalling in cooperation with the Smurf2 ubiquitin ligase

Our earlier study found SIK1 to be a negative regulator of TGFβ receptor signalling by down-regulation of ALK5. SIK1 is a TGFβ target gene, together with the Smad7 and the Smurf2 genes. In paper III, we elucidated the mechanism of SIK1 gene induction by TGFβ. We found that SIK1 induction is dependent on Smad2/3/4 binding to a putative enhancer region located 3’ from the last exon of the gene.

In addition, we provide evidence that Smurf2 cooperates with SIK1 in a complex with the ALK5 receptor to promote receptor degradation. In this process, both Smurf2 ubiquitin ligase activity and SIK1 kinase activity are required. On the other hand, blocking SIK1 function by siRNA showed increased physiological TGFβ signalling and enhanced epithelial cell growth arrest (Figure 7).

Paper IV. Transcriptional induction of NUAK2 by TGFβ leads to phosphorylation and inhibition of the translational regulator Raptor.

Our finding that SIK1 modulates TGFβ signalling motivated us to screen for additional kinases within the AMPK family that are also induced by TGFβ signalling. We found that NUAK2 is strongly induced by TGFβ stimulation at both the mRNA and protein levels. We were able to confirm that NUAK2 induction happens at the transcriptional level by utilising Smad2/3/4 transactivation activity on NUAK2 putative enhancer element located within the first intron of the gene.

TGFβ-induced NUAK2 induction was linked to its association with LKB1 and Raptor phosphorylation. Raptor phosphorylation is known to negatively regulate mTORC1 and we obtained supportive evidence that knock down of NUAK2 leads to increased 4E-BP1 phosphorylation downstream of mTORC1 and enhanced cyclinD1 level, an established protein produced by the mTORC1- 4E-BP1 translational pathway. Cyclin D1 is an important factor in cell cycle progression, thus it was not surprising that TGFβ-induced NUAK2 finally promoted epithelial cell growth inhibition, a well established physiological response to TGFβ signalling.

43 TGFβ BMP BMP TGFβ Plasma membrane β RI RII RI BMP β TGF BMPRII

TGF P P

SIK1 degradation Proteasomal

Mo STRAD 25

LKB1 P

Smad4 P

Smad1/5/8 Smad4 Smad4 P P Smad2/3 TF X TF X CAGA BRE

Figure 7. LKB1-SIK1 negative regulation of the TGFβ/BMP signalling pathways. LKB1 phosphorylates Smad4 (blue P circle) which suppresses Smad4 binding to DNA and TGFβ/BMP-dependent transcription. LKB1 further down-regulates BMP type I receptor (BMPRI) by forming a protein complex with Smad7 and the receptor, promoting receptor proteasomal degradation. TGFβ signaling induces SIK1 level transcriptionally and SIK1 acts as a negative feedback to the signaling pathway by cooperating with Smurf2 and Smad7 to increase TGFβ type I receptor (TGFβRI) turnover and degradation.

44 Future perspectives

Our attempts to answer some questions regarding the link between TGFβ/BMP signalling and LKB1/AMPK signalling will surely lead to many more questions to be addressed in the future. From my small exploration, I can appreciate very well the beauty and complexity of this intracellular network of communication and certainly the vastness of the unknown. Here are some thoughts on the continuation of my studies.

Paper I. Negative regulation of TGFβ signalling by the kinase LKB1 and the scaffolding protein LIP1

In this first paper we described how LKB1 affects Smad4 DNA binding ca- pacity and TGFβ signalling via the mediator protein LKB1-interacting protein 1 (LIP1). Even though we put a special emphasis on LKB1, the role of LIP1 in this cross-talk may also be as important and currently very little is known about the function of LIP1.

Since the complex formation between LKB1 and Smad4 requires LIP1 and LIP1 gain or loss of function showed similar effects as gain or loss of LKB1 in TGFβ/BMP signalling readouts, it would be interesting to investigate how TGFβ/BMP signalling promotes LIP1 activity/function to bring LKB1 and Smad4 together in a complex. Our preliminary results suggest that LKB1 phosphorylates LIP1 and LIP1 forms a dimer. Is this phosphorylation important for LIP1 dimerisation and for bringing LKB1 and Smad4 in close prox- imity?

Since LIP1 is a cytoplasmic protein (Smith, Rayter et al. 2001) some other possible questions include whether LIP1 tethers LKB1 in the cytoplasm to phosphorylate Smad4 and if this would lead to Smad4 cytoplasmic retention and prevent its association with DNA.

In advanced stages of cancer, TGFβ often acts as a tumour promoter partly by supporting cell migration and invasion during metastasis. In our study, we observed that TGFβ enhanced the invasion of metastatic breast cancer cell in vitro and this effect was blocked by LKB1 gain of function. It would be important to test LKB1 and also LIP1 gain or loss of function in vivo, to further

45 ascertain that the effect of LKB1 on TGFβ signalling is mediated by LIP1 and complex formation with Smad4.

Paper II. The tumour suppressor kinase LKB1 negatively regulates bone morphogenetic protein receptor signalling

We obtained clear evidence that LKB1 plays a negative role in BMP signalling, by affecting Smad4 and the BMP type I receptor turnover. It is likely that LKB1 affects BMP receptor signalling in a direct manner, as LKB1 forms functional interacting complexes with both the receptor and Smad7. Finally, LKB1 induces BMP receptor degradation via the proteasome.

Proteasomal degradation could occur in two ways, via ubiquitin-dependent or ubiquitin-independent pathways. Classically, the TGFβ family of receptors are degraded via the ubiquitin-dependent pathway, which is supported by Smad7 recruitment of ubiquitin ligases to the receptor (Ebisawa, Fukuchi et al. 2001). Therefore, it is essential to examine if any of the known ubiquitin ligases are involved in the LKB1-induced BMP receptor degradation pathway.

However, it is also worth investigating whether LKB1 induces BMP receptor degradation via the ubiquitin-independent pathway, as this is a process that does not require ATP and LKB1 is a metabolic enzyme that promotes the survival of cells during low ATP or energy state (Alessi, Sakamoto et al. 2006). This pathway involves an alternate proteasomal activator REGγ which binds to target proteins and induces their proteasomal degradation (Chen, Barton et al. 2007).

Currently, the role of LKB1 in BMP signalling is being tested during wing development in Drosophila. The results so far show agreement with our in vitro findings, that LKB1 inhibits BMP signalling during pupal stage wing development. This is manifested as thickened vein tips in the adult fly wing. Although at this stage we cannot pinpoint the precise mechanism, one possibility could be that LKB1 promotes a loss of Thickvein receptor (BMP type I receptor in Drosophila) and as a consequence, the established restriction of Dpp (BMP ligand for Tkv receptor) to the vein line is reduced and Dpp spreads out, forming a thicker vein.

As a further attempt, we are trying to correlate our signalling cross-talk in a disease model such as lung cancer, as LKB1 mutations have been linked to sporadic lung cancer.

46 Paper III. Transcriptional induction of salt-inducible kinase 1 by transforming growth factor β leads to negative regulation of type I receptor signalling in cooperation with the Smurf2 ubiquitin ligase

SIK1 is a TGFβ target gene but it is also a downstream kinase that is activated by LKB1. It would be interesting to find out if LKB1 acts upstream of SIK1 in the process of TGFβ receptor down-regulation, in addition to the Smad4 mechanism.

Moreover, we could also examine how the ALK5 receptor is degraded by SIK1. Does it involve phosphorylation of Smad7 or Smurf2? SIK1 is a nuclear protein (Katoh, Takemori et al. 2002), how does it move to the plasma membrane to degrade the receptor? Does it bind and move together with Smad7 from the nucleus to the plasma membrane? Would SIK push the in- ternalisation of ALK5 receptor towards the lipid raft degradation pathway instead of the caveolin signalling pathway?

Regarding the biological role of SIK1, so far SIK1 decreases TGFβ- dependent cell cycle arrest in epithelial cells. It would be nice to extend this to an epithelial cancer cell model where the TGFβ cytostatic pathway is functional and see if SIK1 gain of function would increase tumour burden in a mouse model.

Finally, in view of LKB1 and SIK1 regulation in TGFβ/BMP signalling, right now we have only looked at effects on epithelial or mesenchymal cells. Tissues consist of combinations of many types of cells and there is functional communication between different cell types. Therefore, it is also interesting to study the function of this cross-talk in the context of communication between different cell types.

Paper IV. Transcriptional induction of NUAK2 by TGFβ leads to phosphorylation and inhibition of the translational regulator Raptor.

Even though NUAK2 has been reported to interact with TGFβRI, Smad2 and Smad4 (Barrios-Rodiles, Brown et al. 2005), we were not able to see any effects of NUAK2 knockdown on classical TGFβ gene targets. This led us to investigate the translational pathway that is affected by the LKB1- AMPK pathway. Nonetheless, we should keep an open mind about this possibility as we have observed as well that LKB1 forms a complex with phosphorylated Smad2 together with NUAK2 (unpublished result).

There are many more questions we could address in this study. TGFβ- induced NUAK2 is linked to Raptor phosphorylation and it is necessary to test if NUAK2 interacts and phosphorylates Raptor directly. Raptor phos-

47 phorylation is inhibitory for the function of mTORC1 so we could test also if the presence of NUAK2 would abolish Raptor interaction with mTOR to confirm the negative regulation on mTORC1.

The function of mTORC1 and mTORC2 has been related to cell motility (Gulhati, Bowen et al. 2011). Rapamycin disrupts mTOR-Raptor complexes and inhibits tumour cell motility by preventing F-actin reorganisation and phosphorylation of focal adhesion proteins (Liu, Chen et al. 2008). In addition, rapamycin also blocks activation of S6K1 and 4E-BP1 pathways (Liu, Li et al. 2006) resulting in suppression of RhoA expression and activity (Liu, Luo et al. 2010) for negative regulation of cell motility.

NUAK2 has been reported to influence the actin cytoskeleton (Vallenius, Vaahtomeri et al. 2011). Therefore, besides looking at the effects of NUAK2 on TGFβ-induced cell cycle arrest via down-regulation of cyclin D1, it is intriguing to see if Raptor phosphorylation by NUAK2 would affect the actin cytoskeleton in a direct way or indirect way via inhibition of S6K1/4E-BP1 pathways and RhoA.

In conclusion, our study in this thesis work has provided some important links between the LKB1/AMPK signalling and TGFβ/BMP pathways. How- ever, there is always room for deepening our knowledge and for new findings to come. I sincerely hope that our work could serve as a good resource and ground for future studies and applications in the fields of TGFβ, BMP, LKB1 and AMPK signal transduction.

48 References

Akiyama, S., T. Katagiri, et al. (1997). "Constitutively active BMP type I receptors transduce BMP-2 signals without the ligand in C2C12 myoblasts." Exp. Cell Res. 235(2): 362-9. Alarcon, C., A. I. Zaromytidou, et al. (2009). "Nuclear CDKs drive Smad transcriptional activation and turnover in BMP and TGF-beta pathways." Cell 139(4): 757-69. Alessi, D. R., K. Sakamoto, et al. (2006). "LKB1-Dependent Signaling Pathways." Annu. Rev. Biochem. 75: 137-163. Alliston, T., E. Piek, et al., Eds. (2008). TGF-β family signaling in skeletal development, maintenance and disease. The TGF-β family. Cold Spring Harbour, New York, Cold Spring Harbour Laboratory Press. Bannister, A. J. and T. Kouzarides (1996). "The CBP co-activator is a histone acetyltransferase." Nature 384(6610): 641-3. Bardeesy, N., M. Sinha, et al. (2002). "Loss of the Lkb1 tumour suppressor provokes intestinal polyposis but resistance to transformation." Na- ture 419(6903): 162-7. Barrios-Rodiles, M., K. R. Brown, et al. (2005). "High-throughput mapping of a dynamic signaling network in mammalian cells." Science 307(5715): 1621-5. Beers, M. F., K. O. Solarin, et al. (1998). "TGF-beta1 inhibits surfactant component expression and epithelial cell maturation in cultured human fetal lung." Am J Physiol 275(5 Pt 1): L950-60. Bellusci, S., R. Henderson, et al. (1996). "Evidence from normal expression and targeted misexpression that bone morphogenetic protein (Bmp- 4) plays a role in mouse embryonic lung morphogenesis." Develop- ment 122(6): 1693-702. Berdeaux, R., N. Goebel, et al. (2007). "SIK1 is a class II HDAC kinase that promotes survival of skeletal myocytes." Nat Med 13(5): 597-603. Bhowmick, N. A., A. Chytil, et al. (2004). "TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia." Science 303(5659): 848-51. Brosens, L. A., D. Langeveld, et al. (2011). "Juvenile polyposis syndrome." World J Gastroenterol 17(44): 4839-44. Brown, C. B., A. S. Boyer, et al. (1999). "Requirement of type III TGF-beta receptor for endocardial cell transformation in the heart." Science 283(5410): 2080-2. Brown, K. A., K. J. McInnes, et al. (2011). "LKB1 expression is inhibited by estradiol-17beta in MCF-7 cells." J Steroid Biochem Mol Biol 127(3-5): 439-43.

49 Chan, M. C., P. H. Nguyen, et al. (2007). "A novel regulatory mechanism of the bone morphogenetic protein (BMP) signaling pathway involving the carboxyl-terminal tail domain of BMP type II receptor." Mol. Cell. Biol. 27(16): 5776-89. Chau, M. D., J. Gao, et al. (2010). "Fibroblast growth factor 21 regulates energy metabolism by activating the AMPK-SIRT1-PGC-1alpha pathway." Proc Natl Acad Sci U S A 107(28): 12553-8. Chen, C. R., Y. Kang, et al. (2002). "E2F4/5 and p107 as Smad cofactors linking the TGFbeta receptor to c-myc repression." Cell 110(1): 19- 32. Chen, D., X. Ji, et al. (1998). "Differential roles for bone morphogenetic protein (BMP) receptor type IB and IA in differentiation and speci- fication of mesenchymal precursor cells to osteoblast and adipocyte lineages." J. Cell Biol. 142(1): 295-305. Chen, X., L. F. Barton, et al. (2007). "Ubiquitin-independent degradation of cell-cycle inhibitors by the REGgamma proteasome." Mol Cell 26(6): 843-52. Chen, X., M. J. Rubock, et al. (1996). "A transcriptional partner for MAD proteins in TGF-beta signalling." Nature 383(6602): 691-6. Cheng, H., P. Liu, et al. (2009). "SIK1 couples LKB1 to p53-dependent anoikis and suppresses metastasis." Sci Signal 2(80): ra35. Claassen, G. F. and S. R. Hann (2000). "A role for transcriptional repression of p21CIP1 by c-Myc in overcoming transforming growth factor beta -induced cell-cycle arrest." Proc Natl Acad Sci U S A 97(17): 9498-503. Conery, A. R., Y. Cao, et al. (2004). "Akt interacts directly with Smad3 to regulate the sensitivity to TGF-beta induced apoptosis." Nat Cell Bi- ol 6(4): 366-72. Cordenonsi, M., S. Dupont, et al. (2003). "Links between tumor suppressors: p53 is required for TGF-beta gene responses by cooperating with Smads." Cell 113(3): 301-14. Daly, A. C., R. A. Randall, et al. (2008). "Transforming growth factor beta- induced Smad1/5 phosphorylation in epithelial cells is mediated by novel receptor complexes and is essential for anchorage-independent growth." Mol Cell Biol 28(22): 6889-902. Datta, P. K. and H. L. Moses (2000). "STRAP and Smad7 synergize in the inhibition of transforming growth factor beta signaling." Mol Cell Biol 20(9): 3157-67. de Larco, J. E. and G. J. Todaro (1978). "Growth factors from murine sarcoma virus-transformed cells." Proc Natl Acad Sci U S A 75(8): 4001-5. Deheuninck, J. and K. Luo (2009). "Ski and SnoN, potent negative regulators of TGF-beta signaling." Cell Res 19(1): 47-57. Deng, Z., J. H. Morse, et al. (2000). "Familial primary pulmonary hypertension (gene PPH1) is caused by mutations in the bone morphogenetic protein receptor-II gene." Am. J. Hum. Genet. 67(3): 737-44.

50 Dennler, S., S. Itoh, et al. (1998). "Direct binding of Smad3 and Smad4 to critical TGF beta-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene." EMBO J 17(11): 3091-100. Di Guglielmo, G. M., C. Le Roy, et al. (2003). "Distinct endocytic pathways regulate TGF-beta receptor signalling and turnover." Nat Cell Biol 5(5): 410-21. Ding, R., D. C. Darland, et al. (2004). "Endothelial-mesenchymal interactions in vitro reveal molecular mechanisms of smooth muscle/pericyte differentiation." Stem Cells Dev 13(5): 509-20. Djouder, N., R. D. Tuerk, et al. (2010). "PKA phosphorylates and inactivates AMPKalpha to promote efficient lipolysis." EMBO J 29(2): 469-81. Dorfman, J. and I. G. Macara (2008). "STRADα regulates LKB1 localization by blocking access to importin-α, and by association with Crm1 and exportin-7." Mol. Biol. Cell 19(4): 1614-26. Ducy, P., R. Zhang, et al. (1997). "Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation." Cell 89(5): 747-754. Ebisawa, T., M. Fukuchi, et al. (2001). "Smurf1 interacts with transforming growth factor-beta type I receptor through Smad7 and induces receptor degradation." J Biol Chem 276(16): 12477-80. Eggers, C. M., E. R. Kline, et al. (2012). "STE20-related Kinase Adaptor Protein alpha (STRADalpha) Regulates Cell Polarity and Invasion through PAK1 Signaling in LKB1-null Cells." J Biol Chem 287(22): 18758-68. Eneling, K., L. Brion, et al. (2012). "Salt-inducible kinase 1 regulates E- cadherin expression and intercellular junction stability." FASEB J. Feeley, B. T., S. C. Gamradt, et al. (2005). "Influence of BMPs on the formation of osteoblastic lesions in metastatic prostate cancer." J. Bone Miner. Res. 20(12): 2189-99. Feng, X. H. and R. Derynck (1997). "A kinase subdomain of transforming growth factor-beta (TGF-beta) type I receptor determines the TGF- beta intracellular signaling specificity." EMBO J 16(13): 3912-23. Fernandez, L. A., F. Sanz-Rodriguez, et al. (2006). "Hereditary hemorrhagic telangiectasia, a vascular dysplasia affecting the TGF-beta signaling pathway." Clin Med Res 4(1): 66-78. Filippi, B. M., P. de los Heros, et al. (2011). "MO25 is a master regulator of SPAK/OSR1 and MST3/MST4/YSK1 protein kinases." EMBO J 30(9): 1730-41. Forcet, C., S. Etienne-Manneville, et al. (2005). "Functional analysis of Peutz-Jeghers mutations reveals that the LKB1 C-terminal region exerts a crucial role in regulating both the AMPK pathway and the cell polarity." Hum. Mol. Genet. 14(10): 1283-92. Fu, D., Y. Wakabayashi, et al. (2011). "Bile acid stimulates hepatocyte polarization through a cAMP-Epac-MEK-LKB1-AMPK pathway." Proc Natl Acad Sci U S A 108(4): 1403-8.

51 Fujii, M., K. Takeda, et al. (1999). "Roles of bone morphogenetic protein type I receptors and Smad proteins in osteoblast and chondroblast differentiation." Mol. Biol. Cell 10(11): 3801-3813. Gan, B., J. Hu, et al. (2010). "Lkb1 regulates quiescence and metabolic homeostasis of haematopoietic stem cells." Nature 468(7324): 701-4. Godlewski, J., M. O. Nowicki, et al. (2010). "MicroRNA-451 regulates LKB1/AMPK signaling and allows adaptation to metabolic stress in glioma cells." Mol Cell 37(5): 620-32. Gordon, K. J., K. C. Kirkbride, et al. (2009). "Bone morphogenetic proteins induce pancreatic cancer cell invasiveness through a Smad1- dependent mechanism that involves matrix metalloproteinase-2." Carcinogenesis 30(2): 238-48. Gorelik, L. and R. A. Flavell (2001). "Immune-mediated eradication of tumors through the blockade of transforming growth factor-beta signaling in T cells." Nat Med 7(10): 1118-22. Goto, K., Y. Kamiya, et al. (2007). "Selective inhibitory effects of Smad6 on bone morphogenetic protein type I receptors." J Biol Chem 282(28): 20603-11. Goumans, M. J., Z. Liu, et al. (2009). "TGF-beta signaling in vascular biolo- gy and dysfunction." Cell Res 19(1): 116-27. Goumans, M. J., G. Valdimarsdottir, et al. (2003). "Activin receptor-like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling." Mol Cell 12(4): 817-28. Goumans, M. J., G. Valdimarsdottir, et al. (2002). "Balancing the activation state of the endothelium via two distinct TGF-beta type I receptors." EMBO J 21(7): 1743-53. Gronroos, E., U. Hellman, et al. (2002). "Control of Smad7 stability by competition between acetylation and ubiquitination." Mol Cell 10(3): 483-93. Guicheux, J., J. Lemonnier, et al. (2003). "Activation of p38 mitogen- activated protein kinase and c-Jun-NH2-terminal kinase by BMP-2 and their implication in the stimulation of osteoblastic cell differentiation." J. Bone Miner. Res. 18(11): 2060-8. Guldberg, P., P. thor Straten, et al. (1999). "Somatic mutation of the Peutz- Jeghers syndrome gene, LKB1/STK11, in malignant melanoma." Oncogene 18(9): 1777-80. Gulhati, P., K. A. Bowen, et al. (2011). "mTORC1 and mTORC2 regulate EMT, motility, and metastasis of colorectal cancer via RhoA and Rac1 signaling pathways." Cancer Res 71(9): 3246-56. Guo, X. and X. F. Wang (2009). "Signaling cross-talk between TGF- beta/BMP and other pathways." Cell Res 19(1): 71-88. Gurumurthy, S., S. Z. Xie, et al. (2010). "The Lkb1 metabolic sensor main- tains haematopoietic stem cell survival." Nature 468(7324): 659-63. Gwinn, D. M., D. B. Shackelford, et al. (2008). "AMPK phosphorylation of raptor mediates a metabolic checkpoint." Mol Cell 30(2): 214-26. Halbrooks, P. J., R. Ding, et al. (2007). "Role of RGM coreceptors in bone morphogenetic protein signaling." J Mol Signal 2: 4.

52 Hartung, A., K. Bitton-Worms, et al. (2006). "Different routes of bone morphogenic protein (BMP) receptor endocytosis influence BMP signaling." Mol. Cell. Biol. 26(20): 7791-7805. Hawley, S. A., J. Boudeau, et al. (2003). "Complexes between the LKB1 tumor suppressor, STRAD α/β and MO25 α/β are upstream kinases in the AMP-activated protein kinase cascade." J. Biol. 2(4): 28. Hayashi, H., S. Abdollah, et al. (1997). "The MAD-related protein Smad7 associates with the TGFbeta receptor and functions as an antagonist of TGFbeta signaling." Cell 89(7): 1165-73. Heldin, C. H., M. Vanlandewijck, et al. (2012). "Regulation of EMT by TGFbeta in cancer." FEBS Lett 586(14): 1959-70. Hemminki, A., D. Markie, et al. (1998). "A serine/threonine kinase gene defective in Peutz-Jeghers syndrome." Nature 391(6663): 184-7. Hill, C. S. (2009). "Nucleocytoplasmic shuttling of Smad proteins." Cell Res 19(1): 36-46. Howe, J. R., J. L. Bair, et al. (2001). "Germline mutations of the gene encoding bone morphogenetic protein receptor 1A in juvenile polyposis." Nat. Genet. 28(2): 184-187. Huse, M., Y. G. Chen, et al. (1999). "Crystal structure of the cytoplasmic domain of the type I TGF beta receptor in complex with FKBP12." Cell 96(3): 425-36. Imamura, T., M. Takase, et al. (1997). "Smad6 inhibits signalling by the TGF-beta superfamily." Nature 389(6651): 622-6. Inoue, Y., Y. Itoh, et al. (2007). "Smad3 is acetylated by p300/CBP to regulate its transactivation activity." Oncogene 26(4): 500-8. Itoh, S. and P. ten Dijke (2007). "Negative regulation of TGF-beta receptor/Smad signal transduction." Curr Opin Cell Biol 19(2): 176-84. Jaitovich, A. and A. M. Bertorello (2010). "Intracellular sodium sensing: SIK1 network, hormone action and high blood pressure." Biochim Biophys Acta 1802(12): 1140-9. Jang, C., G. Lee, et al. (2008). "LKB1 induces apical trafficking of Silnoon, a monocarboxylate transporter, in Drosophila melanogaster." J. Cell Biol. 183(1): 11-7. Jang, C. W., C. H. Chen, et al. (2002). "TGF-beta induces apoptosis through Smad-mediated expression of DAP-kinase." Nat Cell Biol 4(1): 51- 8. Janknecht, R., N. J. Wells, et al. (1998). "TGF-beta-stimulated cooperation of smad proteins with the coactivators CBP/p300." Genes Dev 12(14): 2114-9. Jeon, S. M., N. S. Chandel, et al. (2012). "AMPK regulates NADPH homeostasis to promote tumour cell survival during energy stress." Na- ture 485(7400): 661-5. Ji, H., M. R. Ramsey, et al. (2007). "LKB1 modulates lung cancer differentiation and metastasis." Nature 448(7155): 807-10. Jones, R. G., D. R. Plas, et al. (2005). "AMP-activated protein kinase induces a p53-dependent metabolic checkpoint." Mol Cell 18(3): 283-93.

53 Joseph, H., A. E. Gorska, et al. (1999). "Overexpression of a kinase-deficient transforming growth factor-beta type II receptor in mouse mammary stroma results in increased epithelial branching." Mol Biol Cell 10(4): 1221-34. Kang, J. S., T. Alliston, et al. (2005). "Repression of Runx2 function by TGF-beta through recruitment of class II histone deacetylases by Smad3." EMBO J 24(14): 2543-55. Kang, J. S., E. F. Saunier, et al. (2008). "The type I TGF-beta receptor is covalently modified and regulated by sumoylation." Nat Cell Biol 10(6): 654-64. Kanomata, K., S. Kokabu, et al. (2009). "DRAGON, a GPI-anchored membrane protein, inhibits BMP signaling in C2C12 myoblasts." Genes Cells 14(6): 695-702. Katagiri, T., S. Akiyama, et al. (1997). "Bone morphogenetic protein-2 inhibits terminal differentiation of myogenic cells by suppressing the transcriptional activity of MyoD and myogenin." Exp. Cell Res. 230(2): 342-51. Katagiri, T., M. Imada, et al. (2002). "Identification of a BMP-responsive element in Id1, the gene for inhibition of myogenesis." Genes Cells 7(9): 949-960. Katagiri, T., T. Suda, et al., Eds. (2008). The bone morphogenetic proteins. The TGF-β family. Cold Spring Harbor, Cold Spring Harbor La- boartory Press. Katajisto, P., K. Vaahtomeri, et al. (2008). "LKB1 signaling in mesenchymal cells required for suppression of gastrointestinal polyposis." Nat Ge- net 40(4): 455-9. Katajisto, P., T. Vallenius, et al. (2007). "The LKB1 tumor suppressor kinase in human disease." Biochim Biophys Acta 1775(1): 63-75. Katoh, Y., H. Takemori, et al. (2002). "Identification of the nuclear localization domain of salt-inducible kinase." Endocr Res 28(4): 315-8. Katsuno, Y., A. Hanyu, et al. (2008). "Bone morphogenetic protein signaling enhances invasion and bone metastasis of breast cancer cells through Smad pathway." Oncogene 27(49): 6322-33. Kavsak, P., R. K. Rasmussen, et al. (2000). "Smad7 binds to Smurf2 to form an E3 ubiquitin ligase that targets the TGF beta receptor for degradation." Mol Cell 6(6): 1365-75. Kim, D. W. and A. B. Lassar (2003). "Smad-dependent recruitment of a histone deacetylase/Sin3A complex modulates the bone morphogenetic protein-dependent transcriptional repressor activity of Nkx3.2." Mol Cell Biol 23(23): 8704-17. Kirsch, T., W. Sebald, et al. (2000). "Crystal structure of the BMP-2-BRIA ectodomain complex." Nat Struct Biol 7(6): 492-6. Kishi, M., Y. A. Pan, et al. (2005). "Mammalian SAD kinases are required for neuronal polarization." Science 307(5711): 929-32. Knockaert, M., G. Sapkota, et al. (2006). "Unique players in the BMP pathway: small C-terminal domain phosphatases dephosphorylate Smad1

54 to attenuate BMP signaling." Proc Natl Acad Sci U S A 103(32): 11940-5. Kodach, L. L., E. Wiercinska, et al. (2008). "The bone morphogenetic protein pathway is inactivated in the majority of sporadic colorectal cancers." Gastroenterology 134(5): 1332-41. Koh, H. J., T. Toyoda, et al. (2010). "Sucrose nonfermenting AMPK-related kinase (SNARK) mediates contraction-stimulated glucose transport in mouse skeletal muscle." Proc Natl Acad Sci U S A 107(35): 15541-6. Koinuma, D., M. Shinozaki, et al. (2003). "Arkadia amplifies TGF-beta superfamily signalling through degradation of Smad7." EMBO J 22(24): 6458-70. Kokudo, T., Y. Suzuki, et al. (2008). "Snail is required for TGFbeta-induced endothelial-mesenchymal transition of embryonic stem cell-derived endothelial cells." J Cell Sci 121(Pt 20): 3317-24. Korchynskyi, O. and P. ten Dijke (2002). "Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter." J. Biol. Chem. 277(7): 4883-4891. Kowanetz, M., P. Lonn, et al. (2008). "TGFbeta induces SIK to negatively regulate type I receptor kinase signaling." J Cell Biol 182(4): 655- 62. Kusanagi, K., H. Inoue, et al. (2000). "Characterization of a bone morphogenetic protein-responsive Smad- binding element." Mol. Biol. Cell 11(2): 555-65. Lai, C., J. Robinson, et al. (2011). "Elevation of WNT5A expression in polyp formation in Lkb1+/- mice and Peutz-Jeghers syndrome." J Pathol 223(5): 584-92. Lamouille, S., E. Connolly, et al. (2012). "TGF-beta-induced activation of mTOR complex 2 drives epithelial-mesenchymal transition and cell invasion." J Cell Sci 125(Pt 5): 1259-73. Lamouille, S. and R. Derynck (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(3): 437-51. Lan, F., J. M. Cacicedo, et al. (2008). "SIRT1 modulation of the acetylation status, cytosolic localization, and activity of LKB1. Possible role in AMP-activated protein kinase activation." J. Biol. Chem. 283(41): 27628-35. Larisch, S., Y. Yi, et al. (2000). "A novel mitochondrial septin-like protein, ARTS, mediates apoptosis dependent on its P-loop motif." Nat Cell Biol 2(12): 915-21. Lee, C. G., Y. W. Kim, et al. (2012). "Farnesoid X receptor protects hepato- cytes from injury by repressing miR-199a-3p, which increases levels of LKB1." Gastroenterology 142(5): 1206-1217 e7. Lee, N. Y., K. C. Kirkbride, et al. (2009). "The transforming growth factor- beta type III receptor mediates distinct subcellular trafficking and

55 downstream signaling of activin-like kinase (ALK)3 and ALK6 receptors." Mol Biol Cell 20(20): 4362-70. Liang, J., S. H. Shao, et al. (2007). "The energy sensing LKB1-AMPK pathway regulates p27(kip1) phosphorylation mediating the decision to enter autophagy or apoptosis." Nat. Cell Biol. 9(2): 218-24. Lin, X., Y.-G. Chen, et al., Eds. (2008). Transcriptional control via Smads. The TGF-β family. Cold Spring Harbor, Cold Spring Harbor La- boartory Press. Lin, X., X. Duan, et al. (2006). "PPM1A functions as a Smad phosphatase to terminate TGFbeta signaling." Cell 125(5): 915-28. Lin, X., M. Liang, et al. (2003a). "SUMO-1/Ubc9 promotes nuclear accumulation and metabolic stability of tumor suppressor Smad4." J Biol Chem 278(33): 31043-8. Lin, X., M. Liang, et al. (2003b). "Activation of transforming growth factor- beta signaling by SUMO-1 modification of tumor suppressor Smad4/DPC4." J Biol Chem 278(21): 18714-9. Linher-Melville, K., S. Zantinge, et al. (2012). "Liver kinase B1 expression (LKB1) is repressed by estrogen receptor alpha (ERalpha) in MCF-7 human breast cancer cells." Biochem Biophys Res Commun 417(3): 1063-8. Liu, I. M., S. H. Schilling, et al. (2009). "TGFβ-stimulated Smad1/5 phosphorylation requires the ALK5 L45 loop and mediates the pro- migratory TGFβ switch." EMBO J. 28(2): 88-98. Liu, L., L. Chen, et al. (2008). "Rapamycin inhibits F-actin reorganization and phosphorylation of focal adhesion proteins." Oncogene 27(37): 4998-5010. Liu, L., F. Li, et al. (2006). "Rapamycin inhibits cell motility by suppression of mTOR-mediated S6K1 and 4E-BP1 pathways." Oncogene 25(53): 7029-40. Liu, L., Y. Luo, et al. (2010). "Rapamycin inhibits cytoskeleton reorganization and cell motility by suppressing RhoA expression and activity." J Biol Chem 285(49): 38362-73. Liu, L., F. M. Siu, et al. (2012). "Akt blocks the tumor suppressor activity of LKB1 by promoting phosphorylation-dependent nuclear retention through 14-3-3 proteins." Am J Transl Res 4(2): 175-86. Lizcano, J. M., O. Göransson, et al. (2004). "LKB1 is a master kinase that activates 13 kinases of the AMPK subfamily, including MARK/PAR-1." EMBO J. 23(4): 833-843. Lo, R. S., Y. G. Chen, et al. (1998). "The L3 loop: a structural motif deter- mining specific interactions between SMAD proteins and TGF-beta receptors." EMBO J 17(4): 996-1005. Loffler, A. S., S. Alers, et al. (2011). "Ulk1-mediated phosphorylation of AMPK constitutes a negative regulatory feedback loop." Autophagy 7(7): 696-706. Londesborough, A., K. Vaahtomeri, et al. (2008). "LKB1 in endothelial cells is required for angiogenesis and TGFbeta-mediated vascular smooth muscle cell recruitment." Development 135(13): 2331-8.

56 Long, J., I. Matsuura, et al. (2003). "Repression of Smad transcriptional activity by PIASy, an inhibitor of activated STAT." Proc Natl Acad Sci U S A 100(17): 9791-6. Long, J., G. Wang, et al. (2004). "Repression of Smad4 transcriptional activity by SUMO modification." Biochem J 379(Pt 1): 23-9. Lonn, P., A. Moren, et al. (2009). "Regulating the stability of TGFbeta receptors and Smads." Cell Res 19(1): 21-35. Lonn, P., L. P. van der Heide, et al. (2010). "PARP-1 attenuates Smad- mediated transcription." Mol Cell 40(4): 521-32. Lonn, P., M. Vanlandewijck, et al. (2012). "Transcriptional induction of salt- inducible kinase 1 by transforming growth factor beta leads to negative regulation of type I receptor signaling in cooperation with the Smurf2 ubiquitin ligase." J Biol Chem 287(16): 12867-78. Lopez-Rovira, T., E. Chalaux, et al. (2002). "Direct binding of Smad1 and Smad4 to two distinct motifs mediates bone morphogenetic protein- specific transcriptional activation of Id1 gene." J. Biol. Chem. 277(5): 3176-3185. MacIver, N. J., J. Blagih, et al. (2011). "The liver kinase B1 is a central regulator of T cell development, activation, and metabolism." J Immunol 187(8): 4187-98. Martin, S. G. and D. St Johnston (2003). "A role for Drosophila LKB1 in anterior-posterior axis formation and epithelial polarity." Nature 421(6921): 379-84. Massague, J., J. Seoane, et al. (2005). "Smad transcription factors." Genes Dev 19(23): 2783-810. McInnes, K. J., K. A. Brown, et al. (2011). "Regulation of LKB1 expression by sex hormones in adipocytes." Int J Obes (Lond). Mirouse, V., C. P. Christoforou, et al. (2009). "Dystroglycan and perlecan provide a basal cue required for epithelial polarity during energetic stress." Dev. Cell 16(1): 83-92. Miyama, K., G. Yamada, et al. (1999). "A BMP-inducible gene, dlx5, regulates osteoblast differentiation and mesoderm induction." Dev. Biol. 208(1): 123-133. Moren, A., E. Raja, et al. (2011). "Negative regulation of TGFbeta signaling by the kinase LKB1 and the scaffolding protein LIP1." J Biol Chem 286(1): 341-53. Morikawa, M., D. Koinuma, et al. (2011). "ChIP-seq reveals cell type- specific binding patterns of BMP-specific Smads and a novel binding motif." Nucleic Acids Res 39(20): 8712-27. Moustakas, A. and C. H. Heldin (2005). "Non-Smad TGF-beta signals." J Cell Sci 118(Pt 16): 3573-84. Moustakas, A. and D. Kardassis (1998). "Regulation of the human p21/WAF1/Cip1 promoter in hepatic cells by functional interactions between Sp1 and Smad family members." Proc Natl Acad Sci U S A 95(12): 6733-8.

57 Mu, Y., R. Sundar, et al. (2011). "TRAF6 ubiquitinates TGFbeta type I receptor to promote its cleavage and nuclear translocation in cancer." Nat Commun 2: 330. Nakada, D., T. L. Saunders, et al. (2010). "Lkb1 regulates cell cycle and energy metabolism in haematopoietic stem cells." Nature 468(7324): 653-8. Nakajima, Y., T. Yamagishi, et al. (2000). "Mechanisms involved in valvu- loseptal endocardial cushion formation in early cardiogenesis: roles of transforming growth factor (TGF)-beta and bone morphogenetic protein (BMP)." Anat Rec 258(2): 119-27. Nakao, A., M. Afrakhte, et al. (1997). "Identification of Smad7, a TGFbeta- inducible antagonist of TGF-beta signalling." Nature 389(6651): 631-5. Nakashima, K., T. Takizawa, et al. (2001). "BMP2-mediated alteration in the developmental pathway of fetal mouse brain cells from neurogenesis to astrocytogenesis." Proc. Natl. Acad. Sci. U S A 98(10): 5868-73. Nakashima, K., M. Yanagisawa, et al. (1999). "Synergistic signaling in fetal brain by STAT3-Smad1 complex bridged by p300." Science 284(5413): 479-82. Nakashima, K., X. Zhou, et al. (2002). "The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation." Cell 108(1): 17-29. Namiki, T., A. Tanemura, et al. (2011). "AMP kinase-related kinase NUAK2 affects tumor growth, migration, and clinical outcome of human melanoma." Proc Natl Acad Sci U S A 108(16): 6597-602. Nath-Sain, S. and P. A. Marignani (2009). "LKB1 catalytic activity contributes to estrogen receptor α signaling." Mol. Biol. Cell 20(11): 2785-95. Oakhill, J. S., R. Steel, et al. (2011). "AMPK is a direct adenylate charge- regulated protein kinase." Science 332(6036): 1433-5. Ogryzko, V. V., R. L. Schiltz, et al. (1996). "The transcriptional coactivators p300 and CBP are histone acetyltransferases." Cell 87(5): 953-9. Oh, S. P., T. Seki, et al. (2000). "Activin receptor-like kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis." Proc Natl Acad Sci U S A 97(6): 2626-31. Onichtchouk, D., Y. G. Chen, et al. (1999). "Silencing of TGF-β signalling by the pseudoreceptor BAMBI." Nature 401(6752): 480-485. Otto, F., A. P. Thornell, et al. (1997). "Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development." Cell 89(5): 765-771. Padua, D. and J. Massague (2009). "Roles of TGFbeta in metastasis." Cell Res 19(1): 89-102. Padua, D., X. H. Zhang, et al. (2008). "TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4." Cell 133(1): 66-77.

58 Pardali, K., A. Kurisaki, et al. (2000). "Role of Smad proteins and transcription factor Sp1 in p21(Waf1/Cip1) regulation by transforming growth factor-beta." J Biol Chem 275(38): 29244-56. Partanen, J. I., A. I. Nieminen, et al. (2007). "Suppression of oncogenic properties of c-Myc by LKB1-controlled epithelial organization." Proc Natl Acad Sci U S A 104(37): 14694-9. Partanen, J. I., T. A. Tervonen, et al. (2012). "Tumor suppressor function of Liver kinase B1 (Lkb1) is linked to regulation of epithelial integrity." Proc Natl Acad Sci U S A 109(7): E388-97. Perlman, R., W. P. Schiemann, et al. (2001). "TGF-beta-induced apoptosis is mediated by the adapter protein Daxx that facilitates JNK activation." Nat Cell Biol 3(8): 708-14. Piccirillo, S. G., B. A. Reynolds, et al. (2006). "Bone morphogenetic proteins inhibit the tumorigenic potential of human brain tumour- initiating cells." Nature 444(7120): 761-765. Pouponnot, C., L. Jayaraman, et al. (1998). "Physical and functional interaction of SMADs and p300/CBP." J Biol Chem 273(36): 22865-8. Roelen, B. A., O. S. Cohen, et al. (2003). "Phosphorylation of threonine 276 in Smad4 is involved in transforming growth factor-beta-induced nuclear accumulation." Am J Physiol Cell Physiol 285(4): C823-30. Rowan, A., V. Bataille, et al. (1999). "Somatic mutations in the Peutz- Jeghers (LKB1/STKII) gene in sporadic malignant melanomas." J Invest Dermatol 112(4): 509-11. Ruzinova, M. B. and R. Benezra (2003). "Id proteins in development, cell cycle and cancer." Trends Cell Biol 13(8): 410-8. Sakamoto, K. and G. D. Holman (2008). "Emerging role for AS160/TBC1D4 and TBC1D1 in the regulation of GLUT4 traffic." Am. J. Physiol. Endocrinol. Metab. 295(1): E29-37. Sanchez-Elsner, T., L. M. Botella, et al. (2001). "Synergistic cooperation between hypoxia and transforming growth factor-beta pathways on human vascular endothelial growth factor gene expression." J Biol Chem 276(42): 38527-35. Sapkota, G., M. Knockaert, et al. (2006). "Dephosphorylation of the linker regions of Smad1 and Smad2/3 by small C-terminal domain phosphatases has distinct outcomes for bone morphogenetic protein and transforming growth factor-beta pathways." J Biol Chem 281(52): 40412-9. Sapkota, G. P., A. Kieloch, et al. (2001). "Phosphorylation of the protein kinase mutated in Peutz-Jeghers cancer syndrome, LKB1/STK11, at Ser431 by p90(RSK) and cAMP-dependent protein kinase, but not its farnesylation at Cys(433), is essential for LKB1 to suppress cell growth." J. Biol. Chem. 276(22): 19469-82. Satow, R., A. Kurisaki, et al. (2006). "Dullard promotes degradation and dephosphorylation of BMP receptors and is required for neural induction." Dev. Cell 11(6): 763-74.

59 Schwappacher, R., J. Weiske, et al. (2009). "Novel crosstalk to BMP signalling: cGMP-dependent kinase I modulates BMP receptor and Smad activity." EMBO J. 28(11): 1537-50. Seoane, J., H. V. Le, et al. (2004). "Integration of Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation." Cell 117(2): 211-23. Seoane, J., C. Pouponnot, et al. (2001). "TGFbeta influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b." Nat Cell Biol 3(4): 400-8. Shaw, R. J., M. Kosmatka, et al. (2004). "The tumor suppressor LKB1 kinase directly activates AMP-activated kinase and regulates apoptosis in response to energy stress." Proc. Natl. Acad. Sci. U S A 101(10): 3329-35. Shi, W., C. Chang, et al. (2007). "Endofin acts as a Smad anchor for receptor activation in BMP signaling." J. Cell Sci. 120(Pt 7): 1216-24. Shi, W., C. Sun, et al. (2004). "GADD34-PP1c recruited by Smad7 dephosphorylates TGFbeta type I receptor." J Cell Biol 164(2): 291-300. Shi, Y. and J. Massagué (2003). "Mechanisms of TGF-β signaling from cell membrane to the nucleus." Cell 113: 685-700. Shimo, T., T. Nakanishi, et al. (2001). "Involvement of CTGF, a hypertroph- ic chondrocyte-specific gene product, in tumor angiogenesis." On- cology 61(4): 315-22. Shore, E. M., M. Xu, et al. (2006). "A recurrent mutation in the BMP type I receptor ACVR1 causes inherited and sporadic fibrodysplasia ossificans progressiva." Nat Genet 38(5): 525-7. Siegel, P. M. and J. Massague (2003). "Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer." Nat Rev Cancer 3(11): 807- 21. Simonsson, M., C. H. Heldin, et al. (2005). "The balance between acetylation and deacetylation controls Smad7 stability." J Biol Chem 280(23): 21797-803. Simonsson, M., M. Kanduri, et al. (2006). "The DNA binding activities of Smad2 and Smad3 are regulated by coactivator-mediated acetylation." J Biol Chem 281(52): 39870-80. Smith, D. P., S. I. Rayter, et al. (2001). "LIP1, a cytoplasmic protein func- tionally linked to the Peutz-Jeghers syndrome kinase LKB1." Hum. Mol. Genet. 10(25): 2869-2877. Smith, D. P., S. I. Rayter, et al. (2001). "LIP1, a cytoplasmic protein func- tionally linked to the Peutz-Jeghers syndrome kinase LKB1." Hum Mol Genet 10(25): 2869-77. Song, P., Z. Xie, et al. (2008). "Protein kinase Cζ-dependent LKB1 serine 428 phosphorylation increases LKB1 nucleus export and apoptosis in endothelial cells." J. Biol. Chem. 283(18): 12446-55. Sorrentino, A., N. Thakur, et al. (2008). "The type I TGF-beta receptor en- gages TRAF6 to activate TAK1 in a receptor kinase-independent manner." Nat Cell Biol 10(10): 1199-207.

60 Sowa, H., H. Kaji, et al. (2004). "Menin is required for bone morphogenetic protein 2- and transforming growth factor β-regulated osteoblastic differentiation through interaction with Smads and Runx2." J. Biol. Chem. 279(39): 40267-75. Spicer, J., S. Rayter, et al. (2003). "Regulation of the Wnt signalling component PAR1A by the Peutz-Jeghers syndrome kinase LKB1." Onco- gene 22(30): 4752-6. Staller, P., K. Peukert, et al. (2001). "Repression of p15INK4b expression by Myc through association with Miz-1." Nat Cell Biol 3(4): 392-9. Stroschein, S. L., W. Wang, et al. (1999). "Negative feedback regulation of TGF-beta signaling by the SnoN oncoprotein." Science 286(5440): 771-4. Suzuki, A., A. Raya, et al. (2006). "Nanog binds to Smad1 and blocks bone morphogenetic protein-induced differentiation of embryonic stem cells." Proc Natl Acad Sci U S A 103(27): 10294-9. Tachibana, I., M. Imoto, et al. (1997). "Overexpression of the TGFbeta- regulated zinc finger encoding gene, TIEG, induces apoptosis in pancreatic epithelial cells." J Clin Invest 99(10): 2365-74. Teichert-Kuliszewska, K., M. J. Kutryk, et al. (2006). "Bone morphogenetic protein receptor-2 signaling promotes pulmonary arterial endothelial cell survival: implications for loss-of-function mutations in the pa- thogenesis of pulmonary hypertension." Circ. Res. 98(2): 209-17. Thaiparambil, J. T., C. M. Eggers, et al. (2012). "AMPK regulates mitotic spindle orientation through phosphorylation of myosin regulatory light chain." Mol Cell Biol. Tiainen, M., K. Vaahtomeri, et al. (2002). "Growth arrest by the LKB1 tumor suppressor: induction of p21(WAF1/CIP1)." Hum Mol Genet 11(13): 1497-504. Tseng, Y. H., E. Kokkotou, et al. (2008). "New role of bone morphogenetic protein 7 in brown adipogenesis and energy expenditure." Nature 454(7207): 1000-4. Tsukazaki, T., T. A. Chiang, et al. (1998). "SARA, a FYVE domain protein that recruits Smad2 to the TGFbeta receptor." Cell 95(6): 779-91. Tu, A. W. and K. Luo (2007). "Acetylation of Smad2 by the co-activator p300 regulates activin and transforming growth factor beta response." J Biol Chem 282(29): 21187-96. Urist, M. R. (1965). "Bone: formation by autoinduction." Science 150(698): 893-9. Urist, M. R. and B. S. Strates (1971). "Bone morphogenetic protein." J Dent Res 50(6): 1392-406. Vaahtomeri, K., E. Ventela, et al. (2008). "Lkb1 is required for TGFbeta- mediated myofibroblast differentiation." J Cell Sci 121(Pt 21): 3531- 40. Valderrama-Carvajal, H., E. Cocolakis, et al. (2002). "Activin/TGF-beta induce apoptosis through Smad-dependent expression of the lipid phosphatase SHIP." Nat Cell Biol 4(12): 963-9.

61 Vallenius, T., K. Vaahtomeri, et al. (2011). "An association between NUAK2 and MRIP reveals a novel mechanism for regulation of actin stress fibers." J Cell Sci 124(Pt 3): 384-93. Wang, W., F. V. Mariani, et al. (2000). "Ski represses bone morphogenic protein signaling in Xenopus and mammalian cells." Proc Natl Acad Sci U S A 97(26): 14394-9. Watanabe, Y., S. Itoh, et al. (2010). "TMEPAI, a transmembrane TGF-beta- inducible protein, sequesters Smad proteins from active participation in TGF-beta signaling." Mol Cell 37(1): 123-34. Weaver, M., J. M. Yingling, et al. (1999). "Bmp signaling regulates proximal-distal differentiation of endoderm in mouse lung development." Development 126(18): 4005-15. Vinals, F., T. Lopez-Rovira, et al. (2002). "Inhibition of PI3K/p70 S6K and p38 MAPK cascades increases osteoblastic differentiation induced by BMP-2." FEBS Lett. 510(1-2): 99-104. Wingo, S. N., T. D. Gallardo, et al. (2009). "Somatic LKB1 mutations promote cervical cancer progression." PLoS One 4(4): e5137. Xia, Y., P. B. Yu, et al. (2007). "Repulsive guidance molecule RGMa alters utilization of bone morphogenetic protein (BMP) type II receptors by BMP2 and BMP4." J Biol Chem 282(25): 18129-40. Xie, M., D. Zhang, et al. (2006). "A pivotal role for endogenous TGF-beta- activated kinase-1 in the LKB1/AMP-activated protein kinase energy-sensor pathway." Proc Natl Acad Sci U S A 103(46): 17378-83. Xie, Z., Y. Dong, et al. (2009). "Identification of the serine 307 of LKB1 as a novel phosphorylation site essential for its nucleocytoplasmic transport and endothelial cell angiogenesis." Mol. Cell. Biol. 29(13): 3582-96. Yamamoto, T., O. Kozawa, et al. (2001). "Involvement of p38 MAP kinase in TGF-beta-stimulated VEGF synthesis in aortic smooth muscle cells." J Cell Biochem 82(4): 591-8. Yan, X., Z. Lin, et al. (2009). "Human BAMBI cooperates with Smad7 to inhibit transforming growth factor-beta signaling." J Biol Chem 284(44): 30097-104. Yang, X., L. Long, et al. (2005). "Dysfunctional Smad signaling contributes to abnormal smooth muscle cell proliferation in familial pulmonary arterial hypertension." Circ. Res. 96(10): 1053-63. Yoshikawa, H., T. Nakase, et al. (2004). "Bone morphogenetic proteins in bone tumors." J. Orthop. Sci. 9(3): 334-40. Yu, J., L. Pan, et al. (2010). "MTMR4 attenuates transforming growth factor beta (TGFbeta) signaling by dephosphorylating R-Smads in endosomes." J Biol Chem 285(11): 8454-62. Yu, P. B., D. Y. Deng, et al. (2008). "BMP type I receptor inhibition reduces heterotopic [corrected] ossification." Nat. Med. 14(12): 1363-9. Zawel, L., J. L. Dai, et al. (1998). "Human Smad3 and Smad4 are sequence- specific transcription activators." Mol Cell 1(4): 611-7.

62 Zeqiraj, E., B. M. Filippi, et al. (2009). "Structure of the LKB1-STRAD- MO25 complex reveals an allosteric mechanism of kinase activation." Science 326(5960): 1707-11. Zhang, F., T. Qiu, et al. (2009). "Sustained BMP signaling in osteoblasts stimulates bone formation by promoting angiogenesis and osteoblast differentiation." J Bone Miner Res 24(7): 1224-33. Zhang, L., F. Zhou, et al. (2012). "USP4 is regulated by AKT phosphorylation and directly deubiquitylates TGF-beta type I receptor." Nat Cell Biol 14(7): 717-26. Zhang, Y. W., N. Yasui, et al. (2000). "A RUNX2/PEBP2α A/CBFA1 mutation displaying impaired transactivation and Smad interaction in cleidocranial dysplasia." Proc. Natl. Acad. Sci. U S A 97(19): 10549-10554. Zhao, J., P. J. Sime, et al. (1998). "Epithelium-specific adenoviral transfer of a dominant-negative mutant TGF-beta type II receptor stimulates embryonic lung branching morphogenesis in culture and potentiates EGF and PDGF-AA." Mech Dev 72(1-2): 89-100. Zheng, B. and L. C. Cantley (2007). "Regulation of epithelial tight junction assembly and disassembly by AMP-activated protein kinase." Proc. Natl. Acad. Sci. U S A 104(3): 819-22. Zheng, B., J. H. Jeong, et al. (2009). "Oncogenic B-RAF negatively regulates the tumor suppressor LKB1 to promote melanoma cell proliferation." Mol Cell 33(2): 237-47. Zhong, D., X. Liu, et al. (2008). "LKB1 is necessary for Akt-mediated phosphorylation of proapoptotic proteins." Cancer Res 68(18): 7270-7. Zhou, L., C. R. Dey, et al. (1996). "Arrested lung morphogenesis in trans- genic mice bearing an SP-C-TGF-beta 1 chimeric gene." Dev Biol 175(2): 227-38.

63 Acknowledgements

In the beginning, it was just a dream  but thanks to the support of many people in many different ways, I am here at the finish line and I hope that this is also the beginning of much more to come.

Firstly, I would like to thank my main supervisor, Dr. Aristidis Moustakas. Thank you for being a great mentor on top of being a good scientist, for the huge amount of time you give for discussions, the freedom of exploring our ideas, the trust and also tolerance when we go through some difficult time. You have shared a lot of science and knowledge with me and I really thank you for that.

To my co-supervisor Prof. Carl-Henrik Heldin, I admire your passion for science and tireless effort for it. I will always remember your positive and humble nature as a leader.

I am grateful to the project collaborators. To Anita Morèn, thank you for helping me a lot in the lab when I just started and for our LKB1 joint effort. Thank you to Peter Lönn, Michael Vanlandewijck, and Yukihide Wata- nabe for the SIK paper that we finished recently. To Lars van der Heide, Jon Carthy and Masoud Razmara, for the ongoing NUAK2 collaboration. To Laia Caja for your interest in the LKB1-glioma study. To our students, Ashish Singh and Iryna Voytyuk for some help with the experiments.

Additionally, to Agata Zieba, Karolina Edlund, Dr. Patrick Micke, Robin Vuilleumier and Dr. Giorgos Pyrowolakis, thank you for very en- thusiastic collaboration. A special thanks to Robin for your kindness and truly persistent effort with LKB1 in the drosophila model, it has been really encouraging for me.

To the TGFβ lab members (Rosita Bergström, Sylvie Thuault, Katia Savary, Kaoru Kahata, E-Jean Tan, Stefan Termen, Varun Maturi, Mahsa Dadras), past or present, I thank you all for scientific or non- scientific support.

To many other colleagues and friends at the Ludwig Institute, thank you for making the institute a very warm and friendly place. Next door lab dwellers

64 Anders Sunqvist, Ria Vasilaki and Noopur Thakur, you are my favourite trio  Thanks for your encouragement and for listening to my insanity from time to time 

I now move to the office people. Thank you Peter, Michael, Markus Dahl, Glenda Eger, Merima Mehic, Berit Bernert, Carmen Herrera for a lot of interesting and funny office talk which many times distract me from work frustrations. Additional thanks to Michael, for fixing my computer problems and my bike, which got punctured three times in one autumn.  Speaking of bike, I won’t forget the super friendly bike doctor Prof. Ulf Hellman, thank you for being so cheerful, it’s contagious.

Berit, thank you for sharing faith and reminded me many times when it got tough, about what was important and not. Fatima Burovic, my ‘rol- lercoaster’ friend, through you I met many more wonderful people whom I could call friends.

To my early mentor in Australia, Prof. Barry Iacopetta, thank you for your support and encouragement, that continues even long after I left UWA. I really appreciate all the help you gave me. To my mentors in Singapore, Prof. Go Mei Lin and Prof. Lim Lee Yong, I am very thankful for your kindness and trust which opened the first door for me and allowed me to step further. I learned and appreciate very much your sincere and firm characters.

To friends along the way, Baidah, Kum Chew, Pei Yi, Evelyn, John and Hazel, Monika and Steve, Diana and Christer, Xiang, Tari, Linda, Lia, Tante Anna, Tante Anggraeni. It is not always easy to move from one place to another but knowing you and sharing with you, laughter and strug- gles and faith, have made me stronger. Yukihide, I thank you for your affec- tion, patience and understanding. It has been both fun and meaningful to spend time and do even the simplest things with you. I look forward to learn more.

Finally, to my family in Jakarta, though we are physically far apart, you are always with me. I cannot thank you enough for always supporting me, even when my wishes were strange and you disagreed, and also when time was very tough. I am so lucky to have you. To papa, although you had to leave early, thank you for sharing many life values during the short time we had together. My dear family, I always have your love with me all the time. Soli Deo Gloria