Regulation of transforming growth factor beta (TGFβ) at the level of nuclear entry

Transforming Growth Factor β (TGFβ) has broad reaching biological actions spanning development, homeostasis and disease. Whilst our understanding of many of the molecules involved in the TGFβ signalling cascade is growing, precisely how these factors deliver their messages into the nucleus remain elusive. The aim of this study was to investigate the cellular machinery and mechanisms involved in delivering proteins involved in the TGFβ cascade into the nucleus. Using novel constructs and diverse cell biology tools and techniques, the effects and consequences of TGFβ stimulation on the expression, activity, sub-cellular localisation and molecular associations of various proteins within the cell were assessed. The primary mediators of TGFβ responses from the receptors into the nucleus are the two Smad proteins, Smad2 and Smad3. Despite significant sequence homology and apparently identical receptor regulation and DNA sequence recognition sequences, knockout and overexpression studies indicate each of these proteins have very differing roles in homeostasis, tumour/fibrosis suppression and driving tumour/fibrosis progression (Yamamoto 1999, Santiago 2005, Hoot 2008, Meng 2010). Here (Chapter 4), we identify SNX9 as being required for Smad3 (but not Smad2) nuclear translocation and cell stimulation. SNX9 principally interacts with phosphorylated (p) Smad3 independent of Smad2 or Smad4 and promotes more rapid nuclear delivery from that observed independent of ligand. Although SNX9 does not bind Imp7, Impβ, Nup153 or Nup214, it mediates the association of pSmad3 with Imp8 and the nuclear membrane via Phox Homology (PX) and Bin, Amphiphysin, Rvs (BAR) domain phosphoinositide binding motifs. This facilitates recruitment of pSmad3/SNX9 to the nuclear pore, heterodimerization of Imp8 with Impβ, and nuclear translocation of pSmad3, but not SNX9. The demonstration that Smad3 is regulated in a distinct manner from Smad2 provides the opportunity to develop intervention strategies to enhance or dampen specific aspects of the cellular response to TGFβ. Smad phosphorylation in response to TGFβ stimulation occurs in all cell types expressing TGFβ receptors, however in a number of mesenchymal cell lines TGFβ receptors also activate PAK2, contributing to a distinct fibroblastic TGFβ response (Wilkes 2003, Sato 2013). Within this study (Chapter 5), we document PAK2 phosphorylation of R-Smads at a site distinct from the C terminal SSxS motif recognized by TGFβR1. Furthermore, R-Smad phosphorylation at the PAK2 site prolongs the duration of the receptor-recognized phosphorylation sites, by preventing binding of PPM1A, the nuclear phosphatase that dephosphorylates the R-Smad receptor sites. For many cytokines, the accepted model of signal propogation requires a transmembrane receptor complex that becomes enzymatically active upon binding the extracellular ligand, to transmit the signal through enzymatic modification of soluble, cytoplasmic proteins that are either directly transported to the nucleus, or initiate a cascade of enzymatic reactions that ultimately lead to transcriptional alterations. The accepted model of TGFβ signalling fits well within that paradigm. However, in recent years, a number of plasma membrane-embedded tyrosine kinase receptors have been documented to traffic from the cell surface to the nucleus, in addition to activating signalling cascades in soluble, cytoplasmic proteins. Evidence is accumulating that a pool TGFβ receptors are also trafficking to the nucleus upon ligand stimulation (Mu 2011, Chandra 2012, Gudey 2014). We document (in Chapter 6) nuclear trafficking of TGFβ receptors occurs in normal and transformed cells with both TGFβR1 and TGFβR2 required for nuclear entry. Receptors pass through the Golgi apparatus, COPI vesicles, endoplasmic reticulum, retrotranslocon and nuclear pore en route to the nucleus. Upon nuclear entry, receptors are not soluble, instead residing in the inner nuclear membrane before incorporation into Promyolecytic Leukemia (PML) nuclear bodies. In the nucleus, TGFβR1 phosphorylates a number of transcription factors, including ATF/CREB and are required for the robust induction of a subset or TGFβ/Smad genes. Being that many of the genes we identified as requiring the presence of nuclear TGFβ receptors for TGFβ regulation have previously been reported as being Smad-dependent, we sought to investigate this apparent data conflict. We document (in Chapter 7), TGFβ receptors in the nucleus maintain kinase activity, and the phosphorylation of transcription factors such as ATF2, CREB and sp1 increases the histone acetyltransferase (HAT) activity of these transcription factors, leading to the exposure of Smad Binding Elements (SBEs) in the promoter regions surrounding the phosphorylated ATF2/CREB/sp1. Once exposed, nuclear pSmad2/3 bind these SBEs in the promoters, prompting a full transcriptional response. In this way the presence of nuclear receptors works in co-operation with the Smads, each being required (but not sufficient) for the TGFβ-induced effect. Additionally (in Chapter 6), we report a short region of 14 amino acids in TGFβR2 that binds to the retromer complex is required for nuclear translocation of the receptors. While retromer binding is maintained in TGFβ superfamily members, such as BMPR2 and ActR2, these receptors fail to translocate to the nucleus upon stimulation. We report a single lysine (K488) that is not present in BMPR2 or ActR2 is responsible for conferring nuclear trafficking ability. We also present evidence ubiquitination of this site may be the cue to select the nuclear-bound subset of receptors from the larger pool being degraded. Our aim has been to examine how the various components of the TGFβ pathway are relayed into the nucleus, the routes, the factors and the mechanisms. Along the way we have stumbled onto unexpected results that have been problematic to explain, as well as pieces of information that have helped unite seemingly contradictory data within the field. As with most studies our answers have opened up as many questions as we've answered but, at least in some small part, we have gained insight too. Uncovering differential mechanisms for nuclear translocation of Smad2 and Smad3 gives validity to models claiming cells actively balance these two signals in various cell contexts. Extending the duration of nuclear signalling of Smads by phosphorylation at non-receptor sites by kinases activated only in select cell types also provides a mechanism for cells to balance and fine tune Smad signals to meet the needs of the cell. Uncovering the nuclear translocation of the TGFβ receptors adds another TGFβ activated kinase that contributes to the nuclear message. Reports of differential expression and multiple mechanisms to deliver these active kinases to the nucleus imply this too is actively regulated to balance with Smad and other non-Smad signals. As we move forward in our attempts to manage disease states driven by aberrant TGFβ signalling we might do well if we switch our thinking from general inhibitors of all TGFβ signalling or blocking the activity of single components of the pathway, but rather look to influence interactions that influence the balance of these factors. Admittedly our understanding of the mechanisms regulating the balance between these factors are in their infancy but with concept of tumour microenvironment and stromal interactions already firmly in place amongst cancer researchers, it seems plausible this way of thinking will permeate throughout research and influence those of us advancing TGFβ understanding. While the age of identifying new TGFβ-specific signalling factors may be at its twilight, a new exciting age of exploring the mechanisms that balance these factors and how other cell stimuli cross talk and influence the TGFβ signal is just beginning. We contend a major bottleneck to regulation is importing these factors into and out of the nucleus and factors like SNX9, retromer complex and kinases such as PAK2 play an important role in this process. Insight from these studies sheds light on new regulatory controls utilized to fine tune the TGFβ signalling cascade with only a limited number of core protein components, with the potential to adapt to a wide range of environmental cues across numerous cell types. Modulating the TGFβ signal at the level of nuclear entry or exit not only enhances our knowledge of how this powerful signal impacts the cell, but also provides essential insight into therapeutic strategies for management of the numerous clinical manifestations that occur due to imbalances in TGFβ signalling that regularly occur in human disease.