UNDERSTANDING THE CELLULAR AND MOLECULAR MECHANISMS OF CEREBRAL CAVERNOUS MALFORMATION 3 (CCM3)
Item Type Dissertation
Authors Mansaray-Storms, Zainab Y.
Rights Attribution-NonCommercial-NoDerivatives 4.0 International
Download date 01/10/2021 21:17:52
Item License http://creativecommons.org/licenses/by-nc-nd/4.0/
Link to Item http://hdl.handle.net/20.500.12648/1888
UNDERSTANDING THE CELLULAR AND MOLECULAR MECHANISMS OF CEREBRAL CAVERNOUS MALFORMATION 3 (CCM3)
Zainab Y. Mansaray-Storms
A Dissertation in the Department of Neuroscience and Physiology
Submitted in partial fulfillment of the requirements for the degree of the Doctor of
Philosophy in the College of Graduate Studies of the State University of New York,
Upstate Medical University, Syracuse, NY.
Approved by: ______
(Sponsor’s Signature)
Date ______08/10/2016______
For my Dad, my favorite biologist.
ii
ACKNOWLEDGEMENTS
First, I would like to thank my mentor, Dr. Brian Howell, for his advice,
encouragement and mentorship as I completed this work. Your guidance and support was
invaluable as I developed into the scientist I am today and I am truly grateful for this.
I would like to thank Drs. Olson, Matthews, Lewis, and Viczian for their
comments and critiques through the years as the project developed and for being integral
in my development as a scientist. Thank you to Dr. Paul Barnes for contributing to the
study and for taking the time to read this thesis and serve as the external advisor.
Thank you to previous and current members of the Howell Lab especially Dr.
Tohru Matsuki and Jianhua Chen for all the support. Thank you Katie for your
friendship! I would also like to thank members of the Olson and Matthews lab for all the
support and especially Dr. Chrissa Dwyer for her encouragement and mentorship. Thank
you to all my friends in the Neuroscience Department as well as the Upstate community
for your help, encouragement and friendships.
Finally, I would like to thank my family and friends, both near and far, for the outpouring love and support as I completed this thesis. I will FINALLY be done with school and stop using it as an excuse. Thank you to my husband, Justin, for supporting my goals and sharing in all the pain, joy and excitement of the past years. I would not have done this without you JJ. Last but definitely not least, Elliott. Thank you for making life infinitely happier and enabling me to define my priorities.
iii
DISSERTATION ABSTRACT
UNDERSTANDING THE CELLULAR AND MOLECULAR MECHANISMS OF
CEREBRAL CAVERNOUS MALFORMATION 3 (CCM3)
Zainab Y. Mansaray-Storms
Thesis Advisor: Brian Howell, Ph.D.
Cerebral cavernous malformation 3 (CCM3) is one of three genes which when mutated plays a role in the neurovascular disease, cerebral cavernous malformation. Through a number of diverse binding partners, CCM3 plays a critical role in modulating several processes including cell survival, migration and vascular development. However, how
CCM3 regulates many of these pathways remains unclear. An interaction with Stk25, a serine/threonine kinase with roles in cell polarity suggests CCM3 might play a role in the functions of Stk25. Here we characterize a role for CCM3 in cellular polarity. We identify a function for CCM3 in epithelial polarity through its association with and regulation of the conserved LKB1 signaling pathway. We find that CCM3 associates with
STRADα, the regulatory pseudokinase of the LKB1 complex, and is necessary for
LKB1-mediated cell polarization. To determine whether this novel association of CCM3 with STRADα and LKB1 pathway plays a role in CCM3-phenotype in endothelial cells, we analyzed the gene expression profiles of endothelial cells deficient in CCM3 protein.
We identified changes in gene expression induced by CCM3 knockdown, particularly a significant downregulation in expression of cell adhesion molecules, a dysregulation of extracellular matrix signaling and an activation of p53 signaling pathway. This work
iv
defines a novel regulatory role for CCM3 in epithelial cell polarity and provides preliminary insights into downstream signaling pathways affected by the reduction of
CCM3 in endothelial cells with potential impact in CCM disease pathogenesis.
v
TABLE OF CONTENTS
TITLE PAGE ...... I DEDICATION...... II ACKNOWLEDGEMENTS ...... III DISSERTATION ABSTRACT...... IV TABLE OF CONTENTS ...... VI LIST OF FIGURES ...... VII LIST OF ABBREVIATIONS ...... IX CHAPTER 1: GENERAL INTRODUCTION ...... 12 1. INTRODUCTION ...... 13 2. CEREBRAL CAVERNOUS MALFORMATIONS (CCMS) ...... 14 3. CCM3: STRUCTURE, FUNCTIONS AND SIGNALING PATHWAYS ...... 18 4. THE LKB1 SIGNALING PATHWAY ...... 20 5.THE ROLE OF STK25 AND THE GOLGI APPARATUS IN CELL POLARITY ...... 25 6. OVERVIEW OF CHAPTERS 2 AND 3 ...... 28 REFERENCES ...... 30 FIGURES...... 38 CHAPTER 2: CCM3 BINDS STRADΑ AND REGULATES LKB1-STRAD DEPENDENT EPITHELIAL CELL POLARIZATION...... 45 ABSTRACT ...... 46 INTRODUCTION ...... 47 MATERIALS AND METHODS ...... 50 RESULTS ...... 55 DISCUSSION ...... 63 REFERENCES ...... 67 FIGURES...... 73 SUPPLEMENTAL FIGURES ...... 86 CHAPTER 3: DIFFERENTIAL GENE EXPRESSION PROFILING OF ENDOTHELIAL CELLS DEFICIENT IN CEREBRAL CAVERNOUS MALFORMATION 3 (CCM3) ...... 97 ABSTRACT ...... 98 INTRODUCTION ...... 99 MATERIALS AND METHODS ...... 102 RESULTS ...... 108 DISCUSSION ...... 115 REFERENCES ...... 121 TABLES ...... 125 FIGURES...... 133 SUPPLEMENTAL TABLES AND FIGURES ...... 144 CHAPTER 4: DISCUSSION AND FUTURE DIRECTIONS ...... 148 REFERENCES ...... 165 FIGURES...... 172
vi
LIST OF FIGURES
Figure 1.1. Schematic representation of CCM3 complexes, signaling pathways 39 and cellular functions. Figure 1.2 CCM3 protein structure and binding partners. 41 Figure 1.3. Schematic representation of the LKB1 Signaling Pathway. 43 Figure 2.1. STRADα interacts with both wildtype and mutant CCM3. 74 Figure 2.2. STRADα interacts with the FAT-homology domain of CCM3. 76 Figure 2.3. CCM3 acts on LKB1-STRADα pathway to regulate epithelial 78 polarization. Figure 2.4. LKB1 and STRADα protein expression is CCM3- and proteasome- 80 dependent. Figure 2.5. CCM3 expression differentially regulates expression of LKB1 82 effectors. Figure 2.6. Model of CCM3 as a binding partner of STRADα acting to modulate 84 LKB1-mediated cell polarity. Supplemental Figure 2.1. STK25 does not interact with mutated form of CCM3. 87 Supplemental Figure 2.2. CCM3 deletion enhances the interaction between 89 STK25 and STRADα. Supplemental Figure 2.3. CCM3 shRNA vectors effectively reduce the expression 91 of CCM3 in HEK293T and W4 epithelial cells. Supplemental Figure 2.4. LKB1 and STRADα protein expression are dependent 93 on CCM3 in 293T cells. Supplemental Figure 2.5. Stk25 knockdown does not influence LKB1 or 95 STRADα protein expression in W4 cells. Table 3.1. Only modest changes in gene expression of LKB1, STRAD and 126 downstream effectors in endothelial cells. Table 3.2. A selection of highly differentially expressed genes in CCM3 127 knockdown cells. Table 3.3. Top biological functions altered by depletion of CCM3 in endothelial 130 cells. Table 3.4. p53 signaling genes are differentially expressed in CCM3 knockdown 131 cells Figure 3.1. LKB1 and STK25 protein expressions are dependent on CCM3 in 134 endothelial cells. Figure 3.2. Quantitative RT-PCR analysis of LKB1 and STRAD gene transcripts 136 indicates an increase in expression when CCM3 is knockdown. Figure 3.3. Quantitative RT-PCR confirms the microarray predictions of a subset 138 of selected genes. Figure 3.4 Exploratory bioinformatics pathway analysis of top differential 140
vii
upstream-regulated gene networks. Figure 3.5. Quantitative RT-PCR validation of highly regulated genes on the p53 142 signaling pathway. Supplemental Table 3.1. List of lentivirus shRNA sequences used in this study. 145 Supplemental Table 3.2. List of human primer sequences used for qRT-PCR 146 experiments. Supplemental Table 3.3. List of human primer sequences used for qRT-PCR 147 experiments on p53 signaling. Figure 4.1. Proposed Working Model of CCM3 in LKB1 signaling pathway. 173
viii
LIST OF ABBREVIATIONS
AMPK AMP activated protein kinase aPKC Atypical protein kinase C BMP6 Bone morphogenetic protein 6 CAMKK2 Calcium/calmodulin-dependent protein kinase kinase 2 CCM1 Cerebral cavernous malformation 1 CCM2 Cerebral cavernous malformation 2 CCM3 Cerebral cavernous malformation 3 CCM3Δ MYC-tagged CCM3 deletion mutation (lacks 18 amino acids from dimerization domain) CCMs Cerebral cavernous malformations Cdc2 Cell division control protein 2 Cdc37 Cell division cycle 37 CDKN1A Cyclin-Dependent Kinase Inhibitor 1A CDKN2A Cyclin-Dependent Kinase Inhibitor 2A cDNA Complementary DNA DDIT3 DNA damage inducible transcript 3 DMEM Dulbecco’s modified eagle medium ECL Enhanced chemiluminescence ECM Extracellular matrix EndMT Endothelial-mesenchymal transition ERM Ezrin/radixin/moesin proteins FABP4 Fatty acid binding protein 4 FAP-1 Fas-associated phosphatase 1 FAT-H Focal adhesion targeting-homology FDR False discovery rate FOXO4 Forkhead box O4 GCKIII Germinal center kinase III GM130 Golgi matrix protein 130 HAEC Human aortic endothelial cell HEK293 Human embryonic kidney 293 cell line HIF1 Hypoxia inducible factor 1 HP1 Hydrophobic pocket 1 HPRT Hypoxanthine-guanine phosphoribosyltransferase Hsp90 Heat-shock protein 90 ICAP1 Integrin-cytoplasmic-domain-associated protein 1 IGFBP5 Insulin like growth factor binding protein 5 IP Immunoprecipitation IPA Ingenuity pathway analysis Klf2 Kruppel-like factor 2 KLF4 Kruppel-like factor 4 Krit1 Krev1 interaction trapped gene 1
LKB1 Liver kinase B1 MAPK Mitogen-activated protein kinase MARK Microtubule affinity regulating kinase MEKK3 Mitogen-activated protein kinase kinase kinase 3 MO25 Mouse protein 25 mRNA Messenger RNA MST Mammalian Ste20 kinase MST4 Mammalian Ste20-like protein kinase 4 mTOR Mammalian target of rapamycin NFκB1 Nuclear Factor of Kappa Light Polypeptide Gene Enhancer in B-Cells 1 NP40 Nonyl phenoxypolyethoxylethanol 40 PDCD10 Programmed cell death 10 PI3 kinase Phosphoinositide-3 kinase PJS Peutz-Jeghers syndrome POSTN Periostin, Osteoblast Specific Factor PP2A Protein phosphatase 2A qRT-PCR Quantitative reverse transcription polymerase chain reaction Rap1A Ras-related protein Rap-1A RIPA Radioimmunoprecipitation assay ROCK RhoA-activated kinase RPLP1 Ribosomal protein P1 RRAGD Ras-related GTP binding D RRM2B Ribonucleotide Reductase M2 B SAD Kinases Synapses of amphids defective kinases shRNA short-hairpin RNAs Stk25 Serine/Threonine Kinase 25 STRAD STE20 related adaptor STRIPAK Striatin-interacting phosphatase and kinase TCL Total cell lysate TGFβ Transforming growth factor beta 1 TP53 Tumor protein 53 TrkA Tropomyosin receptor kinase A VEGFR2 Vascular endothelial growth factor receptor 2 XEEK1 Xenopus egg and embryo kinase 1 YSK1 Yeast Sps1/Ste20kinase YWHAZ Tyrosine 3-Monooxygenase/Tryptophan 5- Monooxygenase Activation Protein, Zeta
x
Chapter 1: General Introduction
CHAPTER 1
GENERAL INTRODUCTION
12 Chapter 1: General Introduction
1. Introduction A mutation in cerebral cavernous malformation 3 (CCM3) causes the familial form of cerebral cavernous malformation (CCM), a vascular brain malformation that leads to severe headaches, hemorrhages and stroke. Through its interaction with several binding partners, CCM3 has been shown to play a role in modulating several processes such as angiogenesis, cell survival and most recently cellular polarity. By interacting with serine/threonine kinase 25 (Stk25) and Golgi matrix protein 130 (GM130), CCM3 regulates the orientation of the Golgi apparatus and polarized cell migration in epithelial cells [1, 2]. However, despite the diverse functions of
CCM3, little is known about the molecular mechanisms of CCM3 signaling.
The association of CCM3 with Stk25 and GM130 suggests it might play a role in the functions of Stk25, both in cellular polarity and Golgi maintenance [1, 2]. Recent findings have identified
Stk25 and GM130 as downstream effectors through which LKB1 regulates epithelial cell polarity and neuronal differentiation [3]. The serine/threonine protein kinase Liver Kinase B1
(LKB1) is an evolutionary conserved signaling protein that is implicated in several important biochemical pathways including cell metabolism, cell cycle control, p53-dependent apoptosis, and cell polarity (reviewed in [4]). Clear downstream targets include the SAD kinases and the
AMPK /mTOR-signaling pathway, which are implicated in LKB1-regulated neuronal polarization [5-7]. The mechanism by which LKB1 regulates these signaling pathways and its upstream activators are still being elucidated.
Understanding a role for CCM3 in cell polarity can expand our knowledge of CCM3 signaling, which might have an impact on elucidating CCM disease. In the body of work presented here, we attempt to address the question, what is the function of CCM3 in cell polarity? In Chapter 2, we identify a novel function for CCM3 in epithelial cell polarization through association with the
13 Chapter 1: General Introduction
LKB1 signaling pathway. We find that CCM3 binds to STRADα and regulates the protein expression of LKB1 and STRAD in epithelial cells. In addition, we show that CCM3 is necessary for LKB1-mediated epithelial polarization. In an effort to determine whether our findings in epithelial cells relate to CCM3 phenotype in endothelial cells, we took an initial step to identify changes in gene expression induced by CCM3 knockdown in endothelial cells. We discovered a robust downregulation of cell adhesion molecules and upregulation of several adhesion complex signaling. This work provides an important insight into a novel regulatory role for CCM3 in epithelial cell polarity and provides preliminary insights into downstream cell adhesion signaling pathways affected by the reduction of CCM3 in endothelial cells. i
2. Cerebral Cavernous Malformations (CCMs) 2a: CCM Disease
Cerebral cavernous malformations (CCMs) are discrete vascular malformations that are characterized by dilated leaky blood vessels. Histologically, CCMs are characterized by an endothelium that lacks smooth muscle cells, defective cell-cell junctions and thin vessels without a basal lamina [8]. These lesions, or cavernoma, increase the risk of strokes, focal hemorrhages and seizures. CCMs affect about 0.5% of the population and account for up to 13% of all cerebrovascular malformations. Currently, the only treatment for CCMs is surgical resection, as little is known about the formation of these clusters of caverns. Although significant progress has been made toward identifying CCM genes and resolving their crystal structures, our understanding of their functions remains incomplete. The relevant signaling pathways in which these proteins participate in have not been fully established in endothelial cells. In addition, the
14 Chapter 1: General Introduction significance of these pathways and the roles of CCM proteins in the vasculature remain elusive.
Future work is desperately needed to link these known functions to CCM formation.
CCM arises sporadically via unknown mechanisms or in an autosomal dominant manner in familial form. Lesions characterize both of these forms of CCMs however, the sporadic form tends to have single lesions while the inherited form typically has multiple lesions [9]. In recent years, significant progress has been made in understanding CCM pathogenesis with the identification of three separate genes associated with the familial form. The familial form of
CCM, which constitutes up to 50% of CCM cases [10], is caused by both germline and somatic mutations in one of three genes; CCM1, CCM2 or CCM3. Mutations in these genes give rise to
CCM cases that cannot be clinically distinguished. This is consistent with the discovery that
CCM1, CCM2 and CCM3 form a complex localized to the cytoplasm. Thus suggesting they function in a common or related pathway required for vascular integrity [11] [12]. CCM3 might have functions separate from this complex as mutations in CCM3 often result in a severe form of the disease [13].
2b. CCM Proteins and Signaling Pathways
Expression of CCM proteins is required for both development and maintenance of endothelial cells in the vasculature, which makes them critical in vasculogenesis and survival of endothelial cells [14]. CCM1 interacts directly with CCM2, which recruits CCM3 to the complex (Figure
1.1). In general, this trimeric complex of CCM proteins inhibits angiogenesis and attenuates Rho kinase signaling to maintain vascular permeability [15, 16]. Each of the CCM proteins is a multi- domain adaptor protein, which interacts with a range of signaling proteins and through which
15 Chapter 1: General Introduction
they play a role in a range of cellular processes from cell adhesion, migration to apoptosis
(Figure 1.1). However, there is still a lack of understanding on how loss-of-function of CCM proteins lead to CCM formation.
CCM1 (Krit1), identified as a protein interactor for Rap1A, co-localizes to microtubules and functions as a link between the cytoskeleton and extracellular matrix (ECM) [17]. In CCM1 mutations, there is significant vascular hyperpermeability because of the Krit1/Rap1 signaling pathway, which leads to dysfunctional tight junction assembly. In addition, loss of the CCM1 gene results in hyperactivation of RhoA and the RhoA-activated kinase (ROCK) that also results in compromised cell-cell junctions [18]. Krit1 also interacts with integrin-cytoplasmic-domain- associated protein 1 (ICAP1) to inhibit β1 integrin signaling [19, 20]. The CCM complex of proteins, through CCM1, represses β1 integrin’s role in angiogenesis and vascular development by overexpression of a transcription factor, klf2, which when upregulated promotes angiogenesis
[21].
CCM2 (malcavernin) functions as a scaffolding protein and it regulates the osmotic stress-
induced p38 MAPK signaling pathway [1]. It has no enzymatic activity and is ubiquitously
expressed in early embryogenesis with increased expression in endothelial cells in large vessels
[22, 23]. CCM2 is the hub of the CCM complex as it interacts directly with CCM1 and CCM3
simultaneously [16]. Loss of CCM2 causes overactive RhoA/ROCK signaling which is believed
to be mediated by CCM1, as CCM2’s ability to localize to cell-cell junctions is lost when CCM1
is absent at these junctions [18]. CCM2 also interacts with TrkA in nerve cells where this
interaction induces cell death [24]. It is hypothesized that CCM3 and Stk25, a member of the
16 Chapter 1: General Introduction
GCKIII family of serine/threonine kinases, mediate this function, however it is still unclear how
CCM2 induces cell death and whether the induction of cell death contributes to the CCM
phenotype seen in CCM2 mutations.
Of the known CCM genes, CCM3 or PDCD10 (programmed cell death 10) is the least
understood and was first identified in a screen for genes expressed during apoptosis induction in
a premyeloid cell line [25]. CCM3 is mapped to 3q26.1 and encodes a small protein composed of
212 amino acids that is up regulated in apoptotic cells and functions as an anti-apoptotic agent
[13, 26, 27]. It is highly conserved in vertebrates and has orthologues in invertebrates such as C. elegans and Drosophila. As well as being expressed in the vasculature, CCM3 protein is ubiquitously expressed in the developing nervous system including pyramidal neurons in the hippocampus, granule neurons in the dentate gyrus and Purkinje cells in the cerebellum [28].
Loss-of-function mutations in CCM3 cause defective vascular development in Drosophilia [29] and in mammals; CCM3 is required for cardiovascular development, cell migration and cell survival [10, 30]. Interestingly, these functions in vascular development, cell motility and survival are mediated through a separate mechanism involving the germinal center kinase III
(GCKIII) kinases [10, 31, 32]
Given its ability to act as an adaptor protein and play a key role in modulation of several signaling pathways, understanding the role of CCM3 would shed light on important mechanisms by which loss of CCM3 leads to CCM formation. Work presented in this thesis characterizes a novel function of CCM3 in epithelial cell polarity as part of the LKB1 signaling pathway
(Chapter 2). In an effort to determine whether this novel interaction contributes to the CCM3
17 Chapter 1: General Introduction phenotype seen in endothelial cells, we utilized microarray analysis to perform a global gene expression study comparing gene expression in normal endothelial cells to those with depleted
CCM3 (in Chapter 3).
3. CCM3: Structure, Functions and Signaling Pathways
CCM3 is the smallest of the CCM proteins and like the other two, is ubiquitously expressed in endothelial cells, vasculature in addition to the developing nervous system [28]. The crystal structure of CCM3 shows a α-helical protein with an N-terminal dimerization domain and C- terminal focal adhesion targeting-homology (FAT-H) domain [33] (Figure 1.2A). Through these domains, CCM3 interacts with a variety of proteins including CCM2 [1, 34], Stk25 [1, 2], MST4
[31, 35], Paxillin [33], FAP-1 [1] (Voss et al 2007), VEGFR2 [36] and Striatin [35, 37] (Figure
1.2B).
CCM3 also plays a role in the regulation of cell survival. The integrity of endothelial cells is poor in CCM and the proper regulation of cell death is critical in maintaining endothelial integrity. Evidence for CCM3’s role in pro-survival functions includes the regulation of
VEGFR2 stability [36] and promoting cell survival at the cell periphery after oxidative stress by activating the ERM proteins [38]. Conversely, CCM expression has also been linked to cell death in several studies as loss of CCM3 increases survival and proliferation [13, 39]. The specific role of CCM3 in cell survival is unclear. Moving forward it is important to reconcile these conflicting observations and understand how regulation of cell survival contributes to the maintenance of endothelial vascular integrity.
18 Chapter 1: General Introduction
Interestingly, Stk25 has been shown to be downstream of the trimeric CCM complex and
necessary for cardiovascular development [10]. These findings suggest that CCM3 and Stk25
may interact to regulate angiogenesis and vascular integrity. CCM3 associates with the germinal
center kinases (GCK) III family (MST4, Stk24 and Stk25) kinases both in yeast two-hybrid assays and upon over-expression in mammalian cells [1, 26]. It is believed that the GCKIII family acts as a novel class of downstream CCM signaling effectors required in cardiovascular development in vivo [10]. CCM3 is also implicated in the mitogen-activated protein kinase
(MAPK) pathway in vitro, in part because of its association with Stk25 [36].
Outside of the CCM complex, CCM3 is predominantly found as a component of a large protein
complex, STRIPAK (striatin-interacting phosphatase and kinase), which assembles phosphatases
and kinases arranged around a striatin core [35]. PP2A (protein phosphatase 2A), found in
STRIPAK, is a multifunctional serine/threonine phosphatase, and like most protein
phosphatases, it is targeted to specific subcellular locations and substrates via interactions with regulatory proteins. Through its associations, PP2A is linked to many aspects of cell function
including control of cell growth, proliferation and differentiation [35, 37]. CCM3 interacts with
PP2A regulatory subunit B, and recruits members of the GCKIII kinases (MST4, Stk24 and
Stk25) to this complex [37], acting as a bridge between the phosphatase and kinase component of
STRIPAK. Interestingly, PP2A is also the phosphatase responsible for dephosphorylating
GM130 at Serine25 [40]. In addition, in an attempt to elucidate the biological function of
STRIPAK complex, Kean et al. (2010) determined that MST4 localization and interaction with
GM130 were regulated by CCM3 and Striatins [37] (Figure 1.1).
19 Chapter 1: General Introduction
Taken together, these studies suggest that CCM3 acts as a scaffolding protein and through its association with the STRIPAK complex, it may be involved in regulating the localization of
Stk25 and PP2A to the Golgi apparatus. At the Golgi, CCM3 regulates Stk25 functions in Golgi morphology and positioning and cell polarization [2].
4. The LKB1 Signaling Pathway The ability of cells to polarize is an essential biological function that is required for several developmental processes including cell migration, cell division and organogenesis. Cell polarity is highly regulated by a family of conserved proteins and signaling pathways. In particular, the coordinated action of three main protein complexes and their association with cell-cell adhesion receptors and Rho family GTPases mediate epithelial cell polarity and the acquisition of apicobasal polarity. These so called ‘polarity complexes’ include the Crumbs, partitioning defective (Par) and Scribble complexes which modulate downstream signaling of GTPases to impact cytoskeletal changes leading to cellular asymmetry. Of interest to us is the Par complex, which is comprised of Par3, Par6 and atypical protein kinase C (aPKC) [41, 42]. One of the downstream effectors of the Par complex is LKB1, a serine/threonine master kinase that plays a role in several biological functions including the regulation of epithelial cell and neuronal polarity.
LKB1 was first identified as the main gene that when mutated is responsible for the familial form of Peutz-Jeghers cancer syndrome, (PJS) [43]. LKB1, maps onto chromosome 19p13.3, encodes a 433-amino-acid protein that shows strong sequence similarity to the serine/threonine protein
20 Chapter 1: General Introduction
kinases [44]. The deletion of both alleles of LKB1 is embryonically lethal between embryonic
days 8 and 11 (E8-E11), however the heterozygous mice develop a PJS-like disease [45]. Several
somatic studies with alterations of LKB1 support the description of LKB1 as a tumor suppressor
protein and about ~ 80% of mutations in PJS, patients are predicted to disrupt the kinase activity
[46]. In addition, the ability of the overexpression of LKB1 to suppress growth by G1 cell cycle
arrest in cancer cells that lack LKB1, suggests that LKB1 functions as a tumor suppressor [47,
48].
LKB1 encodes a highly conserved protein and its functions beyond tumor suppression have been
described in Caenorhabditis elegans (C. elegans), Xenopus and Drosophila. Drosophila LKB1,
with 44% similarity to the human homolog kinase domain, is required for the early anterior-
posterior polarity of the oocyte and for repolarization of the cytoskeleton, which defines the
embryonic axis [49]. In C. elegans, mutations in the par genes disrupt several asymmetries established during early cell cycle of embryogenesis, and the products (of which six have been identified) are required for establishing polarity [50]. The LKB1 ortholog par-4 regulates both cell growth/proliferation and cell polarity and is essential for appropriate partitioning of developmental determinants during the first cell division in the zygote [51]. Par-4 encodes a putative serine/threonine kinase with similarity to the human LKB1 and Xenopus egg and embryo kinase (XEEK1) [50].
Earlier investigations into the role of LKB1 in mammalian cells identified two cellular substrates
that interact with LKB1. In a seminal publication, Bass & colleagues (2003) identified and
characterized an LKB1-specific adaptor protein and substrate, STRAD (STe20 Related ADaptor)
21 Chapter 1: General Introduction
as an upstream activator of LKB1. Both isoforms of STRAD (α and β) bind to LKB1, activate its
catalytic activity and induce an autophosphorylation of LKB1 at Thr336 and Thr363 and
phosphorylation of STRAD at Thr329 and Thr419 [15][46]. Binding of STRAD to LKB1 re-
localizes LKB1 from the nucleus to the cytoplasm and enhances LKB1 in vivo activity [46, 52],
emphasizing a key role for STRAD in LKB1 function regulation. The LKB1-STRAD complex
also associates with a highly evolutionarily conserved protein, Mouse protein-25 (MO25),
through the interaction of the last three residues of STRAD [45]. MO25 enhances the formation
of the LKB1-STRAD complex in vivo by regulating the conformational active states of STRAD
and LKB1, which is stabilized by MO25 interacting with the LKB1 activation loop [45, 53, 54].
Together, STRAD and MO25 anchor LKB1 in the cytoplasm and play a crucial role in regulating
LKB1 activity and cellular localization (Figure 1.3).
The identification of the active LKB1 complex was followed by a search of its downstream
effectors and signaling pathways. Three separate groups identified LKB1 as the upstream kinase
of 5’-AMP-activated kinase (AMPK) [55-57]. LKB1 phosphorylates AMPK in vitro at Thr172
[55, 58] and the ability of LKB1 to activate AMPK increases 100 fold in the presence of STRAD
and MO25 in complex [55]. Additionally, in HeLa cells, which lack LKB1 expression, AMPK
could not be activated by agonists [56]. Taken together, these studies place the LKB1-STRAD-
MO25 complex as upstream kinases that activate AMPK and regulate AMPK cellular functions in cell metabolism and cell polarity (Figure 1.3).
There are 12 human kinases that are related to AMPK, and LKB1 can phosphorylate the T-loop of all the members of this subfamily [59], including SAD kinase and the MARK family of
22 Chapter 1: General Introduction
kinases. These findings suggest that LKB1 functions as a master kinase, to activate many
members of the AMPK-related protein kinase. It has been demonstrated that increasing the
activation of MARK leads to microtubule breakdown and cell death in addition to its functions in
cell polarity development and neurite outgrowth [60] [61].
The mammalian homolog of LKB1 in C. elegans (Par4) and Drosophila (Lkb1) are required for
establishing cell polarity and anterior-posterior axis formation respectively [49, 50]. Yet, a role
for mammalian LKB1 in the establishment of polarity in cells was not established until the
engineering of a cell line that provided an experimental model in which polarity is rapidly and
actively induced. W4 intestinal epithelial cells were constructed to stably express LKB1, which
is activated upon inducible STRAD expression in response to doxycycline [62]. Upon activation
of LKB1, the cells polarize to form an apical brush border by reorganizing their actin
cytoskeleton [62]. Further, in the developing cortex, LKB1 is critical for axon initiation and
differentiation [7] [63] and loss of LKB1 in vivo impairs neuronal migration and centrosome
positioning that is linked with neuronal polarity [64, 65]. Together, these results describe LKB1
as a prominent regulator of cell polarity.
In summary, LKB1 forms a heterotrimeric complex with the pseudokinase STRAD and
scaffolding protein MO25 requiring both for proper localization and for kinase activity (Figure
3). Clear downstream targets include the SAD kinases and the AMPK /mTOR-signaling pathway, which are implicated in LKB1-regulated neuronal polarization by interfering with PI 3- kinase localization [6]. However how LKB1 regulates these signaling pathways and its upstream
23 Chapter 1: General Introduction activators are still being elucidated. Recent findings have identified Stk25 and GM130 as possible downstream effectors through which LKB1 regulates neuronal polarization [3].
4a. STRAD regulates LKB1 stability
As the name implies, the gene STRAD encodes a protein that belongs to the STE20-like kinase family, which are upstream kinases of the mitogen-activated protein kinase (MAPK) pathway.
STRAD lacks several indispensable residues in its kinase domain characterizing it as a pseudokinase, despite its ability to bind to ATP and display a closed conformation and ordered activation loop typical of active protein kinases [46, 54]. The two isoforms of STRAD, STRADα and STRADβ, have both been shown to allosterically regulate the catalytic activity of LKB1 and localization of LKB1[45, 46].
STRAD plays an important role in brain development as homozygous mutations in STRADα result in polyhydramnios, megalocephaly, and symptomatic epilepsy (PMSE) syndrome [66].
PMSE is a rare autosomal-recessive disorder that is characterized by abnormal brain development, cognitive disability and intractable infantile-onset. Both isoforms of STRAD are sufficient for axonogenesis and promoting cell survival in the developing cortex [52]. However, only STRADα is responsible for LKB1 protein stability [52]. Given the critical role in regulating
LKB1, a key regulator of cell polarity, cell metabolism and tumorigenesis, understanding the regulatory mechanisms of STRAD is important.
24 Chapter 1: General Introduction
5.The Role of Stk25 and the Golgi apparatus in Cell Polarity 5a. The Golgi apparatus regulates cell polarity
Many polarity proteins exert their effects by acting at sites of active growth (e.g. leading edge of a migrating cell or axonal growth cone) where they regulate the dynamics of the cytoskeleton
[67]. Changes in the cytoskeletal organization needs to be coordinated with the addition of new membrane at sites of active growth. Most of the membranes needed to elongate the growing dendrite are derived from the Golgi secretory pathway and Golgi outposts [68].
The Golgi apparatus in mammalian cells undergoes morphological changes during the cell cycle.
In particular, during mitosis the Golgi is fragmented [69]. Mitotic Golgi fragmentation studies led to the discovery of Golgi matrix protein GM130, a tightly associated Golgi membrane protein and a component of the Golgi matrix that maintains the structure of the Golgi apparatus
[70]. Phosphorylation of GM130 at serine 25 by Cdc2 inhibits the binding of p115, a vesicular tethering molecule and is responsible for mitotic Golgi fragmentation [69]. In addition to being a popular Golgi marker, GM130 has critical functions in the Golgi, ranging from vesicular transport pathway to emerging roles in the regulation of cell cycle, cell migration and cell polarization (ref).
The position of the Golgi apparatus has also been implicated in different aspects of cell polarity.
Disruption of the positioning of the Golgi in epithelial cells led to defects in polarization and inhibited directional migration in wound assays [71] suggesting that Golgi positioning plays a primary role in cell polarity and cell migration. The position of the Golgi and the attached centrosome may predict the site of axonogenesis in cultured hippocampal neurons. Early in neuronal polarization, the centrosomes and Golgi apparatus cluster at one pole of
25 Chapter 1: General Introduction undifferentiated neurons and this clustering mark the site of axon formation [65, 72]. Later, the
Golgi is polarized towards the longest and more complex dendrite where it is required for dendritic outgrowth [73] and dendritic arbor formation [68]. Dispersion of the Golgi apparatus away from the apical pole leads to a loss of dendrite asymmetry [73]. Despite the importance of the Golgi’s localization in polarity, the regulation of Golgi positioning within epithelial cells or neurons remains undetermined. Recent findings elucidating the association of GM130 with the serine/threonine kinase 25 (Stk25) provides some insight into the importance of the Golgi apparatus in cell polarization.
5b. Stk25, a regulator of cellular polarity
In an effort to clarify the upstream regulatory mechanism of MAPK, Osada and colleagues
(1997, [74]) isolated cDNA clones encoding a novel protein kinase, yeast Sps1/Ste20kinase
(YSK1) or Stk25 that is structurally and evolutionarily related to Ste20-related kinases and does not participate in any of the known MAPK pathways. The Ste20 family of serine/threonine protein kinases is implicated in several signaling pathways including cell migration and polarity
[75]. There are over thirty Ste20 kinases in mammals, which are divided into two groups: p21- activated kinases and germinal center kinases (GCKs). Stk25 belongs to the latter and a further subgroup classification places it in subgroup III with the mammalian Ste20 (MST) kinases [75].
These proteins are involved in regulation of cell proliferation, cell death and cytoskeleton rearrangements in response to cellular stress [76].
Stk25 is broadly localized in the cytosol and Golgi apparatus in cells where it associates with
GM130 [52]. Preisinger and colleagues (2004) showed that depletion of Stk25 caused Golgi
26 Chapter 1: General Introduction dispersion to the cell periphery, interrupted Golgi location and cell migration which suggest that
Stk25 signaling through the Golgi is required for normal Golgi polarization and cell migration
[77]. As the position of the Golgi is important for cell polarization, our laboratory recently characterized the role of Stk25 in nervous system development. Stk25 knockdown in neurons fragmented the Golgi apparatus, inhibited axon initiation, and promoted dendrite asymmetry.
Interestingly, GM130 depletion in cultured neurons lead to reduced axon formation, similar to that observed for Stk25 suggesting that GM130 and the Golgi apparatus are required for axon initiation [38]. In addition, failure of Stk25 to rescue the phenotype of reduced axon numbers of
GM130 depletion places GM130 downstream of Stk25 in the regulation of axon initiation. These studies strongly suggest that Stk25 and GM130 are required for normal Golgi morphology in neurons.
Knockdown of Stk25 expression levels prevented axon formation while overexpression of Stk25 in neurons caused multiple axons to form [3], resembling the LKB1 phenotypes in neurons [5,
7]. Matsuki and colleagues (2010) further demonstrated that Stk25 associates with STRAD and
Stk25 was required for LKB1-STRAD regulated intestinal cell polarization and neuronal polarity
[3]. This initial finding places Stk25 as a regulator of cellular polarization and begins to unravel the signaling pathway downstream of LKB1.
A novel binding partner of Stk25, cerebral cavernous malformation 3 (CCM3), was recently shown to stabilize Stk25 and regulate Golgi orientation [2]. These functions of CCM3 on modulating Stk25 activity suggested to us that CCM3 could be the molecule that mediates Stk25
27 Chapter 1: General Introduction
and STRAD association, which would link CCM3 to the LKB1 signaling pathway and establish
a potential role for CCM3 in cell polarity.
6. Overview of Chapters 2 and 3
As previously stated, LKB1 signaling cascade has a conserved role in the regulation of cell
polarization, including neuronal polarity [5, 7]. In addition, LKB1 kinase activity acts as a tumor
suppressor and functions as a therapeutic target for several cancers. On the other hand, CCM3 is
one of three genes mutated in CCMs, a vascular disease where CCM3 functions as a key
regulator of VEGF signaling in cell survival and vascular development [2]. Our studies are
designed to characterize the molecular role of CCM3 in cellular polarization; in particular,
clarify the function of the association with Stk25, which regulates LKB1-mediated polarity, and determine if CCM3 is part of the LKB1 signaling pathway of polarity.
In Chapter 2, we investigate whether CCM3 plays a role in cellular polarity in an LKB1- dependent manner. Recent findings place CCM3 at the Golgi complex where it interacts with the polarity kinase Stk25 and regulates the orientation of the Golgi and cell migration. Since Stk25 is a binding partner of CCM3, we hypothesized that CCM3 may play an important role in the
LKB1 signaling pathway, a known function of Stk25. Protein expression studies in combination with immunoprecipitation assays reveal that CCM3 interacts with STRADα, a key regulator of
LKB1 activity and protein stability. In addition, we show that CCM3 is necessary for LKB1- mediated epithelial cell polarity. The new evidence we present in these studies shows that CCM3
28 Chapter 1: General Introduction is a regulator of the LKB1 pathway, possibly through STRADα, which opens new avenues for
LKB1 pathway regulation.
In Chapter 3, we take the initial step to examine a possible role for LKB1 and STRADα in the
CCM3 phenotype in endothelial cells. We performed a global gene expression survey to compare gene expression in normal endothelial cells and CCM3 knockdown cells. Despite the lack of evidence of a contribution of the LKB1/STRAD signaling pathway in CCM phenotype, we uncover several other dysregulated signaling pathways in endothelial cells when CCM3 is absent, including a downregulation of cell adhesion molecules and an activation of p53 signaling pathway.
Together, these studies describe novel interactions of CCM3 and the signaling pathways it regulates. The work presented here furthers our understanding of the cellular and molecular mechanisms of CCM3 and sheds light on several cellular functions that are altered in the CCM phenotype.
29 Chapter 1: General Introduction
REFERENCES 1. Voss, K., et al., CCM3 interacts with CCM2 indicating common pathogenesis for
cerebral cavernous malformations. Neurogenetics, 2007. 8(4): p. 249-56.
2. Fidalgo, M., et al., CCM3/PDCD10 stabilizes GCKIII proteins to promote Golgi
assembly and cell orientation. J Cell Sci, 2010. 123(Pt 8): p. 1274-84.
3. Matsuki, T., et al., Reelin and stk25 have opposing roles in neuronal polarization and
dendritic Golgi deployment. Cell, 2010. 143(5): p. 826-36.
4. Alessi, D.R., K. Sakamoto, and J.R. Bayascas, LKB1-dependent signaling pathways.
Annu Rev Biochem, 2006. 75: p. 137-63.
5. Shelly, M., et al., LKB1/STRAD promotes axon initiation during neuronal polarization.
Cell, 2007. 129(3): p. 565-77.
6. Amato, S., et al., AMP-activated protein kinase regulates neuronal polarization by
interfering with PI 3-kinase localization. Science, 2011. 332(6026): p. 247-51.
7. Barnes, A.P., et al., LKB1 and SAD kinases define a pathway required for the
polarization of cortical neurons. Cell, 2007. 129(3): p. 549-63.
8. Dibble, C.F., et al., Defining the functional domain of programmed cell death 10 through
its interactions with phosphatidylinositol-3,4,5-trisphosphate. PLoS One, 2010. 5(7): p. e11740.
9. Awad, I.A., Cavernous malformations. J Neurosurg, 2012. 116(1): p. 119-20; discussion
121.
10. Zheng, X., et al., CCM3 signaling through sterile 20-like kinases plays an essential role during zebrafish cardiovascular development and cerebral cavernous malformations. J Clin
Invest, 2010. 120(8): p. 2795-804.
30 Chapter 1: General Introduction
11. Stahl, S., et al., Novel CCM1, CCM2, and CCM3 mutations in patients with cerebral cavernous malformations: in-frame deletion in CCM2 prevents formation of a
CCM1/CCM2/CCM3 protein complex. Hum Mutat, 2008. 29(5): p. 709-17.
12. Cavalcanti, D.D., et al., Cerebral cavernous malformations: from genes to proteins to disease. J Neurosurg, 2012. 116(1): p. 122-32.
13. Zhu, Y., et al., Differential angiogenesis function of CCM2 and CCM3 in cerebral cavernous malformations. Neurosurg Focus, 2010. 29(3): p. E1.
14. Li, X., et al., Molecular recognition of leucine-aspartate repeat (LD) motifs by the focal adhesion targeting homology domain of cerebral cavernous malformation 3 (CCM3). J Biol
Chem, 2011. 286(29): p. 26138-47.
15. Faurobert, E. and C. Albiges-Rizo, Recent insights into cerebral cavernous malformations: a complex jigsaw puzzle under construction. FEBS J, 2010. 277(5): p. 1084-96.
16. Fisher, O.S., et al., Structural studies of cerebral cavernous malformations 2 (CCM2) reveal a folded helical domain at its C-terminus. FEBS Lett, 2013. 587(3): p. 272-7.
17. Serebriiskii, I., et al., Association of Krev-1/rap1a with Krit1, a novel ankyrin repeat- containing protein encoded by a gene mapping to 7q21-22. Oncogene, 1997. 15(9): p. 1043-9.
18. Stockton, R.A., et al., Cerebral cavernous malformations proteins inhibit Rho kinase to stabilize vascular integrity. J Exp Med, 2010. 207(4): p. 881-96.
19. Zhang, J., et al., Interaction between krit1 and icap1alpha infers perturbation of integrin beta1-mediated angiogenesis in the pathogenesis of cerebral cavernous malformation. Hum Mol
Genet, 2001. 10(25): p. 2953-60.
31 Chapter 1: General Introduction
20. Zawistowski, J.S., et al., KRIT1 association with the integrin-binding protein ICAP-1: a new direction in the elucidation of cerebral cavernous malformations (CCM1) pathogenesis.
Hum Mol Genet, 2002. 11(4): p. 389-96.
21. Renz, M., et al., Regulation of beta1 integrin-Klf2-mediated angiogenesis by CCM proteins. Dev Cell, 2015. 32(2): p. 181-90.
22. Fisher, O.S., et al., Structural basis for the disruption of the cerebral cavernous malformations 2 (CCM2) interaction with Krev interaction trapped 1 (KRIT1) by disease- associated mutations. J Biol Chem, 2015. 290(5): p. 2842-53.
23. Draheim, K.M., et al., CCM2-CCM3 interaction stabilizes their protein expression and permits endothelial network formation. J Cell Biol, 2015. 208(7): p. 987-1001.
24. Costa, B., et al., STK25 protein mediates TrkA and CCM2 protein-dependent death in pediatric tumor cells of neural origin. J Biol Chem, 2012. 287(35): p. 29285-9.
25. Bergametti, F., et al., Mutations within the programmed cell death 10 gene cause cerebral cavernous malformations. Am J Hum Genet, 2005. 76(1): p. 42-51.
26. Ma, X., et al., PDCD10 interacts with Ste20-related kinase MST4 to promote cell growth and transformation via modulation of the ERK pathway. Mol Biol Cell, 2007. 18(6): p. 1965-78.
27. Ji, B.H., et al., A novel deletion mutation in CCM1 gene (krit1) is detected in a Chinese family with cerebral cavernous malformations. Yi Chuan Xue Bao, 2006. 33(2): p. 105-10.
28. Petit, N., et al., Patterns of expression of the three cerebral cavernous malformation
(CCM) genes during embryonic and postnatal brain development. Gene Expr Patterns, 2006.
6(5): p. 495-503.
29. Song, Y., M. Eng, and A.S. Ghabrial, Focal defects in single-celled tubes mutant for
Cerebral cavernous malformation 3, GCKIII, or NSF2. Dev Cell, 2013. 25(5): p. 507-19.
32 Chapter 1: General Introduction
30. Louvi, A., S. Nishimura, and M. Gunel, Ccm3, a gene associated with cerebral cavernous malformations, is required for neuronal migration. Development, 2014. 141(6): p.
1404-15.
31. Zhang, M., et al., Structural mechanism of CCM3 heterodimerization with GCKIII kinases. Structure, 2013. 21(4): p. 680-8.
32. Yoruk, B., et al., Ccm3 functions in a manner distinct from Ccm1 and Ccm2 in a zebrafish model of CCM vascular disease. Dev Biol, 2012. 362(2): p. 121-31.
33. Li, X., et al., Crystal structure of CCM3, a cerebral cavernous malformation protein critical for vascular integrity. J Biol Chem, 2010. 285(31): p. 24099-107.
34. Hilder, T.L., et al., Proteomic identification of the cerebral cavernous malformation signaling complex. J Proteome Res, 2007. 6(11): p. 4343-55.
35. Goudreault, M., et al., A PP2A phosphatase high density interaction network identifies a novel striatin-interacting phosphatase and kinase complex linked to the cerebral cavernous malformation 3 (CCM3) protein. Mol Cell Proteomics, 2009. 8(1): p. 157-71.
36. He, Y., et al., Stabilization of VEGFR2 signaling by cerebral cavernous malformation 3 is critical for vascular development. Sci Signal, 2010. 3(116): p. ra26.
37. Kean, M.J., et al., Structure-function analysis of core STRIPAK Proteins: a signaling complex implicated in Golgi polarization. J Biol Chem, 2011. 286(28): p. 25065-75.
38. Fidalgo, M., et al., Adaptor protein cerebral cavernous malformation 3 (CCM3) mediates phosphorylation of the cytoskeletal proteins ezrin/radixin/moesin by mammalian Ste20-4 to protect cells from oxidative stress. J Biol Chem, 2012. 287(14): p. 11556-65.
39. Chen, L., et al., Apoptotic functions of PDCD10/CCM3, the gene mutated in cerebral cavernous malformation 3. Stroke, 2009. 40(4): p. 1474-81.
33 Chapter 1: General Introduction
40. Lowe, M., N.K. Gonatas, and G. Warren, The mitotic phosphorylation cycle of the cis-
Golgi matrix protein GM130. J Cell Biol, 2000. 149(2): p. 341-56.
41. Kemphues, K.J., et al., Identification of genes required for cytoplasmic localization in early C. elegans embryos. Cell, 1988. 52(3): p. 311-20.
42. Tepass, U., C. Theres, and E. Knust, crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia. Cell,
1990. 61(5): p. 787-99.
43. Hemminki, A., et al., A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature, 1998. 391(6663): p. 184-7.
44. Boudeau, J., G. Sapkota, and D.R. Alessi, LKB1, a protein kinase regulating cell proliferation and polarity. FEBS Lett, 2003. 546(1): p. 159-65.
45. Boudeau, J., et al., MO25alpha/beta interact with STRADalpha/beta enhancing their ability to bind, activate and localize LKB1 in the cytoplasm. EMBO J, 2003. 22(19): p. 5102-14.
46. Baas, A.F., et al., Activation of the tumour suppressor kinase LKB1 by the STE20-like pseudokinase STRAD. EMBO J, 2003. 22(12): p. 3062-72.
47. Ylikorkala, A., et al., Vascular abnormalities and deregulation of VEGF in Lkb1- deficient mice. Science, 2001. 293(5533): p. 1323-6.
48. Tiainen, M., A. Ylikorkala, and T.P. Makela, Growth suppression by Lkb1 is mediated by a G(1) cell cycle arrest. Proc Natl Acad Sci U S A, 1999. 96(16): p. 9248-51.
49. Martin, S.G. and D. St Johnston, A role for Drosophila LKB1 in anterior-posterior axis formation and epithelial polarity. Nature, 2003. 421(6921): p. 379-84.
50. Watts, J.L., et al., The C. elegans par-4 gene encodes a putative serine-threonine kinase required for establishing embryonic asymmetry. Development, 2000. 127(7): p. 1467-75.
34 Chapter 1: General Introduction
51. Narbonne, P., et al., Differential requirements for STRAD in LKB1-dependent functions
in C. elegans. Development, 2010. 137(4): p. 661-70.
52. Veleva-Rotse, B.O., et al., STRAD pseudokinases regulate axogenesis and LKB1 stability. Neural Dev, 2014. 9: p. 5.
53. Zeqiraj, E., et al., Structure of the LKB1-STRAD-MO25 complex reveals an allosteric mechanism of kinase activation. Science, 2009. 326(5960): p. 1707-11.
54. Zeqiraj, E., et al., ATP and MO25alpha regulate the conformational state of the
STRADalpha pseudokinase and activation of the LKB1 tumour suppressor. PLoS Biol, 2009.
7(6): p. e1000126.
55. Hawley, S.A., et al., Complexes between the LKB1 tumor suppressor, STRAD alpha/beta and MO25 alpha/beta are upstream kinases in the AMP-activated protein kinase cascade. J Biol,
2003. 2(4): p. 28.
56. Woods, A., et al., LKB1 is the upstream kinase in the AMP-activated protein kinase cascade. Curr Biol, 2003. 13(22): p. 2004-8.
57. Shaw, R.J., et al., The LKB1 tumor suppressor negatively regulates mTOR signaling.
Cancer Cell, 2004. 6(1): p. 91-9.
58. Shaw, R.J., et al., 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, 2004.
101(10): p. 3329-35.
59. Lizcano, J.M., et al., LKB1 is a master kinase that activates 13 kinases of the AMPK subfamily, including MARK/PAR-1. EMBO J, 2004. 23(4): p. 833-43.
60. Kemphues, K., PARsing embryonic polarity. Cell, 2000. 101(4): p. 345-8.
35 Chapter 1: General Introduction
61. Biernat, J., et al., Protein kinase MARK/PAR-1 is required for neurite outgrowth and
establishment of neuronal polarity. Mol Biol Cell, 2002. 13(11): p. 4013-28.
62. Baas, A.F., et al., Complete polarization of single intestinal epithelial cells upon
activation of LKB1 by STRAD. Cell, 2004. 116(3): p. 457-66.
63. Shelly, M. and M.M. Poo, Role of LKB1-SAD/MARK pathway in neuronal polarization.
Dev Neurobiol, 2011. 71(6): p. 508-27.
64. Asada, N., K. Sanada, and Y. Fukada, LKB1 regulates neuronal migration and neuronal
differentiation in the developing neocortex through centrosomal positioning. J Neurosci, 2007.
27(43): p. 11769-75.
65. de Anda, F.C., et al., Centrosome motility is essential for initial axon formation in the
neocortex. J Neurosci, 2010. 30(31): p. 10391-406.
66. Puffenberger, E.G., et al., Polyhydramnios, megalencephaly and symptomatic epilepsy caused by a homozygous 7-kilobase deletion in LYK5. Brain, 2007. 130(Pt 7): p. 1929-41.
67. Wiggin, G.R., J.P. Fawcett, and T. Pawson, Polarity proteins in axon specification and synaptogenesis. Dev Cell, 2005. 8(6): p. 803-16.
68. Ye, B., et al., Growing dendrites and axons differ in their reliance on the secretory pathway. Cell, 2007. 130(4): p. 717-29.
69. Lowe, M., et al., Cdc2 kinase directly phosphorylates the cis-Golgi matrix protein
GM130 and is required for Golgi fragmentation in mitosis. Cell, 1998. 94(6): p. 783-93.
70. Nakamura, N., et al., The vesicle docking protein p115 binds GM130, a cis-Golgi matrix protein, in a mitotically regulated manner. Cell, 1997. 89(3): p. 445-55.
71. Yadav, S., S. Puri, and A.D. Linstedt, A primary role for Golgi positioning in directed secretion, cell polarity, and wound healing. Mol Biol Cell, 2009. 20(6): p. 1728-36.
36 Chapter 1: General Introduction
72. de Anda, F.C., et al., Centrosome localization determines neuronal polarity. Nature,
2005. 436(7051): p. 704-8.
73. Horton, A.C., et al., Polarized secretory trafficking directs cargo for asymmetric dendrite growth and morphogenesis. Neuron, 2005. 48(5): p. 757-71.
74. Osada, S., et al., YSK1, a novel mammalian protein kinase structurally related to Ste20 and SPS1, but is not involved in the known MAPK pathways. Oncogene, 1997. 14(17): p. 2047-
57.
75. Dan, I., N.M. Watanabe, and A. Kusumi, The Ste20 group kinases as regulators of MAP kinase cascades. Trends Cell Biol, 2001. 11(5): p. 220-30.
76. Nogueira, E., et al., SOK1 translocates from the Golgi to the nucleus upon chemical anoxia and induces apoptotic cell death. J Biol Chem, 2008. 283(23): p. 16248-58.
77. Preisinger, C., et al., YSK1 is activated by the Golgi matrix protein GM130 and plays a role in cell migration through its substrate 14-3-3zeta. J Cell Biol, 2004. 164(7): p. 1009-20.
37 Chapter 1: General Introduction
FIGURES
38 Chapter 1: General Introduction
Figure 1.1. Schematic representation of CCM3 complexes, signaling pathways and cellular functions.
A. CCM3 associates with CCM2 in the CCM Protein Complex to maintain the integrity of the endothelial vasculature through RhoA signaling.
B. Within the STRIPAK complex, CCM3 bridges the interaction between Striatin and the
GCKIII kinases to regulate the orientation of the Golgi apparatus and polarity of migrating cells.
C. In this work, we show that CCM3 interacts with STRAD in the LKB1 complex and that
CCM3 is necessary for epithelial cell polarization.
39 Chapter 1: General Introduction
40 Chapter 1: General Introduction
Figure 1.2 CCM3 protein structure and binding partners.
A. The crystal structure of CCM3 is an all α-helical protein with two domains, an N-terminal dimerization domain and a C-terminal focal adhesion targeting (FAT)-homology domain.
Image is reproduced here with permission from journal. Originally published in The Journal of
Biological Chemistry.. Xiaofeng Li,, Rong Zhang, Haifeng Zhang, Yun He, Weidong Ji, Wang
Min and Titus J. Boggon. Crystal Structure of CCM3, a Cerebral Cavernous Malformation
Protein Critical for Vascular Integrity. JBC 285, 24099-24107. © The American Society for
Biochemistry and Molecular Biology.
B. A schematic representation of the two functional domains of CCM3 and their main interactors. The N-terminal dimerization domain interacts with CCM3, the GCKIII proteins
(STK25 and Mst4). The FAT-homology domain interacts with several proteins including CCM2,
Straitins, Paxillin and STRADα* (data presented in this Thesis).
41 Chapter 1: General Introduction
42 Chapter 1: General Introduction
Figure 1.3. Schematic representation of the LKB1 Signaling Pathway
The heterotrimeric complex of LKB1 with STRAD and MO25 is the active form.
The fully active LKB1 complex activates several downstream targets involved in the regulation
of cell polarity, cellular metabolism and survival. Stk25 associates with STRAD and regulates
LKB1-mediated cell polarity. Stk25 has roles in cell survival and metabolism however, it remains undetermined if these are dependent on LKB1.
43 Chapter 1: General Introduction
44 Chapter 2: CCM3 regulates epithelial cell polarity
CHAPTER 2
CCM3 BINDS STRADα AND REGULATES LKB1-STRAD DEPENDENT
EPITHELIAL CELL POLARIZATION
This chapter is an original copyright and has not been previously published: Mansaray-
Storms, ZY; Howell, BH (2016)
45 Chapter 2: CCM3 regulates epithelial cell polarity
ABSTRACT The small adaptor protein cerebral cavernous malformation 3 (CCM3) is one of three
genes which when mutated is responsible for the familial form of cerebral cavernous
malformations (CCMs) in the central nervous system. It coordinates the protein levels
and activity of multiple signaling proteins including Stk25, which is required for cellular
polarity. CCM3 also plays a role in several cellular functions including Golgi
morphology, cell survival, angiogenesis and cell migration. However, despite the diverse
functions of CCM3, little is known about the functional and signaling importance of
CCM3 in cell polarity and its contribution to CCMs. Here we demonstrate that CCM3
regulates the expression of proteins in the conserved protein liver kinase B1 (LKB1)
pathway through an association with STRADα, the pseudokinase that regulates the
stability of LKB1. We identify the FAT-homology domain of CCM3 as the site of this
interaction. Our data suggests CCM3 is required for LKB1-STRADα-dependent epithelial cell polarization. CCM3 overexpression prevents the proteolytic degradation of
STRADα and augments the protein expression of LKB1, likely through STRADα.
Furthermore, we observe that CCM3 knockdown affects the expression of the downstream LKB1 effector kinases MARK1 and AMPK required for LKB1-mediated epithelial cell polarization. These findings identify CCM3 as a novel interaction partner of STRADα and modulator of the LKB1-STRAD cell-signaling cascade.
46 Chapter 2: CCM3 regulates epithelial cell polarity
INTRODUCTION Human genetics studies have implicated CCM3 as playing a particularly important role in
the formation of the cranial vasculature. Autosomal dominant loss of CCM3 function
leads to cerebral cavernous malformations, a condition observed in up to 0.5% of the
human population [1-4]. CCM3 along with two additional genes implicated in these
vascular malformations (CCM1 and CCM2) regulate the apical endothelial cell-cell
junctions in part by inhibiting RhoA [5-7]. CCM3 has been shown to form complexes
with a growing list of proteins, including CCM1 and CCM2, Paxillin, Striatin, GCKIII
kinases (like Stk25) and VEGFR2, which CCM3 has been shown to protect from
proteolytic degradation [8-13]. In addition, CCM3 has been implicated in several
signaling pathways potentially relevant to the formation and maintenance of polarized
adhesion complexes [10, 11, 14, 15]. CCM3 and the GCKIII kinases have been shown to
regulate disparate cellular events, such as cell proliferation and Golgi assembly, as well
as pro-survival and pro-apoptotic events [13, 16-20]. However, functional significance of
these interactions is still being elucidated as it is undetermined how they contribute to
CCM3 functions within and outside of CCM biology. A novel interaction of CCM3 with
Stk25, the GCKIII kinase that is part of the newly identified LKB1-STK25-GM130 cell
polarity complex, suggests a role for CCM3 in cell polarity [17].
LKB1 is an evolutionarily conserved protein kinase that regulates key signaling events
such as driving and maintaining cell polarity in a variety of tissues [21]. Mutations in
LKB1 cause the hamartoma syndrome Peutz-Jeghers in which patients display an increased susceptibility to several cancers [22, 23]. As a master kinase, LKB1 phosphorylates and activates more than a dozen AMPK-related kinases and therefore
47 Chapter 2: CCM3 regulates epithelial cell polarity coordinates multiple cellular events including cell metabolism, proliferation and polarization. In the basal state LKB1 is retained in the nucleus where it is subject to degradation, but it is rendered more stable by binding the coactivator STRAD, which exports it to the cytoplasm and also allosterically activates its kinase activity [24-26].
Heat-shock protein 90 and STRADα regulate LKB1 protein stability by preventing degradation and promoting cytoplasmic compartmentalization respectively [27, 28].
However, not much is known about the regulation of STRADα.
Our lab previously demonstrated that the GCKIII kinase Stk25 associates with STRADα
[17, 29] and that Stk25 knockdown disrupts the LKB1-STRADα-dependent polarization of W4 epithelial cells [17, 30]. The role of Stk25 in polarity regulation extends to developing neurons as Stk25 knockdown reduces axon initiation, a phenotype paralleling that caused by LKB1 elimination [17, 31, 32]. Conversely, overexpressing Stk25 partially rescues the neuronal polarization defects caused by LKB1 knockdown [17]. Stk25 is not known to be an LKB1 substrate and it is unknown how it participates in this pathway.
These findings suggest Stk25 may function as an accessory adaptor in the LKB1-
STRADα signaling complex, as Stk25 has multiple binding partners [8, 14, 17, 18]. Here we investigated whether one such Stk25-binding protein, CCM3, might also contribute to regulating the LKB1-STRAD pathway.
As an adaptor protein with key roles in the modulation of multiple signaling pathways, we investigated whether CCM3, a STK25-binding protein, contributes to the regulation of LKB1-STRAD pathway. We demonstrate that CCM3 interacts with the pseudokinase
48 Chapter 2: CCM3 regulates epithelial cell polarity
STRADα and participates in LKB1-STRADα-dependent cell polarization. CCM3 protects STRADα from proteolytic degradation and increases the protein levels of both
LKB1 and the downstream LKB1 effector kinase MARK1, but not AMPK. We show that
CCM3 exerts some of its control over these signaling systems through the regulation of protein expression. Our data suggests a role for CCM3 in LKB1-STRADα-mediated epithelial cell polarization.
49 Chapter 2: CCM3 regulates epithelial cell polarity
MATERIALS AND METHODS Antibodies, expression constructs and drugs
For immunoprecipitation, Western blotting and immunocytochemistry the following
antibodies were used: anti-PDCD10 for CCM3 (Sigma Aldrich), anti-LKB1 (Cell
Signaling; Western blotting), anti-LKB1 (Upstate; immunocytochemistry), anti-STRADα
(Sigma Aldrich), anti-β-actin (AC-15, Sigma Aldrich), anti-YSK1 for Stk25 (Santa
Cruz), anti-HA 3F10 (Roche), anti-MYC 71D10 (Cell Signaling; Western blotting), anti-
MYC 9E10 (BD Transduction Laboratories; immunoprecipitation), anti-FLAG M2
(Sigma Aldrich); phalloidin conjugated to Alexa Fluor 568 (Invitrogen). AMPK,
phospho-AMPK and MARK1, (Cell Signaling).
The W4 cell line [30], plasmid pCMV5-HA-HA-CCM3 [13] and HA-STRADα constructs were kind gifts from Dr. Clevers, Dr. Zalvide and Dr. Barnes respectively.
CCM3 mutation constructs (pcDNA-myc-CCM3 and pcDNA-myc-CCM3Δ K4D) were made as follows: Mutations were generated by PCR using standard techniques. The N- terminal mutant of CCM3 (pcDNA-myc-CCM3Δ) lacks 18 amino acids in the N- terminus dimerization domain corresponding to binding site of the GCKIII kinases
(previously described by [14]). The double mutant (pcDNA-myc-CCM3Δ K4D) contains the N-terminal deletion described above and mutations of quadruple Lysine to Aspartate
(K132D, K139D, K172D, K179D) changes at the C-terminal as previously described
[33]. Mutant CCM3 were subcloned into the pcDNA3.1 vector using Gibson Assembly kit (New England BioLabs) for expression in mammalian cells. Other expression constructs have been previously described [17].
Doxycycline, puromycin, and lactacystin were purchased from Sigma-Aldrich.
50 Chapter 2: CCM3 regulates epithelial cell polarity
shRNA-mediated knockdown
Control and CCM3 lentiviral short-hairpin RNA (shRNAs) constructs in the pLKO.1-
puro vector were obtained from Sigma-Aldrich. The CCM3 shRNA target 1 (Fig. S3;
GCTGATGATGTAGAAGAG) was used throughout this study. The lentivirus production was described previously [17]. Briefly, recombinant lentiviral particles containing shRNA carrying plasmids were produced by co-transfecting HEK293FT cells with empty pLKO.1-puro control vector or pLKO.1-puro/CCM3 shRNA plasmids, the envelope vector pCMV-VSVG and packaging vector pLP1 [34]. Infectious lentiviral particles were harvested by collecting the culture media 48h later. W4 cells were infected by adding 1ml lentivirus media to cells plated in a six-well dish. Twenty-four hours later, media were replaced with standard growth medium containing 2μg/ml puromycin, and cells were maintained in this selective medium for one week prior to being used in experiments. The ability of lentivral delivered shRNAs to reduce CCM3 expression was evaluated by Western blotting (Fig. S3).
Cell culture and Drug treatments
HEK293FT and W4 cells were cultured in Dubelcco’s modified Eagle’s medium
(DMEM) supplemented with 10% fetal bovine serum and 1% Penicillin-Streptomycin-
Glutamine (Gibco) at 37 °C in 5% CO2.
Lactacystin (10μM) was added to culture for 4h before collecting cell lysates for protein
analysis.
51 Chapter 2: CCM3 regulates epithelial cell polarity
Immunoprecipitation assay
To assay for protein-protein interactions, we performed coimmunoprecipitation assays
essentially as described previously [17]. Briefly, HEK293FT cells (seeded at 3 x105
cells/cm2 in six- well dishes) were transfected with 2μg of each indicated expression
vector using 3μl X-tremeGENE HP DNA transfection reagent (Roche). Cells were lysed
48h after transfection in NP40 immunoprecipitation buffer (NP40IB; 137 mM NaCl, 2.7 mM KCl, 1% NP-40, 25 mM Tris-HCl[pH 7.4], 2mM EDTA, 50mM NaF, protease inhibitor [complete mini, EDTA free; Roche]). Lysates were cleared by centrifugation at
14,000 x g for 20 minutes and then MYC-STRAD was immunoprecipitated by incubating with 1µg of anti-MYC (9E10) incubated for 90 minutes. Protein antibody complexes were immunoprecipitated by incubation with protein A/G plus Sepharose beads for 1 hr at 4oC (Santa Cruz). Immunocomplexes were collected by centrifugation, washed four
times with NP40 IB, eluted by boiling in 2X SDS sample buffer (4% SDS, 0.125M Tris
HCl, [pH 6.8], 50 mM dithiothreitol (DTT), 20% sucrose, 0.02% bromophenol blue).
Immunoprecipitated proteins were analyzed by Western blotting as described below.
Immunoblotting
To examine protein levels in W4 cells, transduced cells stably expressing CCM3 or
control shRNAs were seeded on collagen-coated six-well dishes at 3 x 105 cells/cm2 and
grown for 16h-24h. The cells were then treated with doxycycline (1μg/μl) for 24h or
otherwise indicated times. Homogenates were prepared by rinsing cells with ice cold PBS
prior to being lysed with RIPA lysis buffer (137mM NaCl, 2.7mM KCl, 1% NP-40,
52 Chapter 2: CCM3 regulates epithelial cell polarity
25mM Tris-HCl[pH 7.4], 2mM EDTA, 1% sodium deoxycolate, 0.1% sodium dodecyl
sulfate (SDS), 50mM NaF, 0.1% 2-mercaptoethanol, protease inhibitor [complete mini,
EDTA free; Roche, and Phosphatase Inhibitor Cocktail I, [Sigma]). Lysates were
incubated on ice and cleared by centrifugation at 14,000 x g for 20 min.
SDS-Page and Western Blotting
Cell lysates were mixed with an equal volume of 2X SDS sample buffer; heat denatured
at 100 °C for 5 min and resolved on 4-12% Bis-Tris gradient gels (Invitrogen). The
proteins were transferred to PVDF membranes (Thermo Scientific), blocked with 5%
milk in Tris-buffered saline containing 0.1% Tween 20 (Sigma), incubated with the
indicated primary antibodies followed by horseradish peroxidase-conjugated secondary
antibodies, and visualized with X-ray film using ECL chemiluminescence substrate
(Thermo Scientific). Blots were scanned using identical parameters, and band intensities were quantified using ImageJ (NIH). The bands were normalized to actin as a loading control. Statistical significance was determined using the Student’s unpaired two-tailed t-
test.
Cell polarization assay
Transduced W4 cells stably expressing shRNA constructs were seeded on collagen-
coated coverslips (0.1mg/ml, BD #354236) at a density of 2 x 104 cells/cm2 and grown
for 24h. Cells were either untreated or treated with doxycycline (1μg/ml) for 24h to
induce STRAD expression and subsequently LKB1-induced polarization [30]. Cells were
then fixed (4% paraformaldehyde) for 10 minutes at room temperature followed by
53 Chapter 2: CCM3 regulates epithelial cell polarity permeabilization (0.1% saponin in PBS) as previously described [17]. Cells were stained with a polyclonal anti-LKB1 antibody (Upstate) and actin filaments were visualized with phalloidin-Alexa Fluor 568 (Invitrogen). DNA was stained with DAPI (Sigma). Images were collected on a DeltaVision Deconvolution Microscope (Applied Precision) at 60X magnification.
54 Chapter 2: CCM3 regulates epithelial cell polarity
RESULTS CCM3 binds to STRADα independent of STK25
We have previously shown that Stk25 forms a stable complex with STRADα and that it
has a role in LKB1-STRADα-regulated cell polarization [17]. However, we were unable
to determine whether this interaction was direct and we sought to determine if CCM3, a
well-established Stk25 binding partner, bridges the complex formation between Stk25 and STRADα.
We created a MYC-tagged CCM3 deletion mutant (CCM3Δ) that lacks portions of the N- terminal dimerization domain (18 amino acids), the domain that binds to the GCKIII family of protein kinases including STK25 [13, 35] and used it to determine whether
CCM3 associated with STRAD and if this was dependent on Stk25. We confirmed that this CCM3 deletion mutant (MYC-CCM3Δ) fails to interact with STK25 (Figure. S2.1).
We expressed combinations of MYC-tagged CCM3 wild-type (myc-CCM3), MYC- tagged CCM3Δ, and HA-tagged STRADα in HEK293T cells, immunoprecipitated
CCM3 or CCM3Δ with anti-MYC tag antibody and assayed for coimmunoprecipitation by Western blotting (Figure. 2.1A). HA-STRADα co-precipitates with both MYC-CCM3 and MYC-CCM3Δ (Figure. 2.1A, B) with STRADα co-immunoprecipitating significantly stronger with CCM3Δ.
Our finding suggests CCM3 interacts with STRADα and does not require STK25 for this interaction. The increased binding observed to CCM3Δ is potentially due to conformational changes induced by the inability of CCM3 to homodimerize, thus
55 Chapter 2: CCM3 regulates epithelial cell polarity
enhancing binding to STRADα. It is also possible that with STK25 unable to bind to
CCM3, STRAD can easily access CCM3 due to less steric hindrance.
Given that Stk25 associates with STRADα [17] and CCM3 forms a complex with Stk25
at the Golgi apparatus [13], it is possible that CCM3 is necessary for the formation of the
Stk25-STRADα complex. To determine whether CCM3 was necessary for the formation
of Stk25-STRADα complex, we expressed combinations of MYC-tagged STRADα and
HA-Stk25 in control and CCM3 shRNA- expressing HEK293T cells. We
immunoprecipitated STRADα with anti-MYC tag antibody and assayed for
coimmunoprecipitation by Western blotting (Figure S2.2). HA-Stk25 co-precipitates with
MYC-STRADα in control and CCM3 shRNA expressing cells. Interestingly, the
association between Stk25 and STRADα is augmented when CCM3 is depleted
suggesting that the presence of CCM3 is not required for the interaction and instead
impedes the formation of the Stk25-STRAD complex.
CCM3-FAT homology domain interacts with STRADα
CCM3 has two domains; an N-terminal dimerization domain involved in homodimer formation by binding adjacent CCM3 and a C-terminal focal adhesion targeting (FAT)-
homology domain [33]. This FAT-homology domain is necessary for proper folding and stability of the protein and is a binding site for multiple partners including CCM2,
Paxillin and Striatins [12, 35, 36]. Given STRADα interacts strongly with a CCM3 mutant that lacks the N-terminal dimerization domain, we hypothesized that STRAD associates with the FAT-homology domain on CCM3.
56 Chapter 2: CCM3 regulates epithelial cell polarity
To test this hypothesis, we created a MYC-tagged CCM3 double mutant (CCM3Δ K4D) that disrupts the highly conserved FAK-like hydrophobic pocket (HP1) of the FAT- homology domain. We introduced a quadruple lysine-to-aspartate amino acid change at the HP1 binding site, which alters the hydrophobicity of the HP1 site on the CCM3 FAT- homology domain (as previously described by [33]). This mutation has previously been shown to disrupt interactions of Paxillin and CCM2 to the FAT-homology domain.
We expressed combinations of MYC-tagged CCM3Δ, CCM3Δ K4D and HA-tagged
STRADα in HEK293T cells, immunoprecipitated CCM3Δ or CCM3Δ K4D via the MYC tag (Figure 2.2). We find that STRADα fails to co-immunoprecipitate with CCM3Δ K4D, suggesting STRAD interacts with the FAT-homology domain of CCM3 (Figure 2.2). We observed that the protein expression levels of CCM3Δ K4D were lower than those of
CCM3Δ and this was a concern in interpreting the data. However, the STRADα signal in the IP was stronger with CCM3Δ and undetectable at any exposure with CCM3Δ K4D suggesting that the FAT domain is required for the interaction (Figure 2.2).
CCM3 is necessary for LKB1-STRADα-regulated epithelial cell polarization
The activation of LKB1 is sufficient to polarize isolated intestinal epithelial cells and this has been elegantly demonstrated using W4 intestinal epithelial cells (3). These cells are engineered to express FLAG-tagged STRADα under the control of a doxycycline inducible promoter in addition to constitutive LKB1 expression [30]. Doxycycline treatment leads to an increase in STRADα levels within 24h and a concomitant increase in LKB1 levels in these cells. The development of polarized actin brush borders, as
57 Chapter 2: CCM3 regulates epithelial cell polarity
visualized using phalloidin, is apparent in isolated epithelial cells at 24h of treatment [29,
30].
In this study, we employed this W4 intestinal cell paradigm to ask whether CCM3 is
required for LKB1-STRADα-regulated cell polarization. We engineered a lentiviral
delivery system to transduce W4 cells with shRNA against CCM3 to stably knockdown
CCM3 expression (Figure S2.3). After 24h of doxycycline treatment, only approximately
16% (+/- 2%) of CCM3 knockdown cells had a polarized actin-rich brush border (Figure
2.3A, 3B; phalloidin stain, red), whereas approximately 40% (+/- 9%) of control vector-
expressing cells were polarized (Figure 2.3B). To determine if the effect of the
knockdown was specific to CCM3, we generated an adenovirus that overexpresses an
shRNA-resistant version of CCM3 (HA-CCM3*). HA-CCM3* expression significantly rescued the polarization defect caused by CCM3 shRNA expression (Figure 2.3B). It did not, however, augment the polarization of control lentivirus-infected cells, suggesting that endogenous CCM3 expression is above a required threshold for polarization in these cells and that increasing CCM3 levels above endogenous levels has no additional polarizing effect (Figure 2.3B). In these cells, LKB1 levels were observed to rise in doxycycline-treated control cells as expected (Figure 2.3A). However, in cells infected with the CCM3 shRNA-expressing lentivirus, LKB1 levels appeared lower.
LKB1 and STRADα expression are dependent on CCM3 but not Stk25
It has previously been shown that CCM3 regulates the stability of protein levels of
binding partners, including Stk25 and VEGFR2 [11, 13]. Therefore we examined the
58 Chapter 2: CCM3 regulates epithelial cell polarity
possibility that CCM3 also modulates protein levels of LKB1 and/or STRADα. To
address this, we prepared lysates from untreated and doxycycline-treated control and
CCM3 shRNA-expressing W4 cells and measured protein levels by Western blotting.
Doxycycline induces STRADα expression, which increases LKB1 levels and induces
polarization in W4 cells within 24h (Figure 2.4A, 2.3A). LKB1 and STRADα protein
levels increase upon doxycycline treatment for both control and CCM3 shRNA virus-
infected cells (Figure 2.4A). However, the levels of LKB1 and STRADα were
significantly lower in CCM3 knockdown cells, both doxycycline treated and untreated
(Figure 2.4A, C). Stk25 levels were also reduced in CCM3 knockdown cells consistent
with previous findings (Figure 2.4A, C) [13].
We observed that re-expression of CCM3 with the HA-CCM3*-expressing adenovirus
was capable of restoring cell polarization to CCM3 shRNA-expressing cells (Figure
2.3B). Thus, we examined whether it would also rescue LKB1 and STRADα protein
levels. Before doxycycline induction, STRADα was not detected in our blots and
therefore not quantified. We observed statistical differences in the protein levels of both
LKB1 and STRADα in the absence of HA-CCM3* after doxycycline induction (Figure
2.4B, C). In cells expressing HA-CCM3* the amount of LKB1 and STRADα was rescued to a level where there was no longer any statistical significance between CCM3 knockdown and control cells (Figure 2.4C), though the levels of LKB1 and STRADα were only partially restored to control levels. It is important to note that HA-CCM3* was expressed at a lower level than the endogenous CCM3 protein, thus the inability to
59 Chapter 2: CCM3 regulates epithelial cell polarity
restore protein levels to wildtype levels may reflect this disparity (Figure 2.4B, C). Our
results suggest that CCM3 enhances protein levels of LKB1 and STRADα expression.
Since Stk25 is a binding partner for CCM3 and STRAD, we hypothesized that Stk25 may
play an important role in the LKB1-STRAD pathway by regulating protein stability.
When Stk25 levels were reduced with a lentiviral transducing shRNA, we did not observe any discernable reduction in either LKB1 or STRADα protein levels (Figure S2.5). This suggests that Stk25 may not operate with CCM3 to regulate LKB1 and STRADα levels, but may have a different role on the LKB1-STRAD pathway.
CCM3 expression regulates protein stability of STRADα and LKB1
STRADα recruits LKB1 to the cytoplasm where it is protected from proteasomal degradation through a process that is thought to involve Hsp90 and its kinase-specific targeting subunit Cdc37 [27, 37]. STRADα stability has been shown to be regulated by
LKB1 [28, 38, 39], yet it is unknown if additional LKB1-independent mechanisms exist to control levels of STRADα.
To determine if CCM3 regulates STRADα and/or LKB1 protein stability, we blocked proteasome activity by lactacystin treatment. This resulted in a restoration of both LKB1 and STRADα protein expression in CCM3 knockdown cells up to control cell levels
(compare Figure 2.4A and 2.4D). Interestingly, CCM3 knockdown appears to reduce
LKB1 protein expression in the absence of induced STRADα (Figure 2.4A, 2.4D, 4C) suggesting a potential role for CCM3 in LKB1 protein regulation independent of STRAD
60 Chapter 2: CCM3 regulates epithelial cell polarity expression. These results confirm CCM3 regulates LKB1 and STRADα protein levels by affecting degradation through the proteasome.
CCM3 knockdown has differential effects on pathways downstream of LKB1
LKB1 activates several downstream kinases involved in energy homeostasis and cell polarization, including AMP-activated protein kinases (AMPKs) and microtubule affinity- regulating kinases (MARKs 1-4) through the phosphorylation of their activation loops [26, 40-42]. LKB1 is required to activate AMPK in response to energy stress [26,
40, 41, 43]. AMPK activation leads to the phosphorylation of effectors resulting in increased energy uptake by the cell and reduced energy consumption [44]. The MARK kinases are thought to regulate cell polarity in part by phosphorylating microtubule- binding proteins (MAPs) such as Tau on the microtubule binding repeats (KXGS), leading to detachment of tau or equivalent MAPs and altered intracellular trafficking
[45].
To further characterize the effects of CCM3 on the LKB1 pathway, we examined how
CCM3 knockdown influenced MARK1 and AMPK. We found that CCM3 knockdown led to a dramatic reduction in MARK1 protein levels as compared to control samples
(Figure 2.5A, B). HA-CCM3* adenovirus infection rescued expression of MARK1 in
CCM3 shRNA-expressing cells, indicating that the observed effect is the result of reduced CCM3 levels (Figure 2.5C, D).
Surprisingly, CCM3 knockdown led to an increase in both AMPK protein expression and a corresponding increase in its phosphorylation in uninduced cells as compared to control
61 Chapter 2: CCM3 regulates epithelial cell polarity
cells (Figure 2.5A, B). In lysates of doxycycline-induced cells, the levels of AMPK were elevated but phospho-AMPK Ser79 were similar between CCM3 shRNA- and control-
infected cells. In addition to LKB1 other kinases such as CAMKK2 phosphorylate
AMPK at Thr172 [46, 47] and it is possible these other kinases are regulating the protein
levels of AMPK phosphorylation.
62 Chapter 2: CCM3 regulates epithelial cell polarity
DISCUSSION The mechanisms by which CCM3 modulates several signaling pathways remain unclear and how these diverse functions influence its roles in the formation of CCM are currently under investigation. Our findings are consistent with the notion that CCM3 can function independently of CCM1 and CCM2 with roles outside of the CCM complex. Here we describe a novel role for CCM3 in LKB1-STRADα signaling pathway. CCM3 is necessary for LKB1-STRADα dependent epithelial cell polarization. We show that
CCM3 binds to the pseudokinase STRADα and modulates protein levels of LKB1 and
STRADα by decreasing degradation through the proteasome. In addition, CCM3 has differential effects in pathways downstream of LKB1. These findings place CCM3 as a key regulator of the conserved LKB1 signaling pathway and define a role for CCM3 in epithelial polarity (Figure 2.6).
An emerging theme in CCM3 function seems to be its ability to elevate or stabilize protein levels of direct and indirect binding partners. CCM3 regulates the protein levels of activated VEGFR2 and the GCKIII kinases including Stk25 [11, 13]. Here we show that depletion of CCM3 decreases the protein levels of STRADα and CCM3 partially protects LKB1 and STRADα from proteasome degradation (Figure 2.4). It remains to be determined how this is achieved. For instance, CCM3 could regulate a post-translational modification of bound proteins that influences protein stability, such as ubiquitination.
Alternatively it could regulate the activity of Hsp90 or other chaperone proteins towards client proteins. Both LKB1 and VEGFR2 are known HSP90 clients, whereas it remains unclear whether CCM3 or Stk25 interact with HSP90. The demonstration that CCM3 regulates STRADα levels to influence LKB1 signaling opens up the possibility that this
63 Chapter 2: CCM3 regulates epithelial cell polarity is a means to activating the pathway. It will be interesting to determine if upstream LKB1 regulators affect the CCM3-STRADα interaction or regulate CCM3 protein levels.
CCM3 plays an important role as a regulator of multiple signaling pathways. Previously,
CCM3 has been shown to have both pro- and anti-apoptotic functions [16, 19, 20], regulate the integrity and polarity of the Golgi apparatus [13], regulate VEGF receptor signaling [11] and regulate GCKIII kinase function [14, 15, 48]. The ability of CCM3 to differentially regulate the protein levels of two kinases downstream of LKB1 (Figure 2.5) suggests a diverse function in LKB1-mediated biological functions. It is possible that these regulations are cell type specific or depending on cell cycle. Animal models of cell- type specific LKB1 knockdowns will be useful to determine a role for CCM3 in diverse
LKB1 functions.
Loss of function mutations in CCM3 cause cerebral cavernous malformations in humans
[1-3]. The molecular and cellular mechanisms that lead to this vascular disease are not fully understood. Interestingly, LKB1-null mutant mice die at embryonic day 9.5 with vascular defects and deregulated VEGF levels [49]. It remains to be determined if LKB1 has a role downstream of VEGF receptor signaling. Findings in this study suggest that this is worth evaluating. Specifically, it would be interesting to determine whether the polarity of endothelial cells in the cavernous vasculature are deregulated and whether
LKB1 might play a role in this disease phenotype.
CCM3 acts as an adaptor protein to link signaling proteins in a variety of signaling complexes. One of the best characterized is known as the STRIPAK complex in which
64 Chapter 2: CCM3 regulates epithelial cell polarity
CCM3 links the GCKIII kinases with the phosphatase PP2A [8, 50, 51]. Interestingly,
CCM3 depletion reduces the binding of the GCKIII kinase Mst4 with PP2A but increases its interaction with the Golgi apparatus [35]. This suggests that CCM3 acts to regulate the binding partners of the GCKIII kinases.
Here we find that CCM3 interacts with STRADα, an allosteric activator of LKB1. The interaction between CCM3 and STRADα is augmented when STK25 is unable to associate with CCM3 (Figure 2.1). However, there was no impact on STRADα protein levels when Stk25 levels were depleted (Figure S2.5) suggesting this increase in interaction is not due to increased protein levels. However, it is possible that the inability of STk25 to interact with CCM3 opens up more room for STRAD binding. It is also possible that Stk25 will have a different function on the pathway, such as regulation of the Golgi apparatus. Indeed, we previously showed that LKB1 knockdown leads to disruption of the Golgi apparatus in embryonic neurons in a similar fashion to Stk25 knockdown [17]. LKB1-Stk25-GM130 signaling regulates Golgi morphology during cell polarization and at this point in time, the GCKIII kinases and CCM3 are the most plausible link between LKB1 and the Golgi apparatus.
As mentioned previously, the levels of LKB1 effectors were differentially regulated in
CCM3 knockdown cells (Figure 2.5). AMPK, which is activated by phosphorylation and
AMP binding during energy stress [46], was expressed and phosphorylated in the activation loop (Thr172) at levels equal to or above control samples (Fig. 5A, 5B). In addition to LKB1, CaMKKβ also phosphorylates Thr172 [47], and therefore it may maintain AMPK phosphorylation in the absence of LKB1. In contrast, the expression of
MARK1, which is the mammalian par-1 homolog that acts downstream of LKB1 to
65 Chapter 2: CCM3 regulates epithelial cell polarity regulate cell polarization [52], is reduced in CCM3 knockdown cells, and this defect was rescued by re-expression of CCM3 (Fig. 5A, 5B, 5C, 5D). It remains to be determined how CCM3 regulates MARK1 protein levels. It could, for instance, directly stabilize
MARK1 or stabilize another protein that regulates MARK1 protein levels.
This study identifies a new function for CCM3 with a novel mechanism by which CCM3, in association with STRADα, influences LKB1 signaling to regulate epithelial cell polarization (Figure 6). CCM3 concurrently exists in two known complexes; CCM1-
CCM2 complex and STRIPAK complex. In association with STRIPAK, CCM3 maintains Golgi stability and endocytic recycling [35] [53]. In contrast, interaction with the CCM1-CCM2 complex maintains the stability of the plasma membrane and regulates the activation of RhoA signaling. Our findings that CCM3 exists in a third complex to maintain cell polarity suggest a complex balance in the diverse functions as well as the importance of CCM3 in our understanding of the intrinsic mechanisms by which cavernous malformations develop in CCM disease.
66 Chapter 2: CCM3 regulates epithelial cell polarity
REFERENCES 1. Labauge, P., et al., Genetics of cavernous angiomas. Lancet Neurol, 2007. 6(3): p.
237-44.
2. Bergametti, F., et al., Mutations within the programmed cell death 10 gene cause cerebral cavernous malformations. Am J Hum Genet, 2005. 76(1): p. 42-51.
3. Plummer, N.W., J.S. Zawistowski, and D.A. Marchuk, Genetics of cerebral cavernous malformations. Curr Neurol Neurosci Rep, 2005. 5(5): p. 391-6.
4. Maddaluno, L., et al., EndMT contributes to the onset and progression of cerebral cavernous malformations. Nature, 2013. 498(7455): p. 492-6.
5. Borikova, A.L., et al., Rho kinase inhibition rescues the endothelial cell cerebral cavernous malformation phenotype. J Biol Chem, 2010. 285(16): p. 11760-4.
6. Stockton, R.A., et al., Cerebral cavernous malformations proteins inhibit Rho kinase to stabilize vascular integrity. J Exp Med, 2010. 207(4): p. 881-96.
7. Whitehead, K.J., et al., The cerebral cavernous malformation signaling pathway promotes vascular integrity via Rho GTPases. Nat Med, 2009. 15(2): p. 177-84.
8. Goudreault, M., et al., A PP2A phosphatase high density interaction network identifies a novel striatin-interacting phosphatase and kinase complex linked to the cerebral cavernous malformation 3 (CCM3) protein. Mol Cell Proteomics, 2009. 8(1): p.
157-71.
9. Voss, K., et al., CCM3 interacts with CCM2 indicating common pathogenesis for cerebral cavernous malformations. Neurogenetics, 2007. 8(4): p. 249-56.
10. Ma, X., et al., PDCD10 interacts with Ste20-related kinase MST4 to promote cell growth and transformation via modulation of the ERK pathway. Mol Biol Cell, 2007.
18(6): p. 1965-78.
67 Chapter 2: CCM3 regulates epithelial cell polarity
11. He, Y., et al., Stabilization of VEGFR2 signaling by cerebral cavernous
malformation 3 is critical for vascular development. Sci Signal, 2010. 3(116): p. ra26.
12. Li, X., et al., Molecular recognition of leucine-aspartate repeat (LD) motifs by the focal adhesion targeting homology domain of cerebral cavernous malformation 3
(CCM3). J Biol Chem, 2011. 286(29): p. 26138-47.
13. Fidalgo, M., et al., CCM3/PDCD10 stabilizes GCKIII proteins to promote Golgi assembly and cell orientation. J Cell Sci, 2010. 123(Pt 8): p. 1274-84.
14. Voss, K., et al., Functional analyses of human and zebrafish 18-amino acid in- frame deletion pave the way for domain mapping of the cerebral cavernous malformation
3 protein. Hum Mutat, 2009. 30(6): p. 1003-11.
15. Yoruk, B., et al., Ccm3 functions in a manner distinct from Ccm1 and Ccm2 in a
zebrafish model of CCM vascular disease. Dev Biol, 2012. 362(2): p. 121-31.
16. Chen, L., et al., Apoptotic functions of PDCD10/CCM3, the gene mutated in
cerebral cavernous malformation 3. Stroke, 2009. 40(4): p. 1474-81.
17. Matsuki, T., et al., Reelin and stk25 have opposing roles in neuronal polarization
and dendritic Golgi deployment. Cell, 2010. 143(5): p. 826-36.
18. Preisinger, C., et al., YSK1 is activated by the Golgi matrix protein GM130 and
plays a role in cell migration through its substrate 14-3-3zeta. J Cell Biol, 2004. 164(7):
p. 1009-20.
19. Pombo, C.M., et al., Activation of the Ste20-like oxidant stress response kinase-1
during the initial stages of chemical anoxia-induced necrotic cell death. Requirement for
dual inputs of oxidant stress and increased cytosolic [Ca2+]. J Biol Chem, 1997.
272(46): p. 29372-9.
68 Chapter 2: CCM3 regulates epithelial cell polarity
20. Lauenborg, B., et al., Programmed cell death-10 enhances proliferation and
protects malignant T cells from apoptosis. APMIS, 2010. 118(10): p. 719-28.
21. Baas, A.F., L. Smit, and H. Clevers, LKB1 tumor suppressor protein: PARtaker in
cell polarity. Trends Cell Biol, 2004. 14(6): p. 312-9.
22. Hemminki, A., et al., A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature, 1998. 391(6663): p. 184-7.
23. Alessi, D.R., K. Sakamoto, and J.R. Bayascas, LKB1-dependent signaling pathways. Annu Rev Biochem, 2006. 75: p. 137-63.
24. Boudeau, J., et al., MO25alpha/beta interact with STRADalpha/beta enhancing their ability to bind, activate and localize LKB1 in the cytoplasm. EMBO J, 2003. 22(19): p. 5102-14.
25. Baas, A.F., et al., Activation of the tumour suppressor kinase LKB1 by the STE20- like pseudokinase STRAD. EMBO J, 2003. 22(12): p. 3062-72.
26. Hawley, S.A., et al., Complexes between the LKB1 tumor suppressor, STRAD alpha/beta and MO25 alpha/beta are upstream kinases in the AMP-activated protein kinase cascade. J Biol, 2003. 2(4): p. 28.
27. Nony, P., et al., Stability of the Peutz-Jeghers syndrome kinase LKB1 requires its binding to the molecular chaperones Hsp90/Cdc37. Oncogene, 2003. 22(57): p. 9165-75.
28. Veleva-Rotse, B.O., et al., STRAD pseudokinases regulate axogenesis and LKB1 stability. Neural Dev, 2014. 9: p. 5.
29. ten Klooster, J.P., et al., Mst4 and Ezrin induce brush borders downstream of the
Lkb1/Strad/Mo25 polarization complex. Dev Cell, 2009. 16(4): p. 551-62.
69 Chapter 2: CCM3 regulates epithelial cell polarity
30. Baas, A.F., et al., Complete polarization of single intestinal epithelial cells upon
activation of LKB1 by STRAD. Cell, 2004. 116(3): p. 457-66.
31. Barnes, A.P., et al., LKB1 and SAD kinases define a pathway required for the
polarization of cortical neurons. Cell, 2007. 129(3): p. 549-63.
32. Shelly, M., et al., LKB1/STRAD promotes axon initiation during neuronal
polarization. Cell, 2007. 129(3): p. 565-77.
33. Li, X., et al., Crystal structure of CCM3, a cerebral cavernous malformation
protein critical for vascular integrity. J Biol Chem, 2010. 285(31): p. 24099-107.
34. Rubinson, D.A., et al., A lentivirus-based system to functionally silence genes in primary mammalian cells, stem cells and transgenic mice by RNA interference. Nat
Genet, 2003. 33(3): p. 401-6.
35. Kean, M.J., et al., Structure-function analysis of core STRIPAK Proteins: a signaling complex implicated in Golgi polarization. J Biol Chem, 2011. 286(28): p.
25065-75.
36. Zhang, M., et al., Structural mechanism of CCM3 heterodimerization with
GCKIII kinases. Structure, 2013. 21(4): p. 680-8.
37. Boudeau, J., et al., Heat-shock protein 90 and Cdc37 interact with LKB1 and regulate its stability. Biochem J, 2003. 370(Pt 3): p. 849-57.
38. Eggers, C.M., et al., STE20-related kinase adaptor protein alpha (STRADalpha) regulates cell polarity and invasion through PAK1 signaling in LKB1-null cells. J Biol
Chem, 2012. 287(22): p. 18758-68.
70 Chapter 2: CCM3 regulates epithelial cell polarity
39. Guerreiro, A.S., et al., A sensitized RNA interference screen identifies a novel role
for the PI3K p110gamma isoform in medulloblastoma cell proliferation and
chemoresistance. Mol Cancer Res, 2011. 9(7): p. 925-35.
40. Woods, A., et al., LKB1 is the upstream kinase in the AMP-activated protein kinase cascade. Curr Biol, 2003. 13(22): p. 2004-8.
41. Shaw, R.J., et al., 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, 2004. 101(10): p. 3329-35.
42. Lizcano, J.M., et al., LKB1 is a master kinase that activates 13 kinases of the
AMPK subfamily, including MARK/PAR-1. EMBO J, 2004. 23(4): p. 833-43.
43. Shaw, R.J., et al., The kinase LKB1 mediates glucose homeostasis in liver and therapeutic effects of metformin. Science, 2005. 310(5754): p. 1642-6.
44. Hardie, D.G. and K. Sakamoto, AMPK: a key sensor of fuel and energy status in skeletal muscle. Physiology (Bethesda), 2006. 21: p. 48-60.
45. Timm, T., et al., MARKK, a Ste20-like kinase, activates the polarity-inducing kinase MARK/PAR-1. EMBO J, 2003. 22(19): p. 5090-101.
46. Hardie, D.G., AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nat Rev Mol Cell Biol, 2007. 8(10): p. 774-85.
47. Steinberg, G.R. and B.E. Kemp, AMPK in Health and Disease. Physiol Rev,
2009. 89(3): p. 1025-78.
48. Zheng, X., et al., CCM3 signaling through sterile 20-like kinases plays an essential role during zebrafish cardiovascular development and cerebral cavernous malformations. J Clin Invest, 2010. 120(8): p. 2795-804.
71 Chapter 2: CCM3 regulates epithelial cell polarity
49. Ylikorkala, A., et al., Vascular abnormalities and deregulation of VEGF in Lkb1- deficient mice. Science, 2001. 293(5533): p. 1323-6.
50. Ceccarelli, D.F., et al., CCM3/PDCD10 heterodimerizes with germinal center kinase III (GCKIII) proteins using a mechanism analogous to CCM3 homodimerization. J
Biol Chem, 2011. 286(28): p. 25056-64.
51. Gordon, J., et al., Protein Phosphatase 2A (PP2A) Binds Within the
Oligomerization Domain of Striatin and Regulates the Phosphorylation and Activation of the Mammalian Ste20-Like Kinase Mst3. BMC Biochem, 2011. 12(1): p. 54.
52. Biernat, J., et al., Protein kinase MARK/PAR-1 is required for neurite outgrowth and establishment of neuronal polarity. Mol Biol Cell, 2002. 13(11): p. 4013-28.
53. Lant, B., et al., CCM-3/STRIPAK promotes seamless tube extension through endocytic recycling. Nat Commun, 2015. 6: p. 6449.
72 Chapter 2: CCM3 regulates epithelial cell polarity
FIGURES
73 Chapter 2: CCM3 regulates epithelial cell polarity
Figure 2.1. STRADα interacts with both wildtype and mutant CCM3.
A) Cell lysates from HEK293T cells transfected with empty vector (EV), MYC-CCM3
(wildtype), MYC-CCM3Δ (mutant) or HA-STRADα as indicated, were immunoprecipitated with anti-MYC antibody 9E10. The immunoprecipitates (IP; top panels) and total cell lysates (TCL; bottom panels) were immunoblotted with the indicated antibodies on the left. HA-STRADα was coimmunoprecipitated when both
MYC-CCM3 and MYC-CCM3Δ were present in the cell lysates, suggesting a complex is formed with CCM3
B) Level of immunoprecipitation HA-STRAD was quantified by densitometry of the
Western blot images. The averages of three independent experiments are shown. Each sample was normalized to actin. The background value for the HA-STRAD alone control was set to zero (**p < 0.01, *p<0.05; Student’s two tailed paired t-test).
74 Chapter 2: CCM3 regulates epithelial cell polarity
75 Chapter 2: CCM3 regulates epithelial cell polarity
Figure 2.2. STRADα interacts with the FAT-homology domain of CCM3.
Cell lysates from HEK293T cells transfected with empty vector (EV), MYC- MYC-
CCM3Δ (mutant), MYC-CCM3ΔK4D (double mutant) or HA-STRADα as indicated, were immunoprecipitated with anti-MYC antibody 9E10. The immunoprecipitates (IP; top panels) and total cell lysates (TCL; bottom panels) were immunoblotted with the indicated antibodies on the left. HA-STRADα was coimmunoprecipitated when MYC-
CCM3Δ was present in the cell lysate but not when MYC-CCM3ΔK4D was present, suggesting STRAD interacts with the c-terminus domain of CCM3.
76 Chapter 2: CCM3 regulates epithelial cell polarity
77 Chapter 2: CCM3 regulates epithelial cell polarity
Figure 2.3. CCM3 acts on LKB1-STRADα pathway to regulate epithelial
polarization.
A) Empty vector (EV) or CCM3 shRNA-transfected W4 epithelial cells were untreated
(No Dox) or treated with doxycycline (+Dox; 24h) to induce STRADα expression.
Polarized brush borders were visualized by staining filamentous actin with phalloidin
toxin (red). LKB1 (green) was visualized using an anti-LKB1 antibody (Upstate). B) The
requirement for CCM3 expression on STRADα-induced epithelial cell polarization was
assessed in untreated and doxycycline-treated cultures by quantifying the percentage of cells that had polarized actin caps. CCM3 knockdown led to a statistically significant loss of polarization in STRADα-expressing cells. The polarization was rescued by expression of the shRNA-resistant HA-CCM3* (****p<0.0001,**p < 0.01; Student’s two tailed paired t-test, Error bars represent the SEM).
78 Chapter 2: CCM3 regulates epithelial cell polarity
79 Chapter 2: CCM3 regulates epithelial cell polarity
Figure 2.4. LKB1 and STRADα protein expression is CCM3- and proteasome- dependent.
A) W4 cells infected with empty vector (EV) or CCM3 shRNA (CCM3 shRNA) viruses were untreated (No Dox) or treated with doxycycline (+Dox; 24h) to induce STRADα expression and polarization. Immunoblotting with anti-LKB1, anti-Stk25, anti-STRADα and anti-CCM3 antibodies (indicated to the left of the panels) revealed that CCM3 knockdown led to reduced protein expression of LKB1, Stk25 and STRAD. B) Same as
A except the cells were infected with HA-CCM3*-overexpressing virus. The positions of
CCM3 and HA-CCM3* are indicated to the right of the CCM3 panel. C, E) For all panels excepting CCM3*/CCM3, the densitometry of each lane was normalized to actin, set relative to the EV control and averaged from at least three independent experiments. In the CCM3*/CCM3 panel, the normalized value of CCM3* was divided by the normalized value of endogenous CCM3 in the EV control lane. D) Same as A except the cells were treated with lactacystin, a proteasome inhibitor for 4 hours. (**** p < 0.0001,
***p<0.005 ,**p= 0.01; * p<0.05; Student’s two tailed paired t-test)
80 Chapter 2: CCM3 regulates epithelial cell polarity
81
Figure 2.5. CCM3 expression differentially regulates expression of LKB1 effectors.
(A) CCM3 knockdown reduces expression of MARK1 in both untreated (No Dox) and
doxycycline-treated (+Dox) samples. In contrast CCM3 knockdown increased AMPK
levels and AMPK Thr172 phosphorylation in cells not treated with doxycycline. (B)
Densitometry of protein bands averaged from three independent experiments. Protein
levels were normalized to actin expression and expressed relative to the EV control
sample. (C) HA-CCM3* expression rescued MARK1 protein levels in CCM3 shRNA- expressing cells. However, AMPK levels were not returned to baseline in CCM3 shRNA- expressing, doxycycline untreated cells. D) Densitometry results for HA-CCM3*- expressing cells. (**** p < 0.0001, **p< 0.01; * p<0.05; Student’s two tailed paired t- test) Chapter 2: CCM3 regulates epithelial cell polarity
83 Chapter 2: CCM3 regulates epithelial cell polarity
Figure 2.6. Model of CCM3 as a binding partner of STRADα acting to modulate
LKB1-mediated cell polarity.
LKB1 is known to act in a complex with STRAD and Stk25 to regular cellular polarity
[17, 23]. CCM3 interacts with STK25 to regulate Golgi orientation [13]. CCM3
coimmunoprecipitates with STRADα, independent of Stk25 ability to bind CCM3
(Figure 2.1). Knockdown of CCM3 causes a reduction in LKB1-STRAD mediated epithelial cell polarization (Figure 2.3A) suggesting a role for CCM3 in LKB1-mediated
cell polarization. The overexpression of CCM3 rescues these effects. CCM3 depletion
differentially modulates the protein expression of AMPK and MARK1, both downstream
effectors of LKB1 polarization signaling pathway (Figure 2.5).
84 Chapter 2: CCM3 regulates epithelial cell polarity
85 Chapter 2: CCM3 regulates epithelial cell polarity
SUPPLEMENTAL FIGURES
86 Chapter 2: CCM3 regulates epithelial cell polarity
Supplemental Figure 2.1. STK25 does not interact with mutated form of CCM3.
A) Cell lysates from HEK293T cells transfected with HA-STK25, HA-STRADα, MYC-
CCM3 (wildtype) or MYC-CCM3Δ (mutant) as indicated, were immunoprecipitated with anti-MYC antibody 9E10. The immunoprecipitates (IP; top panels) and total cell lysates
(TCL; bottom panels) were immunoblotted with the indicated antibodies. HA-ST25 was coimmunoprecipitated when MYC-CCM3 but not MYC-CCM3Δ was present in the cell lysates.
87 Chapter 2: CCM3 regulates epithelial cell polarity
88 Chapter 2: CCM3 regulates epithelial cell polarity
Supplemental Figure 2.2. CCM3 deletion enhances the interaction between STK25 and STRADα.
Cell lysates from empty vector shRNA (293T Control) or CCM3 shRNA-transfected
(293T CCM3 shRNA) HEK293T cells transfected with empty vector (EV), MYC-
STRADα, or HA-STK25, were immunoprecipitated with anti-MYC antibody 9E10. The immunoprecipitates (IP; top panels) and total cell lysates (TCL; bottom panels) were immunoblotted with the indicated antibodies. HA-STK25 was coimmunoprecipitated when MYC-STRADα was present in the cell lysates of both 293T Control cells and 293T
CCM3 shRNA, suggesting a complex is formed with STK25. The absence of CCM3 in
293T CCM3 shRNA enhanced the interaction between STK25 and STRADα.
89 Chapter 2: CCM3 regulates epithelial cell polarity
90 Chapter 2: CCM3 regulates epithelial cell polarity
Supplemental Figure 2.3. CCM3 shRNA vectors effectively reduce the expression of
CCM3 in HEK293T and W4 epithelial cells. W4 (A) and HEK293T (B) cells were
infected with empty vector control (EV) or CCM3 shRNA viruses (1-3) for 24 h and
selected in puromycin. Western blot analysis of the cell lysates with anti-CCM3 antibody showed that target 1 shRNA was the most effective and reduced CCM3 expression by greater than 90% as compare to uninfected cells (no virus). β-Actin was used as a loading control.
91 Chapter 2: CCM3 regulates epithelial cell polarity
92 Chapter 2: CCM3 regulates epithelial cell polarity
Supplemental Figure 2.4. LKB1 and STRADα protein expression are dependent on
CCM3 in 293T cells.
A) HEK293T cells infected with empty vector (EV) or CCM3 shRNA and selected in
puromycin. Immunoblotting with anti-LKB1, anti-STRADα and anti-CCM3 antibodies
(indicated to the left of the panels) revealed that CCM3 knockdown led to reduced protein expression of LKB1 and STRAD. The cells were also infected with HA-CCM3*- expressing virus. The positions of CCM3 and HA-CCM3* are indicated to the right of the CCM3 panel. The levels of LKB1 and STRAD were rescued in the CCM3 shRNA lysates.
B) Densitometry of protein bands averaged from three independent experiments. Protein levels were normalized to actin expression (**p<0.01; *p<0.05; Student’s two tailed paired t-test).
93 Chapter 2: CCM3 regulates epithelial cell polarity
94 Chapter 2: CCM3 regulates epithelial cell polarity
Supplemental Figure 2.5. Stk25 knockdown does not influence LKB1 or STRADα
protein expression in W4 cells. Knockdown of Stk25 by infection with a Stk25 shRNA
expressing lentivirus (pLL3.7 hStk25 shRNA [17]), did not reduce LKB1 or STRADα levels compared to control infected cells (representative gel from three separate experiments). The Stk25-shRNA infected cells were enriched by FAC sorting for GFP, which is also expressed by the virus.
Quantify the HA
95 Chapter 2: CCM3 regulates epithelial cell polarity
96 Chapter 3: Differential Gene Expression Profiling
CHAPTER 3
DIFFERENTIAL GENE EXPRESSION PROFILING OF ENDOTHELIAL CELLS DEFICIENT IN CEREBRAL CAVERNOUS MALFORMATION 3 (CCM3)
This chapter is an original copyright and has not been previously published: Mansaray-
Storms, ZY; Howell, BH (2016)
Chapter 3: Differential Gene Expression Profiling
ABSTRACT Cerebral cavernous malformation (CCM) is a neurovascular disease that leads to vascular lesions, hemorrhages and strokes. Cerebral cavernous malformation 3 (CCM3) is one of three genes which when mutated are responsible for CCM. The signaling pathways downstream of these three proteins are not fully understood especially how they contribute to CCM. The protein encoded by CCM3, is a small adaptor protein that seems to play a role in several signaling pathways involved in cell polarity, cell proliferation and apoptosis. The aim of this study was to probe more broadly for alterations in functional gene networks and identify novel target genes/molecules for CCM3 in endothelial cells. We depleted CCM3 levels using a lentivirus shRNA in human aortic endothelial cell (HAEC) cultures and isolated RNA from these cells. RNA was subjected to transcriptional analysis on microarrays and significant altered genes were analyzed using Qiagen’s Ingenuity Pathway Analysis (IPA) to understand the functional networks these genes play a role in. Furthermore, selected regulated transcript expressions were independently validated by quantitative RT-PCR. We found several signaling pathways and molecules that were dysregulated by at least two-fold change (50% change) within
CCM3 down-regulated cells, including the activation of p53 signaling and inhibition of
VEGF signaling. The top functional networks that were altered included vasculogenesis, angiogenesis, cell proliferation, apoptosis and cell movement; all of which were significantly inhibited or decreased. We identified novel CCM3-dependent regulation of key endothelial transcripts as well as novel upstream-regulated gene networks that may contribute to the endothelial specific phenotypes that accompany CCM disease.
98 Chapter 3: Differential Gene Expression Profiling
INTRODUCTION Cerebral cavernous malformations (CCMs) are vascular lesions affecting the
CNS. These lesions are abnormally large hamartomas formed by a simple layer of
capillary endothelial cells [1]. However, the cellular and molecular mechanisms of CCM
remain poorly understood. Heritable forms of CCM are caused by mutations in CCM1,
CCM2 and CCM3 [2, 3]. The loss-of-function mutations in CCM3 cause a more aggressive form of the disease than in CCM1 or CCM2 [4, 5] with symptoms such as focal neurological defects; headaches and stroke detected much earlier. This suggests the signaling pathways in which CCM3 is involved in might be critical in understanding the etiology of CCM disorders.
CCM3 binds to a variety of proteins. It binds to growth factor receptors
(VEGFR2) [6], cell adhesion molecules (GCKIII kinases; STK24, STK25 and MST4,
Paxillin, protocadherin-c ) [7] [8] [9] [10] [11], and acts as protein adaptor in protein complexes (CCM2, Striatins, STRADα) [12] [13] [14] Chapter 2. Through these interactions, CCM3 plays a role in the regulation of diverse cellular functions including cell survival, cell migration, cell polarity, angiogenesis and vascular remodeling, Golgi assembly and cell orientation. In many cases the biological significance of CCM3 binding to specific partner proteins is still incompletely understood.
CCM3 is involved in the regulation of apoptosis, however, the exact mechanisms remain inconclusive as several studies have demonstrated both pro-apoptotic and anti- apoptotic functions in vitro. It remains unclear whether CCM3 modulates apoptotic pathways involved in regulating cell survival or cell proliferation. It was recently reported that CCM3 is necessary for the induction of senescence, a process that is characterized by the activation of p53 [15, 16] suggesting a potential link between CCM3
99 Chapter 3: Differential Gene Expression Profiling
and the p53-apoptotic signaling pathway. p53 is a DNA-binding protein and stress-
activated transcription factor that regulates the expression of hundreds of genes
implicated in several biological processes including DNA repair, cell cycle, apoptosis and
tumor biology. There has been no link of p53 signaling with CCM3 biology, however
recent evidence demonstrated that loss of p53 sensitizes CCM1 mutant mice to
developing vascular lesions [17].
CCM3 regulates the stability of its binding partners including VEGFR2, STK25 and now STRADα [6, 10] (Chapter 2). The ability of CCM3 to bind to a variety of proteins and regulate the phosphorylation of master regulatory kinases suggests it may function to globally regulate gene transcription. In these studies, we set out to determine whether there is evidence to suggest that CCM3 regulates LKB1 and STRADα activity in
endothelial cells by looking at protein level and gene expressions. In addition, to further
understand how loss of CCM3 function contributes to the etiology of CCMs, we took a
functional genomics approach to identify gene networks alterations and identify novel
target genes or molecules in the CCM3 signaling pathway.
Here, we knockdown the expression of CCM3 in human aortic endothelial cells
(HAEC) and performed microarray analysis as well as RT-PCR validations of selected
genes of interest. We used Qiagen’s Ingenuity Pathway Analysis (IPA) to map out
functional gene networks that were differentially altered between control cells and CCM3
knockdown. We find that knockdown of CCM3 results in a global dysregulation of
transcription factors and proteins involved in signaling pathways that regulate cell
adhesion, cell proliferation, cell survival and apoptosis. Over 800 genes were
100 Chapter 3: Differential Gene Expression Profiling significantly dysregulated with a 2-fold or greater change within CCM3 knockdown cells, including genes that lead to an activation of p53 signaling and inhibition of VEGF activity. We selected a few genes of interest and conducted an independent quantitative
RT-PCR validation. Our RT-PCR results however confirmed the prediction for half of the genes tested while the rest were in opposition to the microarray. Despite the discrepancy seen between the two techniques, this microarray screen is an important beginning step to understand broadly what signaling messages are disrupted in CCM3 phenotypes.
101 Chapter 3: Differential Gene Expression Profiling
MATERIALS AND METHODS shRNA-mediated knockdown
Control and CCM3 lentiviral shRNA constructs in the pLKO.1-puro vector were obtained
from Sigma-Aldrich (Table S3.1). The CCM3 shRNA target 1 (Table S3.1) was used
throughout this study. The control lentiviral shRNA construct is an empty vector that
underwent the same viral production process. The lentivirus production was described
previously [18] and shown to effectively reduce the expression of CCM3 in HAEC cells
(Fig 3.1).
Cell Culture
Human aortic endothelia cells (HAEC, Life Sciences) were grown in Media 200
supplemented with 5% LSGS according to manufacturer’s instructions at 37 °C in 5%
CO2. For all experiments, cells were used between passage number 3- 6. HAEC cells
were infected by adding 1ml lentivirus media to cells plated in a six-well dish. Twenty- four hours later, media were replaced with standard growth medium containing 2 μg/ml puromycin, and cells were maintained in this selective medium until experiments. The knockdown viruses were assayed for CCM3 expression by Western blotting (Figure 3.1).
For RNA experiments, biological triplicates of control and CCM3 shRNA expressing
HAEC cells were cultured in puromycin selection media for a minimum of a week and then plated onto 6-well dishes at 3 x 105 cells density. RNA extraction was performed 5 days after plating.
102 Chapter 3: Differential Gene Expression Profiling
Antibodies
For Western blotting the following antibodies were used: anti-PDCD10 for CCM3
(Sigma), anti-LKB1 (Cell Signaling), anti-β-actin (AC-15, Sigma) and anti-YSK1 for
Stk25 (Santa Cruz).
Immunoblotting
To examine protein levels in HAEC cells, cells stably expressing CCM3 shRNA or
control lentiviruses were seeded on Attachment Factor (Invitrogen) coated six-well dishes
at 3 x 105 cells/cm2 and grown until confluent in puromycin selection media (2ug/ml).
Homogenates were prepared by rinsing cells with ice cold PBS prior to being lysed with
RIPA lysis buffer (137mM NaCl, 2.7mM KCl, 1% NP-40, 25mM Tris-HCl[pH 7.4],
2mM EDTA, 1% sodium deoxycolate, 0.1% sodium dodecyl sulfate (SDS), 50mM NaF,
0.1% 2-mercaptoethanol, protease inhibitor [complete mini, EDTA free; Roche, and
Phosphatase Inhibitor Cocktail I, [Sigma]). Lysates were incubated on ice, sonicated and cleared by centrifugation at 14,000 x g for 20 minutes.
Cell lysates were mixed with an equal volume of 2X SDS sample buffer; heat denatured at 100 °C for 5 minutes and resolved on 4-12% Bis-Tris gradient gels (Invitrogen). The proteins were transferred to PVDF membranes (Thermo Scientific), blocked with 5% milk in Tris-buffered saline containing 0.1% Tween 20 (Sigma), incubated with the indicated primary antibodies followed by horseradish peroxidase-conjugated secondary
103 Chapter 3: Differential Gene Expression Profiling
antibodies, and visualized with X-ray film using ECL chemiluminescence substrate
(Thermo Scientific).
RNA extraction and cDNA synthesis
RNA was harvested from cells using RNEasy lysis buffer (Qiagen) and Qiashredder columns (Qiagen) according to the manufacturer’s instructions. RNA was then purified following the RNeasy Mini Protocol Kit (Qiagen). cDNA was synthesized using equal amounts of purified RNA using QuantiTect Reverse Transcription Kit (Qiagen) following manufacturer’s instructions.
Primer Design
PCR primer sequences were designed using Primer3 and NCBI primer-BLAST or sequences obtained from Primer Bank (Harvard University). All primers used were HPSF purified and purchased from IDT Technologies. A complete list of primer sequences used in quantitative RT-PCR experiments is listed in Table S3.2 and Table S3.3.
Microarray Analysis:
For microarray analysis, biological triplicates of HAEC cells stably expressing control or
CCM3 shRNA lentivirus were plated at equal densities on Attachment Factor
(Invitrogen) precoated dishes. RNA was harvested from cells 5 days after plating using
RNeasy Mini Kit (Qiagen) according to manufacturer’s protocol. Samples were then processed for microarray analysis using Human GeneChip 2.0 ST Array (Affymetrix) at
104 Chapter 3: Differential Gene Expression Profiling
the SUNY Molecular Analysis Core (SUNYMAC), Upstate Medical University,
Syracuse NY. Briefly, RNA quality and concentration was determined using the RNA
Nanochips on an Agilent Technologies 2100 Bioanalyzer (Santa Clara, CA). Equal
amounts of RNA from each sample in a group were used in the microarray screen according to the Affymetrix GeneChip Expression Analysis Technical Manual.
Hybridization and washing steps were performed in accordance with the manufacturer’s instructions of the Affymetrix GeneChip Hybridization, Wash and Stain Kit. The array was scanned using an Affymetrix GeneChip Scanner 7G Plus and the resulting images analyzed with the Affymetrix GeneChip Operating System to convert the image signal intensities into expression level estimates. Fold changes in expression between groups were determined and normalized in a base log2 scale. Over 38,000 genes were identified
which were then filtered based on statistical significance (p < 0.05) and threshold of fold
change (≥ 50%). We used a filtered set of ~800 genes that were significantly altered
between the groups, using a False Discovery Rate (FDR) adjusted p-value from a
Student’s test < 0.05 and greater than a 50% change between control and CCM3 shRNA
for all our analyses.
Bioinformatic Pathway Analysis
For a broader probe on functional gene networks and pathways that may be altered by the
depletion of CCM3, we used the Analysis tools in Ingenuity Pathway Analysis (IPA)
software (Qiagen). We used the Core Analysis workflow of IPA to identify enriched gene
networks considering only relationships that were experimentally observed. We used a
filtered set of our data (FDR adjusted p-Value < 0.05 and Fold Change > [1.5-fold]) with
105 Chapter 3: Differential Gene Expression Profiling
~800 genes to run these analyses. We mapped genes within the well-annotated biological
function and Upstream-Regulated gene networks in the network query. For each
comparison, we utilized the top molecular and cellular functions identified and these
were overlaid with the complete normalized expression value for each comparison.
Quantitative reverse transcription –polymerase chain reaction (qRT-PCR)
qRT-PCR was used to perform independent validation experiments on a selection of
genes from all of the samples in each group. RNA was harvested as described above and
cDNA was synthesized using QuantiTect Reverse Transcription Kit (Qiagen) following
the manufacturer’s protocol. qRT-PCR was then carried out using Roche SYBR Green
Master Mix on a LightCycler 480 instrument (Roche Appliances). Each cDNA sample
was run in triplicate for each of the gene targets and 3 reference genes. For each, these
reagents were used: 10 µl SYBRGreen Master Mix, 2 µl of each 5 µM forward primer
and reverse primer, 4 µl water and 2 µl of cDNA. The qPCR reactions used the following program: 10 minutes at 95oC, 40 cycles of: 10 seconds at 95oC, 20 seconds at 58oC, 30
seconds at 72oC, ending with a melt curve of 5 seconds at 95oC, 1 minute at 65oC and
97oC. The three reference genes used were Hypoxanthine-guanine
phosphoribosyltransferase (HPRT) 60S acidic ribosomal protein P1 (RPLP1) and
Tyrosine 3-Monooxygenase/Tryptophan 5-Monooxygenase Activation Protein, Zeta
(YWHAZ) (Table S2).
Statistical Analysis
106 Chapter 3: Differential Gene Expression Profiling
qRT-PCR Data: For analysis of differential expression, RNA from for each group
(control vs. CCM3shRNA) were used in three biological replicates in qRT-PCR assays for each gene. The data were normalized to the calculated geometric mean of the three reference genes and the relative expression was determined from the obtained Ct values with the 2−ΔΔCt method [19]. Effects for individual genes were determined using a
Student’s T-test. The mean relative expression was expressed as mean +/- standard deviation of 3 independent experiments.
Correlation: A Pearson correlation was calculated for the control and CCM3 shRNA expressing groups between the microarray and qRT-PCR data. The significance of the correlations was calculated using a Fischer’s R to Z transformation.
Clustering: We combined all the significant gene transcripts with p < 0.05 and 50% change in expression levels for hierarchical clustering. This was performed using
MetaboAnalyst Software (http://www.metaboanalyst.ca; McGill University). Data were not renormalized when reloaded. Clusters were created using the average linkage and the
Pearson distance metric.
107 Chapter 3: Differential Gene Expression Profiling
RESULTS LKB1 and STK25 protein levels are dependent on CCM3 in endothelial cells.
We have shown that CCM3 is required for LKB1 protein stability and regulates
LKB1-STRADα regulated epithelial cell polarity (Chapter 2). With a novel interaction with STRADα and an association with the conserved LKB1 signaling pathway in epithelial cells, we wanted to determine whether LKB1 and STRAD were involved in
CCM3 signaling in endothelial cells. To determine whether CCM3 regulates LKB1 and
STRAD in endothelial cells, we used HAEC; a cell line with intact CCM3 expression and
CCM signaling pathways. We surveyed for protein expression of LKB1, STRAD and
STK25. Using Western blot analysis, we found that LKB1 and STK25 proteins were expressed in HAEC cells (Figure 3.1A). Using the best commercially available antibody for STRAD, we could barely detect STRAD protein expression in HAEC cells (Figure
3.1C) suggesting STRAD protein is expressed in these cells.
The protein expression of LKB1 is dependent on CCM3 (Chapter 2, Figure 2.4) in epithelial cells where CCM3 interacts with STRADα and stabilizes LKB1 expression.
Thus we investigated whether this was true in endothelial cells. We infected HAEC cells with control and CCM3 shRNA lentivirus previously shown to reduce the expression of
CCM3 (Chapter 2 Figure S2.3). We find that protein levels of LKB1 and STK25 are significantly reduced in CCM3 shRNA expressing cells (Figure 3.1A, B). Even though we could detect STRAD in HAEC cells, we were unable to quantify this data due to the
108 Chapter 3: Differential Gene Expression Profiling
low level of detection (Figure 3.1C). Together, these data suggest that the protein levels
of LKB1 and STK25 are dependent on CCM3 expression in endothelial cells.
LKB1 and STRAD transcript expressions are upregulated when CCM3 is depleted in
endothelial cells.
CCM3 stabilizes and regulates the expression of proteins it interacts with [6, 10,
20, 21]. We sought to investigate whether CCM3 also regulates transcript expression of
LKB1 and STRAD given previous data suggesting it is necessary for the protein
expression of LKB1 and STK25 (Figure 3.1A, B). We used HAEC cells with depleted
CCM3 protein levels and confirmed reduction in mRNA expression (Figure 3.2) and
compared the genome wide gene expression profiles of CCM3-deficient cells to control
cells in a microarray. Using stringent criteria for differentially expressed, we considered
genes to be differentially expressed only when transcript expression is statistically
significant (p < 0.05) AND greater than 2- fold change between control and CCM3
shRNA cells. We found that depletion of CCM3 does not reduce the mRNA expression
of LKB1, STRAD or STK25 gene transcripts as determined by microarray screen (Table
3.1). It should be noted that LKB1 gene expression was statistically significant (p =
0.035) but the experimental fold change in gene expression of 26% higher in
CCM3shRNA when compared to control did not meet our criterion of at least 2-fold
change (Table 3.1). We further looked at some downstream effectors of LKB1 to
determine whether their transcript expression was downregulated by absence of CCM3 as hypothesized. We looked at the isoforms of MARK1 and AMPK since they play a significant role in LKB1-mediated polarity and we had previously shown to be dependent
109 Chapter 3: Differential Gene Expression Profiling on CCM3 expression (Chapter 2, Figure 2.5). We found that neither MARK1 nor AMPK transcript expressions were significantly altered in CCM3 depleted cells when compared to controls (Table 3.1).
To independently validate the microarray predicted changes, we performed quantitative RT-PCR in individual RNA samples for LKB1, STRAD and STK25 (Figure
3.2). The mRNA expression of LKB1 and STRAD were significantly higher in CCM3 shRNA expressing cells than in controls (Figure 3.2). There was a slight increase in
STK25 expression, however it was not statistically significant (Figure 3.2). As mentioned earlier, we observed a slight increase in LKB1 transcript expression in our microarray screen but this was not considered significant based on our stringent criteria, however the
RT-PCR results suggest there is a significant increase in the transcript expression of
LKB1 and STRAD. Interestingly, this increase in LKB1 and STRAD gene expression in
CCM3 shRNA cells is opposite of what we observed at the protein level (Figure 3.1) suggesting that CCM3 may have several mechanisms by which it regulates LKB1 and
STRAD.
Microarray analysis of CCM3 knockdown endothelial cells reveal global downregulation of cell adhesion molecules and an upregulation of cell signaling molecules.
To determine what other signaling pathways might be affected by the depletion of
CCM3 in endothelial cells, we looked at the top differentially expressed molecules from our microarray screen. Table 3.2 shows a selection of the most highly differentially
110 Chapter 3: Differential Gene Expression Profiling expressed genes in CCM3 knockdown cells compared to control cells. We noticed that a significant number of the downregulated genes (green) are cell adhesion proteins and extracellular matrix proteins with functions in maintaining cell integrity (Table 3.2).
Examples include protocadherin beta 13 (PCDHB13), carbohydrate N- acetylgalactosamine-4-sulfate 6-O sulfotransferase 15 (CHST15) and cell adhesion molecule 3 (CADM3) with significant fold changes of -19.19, -8.53 and -7.87 respectively (Table 3.2). In contrast, most of the downregulated genes (red) are enzymes, transcription factors or cytokines involved in diverse cell signaling including cell growth.
Examples include insulin like growth factor binding protein 5 (IGFBP5), Ras-related
GTP binding D (RRAGD) and DNA damage inducible transcript 3 (DDIT3) with significant fold changes of 12.30, 3.73 and 2.54 respectively (Table 3.2). This data suggests that lack of CCM3 in endothelial cells impairs a diverse group of signaling molecules with roles in cell adhesion and maintaining the integrity of endothelial cells.
Interestingly, a common feature of CCMs is a disorganization of the endothelial lining in cerebrovasculature and it has been suggested that this is as a results of dysfunction in tight junction proteins (refs).
To independently validate some of these microarray predictions of the top differentially expressed genes, we performed qRT-PCR on a selected number of genes
(Figure 3.3). About 60% of the genes we tested were validated in the RT-PCR were in the direction and degree of fold change predicted by the array. These genes were either significantly increased or decreased in CCM3 shRNA expressing cells in agreement with the microarray predictions (Figure 3.3A). These included the transporter fatty acid
111 Chapter 3: Differential Gene Expression Profiling
binding protein 4 (FABP4), the extracellular matrix protein periostin osteoblast specific
factor (POSTN), IGFBP5 and RRAGD (Figure 3.3A).
However, not all of the genes tested using qRT-PCR were validated as predicted
by the microarray. In fact, about 40% of these genes had no significant mRNA gene
expressions (Figure 3.3B). Even though the changes were not significant, the mRNA
expression levels trended in the opposite direction of predictions from the microarray. To
determine the correlation between the fold changes observed in methods, we calculated
the correlation between these two techniques. This coefficient helps us determine whether
the microarray predictions were correlated to the qRT-PCR fold changes (Figure 3.3C).
Looking at the fold changes of all the genes tested in the qRT-PCR, there was a significant correlation of the qRT-PCR fold changes to those of the microarray (R2 =
0.8009; p-value (2 tailed) = 0.0168) (Figure 3.3C). This data suggests that the two techniques are highly correlated (R2= 0.8) and the differences we observed could be due
to differences in the methods. Despite the hit rate of validated genes, the microarray
screen serves as a beginning step to understand broadly what signaling messages are
disrupted in CCM3 phenotypes.
Pathway analysis of differentially altered upstream molecules and biological functions
reveal cell death & survival, cell growth & proliferation and cell movement as the most
significantly affected in CCM3 deficient cells.
To evaluate functional gene networks and pathways that may be altered by the
depletion of CCM3, we used Qiagen’s Ingenuity Pathway Analysis (IPA) software to
further analyze the microarray data. To understand the gene networks of highly regulated
112 Chapter 3: Differential Gene Expression Profiling
genes, we performed a systems level analysis where we mapped out the top differentially
affected gene networks using IPA software. This analysis highlighted several upstream-
regulated gene networks and biological function networks that were significantly altered
in our microarray data from CCM3 depletion in endothelial cells when compared to
control cells (Figure 3.4, Table 3.3). Not surprising, Cell Growth & Proliferation, Cell
Death & Survival and Cellular Movement were among the top significantly
downregulated biological functions (Table 3.3), suggesting that lack of CCM3 induces
gene expression changes that affect these functions in endothelial cells.
Furthermore, we explored upstream-regulated gene networks and predictions of
how knockdown of CCM3 affected these networks (Figure 3.4) using the upstream
regulator analysis tool in IPA. This tool identifies upstream regulators that are predicted
to be activated or inhibited based on an activation z-score (a z-score algorithm produces
either a prediction of activation or inhibition (or no prediction) based on significant
expression values) and is used to identify molecules of interest for future study.
Inspection of these upstream networks revealed a few key molecules that were at the
intersection of the common gene targets and networks affected by lack of CCM3. These
included p53, VEGF and NFκB1 (Figure 3.4), which are all key mediators of cell proliferation, survival and apoptosis, and which were either generally activated or inhibited. The upstream regulated gene networks included the activation of CDKN2A-
regulated genes, KLF2-regulated genes and VEGF-regulated genes (Figure 3.4A). In
addition, NFκB1-regulated genes and HIF1A-regulated genes were significantly inhibited
(Figure 3.4B). The analysis revealed that these upstream regulators are predicted to be
113 Chapter 3: Differential Gene Expression Profiling affected in endothelial cells lacking CCM3 based on differentially regulated genes in our experiment. Besides canonical VEGF gene network, these upstream regulated gene networks are novel implications for CCM3 functions.
Depletion of CCM3 in endothelial cells leads to an activation of p53 signaling
During our upstream network analysis, we noticed a robust number of genes within the p53 signaling pathway that were significantly downregulated in our microarray screen (Table 3.4) because of the activation of p53. In fact, TP53 was one of the top differentially expressed upstream molecules, with a significant fold change increase in its expression. As a result, canonical p53 signaling pathway was significantly altered with a robust decrease in the expression of genes in p53 signaling pathway (Table 3.4).
We identified over 100 transcripts in the p53 signaling pathway that were differentially expressed (≥2-fold change; p <0.05) in the absence of CCM3 in our data set
(Table 3.4, showing a subset). Many of these genes are transcriptional regulators that regulate apoptosis and DNA repair including DNA damage inducible transcript 3
(DDIT3), forkhead box O4 (FOXO4) and hypoxia inducible factor 1 (HIF1). We independently examined 12 of these genes using quantitative RT-PCR to validate predictions from the array (Figure 3.5A). Of the 12 transcripts examined, eight were significantly altered as predicted by the array (Figure 3.5A), many of which were significantly decreased in expression. A few of the transcripts we validated did not follow the changes predicted by the microarray (Figure 3.5B) which could be due to the difference in the methods and use of different probes versus primers. This data suggests
114 Chapter 3: Differential Gene Expression Profiling that depletion of CCM3 in endothelial cells activates p53 signaling pathway (and mRNA expression of p53) which results in the down-regulation of downstream signaling molecules resulting in modulation of cell growth, apoptosis and cell survival.
DISCUSSION To understand the effects of the newly identified LKB1-STRAD-CCM3 complex in endothelial cell signaling, we depleted the expression of CCM3 (protein and mRNA) in human aortic endothelial cells. In endothelial cells, LKB1, STK25 and STRAD protein expression levels are dependent on CCM3, as we had previously shown in epithelial cells
(Figures 3.1, 2.4). With quantitative RT-PCR, we unexpectedly find that transcript expressions of LKB1 and STRAD are significantly increased in CCM3 shRNA expressing cells (Figure 3.2). Further, we performed gene expression profiling using
DNA microarray and discovered that endothelial cells lacking CCM3 expression drastically lose their expression of cell adhesion molecules (Table 3.2) and have dysregulated cell adhesion signaling. Using a functional genome wide analysis, we show that p53 signaling activity in endothelial cells lacking CCM3 is significantly upregulated as compared to control. As a result of this activation, several downstream effectors of p53 are significantly downregulated in expression that leads to a dysregulation of cell growth, cell survival and apoptotic functions (Table 3.4, Figure 3.5). These novel findings provide an important first step to understanding the molecular mechanisms of CCM3 phenotype in endothelial cells and may contribute to the understanding of signaling pathways involved in CCM disease pathogenesis.
115 Chapter 3: Differential Gene Expression Profiling
The identification of a role for CCM3 in LKB1-mediated epithelial polarity and
the regulation of LKB1 and STRAD protein levels, places CCM3 as a key regulator of
the LKB1-STRAD pathway. Further, given the functions of CCM3 in endothelial cell and
CCM pathology, understanding a role for LKB1-STRAD in endothelial cell signaling
will be useful in CCM disease research. Our data suggests that CCM3 regulates the
expressions of LKB1 and STRAD in endothelial cells, both at the protein level and the
mRNA transcript level. At the protein level, CCM3 is necessary for LKB1 and STRAD
expression as we see a significant decrease in LKB1 levels when CCM3 is depleted in endothelial cells (Figure 3.1). However, using microarray followed by quantitative PCR in these cells, we find that the absence of CCM3 increases significantly the mRNA expression of LKB1 and STRAD, suggesting that CCM3 mRNA might play a role in the
regulation of transcription of these genes. It is undetermined why LKB1 transcript levels
are increased in endothelial cells when CCM3 is depleted despite a reduction in protein
levels. It is possible LKB1 and STRAD are under the regulation of a feedback loop where
in at low levels of LKB1 and STRAD protein, the cell is stimulated to increase
transcription of these genes which results in higher transcript messenger available for
translation. However, this pattern of regulation of LKB1 has also been reported. Veleva-
Rotse et al. (2014) recently reported an increase in LKB1 mRNA and a decrease in
protein expressions in STRADα null mice [22]. Further studies are needed to elucidate
the role of LKB1 and STRAD in CCM3-phenotype in endothelial cells and understand
the mechanisms by which CCM3 regulates the protein levels and mRNA expression of
these genes in endothelial cells.
116 Chapter 3: Differential Gene Expression Profiling
In our microarray and qRT-PCR studies we experienced a reproduction rate of
about 60% as not all of the microarray change predictions were validated using qPCR
(Figures 3.3, 3.5). There are several reasons why these two methods might not align and
one of these involves the differences in the probe and primers used. Many genes in the
human genome have a wide variety of transcript variants, which results from alternative
exon splicing and these different transcripts can function differently in human cells [23].
This could explain why not all of our microarray predictions were validated using qRT-
PCR as it possible the two techniques might be measuring at different locations in a transcript. It would be useful to extract exon level data from the microarray screen and design primers to specific regions of the transcript that were detected in the array in order to avoid the possibility of splice variants. The primers used in these studies were not designed as such potentially reading at different locations than the array probe. In addition, using transcriptome profiling with RNA-seq will be a useful step forward since
RNA-seq has greater sensitivity, range and accuracy than microarray. Others have observed that methodologically, RNA-seq identifies a much larger number of differentially expressed genes with much better correlation to qRT-PCR results than the
microarray [24]. The two assays would potentially generate different but complementary hypotheses for further evaluation. Despite these inconsistencies between methods, our microarray study serves as a beginning point for understanding broadly, what signaling messages are disrupted in CCM3 phenotype.
The signaling mechanisms involved in CCM are beginning to unravel as the
structure of the proteins, CCM1, CCM2 and CCM3, responsible for the familial form are
117 Chapter 3: Differential Gene Expression Profiling being elucidated [7, 21, 25]. However, given their unique protein structures and diverse binding partners, it has been difficult to determine a key signaling pathway that is dysregulated and by which the endothelial vascular phenotypes of CCM disease arise.
CCM3 causes a more severe phenotype and interacts with several protein targets beyond the CCM1-CCM2-CCM3 protein complex and as such is believed to play a unique role in
CCM pathology [5]. Understanding the broad interactions of CCM3 will shed light on the disease pathology. In our study, we identified several novel genes and transcription regulators including HIF1A, FOXM1 and CDKN2A that are significantly altered in expression by lack of CCM3 in endothelial cells. This vast number of genes and networks will need to be further characterized to understand how CCM3 is affecting some of these pathways and specifically focusing on disease phenotypes such as dilated and leaky vasculatures and tight junction formation that are more severe in patients with CCM3 mutations [26].
In fact, one of the observed features of CCM is the loss of vascular integrity in the endothelial lining [27], leaving the vessels thin and leading to the formations of lesions and leakiness. It is postulated that the CCM proteins (CCM1, 2 and 3) regulate endothelial cell integrity by interacting directly or indirectly with adhesion junction stability proteins such as VE-cadherin, HEG1, integrin-β1 and VEGFR2 to connect extracellular cues to intracellular signaling pathways [28]. In our microarray study, we find that a significant number of cell adhesion molecule genes were downregulated in the absence of CCM3 (Table 3.3) including protocadherin beta 13 (PCDH13) and cell adhesion molecule 3 (CADM3). Interestingly, we also observed an increase in the transcript expression of extracellular matrix signaling molecules (Table 3.4). Loss of
118 Chapter 3: Differential Gene Expression Profiling adhesion molecules when CCM3 is absent in endothelial cells suggests CCM3 functions to maintain the integrity of endothelial linings and loss of CCM3 in CCM disease could be responsible for the loss of vascular integrity that is one of the hallmark of the disease.
The loss of adhesion molecules at tight junctions in endothelial cells that are null for
CCM3 could lead to aggregation of cells and formation of lesions.
The cell stress sensor and tumor suppressor, p53, plays a key role in cell stress response, cell survival and tumorigenesis. A previous link to CCM pathology suggest a potential role for p53 in CCM biology (Plummer et al 2004), however it remains undetermined what role p53 may play in CCMs. Loss of p53 sensitizes mice with a mutation in CCM1 and the formation of vascular lesions in the brain is significantly higher [17]. It is possible that p53 plays a direct role in the formation of vascular lesions in CCM. We observed a significant increase in p53 activation (increased activity and increased mRNA expression) in endothelial cells deficient in CCM3 when compared to control cells (Table 3.5, Figure 3.5). This dysregulation of p53 signaling pathway in
CCM3 phenotype suggests a role for p53 in CCM.
Recently, patients with a heterozygous CCM3 mutation have been shown to be at a high risk of developing meningioma [5, 29]. This finding suggests a potential link between CCM3 and tumor biology, hinting at a role for CCM3 as a tumor suppressor.
The mechanism for this effect has not been proposed and it would be interesting to determine whether it is in association with p53, a well-characterized tumor suppressor.
We show that absence of CCM3 downregulates the expression of hundreds of genes that result from the activation of p53 signaling (Table 4, Fig. 5) indicating a widespread effect
119 Chapter 3: Differential Gene Expression Profiling of CCM3 on p53 activity. It is possible that CCM3 has tumor suppressing activity through its association with the p53 pathway and it would be worthwhile to design future studies focused on answering this question.
To the best of our knowledge, this is the first report of gene expression changes in
CCM3 depleted endothelial cells. Data from our microarray screen broadens our understanding of the mechanisms by which CCM3 deficiency might result in CCM disease and this opens up new avenues of research in both molecular functions of CCM3 and CCM disease pathology. Large genome screens provide a wealth of information and are an initial step in identifying new molecular pathways and biological functions altered in diseases. A screen of such depth would be beneficial using CCM patient samples or microvasculature cells from CCM animal disease models such as endothelial cells from the CCM3 endothelial-specific conditional mouse knockout [6, 30] to gain insights into the molecular mechanisms perturbed by CCM3 inactivation.
120 Chapter 3: Differential Gene Expression Profiling
REFERENCES
1. Dibble, C.F., et al., Defining the functional domain of programmed cell death 10
through its interactions with phosphatidylinositol-3,4,5-trisphosphate. PLoS One, 2010.
5(7): p. e11740.
2. Bergametti, F., et al., Mutations within the programmed cell death 10 gene cause cerebral cavernous malformations. Am J Hum Genet, 2005. 76(1): p. 42-51.
3. Riant, F., et al., Recent insights into cerebral cavernous malformations: the molecular genetics of CCM. FEBS J, 2010. 277(5): p. 1070-5.
4. Denier, C., et al., Genotype-phenotype correlations in cerebral cavernous malformations patients. Ann Neurol, 2006. 60(5): p. 550-6.
5. Riant, F., et al., CCM3 Mutations Are Associated with Early-Onset Cerebral
Hemorrhage and Multiple Meningiomas. Mol Syndromol, 2013. 4(4): p. 165-72.
6. He, Y., et al., Stabilization of VEGFR2 signaling by cerebral cavernous malformation 3 is critical for vascular development. Sci Signal, 2010. 3(116): p. ra26.
7. Li, X., et al., Crystal structure of CCM3, a cerebral cavernous malformation
protein critical for vascular integrity. J Biol Chem, 2010. 285(31): p. 24099-107.
8. Zhang, M., et al., Structural mechanism of CCM3 heterodimerization with
GCKIII kinases. Structure, 2013. 21(4): p. 680-8.
9. Voss, K., et al., Functional analyses of human and zebrafish 18-amino acid in- frame deletion pave the way for domain mapping of the cerebral cavernous malformation
3 protein. Hum Mutat, 2009. 30(6): p. 1003-11.
10. Fidalgo, M., et al., CCM3/PDCD10 stabilizes GCKIII proteins to promote Golgi
assembly and cell orientation. J Cell Sci, 2010. 123(Pt 8): p. 1274-84.
121 Chapter 3: Differential Gene Expression Profiling
11. Lin, C., et al., PDCD10/CCM3 acts downstream of {gamma}-protocadherins to
regulate neuronal survival. J Biol Chem, 2010. 285(53): p. 41675-85.
12. Kean, M.J., et al., Structure-function analysis of core STRIPAK Proteins: a
signaling complex implicated in Golgi polarization. J Biol Chem, 2011. 286(28): p.
25065-75.
13. Goudreault, M., et al., A PP2A phosphatase high density interaction network
identifies a novel striatin-interacting phosphatase and kinase complex linked to the
cerebral cavernous malformation 3 (CCM3) protein. Mol Cell Proteomics, 2009. 8(1): p.
157-71.
14. Voss, K., et al., CCM3 interacts with CCM2 indicating common pathogenesis for cerebral cavernous malformations. Neurogenetics, 2007. 8(4): p. 249-56.
15. Guerrero, A., et al., The cerebral cavernous malformation 3 gene is necessary for senescence induction. Aging Cell, 2015. 14(2): p. 274-83.
16. Campisi, J. and F. d'Adda di Fagagna, Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol, 2007. 8(9): p. 729-40.
17. Plummer, N.W., et al., Loss of p53 sensitizes mice with a mutation in Ccm1
(KRIT1) to development of cerebral vascular malformations. Am J Pathol, 2004. 165(5):
p. 1509-18.
18. Matsuki, T., et al., Reelin and stk25 have opposing roles in neuronal polarization
and dendritic Golgi deployment. Cell, 2010. 143(5): p. 826-36.
19. Livak, K.J. and T.D. Schmittgen, Analysis of relative gene expression data using
real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 2001. 25(4):
p. 402-8.
122 Chapter 3: Differential Gene Expression Profiling
20. Zhang, Y., et al., A network of interactions enables CCM3 and STK24 to coordinate UNC13D-driven vesicle exocytosis in neutrophils. Dev Cell, 2013. 27(2): p.
215-26.
21. Zhang, R., X. Li, and T.J. Boggon, Structural analysis of the KRIT1 ankyrin repeat and FERM domains reveals a conformationally stable ARD-FERM interface. J
Struct Biol, 2015. 192(3): p. 449-56.
22. Veleva-Rotse, B.O., et al., STRAD pseudokinases regulate axogenesis and LKB1 stability. Neural Dev, 2014. 9: p. 5.
23. Kim, D.S. and Y. Hahn, Identification of human-specific transcript variants induced by DNA insertions in the human genome. Bioinformatics, 2011. 27(1): p. 14-21.
24. Lahiry, P., et al., Transcriptional profiling of endocrine cerebro-osteodysplasia using microarray and next-generation sequencing. PLoS One, 2011. 6(9): p. e25400.
25. Fisher, O.S., et al., Structural studies of cerebral cavernous malformations 2
(CCM2) reveal a folded helical domain at its C-terminus. FEBS Lett, 2013. 587(3): p.
272-7.
26. Riant, F., et al., CCM molecular screening in a diagnosis context: novel unclassified variants leading to abnormal splicing and importance of large deletions.
Neurogenetics, 2013. 14(2): p. 133-41.
27. Burkhardt, J.K., et al., Upregulation of transmembrane endothelial junction proteins in human cerebral cavernous malformations. Neurosurg Focus, 2010. 29(3): p.
E3.
28. Storkebaum, E., et al., Cerebrovascular disorders: molecular insights and therapeutic opportunities. Nat Neurosci, 2011. 14(11): p. 1390-7.
123 Chapter 3: Differential Gene Expression Profiling
29. Labauge, P., et al., Multiple dural lesions mimicking meningiomas in patients with
CCM3/PDCD10 mutations. Neurology, 2009. 72(23): p. 2044-6.
30. Zhou, Z., et al., Cerebral cavernous malformations arise from endothelial gain of
MEKK3-KLF2/4 signalling. Nature, 2016. 532(7597): p. 122-6.
124 Chapter 3: Differential Gene Expression Profiling
TABLES
125 Chapter 3: Differential Gene Expression Profiling
Table 3.1. Only modest changes in gene expression of LKB1, STRAD and
downstream effectors in endothelial cells. Table shows the gene expression changes of
LKB1, STRAD and downstream effectors in fold changes and the associated p-value.
Our stringent definition of significant differential expression indicates that gene expression needs to be statistically significant (p < 0.05) AND greater than 50% (1.5 fold) change between control and CCM3 shRNA cells. This indicates no significant change in LKB1, STRAD and downstream effector transcript expressions.
Gene Symbol p-Value Fold Change
STK11 (LKB1) 0.0356563 1.26
STRADA 0.817021 1.01
STRADB 0.196593 1.11
STK25 0.147448 -1.08
MARK1 0.301855 -1.13
MARK2 0.709135 1.04
MARK3 0.542794 1.05
MARK4 0.0198508 1.10
PRKAA2 (AMPK) 0.0565388 1.29
PRKAA1 (AMPK) 0.177629 -1.08
126 Chapter 3: Differential Gene Expression Profiling
Table 3.2. A selection of highly differentially expressed genes in CCM3 knockdown cells. Table contains genes and molecules significantly altered (p <0.05) with a fold change of 50% or higher (>2- fold change) by CCM3 depletion. The relative expression levels (in fold change) and p-value were calculated from microarray analysis data. The genes were classified into functional categories using Ingenuity IPA Core analysis
(Qiagen). The abbreviated names of genes are from GenBank database. Green represents genes that were upregulated and Red those that were downregulated.
127 Chapter 3: Differential Gene Expression Profiling
Gene ID Gene Name Gene Category/ p-value Fold Symbol Functions Change 17078592 fatty acid binding protein 4, FABP4 Transporter 2.13E-06 -51.53 adipocyte 17083433 interleukin 33 IL33 Cytokine 4.32E-09 -33.51 periostin, osteoblast specific POSTN ECM Protein 2.23E-06 -30.18 16778241 factor PCDHB1 Cell Adhesion 4.31E-05 -19.19 16990312 protocadherin beta 13 3 Protein 17074259 angiopoietin 2 ANGPT2 Growth factor 7.44E-06 -15.60 PCDHB1 Cell Adhesion 7.97E-05 -13.09 16990331 protocadherin beta 15 5 Protein sulfotransferase family 1E SULT1E1 Enzyme 6.39E-05 -11.01 16976615 member 1 carbohydrate (N- CHST15 Enzyme 9.62E-06 -8.53 acetylgalactosamine 4- sulfate 6-O) 16719217 sulfotransferase 15 LIPG Enzyme 0.000102 -8.52 16852296 lipase, endothelial 964 RAS guanyl releasing RASGRP3 other 0.000239 -7.89 protein 3 (calcium and 212 16878994 DAG-regulated) 16672412 cell adhesion molecule 3 CADM3 other 3.56E-05 -7.87 integral membrane protein ITM2A other 1.58E-05 -7.81 17112364 2A cytochrome P450 family 1 CYP1A1 enzyme 1.55E-05 -7.74 16811684 subfamily A member 1 cerebellar degeneration CDR1 other 0.000180 -7.55 17114701 related protein 1 582 16916901 Ras association RASSF2 kinase 5.57E-06 -7.36 (RalGDS/AF-6) domain family member 2 16990257 protocadherin beta 2 PCDHB2 other 8.44E-05 -7.36 16698801 CD34 other 0.000179 -6.99 CD34 molecule 917 16743707 matrix metallopeptidase 10 MMP10 peptidase 3.65E-05 -6.82 16990261 protocadherin beta 3 PCDHB3 other 7.82E-05 -6.78 16990267 protocadherin beta 5 PCDHB5 Other 1.85E-06 -6.50
16908197 insulin like growth factor IGFBP5 other 3.66E-06 12.30 binding protein 5 17115782 chloride intracellular CLIC2 enzyme 1.24E-06 10.16 channel 2 16696416 tumor necrosis factor TNFSF18 cytokine 4.43E-05 9.48 superfamily member 18
16929562 heme oxygenase 1 HMOX1 enzyme 3.00E-06 5.42 16949264 NmrA-like family domain LOC3448 other 1.86E-05 4.97 containing 1 pseudogene 87 16896561 cytochrome P450 family 1 CYP1B1 enzyme 8.01E-05 4.52 subfamily B member 1
128 Chapter 3: Differential Gene Expression Profiling
17021596 Ras-related GTP binding D RRAGD enzyme 0.000222 3.73 99 16726081 retinoic acid receptor RARRES3 enzyme 6.15E-05 3.29 responder (tazarotene induced) 3 16766578 DNA damage inducible DDIT3 Transcription 3.74E-05 2.54 transcript 3 regulator
129 Chapter 3: Differential Gene Expression Profiling
Table 3.3. Top biological functions altered by depletion of CCM3 in endothelial
cells. Ingenuity IPA analysis of significantly changed gene transcripts in microarray show top functional groups affected by depletion of CCM3 in endothelial cells.
130 Chapter 3: Differential Gene Expression Profiling
Table 3.4. p53 signaling genes are differentially regulated in CCM3 knockdown cells. Table represents the top highly regulated genes in canonical p53 signaling pathway that were significantly altered (p <0.05) with a fold change of 50% or higher in CCM3 depleted cells. The relative expression levels (Fold Change) were calculated from microarray analysis data. Green represents genes that were upregulated and Red those that were downregulated. Bolded genes were selected for validations by qRT-PCR.
131 Chapter 3: Differential Gene Expression Profiling
Genes Exp Fold Change Gene ID 16908197 IGFBP5 12.300 16929562 HMOX1 5.420 16837938 ITGB4 3.290 16726081 RARRES3 3.290 16766578 DDIT3 2.540 16715793 KCNMA1 2.350 16990848 ADRB2 2.260 16679411 EXO1 -2.020 16935228 PDGFB -2.020 16799637 RAD51 -2.040 16914315 UBE2C -2.160 17000439 CDC25C -2.170 16745236 H2AFX -2.170 16944738 PDIA5 -2.220 16840982 MYH10 -2.280 16914791 SNAI1 -2.320 16713187 NRP1 -2.330 17011279 PRDM1 -2.330 16977868 ABCG2 -2.360 17104313 AR -2.410 16984689 ITGA2 -2.460 16877019 RRM2 -2.540 16965798 PCDH7 -2.550 16792954 NID2 -2.580 16957396 CCDC80 -2.610 16802519 KIF23 -2.610 16707551 CEP55 -2.670 16840902 AURKB -2.790 16989736 EGR1 -2.950 16671642 EFNA1 -3.010 16878947 LTBP1 -3.020 17048563 PEG10 -3.030 16986913 VCAN -3.100 17049904 LRRC17 -3.220 16919703 PLTP -3.250 16882975 NCAPH -3.270 16768270 KITLG -3.320 16961616 TNFSF10 -3.320 16967771 CXCL8 -3.350 17021217 ME1 -3.480 16913957 MYBL2 -3.490 16817017 PLK1 -3.520 16787135 FLRT2 -3.770 16692846 CTSK -3.830 16761012 A2M -3.890 16916901 RASSF2 -7.360 16811684 CYP1A1 -7.740 16778241 POSTN -30.180 17078592 FABP4 -51.530
132 Chapter 3: Differential Gene Expression Profiling
FIGURES
133 Chapter 3: Differential Gene Expression Profiling
Figure 3.1. LKB1, STRAD and STK25 protein expressions are dependent on CCM3 in endothelial cells. A) Representative Immunoblotting of HAEC cells infected with control vector (CTL shRNA) or CCM3 shRNA lentivirus (CCM3 shRNA) with anti-
LKB1, anti-STK25 and anti-CCM3 antibodies (indicated to the left of the panels). B)
Quantification of three independent experiments revealed that CCM3 knockdown significantly reduces the protein expression of LKB1 and STK25. (* Student’s T-test, p <
0.05). C) Immunoblotting of HAEC cells infected with control vector (CTL shRNA) or
CCM3 shRNA (CCM3 shRNA) with anti-STRAD antibody indicates a reduction in
STRAD protein expression levels.
134 Chapter 3: Differential Gene Expression Profiling
135 Chapter 3: Differential Gene Expression Profiling
Figure 3.2. Quantitative RT-PCR analysis of LKB1 and STRAD gene transcripts indicates an increase in expression when CCM3 is knockdown. Quantitative RT-PCR was performed in triplicates using cDNAs obtained from control shRNA or CCM3 shRNA expressing with custom primers (Table S3.3). LKB1 and STRAD mRNA expression are significantly increased in CCM3 shRNA expressing cells. STK25 increase but not significantly. Results expressed as mean +/- SD of 3 independent experiments. * p < 0.05 ** p< 0.005 ***p< 0.001
136 Chapter 3: Differential Gene Expression Profiling
137 Chapter 3: Differential Gene Expression Profiling
Figure 3.3. Quantitative RT-PCR confirms the microarray predictions of a subset of selected genes. Quantitative RT-PCR was performed in triplicates using cDNAs obtained from control shRNA or CCM3 shRNA expressing (Table S3.1) HAEC cells. Data was normalized to reference genes and the relative expression determined using the 2-ΔΔCt
method. CCM3 mRNA expression was significantly reduced in CCM3 shRNA
expressing cells. Results expressed as mean +/- SD of 3 independent experiments. * p <
0.05. A) Selected highly regulated genes were validated using qRT-PCR with custom primers for each gene (Table S3.2). Genes presented were confirmed to either increase or decrease in expression as predicted by the microarray. * indicates a p-value < 0.05, ** indicates a p-value <0.001, *** indicates a p-value p<0.0001. FABP4: fatty acid binding protein 4, POSTN: periostin, osteoblast specific factor, IGFBP5: insulin like growth factor binding protein 5, RRAGD: G-protein Ras-related GTP binding D. B) These selected highly regulated genes were not validated by qRT-PCR (Table S3.2). The mRNA expression of genes presented are not statistically significant and the gene expression trends are in opposition to predictions.
C) Correlation graph of the fold change of the selected genes from the microarray. The
fold-change values from the microarray data (x-axis) are plotted against the fold change
value from the qRT-PCR experiments (y-axis), with significant correlation (R2 =
0.800917947; p-value (2 tailed) = p = 0.0168)
138 Chapter 3: Differential Gene Expression Profiling
139 Chapter 3: Differential Gene Expression Profiling
Figure 3.4. Exploratory bioinformatics pathway analysis of top differential upstream-regulated gene networks. Filtered data from microarray analysis was used to evaluate comparative changes in gene expression in CCM3 shRNA expressing cells compared to control. Ingenuity IPA analysis of significantly changed genes found gene networks that were differentially altered in CCM3 depleted cells compared to control. A)
Represented are six top differential upstream gene networks that were activated or B) inhibited by the loss of CCM3. CDKN2A: Cyclin-Dependent Kinase Inhibitor 2A, KLF2:
Kruppel-Like Factor 2, VEGF: Vascular Endothelial Growth Factor, HIF1A: Hypoxia
Inducible Factor 1, Alpha Subunit, NFKB1: Nuclear Factor Of Kappa Light Polypeptide
Gene Enhancer In B-Cells 1, FOXM1: Forkhead Box M1.
140 Chapter 3: Differential Gene Expression Profiling
141 Chapter 3: Differential Gene Expression Profiling
Figure 3.5. Quantitative RT-PCR validation of highly regulated genes on the p53 signaling pathway. A, B). Quantitative RT-PCR was performed in triplicates using cDNAs obtained from control shRNA or CCM3 shRNA expressing HAEC cells using custom primers for each gene (Table S3). Data is normalized to reference genes and the relative expression determined using the 2-ΔΔCt method. All genes represented were
validated and matched the microarray results. Results expressed as mean +/- SD of 3 independent experiments. Panel A represents genes that were validated and panel B represents genes that were not validated. * p < 0.01 ; ** p<0.001; *** p<0.0001. TP53:
Tumor Protein 53, POSTN: Periostin Osteoblast Specific Factor, KIT: V-Kit Hardy-
Zuckerman 4 Feline Sarcoma Viral Oncogene Homolog, RRM2B: Ribonucleotide
Reductase M2 B (TP53 Inducible)), IGFBP5: (Insulin-Like Growth Factor Binding
Protein 5, CDKN1A: Cyclin-Dependent Kinase Inhibitor 1A.
142 Chapter 3: Differential Gene Expression Profiling
143 Chapter 3: Differential Gene Expression Profiling
SUPPLEMENTAL TABLES AND FIGURES
144 Chapter 3: Differential Gene Expression Profiling
Supplemental Table 3.1. List of lentivirus shRNA sequences used in this study. shRNA Species Sequences
Control (Empty Vector) Human No shRNA Insert
Target 1 Human CCGGGCTGATGATGTAGAAGAGTATCTCG
(TRCN0000122022)* AGATACTCTTCTACATCATCAGCTTTTTTG
Target 3 Human CCGGCGTAAGTGCCAACCGACTAATCTCG
(TRCN0000140317)* AGATTAGTCGGTTGGCACTTACGTTTTTTG
* : MISSION® pLKO.1-puro CCM3 shRNA Plasmid DNA (Sigma-Aldrich)
145 Chapter 3: Differential Gene Expression Profiling
Supplemental Table 3.2. List of human primer sequences used for qRT-PCR experiments.
Primer (human) Forward (5'-3') Reverse (5'-3') CCM3 AAATGTGTTCGTAAGTGCCAACC CCACAGTTTTGAAGGTCTGAAGT PCDHB5 CACAATCAGCCCAAATTCACAC GGGGCGTTGTCATTATTATCCAA PCDHB13 CACCCTACTAACGGAGAGACC GTCATTGACATCGGCGATCAG FABP4 ACTGGGCCAGGAATTTGACG CTCGTGGAAGTGACGCCTT RASGRP3 CTCTGCATGTATCGAAATGCCA CTACTTCCCGAAATTCCTCAGTC IGFBP5 ACCTGAGATGAGACAGGAGTC GTAGAATCCTTTGCGGTCACAA RRAGD CTAGCGGACTACGGAGACG ATGAGCAGGATTCTCGGCTTC CADM3 CAAGTGCCAAGTGAAAGATCACG CCCAAAGTAGAGAGTCTGCTGAG AMPK TTGAAACCTGAAAATGTCCTGCT GGTGAGCCACAACTTGTTCTT CCM3 AAATGTGTTCGTAAGTGCCAACC CCACAGTTTTGAAGGTCTGAAGT HPRT1 ACCCTTTCCAAATCCTCAGC TCCTCCTCCTGAGCAGTCA LKB1 TGTCGGTGGGTATGGACAC CGGCTGGTAGATGACCTCG MARK1 GAGCGGGACACGGAAAATCAT TGCTACTCGACTTGGTAGGCT RPLP1 ATGAGGCTCCCAATGTTGAC CGGAGGATAAGATCAATGCC STK25 ATCACCCGCTACTTTGGCTC TGGCAATGTATGTCTCCTCCAG STRADA CAGGAGAGTACGTGACTGTACG CGATATGGCACGATATTGGGATG YWHAZ CCTGCATGAAGTCTGTAACTGAG GACCTACGGGCTCCTACAACA
146 Chapter 3: Differential Gene Expression Profiling
Supplemental Table 3.3. List of human primer sequences used for qRT-PCR experiments on p53 signaling.
Primer (human) Forward (5'-3') Reverse (5'-3') TP53 ATG GAG GAG CCG CAG TCA GAT GCA GCG CCT CAC AAC CTC CGT C AURKB CAGTGGGACACCCGACATC GTACACGTTTCCAAACTTGCC DDIT3 GGAAACAGAGTGGTCATTCCC CTGCTTGAGCCGTTCATTCTC KIF23 AGTCAGCGAGAGCTAAGACAC GGTTGAGTCTGTAGCCCTCAG KITLG AATCCTCTCGTCAAAACTGAAGG CCATCTCGCTTATCCAACAATGA KIT CGTTCTGCTCCTACTGCTTCG CCCACGCGGACTATTAAGTCT ME1 CTGCTGACACGGAACCCTC GATCTCCTGACTGTTGAAGGAAG PDGFRB AGCACCTTCGTTCTGACCTG TATTCTCCCGTGTCTAGCCCA PLK1 AAAGAGATCCCGGAGGTCCTA GGCTGCGGTGAATGGATATTTC POSTN CTCATAGTCGTATCAGGGGTCG ACACAGTCGTTTTCTGTCCAC RARRES3 GAGATTTTCCGCCTTGGCTAT CCGGGGTACTCACTTGGAG RRM2B AGAGGCTCGCTGTTTCTATGG GCAAGGCCCAATCTGCTTTTT CDC25C TCTACGGAACTCTTCTCATCCAC TCCAGGAGCAGGTTTAACATTTT FABP4 ACTGGGCCAGGAATTTGACG CTCGTGGAAGTGACGCCTT CDKN1A TGTCCGTCAGAACCCATGC AAAGTCGAAGTTCCATCGCTC
147 Chapter 4. General Discussion and Future Directions
CHAPTER 4
DISCUSSION AND FUTURE DIRECTIONS
148 Chapter 4. General Discussion and Future Directions
Cerebral cavernous malformation is one of the most prevalent neurovascular
malformations and it is caused by mutations in three genes, CCM1, CCM2 and CCM3.
The CCM proteins are implicated in several biological functions through many molecular
pathways including angiogenesis, formation of stable junctions, cell death, ECM
remodeling and Golgi assembly and orientation [1-6]. Through a diverse range of interactions, CCM proteins regulated several signaling pathways including β1-integrin signaling, MEKK3 signaling, RhoA, Rap1, VEGF signaling, and β-catenin [7-11]. A few of these signaling pathways suggest a role for the CCM proteins, specifically CCM1 and
CCM3, in polarity and maintenance of cell integrity. However, the molecular mechanisms of most of these interactions are still being elucidated. Here, we present evidence for CCM3 in the canonical LKB1-STRAD signaling pathway that regulates epithelial cell polarity and neuronal polarity. CCM3 interacts with STRADα, the pseudokinase that regulates LKB1 activity and is required for LKB1-STRAD-dependent epithelial polarity (Chapter 2, Figure 2.1, 2.3). In endothelial cells, LKB1 and STRAD mRNA expressions are regulated by CCM3 as absence of CCM3 increases the expression of these transcripts (Chapter 3, Figure 3.2). Further, CCM3 phenotype in endothelial cells show a vast dysregulation in gene expression of intracellular signaling and downregulation of adhesion molecules (Chapter 3, Table 3.2). Our evidence suggests a role for CCM3 regulating epithelial cell polarity through the LKB1 signaling pathway and a role in maintaining cell integrity by regulating the expression of cell adhesion molecules in endothelial cells.
Previous studies in our lab have shown that Stk25 is part of an LKB1 cell polarization pathway [12]. In an effort to elucidate the molecular mechanisms by which
149 Chapter 4. General Discussion and Future Directions
Stk25 regulates cell polarity via the LKB1 signaling, we examined a role for the Stk25 binding partner CCM3 (Chapter 2, Figure 2.1). We found that CCM3 also binds to
STRADα. The interaction between CCM3 and STRADα occurred in the absence of
CCM3 binding to Stk25. In fact, a CCM3 mutant that fails to interact with Stk25 immunoprecipitates more STRADα than wild-type CCM3. In addition, Stk25 overexpression reduces the CCM3-STRADα association. These experiments suggest that Stk25-CCM3 binding may reduce binding of CCM3 to STRAD and the regulation of
LKB1 by these molecules during polarization is likely complex. Furthermore, in the future it will be important to determine if CCM3-STRAD and Stk25-STRAD complexes have separable roles in cell polarization pathways.
Are Stk25 and CCM3 upstream or downstream of LKB1-STRAD complex? Do they both interact with STRAD at the same time? To better dissect this further, future studies using GST pull down techniques, mutant binding proteins and binding competition assays would be beneficial to determine the nature of the complex and the binding domains. Preliminary data shows that CCM3 is not needed for Stk25-STRAD interaction as they co-immunoprecipitate more robustly from HEK 293 cells expressing
CCM3 shRNA as compared to control cells (Chapter 2, Figure S2.2). This suggests that the presence of CCM3 impedes the interaction of Stk25 and STRAD. The ability of these two proteins to affect the binding of one another suggests they are not independent as the absence of one enhances the binding of the other to STRAD.
Given this data, we propose a working model of the complex (Figure 4.1). We believe that CCM3 interacts directly with STRAD through its C-terminal focal adhesion
150 Chapter 4. General Discussion and Future Directions
targeting domain. CCM3 is required for LKB1-mediated epithelial cell polarity and
differentially regulates the downstream effectors, MARK1 and AMPK, suggesting a role
in LKB1 functions downstream. However, we do not have evidence suggesting whether
CCM3 is upstream or downstream of LKB1 complex. We hypothesize it is upstream of
LKB1 given that CCM3 regulates LKB1 protein expression, STRAD protein expression
and the mRNA expression of both genes. However, further studies such as rescue
experiments need to be performed to determine whether CCM3 is acting upstream or
downstream of LKB1. In our model, Stk25 also associates with STRAD as previously
determined by Matsuki and colleagues [12] (Figure 4.1) but it is undetermined whether
this interaction is direct or at the same time as CCM3. It is possible that these interactions
play different roles on LKB1 functions. For example, LKB1-STRAD-STK25-CCM3 may be involved in the regulation of Golgi morphology while LKB1-STRAD-CCM3 complex is involved in cell polarity and cell metabolism regulations.
In the light of this novel connection, it becomes imperative that the biological relevance of CCM3 and LKB1 signaling pathway be examined in endothelial cells to determine whether LKB1 contributes to a cell polarity defect in CCM disease. As a first step to examining a possible role for LKB1 in the CCM3 phenotype in endothelial cells, we compared global gene expression in normal endothelial cells to those expressing
CCM3 shRNA lentivirus (Chapter 3). There was no significant change in LKB1 signaling pathway molecules; however, we observed significant but small increase in LKB1 expression (Table 3.1). In addition, several genes not associated with the LKB1 pathway are significantly and dramatically misregulated in CCM3 knockdown cells (Table 3.2).
151 Chapter 4. General Discussion and Future Directions
Finally, using western blot to examine protein expression levels, we show that LKB1 and
STK25 protein levels are reduced when CCM3 is absent from endothelial cells. STRAD
levels were barely detectable in these cells and we were not able to determine if its levels
also decreased. This data suggest CCM3 regulates LKB1 and STRAD expression at
different levels, potentially by both direct and indirect mechanisms, which might be cell
type dependent. Interestingly, Veleva-Rotse et al (2014, [13]) saw a similar mode of regulation of LKB1 when STRADα was deleted from cells. LKB1 protein expression is reduced in STRADα knockout mice but LKB1 gene expression was higher, similar to our observations in CCM3 knockdown endothelial cells [13]. It is currently unknown whether this mode of regulation is significant for LKB1 but future studies will focus on determining the mechanisms by which differential regulation of LKB1 is made possible.
One possible reason might be due to the localization of inactive LKB1 protein in the nucleus. Inactive LKB1 resides in the nucleus where upon binding to STRAD is activated and shuttled to the cytoplasm [14-16]. This unusual mode of regulation might play a role
in the ability of STRAD and CCM3 to regulate LKB1 protein and mRNA expression in
different ways. It would be interesting to determine the functions (if any) of active or
inactive LKB1 in the nucleus. Is LKB1 regulating the expression of transcription factors?
Moreover, is CCM3 or STRAD involved in this regulation?
A Role for CCM3 in LKB1-dependent Cellular Polarity
Cell polarity is an essential process in the development of multicellular organisms
as it is crucial for development and maintenance of tissue integrity [17]. The
establishment of cell polarity is dependent on several signaling pathways that are
152 Chapter 4. General Discussion and Future Directions
responsible for executing asymmetry in cell shape, organelle distribution and cell
functions that characterize a polarized cell [18]. Conserved polarity signaling pathways
include LKB1 signaling, which has an essential role in establishing polarity in many cell
types [19-22]. We find that CCM3 binds to STRADα, a pseudokinase that binds to LKB1
and regulates LKB1. CCM3 regulates the protein levels of STRADα and LKB1. In
addition, CCM3 knockdown reduces expression of Mark1, a downstream polarity
effector of LKB1 (Chapter 2, Figures 2.1, 2.3, 2.5). Our data places CCM3 in the LKB1
signaling pathway through interactions with STRADα and demonstrates CCM3 regulates
epithelial cell polarity. In addition, by differentially modulating the protein expression of
downstream effector kinases, CCM3 may play an important role in LKB1 signaling. Our
findings suggest deficits in CCM3 expression may alter the activation of LKB1-mediated signaling and cell polarity control. Therefore, it would be interesting to probe further, how CCM3 regulates LKB1-dependent polarization.
A polarized epithelial cell has several prominent features that are dependent on polarization, including a microvilli-studded apical membrane, junctional protein complexes forming tight junctions and adherens junctions. These are regulated in part by the sorting of surface proteins to either the apical or basolateral domains [24]. Single intestinal W4 epithelial cells have hallmarks of a polarized cell including actin-rich apical brush borders. Interestingly, upon the activation of LKB1 they also redistribute junctional proteins despite lacking adjacent cells to support cell-to-cell adhesions [24]. The molecular mechanisms by which LKB1 mediates these activities are not completely understood. ten Klooster and colleagues (2009) suggest the ERM (Ezrin, Radixin,
153 Chapter 4. General Discussion and Future Directions
Moesin) pathway is downstream of LKB1 and could have critical roles to set up polarity
in epithelial cells [25]. The authors show that upon activation of the LKB1 complex in
epithelial cells, Mst4 becomes partially translocated to the apical domain, where it
presumably phosphorylates Ezrin [25]. However, it is not known how Mst4 is activated
or what aspects of epithelial polarity it regulates at the apical domain.
CCM3 regulates LKB1 and STRAD protein expression which would suggest it
regulates the activity of the complex. It would be interesting to determine what aspects of
epithelial cell polarity CCM3 regulates. Does it influence the polarity of adherens
junction proteins, the distribution of surface proteins or the formation of microvilli?
These experiments could be performed in W4 cells deficient in CCM3 to determine
whether apical markers such as CD66 (carcinoembryonic antigen/CEA), CD13
(dipeptidyl peptidase IV), and CD26 (amino peptidase-N) are localized to the microvilli.
Or whether tight junction components such as ZO-1 are correctly localized upon STRAD induction by doxycycline treatment. In addition, might CCM3 overexpression rescue the polarization defects in W4 cells lacking LKB1? Overexpressing CCM3 in W4 cells with reduced LKB1 will help resolve whether CCM3 acts upstream or downstream of LKB1.
For instance, if CCM3 acts upstream of LKB1, complete loss of LKB1 would not likely be rescued by CCM3 overexpression. Further, are these polarization effects dependent on Stk25 or CCM3 association with the complex? Stk25 belongs to the GCKIII family of proteins, which includes Mst4. The two proteins have been demonstrated to have similar functions at the Golgi apparatus and could compensate for each other [1, 26]. If the above
Mst4-ERM signaling pathway is indeed a molecular mechanism downstream of LKB1, future experiments should focus on understanding whether CCM3 is required for Mst4-
154 Chapter 4. General Discussion and Future Directions
ERM and potentially Stk25-ERM signaling. In addition, it would be interesting to determine whether CCM3 regulates polarity in endothelial cells given our gene expression data suggesting a significant downregulation of adhesion molecules and dysregulation of the extracellular matrix signaling in CCM3 knockdown endothelial cells.
Interestingly, LKB1 null mice manifest several developmental abnormalities including vascular abnormality where angiogenesis and the expression of VEGF are severely defective [27]. It would be critical to determine whether the vascular defects manifested in familial CCM due to loss of CCM3, also show a reduction in LKB1 protein levels, as our data suggests. Attempting to rescue vascular lesions in CCM3 knock out mice [28] by artificially overexpressing LKB1 could provide insights into the contributions of LKB1 in CCM3-dependent endothelial cell functions.
The establishment of endothelial polarity is crucial for vascular-lumen organization, which is defective in CCM disease. Several lines of evidence, including our data and those from other laboratories, suggest that CCM3 may regulate endothelial cell polarity via LKB1 signaling. CCM3 has been shown to regulate Golgi orientation and polarized migration, a function that relies heavily on its association with Stk25 in epithelial cells [1]. Stk25 regulates Golgi morphology as part of an LKB1-Stk25-GM130 signaling pathway [12]. In addition to roles in Golgi morphology and cell polarity, loss of
LKB1 in endothelial cells disrupts VEGF and TGFβ signaling, leading to deficits in angiogenesis [29]. The TGFβ pathway is a well-known mediator of signaling in endothelial cells and an established signaling pathway that specifies polarity in brain development [30]. Intact TGFβ signaling in endothelial cells is necessary for proper vascular formation [31]. Interestingly, loss of LKB1 in endothelial cells and fibroblasts
155 Chapter 4. General Discussion and Future Directions
leads to defective TGFβ signaling and the defects related to loss of LKB1 can be rescued
by exogenous TGFβ expression [29, 32] indicating an intricate role for LKB1 in vascular
development especially in the establishment of endothelial cells in their environment.
However, whether CCM3 might play a role in LKB1 signaling in endothelial cells
remains to be determined.
It would be interesting to test the hypothesis that a loss-of-function mutation of CCM3
resulting in the disruption of endothelial lining integrity is caused by the reduction and
disruption of LKB1 and TGFβ signaling, preventing endothelial cell polarity. Designing
experimental models such as cell culture models of endothelial cells lacking CCM3 to
test this hypothesis will enhance our understanding of CCM3 cellular and molecular
mechanisms as well as the pathogenesis of CCM. Further, using CCM3 deficient
endothelial cells from a homozygous CCM3 conditional KO mouse carrying the Tie2-Cre
transgene would be a useful in vitro model to determine whether loss of CCM3 in
endothelial cells results in dysregulation of LKB1 and TGFβ signaling. Furthermore, it
could be easily determined whether LKB1 overexpression rescued the observed cellular
phenotypes.
Louvi et al (2014) provide another example of a potential role for CCM3 in
cellular polarization. In a recent study, they demonstrate that CCM3 is necessary for neuronal migration and orientation of radial glial cells in early cortical development through a RhoA-dependent mechanism [33]. There is an established link between CCM3 and RhoA signaling in vascular development and this novel function in cortical development suggests a function in polarity outside of endothelial cells. However, even
156 Chapter 4. General Discussion and Future Directions
though it is established that LKB1 regulates axonogenesis, brain lamination and
migration [12, 19, 21, 34], this study does not link CCM3 to the LKB1 pathway in
cortical development. Future work would focus on understanding the molecular
mechanisms of CCM3 in LKB1-dependent early cortical development.
Potential role for CCM3 in other LKB1-dependent processes.
The Golgi apparatus is implicated in several aspects of cell polarity, in both epithelial cells and neuronal polarity [35-37]. It has been previously shown that CCM3 in association with the GCKIII family of proteins and Golgi matrix protein GM130 regulates the morphology and polarization of the Golgi apparatus [1, 38]. In addition,
Stk25, a CCM3 interacting protein, regulates Golgi morphology in association with
LKB1 and GM130 [12]. However, it is undetermined whether CCM3 regulates cell polarity through regulation of the Golgi apparatus. Given its role at the Golgi apparatus it is possible that CCM3 recruits LKB1 to the Golgi to regulate morphology and function and this could ultimately regulate protein trafficking during cell polarization. This may also involve Stk25, however the complex interaction between Stk25, CCM3 and STRAD need to be clarified (Figure 4.1). Future studies should focus on the resolution of the mechanistic details involving CCM3, Stk25 and STRAD regulation of Golgi function.
LKB1 is a master kinase with well-defined roles in cellular polarity, maintaining cell metabolism and tumor suppression. The activation of the LKB1-STRAD complex leads to the activation of several downstream kinases, most in the AMPK family of kinases [39]. The regulation of these downstream kinases enables LKB1 to control
157 Chapter 4. General Discussion and Future Directions
diverse cellular processes. Interestingly, in our data, we observed a differential regulation
of LKB1 downstream kinases by CCM3 (Chapter 2, Figure 2.5). Specifically, CCM3
differentially regulated the protein expression levels of MARK1 and AMPK. When
CCM3 is depleted in W4 cells, MARK1 protein levels significantly reduced while total
AMPK and phosphorylated AMPK levels significantly increased (Chapter 2, Figure 2.5).
This observed differential regulation not only suggests a role for CCM3 downstream of the LKB1-STRAD complex, but it also suggests potential biological functions of CCM3
in other LKB1-mediated processes beyond epithelial cell polarity. Both MARK1 and
AMPK play a critical role in the establishment of neuronal polarity [40, 41] and would
suggest CCM3 might regulate their activities in neuronal polarity downstream of LKB1.
Further, the differential regulations of MARK1 and AMPK suggests CCM3 might play a
role in selecting which kinases downstream of LKB1 are activated in the cell. Perhaps
CCM3 acts to regulate a switch between cell polarity and cell metabolism during cellular
stress. This phenomenon would be interesting to investigate further as it would have
implications beyond cellular polarity. To begin these future explorations, the use of
MARK1 and AMPK knockout mouse models would be beneficial to assess CCM3
functions and determine whether CCM3 plays a role in other MARK1 and AMPK
functions (such as regulating cell metabolism).
Does CCM3 modulate p53-dependent apoptosis in endothelial cells?
p53, the protein product of tumor suppressor protein TP53, is one of the most
mutated or inactivated genes in cancer [42]. p53 also causes cell-cycle arrest and
apoptotic cell death in response to cellular stress and DNA damage. Active p53
158 Chapter 4. General Discussion and Future Directions selectively influences the expression of several downstream target genes however the molecular basis of this regulation has been the subject of intensive studies. We show that reducing CCM3 in HAEC cells causes alterations in the expressions of several genes downstream of p53-apoptotic pathway. Overall, there is a significant activation of p53 activity resulting in a down-regulation of many p53 effectors regulating cell cycle arrest and apoptosis (Chapter 3, Table 3.4). This finding probes several questions including whether the presence of CCM3 inhibits the induction of apoptosis and cell death in a p53- dependent manner in endothelial cells as well as whether lack of CCM3 (due to a loss-of- function mutation) in the vascular endothelium could lead to increased apoptosis in
CCM.
The CCM proteins interact with several protein kinases implicating them in several cellular pathways. However, the link between CCM proteins and cell death pathways is weak. CCM2 interacts with TrkA receptor kinase which signals to cell death pathways in an unknown mechanism [43]. It was suggested that since CCM3 and GCKIII kinases were recruited to TrkA [44], CCM3 could recruit CCM2 to TrkA. It has also been proposed that aberrant apoptosis in the underlining endothelium of the neurovasculature leads to the formation of lesions in CCM [2] but a molecular mechanism has yet to been proposed. If poor endothelial integrity and increased apoptosis in the endothelium is as a result of loss-of-function in one of the CCM proteins, it is imperative we understand the molecular mechanisms leading to increased cell death.
CCM3 was first identified in a screen for genes up-regulated during induction of apoptosis in a premyeliod cell line [45, 46] and is thought to play a role in cell death.
159 Chapter 4. General Discussion and Future Directions
However, the specific role of CCM3 in cell survival is yet to be clearly determined as it
has both pro-survival and pro-apoptotic effects on cells, which suggests a cell-specific
response is at play. The significant activation of p53 signaling observed in cultured
endothelial cells is relevant to understanding the intersection of CCM3 molecular
mechanisms with CCM disease phenotypes. Prior to this study, the functions of the tumor
suppressor p53, which includes inhibiting cell cycle, cell proliferation and as well as
transcription regulation have been linked to CCM1 and interestingly LKB1, both of
which exist in complexes with CCM3. The loss of p53 sensitizes mice with mutations in
CCM1 to form lesions as seen in CCMs [47] that suggests that p53 might play a direct
role in the formation of cavernomas seen in the vasculature. In addition, it was recently
reported that p53 binds to one of the four putative p53 binding sites on LKB1 promoter
region [48] regulating the expression of LKB1. It has also been shown that endogenous
reduction of LKB1 accelerates cell cycle progression in a p53-dependent manner [49]. It is possible that CCM3 regulates apoptosis and cell cycle progression through an
LKB1/p53-dependent mechanism. Testing this hypothesis in a primary endothelial cell line that lacks CCM3 would be beneficial. In addition, using a disease mutation of CCM3 that is commonly seen in CCM would be ideal, as the evidence from these studies would contribute to the molecular mechanisms of CCM formation. While in our studies, we used human aortic endothelial cells, it would be beneficial for future studies to utilize endothelial cells found in the microvasculature of the brain as their morphologies and properties are in line with the environment of cavernomas in CCM disease.
CCM3 regulates endothelial cell integrity
160 Chapter 4. General Discussion and Future Directions
The integrity of endothelial cells is severely affected in CCM disease and is
critical for lesion development. The endothelial cells lining the vasculature lose their
adhesion molecules at tight junctions in cells that are null for CCM3 [23]. Adherens
junctions and VE-cadherin are required for endothelial apicobasal polarity in vitro and
during embryonic development [50]. As alluded to earlier, we observed a significant
decrease in the gene expressions of adhesion molecules in endothelial cells lacking
CCM3 (Chapter 3, Table 3.2). Whether this phenotype occurs in CCM disease when
CCM3 is mutated remains to be determined, however our gene expression study suggests
this dysregulation would be worth investigating further. In addition to our evidence, a
recent study by Stamatovic et al. (2015) highlights a potential role of CCM3 in regulating
tight junction complex organization [8]. Through crosstalk between and ERK1/2-
cortactin, CCM3 maintains endothelial tight junctions and the integrity of the blood brain
barrier [8] in human brain endothelial cells.
From these findings, it is possible that loss of vascular adhesion molecules due to a loss-of-function mutation in CCM3 prevents the formation of functional cell-cell contacts in endothelial cells. This could lead to the aggregation of endothelial cells and the formation of lesions characterized in CCM disease. Future studies focused on understanding the molecular mechanisms underlining this important function of CCM3 would be beneficial for CCM disease pathogenesis. In addition, the novel interacting partner STRADα and link to LKB1 pathway we established in our studies could play an important role in CCM3’s functions in maintaining the integrity of the endothelial lining.
LKB1 promotes localization of junctional proteins in epithelial polarity and in neuronal
161 Chapter 4. General Discussion and Future Directions polarity, the distribution of the Golgi apparatus into axons [12, 22] and it is possible
CCM3 in association with LKB1 could be maintaining the expression of cell adhesion molecules. It would be interesting to determine in human brain endothelial cells lacking
CCM3 if there is a dysregulation of adhesion molecule localizations in conjunction with their downregulation. In addition, are these functions of CCM3 in maintaining cell integrity independent of CCM1 or CCM2? CCM1 and CCM2 interact with Rap1/ β- integrin complex at endothelial cell junctions and it is believed that they help maintain the integrity and permeability of the endothelial lining [51-53]. CCM1 interacts with VE-
Cadherin and directs the organization of adherens junctions and adherens junction association with the polarity complex [54]. Is CCM3 upstream of this signaling pathway?
From preliminary evidence from our gene expression studies, we believe CCM3 plays a critical role in the maintenance of endothelial cell integrity and this is one of the mechanisms impacted by CCM3 mutation leading to lesion formation and CCM disease.
Unresolved Complexities with CCM3, Stk25 and LKB1 signaling
Interestingly both CCM3 and Stk25 are required for efficient LKB1-dependent polarization of W4 cells. However, Stk25 appears to hinder CCM3 binding to STRADα.
The CCM3 mutant defective for Stk25 binding coimmunoprecipitates STRADα more effectively. In addition, Stk25 overexpression reduces CCM3-STRADα coimmunoprecipitation. Thus although both CCM3 and Stk25 seem required for polarization and both bind STRAD they do not act in a synergistic manner. Polarization is a complex process that requires a multitude of cellular reorganization events that once established need to be maintained. It is possible that our observations reflect this
162 Chapter 4. General Discussion and Future Directions
complexity. CCM3 and Stk25 may have different temporal or spatial roles in polarization
both of which are necessary and require LKB1-STRADα, but which are not necessarily
synergistic. Future work therefore should be designed to resolve if CCM3-STRADα and
Stk25-STRADα complexes have distinct functional consequences in polarization or if a
trimeric STRADα-CCM3-Stk25 complex is required.
Our observations that CCM3 is required for LKB1-dependent polarization were made in an intestinal epithelial cell line that was specifically engineered to address
LKB1-STRADα dependent polarization [22]. Since endothelial cells represent the cell type most likely defective in CCMs, we tried to determine if CCM3 regulates LKB1 in these cells also. Surprisingly although knocking down CCM3 did reduce LKB1 protein levels, STRADα protein was barely detectible and not clearly altered between CCM3 expressing and nonexpressing cells. This begs the question whether another molecule substitutes for STRADα in endothelial cells. One possibility is that STRADβ fulfills that role. Unfortunately, there currently is not an adequate STRADβ antibody to resolve this issue. In the future, it will also be interesting to determine if CCM3 regulates LKB1 levels by regulating STRAD or if its effects on LKB1 are STRAD independent.
Conclusions
This work represents an important advancement in our insight into CCM3’s role in cell polarity and possible contribution to CCM pathogenesis. Despite the novel molecular association between CCM3 and STRAD in LKB1-dependent polarization, there is still much to discover to fully understand the mechanisms by which CCM3
163 Chapter 4. General Discussion and Future Directions regulates LKB1 protein in addition to how CCM3 functions in brain vascular malformations. The data generated during the course of this project suggests that CCM3 regulates the protein expression of LKB1 along with some downstream effectors of the
LKB1 signaling pathway in an effort to regulate epithelial cell polarity. These mechanisms remain to be studied in brain microvascular endothelial cells and in CCM animal models given the role of LKB1 in vascular development, it would be interesting to examine a role for LKB1-CCM3 in endothelial cells. Understanding the effects of
CCM3-STRAD interaction in the future is essential for and will greatly expand our understanding of the molecular mechanisms of CCM3 in addition to understanding where this signaling pathway cross-talks with other CCM proteins. We believe this work is a critical first step on this path and will enable the elucidation of the role of CCM3 in signaling pathways that are critical for the development of lesions in CCM disease pathogenesis, particularly endothelial cell integrity maintenance.
The current treatment for CCM remains surgical resection of the lesions [45], however sometimes the locations of the cavernomas are impossible to remove by surgical means. Therefore, future treatments are focused on inhibiting the RhoA-ROCK signaling pathway that is activated in CCM patients with mutations in CCM1, CCM2 and CCM3
[5, 50, 53, 55, 56]. The need for more viable and relevant targets that are specific to the formation and progression of vascular lesions in endothelial cells would be more effective for patients with CCM and the understanding of CCM3 cellular and molecular biology would be key to this.
164 Chapter 4. General Discussion and Future Directions
REFERENCES 1. Fidalgo, M., et al., CCM3/PDCD10 stabilizes GCKIII proteins to promote Golgi assembly and cell orientation. J Cell Sci, 2010. 123(Pt 8): p. 1274-84.
2. Chen, L., et al., Apoptotic functions of PDCD10/CCM3, the gene mutated in cerebral cavernous malformation 3. Stroke, 2009. 40(4): p. 1474-81.
3. Lin, C., et al., PDCD10/CCM3 acts downstream of {gamma}-protocadherins to regulate neuronal survival. J Biol Chem, 2010. 285(53): p. 41675-85.
4. Liu, H., et al., Ccm1 Assures Microvascular Integrity During Angiogenesis.
Transl Stroke Res, 2010. 1(2): p. 146-153.
5. Zheng, X., et al., CCM3 signaling through sterile 20-like kinases plays an essential role during zebrafish cardiovascular development and cerebral cavernous malformations. J Clin Invest, 2010. 120(8): p. 2795-804.
6. Zhu, Y., et al., Differential angiogenesis function of CCM2 and CCM3 in cerebral cavernous malformations. Neurosurg Focus, 2010. 29(3): p. E1.
7. Voss, K., et al., CCM3 interacts with CCM2 indicating common pathogenesis for cerebral cavernous malformations. Neurogenetics, 2007. 8(4): p. 249-56.
8. Stamatovic, S.M., et al., PDCD10 (CCM3) regulates brain endothelial barrier integrity in cerebral cavernous malformation type 3: role of CCM3-ERK1/2-cortactin cross-talk. Acta Neuropathol, 2015. 130(5): p. 731-50.
9. Draheim, K.M., et al., CCM2-CCM3 interaction stabilizes their protein expression and permits endothelial network formation. J Cell Biol, 2015. 208(7): p. 987-
1001.
165 Chapter 4. General Discussion and Future Directions
10. You, C., et al., Loss of CCM3 impairs DLL4-Notch signalling: implication in
endothelial angiogenesis and in inherited cerebral cavernous malformations. J Cell Mol
Med, 2013. 17(3): p. 407-18.
11. Ma, X., et al., PDCD10 interacts with Ste20-related kinase MST4 to promote cell
growth and transformation via modulation of the ERK pathway. Mol Biol Cell, 2007.
18(6): p. 1965-78.
12. Matsuki, T., et al., Reelin and stk25 have opposing roles in neuronal polarization
and dendritic Golgi deployment. Cell, 2010. 143(5): p. 826-36.
13. Veleva-Rotse, B.O., et al., STRAD pseudokinases regulate axogenesis and LKB1
stability. Neural Dev, 2014. 9: p. 5.
14. Boudeau, J., G. Sapkota, and D.R. Alessi, LKB1, a protein kinase regulating cell
proliferation and polarity. FEBS Lett, 2003. 546(1): p. 159-65.
15. Boudeau, J., et al., Analysis of the LKB1-STRAD-MO25 complex. J Cell Sci, 2004.
117(Pt 26): p. 6365-75.
16. Zeqiraj, E., et al., Structure of the LKB1-STRAD-MO25 complex reveals an allosteric mechanism of kinase activation. Science, 2009. 326(5960): p. 1707-11.
17. Nelson, W.J., Epithelial cell polarity from the outside looking in. News Physiol
Sci, 2003. 18: p. 143-6.
18. Li, P., et al., Epithelial cell polarity determinant CRB3 in cancer development. Int
J Biol Sci, 2015. 11(1): p. 31-7.
19. Shelly, M., et al., LKB1/STRAD promotes axon initiation during neuronal polarization. Cell, 2007. 129(3): p. 565-77.
166 Chapter 4. General Discussion and Future Directions
20. Forcet, C., et al., 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, 2005. 14(10): p. 1283-92.
21. Barnes, A.P., et al., LKB1 and SAD kinases define a pathway required for the polarization of cortical neurons. Cell, 2007. 129(3): p. 549-63.
22. Baas, A.F., et al., Complete polarization of single intestinal epithelial cells upon activation of LKB1 by STRAD. Cell, 2004. 116(3): p. 457-66.
23. Burkhardt, J.K., et al., Upregulation of transmembrane endothelial junction proteins in human cerebral cavernous malformations. Neurosurg Focus, 2010. 29(3): p.
E3.
24. Baas, A.F., et al., Activation of the tumour suppressor kinase LKB1 by the STE20- like pseudokinase STRAD. EMBO J, 2003. 22(12): p. 3062-72.
25. ten Klooster, J.P., et al., Mst4 and Ezrin induce brush borders downstream of the
Lkb1/Strad/Mo25 polarization complex. Dev Cell, 2009. 16(4): p. 551-62.
26. Ceccarelli, D.F., et al., CCM3/PDCD10 heterodimerizes with germinal center kinase III (GCKIII) proteins using a mechanism analogous to CCM3 homodimerization. J
Biol Chem, 2011. 286(28): p. 25056-64.
27. Ylikorkala, A., et al., Vascular abnormalities and deregulation of VEGF in Lkb1- deficient mice. Science, 2001. 293(5533): p. 1323-6.
28. He, Y., et al., Stabilization of VEGFR2 signaling by cerebral cavernous malformation 3 is critical for vascular development. Sci Signal, 2010. 3(116): p. ra26.
167 Chapter 4. General Discussion and Future Directions
29. Londesborough, A., et al., LKB1 in endothelial cells is required for angiogenesis
and TGFbeta-mediated vascular smooth muscle cell recruitment. Development, 2008.
135(13): p. 2331-8.
30. Yi, J.J., et al., TGF-beta signaling specifies axons during brain development. Cell,
2010. 142(1): p. 144-57.
31. Lebrin, F., et al., Endoglin promotes endothelial cell proliferation and TGF-
beta/ALK1 signal transduction. EMBO J, 2004. 23(20): p. 4018-28.
32. Vaahtomeri, K., et al., Lkb1 is required for TGFbeta-mediated myofibroblast
differentiation. J Cell Sci, 2008. 121(Pt 21): p. 3531-40.
33. Louvi, A., S. Nishimura, and M. Gunel, Ccm3, a gene associated with cerebral cavernous malformations, is required for neuronal migration. Development, 2014.
141(6): p. 1404-15.
34. Asada, N., K. Sanada, and Y. Fukada, LKB1 regulates neuronal migration and neuronal differentiation in the developing neocortex through centrosomal positioning. J
Neurosci, 2007. 27(43): p. 11769-75.
35. de Anda, F.C., et al., Centrosome motility is essential for initial axon formation in
the neocortex. J Neurosci, 2010. 30(31): p. 10391-406.
36. de Anda, F.C., et al., Centrosome localization determines neuronal polarity.
Nature, 2005. 436(7051): p. 704-8.
37. Sylvie Deborde, D.G., Aparna Lakkaraju, Enrique Rodriguez-Boulan, Golgi
apparatus and epithelial cell polarity, in The Golgi Apparatus: State of the art 110 years
after Camillo Golgi's discovery, A.A. Mironov, Pavelka, Margit (Eds.), Editor. 2008,
Springer.
168 Chapter 4. General Discussion and Future Directions
38. Kean, M.J., et al., Structure-function analysis of core STRIPAK Proteins: a signaling complex implicated in Golgi polarization. J Biol Chem, 2011. 286(28): p.
25065-75.
39. Lizcano, J.M., et al., LKB1 is a master kinase that activates 13 kinases of the
AMPK subfamily, including MARK/PAR-1. EMBO J, 2004. 23(4): p. 833-43.
40. Amato, S., et al., AMP-activated protein kinase regulates neuronal polarization by interfering with PI 3-kinase localization. Science, 2011. 332(6026): p. 247-51.
41. Biernat, J., et al., Protein kinase MARK/PAR-1 is required for neurite outgrowth and establishment of neuronal polarity. Mol Biol Cell, 2002. 13(11): p. 4013-28.
42. Hollstein, M., et al., p53 mutations in human cancers. Science, 1991. 253(5015): p. 49-53.
43. Harel, L., et al., CCM2 mediates death signaling by the TrkA receptor tyrosine kinase. Neuron, 2009. 63(5): p. 585-91.
44. Costa, B., et al., STK25 protein mediates TrkA and CCM2 protein-dependent death in pediatric tumor cells of neural origin. J Biol Chem, 2012. 287(35): p. 29285-9.
45. Bergametti, F., et al., Mutations within the programmed cell death 10 gene cause cerebral cavernous malformations. Am J Hum Genet, 2005. 76(1): p. 42-51.
46. Wang, Y., et al., cDNA cloning and expression of an apoptosis-related gene, humanTFAR15 gene. Sci China C Life Sci, 1999. 42(3): p. 323-9.
47. Plummer, N.W., et al., Loss of p53 sensitizes mice with a mutation in Ccm1
(KRIT1) to development of cerebral vascular malformations. Am J Pathol, 2004. 165(5): p. 1509-18.
169 Chapter 4. General Discussion and Future Directions
48. Co, N.N., et al., Loss of LKB1 in high-grade endometrial carcinoma: LKB1 is a
novel transcriptional target of p53. Cancer, 2014. 120(22): p. 3457-68.
49. Liang, X., et al., Endogenous LKB1 knockdown accelerates G(1)/S transition through p53 and p16 pathways. Cancer Biol Ther, 2010. 9(2): p. 156-60.
50. Whitehead, K.J., et al., The cerebral cavernous malformation signaling pathway promotes vascular integrity via Rho GTPases. Nat Med, 2009. 15(2): p. 177-84.
51. Zhang, J., et al., Krit1 modulates beta 1-integrin-mediated endothelial cell proliferation. Neurosurgery, 2008. 63(3): p. 571-8; discussion 578.
52. Zawistowski, J.S., et al., KRIT1 association with the integrin-binding protein
ICAP-1: a new direction in the elucidation of cerebral cavernous malformations (CCM1) pathogenesis. Hum Mol Genet, 2002. 11(4): p. 389-96.
53. Glading, A., et al., KRIT-1/CCM1 is a Rap1 effector that regulates endothelial cell cell junctions. J Cell Biol, 2007. 179(2): p. 247-54.
54. Lampugnani, M.G., et al., CCM1 regulates vascular-lumen organization by inducing endothelial polarity. J Cell Sci, 2010. 123(Pt 7): p. 1073-80.
55. Crose, L.E., et al., Cerebral cavernous malformation 2 protein promotes smad ubiquitin regulatory factor 1-mediated RhoA degradation in endothelial cells. J Biol
Chem, 2009. 284(20): p. 13301-5.
56. Stockton, R.A., et al., Cerebral cavernous malformations proteins inhibit Rho kinase to stabilize vascular integrity. J Exp Med, 2010. 207(4): p. 881-96.
57. Maddaluno, L., et al., EndMT contributes to the onset and progression of cerebral cavernous malformations. Nature, 2013. 498(7455): p. 492-6.
170 Chapter 4. General Discussion and Future Directions
58. Cuttano, R., et al., KLF4 is a key determinant in the development and progression of cerebral cavernous malformations. EMBO Mol Med, 2016. 8(1): p. 6-24.
171 Chapter 4. General Discussion and Future Directions
FIGURES
172 Chapter 4. General Discussion and Future Directions
Figure 4.1. Proposed Working Model of CCM3 in LKB1 signaling pathway.
The active heterotrimeric complex of LKB1 includes STRAD and MO25, both of which stabilize LKB1. Active LKB1 complex then goes on to activate several downstream kinases to regulate cell polarity, cell survival and metabolism. We show that CCM3 interacts with STRAD and regulate LKB1-mediated epithelial cell polarity. Preliminary data suggests that through𝛼𝛼 its association with STRADα, CCM3 could regulate AMPK and MARK1 protein level expression.
173 Chapter 4. General Discussion and Future Directions
174