Health Science Campus
FINAL APPROVAL OF DISSERTATION Doctor of Philosophy in Biomedical Sciences
Transcriptional Regulation of RKIP in Prostate Cancer Progression
Submitted by: Sandra Marie Beach
In partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biomedical Sciences
Examination Committee
Major Advisor: Kam Yeung, Ph.D.
Academic William Maltese, Ph.D. Advisory Committee: Sonia Najjar, Ph.D.
Han-Fei Ding, M.D., Ph.D.
Manohar Ratnam, Ph.D.
Senior Associate Dean College of Graduate Studies Michael S. Bisesi, Ph.D.
Date of Defense: May 16, 2007
Transcriptional Regulation of RKIP in Prostate Cancer Progression
Sandra Beach
University of Toledo
ACKNOWLDEGMENTS
I thank my major advisor, Dr. Kam Yeung, for the opportunity to pursue my
degree in his laboratory. I am also indebted to my advisory committee members past
and present, Drs. Sonia Najjar, Han-Fei Ding, Manohar Ratnam, James Trempe, and
Douglas Pittman for generously and judiciously guiding my studies and sharing reagents
and equipment. I owe extended thanks to Dr. William Maltese as a committee member
and chairman of my department for supporting my degree progress. The entire
Department of Biochemistry and Cancer Biology has been most kind and helpful to me.
Drs. Roy Collaco and Hong-Juan Cui have shared their excellent technical and practical
advice with me throughout my studies. I thank members of the Yeung laboratory, Dr.
Sungdae Park, Hui Hui Tang, Miranda Yeung for their support and collegiality.
The data mining studies herein would not have been possible without the helpful
advice of Dr. Robert Trumbly. I am also grateful for the exceptional assistance and
shared microarray data of Dr. Mohan Dhanasekaran and Jianjun Yu who both work in the
laboratory of Dr. Arul Chinnaiyan at the University of Michigan. Thank you to Dr. Sadik
Khuder and Peter Basely in Bioinformatics for data processing discussions and statistical assistance.
There have been many special people who have helped me with their scholastic advice, academic assistance, and moral support throughout my years of study. I will not spoil their humility by naming them, but I thank you all from the bottom of my heart.
ii TABLE OF CONTENTS
ACKNOWLEDGEMENTS...... ii
TABLE OF CONTENTS...... iii
INTRODUCTION ...... 1
LITERATURE...... 3
MATERIALS AND METHODS...... 29
RESULTS ...... 39
DISCUSSION...... 97
SUMARY...... 104
BIBLIOGRAPHY...... 105
ABSTRACT...... 122
iii INTRODUCTION
Raf kinase inhibitor protein (RKIP) is a conserved, multifunctional protein that seems to have a role in the metastatic process in cancer. RKIP, or PEBP1, has been described as having functions in neuronal pathways, gonadal tissues, and cell signaling pathways [reviewed in (Keller et al., 2005; Odabaei et al., 2004)]. It was originally identified as a negative regulator of the mitogen-activated protein kinase cascade initiated by Raf-1 (Yeung et al., 1999). RKIP is also able to inhibit NF-kB activation and signaling (Yeung et al., 2001). G-protein signaling can also be facilitated by RKIP.
RKIP inhibits GRK-2, which is involved in inactivating G-protein signaling [reviewed in
(Goel and Baldassare, 2004)]. Recently, RKIP has been found to be downregulated in various cancers, including melanoma, breast, hepatocellular, and prostate (Chatterjee et al., 2004; Hagan et al., 2005; Lee et al., 2006; Park et al., 2005). Breast and prostate cancer metastases have been correlated with RKIP expression in human tissue samples as well (Fu et al., 2006; Hagan et al., 2005).
Little is currently known about RKIP regulation. Putative transcription factors such as AP1 and YY1 were hypothesized to bind the RKIP promoter using computer database analysis, but these have yet to be tested (Odabaei et al., 2004). The downregulation of RKIP in cancer metastases suggests it is acted upon by a repressor that is upregulated during the metastatic process. Such a repressor may be Snail. The zinc finger transcription factor Snail is a potent repressor of E-cadherin and is involved in the
1 epithelial-to-mesenchymal transition phenotype found in cancer progression (Barbera et al., 2004; Barrallo-Gimeno and Nieto, 2005; De Craene et al., 2005).
The aim of this study was to analyze the relationship between Snail and RKIP.
We found that Snail represses RKIP in prostate cancer cell lines in vitro. Furthermore, we correlated RKIP downregulation with Snail in a meta-analysis of prostate cancer microarray studies. Finally, we utilized the microarray datasets to determine genes clustered with RKIP, and are potentially coregulated with RKIP.
2 LITERATURE REVIEW
RKIP is a phosphatidylethanolamine-binding protein.
RKIP in mammals
RKIP belongs to a highly conserved family of phospholipid-binding proteins, the phosphatidylethanolamine-binding proteins (PEBP). PEBPs are represented in eukaryotes, bacteria, and archae with no significant sequence homology to other proteins
(NCBI, 2003; NIH, 2004). Mammalian PEBPs have been categorized into four subfamilies based on sequence identity: PEBP1,2,3, or 4 (Simister et al., 2002).
Humans have two identified PEBPs, hPEBP1 and hPEBP4. hPEBP1 was identified in a yeast two-hybrid screen as a Raf-1 interacting protein and was designated RKIP (Yeung et al., 1999). Human RKIP is located on chromosome 12 (12q24.23) and is composed of four exons. The RKIP mRNA is 1507 base pairs which is translated into a 187 amino acid protein with a molecular mass of 21-23 kDa. A simple search on OMIM search found no human diseases mapping to this location (Hamosh and Hartz, 1966-2004). hPEBP4 has approximately 33 residues in its N-terminus that is not included in hPEBP1, as well as two insertions and one deletion in the amino acid sequence (Simister et al.,
2002). hPEBP4 was recently cloned from human bone marrow stromal cells and found to interfere with Ras/Raf/MEK/ERK signaling. hPEBP4 appears to promote resistance to tumor necrosis factor (TNF) α –induced apoptosis (Wang et al., 2004).
3 PEBP has been discovered and studied in several mammalian species. In 1980 a
protein named “h3” was purified from human brain (and later determined to be hPEBP1)
(Bollengier and Mahler, 1980). The molecule was identified as
phosphatidylethanolamine protein after its discovery in bovine brain as a soluble basic
protein (Bernier and Jolles, 1984; Bernier et al., 1986). The rat homolog of PEBP1 was
found by Grandy et al. by morphine affinity chromatography and determined to be related
to the bovine and human PEBP/h3 (Grandy et al., 1990; Seddiqi et al., 1994). PEBP
members have also been determined in mouse and monkey (Simister et al., 2002).
RKIP is localized in the cytoplasm and at the plasma membrane in many
different cell types (Simister et al., 2002). RKIP and its mammalian homologs are widely
expressed in tissues; it has been detected in lung, oviduct and ovary, mammary glands,
uterus, prostate epithelium, thyroid, mesenteric lymph node, megakaryocytes of the heart,
spleen, liver, and epididymis, testis, spermatids, Leydig cells, steroidogenic cells of the
adrenal gland zona fasiculata, small intestine, plasma cells, and neural cells such as brain
oliodendricytes, Schwann cells, and Pukinje cells (Frayne et al., 1998; Fu et al., 2003;
Katada et al., 1996; Moore et al., 1996; Schoentgen and Jolles, 1995). Theroux et al. measured the highest expression of mouse RKIP in brain and testes tissues (Theroux et al., 2007). The mouse PEBP family includes PEBP2, which is testes-specific and believed to play a role in spermatogenesis (Hickox et al., 2002). Simister et al. had identified mouse PEBP members besides RKIP and PEBP2 (Simister et al., 2002).
However, recent analysis of PEBP expression in the RKIP knockout mouse has shown that only RKIP and PEBP2 exist; other potential gene candidates were determined to be 4 silent pseudogenes as their was no protein or RNA products detected for them (Theroux
et al., 2007).
RKIP in non-mammals
In addition to mammals, PEBP family members have been identified and
described in various other species, but the cellular and molecular function of species-
specific PEBPs is not clear. The yeast Saccharomyces cerevisiae has the
phosphatidylethanolamine-binding protein called Tfs1p, which was identified as a
suppressor of the cdc25-1 mutant (Robinson and Tatchell, 1991). Later reports described
Tfs1 as Ic, an inhibitor of the classical member of the serine carboxypeptidase family,
carboxypeptidase Y (Bruun et al., 1998; Mima et al., 2003). Tfs1 was shown to inhibit
the yeast GTPase-activating protein Ira2, and to directly affect the activation level of the yeast Ras/cAMP/PKA pathway (Chautard et al., 2004). The deletion of TFS1 was
associated with lower levels of Ras activity (Chautard et al., 2004).
Mammalian PEBPs share sequence identity structural similarities with proteins of
several plants species. Orthologous CEN (Antirrhinum), SP (tomato), TFL1
(Arabidopsis) act as antagonists to proteins that promote development of flowers over
shoots, thereby promoting shoot emergence (Amaya et al., 1999; Banfield and Brady,
2000; Bradley et al., 1996; Bradley et al., 1997; Kardailsky et al., 1999; Ohshima et al.,
1997; Pnueli et al., 1998; Weigel and Nilsson, 1995). The crystal structure of CEN
suggests a role for CEN in cell differentiation through interference of regulatory kinase
cascades (Banfield and Brady, 2000). PEBP-related proteins in other species have not 5 been as well described. PEBP homologues found in GenBank include Drosophila,
Escherichia coli, and nematodes (NIH, 2004).
Transcriptional Regulation of RKIP
Extremely little is known about the transcriptional regulation of RKIP. A
computer database analysis (TESS master analysis) of the RKIP promoter revealed
putative binding sites for AP-1, c-Fos, c-Jun, Sp-1, YY1, WT1, Zeste, IK-1, IK-2, TAF-
1, Hb, GAGA factor, AP-4, CP-1, and ATF (Odabaei et al., 2004). Treatment of prostate
carcinoma cell lines with inhibitors for YY1 resulted in an upregulation in RKIP
expression, suggesting that YY1 may be acting as a transcriptional repressor to RKIP
(Huerta-Yepez et al., 2004; Odabaei et al., 2004). YY1 is both an activator and repressor, and Vega et al. speculate that YY1 repression of RKIP may be involved in regulation of survival pathways (Odabaei et al., 2004; Vega et al., 2004).
Kazuki et al. found that a truncated region of human chromosome 21 may contain
an RKIP repressor (Kazuki et al., 2004). These researchers introduced human
chromosome 21 into mouse embryonic stem cells to generate chimeric animals with traits
of Down syndrome (trisomy 21) (Shinohara et al., 2001). The Down syndrome mouse
model had cardiac defects in common with human Down syndrome. RKIP protein
levels were decreased in the hearts of the chimeric mice, but the repression of RKIP was relaxed when the introduced chromosome 21 was truncated at ETS2 (Kazuki et al., 2004).
There was no significant effect on RKIP RNA levels as detected by semi-quantitative
PCR between wild type, chromosome 21 chimeric, and truncated chromosome 21 6 chimeric mouse hearts suggesting that RKIP was post-transcriptionally downregulated in the chimeric hearts (Kazuki et al., 2004). Several genes are included on the truncated region; however, it is not known how any of them contribute to RKIP repression.
The general role of RKIP in the Raf-1 signaling pathway
The mitogen-activated protein (MAP) kinases are important components of pathways controlling embryogenesis, cell differentiation, cell proliferation, and cell death. The MAPK signaling cascades are organized in hierarchical three-tier modules
(Figure 1). Starting at the bottom of the tier, a MAP kinase can have multiple substrates that regulate gene expression. MAP kinases are phosphorylated and activated by MAP kinase kinases (MAPKK), and MAPKKs are activated by MAP kinase kinase kinases
(MAPKKK). The ERK (extracellular-signal-regulated kinase) pathway is a MAPK pathway regulated in part by RKIP. In this pathway the MAPK is ERK, the MAPKK is
MEK1/2, and the MAPKKK is Raf. Raf (Raf-1) is recruited to the plasma membrane by
Ras, a small GTP binding protein that is activated by the exchange of GDP with GTP.
The son of sevenless (SOS) Ras exchange factor, which is a guanine nucleotide exchange factor (GEF), is associated with growth factor receptor binding protein 2 (Grb2). This adapter protein links the GEF to phosphorylated tyrosine kinase receptors to send the cell surface signal down to the nucleus (Gomperts et al., 2002; Hilger et al., 2002; Kolch,
2000; Odabaei et al., 2004; Pearson et al., 2001).
As inferred by its name, RKIP inhibits the Raf/MEK/ERK cascade. Identified as a Raf-1 interacting protein in a yeast two-hybrid screen, RKIP was found to inhibit 7 phosphorylation and activation of MEK by Raf-1 (Yeung et al., 2000). Yeung et al.
suggested RKIP disrupts the cascade by binding to complexed Raf or MEK, not as a
substrate of either kinase, but by acting as a competitive inhibitor. The binding of either
Raf-1 or MEK to RKIP is mutually exclusive, as their binding domains on RKIP overlap
(Yeung et al., 2000). The over expression of RKIP leads to decreased activation of MEK by Raf-1 and the reduction of Raf-1 mediated proliferation and transformation. In contrast, the downregulation of RKIP increases MEK and ERK phosphorylation and AP-
1 dependant transcription (Yeung et al., 2000; Yeung et al., 1999).
Further studies have been conducted on the nature of the RKIP/Raf-1 interaction.
The role of RKIP in signal transduction extends into protein kinase C (PKC) and feedback inhibition for G protein-coupled receptors (GPCRs). Injection of human RKIP into Xenopus laevis oocytes led to an increase in G protein-mediated signaling (Kroslak et al., 2001). Additionally, Kroslak et al. observed an increase in G-protein –mediated inhibition of adenylate cyclase activity (reduction of cAMP) in mouse NIH3T3 cells stably transfected with human RKIP and stimulated with forskolin (Kroslak et al., 2001).
Later, it was discovered that classical and atypical PKCs phosphorylate RKIP at serine
153 in response to epidermal growth factor or 12-O-tetradecanoylphorbol-13-acetate
(TPA) in vitro(Corbit et al., 2003). RKIP phosphorylation at serine 153 causes
dissociation of the Raf-1 kinase domain and RKIP, indicating that PKC can mediate ERK
activation through RKIP (Corbit et al., 2003). PKC activator can upregulate ERK
signaling by phosphorylating RKIP thereby releasing it from Raf-1. The G-protein-
coupled receptor kinase 2 (GRK-2) is in a class of enzymes that phosphorylates seven 8 transmembrane proteins. The major mammalian GRK responsible for phosphorylating
activated receptors and thereby uncoupling them from G proteins and initiating their
internalization is GRK-2. Lorenz et al. demonstrated that RKIP acts as a signal modifier
between Raf-1 and GRK-2 pathways (Lorenz et al., 2003). PKC phosphorylation of
RKIP following GPCR stimulation causes its release from Raf-1 (Corbit et al., 2003;
Lorenz et al., 2003). Once free from Raf-1, RKIP was shown to bind GRK-2 and block
its activity, promoting and enhancing G protein signaling and MEK/ERK signaling
(Figure 1) (Lorenz et al., 2003). Very recently, Huang et al. provided evidence that PKC switches RKIP from Raf-1 to GRK-2 in gastrointestinal smooth muscle cells (Huang et al., 2007). Activation of Gq-coupled receptor through the muscarinic M3 receptor resulted in PKC stimulation and RKIP phosphorylation., then GRK-2 inhibition. The phosphorylation, internalization, and desensitization of VPAC2 receptors, which are
exclusively phosphorylated by GRK-2, were inhibited following this PKC stimulation.
Huang et al. concluded that the desensitization of VPAC2 receptor involved the PKC-
dependent switching of RKIP from Raf-1 to GRK-2 and repression of GRK-2 activity
(Altuwaijri et al., 2003; Huang et al., 2007).
The relationship between RKIP and its substrate Raf-1 has been investigated recently. Several studies, including ours, have shown that RKIP inhibits the
phosphorylation of the N-region of Raf-1 by (p21-activated kinase) Pak and Src family
kinases thereby inhibiting activation of Raf-1 (Park et al., 2006; Trakul et al., 2005).
RKIP blocked Pak phosphorylation of serine338 and Src phosphorylation of tyrosine341
of Raf-1 in a dose dependent manner (Trakul et al., 2005). However, our studies showed 9 that RKIP had no effect on the phosphorylation of a Raf-1 fragment (amino acids 325-
349) by PAK and v-Src in vitro (Park et al., 2006). Trakul et al. also depleted RKIP in cells and stimulated them with epidermal growth factor showing that RKIP could
influence the amplitude and dose response of ERK to epidermal growth factor (Trakul et
al., 2005). We fine-mapped the RKIP binding region to the catalytic domain of Raf-1 containing the N-region phosphorylation sites. We also show regions in the CR1 and
CR2 domains of Raf-1 may bind RKIP (Park et al., 2006). We propose that RKIP re- binding to Raf-1 after dissociation involves the phosphorylation of the Raf-1 N-region, although others suggest that N-region phosphorylation is more of a dissociation signal
(Corbit et al., 2003; Park et al., 2006).
RKIP in the NF-κB signaling pathway
Another central signaling pathway interrupted by RKIP is the nuclear factor
kappa B (NF-κB) signaling pathway (Figure 1). The transcriptional regulator NF-κB is a
dimer made up of members of the Rel family. Cytoplasmic NF-κB is bound to inhibitors
of κB (IκB) in an inactive state until IκB is phosphorylated by an IκB kinase (IKK)
complex. The phosphorylated IκB is ubiquitinated and subsequently degraded by the
proteosome. NF-κB then translocates to the nucleus and promotes transcription of
cytokines and anti-apoptosis genes (Gomperts et al., 2002; Odabaei et al., 2004). TNFα
and IL-1β can trigger the NF-κB signaling pathway, however, RKIP has been shown to
downregulate this pathway in response to these activators. NF-κB –inducing kinase
10 (NIK) and transforming growth factor beta-activated kinase 1 (TAK1) are MAPKKK family members involved in the phosphorylation and activation of IKKs (Gomperts et al.,
2002; Malinin et al., 1997; Ninomiya-Tsuji et al., 1999). RKIP has been shown to associate with NIK and TAK1 to decrease the activation of the IKK complex, thereby inhibiting the NF-κB pathway. RKIP was also demonstrated to interact with the α and β subunits of the IKK complex and interfere with IKK activity (Yeung et al., 2001).
11 A B
GPCR GPCR RTK RTK EGFR EGFR Ras grb2 sos Ras GRK-2 Ras grb2 sos Ras GDP GTP GDP GTP
PKC PKC Raf-1 RKIP Raf-1
(effectors) RKIP
MEK1/2 MEK1/2 GRK-2
ERK1/2 ERK1/2
ERK1/2 Nucleus ERK1/2 Nucleus
Transcription Factor
C D
NIK/TAK1 NIK/TAK1
RKIP IKK IKK
IκB IκB
NF-κB NF-κB
Nucleus Nucleus
NF-κB
Figure 1 - Proposed models of RKIP action in signaling pathways. A) RKIP phosphorylated by PKC inhibits GRK-2 and allowing GPCR to phosphorylate its targets. B) Ras/Raf/MEK signaling is inhibited when unphosphorylated RKIP binds Raf. GRK-2 binds and inhibits GPCR. C) NF-κB signaling pathways without and D) with RKIP inhibiting the NF-κB pathway through interaction with NIK, TAK1, and IKK.
12 RKIP has serine protease activity.
RKIP was detected as a novel thrombin inhibitor in the mouse brain (Hengst et
al., 2001). Purified RKIP was found to inhibit the serine proteases thrombin,
chymotrypsin, and neuropsin. This group attributed serine protease inhibition properties
to mPEBP1 and speculated that PEBP (RKIP) could be important in the outgrowth and
maintenance of neuronal processes (Hengst et al., 2001).
The ability of RKIP to act as a serine protease was tested after looking for a relationship between RKIP and calpain (Chen et al., 2006b). RKIP was assayed for its ability to inhibit the chymotrypsin-like activity of the proteasome in vitro. Proteasome activity was inhibited by high levels of RKIP (Chen et al., 2006b). This inhibition was similar to that observed with synuclein and huntingtin, which were previously shown to be proteasome inhibitors, indicating the RKIP may be a regulator of the proteasome
(Bence et al., 2001; Chen et al., 2006b; Snyder et al., 2003). Chen et al. arrived at this inhibitory action for RKIP while exploring its identity as a calpain substrate in vitro and in situ. It is interesting that neuroblastoma cells treated with Calpain inhibitor 4 resulted in increased protein levels of RKIP (Bian et al., 2004).
RKIP knockout mouse phenotype
These studies indicate that RKIP is an important player in multiple signaling
pathways. An understanding of these systems provides clues about the function of RKIP at the organismal level. RKIP protein knockout mice were very recently generated using
ES cells carrying a gene trap in RKIP intron 1 (Theroux et al., 2007). This gene trap 13 insertion as such may allow for the expression and processing of exon 1, which may
generate a functional hippocampal cholinergic neurostimulatory peptide (HCNP) which
will be discussed later in this chapter. RKIP protein knockout mice were found to be
viable and appeared normal up to ten months of age. The expression of RKIP was found
to be highest in the brain and testes. RKIP expression in the brain was determined by
histochemical detection of β-galactosidase activity as the gene trap contained β- galactosidase. Theroux et al. give a detailed account of brain staining, and overall strong expression was found in limbic formations (Theroux et al., 2007). The mice exhibited depressive behavior and had olfaction deficiencies measurable by four months of age.
The behavioral abnormalities of RKIP protein knockout mice were suggested to be associated with a general learning deficit (Theroux et al., 2007). The RKIP protein knockout mouse is also currently under investigation in our laboratory.
RKIP in neoplastic disease
RKIP and chemosensitization of cancer cells to apoptosis
RKIP expression levels were modulated in breast and prostate cancer cell lines to
demonstrate how chemotherapeutic agents may activate apoptosis using RKIP. RKIP
expression was inversely correlated with tumorigenicity of prostate cancer cell lines.
PC3 and DU145 cells express low levels of RKIP and are tumorigenic in nude mice,
whereas the cell line LNCaP has high levels of RKIP and is non-tumorigenic. Treating
PC3 and DU145 cells with the DNA-damaging agent 9-nitrocamptothecin (9NC) resulted
14 in increased RKIP levels and cell death. Apoptosis could also be triggered in 9NC-
resistant RC1 prostate cell line by over expressing RKIP. When RKIP was knocked
down in DU145 cells 9NC-induced apoptosis was blocked. The low-RKIP-expressing
breast cancer cells 578T and MCF7 did not undergo apoptosis following 9NC treatment,
as RKIP levels did not change, but Taxol treatment did induce RKIP expression and
apoptosis. The breast cancer cells underwent apoptosis following ectopic RKIP addition.
We postulated that RKIP was a novel effector of apoptosis signaling pathways
(Chatterjee et al., 2004).
The potential clinical importance of RKIP was investigated in non-Hodgkin’s
lymphoma (NHL B) cell lines. The chimeric mouse antihuman CD20 monoclonal
antibody rituximab can sensitize non-Hodgkin’s lymphoma cell lines (NHL B) to
apoptosis following paclitaxel treatment. In this study, RKIP expression was upregulated
following rituximab treatment of BHL B cells. Treatment also reduced activity of the
Raf/MEK/ERK pathway and the anti-apoptotic molecule Bcl-2, and sensitized them to
paclitaxel-induced apoptosis (Jazirehi et al., 2004).
RKIP as a metastasis suppressor
Loss of RKIP expression has been associated with cancer metastasis in various
tissues, including the prostate, breast, skin, pancreas, and liver. Initial investigation of
RKIP’s role in cancer metastasis was studied in a prostate cancer mouse xenograft model by Fu et al. (Fu et al., 2003). This investigation examined RKIP expression in a non- metastatic LNCaP cell line and the C4-2B metastatic cell line derived from LNCaP (Fu et 15 al., 2006; Fu et al., 2003). C4-2B cells had substantially less RKIP protein and mRNA
than the LNCaP cells. Modulation of RKIP by over expression in C4-2B cells and
antisense RKIP in LNCaP cells did not change cell proliferation rates, but did result in
invasion differences. In vitro invasion was decreased in C4-2B cells with over expressed
RKIP and increased in LNCaP cells with reduced RKIP. Furthermore, reduced invasion
in C4-2B cells correlated with Raf-1 inhibition, and vice-versa in the LNCaP cells. C4-
2B cells with control or RKIP expression vector were transplanted into mice prostates.
Following injection, primary tumor sizes were similar, but the mice injected with RKIP-
over expressing C4-2B had 85% fewer lung metastases. RKIP was not detectable in the
lung metastases as examined by immunohistochemistry (Fu et al., 2003). The primary
orthotopic tumors of mice injected with C4-2B cells had an 86% decrease in vascular
invasion (number of vessels) if they expressed higher levels of RKIP. The number of
blood vessels found in C4-2B-injected mouse tumors was also decreased if the tumor was
derived from C4-2B-RKIP injection (Fu et al., 2003). RKIP was reduced or non-
detectable in human samples as well. In primary prostate tumors RKIP expression was
reduced compared to normal tissue (Chatterjee et al., 2004). RKIP expression levels
were found to be highest in benign prostate tissues, reduced in cancerous tissue, and weak to non-detectable in metastatic tissue samples (Fu et al., 2006; Fu et al., 2003). Statistical analysis of a tissue microarray not only determined RKIP expression decreased significantly with prostate cancer progression, but that the time to PSA recurrence was significantly delayed in patients with moderate to high RKIP expression in primary tumors versus those with low or negative RKIP in the primary tumor (Fu et al., 2006). 16 These authors concluded that RKIP may be an independent prognostic factor in prostate
cancer (Fu et al., 2006).
RKIP expression level changes were also investigated in melanoma. We have
found the levels of RKIP expression are low in nine different melanoma cancer cell lines
derived from primary and metastasized tumors compared with primary melanocytes.
Moreover, the ectopic expression of RKIP in the melanoma cell lines partially reverted
the oncogenic B-Raf transformed phenotype (Park et al., 2005). Schuierer et al. have also
described RKIP downregulation in several melanoma cell lines compared with normal
melanocytes, and demonstrated diminished or eliminated RKIP in primary malignant
melanoma and metastases. Similar to the prostate cancer findings, melanoma cells over
expressing RKIP had reduced invasive potential (Schuierer et al., 2004). RKIP was
shown to specifically inhibit signaling initiated by B-Raf and could inhibit the B-Raf-
mediated neurite outgrowth of PC12 cells (Park et al., 2005). (Schuierer et al., 2006b)
Schuierer et al. speculated that decreased RKIP expression may predispose
primary human hepatocytes to malignant transformation, and showed that the
downregulation of RKIP may be important in hepatocellular carcinoma progression
(Schuierer et al., 2006a). In hepatocellular carcinoma cells lines there is decreased RKIP mRNA and protein compared to primary hepatocytes (Schuierer et al., 2006a). Lee at al found decreased RKIP expression correlated to differentiation level of hepatocellular carcinoma cell lines (Lee et al., 2006). Normal liver tissue has higher RKIP mRNA and protein levels than hepatocellular carcinoma tumors, also (Schuierer et al., 2006a). Both the Schuierer and Lee studies found that low RKIP levels were correlated with enhanced 17 ERK ½ signaling activity in hepatocellular carcinoma cell lines, and that indeed
modulating RKIP expression by over expression or knockdown resulted in modulated
ERK phosphorylation (Lee et al., 2006; Schuierer et al., 2006a). Importantly, RKIP has
not been shown to harbor mutations that affect its expression. With this clue, Scuierer et
al. hypothesized that RKIP reduction may be due to promoter methylation; however, 5- azacytidine treatment (demethylating agent) did not increase RKIP mRNA levels
(Schuierer et al., 2006a).
Breast cancer metastasis has also been correlated with RKIP expression.
Immunohistochemical examination of breast cancer lymph node metastases showed
significant loss of RKIP protein expression compared to normal breast duct epithelia and
primary tumors (Hagan et al., 2005). Interestingly, there was a weak negative correlation
between RKIP expression and apoptosis in breast tumors that did not have associated
lymph node metastases (Hagan et al., 2005).
Promoter methylation of RKIP was also studied in regards to colorectal cancer.
Minoo et al. reported methylation of the RKIP promoter in normal colonic mucosa and in
patients with hyperplastic polyposis (Minoo et al., 2006). Upon further investigation using colorectal cancer samples, the RKIP promoter was not methylated (Minoo et al.,
2007). The loss of RKIP in colorectal cancer samples of patients was associated with
tumor progression and distant metastasis, as well as poor survival (Minoo et al., 2007).
Another study found that RKIP was expressed in differentiating keratinocytes in
the spinous and granulous layer of epidermis, but not the basal layer (Yamazaki et al.,
2004). Differentiation induced by calcium treatment of normal human epidermal 18 keratinocytes correlated to increases RKIP expression, and the over expression of RKIP
in another keratinocyte cell line induced differentiation. Raf-1 signaling was inhibited
following RKIP expression in keratinocytes also (Yamazaki et al., 2004)
RKIP has been studied for its role in the pancreas. It may play a role in islet
neoplasia as RKIP expression was downregulated in human insulinomas. MAPK signaling was also examined in beta cells; RKIP could block MEK and ERK activation by Raf-1. Beta-cell proliferation was shown to be inhibited by RKIP, also (Zhang et al.,
2004).
RKIP is multifunctional.
The function of phosphatidylethanolamine-binding proteins has been deliberated
since their discovery in the 1980s. One of the interesting properties of some PEBP
family members, including hRKIP, rPEBP3, and mPEBP1, is that they are cleaved at the
N-terminus to release an undecapeptide which has been named hippocampal cholinergic
neurostimulating peptide (HCNP) (Ojika et al., 1992; Tohdoh et al., 1995). As indicated
by its name, HCNP seems to be involved in neuronal differentiation in medial septal
explant culture by enhancing production of choline acetyltransferase (ChAT), and seems
to act with NGF to regulate cholinergic phenotype development in medial septal nuclease
culture. RKIP and HCNP were found in the matrix of the secretary granules of bovine
chromaffin cells and were secreted into the blood with catecholamines (Goumon et al.,
2004). RKIP was also found to be secreted by platelets upon stimulation with thrombin 19 receptor-activating peptide. Goumon et al. postulated that RKIP could be secreted via a
non-classical pathway as it does not contain a signal (Goumon et al., 2004). HCNP can
act on frog cardiac mechanical performance, exerting a negative inotropism. Results of
these experiments suggest that RKIP/HCNP may be a new endocrine factor that regulates
cardiac physiology (Goumon et al., 2004). There is also some evidence that PEBP
downregulation may be associated with the congenital heart disease manifested in Down
syndrome (Kazuki et al., 2004). Recently, RKIP downregulation was found in the rat
right ventricle and in the interventricular septum upon cardiac remodeling (Melle et al.,
2006) Cardiac remodeling comprises scarred myocardia that does not contract, which
leads to secondary volume overload, then to hypertrophy of the left ventricle, then to
ventricular enlargement and heart failure (Melle et al., 2006)(Pfeffer Circulation 81
1161). The implications of this decrease in RKIP in heart disease are not yet known.
Another activity of RKIP has been explored in macrophage differentiation. RKIP was a strongly induced gene in mature macrophages as compared to freshly isolated monocytes as detected in a DNA microarray (Schuierer et al., 2006b). Furthermore,
macrophage maturation markers were increased following RKIP transfection in THP-1
myeloid cells. In this system, NF-κB signaling, not Raf-1/MEK/ERK signaling, was
found to be inhibited by RKIP to contribute to monocyte differentiation (Schuierer et al.,
2006b).
Mitotic progression may be under RKIP influence as well. Recent experiments
suggest that RKIP regulates the chromosomal passenger protein Aurora B via the Raf-
1/MEK/ERK signaling cascade (Eves et al., 2006). Aurora B controls the interactions of 20 microtubules with kinetechores to maintain integrity of the spindle assembly checkpoint
[reviewed in (Rosner, 2007)]. Depletion of RKIP during M phase caused inhibition of
Aurora B kinase activity and led to the spindle checkpoint being overridden, with
subsequent increase in chromosomal defects (after Taxol treatment). Raf-1
hyperactivation produced similar effects as RKIP depletion. Both phosphorylated RKIP
and activated Raf-1 were localized with kinetechores in mitotic cells (Eves et al., 2006).
This work suggests a potentially important new facet of RKIP function.
The prostate cancer cell lines PC3, DU145, and LNCaP contained activators of
OAS (2’,5’-oligoadenylate synthetase) not present in normal prostate epithelial cells
(Molinaro et al., 2006). OAS produces 2-5A (5’-phosphorylated, 2’,5’-linked
oligoadenylate) which binds to RNase L causing RNase L to dimerize into an
endoribonuclease [see (Molinaro et al., 2006)]. RKIP was one RNA identified that could activate OAS. Molinaro et al. speculated that the RKIP may activate OAS in vivo in prostate cancer cells to activate the antitumor functions of RNase L, although RKIP did
not seem to be involved in OAS activation in their prostate cancer cell lines (Molinaro et
al., 2006).
PEBP has been found in the male reproductive tract with implications in the
organization of sperm membranes during spermiogenesis (Frayne et al., 1998). Recently,
Nixon et al. identified RKIP as a decapacitation factor in mouse spermatozoa (Nixon et
al., 2006). RKIP and other proteins inhibited progesterone-induced acrosome reaction
and zona pellucida binding of sperm. RKIP alone was able to significantly reduce the
21 expression of phosphotyrosine residues within the sperm membrane. The mechanism by
which RKIP inhibits capacitation is unknown at this time (Nixon et al., 2006).
The evidence suggesting RKIP is involved in tumor metastasis is growing rapidly.
How RKIP is downregulated in cancer remains to be determined. The regulation of
RKIP is a promising frontier of research. It does not seem that RKIP protein degradation
or mutation is a means of downregulation. The transcriptional regulation of RKIP is therefore an appropriate place to start looking for mechanisms of regulation. The Snail transcription factor family has been extensively studied for its involvement in the repression of epithelial markers and induction of epithelial-to-mesenchymal transition which may lead to metastasis. This family of transcription factors is therefore worth investigating as potential regulators of RKIP.
The Snail family of transcription factors: general information
Snail (Snai1, Snail1) is a member of the Snail superfamily of zinc finger
transcription factors. Other Snail family members in humans include Slug (Snai2) and
Smuc (Snai3) Dozens of other members have been discovered in organisms represented
in vertebrates, protochordates, insects, and molluscs [reviewed in (Barrallo-Gimeno and
Nieto, 2005; Nieto, 2002)]. The Snail family has been extensively studied in various
organisms, especially Drosophila melanogaster where it was first discovered, where the
Snail family plays important roles in normal morphological development during
22 embryogenesis (Grau et al., 1984). In adulthood they are involved in normal processes
such as vascularization and wound healing (Peinado et al., 2004b), (Cano et al., 2000)
(Savagner et al., 2005).
These transcription factors are characterized as having a conserved carboxy-
terminal region containing four to six C2H2-type zinc finger repeats (Knight and
Shimeld, 2001). The fingers function as sequence specific DNA-binding motifs and are stabilized by a zinc ion binding to the conserved cysteine and histidine residues. Many also have a Snail/Gfi (SNAG) domain that enhances repressor activity in mammalian cells (Grimes et al., 1996; Nakayama et al., 1998). Peinado et al. demonstrated that the
SNAG domain is essential for recruitment of histone deacetylase 1 (HDAC1) and
HDAC2, and the corepressor mSin3A (Peinado et al., 2004a). The Snail family binds to a sequence called the E-box, with the consensus of CANNTG, which is also the consensus binding site for basic helix-loop-helix (bHLH) transcription factors (Mauhin et al., 1993; Nakayama et al., 1998). After binding to the E-box, Snail acts as a
transcriptional repressor; there is no evidence to date of Snail having activator properties.
The Snail superfamily is involved in normal embryonic development, cell
division, and cell survival. Their best known role is in cell movement and the induction
of the epithelial-to-mesenchymal transition (EMT). As the name suggests, the epithelial- to-mesenchymal transition involves phenotypic changes in cells with epithelial characteristics to more cells with more migratory properties typical of mesenchymal cells. Abnormalities in the EMT program has been shown to drive cancer cell invasion and metastasis (Thiery, 2002). As reviewed in (Barrallo-Gimeno and Nieto, 2005; De 23 Craene et al., 2005; Nieto, 2002), the Snail family represses epithelial markers and leads to the activation of markers of mesenchymal phenotype and cell survival (anti-apoptosis).
Importantly, Snail expression in stomach, liver, colon, ovarian, and breast cancer cell lines has been shown to induce an invasive, mesenchymal phenotype and correlated with
E-cadherin downregulation (Barrallo-Gimeno and Nieto, 2005; Peinado et al., 2004b;
Thiery and Sleeman, 2006).
The Snail superfamily has many downstream targets.
Snail and EMT
The best known target of Snail is E-cadherin. E-cadherin is central in maintaining the cell-cell adhesion of epithelial cells. Snail directly binds the E-cadherin promoter to strongly repress transcription (Batlle et al., 2000; Cano et al., 2000). The E-cadherin promoter in Snail-expressing cells had deacetylated histones H3 and H4 and dimethylated
H3 at K9 which were critical for Snail repression (Peinado et al., 2004a). Snail was shown not only to directly bind the E-cadherin promoter, but to also recruit HDAC activity (Peinado et al., 2004a). Cano et al. (Batlle et al., 2000; Cano et al., 2000) demonstrated that over expression of Snail in MDCK cells, which have a epithelial phenotype, resulted in a dramatic reduction in E-cadherin and transition to a dedifferentiated fibroblastic phenotype (Cano et al., 2000). Snail-transfected MDCK cells also exhibited a highly migratory and invasive behavior compared to mock- transfected cells. This study also showed high E-cadherin expression in human and mouse cell lines of differentiated carcinomas with low/absent Snail, and low E-cadherin 24 in dedifferentiated metastatic cell lines with high Snail expression. The inverse
relationship between Snail and E-cadherin was established in hepatocellular carcinoma,
oral squamous cell carcinoma, melanoma, and breast cancer (Blanco et al., 2002; Jiao et
al., 2002; Poser et al., 2001; Yokoyama et al., 2001). Slug has also been shown to
downregulate E-cadherin in breast cancer, indeed it may be more responsible than Snail
for this downregulation (Hajra et al., 2002). Slug was also shown to induce EMT in
MDCK cells, similar to that seen by Snail over expression (Bolos et al., 2003). E-
cadherin can also be repressed by delta-EF1, SIP1, and the bHLH transcription factors
E12/E47 and Twist (Cano et al., 2000; Grooteclaes and Frisch, 2000; Yang et al., 2004).
Several epithelial markers have been tied to downregulation by the Snail
transcription factors. The epithelial proteins desmoplakin, cytokeratin-18, Muc-1, occludins, and claudin are also downregulated by Snail (Nieto, 2002; Olmeda et al.,
2007; Thiery, 2003; Thiery and Sleeman, 2006). The mesenchymal markers vimentin
and fibronectin may be indirectly upregulated by Snail (Nieto, 2002; Olmeda et al.,
2007). Snail can also mediate matrix metalloproteinases (MMP)-induced EMT (Olmeda
et al., 2007; Przybylo and Radisky, 2007). MMPs are involved in tissue remodeling,
tumor invasion, metastasis, and neoangiogenesis as they can cleave extracellular matrix
and, break down cell-cell and cell-matrix interactions, and activate growth factors
[reviewed in (Przybylo and Radisky, 2007)].
25 Snail and Slug promote resistance to apoptosis.
Kajita et al. demonstrated that aberrant expression of Snail or Slug increased
resistance to apoptosis in breast carcinoma cells (Kajita et al., 2004). Snail or Slug
expression prevented cell death in MCF7 cells treated with a DNA-damaging drug. The
tumor suppressor gene TP53 promoter contains several E-boxes and its expression was
directly reduced with Snail or Slug expression. MDM2 transcript levels were also
increased in Slug-expressing cells and Snail-expressing cells following drug treatment,
which could increase p53 turnover. Furthermore, in this system ATM kinase was
significantly reduced, which could destabilize p53. p53 targets were also repressed
following Snail or Slug expression. P53 nonresponsive proapoptosis genes were also
repressed, and combined with the aforementioned properties resulted in resistance
chemotherapeutic-drug induced apoptosis (Fujita et al., 2003).
Regulation of Snail
There are numerous signaling cascades leading to Snail or Slug expression. Both
epidermal growth factor (EGF) and fibroblast growth factor (FGF) upregulate Snail
(Barrallo-Gimeno and Nieto, 2005; Mann et al., 2006). Ras/Raf-1/MEK/ERK signaling
also activates Snail transcription through AP-1 (Peinado et al., 2003). TGF-β1 and TGF-
β2 have been shown to upregulate Snail in hepatocytes and Slug in heart development
EMT, respectively (Romano and Runyan, 2000; Spagnoli et al., 2000). TGF-β can induce Snail in hepatocytes, epithelial, and mesothelial cells through activation of the
SMAD pathway (Takano et al., 2007)[reviewed in (Przybylo and Radisky, 2007; Thiery 26 and Sleeman, 2006)]. Cho et al. show that the Snail is activated by TGF-β through
Smad2, Smad3, and Smad4. They also suggest that Snail acts upstream of Akt in TGF-β- induced EMT (Cho et al., 2007). Bone morphogenetic proteins (BMPs) have also been implicated in Slug induction [reviewed in (Nieto, 2002)]. EGF receptor signaling can activate Stat3 , which can activate Liv1, leading to Snail nuclear translocation [reviewed in (Nieto, 2002; Przybylo and Radisky, 2007)]. Pak1 can also promote Snail nuclear translocation by phosphorylating Snail at serine246 [reviewed in (Przybylo and Radisky,
2007)], (Yang et al., 2005b).
Barbera et al. showed Snail promoter upregulation through ERK signaling and also NF-κB/p65 (Barbera et al., 2004). Additionally, insulin-like growth factor receptor signaling to activate NF-κB was shown to induce Snail and EMT in immortalized mammary epithelial cells (Kim et al., 2007).
Wnt signaling also regulates Snail. Yook et al. demonstrated that Wnt signaling
increases Snail protein levels (Yook et al., 2005). The β-catenin-like canonical motifs of
Snail provide for GSK3β-dependent phosphorylation, β-TrCP-directed ubiquitination,
and proteasomal degradation, but Wnt signaling through the Frizzled receptor
downregulates GSK3β (Yook et al., 2005). GSK3β downregulation allows for the
stabilization and nuclear accumulation of Snail as well as β-catenin (also phosphorylated
by GSK3β) (Katoh and Katoh, 2006; Yook et al., 2005). Recently, Yook et al.
demonstrated that the β-catenin-TCF cascade led to Axin2 induction, which regulates
nuclear GSK3β activity (Yook et al., 2006). FGF receptor signaling through PI3K-AKT
may also downregulate GSK3β activity [reviewed in (Katoh and Katoh, 2006)]. Besides 27 GSK3β, another negative regulator of Snail is Snail itself. The Snail promoter contains an E-box which can bind Snail, providing a feedback mechanism for Snail transcription
(Peiro et al., 2006).
Little is known about Snail regulation in prostate cancer in particular. One recent
study did observe the upregulation of Slug by dihydrotestosterone and EGF in LNCaP
cells (Chen et al., 2006a). Prostasin is a serine protease normally secreted in seminal
fluid, but whose expression is lost in advanced prostate cancer (CHEN PAPER and
other). Chen et al. discovered that prostasin is regulated by sterol-regulatory element-
binding proteins and Slug; SREBP-1c and SREBP-2 could upregulate prostasin and both
Slug and Snail repressed prostasin (CHEN). That androgens upregulate SREBP-1c and
SREBP-2 and also indirectly support an increase in mature SREBPs make the regulatory
story of prostasin very interesting.
28 MATERIALS AND METHODS
Cell culture.
The cell lines DU145, PC3, and LNCaP were cultured in RPMI 1640 (CellGro)
supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin solution (Gibco).
MCF7 cells were cultured in DMEM (GellGro) supplemented with 10% FBS (Gibco) and
1% penicillin/streptomycin solution. Cells were cultured at 37°C and 5% CO2. The
pSuper.retro.puro-siSnail vector used in stable transfection of PC3 and Du145 cells was a
gift from the laborator of Stephen Weiss (Yook et al., 2005). The si-Luciferase control vector was used previously (Mc Henry et al., 2008).
Western Blot
Cells were collected from cell culture plates by scraping them off in cold PBS.
Pelleted cells were lysed with TBST (20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA,
and 1% Triton X-100) for 15 minutes on ice and centrifuged for 10 minutes at 4˚C.
Lysate protein concentrations were measured by BCA assay (Pierce). Total cell lysates
(2-30μg) were mixed with 4x sample loading buffer, then loaded onto 12% SDS-
polyacrylamide gels. Gels were electrophoretically transferred onto PVDF membranes
(Millipore) and blocked in 5% non-fat milk solution (in PBS) one hour to overnight.
Membranes were washed in PBST (PBS and Tween) for 5 minutes then rotated with
primary antibody. All primary antibodies were diluted in primary antibody solution
(PBST, 5% bovine serum albumin, 0.002% sodium azide). Flag-tagged proteins were
detected using anti-FLAG, or M2 antibody (Sigma), rabbit anti-RKIP antibody was 29 diluted 1:2000 and mixed with the membrane for one hour at room temperature. Rabbit
anti-E-cadherin antibody (1:1000) (Cell Signaling) and mouse anti-α-tubulin antibody
(1:200) (Sigma) were incubated with rotation for one hour at room temperature. Snail
was detected by iincubating the blot overnight at 4˚C in rabbit anti-Snail (1:1000)
(Abcam). Following three washes in PBST, blots were incubated with the appropriate
horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch
Laboratories, Inc.) in PBST with 5% milk solution for one hour at room temperature.
Blots were visualized using enhanced chemiluminescence and the Bio-Rad Gel Doc
system.
ChIP Assay
LNCaP cells stably infected with flag-Snail S6A were grown in normal culturing
conditions. The Snail-S6A vector was a gift from the laboratory of Stephen Weiss.
Cross-linking was done by adding formaldehyde directly to culture medium to a final
concentration of 1%. Cells were incubated at 37˚C for 15 minutes and the medium was then aspirated. Cells were washed twice with ice cold PBS containing protease inhibitors
(1mM phenylmethylsulfonyl fluoride, 1ug/ml aprotinin, 1ug/ml pepstatin A). Cells were scraped from plates in 2-3ml of PBS with protease inhibitors and pelleted by centrifugation for 4 minutes at 2000 rpm, 4˚C, and the PBS aspirated. Cell pellets were resuspended in 200 ul of Lysis Buffer (1% SDS, 10mM EDTA, 50mM Tris-HCl, pH 8.1) with protease inhibitors and incubated on ice for 10 minutes.
30 Lysates were sonicated on ice (Fisher Sonic Dismembrator, model 550) set to level 3 (10 second burst, 10 second rest done a total of 12 times) to shear DNA to lengths between 200 and 2000 base pairs. Sonicated lysates were centrifuged for 10 minutes at
13,000 rpm, 4˚C. Sonicated cell supernatants were transferred to fresh tubes and DNA concentration was approximated by measuring absorbance at OD260 in a spectrophotometer. Chromatin (100 μg, which was about 200 ul) was diluted 10-fold in
Dilution Buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl pH
8.1, 167 mM NaCl) with protease inhibitors to a total volume of 2 ml. 20 ul of diluted
lysates were removed for input analysis. 80 ul of a 50% protein G-agarose slurry was
washed in Dilution Buffer and then added to the lysates along with 200 ug of sheared salmon sperm DNA (Sigma). Lysates were pre-cleared by rotating for 30 minutes at 4˚C.
Lysates were centrifuged for one minute, 4000 rpm, 4˚C, and supernatants transferred to fresh tubes. Either 20ul of M2 anti-flag antibody covalently atached to protein G agarose beads (Sigma) or 20 ul protein G agarose beads with 2ug anti-HA antibody were added to each 2ml lysate aliquot and rotated overnight at 4˚C.
Beads were washed for 3 minutes at room temperature once in 1ml low salt wash buffer (0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris-HCl pH8.1, 150 mM
NaCl), followed by 1ml high salt wash buffer (0.1% SDS, 1% Triton X-100, 2mM
EDTA, 20mM Tris-HCl pH 8.1, 500mM NaCl) and then by a 1ml LiCl Wash Buffer
(0.25M LiCl, 1% IGEPAL-CA630, 1% deoxycholic acid (sodium salt), 1mM EDTA,
10mM Tris pH 8.1). Beads were then washed twice, 3 minutes each time, at room temperature in 1ml 1X TE. Precipitates were eluted in 500μ freshly prepared elution 31 buffer (1% SDS, 100mM NaHCO3), and crosslinks reversed in both precipitates and
inputs by heating the beads/ precipitate complex in elution buffer with 0.2M NaCl at
65˚C overnight with occasional vortexing. Eluates were digested with 20μg Proteinase K
(with 10mM EDTA, 40mM Tris-HCl pH 6.5) for one hour at 45˚C. DNA was recovered
by phenol/chloroform extraction and ethanol precipitation with addition of glycogen as an inert carrier. Pellets were washed in 70% ethanol and air dried before being resuspended in 55 μl of DNase/RNase free water.
Semi-quantitative PCR
Extracted DNA was amplified using Platinum taq polymerase (Invitrogen) and
visualized by ethidium bromide-stained 2% agarose gels. PCR conditions included a mix
of 1x Platinum taq buffer (Invitrogen), 5 mM MgCl, 0.2mM dNTPs, 0.4μM primers each,
9μl DNA, Platinum taq polymerase (Invitrogen), and water to a total volume of 25μl.
PCR was performed in an Eppendorf Mastercycler Gradient thermocycler with the
following conditions: 95˚C for 5 minutes fand then 30 cycles of 95˚C for 30 seconds,
55˚C for 30 seconds, and 72˚C for 1 minute followed by an extra extension step of 72 ˚C
for 5 minutes. Primers used for the E-cadherin E-box region was: FWD-5’-
AGGTGAACCCTCAGCCAATCAG-3’, and REV-5’-
AAGCTCACAGGTGCTTTGCAGTT-3’. Primers used for E-cadherin non-consensus
region were FWD-5’- GGATAAGAAAGTGAGGTCGGAGGA -3’, and REV-5’-
CGTTCCCTTTCAGTCTCCTTTCTC -3. The E-cadherin primer sequences were shared
by Dr. Paul Wade at the NIH (Kajita et al., 2004). Primers used for RKIP E-box region 1 32 were FWD -5’- ACCAAAACGCAACCAAGGCG -3’, and REV-5’-
GAAGGGCGCCCTGGAATAGG -3’. Primers used for RKIP E-box regions 2, 3, and 4
were FWD -5’- TTGTGTGTGGAGAAAACGGAC -3’, and REV 5’ –
CCAGATTCTCTGTATTTTCGCACGTG- 3’. Primers used for the RKIP region not
containing an E-box consensus region were FWD,5’ –
GGTCAACATGAAGGGCAATGA 3’, and REV-5’- GCCCCGAGCCCACATAA - 3’.
Amplified products were between 75 and 175 base pairs long.
Reverse Transcription
RNA was collected by the Trizol method per the manufacturer’s protocol
(Invitrogen). Diluted total RNA was reverse transcribed in first strand cDNA using
MultiScribe reverse transcriptase (Applied Biosystems) in 1x Multiscribe reverse
transcription buffer (Applied Biosystems), 2mM dNTPs, (Invitrogen) 4 Units of RNAase
inhibitors (Applied Biosystems), 0.5 μg random hexanucleotides (Amersham), and
5.5mM MgCl (Sigma). Mixtures were incubated at 25˚C for 10 minutes, 48˚C for 30 minutes, and 95˚C for 5 minutes in a thermocycler (Eppendorf).
Real-Time PCR
Total RNA was reverse-transcribed into cDNA as described above. cDNAs were
combined with 2x Applied Biosystems SybrGreen Master Mix and 100-200 nM forward
and reverse primers. The primers used were for human sequences of actin, RKIP, Slug,
Snail, and E-cadherin, as listed in Table 1. 33 Name 5’ to 3’ sequence
Actin Forward TCATCACCATTGGCAATGAG
Actin Reverse CACTGTGTTGGCGTACAGGT
RKIP Forward TGATTCAGGGAAGCTCTACACCTT
RKIP Reverse TGTTGACCACCAGGAAATGATG
Snail Forward CCCAATCGGAAGCCTAACT
Snail Reverse GGTCGTAGGGCTGCTGGAA
Slug Forward AGAACTCACACGGGGGAGAAG
Slug Reverse CTCAGATTTGACCTGTCTGCAAA
E-cadherin Forward GAACAGCACGTACACAGCCCT
E-cadherin Reverse GCAGAAGTGTCCCTGTTCCAG
Table 1. Primer sequences used in real-time PCR.
Real-time PCR was performed in an Applied Biosystems 7500 Real Time PCR System and software with the following conditions: 95˚C for 15 minutes followed by 40 cycles of
95˚C for 15 seconds, 60˚C for 30 seconds, and 72˚C for 33 seconds. A dissociation curve was generated immediately following real-time PCR to detect nonspecific amplification.
Fold differences and standard deviations were calculated and plotted in Microsoft Excel.
34 Luciferase Reporter Plasmids
pGL2-E-cadherin wild type and mutated promoter reporter vectors, and also Snail and
Slug expression vectors, were gifts from E.R. Fearon of the University of Michigan
(Hajra et al., 2002). To construct the RKIP promoter reporter plasmids, PCR-amplified
products were ligated into the multiple cloning region of the GL2-Basic Vector
(Promega). A 2200 base pair RKIP promoter fragment was amplified from the Homo sapiens chromosome 12 BAC clone 385C6 (Roswell Park Cancer Institute) using the primers F,5’ – CCGGCACATCCCTCCATACC – 3’ and REV,5’ –
CGGGATCCGGCCAAGCAGAGCGTGCAGCCGGG -3’. The amplified product was ligated into pGL2-Basic to make pGL2-RKIP2.2-luciferase. The plasmid pGL2-
RKIP2.2-luciferase was used as a template to PCR amplify smaller promoter fragments using the same reverse primer (as used for pGL2-RKIP2.2-luciferase) and the following forward primers: 5’- TTGTGTGTGGAGAAAACGGAC -3’ (RKIP1.3), 5’-
CGGGGTACCGAGGCCGAGGTCAGGAGATC -3’ (RKIP0.9), and 5’-
CGGGGTACCGCCAGCTTCGGCCAATCAGAG -3’ (RKIP0.2). These PCR products were ligated into pGL2-Basic to create pGL2-RKIP1.3-luciferase, pGL2-RKIP0.9- luciferase, and pGL2-RKIP0.2-luciferase, respectively.
Transfection and Dual Luciferase Reporter Assay
MCF7 cells were normally cultured and split into 24 well plates. After 24 hours
cells were approximately 75% confluent, and cells were transfected with 1ug total DNA
35 per well including 0.15 μg pGL2-luciferase reporter vector (control), pGL2-E-cadherin
promoter-luciferase, pGL2-E-cadherin mutant promoter-luciferase, pGL2-RKIP1.3-
luciferase, pGL2-RKIP0.9-luciferase, or pGL2-RKIP0.2-luciferase. Cells were
cotransfected with 0.025 μg pRL-TK per well, which contains the herpes simplex virus
thymidine kinase (HSV-TK) promoter (Promega) to express Renilla luciferase as a
transfection efficiency control. 0.75 μg of Snail or Slug expression vector, and 0.075 μg
(with Snail or Slug cotransfection) or 0.825 μg (no Snail or Slug) of filler DNA (puc19)
were also cotransfected. Cells were transfected in serum-free DMEM media with the
Lipofectamine Plus Reagent (Invitrogen) per manufacturer’s instructions using 4μl Plus
Reagent and 2.75 μl Lipofectamine Reagent.. Transfection media was aspirated and
replaced by regular culturing media after 2 hours.
Forty-eight hours post-transfection, cells were gently rinsed with PBS and lysed
with Passive Lysis Buffer from the Dual-Luciferase Reporter 1000 Assay System
(Promega). The activities of both firefly and Renilla luciferases were measured in a
Turner Designs 20/20 Instrument using Dual-Luciferase 1000 Assay System reagents.
The ratio of firefly luciferase levels to internal controls, Renilla luciferase levels, was used to determine E-cadherin or RKIP promoter activity and expressed as relative
luciferase activity. The analysis of each promoter construct was performed in triplicate.
Correlation analyses and statistical tests
Pearson correlation coefficients and p-values were calculated for RKIP and Snail,
Snail and E-cadherin, RKIP and Slug, and RKIP and EZH2 using the statistical program 36 R version 2.4.1 (RDevelopmentCoreTeam, 2006). Box plots and Student’s T-tests were
also generated or performed by the program R. Scatterplots, trendlines, and regression
analyses were created in Excel (Microsoft).
Node correlation values were obtained from hierarchical cluster analysis as
described in the section “Hierarchical Cluster Analysis”. Following the cluster analysis, results were output in the program TreeView, also described in the “Hierarchical Cluster
Analysis” section.
Hierarchical Cluster Analysis
Array data was provided by the laboratory of Arul Chinnaiyan of the University
of Michigan. The data was in the form of normalized Cy3/Cy5 ratios of non-Affymetrix
cDNA. The samples had been categorized as normal prostate tissue, tissue of benign
prostatic hyperplasia samples, primary prostate cancer tumor samples, or samples from
metastasized prostate cancer samples [see (Dhanasekaran et al., 2001; Fu et al., 2006;
Varambally et al., 2005) for examples and further explanation of sample groupling].
These ratios were log2-transformed and then correlated to RKIP across all samples or
across localized prostate tumor samples and metastatic tumors only using the the
statistical function for determining correlations coefficients in the spreadsheet program
Excel (Microsoft). The resulting correlations were sorted according to correlation values
(highest to lowest), again using Excel software, and the 8000 genes most closely
correlated to RKIP were uploaded into the hierarchical clustering program, Cluster3.0.
The 8000-gene cutoff point was used as this was the highest number of entries we could 37 analyze in our available computers using Cluster3. This was deemed acceptable as 8000 genes is still a very high number, and limitations on gene representation were larger from the selection available for use on teh microarray rather than at this point in the analysis.
Cluster and TreeView were created by Michael Eisen at the University of California at
Berkely to process and analyze large micrarray datasets (Eisen et al., 1998). Cluster3.0 was adapted from Eisen’s program by Michiel de Hoon of the University of Tokyo. Data was filtered in Cluster3.0 to eliminate genes with less than 80% expression across samples and adjusted to be centered around the median of genes. Average hierarchical clustering was used on gene- and array-centered data that was filtered and adjusted.
Cluster3.0-processed data was directly imported into TreeView to depict the cluster analyses as a heat map and dendrogram. The TreeView program allowed for the designation of coloration in the dendrograms. TreeView also allowed node correlation values to be viewed, although these numeric values were not reproduced in the dendrogram, the values are denoted by length of connecting lines only. The entire hierarchical clustering analysis was repeated using the same process to analyze Snail hierarchical clustering.
38 RESULTS
RKIP and Snail are differentially expressed in prostate cancer cell lines.
It has been shown previously that prostate cancer cell lines express RKIP to
different extents (Fu et al., 2006; Fu et al., 2003). Under normal culture conditions,
RKIP levels in the low metastatic LNCaP cells are higher than in both PC3 and Du145
cells, derived from bone and brain metastases of prostate adenocarcinoma (Figure 2).
Du145 cells had the lowest RKIP expression levels.
Figure 2 – Protein expression of endogenous RKIP, Snail, and E-cadherin in prostate cancer cell lines. Cells were cultured in normal media and harvested on ice by scraping. Whole cell lysates were run on SDS-PAGE and immuno-blotted with their respective antibodies. Both tubulin and nucleophosmin were used to compare equal loading.
There are many potential transcriptional regulators that could be responsible for
the repression of RKIP. As RKIP expression is clearly downregulated in metastases, we decided to investigate transcriptional repressors of genes relevant to metastasis. Snail
39 and Slug have been shown to downregulate epithelial markers such as occludins,
claudins, E-cadherin, and desmoplakin in tumor cells, but there is a dearth of information
regarding their role in prostate cancer (Barbera et al., 2004; Barrallo-Gimeno and Nieto,
2005; Batlle et al., 2000; Castro Alves et al., 2007; De Craene et al., 2005; Hajra et al.,
1999; Shih et al., 2005). We wanted to examine Snail and Slug expression in the prostate
cancer cell lines LNCaP, PC3, and Du145. Protein expression of Snail was lowest in
LNCaP and highest in Du145 by western blotting (Figure 2). We also observed the highest expression of E-cadherin in LNCaP cells and the lowest in Du145 (Figure 2).
Endogenous levels of RNA for RKIP, E-cadherin, and Snail as measured by real
time RT-PCR were similar to the protein levels (Figure 3), with RKIP and E-cadherin
RNA levels highest in the LNCaP cell line and lowest in DU145. Conversely, Snail was
high in DU145. PC3 E-cadherin levels were unexpectedly low, however, based on observed Snail levels. Slug is also a potent repressor of E-cadherin, and in several breast cancer cell lines Hajra et al. attribute E-cadherin downregulation to Slug primarily .
Western blotting for Slug was not feasible as we could not obtain a reputable anti-Slug antibody. We analyzed Slug levels by real-time quantitative real time RT-PCR (Figure
3d). The PC3 cell line had much higher levels of Slug than LNCaP or Du145, which may
be the reason E-cadherin is still low in PC3 cells.
40
Figure 3 – Endogenous levels of RKIP, Snail, E-cadherin, and Slug in prostate cancer cell lines. Cells were grown in normal culturing condition. Total RNA was extracted using the Trizol method. Following reverse transcription, cDNA levels were quantified in triplicate using SYBR-Green real time PCR. Real time Ct values were normalized to beta-actin expression and referenced to LNCaP expression (=1) to obtain fold difference values for (A) RKIP, (B) E-cadherin, (C) Snail, and (D) Slug. 41 RKIP expression is modulated by Snail in prostate cancer cell lines.
If Snail is responsible for repression of RKIP in prostate cancer cell lines, then
over expressing or reducing Snail expression could change RKIP expression levels. The
LNCaP cell line was stably transfected with either pBabe-flag-Snail, pBabe-flag-Snail
S6A, or empty vector. Snail S6A is a mutant of Snail that is unable to be phosphorylated by GSK3 (then ubiquitinated and degraded) and is subsequently more stable than wild type Snail (Yook et al., 2005). Snail protein expression was increased in the Snail- transfected cells compared to empty vector-transfected cells (Figure 4a). E-cadherin expression was high in the LNCaP cell line, but following Snail over expression it decreased. This downregulation is especially marked in the flag-Snail S6A-transfected cells, as this Snail was not as labile as the wild type. RKIP expression was also
decreased in the cells over expressing Snail, more markedly in the flag-Snail S6A-
transfected LNCaP cells (Figure 4a). Quantitative real time RT-PCR also revealed a
decrease in Snail-transfected cells compared with empty vector control- or flag-EZH2-
transfected LNCaP cells (Figure 4b). Transcript levels were decreased the furthest in
flag-Snail S6A-transfected cells, as was seen with the western blot (Figure 4).
42 A
B
Figure 4 – Snail over expression in LNCaP cells. (A) Western Blot. LNCaP cells stably transfected with either Snail expression vector, the mutated Snail (S6A) expression vector, or an empty vector control were grown in normal culture conditions. Whole cell lysates were subjected to SDS-PAGE and immunoblotted with anti-flag (M2) for Snail detection, anti-E-cadherin antibody, anti-tubulin antibody, and anti-RKIP antibody. (B) Reverse-transcription real-time PCR. The cells were harvested RNA extracted by the Trizol method. Following reverser transcription, cDNA were in triplicate analyzed using SYBR green real time RT-PCR. Real time PCR values were normalized to GAPDH and referenced to the empty vector control.
43 Snail was also downregulated using RNA interference. PC3 and DU145 cells, which had higher endogenous Snail levels than LNCaP (Figure 2), were stably
transfected with siRNA for Snail, or siRNA for luciferase as a control. Figure 5 shows
the knockdown of Snail and upregulation of RKIP in both cell lines by western blot.
Figure 5- Knockdown of Snail in prostate cancer cell lines. PC3 and DU145 cell lines were stably transfected with pSuper.retro.puro-siSnail or pSuper.retro.puro-siLuciferase (control). Following puromycin selection, cells were harvested by scraping and whole cell lysates subjected to SDS-PAGE. Western blotting was done as described in Materials and Methods with specific antibodies.
Snail and Slug repress RKIP promoter activity
Little is known about the RKIP transcriptional regulation. The promoter region of
RKIP does not have a well-defined TATA box or initiator (Odabaei et al., 2004). The
promoter is of the high CpG content type in that is has a CpG island of at least 500 base
pairs in length with 50% or higher GC content and 0.6 or higher observed CpG/ expected 44 CpG (Maglott et al., 2005; Saxonov et al., 2006). The CpG island is centered around the transcriptional start site (Maglott et al., 2005). The transcription factors AP1, SP1, and
YY1 have putative binding sites in the RKIP promoter as predicted by TESS master analysis (Odabaei et al., 2004). There are also three potential androgen receptor elements within 2200 base pairs of the transcriptional start site (Even Keller, personal communication). We searched the RKIP promoter for E-box elements and found many sections that fit the CANNTG consensus. There are 4 putative E-boxes within 1.3 kb of the transcriptional start site (Figure 6). We constructed luciferase reporter vectors of different lengths to include all the E-boxes with 1.3kb of the transcriptional start site
(RKIP1.3-Luc), the E-box closest to the transcriptional start site only (RKIP0.9-Luc), or to exclude all E-boxes (RKIP0.2-Luc).
Figure 6 – RKIP promoter with E-boxes. The RKIP promoter has four E-boxes with 1.3kb of the transcriptional start site. E-box 1 is 407 bp from the start site and E-boxes 2- 4 are 1203 to 1245 bp from the start site.
We tested the effectiveness of Snail and Slug to repress E-box activity by using luciferase reporters containing part of the E-cadherin promoter. The reporters had either a wild-type promoter fragment that included three E-boxes previously shown to be repressed by both Snail and Slug, or the same fragment with all three E-boxes mutated 45 such that neither Snail nor Slug could repress (Hajra et al., 2002). The MCF7 breast
cancer cell line was used in determining Slug and Snail effectiveness as this cell line was easily transfected. Subconfluent MCF7 cells were co-transfected with pGL2-RKIP-
Luciferase (Figure 7b) or pGL2-Ecadherin-luciferase (firefly or Photinus pyralis) reporter vectors, CMV-Snail or CMV-Slug, and a Renilla luciferase internal control reporter. Total transfected DNA amounts were equivalent. Snail or Slug repressed the wild type E-cadherin promoter; however, little effect was seen on the E-box-mutated E- cadherin promoter (Figure 7a). Both Snail and Slug also repressed luciferase activity
substantially in the RKIP promoters containing E-box elements. The 200 base pair
promoter fragment showed little change in expression (Figure 7a).
46 A
B
Figure 7 – Luciferase assay of Snail and Slug repression of RKIP. (A) MCF7 cells were transfected with firefly luciferase reporters pGL2-E-cadherin-luciferase, pGL2-mut- E-cadherin-luciferase, pGL2-RKIP1.3-luciferase, pGL2-RKIP0.9-luciferase, or pGL2RKIP0.2-luciferase; and cotransfected with either CMV-Snail, CMV-Slug, or puc19 (EVC), and also the Renilla control reporter pRL-TK. Each transfection was done in triplicate. 48 hours after transfection cells were lysed with Promega passive lysis buffer and centrifuged to pellet debris. Luciferase activity was measured using the Promega Dual Luciferase kit and instructions. Firefly/Renilla luciferase ratios were averaged for each sample group, and results are reported as percent of wild type E- cadherin reporter activity (wt Ecad-luc = 100%). pGL2-basic (empty vector) luciferase readings were nearly 0. (B) Schematic of the reporter constructs. The red bar represents an E-box location.
47 Snail binds the RKIP promoter
To see whether Snail may be binding the RKIP promoter, we performed
chromatin immunoprecipitation (ChIP) assays. The stable cell line LNCaP flag-Snail-
S6A over expresses a Snail variant that is more stable than wild-type. After cross-linking
with formaldehyde, lysates were sonicated to 200-2000 base pair fragments. Snail was
immunoprecipitated with anti-flag antibodies. As a negative control, lysates were also
immunoprecipitated with anti-HA antibody. The E-cadherin promoter was again used as
a control, as Kajita et al. detected the E-cadherin promoter using PCR primers flanking the E-box elements shown to bind Snail (Kajita et al., 2004). We designed primers flanking the RKIP promoter E-box1, the E-box cluster (E-box 2,3,4),and in a coding region of RKIP (exon3) that does not include E-box elements (see Figure 6).
Semi-quantitative PCR was used to visualize the results of the chromatin
immunoprecipitation. As shown in Figure 8, amplification occurred in the E-cadherin
and RKIP promoter regions containing E-boxes. The RKIP E-box located closest to the
transcriptional start site (Ebox1) amplified more strongly compared to the input than the
E-box “cluster” containing the evolutionarily conserved E-box (E-boxes 2,3,4) as well as
from the non-specific immunoprecipitation (anti-HA). The E-cadherin and RKIP non-E-
box flanking primers produced very little to no amplification. The data supports the
hypothesis that Snail associates with the RKIP promoter at the E-box nearest the
transcriptional start site.
48
Figure 8- ChIP assay and semi-quantitative PCR. LNCaP cells stably transfected with flag-Snail S6A lysed and sonicated as described in Materials and Methods. Pre-cleared lysates were incubated with M2-agarose beads or anti-HA antibody and agarose beads. Immunoprecipitated DNA was reverse crosslinked, treated with proteinase K, and precipitated. Semi-quantitative PCR was carried out on immunoprecipitated and input DNA using primers flanking E-boxes or in non-E-box regions, producing amplicons under 200 base pairs. The E-cad no E-box primers were located in intron 1 of the CDH1 gene.
RKIP and Snail are negatively correlated in human tumor metastases.
The findings in our prostate cancer cell line studies lead us to investigate whether
there was a physiological relationship between RKIP and Snail in prostate cancer
patients. One way to study the expression of thousands of genes in parallel is to use a
microarray. There are many published studies regarding aspects of prostate cancer using
microarray analysis, and some of these are available on public databases such as
Oncomine (www.Oncomine.org) and the Gene Expression Omnibus (GEO) of NCBI
(Bioscience, 2007; Rhodes et al., 2004). It is thus possible to interrogate the microarray
49 data of other laboratories for your gene of interest through these public databases and
possibly perform a meta-analysis.
The Oncomine database included dozens of studies including RKIP in human
prostate cancer. These studies included microarray experiments comparing RKIP levels
in different normal tissues, as well as comparing RKIP expression in prostate cancer with other cancers. Other studies examined gene expression differences in prostate cancer samples in various stages of growth or differentiation, of those with PSA recurrence, following neoadjuvant therapy, and between patients based on two year survival. We focused on those experiments regarding prostate cancer and metastasis. Ten of these studies included analyses of normal and/or benign prostatic hyperplasia, localized prostate cancer, and metastases of prostate cancer. Seven of these studies did not include
Snail. Of the three remaining studies, Varambally et al. dataset showed no significant difference (p<0.05) in RKIP RNA levels with Snail decreasing from normal to primary prostate cancer, then increasing from primary to hormone-refractory cancer, but the changes were not statistically significant (Bioscience, 2007; Varambally et al., 2005).
The Vanaja et al. dataset showed a significant decrease in RKIP levels from primary to metastatic cancer with Snail levels increasing, but the difference was not statistically significant (p<0.05) (Bioscience, 2007).
The Dhanasekaran et al. dataset described in Oncomine revealed an interesting
correlation between RKIP and Snail (Bioscience, 2007; Dhanasekaran et al., 2001).
There was a significant decrease in RKIP from normal to localized prostate cancer to
metastatic prostate cancer. As seen in the Vanaja and Varambally datasets, Snail levels 50 increased as RKIP decreased. Unlike those datasets this increase was significant
(p<0.05). E-cadherin expression in this dataset was decreased between localized cancer
samples and metastatic samples. Oncomine described this dataset graphically as
boxplots, lumping normal adjacent prostate and benign prostatic hyperplasia samples
together. Lists of correlations to any gene in the dataset and heat maps for the
correlations were provided by Oncomine, also. In order to determine if there was any
difference between NAP and BPH samples and in order to perform other analyses, we
downloaded the data used in Oncomine from Dr. Arul Chinnaiyan
(www.pathology.med.umich.edu/chinnaiyain/Nature/Nature.htm). This dataset included
RKIP, Snail, and Slug expression data in normal adjacent prostate tissue (NAP), benign
prostatic hyperplasia (BPH), localized prostate cancer (PCA), and metastatic cancer
samples (MET). The authors referenced the prostate samples against a commercial pool
of normal prostate samples and also against their own pool of normal adjacent prostate
tissues of prostate cancer patients. These sets of microarray data will be referred to as
Dataset 1a for samples referenced against the commercial pool and Dataset 1b for samples referenced against normal adjacent prostate tissue. The data provided were ratios of sample data to the reference. The ratios were derived from values already cleaned and normalized as published previously (Dhanasekaran et al., 2001). We log-transformed the ratios prior to comparing the distributions of the sample groups by box plots. Box plots enable the quick visualization for comparing the distributions of log ratios of genes from several sample groups or microarrays. Gene expression ratios of microarray data are often log-transformed (Stekel, 2003). Log-transformation gives a natural symmetry to 51 the data that is not seen in fold-difference representations. Log-transformation should yield a more even spread of ratios across the data range and the intensities should have a bell-shaped distribution (Stekel, 2003). For example, a 2-fold increase would have a log ratio of 1, no differential expression would have a log ratio of 0, and a 2-fold downregulation would have a log ration of -1. With the Dhanasekaran et al. datasets, the ratio of Cy5 and Cy3 intensities were transformed into the difference between the logarithms (base 2) of the Cy5/Cy3 intensities. Figure 9 shows the differences between the sample groups for RKIP, Snail, and E-cadherin as box plots following data processing. Figures 10 and 11 describe point-by point sample data for Snail and RKIP.
52
A
B
Figure 9- Box plots of RKIP, Snail, and E-cadherin levels in prostate cancer microarray datasets. The box represents the standard deviation of the distribution and the line through that box represents the mean of that distribution. The horizontal lines above and below the box represent the extreme values of the distribution and outliers are denoted with an asterisk. (A) Ratios of prostate cancer samples referenced against a commercial prostate pool of normal prostate samples were log-transformed and plotted using the statistical program R. Normal adjacent prostate (NAP) N=4, benign prostatic hyperplasia (BPH) N=14, localized prostate cancer (PCA) N=14, metastatic prostate cancer (MET) N=20 (B) Ratios of prostate cancer samples referenced against a pool of normal adjacent prostate samples from patients were log-transformed and plotted using the statistical program R. NAP N=3, BPH N=5, PCA N=10, and MET N=7.
53 ratios wereplottedforeachsample. Me Figure 10–RKIPandSnailexpression Normalized Expression Unit Expression Normalized tastatic samples weregroupedbylocation. in eachsample-Dataset1a.RKIPandSnail
Sample Number
54
Sample Number Sample
Normalized Expression Unit
Figure 11 - RKIP and Snail expression in each sample - Dataset 1b. RKIP and Snail ratios were plotted for each sample. Metastatic samples were grouped by location 55 The box plot data reveals an inverse relationship between RKIP and Snail expression. As discussed previously, the epithelial phenotype marker E-cadherin is downregulated by
Snail. E-cadherin expression decreases from PCA to MET samples, as would be expected from increased Snail expression.
The beauty of genomics is how genes can be seen to work together to generate a
phenotype, and the beauty of microarrays is that they can be used to identify genes that
are expressed in a coordinated manner. To quantify the level of relationship between
RKIP and Snail, the data was plotted in a scatterplot. The ratio values of RKIP plotted
against Snail values, and of E-cadherin and Snail, show a negative correlation in both
Datasets 1a and 1b (Figure 12). Pierson correlation, Spearman correlation, and Euclidean distance are popular measures of similarity, or distance measures, used in microarray data
analysis. We measured similarity using Pearson correlation coefficients as a distance
measure (Table 2). Pearson correlation is an often used correlation measure, and offered
advantages over Spearmand coefficients and Euclidean distance. The Pearson correlation
is a powerful measure that can identify both positive and negative correlations. The
Spearman correlation also spots positive and negative correlations and is more robust to
outliers than the Pearson measure, but is not as powerful. Euclidean distance measures
the relationship between gene expression profiles differently than correlation, as a
geometric interpretation, but doesn’t detect negative correlations (Stekel, 2003). We used
Pearson correlation for our analyses as we were specifically attempting to quantify a
negative correlation and it is more powerful than the Spearman distance measure, and our
scatterplot did not reveal outliers that would skew the correlation. As listed in Table 2, 56 A
B
Figure 12- Scatterplot correlation between Snail and E-cadherin or RKIP in datasets 1a and 1b. (A) Scatterplots of Snail versus E-cadherin and Snail versus RKIP of Dataset 1a. (B) Scatterplots of Snail versus E-cadherin and Snail versus RKIP of Dataset 1b. Axis values are in normalized expression units as provided by the Chinnaiyan laboratory. Trendlines and regression were generated in Excel.
57 there was a significant negative correlation between RKIP and Snail across all samples and across just the PCA and MET samples. The Pearson correlation coefficients of Snail and E-cadherin were also significantly negatively correlated across all samples and across just PCA and MET samples (Table 3). As Pearson correlation coefficients range from 1
(strong positive correlation) to -1 (strong negative correlation), with 0 meaning no correlation, the correlations between Snail and RKIP or E-cadherin are moderate to strong negative correlations that are statistically significant.
Dataset Samples N Correlation Coefficient p-value 1a All 52 -0.32 0.023 1a PCA/MET 34 -0.24 0.177 1b All 25 -0.58 0.003 1b PCA/MET 17 -0.52 0.038
Table 2 – Pearson correlation coefficients and p-values of Snail and RKIP. Values were derived from datasets 1a and 1b in the statistical program R
Dataset Samples N Correlation Coefficient p-value 1a All 52 -0.28 0.051 1a PCA/MET 32 -0.30 0.097 1b All 25 -0.57 0.003 1b PCA/MET 17 -0.68 0.004
Table 3- Pearson correlation coefficients and p-values of Snail and E-cadherin. Values were derived from datasets 1a and 1b in the statistical program R
58 The Oncomine datasets [from (Dhanasekaran et al., 2001)] showed a relationship
between RKIP and Snail in one study of prostate cancer progression. To substantiate this
finding, we obtained another dataset from the Chinnaiyan laboratory that used different
arrays and genes, but also included RKIP and Snail. The microarray data was not in the
Oncomine database. This data was also from a non-Affymetric cDNA microarray. This
dataset, which will be referred to as Dataset 2, was subjected to the same analyses as
done to datasets 1a and 1b. As shown in the box plots in Figure 13 and individual point-
by-point graph in Figure 14 RKIP and E-cadherin expression levels were largely
unchanged except for the metastatic samples, which decreased in both genes. Snail levels
increased from PCA to MET samples, as seen in the previous dataset.
Figure 13 – Box plots of RKIP, Snail, and E-cadherin levels in prostate cancer microarray Dataset 2. Data provided by the Chinnaiyan laboratory were log- transformed ratios. Graphs were created in the statistical program R. (B) NAP N=3, BPH N=5, PCA N=10, and MET N=7 (C) NAP N=13, BPH N= 17, PCA N=25, MET N=7
59 Because the Snail and E-cadherin levels remained largely unchanged between
normal, BPH, and PCA samples, the scatterplots and regression trendlines did not reflect significant relationships (Figure 15). However, plotting just the PCA and MET samples revealed that there were negative correlations between Snail and RKIP and Snail and E- cadherin (Figure 15). Statistical analyses confirmed the significant negative correlations between Snail and its target E-cadherin and also RKIP (Table 4).These datasets both show statistically significant negative correlations between Snail and RKIP in PCA and
MET samples.
60 ratios wereplottedforeachsample. Me Figure 14-
RKIP andSnailexpressioninea Normalized Expression Unit Expression Normalized tastatic samples weregroupedbylocation.
ch sample-Dataset1b.
Sample Number
RKIP andSnail
61 A
B
Figure 15- Scatterplot correlation between Snail and E-cadherin or RKIP in Dataset 2. (A) Scatterplots of Snail versus E-cadherin and Snail versus RKIP of all samples. (B) Scatterplots of Snail versus E-cadherin and Snail versus RKIP of only PCA and MET samples. Trendlines and regression were generated in Excel.
Snail versus Sample N Correlation Coefficient p-value RKIP All 62 -0.19 0.145 RKIP PCA/MET 32 -0.36 0.046 E-cadherin All 62 -0.05 0.726 E-cadherin PCA/MET 32 -0.41 0.019
Table 4- Pearson correlation coefficients and p-values of Snail and E-cadherin or RKIP. Values were derived from Dataset 2 using the statistical program R. 62 RKIP and E-cadherin are not negatively correlated with Slug in human tumor metastases
The Snail family member Slug was also able to repress E-cadherin and RKIP promoter activity. However, unlike Snail, Slug levels do not increase from normal, to
BPH, PCA, and MET samples, but decreases in the datasets (Figure 16). Boxplots from dataset 1 that Slug increased between NAP and BPH sample groups, but then fell in PCA and MET sample groups, which is not the expression pattern expected of an EMT repressor. Like Dataset 1, the Dataset 2 boxplot showed a drop in Slug between BPH and
PCA and MET samples. There was a small increase in Slug between PCA and MET samples, but statistical analysis determined that there was not a significant negative correlation between Slug and RKIP in Dataset 2 either across all samples or just PCA and
MET (Table 5). There was not a consistent correlation trend and these data do not support a role for RKIP regulation by Slug in prostate cancer. It also suggests differential regulation of Snail and Slug in prostate cancer; with Snail being more involved in prostate cancer progression than Slug.
63 A B C
Figure 16- Box plots of Slug expression from datasets 1a, 1b, and 2. Data provided by the Chinnaiyan laboratory were log-transformed ratios. Graphs were created in the statistical program R. (A) NAP N=4, BPH N=14, PCA N=14, MET N=20 (B) NAP N=3, BPH N=5, PCA N=10, and MET N=7, (C) NAP N=13, BPH N= 17, PCA N=25, MET N=7
Dataset Sample N Correlation Coefficient p-value 1b All 25 -0.12 0.375 1b PCA/MET 17 -0.14 0.454 2 All 62 0.46 0.022 2 PCA/MET 32 0.20 0.431
Table 5 – Statistical analysis of the correlation between RKIP and Slug. Values from datasets 1b and 2 were analyzed in the statistical program R.
EZH2 is negatively correlated with RKIP in human prostate metastases.
Another transcriptional regulator of interest in EMT is the polycomb group protein, enhancer of zeste EZH2 homolog 2. Saramaki et al. showed EZH2 expression was higher in hormone-refractory prostate tumors than hormone-naive tumors (Saramaki et al., 2006). EZH2 amplification was found to be in half of hormone refractory tumors 64 (Saramaki et al., 2006). EZH2 was recently shown to be highly upregulated in metastatic prostate cancer (Tomlins et al., 2007; Varambally et al., 2002). This upregulation is shown for both Datasets 1 and 2 in Figure 17a. Interestingly, there is also a strong negative correlation between RKIP expression and EZH2 (Figure 17b). In Dataset 1b there is a significant correlation across all samples of -0.67 (p-value = 0.0004) and of -
0.65 across PCA and MET samples (p-value = 0.009). While we do not have other data suggesting EZH2 represses RKIP, this would be an interesting molecule to investigate further.
65 A B C
D
Figure 17 – EZH2 expression and relationship to RKIP. (A) Box plots were generated as described previously for the datasets 1a (NAP N=4, BPH N=14, PCA N=14, MET N=20), (B) Dataset 1b (NAP N=3, BPH N=5, PCA N=10, and MET N=7), and (C) Dataset 2 NAP N=13, BPH N= 17, PCA N=25, MET N=7). (D) Scatterplot and regression analysis of RKIP and EZH2 values from Dataset 1b. Axes values are in normalized expression units that are not log-transformed.
66 Hierarchical clustering analysis reveals genes that may be coregulated with RKIP.
Microarray datamining permits not only the investigation of expression levels and
correlations between molecules, but also of determining which genes have similar
expression patterns. The identification of molecules involved in a specific cellular or
medical event, such as formation of a tumor, is sometimes referred to as gene expression
profiling. The expression “profiles” or “signatures” need to be organized in such a way
to find relationships that may imply significance. Cluster analysis can help elucidate
genes or groups of genes that may be similarly regulated or coexpressed in the samples of
interest. Hierarchical clustering can simplify the enormous amount of data in microarray
experiments and display expression profiles that are similar in the style of a phylogenetic
tree, which is more easily interpreted (Stekel, 2003). Dhanasekar et al. analyzed
microarray data to arrive at a molecular classification of prostate cancer; a profile of
prostate cancer at the molecular level (Dhanasekaran et al., 2001). We accepted the
classification of “normal” prostatic tissue, benign prostatic hyperplasia, and both
localized and metastasized prostate cancer samples used in the studies from the
laboratory of Arul Chinnaiyan. We then endeavored to determine which genes were
clustered with either RKIP or Snail to gain insight into how RKIP or Snail may be
regulated or grouped. Would their expresssion profiles be similar to other genes whose
functions are know? Might Snail induction in metastatic cancer samples match the pattern of other genes that, for example, are found near the Snail locus and may therefore offer some explanation of regulation? Hierarchical cluster analysis was performed across all samples for their correlation to RKIP in all three datasets. In other words, this type of 67 analysis allowed us to see which cDNAs had expression patterns similar to the expressoin
pattern of RKIP. The top 8000 genes correlating to RKIP (lowest correlation was
0.179575) were loaded into Cluster3.0, which did not filter out any genes prior to
clustering. The genes were correlated to RKIP in Excel, and the cutoff of 8000 (genes
most closely correlated to RKIP or Snail) was made as this was the highest number of
entries that the clustering program could handle on the computers available for the study.
Attempts at analyzing all genes resulted in computer failure, and 8000 is a large number
of candidates to examine as it was.
Hierarchical clustering was deemed the best analysis tool to use in organizing
these datasets. K-means clustering uses a different algorithm than hierarchical clustering
and self-organized maps uses methods similar to k-means clustering. Both of the latter
methods require pre-determined numbers of clusters to generate results, which is not
what we desired, and what’s more attempts at these clustering programs did not create
results that made sense biologically – the normal and cancer samples did not group as
expected. Furthermore, both k-means clustering and self-organised maps give different
results when run different times (Stekel, 2003). Hierarchical clustering was selected to
analyze the relationships between the genes and samples as it gives the same results every
time and because we wanted to see how the data “naturally” clustered, not how the data would group, or cluster, when we told the program how many clusters we wanted to see.
Also, Dhanasekaran et al. used average linkage hierarchical clustering to generate their data (Dhanasekaran et al., 2001). Average linkage was used in the hierarchical clustering, as it tends to perform well in many microarray applications (Stekel, 2003). 68 Both single and complete linkage clustering were performed also; however, these
methods resulted in heat maps and dendrograms that were inferior to average linkage, as the gene expression profiles did not look as similar and the samples did not cluster as well by known biological relevance (i.e.- MET samples with MET samples) (data not shown).
In the ensuing analyses, we examined the hierarchical cluster results looking for
various relationships. First we located the cluster containing RKIP and then found the
expressed genes that have expression profiles most similar to RKIP, or “clustered with”
RKIP. We performed the cluster anlaysis for each Dataset 1a, 1b, and 2 separately. Next
we considered whether the clusters created from the three datasets had genes in common.
In order for this line of investigation to make sense, it needs to be explained that the
cluster analysis figures in this dissertation are incomplete. The full dendrogram for each
cluster analysis was far larger than could be accommodated (and clearly labeled) in this
format and therefore only the part of the dendrogram including genes with expression
profiles similar to RKIP were shown. The portion of the dendrogram showing expression
profiles very dissimilar to RKIP were not included. Thus it is not unreasonable to
compare genes with somewhat similar expression profiles to RKIP in different cluster
analyses just to see if there are shared members out of the thousands of possibilities. We
also narrowed, or focused, the cluster analysis to only localized and metastatic prostate
cancer samples and listed those molecules with expression patterns most similar to that of
RKIP. And finally we queried the PCA and MET-sample only clusters for shared genes
as we did for the cluster analyses that included all sample groups. 69 Table 6 lists the genes clustered with RKIP across all samples groups (NAP,
BPH, PCA, MET). Dataset 1a has genes for protein tyrosine phosphatase, receptor type,
F (PTPRF) and guanine nucleotide binding protein (G protein) beta polypeptide 1
(GNB1) most closely correlating to RKIP (Figure 18 and Table 6). PTPRF is a tyrosine
phosphatase involved in beta-catenin signaling and regulating cell to cell contacts at
adherens junctions. The genes glia maturation factor (GMFB), phosphoribosyl
pyrophosphate synthetase 2 (PRPS2), (SID1669), gamma-aminobutyric acid (GABA) B
receptor, 1 (GABBR1) were found to be most closely correlated to RKIP in Dataset 1b
(node correlation = 0.699697) (Figure 19 and Table 6). GMFB is a brain protein that
acts as a neuronal and glial growth and differentiation factor (Kaplan et al., 1991;
Nishiwaki et al., 2001). It is also believed to be involved in signal transduction, and
recent studies implicated GMFB in thymoma T-cell development by inducing
lymphoepithelial interactions (Yamazaki et al., 2005). PRPRS2 is not well-studied, it is
simply believed to catalyze the synthesis of phosphoribosyl pyrophosphate, which is a
substrate in the synthesis of nucleotides (NCBI). GABBR1 is a receptor for GABA,
which mediates inhibitory signaling in the central nervous system. It has recently been shown that advanced hippocampal pathology may be associated with decreased levels of
GABBR1 in this region, which precedes neuronal cell death and may contribute to the
dysfunctional of hippocampal circuitry in Alzheimer’s Disease (Iwakiri et al., 2005).
Dataset 2 had RKIP most closely clustered with cancer susceptibility candidate
4(CASC4), zinc finger protein 579 (ZNF579), and RING finger protein 141 (RNF141)
with a node correlation of 0.690065 (Figure 20 and Table 6). This cluster was most 70 interesting, as ZNF579 and RNF141 contain motifs known to be involved in transcription regulation. Little is known about ZNF579 except for what can be deduced from its sequence (Maglott et al., 2005). RNF141 has a mouse counterpart which suggests the function of a testis-specific transcription factor during spermatogenesis (Qiu et al., 2003).
In humans, this protein was found to be expressed in the testicular tissue of fertile men but not in azoospermic men (Zhang et al., 2001). The RING finger is a motif involved in protein-protein and protein-DNA interactions. There is a finger of about 50 residues that binds two atoms of zinc (Maglott et al., 2005); (Qiu et al., 2003). CASC4 has been associated with HER2/neu proto-oncogene over expression. Increased CASC4 amplifies
Her2/neu, and over expressed Her2/neu is found in about 30% of human breast and 20% of human ovarian cancers (Maglott et al., 2005; Scheer et al., 2003).
In addition to analyzing the genes most closely clustered to RKIP, we sought to compare the clusters for shared genes. Ideally datasets 1a and 1b would have nearly identical clusters as they are made of the same experimental data (with different references), but they were not perfectly matched although they did share many genes.
These genes shared between Dataset 1a and 1b are not listed here as there were very many. Instead, genes shared between two experiments Dataset 1a and Dataset 2, or
Dataset 1b and Dataset 2, or all three, were investigated (Table 7). The three clustered datasets for RKIP across all samples included shared genes, which may signify a more significant relationship to RKIP. These genes included transmembrane 9 super family member 2 (TM9SF2), testis enhanced gene transcript (TEGT), DAZAP2, twifilin, 71 Gene Node Study Summary Name Cor. GNB1 0.774 1a Guanine nucleotide binding protein (G protein) beta polypeptide 1 PTPRF 0.774 1a Protein tyrosine phosphatase, receptor type F GMFB 0.899 1b Glia maturation factor, beta PRPS2 0.803 1b Phosphoribosyl pyrophosphate synthetase 2 SID1669 0.803 1b Ring finger protein 11 GABBR1 0.700 1b Gamma-aminobutyric acid (GABA) B receptor, 1 CASC4 0.690 2 Cancer susceptibility candidate 4 ZNF579 0.690 2 Zinc finger protein 579 RNF141 0.690 2 Ring finger protein 141
Table 6 – Genes most closely clustered with RKIP across all samples. Summary of cluster analyses of datasets 1a, 1b, and 2 including all sample groups, NAP, BPH, PCA, and MET
actin –binding protein, homolog 1 (PTK9), transcription elongation factor, 1 (TCEA1 or
SII), transmembrane 9 superfamily member 2 (TM9SF2), YME-like 1 (YME1L1), hypothetical protein FLJ13611, NADH dehydrogenase (ubiquinone) 1 alpha subcomplex,
5 (NDUFA5), caspase 7 (CASP7), and member of the Ras oncogene family (RAB6A).
TEGT was named such as the rat protein was found to be highly abundant in rat
postpubertal testis (Walter et al., 1994). TEGT is also known as Bax inhibitor 1 or BI1,
and was found to suppress the proapoptotic action of Bax (Xu and Reed, 1998).
Interestingly, DNA microarray analysis showed TEGT is upregulated in primary breast
tumors and malignant prostate tissue (Grzmil et al., 2003; van 't Veer et al., 2002; Welsh
et al., 2001). The Welsh microarray study showed that there was no significant
difference in RKIP expression between normal and cancerous prostate tissue (Snail was
72 Gene Name Study Summary TEGT 1a, 2 Testis enhanced gene transcript (BAX inhibitor 1) DAZAP2 1a, 2 DAZ associated protein 2 COPB 1b, 2 Coatomer protein complex, subunit beta 1 PTK9 1b, 2 Twifilin, actin-binding protein, homolog 1 (Drosophila) TCEA1 1b, 2 Transcription elongation factor A, 1 (SII) TM9SF2 1a,1b,2 Transmembrane 9 superfamily member 2 YME1L1 1a,2 YME1-like 1 (S. cerevisiae) FLJ13611 1a,2 Hypothetical protein NDUFA5 1a,2 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 5 CASP7 1a,2 Caspase 7, apoptosis-related cysteine peptidase RAB6A 1a,2 Member of the RAS oncogene family
Table 7– Genes shared in datasets 1a and 2, or 1b and 2 when clustered to RKIP across all samples. The heat maps and dendrograms of Figures 15-17.
not included) (Welsh et al., 2001). Grzmil and colleagues also showed that knockdown of TEGT increased spontaneous apoptosis in prostate cancer cell lines (Grzmil et al.,
2003). Like DAZAP2, TEGT initially grabbed our attention as it is located on the same chromosome as RKIP. TEGT is located at 12q12-q13 (Ensemble genes on sequence map location= 48.4 Mbp, deCODE map= 63.7 cM), DAZAP2 is located at 12q13.13
(Ensemble genes on sequence map location = 49.9 Mbp, deCODE map = 66 cM), and
RKIP is located at 12q24.23 (Ensemble genes on sequence map = 117 Mbp, deCODE map = 137 cM) (Kong et al., 2002; Maglott et al., 2005). It is tempting to speculate that these three genes are co-regulated; however, RKIP is not very close to DAZAP2 and
TEGT.
COPB1 is a part of non-clathrin coated vesicles purportedly involved in intra-
Golgi transport and bulk transport from the ER to the Golgi complex (Duden et al.,
73 1991). PTK9, or TWF1, is an actin-binding protein that is involved in endocytosis in
mammalian cells (Helfer et al., 2006). It appears to be regulated by Rac1 and Cdc42 suggesting TWF1 may be involved in polarized growth and motility (Vartiainen et al.,
2003). PTK9 is located on chromosome 12 at 12q12 (42.5 Mbp, 58.6 cM TCEA1 helps
RNA polymerase II bypass specific blocks to transcript elongation, enabling pol II to transcribe faster [reviewed in (Wind and Reines, 2000)]. TM9SF2, also known as p76, is highly evolutionarily conserved and found predominantly in endosomes (Schimmoller et al., 1998). Little is known about YME1L1. This protein is localized in the mitochondria and is the human ortholog to the yeast mitochondrial AAA protease Yme1p (Coppola et al., 2000). NDUFA5 codes for the B13 subunit of complex I of the respiratory chain which is localized to the inner mitochondrial membrane (Maglott et al., 2005). Caspase 7 is crucial in the execution phase of apoptosis (Lakhani et al., 2006). Inactivating mutations in CASP7 have been linked to the pathogenesis of some human solid cancers
(Soung et al., 2003). The CASP7 promoter was able to be activated by p53 after prohibitin over expression through p53-binding sites (Joshi et al., 2007). Rab6A is a small GTPase that regulates retrograde transportation connecting early endosomes to the endoplasmic reticulum (Martinez et al., 1997). Rab6 may also regulate the inactivation of the Mad2-spindle checkpoint (Miserey-Lenkei et al., 2006). Most other genes listed in Table 6 were shared between Dataset 1a and 1b, which is not surprising as these include the same samples.
We also performed hierarchical cluster analyses on genes correlated with RKIP in
just PCA and MET samples. It appeared in Dataset 2 that the correlation between RKIP 74 and Snail was strong between these two sample groups, and we wanted to see what genes would have expression profiles most similar to RKIP between just these two groups
(Table 8). In Dataset 1a, the genes DAZ associated protein 2(DAZAP2), PEST proteolytic signal containing nuclear protein (PCNP), and CD46 molecule, complement regulatory protein (MCP or CD46) were most closely clustered with RKIP, with a node correlation of 0.736259 (Figure 21). Very little is known about PCNP, except that it can be ubiquitinated by NIRF (Maglott et al., 2005; Mori et al., 2004). DAZAP2 protein is ubiquitously expressed protein that has been shown to bind to DAZ and DAZL in a yeast two-hybrid system (Maglott et al., 2005; Tsui et al., 2000). DAZ (deleted in azoospermia) is expressed solely in the testes and its deletion from the Y chromosome is associated with azoospermia (reviewed in (Yen, 2004). Also, DAZAP2 gene expression is downregulated in multiple myeloma, but little is know about this role (Shi et al., 2004).
DAZAP2 is located on chromosome 12 (12q12) as is RKIP (12q24.3) (Maglott et al.,
2005). CD46 protein is a membrane protein and a complement regulatory receptor for
C3b and C4b, as well as a receptor for many pathogens. It may also be involved in the fusion of egg and sperm during fertilization (Maglott et al., 2005); reviewed in
(Liszewski et al., 2005).
Dataset 1b showed RKIP most closely clustered with glia maturation factor, beta
(GMFB) with a node correlation 0.882359, and Runt-related transcription factor 1
(RUNX1) with a node correlation of 0.74877 (Figure 22 and Table 8). GMFB was described previously as it clustered closely with RKIP across all samples, also. RUNX1
75 is a transcription factor that can function as an activator or repressor, presumably depending on its interaction with coregulators. Deregulation by mutation, haploinsufficiency, translocation, and amplification of RUNX1 have all been linked to
Gene Cor. Study Summary DAZAP2 0.736 1a DAZ associated protein 2 PCNP 0.736 1a PEST proteolytic signal containing nuclear protein MCP 0.736 1a Membrane cofactor protein (CD46, trophoblast-lymphocyte cross-reactive antigen) GMFB 0.882 1b Glia maturation factor, beta RUNX1 0.748 1b Runt-related transcription factor 1 (acute myeloid leukemia 1; aml1 oncogene) CLTB 0.857 2 Clathrin, light chain (Lcb) UBE2D3 0.788 2 Ubiquitin-conjugating enzyme E2D 3 RAB2 0.788 2 Member of Ras oncogene family PSMA4 0.788 2 Proteasome (prosome, macropain) subunit, alpha type, 4 DRAP1 0.788 2 Dr1-associated protein 1 (negative cofactor 2 alpha) SRP54 0.788 2 Signal recognition particle of 54kDa RNF13 0.788 2 Ring finger protein 13 MAP1B 0.788 2 Micotubule-associated protein 1B CD46 0.788 2 CD46 molecule, complement regulatory protein
Table 8 – Gene most closely clustered with RKIP across only PCA and MET samples. Summary of cluster analyses of datasets 1a, 1b, and 2 including sample groups PCA, and MET.
several types of leukemia, making it a critical target in treatment (Maglott et al.,
2005{Mikhail, 2006 #600; Mikhail et al., 2006).
Coincidentally, RUNX1 is also known as PEBP2aB, and other family members such as RUNX2 and RUNX3 are known as PEBP2a1/A2/aA/aA1 and PEBP2aC,
76 respectively. In these names PEBP is short for polyomavirus enhancer-binding protein, not phosphatidylethanolamine-binding protein.
The genes clustered with RKIP in Dataset 2 were clathrin, light chain (CLTB), ubiquitin-conjugating enzyme E2D3 (UBE2D3), RAB2, proteasome subunit, alpha type,
4 (PMSA4), Dr1-associated protein 1 (DRAP1), signal recognition particle of 54 kDa
(SRP54), RING finger protein 13 (RNF13), microtubule-associated protein 1B (MAP1B), and CD46 molecule (CD46) (Figure 23 and Table 8). CLTB is believed to function as regulatory elements as part of clathrin (NCBI). Clathrin-coated pits internalize membrane receptors and their ligands in receptor-mediated endocytosis affects membrane dynamics, signal transduction, and intracellular cytoplasmic organization. UBE2D3 is a member of the E2 ubiquitin-conjugating enzyme family. Saville et al. (2004) showed evidence that UBE2D3 catalyzes the ubiquitination of the tumor-suppressor protein p53.
UBE2D3 was also shown to be able to catalyze the polyubiquitination of NEMO (also known as Inhibitor of KappaB Kinase, Gamma) (Tang et al., 2003). RAB2 is a small
GTPase involved in membrane transport between the endoplasmic reticulum and the
Golgi complex (Tisdale et al., 1992). DRAP1 is a corepressor of transcription, requiring interaction with DR1 (NC2) (Mermelstein et al., 1996). NC2 is a bifunctional transcription factor that has distinct activation and repression functions (Willy et al.,
2000). The 54 kDa subunit of the signal recognition particle binds to the signal sequences of nascent secretory and transmembrane proteins, stalls translation, and facilitates translocation of the nascent polypeptide into the endoplasmic reticulum lumen
(Walter and Blobel, 1981; Walter and Johnson, 1994; Wild et al., 2004). RNF13 is not 77 well characterized. It contains a RING zinc finger and its homolog in the chicken is
suspected of being a transcriptional regulator (Maglott et al., 2005; Tranque et al., 1996).
Members of the microtubule-associated protein family, to which MAP1B belongs, are
believed to be involved in microtubule assembly (Maglott et al., 2005). It is involved in
normal brain development and was found to have a role in axon guidance and elongation
(Bouquet et al., 2004; Takei et al., 1997). PMSA4 will be described in regards to Table
9.
In addition to genes shared in datasets clustered around RKIP expression in all
samples, there are some clustered around RKIP in just PCA and MET samples. The genes UBE2D3, SRP54, RAB2, and PSMA4 are included in this group as well as in the group of genes most closely clustered to RKIP across PCA and MET samples (Table 9).
TCEA1 is a gene shared in RKIP clusters both across all samples and just across
PCA and MET. Genes in Table 9 not previously described include: Ras family member
(RAB5A), ribosomal protein L10 (RPL10 or QM), proteasome subunits, alpha type, 2, 4,
and 6 (PSMA2, PSMA4, PSMA6), Rad23 homolog (RAD23), NADH dehydrogenase
(ubiquinone) Fe-S protein 4 (NADH-coenzyme Q reductase), serpin peptidase inhibitor,
clade B (ovalbumin), member 6 (SERPINB6), DEAD box polypeptide 1 (DDX1), SET
translocation (myeloid leukemia associated) (SET), reticulocalbin 2, EF-hand calcium
binding domain (RCN2), G1 to S phase transition 2 (GSPT2), chaperonin containing
TCP1, subunit 4 (delta) (CCT4), retinol binding protein 4, plasma (RBP4), protein
tyrosine phosphatase type IVA, member 2 (PTPA2), and Meis homeobox 2 (MEIS2).
78 Gene Study Summary Name UBE2D3 1b,2 Ubiquitin-conjugating enzyme E2D 3 RAB2 1a,1b,2 Member of Ras oncogene family RAB5A 1b, 2 Member of Ras oncogene family SRP54 1b, 2 Signal recognition particle, 54 kDa RPL10 1a, 2 Ribosomal protein L10 (QM) TCEA1 1b,2 Transcription elongation factor A (SII), 1 PSMA6 1a,1b,2 Proteasome (prosome, macropain) subunit, alpha type, 6 PSMA2 1a,1b,2 Proteasome (prosome, macropain) subunit, alpha type, 2 PSMA4 1b, 2 Proteasome (prosome, macropain) subunit, alpha type, 4 RAD23B 1b, 2 RAD23 homolog B (S. cerevisiae) NDUFS4 1b, 2 NADH dehydrogenase (ubiquinone) Fe-S protein 4 (NADH- coenzyme Q reductase) SERPINB6 1a, 2 Serpin peptidase inhibitor, clade B (ovalbumin), member 6 DDX1 1b, 2 DEAD (Asp-Glu-Ala-Asp) box polypeptide 1 SET 1b, 2 SET translocation (myeloid leukemia-associated) RCN2 1b, 2 Reticulocalbin 2, EF-hand calcium binding domain GSPT2 1b, 2 G1 to S phase transition 2 CCT4 1b, 2 Chaperonin containing TCP1, subunit 4 (delta) RBP4 1a,1b,2 Retinol binding protein 4, plasma PTP4A2 1b, 2 Protein tyrosine phosphatase type IVA, member 2 MEIS2 1b, 2 Meis homeobox 2
Table 9 - Genes shared in datasets 1a and 2, or 1b and 2 when clustered to RKIP across PCA and MET samples. Summary of cluster analyses of datasets 1a, 1b, and 2 including sample groups PCA and MET.
The functions or points-of-interest regarding the PCA-MET shared genes listed in
Table 9 will be summarized briefly. Rab5 is a GTPase that acts by recruiting effector proteins onto early endosomes and controlling their organization in distinct membrane subdomains (Rybin et al., 1996). Shin et al. (Shin et al., 2005) has also shown that Rab5 coordinates signaling with organelle homeostasis by regulating phosphoinositide generation and turnover through phosphoinositide kinases and phosphatases. Rpl10 is a
79 component of the 60S subunit of the ribosome (Maglott et al., 2005). Rpl10 can interact
with c-Jun to inhibit transactivation of AP-1-regulated promoters (Imafuku et al., 1999;
Monteclaro and Vogt, 1993). Downregulation of Rpl10 has been correlated with tumor
grade in prostate cancer and may also be associated with ovarian cancer (Altinok et al.,
2006; Shen et al., 2006). PMSA2, PMSA4 and PMSA6 all encode for proteins that are
part of the 20S core alpha subunit of the proteasome and members of the peptidase T1A
family (Maglott et al., 2005). Unexpectedly, researchers have found PMSA2 to be over-
expressed in multiple myeloma (compared to healthy controls) patient samples and
PMSA6 was upregulated in hepatocellular carcinomas of HBx gene knock in mice and in the breast cancer cells of women (Abraham et al., 2005; Bhui-Kaur et al., 1998; Cui et al.,
2006). Rad23B is a nuclear excision repair protein (Maglott et al., 2005). It has been shown to increase the nucleotide excision activity of MPG, indicating a possible role in
DNA damage recognition (Miao et al., 2000). A splice variant it is highly expressed in
the testis and correlated with spermatogenesis (Huang et al., 2004). It has also been
shown to bind polyubiquitinylated p53 and “shield” it from deubiquitylation (Glockzin et
al., 2003). The protein of NDUFS4 is the 18 kDa IP subunit of the mammalian complex I
of the respiratory chain, playing a vital role in ATP manufacture [reviewed in (Papa et al.,
2002)]. Serpinb6 inhibits serine proteinases, including kallikreins in the prostate and
ovary (Bird, 1999; Mikolajczyk et al., 1999). DDX1 is an RNA helicase. Several studies
have examined the amplification and over expression of DDX1 in neuroblastoma , and in
general the DDX family of helicases have a differentiation and carcinogensis
(Abdelhaleem et al., 2003; Amler et al., 1996; Manohar et al., 1995; Noguchi et al., 1996; 80 Squire et al., 1995). The oncoprotein SET (also known as TAF-Iβ and INHAT) is a histone chaperone. It has been attributed several additional functions including interacting with proteases and DNA-binding proteins, regulating transcription, replication, apoptosis, and inhibiting histone acetyltransferase activity [discussed in
(Muto et al., 2007)]. Reticulocalbin 2 bears similarity to the EF-hand, which is a high affinity Ca+2-binding motif (Maglott et al., 2005). Cavallo et al. identified Rcn2 as a
Tumor Associated Antigen whose expression is linearly related to mouse tumor mass
increase in mammary tissue (Cavallo et al., 2005). Little is known about the protein
product of GSPT2. The product is known as eRF3b, an isotype of eRF3 proteins which
are responsible for correct termination of translation in eukaryotes (Hoshino et al., 1998;
Maglott et al., 2005). CCT4 encodes for subunit 4 of the TCP-1 ring complex (TRiC)
that mediates and promotes the proper folding of a wide spectrum of proteins, including
von Hippel-Lindau protein [reviewed in (Dunn et al., 2001; Melville et al., 2003)].
Retinol-binding protein 4 is a member of the lipocalin family of extracellular proteins
that are able to bind lipophiles. It is the specific carrier for retinol in the blood (Maglott
et al., 2005). In regards to disease, RBP4 is associated with insulin resistance, and was
found to be downregulated in acute myeloid leukemia (Kwak et al., 2004; Yang et al.,
2005a). PTP4A2 encodes for a protein tyrosine phosphatase often referred to as PRL2.
The biological function of members of this small protein class is not clear; however,
PRL2 over expression in hamster pancreatic ductal epithelial cells resulted in a loss of
contact inhibition [reviewed in (Stephens et al., 2005)]. Meis2 is a TALE homeobox
transcriptional regulator (Maglott et al., 2005). It has been largely studied in regards to 81 embryonic mesencephalon development (Shim et al., 2007). Meis proteins can bind to
Hox or Pbx proteins, which may stabilize the homeoprotein/DNA complex. They are
considered to be able to activate specific genes by penetrating repressive chromatin
(Schnabel et al., 2000).
Hierarchical clustering analysis of Snail reveals genes that may be coregulated.
We were interested investigating how hierarchical clustering may help highlight
the mechanism or pathway of Snail upregulation. Genes clustering with Snail may be
part of its regulatory network, as activators of Snail may be upregulated with it. The
datasets were also hierarchically clustered according to their correlation to Snail. The methods used were the same as described for the RKIP hierarchical cluster analyses.
Again, the cluster analyses were performed across all samples and then across PCA and
MET samples only. The cluster analyses results are listed below; however, gene
descriptions are not as verbose because we were primarily interested in genes clustering
with RKIP.
As listed in Table 10, genes clustered with Snail across all samples included
UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 2
(GALNT2 or GalNAc-T2) and diacylglycerol O-acyltransferase homolog 1 (DGAT) in
Dataset 1a (Figure 24 and Table 10). Dataset 1b (Figure 25 and Table 10) had tyrosine
kinase with immunoglobin-like and EGF-like domains 1 (TIE) and FK506 binding
protein 6 (FKBP6) most closely clustered with Snail, and Dataset 2 had signal sequence
receptor, delta (SSR4) and polycomb group ring finger 3 (PCGF3) most closely clustered 82 with Snail (Figure 26 and Table 10). Table 10 summarizes the genes found to be most
closely correlated to Snail across all samples. Dataset 2 has a low correlation for Snail-
clustered genes, and is unfortunately probably not very meaningful. For this reason,
shared genes were not evaluated across all sample groups and clusters, as was done for
RKIP in Table 7.
Gene Cor. Study Summary Name GALNT2 0.720 1a UDP-N-acetyl-alpha-D-galactosamine: polypeptide N- acetylgalactosaminyltransferase 2 (GalNAc-T2) DGAT 0.720 1a Diacylglycerol O-acyltransferase homolog 1 TIE 0.840 1b Tyrosine kinase with immunoglobin-like and EGF-like domains 1 FKBP6 0.805 1b FK506 binding protein 6 SSR4 0.440 2 Signal sequence receptor, delta (translocon-associated protein delta) PCGF3 0.440 2 Polycomb group ring finger 3
Table 10- Genes most closely clustered with Snail across all samples. Summary of cluster analyses of datasets 1a, 1b, and 2 including all sample groups, NAP, BPH, PCA, and MET.
Genes most closely clustered with Snail across only PCA and MET samples are
listed in Tale 11. These genes included solute carrier family 26, member 6 (SLC26A6) in
Dataset 1a (Figure 27and Table 11), and FKBP6, tumor necrosis factor, alpha-induced protein 1 (TNFAIP1), and coagulation factor X (F10) in Dataset 1b (Figure 28 and Table
11). Interestingly, WNT5 is also clustered with Snail in Dataset 1a, although not as 83 closely as SLC26A6. Also, in Dataset 1b the genes PDGFB and Smad3 (MADH3) were
clustered near Snail. Dataset 2 had haloacid dehalogenase-like hydrolase domain containing 1a (HDHD1A) and Fancomi anemia, complementation group D2 (FANCD2) most closely clustered with Snail (Figure 29 and Table 11). In this well-annotated dataset, the genes NR0B2, ACMSD, GPR120, CPLX2, SNAP25, CENPE, and TAF7L are the next closely correlated genes to Snail. Although it is not the most closely correlated gene, EGFR is found in both the Snail clusters (PCA and MET only) for datasets 1a and 2.
Gene Cor. Study Summary Name SLC26A6 0.76 1a Solute carrier family 26, member 6 FKBP6 0.947 1b FK506 binding protein 6 RAB7 0.759 1b RAB7A, member of RAS oncogene family AK1 0.759 1b Adenylate kinase 1 HDHD1A 0.694 2 Haloacid dehalogenase-like hydrolase domain containing 1A FANCD2 0.694 2 Fancomi anemia, complementation group D2
Table 11 - Genes most closely clustered with Snail across PCA and MET samples. Summary of cluster analyses of datasets 1a, 1b, and 2 including sample groups PCA and MET.
In addition to analyzing genes that may be related to Snail via cluster analysis, we
looked for correlations with candidate genes gathered from the literature. Candidates
included FGF and EGF signaling pathways, the Wnt signaling pathway, and NF-κB
signaling. These signaling pathways are very complex. Significant correlations for Snail-
signaling-related molecules across PCA and MET samples are listed in Table 12. 84
Table 12 – Correlations of Snail-signaling related molecules. Pearson correlation coefficients were generated from Dataset 2 comparing Snail to other selected molecules. Statistical analysis was done in the program R.
The only signaling pathway that had a significant correlation to Snail (that made
any sense) was EGFR. EGFR signaling is well established in cancer. Interestingly,
EGFR is also rather closely correlated with Snail in the cluster analysis making it
tempting to speculate as to how this pathway is the most involved in Snail regulation in
prostate cancer. It is unlikely that this signaling pathway is solely responsible for the
regulation of Snail. Snail is regulated post-transcriptionally by GSK-3B and although
there is no correlation in the cDNA levels of these genes, there was no measure of the
protein levels or activity of GSK-3B. Certainly Wnt signaling can not be ruled out of
Snail regulation in prostate cancer as it was not completely examined. There were select
few Wnt genes in the microarrays and this pathway would be more involved in the post-
transcriptional regulation of Snail.
85
Figure 18 - Heat map and dendrogram Dataset 1a. Log-transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering.
86
Figure 19 - Heat map and dendrogram Dataset 1b Log-transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering
87
Figure 20 - Heat map and dendrogram Dataset 2. Log-transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering.
88
Figure 21 - Heat map and dendrogram Dataset 1a, PCA and MET only. Log- transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering.
89 Decreased expression, downregulated Increased expression, upregulated No change in expression Missing
Figure 22 - Heat map and dendrogram Dataset 1b, PCA and MET only. Log- transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering. 90
Figure 23 - Heat map and dendrogram Dataset 2, PCA and MET only. Log- transformed ratios were correlated to RKIP and subjected to average linkage hierarchical clustering.
91
Figure 24 - Heat map and dendrogram Dataset 1a,. Log-transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering. 92
Figure 25 - Heat map and dendrogram Dataset 1b. Log-transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering. 93
Figure 26 - Heat map and dendrogram Dataset 2. Log-transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering.
94
Figure 27 - Heat map and dendrogram Dataset 1a, PCA and MET only. Log- transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering.
95
Figure 28 - Heat map and dendrogram Dataset 1b, PCA and MET only. Log- transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering.
96
Figure 29 - Heat map and dendrogram Dataset 2, PCA and MET only. Log- transformed ratios were correlated to Snail and subjected to average linkage hierarchical clustering.
97 DISCUSSION
Raf kinase inhibitor protein is emerging as a protein of extreme interest in
the fields of cancer biology and neurobiology. RKIP is central in RAF/MEK/ERK
signaling and as such may be a very important player in a number of physiological processes. As a signaling molecule that can sensitize cancer cells to apoptosis and is
involved in cancer cell invasion, RKIP must be investigated for its role in cancer and
potential therapeutic benefits.
RKIP was recently reported as being downregulated in metastatic prostate cancer
(Fu et al., 2006; Fu et al., 2003). Keller et al. has proposed that RKIP functions as a
metastasis suppressor gene (Fu et al., 2006). It is not known how RKIP is
downregulated in cancer, or how it is transcriptionally regulated at all. We have
endeavored here to investigate the role of the repressors Snail and Slug on RKIP and to
possibly identify other genes co-regulated with RKIP.
The transcription factors Snail and Slug can repress the cell adhesion molecule E-
cadherin (Barrallo-Gimeno and Nieto, 2005; Peinado et al., 2004b; Thiery and Sleeman,
2006). Their role in the epithelial-to-mesenchymal transition has been implicated in cancer cell invasion and metastasis (Thiery, 2002). We observed that RKIP was highly
expressed in the low metastatic cell line LNCaP, but reduced in the DU145 and PC3 cell
lines. RKIP expression in these cell lines was similar to E-cadherin expression, and
inverse to Snail expression. Modulation of Snail could alter RKIP expresssion prostate
cancer cells lines. E-cadherin repression by Snail overexpression had been reported in 98 other cell lines, and here we reproduced this effect in the prostate cancer cell line LNCaP
(Batlle et al., 2000; Cano et al., 2000; Hajra et al., 2002; Peinado et al., 2004a). RKIP was also repressed by Snail overexpression in LNCaP,; and modulating Snail by RNA interfernce produced the opposite effect – RKIP upregulation – in DU145 and PC3 cells.
Thus far, Snail has been shown to induce mesenchymal markers such as vimentin and fibronectin, although there is no evidence that this upregulation is a direct effect of Snail on those molecules’ promoters (De Craene et al., 2005; Olmeda et al., 2007{De Craene,
2005 #451). The direct binding and repression activity of Snail on RKIP is supported by
ChIP assay, in which E-boxes proximal to the RKIP transcriptional start site appear to interact with Snail. The extent to which each of the four E-boxes tested near the promoter is responsible can not be answered by these experiments. Nor was the effect of several other putative E-boxes within 2kb of the RKIP transcriptional start site measured.
It is possible that the repression caused by Snail is from cumulative binding at multiple
E-boxes. The assay also did not attempt to detect other members of the complex to
which Snail is believed to belong. . Recently it has been shown that Snail is part of a
complex with HDAC1/HDAC2 and mSinA, through which it is believed Snail exerts its repressive effects (Peinado et al., 2004a). Histone deacetylases (HDACs), acting with corepressors, may strip acetylated histones H3 and H4 to confer transcriptional inactivation on a chromosomal region. Chromatin remodeling and histone modification are important means of gene expression control.
We focused here on RKIP expression regulation through a transcription factor,
Snail, although other factors, and other means of regulation have not been entirley rules 99 out. For example, promoter methylation may be responsible for RKIP downregulation.
Previous studies in colorectal and hepatocellular carcinomas failed to show RKIP
expression to be induced, or repression relieved, following treatment with a
demethylating agent (Minoo et al., 2007; Schuierer et al., 2006a). Based on these studies,
we did not focus on promoter methylation; however, future studies are needed to
specifically address this issue in prostate cancer.
Regulation of RKIP may not occur through promoter methylation but epigenetic transcriptional control may occur through the polycomb group protein EZH2. The polycomb group protein EZH2 is part of a gene regulatory mechanism that can determine cell fate in pathogenesis. EZH2 is part of polycomb repressive complexes 2,3, and 4
(Kuzmichev et al., 2005). Recruited to chromosomal DNA, PRC2 deacetlyates histone
H3 and EZH2 methlates K27. This modification can then recruit PRC1, which inhibits transcription (Dellino et al., 2004; Gil et al., 2005; Min et al., 2003; Ringrose et al., 2003; van der Vlag and Otte, 1999). Deregulation of EZH2 and other polycomb group proteins has been linked with cancer cell proliferation (Gil et al., 2005; Saramaki et al., 2006;
Varambally et al., 2002). EZH2 was also strongly negatively correlated with RKIP in the microarray analysis. This is not surprising, as EZH2 gene amplification has been found to lead to EZH2 over expression in late-stage prostate tumors. Furthermore, EZH2 amplification has been studied as a prostate cancer biomarker by the laboratory from which we obtained our microarray data (Varambally et al., 2002). We are currently investigating EZH2 as an RKIP repressor, also.
100 Slug is also a potent repressor of E-cadherin. Hajra et al. found Slug RNA levels in various breast cancer cell lines to correlate with loss of E-cadherin more than did Snail
RNA levels (Hajra et al., 2002). In this study, Slug RNA levels seem to be correlated with loss of E-cadherin, in the prostate cancer cell line PC3. The microarray meta- analysis in prostate cancer tissue samples suggests that Slug is not involved in E-cadherin repression in vivo. Our analysis suggests that Snail is the more likely candidate for both
E-cadherin and RKIP repression in prostate cancer progression. However, it is entirely possible that Slug and Snail may be differentially regulated according to cell type and stage. The transcriptional regulation of prostasin provides a very interesting example.
Prostasin is a glycosylphosphatidylinositol-anchored protein on the cell surface and is also found in seminal fluid (Chen et al., 2001b; Yu et al., 1994). Prostasin, like
RKIP has serine protease activity (Yu et al., 1994). Prostasin is expressed in the low- metastatic cell line LNCaP and absent in the invasive cell lines PC3 and DU145. The invasiveness of PC3 and DU145 were reduced by forced prostasin expression. In another interesting parallel to RKIP, prostasin expression is also lost in advanced prostate cancer
(Chen et al., 2001a). Prostasin promoter methylation has been shown to be a means of downregulation of prostasin in prostate and breast cancer cell lines (Chen and Chai, 2002;
Chen et al., 2004{Chen, 2002 #697). Recently, Chen et al. identified Slug and Snail as prostasin repressors in vitro, while investigating the role of androgens on prostasin regulation (Chen et al., 2006a). Prostasin promter activity, as measured by luciferase assay, was increased by the sterol-regulatory element-binding proteins SREBP-1c and
SREBP-2. However, treating LNCaP cells with dihydrotestosterone (DHT), which 101 upregulates SREBP-1c and SREBP-2, did not cause a concommitant increase in prostasin levels (Chen et al., 2006a). DHT was found to upregulate Slug in LNCaP cells, and the repressive action of Slug negated any induction of prostasin by SREBP-1c and SREBP-2.
This raises interesting possibilities in the regulation of RKIP. RKIP has putative androgen response elements (AREs) and sterol regulatory elements in its promoter
(information not yet published). RKIP transcriptioanal activation is unknown, and androgen induction of RKIP through either AREs or alternatively through SREBP-1c or
SREBP-2s would be very enlightening areas of investigation. Furthermore, Chen et al. established that Slug could be upregulated in androgen receptor-lacking DU145 cell line, and posited that hormone-refractory prostate cancers may activate Slug through EGF
(Chen et al., 2006a). Following EGF treatment, mRNA levels of Slug, and to a lesser extent of Snail, had increased (Chen et al., 2006a). Signal cascades leading from EGF is believed to possibly upregulate Snail through Stat3 and Liv1 {Barrallo-Gimeno, 2005
#448}(Mann et al., 2006; Nieto, 2002; Przybylo and Radisky, 2007). RKIP may be differentially regulated by Snail and Slug, as well as SREBPs, androgen, and EGF signaling during normal cell processes and cancer dysregulation.
The many “players” in cancer signaling makes unravelling the signaling pathways of Snail and RKIP regulation a daunting task. It makes sense to attempt to concentrate on
areas believed to be most important while planning bench experiments. We have attempted to find important relationships between molecules using microarray data analysis, using data and equipment that was freely, or inexpensively available. We had hoped to clarify a group or cluster of genes to which RKIP belonged that clearly pointed 102 to function or means of regulation. The picture created through hierarchical cluster
analysis did indeed show RKIP clustering with other genes, but there was no consensus that, for example, placed RKIP with other serine proteases as a group. Despite the many similarites between RKIP and prostasin, these genes did not cluster closely together.
Prostasin is represented in two of the three datasets utilized here, and initial observation is that it is positively correlated to RKIP; however, this study stopped short of investigating this correlation. Several transcription factors were clustered with RKIP in our analysis, including RUNX1, shown to be an inhibitor of cell proliferation (Table 8)
(Himes et al., 2005) While RUNX1 has mostly been studied as a trascriptional activator, it can also associate with mSin3A and histone deacetylases to suppress transcription
(Lutterbach et al., 1998; Yamagata et al., 2005). The hierarchical clustering study provided several canditates for further exploration as RKIP regulators, and indeed, clustering is often used to generate hypotheses. This type of analysis was less illuminating for Snail. Fewer genes clustered closely with Snail than had with RKIP, and
Dataset 2 gave disappointingly low node correlation values. EGFR was found clustered with Snail, albeit it was not in the most closely correlated group. In a different approach to finding possible correlations between Snail and its regulators, we calculated correlation coefficients for specific genes found in the microarrays based on information in the literature. We looked for statistically significant correlations between Snail and factors in
Wnt signaling in particular, as dysregulated Wnt signaling can trigger EMT and Snail is part of a “β-catenin-TCF-regulated Axin2-GSK3β-Snail axis” (Yook et al., 2006). No significant correlation was found between Snail and members of the Wnt signaling 103 network represented in the arrays. EGFR was the only molecule that was both represented in the arrays and a known part of a snail regulatory trail that had a significant positive correlation to Snail (Table 11). It is difficult to make conclusions about Snail regulation in prosate cancer using this type of analysis, but it is interesting that EGF signaling was previously demonstrated by Chen et al. to be able to upregulate Snail and
Slug in a hormone-refractive prostate cancer cel line, and again we find it in the array analysis (Chen et al., 2006a).
This work identifies Snail as a negative transcriptional regulator of RKIP and supports evidence that RKIP downregulation is involved in prostate cancer progression.
By extracting specific expression profiles from the datasets we were able to analyze correlations and relationships between genes of interest. This is particularly important as most microarray experiments report only those genes with the highest fold difference in their experiment or area of interest. For example, RKIP was not reported as one of the genes with the biggest reduction in expression amongst the thousands included on the array (Dhanasekaran et al., 2001; Varambally et al., 2002). Nor was Snail a gene with the highest upregulation in these arrays, but that does not preclude them from having a significant role in prostate cancer.
104
SUMMARY
The molecular and bioinformatic experiments in this study demonstrate how Snail is related to RKIP transcriptional repression in prostate cancer. Endogenous RKIP
expression levels are inverse to that of Snail and can be repressed or relieved by
modulation of Snail in prostate cancer cell lines. Luciferase reporter assays show that
RKIP promoter activity can be inhibited by Snail and Slug, and ChIP assays have revealed that Snail may be directly binding E-box elements in the RKIP promoter.
Further study should reveal the exact E-box element(s) necessary for the repression by
Snail. Snail was found to be significantly negatively correlated to RKIP in prostate cancer microarray analysis, suggesting that Snail is a physiologically relevant regulator of
RKIP in prostate cancer progression from primary tumor to metastasic cancer.
Hierarchical clustering analyses identified dozens of genes that may be involved in the
regulation of either RKIP or Snail in prostate cancer.
105 BIBLIOGRAPHY
Abdelhaleem, M., Maltais, L., and Wain, H. (2003). The human DDX and DHX gene families of putative RNA helicases. Genomics 81, 618-622. Abraham, R. S., Ballman, K. V., Dispenzieri, A., Grill, D. E., Manske, M. K., Price- Troska, T. L., Paz, N. G., Gertz, M. A., and Fonseca, R. (2005). Functional gene expression analysis of clonal plasma cells identifies a unique molecular profile for light chain amyloidosis. Blood 105, 794-803. Altinok, G., Powell, I. J., Che, M., Hormont, K., Sarkar, F. H., Sakr, W. A., Grignon, D., and Liao, D. J. (2006). Reduction of QM protein expression correlates with tumor grade in prostatic adenocarcinoma. Prostate Cancer Prostatic Dis 9, 77-82. Altuwaijri, S., Lin, H. K., Chuang, K. H., Lin, W. J., Yeh, S., Hanchett, L. A., Rahman, M. M., Kang, H. Y., Tsai, M. Y., Zhang, Y., et al. (2003). Interruption of nuclear factor kappaB signaling by the androgen receptor facilitates 12-O- tetradecanoylphorbolacetate-induced apoptosis in androgen-sensitive prostate cancer LNCaP cells. Cancer Res 63, 7106-7112. Amaya, I., Ratcliffe, O. J., and Bradley, D. J. (1999). Expression of CENTRORADIALIS (CEN) and CEN-like genes in tobacco reveals a conserved mechanism controlling phase change in diverse species. Plant Cell 11, 1405-1418. Amler, L. C., Schurmann, J., and Schwab, M. (1996). The DDX1 gene maps within 400 kbp 5' to MYCN and is frequently coamplified in human neuroblastoma. Genes Chromosomes Cancer 15, 134-137. Banfield, M. J., and Brady, R. L. (2000). The structure of Antirrhinum centroradialis protein (CEN) suggests a role as a kinase regulator. J Mol Biol 297, 1159-1170. Barbera, M. J., Puig, I., Dominguez, D., Julien-Grille, S., Guaita-Esteruelas, S., Peiro, S., Baulida, J., Franci, C., Dedhar, S., Larue, L., and Garcia de Herreros, A. (2004). Regulation of Snail transcription during epithelial to mesenchymal transition of tumor cells. Oncogene 23, 7345-7354. Barrallo-Gimeno, A., and Nieto, M. A. (2005). The Snail genes as inducers of cell movement and survival: implications in development and cancer. Development 132, 3151-3161. Batlle, E., Sancho, E., Franci, C., Dominguez, D., Monfar, M., Baulida, J., and Garcia De Herreros, A. (2000). The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nat Cell Biol 2, 84-89. Bence, N. F., Sampat, R. M., and Kopito, R. R. (2001). Impairment of the ubiquitin- proteasome system by protein aggregation. Science 292, 1552-1555. Bernier, I., and Jolles, P. (1984). Purification and characterization of a basic 23 kDa cytosolic protein from bovine brain. Biochim Biophys Acta 790, 174-181. Bernier, I., Tresca, J. P., and Jolles, P. (1986). Ligand-binding studies with a 23 kDa protein purified from bovine brain cytosol. Biochim Biophys Acta 871, 19-23.
106 Bhui-Kaur, A., Therwath, A., Henry, L., Chiesa, J., Kurkure, A., Scherrer, K., and Bureau, J. P. (1998). Increased prosomal proteins in breast cancer cells and in neighboring normal cells in Parsi and non-Parsi populations. J Cancer Res Clin Oncol 124, 117-126. Bian, D., Su, S., Mahanivong, C., Cheng, R. K., Han, Q., Pan, Z. K., Sun, P., and Huang, S. (2004). Lysophosphatidic Acid Stimulates Ovarian Cancer Cell Migration via a Ras-MEK Kinase 1 Pathway. Cancer Res 64, 4209-4217. Bioscience, C. (2007). Oncomine Research (Compendia Bioscience). Bird, P. I. (1999). Regulation of pro-apoptotic leucocyte granule serine proteinases by intracellular serpins. Immunol Cell Biol 77, 47-57. Blanco, M. J., Moreno-Bueno, G., Sarrio, D., Locascio, A., Cano, A., Palacios, J., and Nieto, M. A. (2002). Correlation of Snail expression with histological grade and lymph node status in breast carcinomas. Oncogene 21, 3241-3246. Bollengier, F., and Mahler, A. (1980). Isolation an dpartial biochemical characrerizaion of an unknown polypeptide from white and grey brain material. Neuropeptides 1, 119-135. Bolos, V., Peinado, H., Perez-Moreno, M. A., Fraga, M. F., Esteller, M., and Cano, A. (2003). The transcription factor Slug represses E-cadherin expression and induces epithelial to mesenchymal transitions: a comparison with Snail and E47 repressors. J Cell Sci 116, 499-511. Bouquet, C., Soares, S., von Boxberg, Y., Ravaille-Veron, M., Propst, F., and Nothias, F. (2004). Microtubule-associated protein 1B controls directionality of growth cone migration and axonal branching in regeneration of adult dorsal root ganglia neurons. J Neurosci 24, 7204-7213. Bradley, D., Carpenter, R., Copsey, L., Vincent, C., Rothstein, S., and Coen, E. (1996). Control of inflorescence architecture in Antirrhinum. Nature 379, 791-797. Bradley, D., Ratcliffe, O., Vincent, C., Carpenter, R., and Coen, E. (1997). Inflorescence commitment and architecture in Arabidopsis. Science 275, 80-83. Bruun, A. W., Svendsen, I., Sorensen, S. O., Kielland-Brandt, M. C., and Winther, J. R. (1998). A high-affinity inhibitor of yeast carboxypeptidase Y is encoded by TFS1 and shows homology to a family of lipid binding proteins. Biochemistry 37, 3351- 3357. Cano, A., Perez-Moreno, M. A., Rodrigo, I., Locascio, A., Blanco, M. J., del Barrio, M. G., Portillo, F., and Nieto, M. A. (2000). The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat Cell Biol 2, 76-83. Castro Alves, C., Rosivatz, E., Schott, C., Hollweck, R., Becker, I., Sarbia, M., Carneiro, F., and Becker, K. F. (2007). Slug is overexpressed in gastric carcinomas and may act synergistically with SIP1 and Snail in the down-regulation of E-cadherin. J Pathol 211, 507-515. Cavallo, F., Astolfi, A., Iezzi, M., Cordero, F., Lollini, P. L., Forni, G., and Calogero, R. (2005). An integrated approach of immunogenomics and bioinformatics to
107 identify new Tumor Associated Antigens (TAA) for mammary cancer immunological prevention. BMC Bioinformatics 6 Suppl 4, S7. Chatterjee, D., Bai, Y., Wang, Z., Beach, S., Mott, S., Roy, R., Braastad, C., Sun, Y., Mukhopadhyay, A., Aggarwal, B. B., et al. (2004). RKIP sensitizes prostate and breast cancer cells to drug-induced apoptosis. J Biol Chem 279, 17515-17523. Chautard, H., Jacquet, M., Schoentgen, F., Bureaud, N., and Benedetti, H. (2004). Tfs1p, a member of the PEBP family, inhibits the Ira2p but not the Ira1p Ras GTPase- activating protein in Saccharomyces cerevisiae. Eukaryot Cell 3, 459-470. Chen, L. M., and Chai, K. X. (2002). Prostasin serine protease inhibits breast cancer invasiveness and is transcriptionally regulated by promoter DNA methylation. Int J Cancer 97, 323-329. Chen, L. M., Hodge, G. B., Guarda, L. A., Welch, J. L., Greenberg, N. M., and Chai, K. X. (2001a). Down-regulation of prostasin serine protease: a potential invasion suppressor in prostate cancer. Prostate 48, 93-103. Chen, L. M., Skinner, M. L., Kauffman, S. W., Chao, J., Chao, L., Thaler, C. D., and Chai, K. X. (2001b). Prostasin is a glycosylphosphatidylinositol-anchored active serine protease. J Biol Chem 276, 21434-21442. Chen, L. M., Zhang, X., and Chai, K. X. (2004). Regulation of prostasin expression and function in the prostate. Prostate 59, 1-12. Chen, M., Chen, L. M., and Chai, K. X. (2006a). Androgen regulation of prostasin gene expression is mediated by sterol-regulatory element-binding proteins and SLUG. Prostate 66, 911-920. Chen, Q., Wang, S., Thompson, S. N., Hall, E. D., and Guttmann, R. P. (2006b). Identification and characterization of PEBP as a calpain substrate. J Neurochem 99, 1133-1141. Cho, H. J., Baek, K. E., Saika, S., Jeong, M. J., and Yoo, J. (2007). Snail is required for transforming growth factor-beta-induced epithelial-mesenchymal transition by activating PI3 kinase/Akt signal pathway. Biochem Biophys Res Commun 353, 337-343. Coppola, M., Pizzigoni, A., Banfi, S., Bassi, M. T., Casari, G., and Incerti, B. (2000). Identification and characterization of YME1L1, a novel paraplegin-related gene. Genomics 66, 48-54. Corbit, K. C., Trakul, N., Eves, E. M., Diaz, B., Marshall, M., and Rosner, M. R. (2003). Activation of Raf-1 Signaling by Protein Kinase C through a Mechanism Involving Raf Kinase Inhibitory Protein. J Biol Chem 278, 13061-13068. Cui, F., Wang, Y., Wang, J., Wei, K., Hu, J., Liu, F., Wang, H., Zhao, X., Zhang, X., and Yang, X. (2006). The up-regulation of proteasome subunits and lysosomal proteases in hepatocellular carcinomas of the HBx gene knockin transgenic mice. Proteomics 6, 498-504. De Craene, B., van Roy, F., and Berx, G. (2005). Unraveling signalling cascades for the Snail family of transcription factors. Cell Signal 17, 535-547. Dellino, G. I., Schwartz, Y. B., Farkas, G., McCabe, D., Elgin, S. C., and Pirrotta, V. (2004). Polycomb silencing blocks transcription initiation. Mol Cell 13, 887-893. 108 Dhanasekaran, S. M., Barrette, T. R., Ghosh, D., Shah, R., Varambally, S., Kurachi, K., Pienta, K. J., Rubin, M. A., and Chinnaiyan, A. M. (2001). Delineation of prognostic biomarkers in prostate cancer. Nature 412, 822-826. Duden, R., Griffiths, G., Frank, R., Argos, P., and Kreis, T. E. (1991). Beta-COP, a 110 kd protein associated with non-clathrin-coated vesicles and the Golgi complex, shows homology to beta-adaptin. Cell 64, 649-665. Dunn, A. Y., Melville, M. W., and Frydman, J. (2001). Review: cellular substrates of the eukaryotic chaperonin TRiC/CCT. J Struct Biol 135, 176-184. Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. (1998). Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A 95, 14863-14868. Eves, E. M., Shapiro, P., Naik, K., Klein, U. R., Trakul, N., and Rosner, M. R. (2006). Raf kinase inhibitory protein regulates aurora B kinase and the spindle checkpoint. Mol Cell 23, 561-574. Frayne, J., McMillen, A., Love, S., and Hall, L. (1998). Expression of phosphatidylethanolamine-binding protein in the male reproductive tract: immunolocalisation and expression in prepubertal and adult rat testes and epididymides. Mol Reprod Dev 49, 454-460. Fu, Z., Kitagawa, Y., Shen, R., Shah, R., Mehra, R., Rhodes, D., Keller, P. J., Mizokami, A., Dunn, R., Chinnaiyan, A. M., et al. (2006). Metastasis suppressor gene Raf kinase inhibitor protein (RKIP) is a novel prognostic marker in prostate cancer. Prostate 66, 248-256. Fu, Z., Smith, P. C., Zhang, L., Rubin, M. A., Dunn, R. L., Yao, Z., and Keller, E. T. (2003). Effects of raf kinase inhibitor protein expression on suppression of prostate cancer metastasis. J Natl Cancer Inst 95, 878-889. Fujita, N., Jaye, D. L., Kajita, M., Geigerman, C., Moreno, C. S., and Wade, P. A. (2003). MTA3, a Mi-2/NuRD complex subunit, regulates an invasive growth pathway in breast cancer. Cell 113, 207-219. Gil, J., Bernard, D., and Peters, G. (2005). Role of polycomb group proteins in stem cell self-renewal and cancer. DNA Cell Biol 24, 117-125. Glockzin, S., Ogi, F. X., Hengstermann, A., Scheffner, M., and Blattner, C. (2003). Involvement of the DNA repair protein hHR23 in p53 degradation. Mol Cell Biol 23, 8960-8969. Goel, R., and Baldassare, J. J. (2004). Killing two birds with one RKIP. Trends Endocrinol Metab 15, 91-92. Gomperts, B. D., Tatham, P. E. R., and Kramer, I. M. (2002). Signal Transduction, Academic Press/Elsevier). Goumon, Y., Angelone, T., Schoentgen, F., Chasserot-Golaz, S., Almas, B., Fukami, M. M., Langley, K., Welters, I. D., Tota, B., Aunis, D., and Metz-Boutigue, M. H. (2004). The hippocampal cholinergic neurostimulating peptide, the N-terminal fragment of the secreted phosphatidylethanolamine-binding protein, possesses a new biological activity on cardiac physiology. J Biol Chem 279, 13054-13064.
109 Grandy, D. K., Hanneman, E., Bunzow, J., Shih, M., Machida, C. A., Bidlack, J. M., and Civelli, O. (1990). Purification, cloning, and tissue distribution of a 23-kDa rat protein isolated by morphine affinity chromatography. Mol Endocrinol 4, 1370- 1376. Grau, Y., Carteret, C., and Simpson, P. (1984). Mutations and Chromosomal Rearrangements Affecting the Expression of Snail, a Gene Involved in Embryonic Patterning in DROSOPHILA MELANOGASTER. Genetics 108, 347-360. Grimes, H. L., Chan, T. O., Zweidler-McKay, P. A., Tong, B., and Tsichlis, P. N. (1996). The Gfi-1 proto-oncoprotein contains a novel transcriptional repressor domain, SNAG, and inhibits G1 arrest induced by interleukin-2 withdrawal. Mol Cell Biol 16, 6263-6272. Grooteclaes, M. L., and Frisch, S. M. (2000). Evidence for a function of CtBP in epithelial gene regulation and anoikis. Oncogene 19, 3823-3828. Grzmil, M., Thelen, P., Hemmerlein, B., Schweyer, S., Voigt, S., Mury, D., and Burfeind, P. (2003). Bax inhibitor-1 is overexpressed in prostate cancer and its specific down-regulation by RNA interference leads to cell death in human prostate carcinoma cells. Am J Pathol 163, 543-552. Hagan, S., Al-Mulla, F., Mallon, E., Oien, K., Ferrier, R., Gusterson, B., Garcia, J. J., and Kolch, W. (2005). Reduction of Raf-1 kinase inhibitor protein expression correlates with breast cancer metastasis. Clin Cancer Res 11, 7392-7397. Hajra, K. M., Chen, D. Y., and Fearon, E. R. (2002). The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res 62, 1613-1618. Hajra, K. M., Ji, X., and Fearon, E. R. (1999). Extinction of E-cadherin expression in breast cancer via a dominant repression pathway acting on proximal promoter elements. Oncogene 18, 7274-7279. Hamosh, A., and Hartz, P. A. (1966-2004). Online Mendelian Inheritance in Man (Johns Hopkins University). Helfer, E., Nevalainen, E. M., Naumanen, P., Romero, S., Didry, D., Pantaloni, D., Lappalainen, P., and Carlier, M. F. (2006). Mammalian twinfilin sequesters ADP- G-actin and caps filament barbed ends: implications in motility. Embo J 25, 1184- 1195. Hengst, U., Albrecht, H., Hess, D., and Monard, D. (2001). The phosphatidylethanolamine-binding protein is the prototype of a novel family of serine protease inhibitors. J Biol Chem 276, 535-540. Hickox, D. M., Gibbs, G., Morrison, J. R., Sebire, K., Edgar, K., Keah, H. H., Alter, K., Loveland, K. L., Hearn, M. T., de Kretser, D. M., and O'Bryan, M. K. (2002). Identification of a novel testis-specific member of the phosphatidylethanolamine binding protein family, pebp-2. Biol Reprod 67, 917-927. Hilger, R. A., Scheulen, M. E., and Strumberg, D. (2002). The Ras-Raf-MEK-ERK pathway in the treatment of cancer. Onkologie 25, 511-518. Himes, S. R., Cronau, S., Mulford, C., and Hume, D. A. (2005). The Runx1 transcription factor controls CSF-1-dependent and -independent growth and survival of macrophages. Oncogene 24, 5278-5286. 110 Hoshino, S., Imai, M., Mizutani, M., Kikuchi, Y., Hanaoka, F., Ui, M., and Katada, T. (1998). Molecular cloning of a novel member of the eukaryotic polypeptide chain-releasing factors (eRF). Its identification as eRF3 interacting with eRF1. J Biol Chem 273, 22254-22259. Huang, J., Mahavadi, S., Sriwai, W., Grider, J. R., and Murthy, K. S. (2007). Cross- regulation of VPAC(2) receptor desensitization by M(3) receptors via PKC- mediated phosphorylation of RKIP and inhibition of GRK2. Am J Physiol Gastrointest Liver Physiol 292, G867-874. Huang, X., Wang, H., Xu, M., Lu, L., Xu, Z., Li, J., Zhou, Z., and Sha, J. (2004). Expression of a novel RAD23B mRNA splice variant in the human testis. J Androl 25, 363-368. Huerta-Yepez, S., Vega, M., Jazirehi, A., Garban, H., Hongo, F., Cheng, G., and Bonavida, B. (2004). Nitric oxide sensitizes prostate carcinoma cell lines to TRAIL-mediated apoptosis via inactivation of NF-kappa B and inhibition of Bcl- xl expression. Oncogene 23, 4993-5003. Imafuku, I., Masaki, T., Waragai, M., Takeuchi, S., Kawabata, M., Hirai, S., Ohno, S., Nee, L. E., Lippa, C. F., Kanazawa, I., et al. (1999). Presenilin 1 suppresses the function of c-Jun homodimers via interaction with QM/Jif-1. J Cell Biol 147, 121- 134. Iwakiri, M., Mizukami, K., Ikonomovic, M. D., Ishikawa, M., Hidaka, S., Abrahamson, E. E., DeKosky, S. T., and Asada, T. (2005). Changes in hippocampal GABABR1 subunit expression in Alzheimer's patients: association with Braak staging. Acta Neuropathol (Berl) 109, 467-474. Jazirehi, A. R., Vega, M. I., Chatterjee, D., Goodglick, L., and Bonavida, B. (2004). Inhibition of the Raf-MEK1/2-ERK1/2 signaling pathway, Bcl-xL down- regulation, and chemosensitization of non-Hodgkin's lymphoma B cells by Rituximab. Cancer Res 64, 7117-7126. Jiao, W., Miyazaki, K., and Kitajima, Y. (2002). Inverse correlation between E-cadherin and Snail expression in hepatocellular carcinoma cell lines in vitro and in vivo. Br J Cancer 86, 98-101. Joshi, B., Rastogi, S., Morris, M., Carastro, L. M., DeCook, C., Seto, E., and Chellappan, S. P. (2007). Differential regulation of human YY1 and caspase 7 promoters by prohibitin through E2F1 and p53 binding sites. Biochem J 401, 155-166. Kajita, M., McClinic, K. N., and Wade, P. A. (2004). Aberrant expression of the transcription factors snail and slug alters the response to genotoxic stress. Mol Cell Biol 24, 7559-7566. Kaplan, R., Zaheer, A., Jaye, M., and Lim, R. (1991). Molecular cloning and expression of biologically active human glia maturation factor-beta. J Neurochem 57, 483- 490. Kardailsky, I., Shukla, V. K., Ahn, J. H., Dagenais, N., Christensen, S. K., Nguyen, J. T., Chory, J., Harrison, M. J., and Weigel, D. (1999). Activation tagging of the floral inducer FT. Science 286, 1962-1965.
111 Katada, E., Mitake, S., Matsukawa, N., Otsuka, Y., Tsugu, Y., Fujimori, O., and Ojika, K. (1996). Distribution of hippocampal cholinergic neurostimulating peptide (HCNP)-like immunoreactivity in organs and tissues of young Wistar rats. Histochem Cell Biol 105, 43-51. Katoh, M., and Katoh, M. (2006). Cross-talk of WNT and FGF signaling pathways at GSK3beta to regulate beta-catenin and SNAIL signaling cascades. Cancer Biol Ther 5, 1059-1064. Kazuki, Y., Kimura, M., Nishigaki, R., Kai, Y., Abe, S., Okita, C., Shirayoshi, Y., Schulz, T. C., Tomizuka, K., Hanaoka, K., et al. (2004). Human chromosome 21q22.2-qter carries a gene(s) responsible for downregulation of mlc2a and PEBP in Down syndrome model mice. Biochem Biophys Res Commun 317, 491-499. Keller, E. T., Fu, Z., and Brennan, M. (2005). The biology of a prostate cancer metastasis suppressor protein: Raf kinase inhibitor protein. J Cell Biochem 94, 273-278. Kim, H. J., Litzenburger, B. C., Cui, X., Delgado, D. A., Grabiner, B. C., Lin, X., Lewis, M. T., Gottardis, M. M., Wong, T. W., Attar, R. M., et al. (2007). Constitutively active type I insulin-like growth factor receptor causes transformation and xenograft growth of immortalized mammary epithelial cells and is accompanied by an epithelial-to-mesenchymal transition mediated by NF-kappaB and snail. Mol Cell Biol 27, 3165-3175. Knight, R. D., and Shimeld, S. M. (2001). Identification of conserved C2H2 zinc-finger gene families in the Bilateria. Genome Biol 2, RESEARCH0016. Kolch, W. (2000). Meaningful relationships: the regulation of the Ras/Raf/MEK/ERK pathway by protein interactions. Biochem J 351 Pt 2, 289-305. Kong, A., Gudbjartsson, D. F., Sainz, J., Jonsdottir, G. M., Gudjonsson, S. A., Richardsson, B., Sigurdardottir, S., Barnard, J., Hallbeck, B., Masson, G., et al. (2002). A high-resolution recombination map of the human genome. Nat Genet 31, 241-247. Kroslak, T., Koch, T., Kahl, E., and Hollt, V. (2001). Human phosphatidylethanolamine- binding protein facilitates heterotrimeric G protein-dependent signaling. J Biol Chem 276, 39772-39778. Kuzmichev, A., Margueron, R., Vaquero, A., Preissner, T. S., Scher, M., Kirmizis, A., Ouyang, X., Brockdorff, N., Abate-Shen, C., Farnham, P., and Reinberg, D. (2005). Composition and histone substrates of polycomb repressive group complexes change during cellular differentiation. Proc Natl Acad Sci U S A 102, 1859-1864. Kwak, J. Y., Ma, T. Z., Yoo, M. J., Choi, B. H., Kim, H. G., Kim, S. R., Yim, C. Y., and Kwak, Y. G. (2004). The comparative analysis of serum proteomes for the discovery of biomarkers for acute myeloid leukemia. Exp Hematol 32, 836-842. Lakhani, S. A., Masud, A., Kuida, K., Porter, G. A., Jr., Booth, C. J., Mehal, W. Z., Inayat, I., and Flavell, R. A. (2006). Caspases 3 and 7: key mediators of mitochondrial events of apoptosis. Science 311, 847-851.
112 Lee, H. C., Tian, B., Sedivy, J. M., Wands, J. R., and Kim, M. (2006). Loss of Raf kinase inhibitor protein promotes cell proliferation and migration of human hepatoma cells. Gastroenterology 131, 1208-1217. Liszewski, M. K., Kemper, C., Price, J. D., and Atkinson, J. P. (2005). Emerging roles and new functions of CD46. Springer Semin Immunopathol 27, 345-358. Lorenz, K., Lohse, M. J., and Quitterer, U. (2003). Protein kinase C switches the Raf kinase inhibitor from Raf-1 to GRK-2. Nature 426, 574-579. Lutterbach, B., Westendorf, J. J., Linggi, B., Patten, A., Moniwa, M., Davie, J. R., Huynh, K. D., Bardwell, V. J., Lavinsky, R. M., Rosenfeld, M. G., et al. (1998). ETO, a target of t(8;21) in acute leukemia, interacts with the N-CoR and mSin3 corepressors. Mol Cell Biol 18, 7176-7184. Maglott, D., Ostell, J., Pruitt, K. D., and Tatusova, T. (2005). Entrez Gene: gene-centered information at NCBI. Nucleic Acids Res 33, D54-58. Malinin, N. L., Boldin, M. P., Kovalenko, A. V., and Wallach, D. (1997). MAP3K- related kinase involved in NF-kappaB induction by TNF, CD95 and IL-1. Nature 385, 540-544. Mann, J. R., Backlund, M. G., Buchanan, F. G., Daikoku, T., Holla, V. R., Rosenberg, D. W., Dey, S. K., and DuBois, R. N. (2006). Repression of prostaglandin dehydrogenase by epidermal growth factor and snail increases prostaglandin E2 and promotes cancer progression. Cancer Res 66, 6649-6656. Manohar, C. F., Salwen, H. R., Brodeur, G. M., and Cohn, S. L. (1995). Co-amplification and concomitant high levels of expression of a DEAD box gene with MYCN in human neuroblastoma. Genes Chromosomes Cancer 14, 196-203. Martinez, O., Antony, C., Pehau-Arnaudet, G., Berger, E. G., Salamero, J., and Goud, B. (1997). GTP-bound forms of rab6 induce the redistribution of Golgi proteins into the endoplasmic reticulum. Proc Natl Acad Sci U S A 94, 1828-1833. Mauhin, V., Lutz, Y., Dennefeld, C., and Alberga, A. (1993). Definition of the DNA- binding site repertoire for the Drosophila transcription factor SNAIL. Nucleic Acids Res 21, 3951-3957. Mc Henry, K. T., Montesano, R., Zhu, S., Beshir, A. B., Tang, H. H., Yeung, K. C., and Fenteany, G. (2008). Raf kinase inhibitor protein positively regulates cell- substratum adhesion while negatively regulating cell-cell adhesion. J Cell Biochem 103, 972-985. Melle, C., Camacho, J. A., Surber, R., Betge, S., Von Eggeling, F., and Zimmer, T. (2006). Region-specific alterations of global protein expression in the remodelled rat myocardium. Int J Mol Med 18, 1207-1215. Melville, M. W., McClellan, A. J., Meyer, A. S., Darveau, A., and Frydman, J. (2003). The Hsp70 and TRiC/CCT chaperone systems cooperate in vivo to assemble the von Hippel-Lindau tumor suppressor complex. Mol Cell Biol 23, 3141-3151. Mermelstein, F., Yeung, K., Cao, J., Inostroza, J. A., Erdjument-Bromage, H., Eagelson, K., Landsman, D., Levitt, P., Tempst, P., and Reinberg, D. (1996). Requirement of a corepressor for Dr1-mediated repression of transcription. Genes Dev 10, 1033-1048. 113 Miao, F., Bouziane, M., Dammann, R., Masutani, C., Hanaoka, F., Pfeifer, G., and O'Connor, T. R. (2000). 3-Methyladenine-DNA glycosylase (MPG protein) interacts with human RAD23 proteins. J Biol Chem 275, 28433-28438. Mikhail, F. M., Sinha, K. K., Saunthararajah, Y., and Nucifora, G. (2006). Normal and transforming functions of RUNX1: a perspective. J Cell Physiol 207, 582-593. Mikolajczyk, S. D., Millar, L. S., Marker, K. M., Rittenhouse, H. G., Wolfert, R. L., Marks, L. S., Charlesworth, M. C., and Tindall, D. J. (1999). Identification of a novel complex between human kallikrein 2 and protease inhibitor-6 in prostate cancer tissue. Cancer Res 59, 3927-3930. Mima, J., Narita, Y., Chiba, H., and Hayashi, R. (2003). The multiple site binding of carboxypeptidase Y inhibitor (IC) to the cognate proteinase. Implications for the biological roles of the phosphatidylethanolamine-binding protein. J Biol Chem 278, 29792-29798. Min, J., Zhang, Y., and Xu, R. M. (2003). Structural basis for specific binding of Polycomb chromodomain to histone H3 methylated at Lys 27. Genes Dev 17, 1823-1828. Minoo, P., Baker, K., Goswami, R., Chong, G., Foulkes, W. D., Ruszkiewicz, A. R., Barker, M., Buchanan, D., Young, J., and Jass, J. R. (2006). Extensive DNA methylation in normal colorectal mucosa in hyperplastic polyposis. Gut 55, 1467- 1474. Minoo, P., Zlobec, I., Baker, K., Tornillo, L., Terracciano, L., Jass, J. R., and Lugli, A. (2007). Loss of raf-1 kinase inhibitor protein expression is associated with tumor progression and metastasis in colorectal cancer. Am J Clin Pathol 127, 820-827. Miserey-Lenkei, S., Couedel-Courteille, A., Del Nery, E., Bardin, S., Piel, M., Racine, V., Sibarita, J. B., Perez, F., Bornens, M., and Goud, B. (2006). A role for the Rab6A' GTPase in the inactivation of the Mad2-spindle checkpoint. Embo J 25, 278-289. Molinaro, R. J., Jha, B. K., Malathi, K., Varambally, S., Chinnaiyan, A. M., and Silverman, R. H. (2006). Selection and cloning of poly(rC)-binding protein 2 and Raf kinase inhibitor protein RNA activators of 2',5'-oligoadenylate synthetase from prostate cancer cells. Nucleic Acids Res 34, 6684-6695. Monteclaro, F. S., and Vogt, P. K. (1993). A Jun-binding protein related to a putative tumor suppressor. Proc Natl Acad Sci U S A 90, 6726-6730. Moore, C., Perry, A. C., Love, S., and Hall, L. (1996). Sequence analysis and immunolocalisation of phosphatidylethanolamine binding protein (PBP) in human brain tissue. Brain Res Mol Brain Res 37, 74-78. Mori, T., Li, Y., Hata, H., and Kochi, H. (2004). NIRF is a ubiquitin ligase that is capable of ubiquitinating PCNP, a PEST-containing nuclear protein. FEBS Lett 557, 209- 214. Muto, S., Senda, M., Akai, Y., Sato, L., Suzuki, T., Nagai, R., Senda, T., and Horikoshi, M. (2007). Relationship between the structure of SET/TAF-Ibeta/INHAT and its histone chaperone activity. Proc Natl Acad Sci U S A 104, 4285-4290.
114 Nakayama, H., Scott, I. C., and Cross, J. C. (1998). The transition to endoreduplication in trophoblast giant cells is regulated by the mSNA zinc finger transcription factor. Dev Biol 199, 150-163. NCBI (2003). Mouse Genome Resources (The National Center for Biotechnology Information). Nieto, M. A. (2002). The snail superfamily of zinc-finger transcription factors. Nat Rev Mol Cell Biol 3, 155-166. NIH (2004). Homologene:37637 (The National Center for Biotechnology Information). Ninomiya-Tsuji, J., Kishimoto, K., Hiyama, A., Inoue, J., Cao, Z., and Matsumoto, K. (1999). The kinase TAK1 can activate the NIK-I kappaB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 398, 252-256. Nishiwaki, A., Asai, K., Tada, T., Ueda, T., Shimada, S., Ogura, Y., and Kato, T. (2001). Expression of glia maturation factor during retinal development in the rat. Brain Res Mol Brain Res 95, 103-109. Nixon, B., MacIntyre, D. A., Mitchell, L. A., Gibbs, G. M., O'Bryan, M., and Aitken, R. J. (2006). The identification of mouse sperm-surface-associated proteins and characterization of their ability to act as decapacitation factors. Biol Reprod 74, 275-287. Noguchi, T., Akiyama, K., Yokoyama, M., Kanda, N., Matsunaga, T., and Nishi, Y. (1996). Amplification of a DEAD box gene (DDX1) with the MYCN gene in neuroblastomas as a result of cosegregation of sequences flanking the MYCN locus. Genes Chromosomes Cancer 15, 129-133. Odabaei, G., Chatterjee, D., Jazirehi, A. R., Goodglick, L., Yeung, K., and Bonavida, B. (2004). Raf-1 kinase inhibitor protein: structure, function, regulation of cell signaling, and pivotal role in apoptosis. Adv Cancer Res 91, 169-200. Ohshima, S., Murata, M., Sakamoto, W., Ogura, Y., and Motoyoshi, F. (1997). Cloning and molecular analysis of the Arabidopsis gene Terminal Flower 1. Mol Gen Genet 254, 186-194. Ojika, K., Kojima, S., Ueki, Y., Fukushima, N., Hayashi, K., and Yamamoto, M. (1992). Purification and structural analysis of hippocampal cholinergic neurostimulating peptide. Brain Res 572, 164-171. Olmeda, D., Jorda, M., Peinado, H., Fabra, A., and Cano, A. (2007). Snail silencing effectively suppresses tumour growth and invasiveness. Oncogene 26, 1862-1874. Papa, S., Sardanelli, A. M., Scacco, S., Petruzzella, V., Technikova-Dobrova, Z., Vergari, R., and Signorile, A. (2002). The NADH: ubiquinone oxidoreductase (complex I) of the mammalian respiratory chain and the cAMP cascade. J Bioenerg Biomembr 34, 1-10. Park, S., Rath, O., Beach, S., Xiang, X., Kelly, S. M., Luo, Z., Kolch, W., and Yeung, K. C. (2006). Regulation of RKIP binding to the N-region of the Raf-1 kinase. FEBS Lett 580, 6405-6412. Park, S., Yeung, M. L., Beach, S., Shields, J. M., and Yeung, K. C. (2005). RKIP downregulates B-Raf kinase activity in melanoma cancer cells. Oncogene.
115 Pearson, G., Robinson, F., Beers Gibson, T., Xu, B. E., Karandikar, M., Berman, K., and Cobb, M. H. (2001). Mitogen-activated protein (MAP) kinase pathways: regulation and physiological functions. Endocr Rev 22, 153-183. Peinado, H., Ballestar, E., Esteller, M., and Cano, A. (2004a). Snail mediates E-cadherin repression by the recruitment of the Sin3A/histone deacetylase 1 (HDAC1)/HDAC2 complex. Mol Cell Biol 24, 306-319. Peinado, H., Marin, F., Cubillo, E., Stark, H. J., Fusenig, N., Nieto, M. A., and Cano, A. (2004b). Snail and E47 repressors of E-cadherin induce distinct invasive and angiogenic properties in vivo. J Cell Sci 117, 2827-2839. Peinado, H., Quintanilla, M., and Cano, A. (2003). Transforming growth factor beta-1 induces snail transcription factor in epithelial cell lines: mechanisms for epithelial mesenchymal transitions. J Biol Chem 278, 21113-21123. Peiro, S., Escriva, M., Puig, I., Barbera, M. J., Dave, N., Herranz, N., Larriba, M. J., Takkunen, M., Franci, C., Munoz, A., et al. (2006). Snail1 transcriptional repressor binds to its own promoter and controls its expression. Nucleic Acids Res 34, 2077-2084. Pnueli, L., Carmel-Goren, L., Hareven, D., Gutfinger, T., Alvarez, J., Ganal, M., Zamir, D., and Lifschitz, E. (1998). The SELF-PRUNING gene of tomato regulates vegetative to reproductive switching of sympodial meristems and is the ortholog of CEN and TFL1. Development 125, 1979-1989. Poser, I., Dominguez, D., de Herreros, A. G., Varnai, A., Buettner, R., and Bosserhoff, A. K. (2001). Loss of E-cadherin expression in melanoma cells involves up- regulation of the transcriptional repressor Snail. J Biol Chem 276, 24661-24666. Przybylo, J. A., and Radisky, D. C. (2007). Matrix metalloproteinase-induced epithelial- mesenchymal transition: Tumor progression at Snail's pace. Int J Biochem Cell Biol. Qiu, W., Zhang, S., Xiao, C., Xu, W., Ma, Y., Liu, Y., and Wu, Q. (2003). Molecular cloning and characterization of a mouse spermatogenesis-related ring finger gene znf230. Biochem Biophys Res Commun 306, 347-353. RDevelopmentCoreTeam (2006). R: A language and environment for statistical computing. R Foundation for Statistical Computing (Vienna, Austria, R Foundation for Statistical Computing). Rhodes, D. R., Yu, J., Shanker, K., Deshpande, N., Varambally, R., Ghosh, D., Barrette, T., Pandey, A., and Chinnaiyan, A. M. (2004). ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia 6, 1-6. Ringrose, L., Rehmsmeier, M., Dura, J. M., and Paro, R. (2003). Genome-wide prediction of Polycomb/Trithorax response elements in Drosophila melanogaster. Dev Cell 5, 759-771. Robinson, L. C., and Tatchell, K. (1991). TFS1: a suppressor of cdc25 mutations in Saccharomyces cerevisiae. Mol Gen Genet 230, 241-250. Romano, L. A., and Runyan, R. B. (2000). Slug is an essential target of TGFbeta2 signaling in the developing chicken heart. Dev Biol 223, 91-102.
116 Rosner, M. R. (2007). MAP kinase meets mitosis: A role for Raf Kinase Inhibitory Protein in spindle checkpoint regulation. Cell Div 2, 1. Rybin, V., Ullrich, O., Rubino, M., Alexandrov, K., Simon, I., Seabra, M. C., Goody, R., and Zerial, M. (1996). GTPase activity of Rab5 acts as a timer for endocytic membrane fusion. Nature 383, 266-269. Saramaki, O. R., Tammela, T. L., Martikainen, P. M., Vessella, R. L., and Visakorpi, T. (2006). The gene for polycomb group protein enhancer of zeste homolog 2 (EZH2) is amplified in late-stage prostate cancer. Genes Chromosomes Cancer 45, 639-645. Savagner, P., Kusewitt, D. F., Carver, E. A., Magnino, F., Choi, C., Gridley, T., and Hudson, L. G. (2005). Developmental transcription factor slug is required for effective re-epithelialization by adult keratinocytes. J Cell Physiol 202, 858-866. Saxonov, S., Berg, P., and Brutlag, D. L. (2006). A genome-wide analysis of CpG dinucleotides in the human genome distinguishes two distinct classes of promoters. Proc Natl Acad Sci U S A 103, 1412-1417. Scheer, M., Prange, W., Petmecky, K., Schirmacher, P., Zoller, J. E., and Kubler, A. C. (2003). [Evaluation of her-2/neu amplification/overexpression in OSCC with fluorescence in situ hybridization (FISH) and immunohistochemistry]. Mund Kiefer Gesichtschir 7, 138-145. Schimmoller, F., Diaz, E., Muhlbauer, B., and Pfeffer, S. R. (1998). Characterization of a 76 kDa endosomal, multispanning membrane protein that is highly conserved throughout evolution. Gene 216, 311-318. Schnabel, C. A., Jacobs, Y., and Cleary, M. L. (2000). HoxA9-mediated immortalization of myeloid progenitors requires functional interactions with TALE cofactors Pbx and Meis. Oncogene 19, 608-616. Schoentgen, F., and Jolles, P. (1995). From structure to function: possible biological roles of a new widespread protein family binding hydrophobic ligands and displaying a nucleotide binding site. FEBS Lett 369, 22-26. Schuierer, M. M., Bataille, F., Hagan, S., Kolch, W., and Bosserhoff, A. K. (2004). Reduction in Raf kinase inhibitor protein expression is associated with increased Ras-extracellular signal-regulated kinase signaling in melanoma cell lines. Cancer Res 64, 5186-5192. Schuierer, M. M., Bataille, F., Weiss, T. S., Hellerbrand, C., and Bosserhoff, A. K. (2006a). Raf kinase inhibitor protein is downregulated in hepatocellular carcinoma. Oncol Rep 16, 451-456. Schuierer, M. M., Heilmeier, U., Boettcher, A., Ugocsai, P., Bosserhoff, A. K., Schmitz, G., and Langmann, T. (2006b). Induction of Raf kinase inhibitor protein contributes to macrophage differentiation. Biochem Biophys Res Commun 342, 1083-1087. Seddiqi, N., Bollengier, F., Alliel, P. M., Perin, J. P., Bonnet, F., Bucquoy, S., Jolles, P., and Schoentgen, F. (1994). Amino acid sequence of the Homo sapiens brain 21- 23-kDa protein (neuropolypeptide h3), comparison with its counterparts from
117 Rattus norvegicus and Bos taurus species, and expression of its mRNA in different tissues. J Mol Evol 39, 655-660. Shen, X. J., Ali-Fehmi, R., Weng, C. R., Sarkar, F. H., Grignon, D., and Liao, D. J. (2006). Loss of heterozygosity and microsatellite instability at the Xq28 and the A/G heterozygosity of the QM gene are associated with ovarian cancer. Cancer Biol Ther 5, 523-528. Shi, Y., Luo, S., Peng, J., Huang, C., Tan, D., and Hu, W. (2004). The structure, expression and function prediction of DAZAP2, a down-regulated gene in multiple myeloma. Genomics Proteomics Bioinformatics 2, 47-54. Shih, J. Y., Tsai, M. F., Chang, T. H., Chang, Y. L., Yuan, A., Yu, C. J., Lin, S. B., Liou, G. Y., Lee, M. L., Chen, J. J., et al. (2005). Transcription repressor slug promotes carcinoma invasion and predicts outcome of patients with lung adenocarcinoma. Clin Cancer Res 11, 8070-8078. Shim, S., Kim, Y., Shin, J., Kim, J., and Park, S. (2007). Regulation of EphA8 gene expression by TALE homeobox transcription factors during development of the mesencephalon. Mol Cell Biol 27, 1614-1630. Shin, H. W., Hayashi, M., Christoforidis, S., Lacas-Gervais, S., Hoepfner, S., Wenk, M. R., Modregger, J., Uttenweiler-Joseph, S., Wilm, M., Nystuen, A., et al. (2005). An enzymatic cascade of Rab5 effectors regulates phosphoinositide turnover in the endocytic pathway. J Cell Biol 170, 607-618. Shinohara, T., Tomizuka, K., Miyabara, S., Takehara, S., Kazuki, Y., Inoue, J., Katoh, M., Nakane, H., Iino, A., Ohguma, A., et al. (2001). Mice containing a human chromosome 21 model behavioral impairment and cardiac anomalies of Down's syndrome. Hum Mol Genet 10, 1163-1175. Simister, P. C., Banfield, M. J., and Brady, R. L. (2002). The crystal structure of PEBP-2, a homologue of the PEBP/RKIP family. Acta Crystallogr D Biol Crystallogr 58, 1077-1080. Snyder, H., Mensah, K., Theisler, C., Lee, J., Matouschek, A., and Wolozin, B. (2003). Aggregated and monomeric alpha-synuclein bind to the S6' proteasomal protein and inhibit proteasomal function. J Biol Chem 278, 11753-11759. Soung, Y. H., Lee, J. W., Kim, H. S., Park, W. S., Kim, S. Y., Lee, J. H., Park, J. Y., Cho, Y. G., Kim, C. J., Park, Y. G., et al. (2003). Inactivating mutations of CASPASE- 7 gene in human cancers. Oncogene 22, 8048-8052. Spagnoli, F. M., Cicchini, C., Tripodi, M., and Weiss, M. C. (2000). Inhibition of MMH (Met murine hepatocyte) cell differentiation by TGF(beta) is abrogated by pre- treatment with the heritable differentiation effector FGF1. J Cell Sci 113 ( Pt 20), 3639-3647. Squire, J. A., Thorner, P. S., Weitzman, S., Maggi, J. D., Dirks, P., Doyle, J., Hale, M., and Godbout, R. (1995). Co-amplification of MYCN and a DEAD box gene (DDX1) in primary neuroblastoma. Oncogene 10, 1417-1422. Stekel, D. (2003). Microarray Bioinformatics (New York, NY, USA, Cambridge University Press).
118 Stephens, B. J., Han, H., Gokhale, V., and Von Hoff, D. D. (2005). PRL phosphatases as potential molecular targets in cancer. Mol Cancer Ther 4, 1653-1661. Takano, S., Kanai, F., Jazag, A., Ijichi, H., Yao, J., Ogawa, H., Enomoto, N., Omata, M., and Nakao, A. (2007). Smad4 is Essential for Down-regulation of E-cadherin Induced by TGF-{beta} in Pancreatic Cancer Cell Line PANC-1. J Biochem (Tokyo) 141, 345-351. Takei, Y., Kondo, S., Harada, A., Inomata, S., Noda, T., and Hirokawa, N. (1997). Delayed development of nervous system in mice homozygous for disrupted microtubule-associated protein 1B (MAP1B) gene. J Cell Biol 137, 1615-1626. Tang, E. D., Wang, C. Y., Xiong, Y., and Guan, K. L. (2003). A role for NF-kappaB essential modifier/IkappaB kinase-gamma (NEMO/IKKgamma) ubiquitination in the activation of the IkappaB kinase complex by tumor necrosis factor-alpha. J Biol Chem 278, 37297-37305. Theroux, S., Pereira, M., Casten, K. S., Burwell, R. D., Yeung, K. C., Sedivy, J. M., and Klysik, J. (2007). Raf kinase inhibitory protein knockout mice: expression in the brain and olfaction deficit. Brain Res Bull 71, 559-567. Thiery, J. P. (2002). Epithelial-mesenchymal transitions in tumour progression. Nat Rev Cancer 2, 442-454. Thiery, J. P. (2003). Epithelial-mesenchymal transitions in development and pathologies. Curr Opin Cell Biol 15, 740-746. Thiery, J. P., and Sleeman, J. P. (2006). Complex networks orchestrate epithelial- mesenchymal transitions. Nat Rev Mol Cell Biol 7, 131-142. Tisdale, E. J., Bourne, J. R., Khosravi-Far, R., Der, C. J., and Balch, W. E. (1992). GTP- binding mutants of rab1 and rab2 are potent inhibitors of vesicular transport from the endoplasmic reticulum to the Golgi complex. J Cell Biol 119, 749-761. Tohdoh, N., Tojo, S., Agui, H., and Ojika, K. (1995). Sequence homology of rat and human HCNP precursor proteins, bovine phosphatidylethanolamine-binding protein and rat 23-kDa protein associated with the opioid-binding protein. Brain Res Mol Brain Res 30, 381-384. Tomlins, S. A., Mehra, R., Rhodes, D. R., Cao, X., Wang, L., Dhanasekaran, S. M., Kalyana-Sundaram, S., Wei, J. T., Rubin, M. A., Pienta, K. J., et al. (2007). Integrative molecular concept modeling of prostate cancer progression. Nat Genet 39, 41-51. Trakul, N., Menard, R. E., Schade, G. R., Qian, Z., and Rosner, M. R. (2005). Raf kinase inhibitory protein regulates Raf-1 but not B-Raf kinase activation. J Biol Chem. Tranque, P., Crossin, K. L., Cirelli, C., Edelman, G. M., and Mauro, V. P. (1996). Identification and characterization of a RING zinc finger gene (C-RZF) expressed in chicken embryo cells. Proc Natl Acad Sci U S A 93, 3105-3109. Tsui, S., Dai, T., Roettger, S., Schempp, W., Salido, E. C., and Yen, P. H. (2000). Identification of two novel proteins that interact with germ-cell-specific RNA- binding proteins DAZ and DAZL1. Genomics 65, 266-273. van 't Veer, L. J., Dai, H., van de Vijver, M. J., He, Y. D., Hart, A. A., Mao, M., Peterse, H. L., van der Kooy, K., Marton, M. J., Witteveen, A. T., et al. (2002). Gene 119 expression profiling predicts clinical outcome of breast cancer. Nature 415, 530- 536. van der Vlag, J., and Otte, A. P. (1999). Transcriptional repression mediated by the human polycomb-group protein EED involves histone deacetylation. Nat Genet 23, 474-478. Varambally, S., Dhanasekaran, S. M., Zhou, M., Barrette, T. R., Kumar-Sinha, C., Sanda, M. G., Ghosh, D., Pienta, K. J., Sewalt, R. G., Otte, A. P., et al. (2002). The polycomb group protein EZH2 is involved in progression of prostate cancer. Nature 419, 624-629. Varambally, S., Yu, J., Laxman, B., Rhodes, D. R., Mehra, R., Tomlins, S. A., Shah, R. B., Chandran, U., Monzon, F. A., Becich, M. J., et al. (2005). Integrative genomic and proteomic analysis of prostate cancer reveals signatures of metastatic progression. Cancer Cell 8, 393-406. Vartiainen, M. K., Sarkkinen, E. M., Matilainen, T., Salminen, M., and Lappalainen, P. (2003). Mammals have two twinfilin isoforms whose subcellular localizations and tissue distributions are differentially regulated. J Biol Chem 278, 34347-34355. Vega, M. I., Huerta-Yepaz, S., Garban, H., Jazirehi, A., Emmanouilides, C., and Bonavida, B. (2004). Rituximab inhibits p38 MAPK activity in 2F7 B NHL and decreases IL-10 transcription: pivotal role of p38 MAPK in drug resistance. Oncogene 23, 3530-3540. Walter, L., Dirks, B., Rothermel, E., Heyens, M., Szpirer, C., Levan, G., and Gunther, E. (1994). A novel, conserved gene of the rat that is developmentally regulated in the testis. Mamm Genome 5, 216-221. Walter, P., and Blobel, G. (1981). Translocation of proteins across the endoplasmic reticulum III. Signal recognition protein (SRP) causes signal sequence-dependent and site-specific arrest of chain elongation that is released by microsomal membranes. J Cell Biol 91, 557-561. Walter, P., and Johnson, A. E. (1994). Signal sequence recognition and protein targeting to the endoplasmic reticulum membrane. Annu Rev Cell Biol 10, 87-119. Wang, X., Li, N., Liu, B., Sun, H., Chen, T., Li, H., Qiu, J., Zhang, L., Wan, T., and Cao, X. (2004). A novel human phosphatidylethanolamine-binding protein resists tumor necrosis factor alpha-induced apoptosis by inhibiting mitogen-activated protein kinase pathway activation and phosphatidylethanolamine externalization. J Biol Chem 279, 45855-45864. Weigel, D., and Nilsson, O. (1995). A developmental switch sufficient for flower initiation in diverse plants. Nature 377, 495-500. Welsh, J. B., Sapinoso, L. M., Su, A. I., Kern, S. G., Wang-Rodriguez, J., Moskaluk, C. A., Frierson, H. F., Jr., and Hampton, G. M. (2001). Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer. Cancer Res 61, 5974-5978. Wild, K., Halic, M., Sinning, I., and Beckmann, R. (2004). SRP meets the ribosome. Nat Struct Mol Biol 11, 1049-1053.
120 Willy, P. J., Kobayashi, R., and Kadonaga, J. T. (2000). A basal transcription factor that activates or represses transcription. Science 290, 982-985. Wind, M., and Reines, D. (2000). Transcription elongation factor SII. Bioessays 22, 327- 336. Xu, Q., and Reed, J. C. (1998). Bax inhibitor-1, a mammalian apoptosis suppressor identified by functional screening in yeast. Mol Cell 1, 337-346. Yamagata, T., Maki, K., and Mitani, K. (2005). Runx1/AML1 in normal and abnormal hematopoiesis. Int J Hematol 82, 1-8. Yamazaki, H., Tateyama, H., Asai, K., Fukai, I., Fujii, Y., Tada, T., and Eimoto, T. (2005). Glia maturation factor-beta is produced by thymoma and may promote intratumoral T-cell differentiation. Histopathology 47, 292-302. Yamazaki, T., Nakano, H., Hayakari, M., Tanaka, M., Mayama, J., and Tsuchida, S. (2004). Differentiation induction of human keratinocytes by phosphatidylethanolamine-binding protein. J Biol Chem. Yang, J., Mani, S. A., Donaher, J. L., Ramaswamy, S., Itzykson, R. A., Come, C., Savagner, P., Gitelman, I., Richardson, A., and Weinberg, R. A. (2004). Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 117, 927-939. Yang, Q., Graham, T. E., Mody, N., Preitner, F., Peroni, O. D., Zabolotny, J. M., Kotani, K., Quadro, L., and Kahn, B. B. (2005a). Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature 436, 356- 362. Yang, Z., Rayala, S., Nguyen, D., Vadlamudi, R. K., Chen, S., and Kumar, R. (2005b). Pak1 phosphorylation of snail, a master regulator of epithelial-to-mesenchyme transition, modulates snail's subcellular localization and functions. Cancer Res 65, 3179-3184. Yen, P. H. (2004). Putative biological functions of the DAZ family. Int J Androl 27, 125- 129. Yeung, K., Janosch, P., McFerran, B., Rose, D. W., Mischak, H., Sedivy, J. M., and Kolch, W. (2000). Mechanism of suppression of the Raf/MEK/extracellular signal-regulated kinase pathway by the raf kinase inhibitor protein. Mol Cell Biol 20, 3079-3085. Yeung, K., Seitz, T., Li, S., Janosch, P., McFerran, B., Kaiser, C., Fee, F., Katsanakis, K. D., Rose, D. W., Mischak, H., et al. (1999). Suppression of Raf-1 kinase activity and MAP kinase signalling by RKIP. Nature 401, 173-177. Yeung, K. C., Rose, D. W., Dhillon, A. S., Yaros, D., Gustafsson, M., Chatterjee, D., McFerran, B., Wyche, J., Kolch, W., and Sedivy, J. M. (2001). Raf kinase inhibitor protein interacts with NF-kappaB-inducing kinase and TAK1 and inhibits NF-kappaB activation. Mol Cell Biol 21, 7207-7217. Yokoyama, K., Kamata, N., Hayashi, E., Hoteiya, T., Ueda, N., Fujimoto, R., and Nagayama, M. (2001). Reverse correlation of E-cadherin and snail expression in oral squamous cell carcinoma cells in vitro. Oral Oncol 37, 65-71.
121 Yook, J. I., Li, X. Y., Ota, I., Fearon, E. R., and Weiss, S. J. (2005). Wnt-dependent regulation of the E-cadherin repressor snail. J Biol Chem 280, 11740-11748. Yook, J. I., Li, X. Y., Ota, I., Hu, C., Kim, H. S., Kim, N. H., Cha, S. Y., Ryu, J. K., Choi, Y. J., Kim, J., et al. (2006). A Wnt-Axin2-GSK3beta cascade regulates Snail1 activity in breast cancer cells. Nat Cell Biol 8, 1398-1406. Yu, J. X., Chao, L., and Chao, J. (1994). Prostasin is a novel human serine proteinase from seminal fluid. Purification, tissue distribution, and localization in prostate gland. J Biol Chem 269, 18843-18848. Zhang, L., Fu, Z., Binkley, C., Giordano, T., Burant, C. F., Logsdon, C. D., and Simeone, D. M. (2004). Raf kinase inhibitory protein inhibits beta-cell proliferation. Surgery 136, 708-715. Zhang, S., Qiu, W., Wu, H., Zhang, G., Huang, M., Xiao, C., Yang, J., Kamp, C., Huang, X., Huellen, K., et al. (2001). The shorter zinc finger protein ZNF230 gene message is transcribed in fertile male testes and may be related to human spermatogenesis. Biochem J 359, 721-727.
122 ABSTRACT
Raf Kinase Inhibitor Protein (RKIP), a negative regulator of the Raf/MEK/ERK
and NF-κB signaling pathways, is downregulated in several cancers, including prostate
cancer. RKIP expression is further diminished in metastatic prostate cancer, qualifying it
as a metastasis suppressor gene. The transcriptional regulation of RKIP in prostate
cancer is unknown; however, promoter studies suggest that methylation is not the means of transcriptional inhibition.
The decrease in RKIP at the time of metastatic invasion suggests that factors
involved in the epithelial-to-mesenchymal transition (EMT) may be involved in its
regulation. The Snail superfamily of transcription factors are involved in the EMT
process in normal development and in cancer. Protein levels of Snail were low in the low-mestastatic prostate cancer cell line LNCaP and increased in the PC3 and Du145 cell lines derived from metastases of prostate adenocarcinoma. The endogenous protein levels of E-cadherin, a known target of Snail repression, and RKIP were higher in LNCaP cells and diminished in PC3 and Du145 cells. Overexpressing Snail in LNCaP cells resulted in diminished RKIP levels, and knocking down Snail in PC3 and Du145 cells increased RKIP expression. RKIP promoter activity was repressed by both Snail and
Slug. Chromatin immunoprecipitation assays showed that Snail associates with E-box elements near the RKIP transcriptional start site.
We identify a physiological relationship between RKIP and Snail in prostate
cancer through analysis of publicly available microarray data. There is a significant 123 negative correlation between RKIP and Snail, Snail and E-cadherin, and also RKIP and
EZH2, but not RKIP and Slug. Finally, hierarchical cluster analysis was used to attempt to determine the expression profile to which RKIP and Snail belong and to predict genes that may be involved in the regulation of RKIP.
124