Chapter 2shodhganga.inflibnet.ac.in/bitstream/10603/11669/11... · 2.3 Chromosomal deletion at...

Chapter 2

Identification and validation of novel CNAs in search of

targeted therapy

Identification and validation of novel CNAs in search of targeted therapy

2.1 Characterization of CNAs in cancer

Identification of and functional characterization of chromosomal gains and losses has been

the key to identify novel oncogenes and tumor suppressor genes (TSGs). For instance, a

common deletion observed in many cancers located at 1p36. Characterization of this deletion

led to the identification of a novel TSG, Chromodomain Helicase DNA binding protein 5

(CHD5) and subsequently shown to be involved in proliferation, apoptosis and senescence

via p19/p53 pathways (Bagchi, Papazoglu et al. 2007). In another study of a high density

aCGH analysis of 86 prostate cancer samples revealed a frequent deletion at 5q21.1. Precise

mapping of this deletion identified Chromodomain Helicase DNA binding protein 1 (CHD1)

gene. Functional characterization of this gene in prostate cancer cell lines showed its ability

to inhibit invasiveness and clonogenicity, but not proliferation of cells (Huang, Gulzar et al.

2012). Similarly, a recurrent amplification at 22q11.21 was identified in lung cancer and

functional characterization of the amplification in lung cancer cell lines led to the discovery

of a novel oncogene namely CRKL (Kim, Kwei et al. 2010). Another lung cancer lineage

specific oncogene TITF1 was discovered following validation of a focal amplification at

14q13.3, detected in a high resolution aCGH analysis of 52 non-small cell lung cancer cell

lines and 76 tumor tissues (Kwei, Kim et al. 2008). Another analysis of squamous cell

carcinoma (SCC) of esophagus and lung through high resolution SNP array led to the

identification of an amplification at 3q26.33. Characterization of this amplification led to

identification of SOX2 oncogene. Further functional characterization of SOX2 in esophageal

and lung cancer cell lines revealed growth promoting effect and increased anchorage

independent growth (Bass, Watanabe et al. 2009).


2.2 Characterization of CNAs in PaCa

Though many novel amplifications and deletions were identified in PaCa in last decade, very

few recurrent amplicons and deletions have been characterized for identification of novel

oncogenes and TSGs. For example, an analysis of transcriptome (23,219 transcripts, 20,661

protein-coding genes) from 24 pancreatic tumors through next generation sequencing

revealed 63 altered genes fundamentally belonging to 12 core signaling pathways and in

addition to this, the study also identified several amplification and deletions in many tumor

samples (Jones, Zhang et al. 2008). Further characterization of the deletions and

amplifications is needed to identify novel target(s). In a recent study, chromosomal arm 20q

was shown to be deleted in pancreatic and many other cancers; validation of this deletion in

PaCa cell lines revealed TPX2 as a novel candidate oncogene during pancreatic

carcinogenesis (Warner, Stephens et al. 2009). Another, frequent amplification has been

observed at 7q21-q22 in PDAC and other cancers including malignancies of esophagus,

stomach and melanomas. Functional validation of this amplification discovered two

frequently up regulated genes ARPC1A and ARPC1B, and subsequent validation on a panel of

PaCa cell lines showed an oncogenic effect and the genes were proposed as candidate

oncogenes for therapeutic target in PaCa (Laurila, Savinainen et al. 2009). A high resolution

aCGH and gene expression profiling of 22 PaCa cell lines (Bashyam, Bair et al. 2005) and 37

pancreatic tumor xenografts (Kwei, Bashyam et al. 2008) revealed amplifications at 7q22.1,

11q13, 18q11.2 etc and recurrent homozygous deletions at 6q25, 8p22-p23, 16q23, 18q21-

q23 etc. A novel amplification at 18q was observed in more than 19% of pancreatobilliary

xenografts; and mapping of this amplification revealed a novel oncogene GATA6, a

transcriptional factor known to be involved in development and differentiation (Kwei,

Bashyam et al. 2008). Further characterization of GATA6 in PaCa cell lines showed its ability

to increase proliferation rate, anti-apoptotic effect and colony formation (Kwei, Bashyam et


al. 2008). Another important genomic amplification at 7q22 was identified in PaCa cell lines

(Bashyam, Bair et al. 2005). Further high resolution aCGH analysis of pancreatic tumor

xenografts and cell lines led to the discovery of a novel oncogene SMURF1 (SMAD specific

E3 ubiquitin protein ligase 1) at 7q22.1 (Kwei, Bashyam et al. 2008; Suzuki, Shibata et al.

2008). SMURF1 is an important component of TGFβ signaling pathway. Functional analysis

revealed SMURF1 to have pro-motility effect and elevated potential for colony formation in

PaCa cell lines (Kwei, Shain et al. 2011).

Recurrent homozygous deletions may harbor important TSGs. Given the comparatively lower

focus of researchers, studies on TSGs have become important, especially since oncogenes

have not yielded expected results for targeted therapy. A recent broad mathematical modular

analysis of more than 3,000 genes from more than 90,000 tumors (COSMIC database:

http://www.sanger.ac.uk/genetics/CGP/cosmic/) revealed a significantly higher proportion of

driver genes are TSGs (Bozic, Antal et al. 2010). Moreover, the probability of mutational

inactivation of TSGs is far more than an oncogene (Bozic, Antal et al. 2010). Therefore, in

the current work I focused on two recurrent homozygous deletions for characterization of

possible novel TSGs.

2.3 Chromosomal deletion at 6q25.3 in PaCa

6q25.3 is a susceptible locus for genomic loss in PaCa. Several studies revealed abnormalities

in this locus. For example, an aCGH analysis of 22 PaCa cell lines detected a homozygous

deletion in MiaPaCa2 and suspected single copy loss at this locus in several other PaCa cell

lines (Bashyam, Bair et al. 2005). Deletion of this locus was also observed in another aCGH

study of eight PaCa cell lines (Nissenblatt 2011). Primary tumor tissues have also been

shown to harbor genomic loss at this locus. For instance, genomic loss of 6q25.2-q25.3 was

detected in an aCGH analysis of 23 PDAC tissue samples (Harada, Baril et al. 2007). Another

http://www.sanger.ac.uk/genetics/CGP/cosmic/


recent aCGH analysis of a large number of PDAC tissue samples (72) identified a focal

genomic loss at 6q25.3 (Shain, Giacomini et al. 2012). Thus, analysis of the locus for a

plausible candidate TSG revealed the AT-Rich interactive domain containing protein 1B

(ARID1B) gene, encoding DNA binding component of SWI/SNF chromatin remodeling

complex.

2.3.1 SWI/SNF chromatin remodeling complex

Chromatin is the basic functional and heritable material of eukaryotic cells composed of

nucleosomes. Nucleosome is the DNA-histone complex compacted more than 5000 fold in

eukaryotic cells (Dawson and Kouzarides 2012). This compact chromatin structure is broadly

classified into a compact hetero-chromatin region and a relatively less compact chromatin

called euchromatin. Despite the highly compact structure, both the hetero-chromatin and

euchromatin are dynamic entities involved in various critical cellular processes such as

facilitating binding of transcription factor during transcription, DNA recombination and

repair. These cellular processes are highly regulated and achieved through DNA methylation,

histone modification and ATP dependent chromatin remodeling. There are four different

classes of chromatin remodeling complexes including SWI/SNF (Switch/Sucrose non

fermenter) complex, imitation SWI complex, NuRD (Nucleosome remodeling and

deacetylation) complex and CHRAC (Chromatin accessibility) complex (Wu, Lessard et al.

2009). SWI/SNF complex is emerging as central remodeling complex for regulated gene

expression in normal as well as in diseased condition (Roberts and Orkin 2004). The process

of chromatin remodeling by SWI/SNF involves the sliding of nucleosome on DNA strand

using energy of ATP generated through its ATPase activity (Saha, Wittmeyer et al. 2006).

SWI/SNF complex was first identified in Saccharomyces cerevisiae during genetic screening

of altered gene expression. SNF mutants were identified in the yeast having SUC2 gene


expression defects causing growth suppression on sucrose and hence named as sucrose non-

fermenter (SNF). While SWI genes were identified in mating type switching mutants of yeast

having abnormal HO gene expression and thus this class of genes was named as switch

(SWI). SWI2/SNF2 was identified as transcription activator of these genes using energy from

ATP (Neigeborn and Carlson 1984; Stern, Jensen et al. 1984; Pazin and Kadonaga 1997).

The components of SWI/SNF complex are evolutionarily conserved among eukaryotes and

homologues of it were identified in Drosophila as well as in higher eukaryotes. In Drosophila,

Brahma (BRM/SMARCA2) (yeast SWI2/SNF2 homologue) was identified in the screening of

suppressor of Polycomb mutants which causes body segment defects through regulation of

chromatin structure during transcription activation of homeotic genes (Tamkun, Deuring et

al. 1992; Hargreaves and Crabtree 2011).

In vertebrate genome, SWI/SNF encoding genes are one of the most diverse families of

genes. More than 30 genes were known to encode SWI2/SNF2 component of SWI/SNF

complex (Wu, Lessard et al. 2009). Biochemical purification and reconstitution studies

showed SWI/SNF as a multi-subunit protein complex and revealed to have 8-15 subunits in

mammalian cells (Wang, Cote et al. 1996; Phelan, Sif et al. 1999; Weissman and Knudsen

2009). There are three major components of SWI/SNF complex in human cells: 1) enzymatic

component, 2) core component, 3) accessory component (Fig 2.1).


Figure 2.1: Diagram of the components of SWI/SNF: SWI/SNF complex harbor three

distinct classes of proteins. Enzyme component having two mutually exclusive members

namely BRG1 or BRM. List of known core and accessory components are listed. ARID1A or

ARID1B are mutually exclusive as depicted

.

1) Enzymatic component: Enzymatic components (harboring ATPase activity) of SWI/SNF

include either BRG1 (Brahma Related Gene1) or its homologue BRM. BRG1 was first

identified in humans as the yeast SWI2/SNF2 or Drosophila BRM homologue and shown to

be transcriptionally activate several genes (Khavari, Peterson et al. 1993). Biochemical

purification from several human cell lines revealed BRG1 and BRM as mutually exclusive

subunit of SWI/SNF (Wang, Cote et al. 1996). BRG1 has been shown to be transcriptional

activator as well as repressor in various cellular contexts by targeting set of genes required to

be expressed or silenced (Trotter and Archer 2008).

2. Core component: Core component of SWI/SNF contains several subunits named BRG1 or

BRM associated factors (BAF)s. In this component, SNF5 (also known as INI1, SMARCB1

and BAF47), BAF155 and BAF170 are highly conserved subunits. Other relatively less

conserved subunits include BAF45, BAF53, BAF57, BAF60 and actin (Wang, Cote et al.


1996; Phelan, Sif et al. 1999; Wilson and Roberts 2011). Core components provide stability

to the SWI/SNF complex by maintaining stoichiometric level of its subunits. For example

interaction of BAF57 with BAF155 or BAF170 was shown prevent prteosomal degradation of

BAF57 and also maintains its stoichiometric level in various human cell lines (Chen and

Archer 2005). SNF5 core subunit was initially identified as transcription activator of SUC2

gene in yeast (Hirschhorn, Brown et al. 1992) and later identified in humans as HIV integrase

binding factor (required for viral genome integration in human DNA) in a yeast two hybrid

screening and named as integrase inter-actor1 (INI1) (Kalpana, Marmon et al. 1994). SNF5

was shown to be involved in many cellular processes. For example, cell cycle arrest upon

SNF5 inactivation was observed in murine mouse fibroblast (Isakoff, Sansam et al. 2005;

Wilson and Roberts 2011).

3. Accessory component: Another major component of SWI/SNF is accessory component

which was first identified as RSC (yeast SWI/SNF variant) homologue in humans (Cairns,

Lorch et al. 1996; Xue, Canman et al. 2000). This component was shown to have a unique

evolutionary conserved subunit BAF180, which contains poly-bromo-domain and hence

named as PBAF (Xue, Canman et al. 2000). Further, studies revealed that accessory

component provides specificity to SWI/SNF, targeting specific set of genes required in a

specific cellular context. For example, BAF180 interacts with BAF200 or ARID2 (also known

as PBAF complex) for transcription of ligand activated receptor (in response of retinoic acid)

and interferon responsive genes (Lemon, Inouye et al. 2001; Yan, Cui et al. 2005). A distinct

BAF complex containing BAF250 subunit was identified targeting specifically

Glucocorticoid receptor-dependent gene transcription activation (Nie, Xue et al. 2000).

BAF250 is the largest conserved subunit (SWI1 homologue of yeast) of accessory

component, first identified as BAF subunit in various human cell lines (Wang, Cote et al.

1996). There are two alternative isoforms of BAF250 identified in human (hOsa1 and hOsa2:


drosophila homologue Osa) and shown to have a characteristic DNA binding domain (AT-

interactive domain or ARID) and hence named as ARID genes (Dallas, Pacchione et al. 2000;

Inoue, Furukawa et al. 2002). ARID1A was first identified as 270kda protein (p270) in Hela

cells and shown to have ARID domain which binds DNA in a non-sequence specific manner

(Dallas, Cheney et al. 1998; Dallas, Pacchione et al. 2000). Further, Osa1or ARID1A was

cloned from human brain and was shown to have ARID1A domain (Hurlstone, Olave et al.

2002). However, BAF250 isoform b (BAF250b or ARID1B) was first identified in many

human tissues and demonstrated to be alternative BAF250 component of SWI/SNF (Nie, Yan

et al. 2003). Also, the two isoforms (named ARID1A and ARID1B) are mutually exclusive

components of SWI/SNF similar to BRG1 and BRM as demonstrated in several human cell

lines (Wang, Nagl et al. 2004). Other subunits of accessory component such as BAF60,

BAF45 and BAF53 are relatively less conserved and their functional significance is yet to be

determined (Weissman and Knudsen 2009).

2.3.2 Role of SWI/SNF in cellular function

The diversity of SWI/SNF components and possibility of its immense combinatorial

assembly during various cellular contexts implies SWI/SNF to have key role in selective

target gene expression during critical cellular functions. Basic function of SWI/SNF is

transcription regulation and targeting genes required for specific cellular functions. Set of

genes targeted mainly belong to major known cellular functions (Fig 2.2).


Figure 2.2: Diagram represents cellular functions targeted by SWI/SNF complex.

1) Role of SWI/SNF in embryonic development and stem cell differentiation: SWI/SNF

complex targets the specific genes required for embryonic developmental process and stem

cell differentiation (Wu 2012). For example, deletion of BRG1, BAF155 or SRG3 and SNF5

was demonstrated to cause embryonic lethality in mice (Bultman, Gebuhr et al. 2000;

Klochendler-Yeivin, Fiette et al. 2000; Kim, Huh et al. 2001; Wu 2012). Zygote genome

activation is prerequisite for activation of transcription program just after fertilization.

Molecular determinant for this genome level activation was unknown; however, BRG1 has

been shown to be responsible for zygote genome activation during embryonic cell division in

mice (Bultman, Gebuhr et al. 2000). Stem cell maintenance (pluripotency of stem cells or PS)

and its differentiation into different linage requires specific transcriptional program. This

transcription program is regulated at various levels. For example, transcription network

mediated by a small set transcription factors such as Oct4/POU5F1, SOX2 and NANOG

(Jaenisch and Young 2008; Young 2011). The transcription program of embryonic stem cells

is also regulated at chromatin structure level mediated through various chromatin remodeling

complexes including SWI/SNF (Young 2011). SWI/SNF composition identified from mouse


embryonic stem cells revealed a distinct composition as (named as esBAF) compared to

fibroblast and shown to have BRG1 instead of BRM, BAF60a instead of BAF60c, presence of

BAF155 and absence of BAF170 (Ho, Jothi et al. 2009). BRG1 of esBAF complex was shown

to interact with the genes which are required for maintenance of transcription circuitry of

embryonic stem cell and also shown esBAF co-localization with classical stem cell

transcription factors Oct4, SOX2 and NANOG (Ho, Ronan et al. 2009). Furthermore, both

ARID1A (BAF250a) and ARID1B (BAF250b) knockout mice embryonic stem cells

demonstrated a reduced proliferation (stem cell renewal) and altered differentiation

phenotype (Gao, Tate et al. 2008; Yan, Wang et al. 2008). Induction of pluripotency in stem

cells requires reprogramming of transcription by a small number of transcription factors like

Oct4, SOX2, c-Myc and KLF4; however, the efficiency of reprograming is very less and a

recent study suggested that induction of pluripotency can be enhanced by overexpression of

BRG1 and BAF155 (Singhal, Graumann et al. 2010).

2) SWI/SNF and tissue differentiation:

During embryonic developmental stages, development and differentiation of different tissue

from its progenitor cells require a specific set of genes to be expressed. This specific

transcription program is regulated through tissue specific transcription factors. Regulation at

the level of chromatin is a prerequisite for these transcription factors to interact with gene

specific transcription regulatory elements. SWI/SNF mediated chromatin remodeling has

been shown to be important for this process (Wu 2012).

Role of SWI/SNFF in neuronal development and differentiation of heart and muscle is

relatively well studied. In case of neuronal development, mutant BRG1 and BAF155/SRG3

deficient mice were shown to be defective in brain development in embryonic developmental

stage (Bultman, Gebuhr et al. 2000; Kim, Huh et al. 2001). Further, proteomic studies from


neural progenitor/stem cells and differentiated cells revealed a characteristic difference in

SWI/SNF subunit composition. For instance, BAF45a and BAF53a are present in progenitor

cells, while BAF45b and BAF53b are present in differentiated post mitotic neurons as

demonstrated in a proteomic studies and hence named as neural progenitor BAF or npBAF

and neuron BAF or nBAF (Lessard, Wu et al. 2007; Wu, Lessard et al. 2007; Wu, Lessard et

al. 2009). In the development of heart muscles, SWI/SNF subunit composition plays a

transcription regulator role in differentiation of cardiac progenitor cells into myocardium and

endocardium. For example, BRG1 mediated repression of ADAMTS1 (matrix

metalloproteinase required for formation of cardiac jelly for trabeculae formation and

development of heart ventricles) is required for myocardial development stage (Stankunas,

Hang et al. 2008). In cardiomyocyte development, expression of alpha and beta myosin heavy

chain (α or β MHC) gene (MHCB expression is required in embryonic cardiomycyte stage

and MHCA express in adult myocyte) is regulated by BRG1 as demonstrated in a recent study

(Hang, Yang et al. 2010). In addition BAF180, BAF60c and BAF45c have been shown to be

involved in transcription of genes required at different developmental stage of heart

(Takeuchi, Lou et al. 2011; Wu 2012). During skeletal muscles development, MYOG

(Myogenin) expression is required and shown to be regulated through p38 pathway and

BAF60 was shown to be the target of p38 kinase in myoblasts (Simone, Forcales et al. 2004).

3) SWI/SNF in DNA repair:

Broadly, there can be four types of DNA damages due to genotoxic stress namely, 1) base

damage, 2) helix distortion due to bulky adducts 3) strand breaks (including single strand and

double strand breaks) and 4) mismatch. These DNA damages are restored by various

mechanisms including direct reversal of base damage, base excision repair (BER) pathway,

nucleotide excision repair (NER) pathway, double strand break repair through non


homologous end joining (NHEJ) and homologous recombination (Osley, Tsukuda et al.

2007). Accessibility of DNA repair machinery at the damage site is controlled at the level of

chromatin structure and regulated by various chromatin remodeling complexes including

SWI/SNF (Wuebbles and Jones 2004). For example, SWI/SNF complex has been shown to

improve the BER efficiency in an in vitro study of DNA repair (Menoni, Gasparutto et al.

2007). Various DNA repair proteins were shown to be associated with the components of

SWI/SNF. For example, a study in yeast identified recruitment of SWI/SNF components (e.g

SNF5) at double strand breaks (Chai, Huang et al. 2005). Moreover, in an elegant study from

human cell lines (HeLa), BRCA1 (DNA double strand break repair protein through HR) was

shown to interact directly with BRG1 (Bochar, Wang et al. 2000). Further, SNF5 or BAF47

deficient mouse cells were demonstrated to be more sensitive in response to genotoxic stress

and showed an aberrant mitotic phenotype specific for DNA damage response (Klochendler-

Yeivin, Picarsky et al. 2006). NER is important in ultraviolet (UV) induced DNA damage.

NER machinery recruits UV induced DNA damage response factor XPC at damage site more

efficiently in presence of SNF5; which ultimately activates ATM mediated checkpoint (Ray,

Mir et al. 2009).

4) Immune system and SWI/SNF:

Immune system has many types of specialized cells, which is differentiated from its

progenitor cells. For example, Thymocytes differentiate into mature T-cells through

successive stages of development. Mature T-cells are characterized by the presence of TCR

(CD4+ or CD8

+) along with other co-receptors. Transcription of specific set of genes at every

stage of development is regulated by chromatin remodeling complexes including SWI/SNF

(Wurster and Pazin 2012). For example, mutually exclusive expression of CD4 and CD8 was

shown to be regulated by BAF57 and BRG1 through recruitment of RUNX1 transcription


repressor at repressor site of CD4 gene (Chi, Wan et al. 2002; Wan, Zhang et al. 2009). Viral

infection induces expression of interferon alpha (IFNA) and IFNB. Expression of IFNB was

shown to be regulated through SWI/SNF mediated transcription activation (Agalioti,

Lomvardas et al. 2000). Further, a study in HeLa cells showed BRG1 mediated enhancement

of expression level of IFNA responsive genes through interaction with STAT1 transcription

activation (Huang, Qian et al. 2002). Further, BAF47 inhibition through RNAi suppressed the

expression of IFN responsive genes and also reduced the antiviral activity of cells (Cui,

Tailor et al. 2004). This raises the possibility that cells defective in SWI/SNF complex may

not respond appropriately in viral infection.

5) Functional significance of SWI/SNF in cell cycle and DNA replication

Components of SWI/SNF have been implicated to regulate cell cycle through interaction with

regulators of cell cycle and DNA replication. G1/S cell cycle checkpoint is regulated by the

transcriptional regulation of E2F target genes (through inactivation of retinoblastoma (RB))

which is required for DNA synthesis, repair and mitosis. This pathway of cell cycle

regulation is found to be regulated by defective SWI/SNF components and restoration of

these components was found to be involved in normal cell cycle arrest through interaction

with RB. For example, BRG1/BRM was shown to interact with RB and found to cooperate for

induction of cell cycle arrest along with other cell cycle regulatory proteins (Dunaief, Strober

et al. 1994). Also, RB was shown to be associated with SWI/SNF and HDAC for repression

of Cyclin A and E genes in G1 phase and RB-SWI/SNF association was shown to be

disrupted upon phosphorylation by Cyclin D/CDK and ultimately helps in up-regulation of

Cyclin E and progression in S phase (Zhang, Gavin et al. 2000). Further, cell cycle arrest

through RB was shown to be regulated by BRG1 and facilitates the increase in the CDK

inhibitor p21 levels which ultimately keeps RB hypo-phosphorylated and cell cycle is arrested


in G1 (Kang, Cui et al. 2004). Moreover, SNF5 deficiency leads to aberrant up-regulation of

E2F target genes including p53 (Isakoff, Sansam et al. 2005). p16/CDKN2A and P21 are

well-studied cell cycle arrest genes up-regulated during cell cycle arrest through cyclin-

dependent kinase. SWI/SNF (BRG1/BRM) complex was demonstrated to bind the promoters

of p16/CDKN2A and P21 during differentiation of mouse erythroleukemia cells (Kadam and

Emerson 2003). p16/ CDKN2A has also been demonstrated to be down regulated in SNF5

knock-out mice models (Oruetxebarria, Venturini et al. 2004; Isakoff, Sansam et al. 2005). In

another study, BRM subunit was found to act cooperatively for repression of E2F target genes

(Trouche, Le Chalony et al. 1997; Bracken, Ciro et al. 2004). p53 mediated cell cycle arrest

and apoptosis in case of DNA damage is a well-established phenomenon. SWI/SNF plays an

important role in regulation of p53 mediated transcription of its responsive genes. For

example, interaction of p53 with SWI/SNF components (BAF53, SNF5 and BRG1) has been

observed and a concomitant activation of p53 response genes such as p21 was demonstrated

in human cell lines (Lee, Kim et al. 2002; Wang, Gu et al. 2007). In case of double strand

DNA damage, BRCA1 activated repair process was shown to be mediated by SWI/SNF

complex and BRCA1-SWI/SNF complex has been demonstrated to activate p53 dependent

gene transcription (Bochar, Wang et al. 2000). ARID1A has been implicated in cell cycle

arrest and further its mutually exclusive isoform ARID1B was shown to have opposing role in

cell cycle progression (Nagl, Patsialou et al. 2005).

2.3.3 Function of SWI/SNF as tumor suppressor complex in cancer

SWI/SNF complex has been established as key epigenetic regulator of transcription and also

has been shown to be involved in gene transcription required for key cellular functions as

described in section 2.3.2. SWI/SNF complex has been emerging as a novel tumor suppressor

complex during past few years and all the three components (enzyme, core and accessory) of


this complex have been shown to be associated with tumor suppressor function in a variety of

cancers (Table2.1).

1) Core subunits of SWI/SNF complex in cancer initiation and progression

Core subunits of SWI/SNF are the major targets during tumor initiation and progression. For

example, SNF5/BAF47/INI1 has emerged as a bona fide tumor suppressor in several cancers

including malignant rhabdoid tumors (RTs) in children (Hollmann and Hornick 2011). RTs

are very aggressive cancer known so far and found to have very few genetic aberrations. For

example, SNF5 was found to be inactivated through mutation or homozygous deletion in a

cohort of RT cell lines and primary tissues (Versteege, Sevenet et al. 1998; Jackson, Sievert

et al. 2009). Further, SNF5 driven tumor were demonstrated to be aggressive tumor in mice

models (11 weeks for tumor onset) and found to be phenotypically similar to RT and

lymphomas (Roberts, Galusha et al. 2000; Roberts, Leroux et al. 2002). Tumorigenesis

through SNF5 loss may be driven by activation of genes involved in cell proliferation. In a

study of SNF5 knockout in RT cell line, cells were accumulated in G0/G1 and also shown to

repress p16/CDKN2A and several Cyclin genes (Oruetxebarria, Venturini et al. 2004). Cell

cycle progression was also shown to be stimulated in SNF5 deficient murine model by up-

regulation of p53 and other E2F target genes (Isakoff, Sansam et al. 2005). In a recent study,

another subunit of core component BAF155 was suspected to possess possible tumor

suppressor activity by reduction of cell proliferation rate and induction of replicative

senescence in two cell line from ovarian cancer and colorectal cancer (DelBove, Rosson et al.

2011).

2) Enzyme subunit of SWI/SNF complex in cancer

ATPase subunits of SWI/SNF are also major targets during carcinogenesis as evidenced by

studies in several cancers. BRG1 inactivation through mutation and down-regulation of


expression was documented in a) lung cancer cell lines and primary tissues (Reisman,

Sciarrotta et al. 2003), b) a panel breast cancer, c) prostate cancer d) PaCa cell lines

(Decristofaro, Betz et al. 2001), e) primary tumor tissues from RT (Schneppenheim,

Fruhwald et al. 2010) and tissue samples of medulloblastoma patients (Parsons, Li et al.

2011). However, the frequency of mutation was found to be higher in lung cancer,

specifically in non-small cell lung carcinoma cell lines (Medina, Romero et al. 2008). Also,

coexistence of mutation in KRAS, NRAS, LKB1, p16/CDKN2A and TP53 was identified in

addition to BRG1, which is suggesting a synergistic role of BRG1 in tumor formation

(Medina, Romero et al. 2008). Consistent with the data, BRG1-/-

murine model was

embryonic lethal; however, heterozygous mice showed significant growth of breast tumor in

at least 10% of animals (Bultman, Gebuhr et al. 2000). Interestingly, conditional BRG1

knockout mice revealed a high penetrance of lung tumor development suggesting a role of

BRG1 during lung tumor progression but perhaps not at the initiation stage (Glaros,

Cirrincione et al. 2008). Despite the molecular and functional similarity between BRM

(mutually exclusive subunit of BRG1) and BRG1, BRM has shown different pattern of

influence in tumorigenesis as compared to BRG1. For example, BRM-/-

murine model was

viable and fertile while BRG1-/-

mice was shown to be embryonic lethal (Bultman, Gebuhr et

al. 2000; Bultman, Herschkowitz et al. 2008). Another study in osteoblast cell line revealed

an antagonistic role of BRG1 and BRM for differentiation of osteoblast and shown to both

repress and activate osteocalcin (BGLAP) gene transcription in a temporal fashion (Flowers,

Nagl et al. 2009). Many studies also revealed role of BRM in various cancers. For example,

reduced BRM expression was identified in several cancers including a panel of lung cancer

cell lines and primary tumor tissues (Reisman, Sciarrotta et al. 2003) and primary prostate

cancer samples (Sun, Tawfik et al. 2007). Comparatively more reduction was identified in

undifferentiated primary tumor tissues of gastric carcinoma (Yamamichi, Inada et al. 2007).


Importantly, common mode of BRG1 inactivation was shown to be through mutation, but

BRM was shown to be predominantly inactivated through DNA hyper-methylation in a panel

of lung cancer cell lines (Glaros, Cirrincione et al. 2007).

3) Functional implication of accessory subunits of SWI/SNF complex in cancer

Accessory subunits of SWI/SNF are essential functional component and implicated to be de-

regulated in carcinogenesis. ARID1A (BAF250a) is a DNA binding component of SWI/SNF

complex and has emerged as a lineage specific TSGs in several cancer (Wu and Roberts

2013). Cancer genome sequencing through next generation sequencing identified ARID1A

mutation or deletion in a several cancers. For example, frequent ARID1A mutation was

identified in ovarian clear cell carcinoma (an aggressive subtype of ovarian carcinoma) and in

endometrioid carcinoma (Jones, Wang et al. 2010; Wiegand, Shah et al. 2010); relatively less

frequent ARID1A mutation was observed in a gastric cancer subtype (Wang, Kan et al. 2011;

Zang, Cutcutache et al. 2012), in hepatocellular carcinoma samples (Guichard, Amaddeo et

al. 2012), and in bladder cancer primary tumor samples (Gui, Guo et al. 2011) in exome

sequencing. In addition to this, ARID1A deletion and inactivation was identified through

aCGH in breast cancer cell lines (Mamo, Cavallone et al. 2012), and in PaCa primary tumor

samples (Shain, Giacomini et al. 2012). Another DNA binding subunit of SWI/SNF is

ARID1B that is mutually exclusive to ARID1A as described in section 2.3.1. ARID1B along

with ARID1A was found to be inactivated by mutation in hepatocellular carcinoma tumor

samples in exome sequencing study (Fujimoto, Totoki et al. 2012). Further, in a recent study

of whole exome sequencing of breast tumor tissues, ARID1B mutation was identified as

driver mutation (Stephens, Tarpey et al. 2012). Further, a recent exome sequencing of

childhood cancer neuroblastoma patient samples and neuroblastoma cell lines revealed

several genomic alterations including ARID1B and ARID1A mutations (Sausen, Leary et al.

2013). In addition, ARID1B deletion was documented in PaCa samples and found to de-


regulated along with other SWI/SNF subunits (Shain, Giacomini et al. 2012). BAF180

(having polybromo domain and named as PBRM1) is a SWI/SNF subunit and is important for

interaction with acetylated histone (Zeng and Zhou 2002). In an exome sequencing study of

renal cancer patient samples, BAF180 was revealed to be most frequently mutated gene after

VHL (most frequently mutated gene in renal carcinoma) (Varela, Tarpey et al. 2011). BAF180

is a specific subunit of PBAF complex which is known to target specific set of gene loci in

context dependent manner. Therefore, BAF180 may not be inactivated in wide variety of

cancers but can be targeted in particular cancers viz renal carcinoma (Brownlee, Chambers et

al. 2012).

2.3.4 ARID1A and ARID1B SWI/SNF subunits as possible TSGs in PaCa

As discussed in 2.3.3, SWI/SNF complex is emerging as lineage specific deregulator in

cancer. For example, SNF5 and BRG1 inactivation was documented predominantly in

pediatric cancer and lung cancer respectively. Similarly, ARID1A predominantly inactivated

in ovarian clear cell carcinoma as shown in more than 55% of patient samples (Jones, Wang

et al. 2010) and more than 45% cases in ovarian carcinoma (Wiegand, Shah et al. 2010).

Concomitant loss of ARID1A expression was also observed in ovarian carcinoma in a tissue

microarray study (Wiegand, Lee et al. 2011). ARID1A mutation was also observed in other

types of cancer, but at a lower frequency (Cornen, Adelaide et al. 2012; Zang, Cutcutache et

al. 2012). ARID1B mutation also emerging to be linage specific as found in a highly

aggressive pediatric neuroblastoma patients (Sausen, Leary et al. 2013). Studies have shown

SWI/SNF components including ARID1A and ARID1B to be de-regulated in PaCa. For

example a recent study on PaCa patient samples identified copy number loss or mutational

inactivation of SWI/SNF components including ARID1A, ARID1B, PBRM1, BRG1

(SMARCA4), BRM (SMARCA2) and other core subunits of the complex and functional

characterization of SWI/SNF components in PaCa cell lines suggested a possible tumor


suppressor function (Shain, Giacomini et al. 2012). More importantly, previous aCGH studies

on PaCa patient samples and cell lines identified loss of 6q25 (Bashyam, Bair et al. 2005;

Birnbaum, Adelaide et al. 2011).


Table 2.1: List of SWI/SNF subunits alterations in cancer

Human SWI/SNF in cancer SWI/SNF

component

Subunit Alteration Cancer

involved

Origin

Enzyme BRG1 Mutation1,2,5

Lung cancer Primary tumor

Mutation1,3

NSLC Cell line

Mutation1 Medulloblastoma Primary tumor

Mutation3 Rhabdoid Primary tumor

Mutation1,3

Pancreas, breast

and prostate

Cell line

BRM Loss of expression Lung cancer Primary tumor

Copy number loss Hepatocellular

carcinoma

Primary tumor

Core SNF5/BAF47 Homozygous deletion,

mutation1,2,4

Rhabdoid cancer Primary tumor and

cell lines

Mutation3 Schwanomatosis Primary tumor

Translocation and

homozygous deletion

Hepatoblastoma Primary tumor

Homozygous deletion Epitheloid

sarcoma

Primary tumor

Mutation1 Menigioma Primary tumor

BAF57 Mutation3

and loss of

expression

Breast cancer Primary tumor and

cell lines

Accessory ARID1A/BAF250a Mutation3 Ovarian clear cell

cancer

Primary tumor and

cell lines

Loss of expression Uterine

endometiroid

cancer

Primary tumor

Mutation1,2,3,5,6

Gastric cancer Primary tumor

Mutation1,2,3,5,6

Hepatocellular

carcinoma

Primary tumor

Mutation1,5,6

and copy

number loss

Breast cancer Primary tumor

Mutation1,2,5,6

Bladder cancer Primary tumor

Mutation1 Renal cancer Primary tumor

Mutation3 Medulloblastoma Primary tumor

ARID1B/BAF250b Hemizygous deletion Neuroblastoma Primary tumor and

cell lines

Mutation3 Breast cancer Primary tumor

Copy number loss Pancreatic cancer Primary tumor

Copy number loss Astrocyte cancer Primary tumor

BAF180 Mutation1,2,3,4

Renal cancer Primary tumor and

cell lines

Mutation3 Breast cancer Cell lines

BAF200/ARID2 Mutation2,4,7

Hepatocellular

carcinoma

Primary tumor

1-Missence, 2-Nonsense, 3-Truncating, 4-Frameshift, 5-insertion, 6-Deletion, 7-Splice site

NSLC-Non small cell lung carcinoma


2.4 Chromosomal deletion at 18q23 in PaCa

Chromosomal arm 18q is widely suspected to be deleted in many cancers including

gastrointestinal cancers, head and neck cancers, prostate and ovarian cancers. For example,

head and neck squamous cell carcinoma showed a poor prognosis in patients having 18q

LOH (Agada, Patmore et al. 2009). CGH and cytogenetic studies revealed 18q loss as an

early event in prostate cancer progression (Nupponen and Visakorpi 2000). Frequent loss of

18q was also observed in colorectal cancer and PaCa. Colorectal cancer harbors one of the

most frequent allelic losses at 18q in advanced stage (Migliore, Migheli et al. 2011). Genome

profiling studies of PDAC revealed many frequent chromosomal aberrations including loss of

18q (Karhu, Mahlamaki et al. 2006). In a study of 27 primary pancreatic tissue using high

density aCGH detected 18q loss in more than 60% of samples (Harada, Chelala et al. 2008).

A high resolution aCGH study on 39 PDAC tumor tissues revealed loss of 18q in more than

50 % of samples (Birnbaum, Adelaide et al. 2011). Importantly, chromosome 18q is a highly

gene rich region, but very few genes have been annotated till date

(http://atlasgeneticsoncology.org/Indexbychrom/idxl_18.html). A small number of genes have

however been annotated as TSG (Table 2.2). For example, 18q11.2 harbors RBBP8

(retinoblastoma binding protein 8 also known as CtIP) gene and is shown to have tumor

suppressive function in leukemia mice models (Chen, Liu et al. 2005). Desmoglein-2 gene

maps to chromosome 18q12.1 and was observed to be down regulated in PaCa (Ramani,

Hennings et al. 2008) and gastric cancers (Yashiro, Nishioka et al. 2006). SMAD2 (18q21), a

TGFβ signaling component, is a well-studied TSG found to be inactivated in colorectal

carcinoma (Eppert, Scherer et al. 1996; Takagi, Koumura et al. 1998). SMAD4 (also known

as SMAD4/DPC4) located at 18q21.1 is perhaps the most well studied TSG in this region and

is known to be frequently deleted in PaCa as well as inactivated in other cancers such as

colorectal cancer at a lower frequency (Hahn, Schutte et al. 1996; Miyaki and Kuroki 2003).

http://atlasgeneticsoncology.org/Indexbychrom/idxl_18.html


However, it is possible that the 18q region likely harbors additional important TSGs and

application of genomic techniques might be useful in this regard. An NCBI gene database

search for 18q23 region revealed an important human gene viz partition defective 6 homolog

gamma (PARD6G) (http://www.ncbi.nlm.nih.gov/gene/?term=18q23). PARD6G is known to be

a component of polarity complex.

2.4.1 Cell polarity

Cell polarity is a fundamental feature of cell structure, which determines spatial differences in

shape and function of the cell. Well known examples of cell polarity function include

formation of asymmetric cells during early embryonic development, apical-basal polarity of

epithelial cells, polarity of neurons and polarized cell migration. Polarity is an important

basic requirement in a wide range of cellular functions. For example, asymmetric cell

division is required during early embryonic development to produce progenitor cells of

different fate and lineage for future organogenesis. The phenomenon of cell polarity is

evolutionarily conserved and several studies have been performed on model organisms like

C. elegans and Drosophila to understand its mechanism (Hawkins and Garriga 1998).

Polarity regulators play an important role in determining anterior-posterior body axis during

first asymmetric cleavage of the C. elegans zygote (Schneider and Bowerman 2003). Cell

polarity is an important feature of epithelial cells. In epithelial tissue, the apical membrane of

cells is structurally and functionally different from the basolateral membrane. Apical

membrane communicates with adjacent cells through the tight junction while basolateral

membrane contains adherence junction and gap junctions (Fig 2.3). This asymmetric apical-

basal polarity of epithelial cells is regulated through junction polarity complexes (Shin, Fogg

et al. 2006). Apical-basal polarity is also important for physiological function of epithelial

cells. For example, apical membrane of polarized renal epithelial cells has more secretory

function than absorptive role (Sekine, Miyazaki et al. 2006). Polarity of neurons is very

http://www.ncbi.nlm.nih.gov/gene/?term=18q23


important for directional flow of nerve impulse. Mature neurons have a highly polarized

structure, achieved through polarity regulators and asymmetric distribution of cytoskeletal

components (Horton and Ehlers 2003). Similarly, establishment of polarity is an important

requirement for cell migration. Directional cell migration is an important event during

embryonic morphogenesis, tissue repair and regeneration. A key feature of migrating cell is

the presence of highly polarized structure in terms of cellular components from the leading to

lagging end. For instance, lamellopodia or filopodia formation at the leading edge requires

actin polymerization machinery and interaction of filopodial protein with the extracellular

matrix (Ridley, Schwartz et al. 2003). The sustained polarity of migrating cells is achieved

through various polarity complexes, such as Par-polarity complex and the chemotactic factors

in the surrounding (Ridley, Schwartz et al. 2003).

Figure 2.3: Diagram represents basic structure of epithelial cell in epithelial tissue

2.4.2 Polarity complex

Structural and functional polarity axis (apical-basal polarity) of epithelial cell is characterized

by the presence of two different cell-cell junctions, commonly known as tight junction (Li,

Wientjes et al.) and adherence junction (Are, Colburn et al.). These two junctions divide


plasma membrane of epithelial cells into biochemically and structurally different domains.

One of the important molecular determinants of epithelial cell polarity is polarity complex.

Genetic studies on model organism like C.elegans and Drosophila in last two decades

revealed three major polarity complexes as epithelial cell polarity determinants; 1) “Par”

(partition) complex, 2) Scrib complex and 3) Crb complex. These polarity complexes are

highly conserved from yeast to mammalian cells (Assemat, Bazellieres et al. 2008).

1) “Par” complex

Par complex is a tri-molecular protein complex localized at the anterior compartment of the

cell. Par family of proteins was first identified in C. elegans during screening of maternal

effect lethal mutants for detection of early partition developmental regulated genes

(Kemphues, Priess et al. 1988). These mutants lose apico-basal polarity and asymmetric cell

division of blastomeres in early gastrulating embryonic cells which leads to the formation of

defective body pattern. Six different mutants were identified and named “par”(for partition)

according to their phenotype (par1 to par6); each of the mutants have different roles during

early embryonic development (Kemphues, Priess et al. 1988). For instance, par1 and par4

were shown to have Serine/Threonine kinase activity. Par1 was demonstrated to be localized

at the posterior part of zygote, while par4 exhibited a uniform distribution. Both par1 and

par4 were shown to be responsible for asymmetric cleavage of zygote to produce daughter

cells of different fate (Guo and Kemphues 1996; Watts, Morton et al. 2000). Par3 gene was

characterized and shown to have asymmetric distribution in zygote and early blastomeres and

may be required to regulate polarity in the early embryonic development of C.elegans

(Etemad-Moghadam, Guo et al. 1995). Another important study in C.elegans par-mutant led

to the identification of the atypical protein kinase C (aPKC) as inter-acting partner of par3,

which co-localize together at the periphery of early blastomeres. Also, aPKC mutant studies

demonstrated that interaction of par3-aPKC is required for asymmetric cell division during


early stages of C.elegans embryogenesis (Tabuse, Izumi et al. 1998). Similarly, three par6

mutant of C.elegans exhibited phenotype similar to par3 mutant (Watts, Morton et al. 2000).

Par6 co-localization with par3 at the anterior periphery was shown to be dependent upon the

activity of par3/aPKC complex (Joberty, Petersen et al. 2000). Taken together, these findings

suggested par3/aPKC/par6 as important component of the polarity complex (Hung and

Kemphues 1999; Goldstein and Macara 2007; Assemat, Bazellieres et al. 2008). This

complex is perhaps the first polarity complex shown to be involved in the polarity of

migrating cells. Functional dependence of “par” complex on the small GTPase molecule

CDC42 was discovered and a physical interaction of par6 with CDC42 was demonstrated in

C.elegans embryo (Joberty, Petersen et al. 2000; Gotta, Abraham et al. 2001). Although, this

tri-molecular par-complex was discovered in C.elegans it appears to be conserved from

worms to mammals. In an investigation, three human homologues of PAR6 were identified

and named as PAR6A or PAR6α or PARD6A, PAR6B or PAR6β or PARD6B and PAR6G or

PAR6γ or PARD6G (Noda, Takeya et al. 2001). All the three homologues harbor a conserved

PDZ domain and a comparatively less conserved CRIB domain (Noda, Takeya et al. 2001).

Similarly, two PAR3 human homologues have been identified in humans namely PAR3A or

PAR3α and PAR3B or PAR3β. PAR3α was associated during tight junction establishment in

epithelial cells, but PAR3β lacks aPKC binding domain and is not known to be functionally

active at the tight junction (Kohjima, Noda et al. 2002). Interaction of mammalian aPKC with

PAR3/PAR6 appears to be an important prerequisite for establishment of cell polarity

(Helfrich, Schmitz et al. 2007). Moreover, PAR3/PAR6/aPKC localization was determined at

the tight junction of epithelial cells (Izumi, Hirose et al. 1998; Lin, Edwards et al. 2000).

2) Scribble complex

SCRIB (a component of scribble complex) first identified in Drosophila as a regulator of

septate junction of epithelial cells. SCRIB was shown to act along with two proteins lethal


giant larvae (LGL) and disc large (DLG) (Bilder, Li et al. 2000). Many human homologues of

SCRIB components have been identified. Moreover, this complex was also shown to have

functional interaction with “par” complex in Drosophila as well as in mammals during

establishment of polarity in epithelial cells (Benton and St Johnston 2003; Yamanaka,

Horikoshi et al. 2003). Scribble complex was shown to be important for adherence junction

integrity and cell-cell contact. Scribble complex was also shown to be involved in the PI3K

mediated regulation of E-cadherins function at adherence junction (Laprise, Viel et al. 2004).

3) Crumbs (Crb) complex

CRB gene (a component of Cumbs complex) was originally identified in Drosophila as sub

apical localizing factor, responsible for formation of normal epithelium and cuticle. The

mutant phenotype showed a disorganized epithelium and has been observed a loss of

epithelial cell polarity (Tepass, Theres et al. 1990; Assemat, Bazellieres et al. 2008). A

genetic study on Drosophila crb mutant led to the identification of another related protein

named as sdt (Stardust mutant) (Tepass, Theres et al. 1990). Another functional interactant of

crb was discovered in Drosophila and named as DpatJ, and shown to be important for

epithelial cell polarity (Bhat, Izaddoost et al. 1999). Mammalian homologues of crb complex

were identified and shown to be involved in regulation of tight junction assembly in the

human breast cancer cell line MCF7 (Straight, Shin et al. 2004). Other mammalian

homologues such as PALS1/MPP5 and PATJ/INADL were shown to be associated in

epithelial junction assembly and formation of polarity (Assemat, Bazellieres et al. 2008).

Further, PALS1 in Crb complex is a pre requisite for the establishment of polarity and an

active Crb complex was shown to harbor PALS1 and in direct interaction (Roh, Fan et al.

2003; Shin, Straight et al. 2005; Pieczynski and Margolis 2011).


Figure 2.4: Interaction and localization of polarity complex components in a polarized

epithelial cell. [adapted from (Iden and Collard 2008), reproduce with permission from

Nature Publishing Group], dashed arrow represents translocations.

The three tri-molecular polarity complexes (PAR3/PAR6/aPKC; SCRIB/DLG/LGL;

CRB/PALS1 PATJ) appear to be highly conserved and may play a pivotal role in defining

apico-lateral, apical and basolateral domain of epithelial cells. Polarity complexes localize

epithelial cell-cell junctions (TJ & AJ) and provide a central regulatory pathway for the

establishment of polarity and epithelial functions (Fig 2.4 & (Iden and Collard 2008),

(Martin-Belmonte and Perez-Moreno 2012).

2.4.3 “Par” complex and cell migration

Cell migration is an important physiological feature in many normal physiological conditions

(e.g. stem cell migration for renewal of skin and intestinal cell) as well as in various

pathological conditions such as invasive carcinoma and metastasis. Cell migration is basically

a cyclic process and starts with polarization towards direction of movement in presence of

migration-promoting agents. At the leading side of migration, cells form a protrusion either

as lamellipodia or filopodia which helps the migrating cells in binding to extracellular matrix

(ECM). Simultaneously, detachment from the lagging side allows the cells to move over the

ECM. This protrusive machinery is generated and sustained by the cell cytoskeleton,


microtubule organizing centre MTOC and golgi apparatus through a dynamic process of

polymerization and de-polymerization of actin filaments (Ridley, Schwartz et al. 2003; Bose

and Wrana 2006). Critical regulators of migration are family of small GTPases like RHO,

RAC and CDC42 (Charest and Firtel 2007). The regulatory processes of the three GTPases

are a fundamental requirement for migration. In brief, activation of CDC42 and RAC are

required at the leading edge where actin and microtubule reorganization takes place for the

protrusive activity. Whereas, RHO activity is more at the rear end which helps in the

induction of actomyosin complex, which gives a contractile force for retraction and allows

the forward movement (Etienne-Manneville 2008). As discussed in 2.4.2, polarity complex

plays a key role in establishment of cell polarity. Decades of research have also suggested a

fundamental role of these polarity complexes in cell migration. Specifically, crosstalk of par

proteins with GTPase family suggested a key role of par proteins during polarization of

migrating cells (Aranda, Nolan et al. 2008). The cascade of events regulating cell migration

has been studied in human astrocytes and have shown that integrin activation at the leading

edge of migrating cell activates and recruits CDC42 and the activated CDC42 in turn recruits

and activates PAR6A/aPKC which interacts with PAR3 (Joberty, Petersen et al. 2000;

Etienne-Manneville and Hall 2001). This activated par-complex is necessary for activation of

downstream target such as the dynin motor protein, required for microtubule and actin

polymerization (Joberty, Petersen et al. 2000; Etienne-Manneville and Hall 2001). Further

elucidation of this cascade led to the discovery of a novel link between par-complex and

WNT signaling. Activated aPKC at the leading edge was shown to phosphorylate glycogen

synthase-3β (GSK3β) at the inhibitory site and inactivates it, and the inactivated GSK3β was

shown to promote the clustering of adenomatous polyposis coli (APC) with the plus end of

microtubule and thus stabilizing the growing microtubules in migrating cells (Etienne-

Manneville and Hall 2003). Also, it has been shown that activated PAR6/aPKC facilitates


clustering of the Scrib complex component DLG1 (a trans-membrane protein). DLG1

clustering facilitates interaction of DLG1 with APC at the polarized site of migrating cell

(Etienne-Manneville, Manneville et al. 2005) and ultimately helping in the reorientation of

microtubule organizing center (MTOC) towards the leading edge of migrating cell (Gomes,

Jani et al. 2005).

Figure 2.5: Schematic representation of regulation of cell migration at the leading edge

(Bose and Wrana 2006) reproduced with permission from Elsevier publishing group).

Simultaneously, activated PAR6/aPKC also plays a key role in regulation of RHOA at the

lagging end of migrating cells. SMURF1, (an E3 ubiquitin ligase) was shown to degrade

RHOA at the lagging end which makes lagging end cytoskeleton unstable (Wang, Zhang et al.

2003). While PAR6/aPKC complex was shown to form leading-lagging polarity and regulate

protrusive activity through CDC42; PAR3 mediates cell protrusion through interaction with

RAC GTPases. Dissolution of tight junction is a prerequisite for cell migration. Activation of

PAR3/TIAM1 (TIAM1 is Rac exchange factor1) complex was demonstrated to be the main

event at the leading edge of polarized migrating cells (Chen and Macara 2005). Lagging end

polarity defects of migrating cells were observed in TIAM knockout cells but detailed

mechanism is still not clear (Pegtel, Ellenbroek et al. 2007). Thus, “par” complex has

emerged as the central cellular machinery for generating leading-lagging polarity axis during

cell migration. The process of cell migration however is tightly controlled and is, regulated

by various cellular signaling pathways.


2.4.4 Regulation of “par” complex by cellular signaling

As described in section 2.4.3, cell migration is a highly regulated cellular mechanism co-

ordinated through polarity complex proteins mainly through “par” complex. Deregulation of

this mechanism leads to various pathological conditions including metastasis. There are two

known signaling pathways which regulate this process namely Integrin signaling and TGFβ

signaling (Bose and Wrana 2006). Polarized cell migration requires reorganization of

microtubule organizing centre (MTOC) and Golgi apparatus between leading edge and

nucleus (Gomes, Jani et al. 2005). Integrins are trans-membrane receptors which bind with

extracellular matrix (ECM) during cell motility (Huttenlocher and Horwitz 2011). Activation

of Integrin in focal adhesion complex (protein complex required at cell-ECM adhesion)

activates CDC42 which facilitates the activation of aPKC/PAR3/PAR6 complex. This tri-

molecular activated complex phosphorylate at inhibitory site of GSK3β which ultimately

block phosphorylation of APC and thus un phosphorylated APC can bind to growing end of

microtubule (Fig 2.5 and (Etienne-Manneville and Hall 2001; Etienne-Manneville and Hall

2003; Guo and Giancotti 2004; Bose and Wrana 2006). Another important feature of invasive

phenotype is the induction of epithelial to mesenchymal transition (EMT) (Thiery, Acloque et

al. 2009; Nauseef and Henry 2011). Interestingly, one of the hallmarks of EMT is the loss of

polarity. Loss of polarity during EMT is achieved through deregulation of transcription

program for the polarity related genes in cancer cells (Moreno-Bueno, Portillo et al. 2008).

EMT is a complex phenomenon of cellular plasticity and is tightly regulated by various

signaling pathways. TGFβ signaling pathway is a well-established regulatory pathway for the

induction of EMT (Xu, Lamouille et al. 2009). But the mechanism of TGFβ mediated loss of

cell polarity is not well understood. Studies have shown that TGFβ signaling may occur

through various alternative pathways. For example, canonical pathway through SMAD

mediated induction of transcription of mesenchymal specific genes (Feng and Derynck 2005;


Xu, Lamouille et al. 2009). Another TGFβ signaling for EMT induction is named as non-

canonical or non SMAD TGFβ pathway. This pathway is relatively less explored. However in

an interesting study, non-canonical TGFβ pathway was linked to “par” complex and this

study helped to understand the mechanism of EMT through loss of polarity mediated by non-

canonical TGFβ signaling (Ozdamar, Bose et al. 2005; Bose and Wrana 2006). This study

perhaps for the first time identified novel link between TGFβ signaling and EMT. The study

demonstrated that TGFβ activate TβRII, activated TβRII form hetero dimer with TβRI which

facilitate interaction of Par6/aPKC which ultimately activate and recruit E3 ubiquitin ligase

SMURF1 to RHOA through TβRII (Ozdamar, Bose et al. 2005; Wang, Nie et al. 2008).

Degradation of RHOA ultimately leads to loss of polarity and induction of EMT (Ozdamar,

Bose et al. 2005; Bose and Wrana 2006). In addition to classical EMT signaling such as

Integrin, TGFβ and WNT, small GTPases such as RAS and growth factor mediated signaling

through PIP3K also appear to regulate “par” complex during polarity establishment and cell

migration (Etienne-Manneville 2008). However, the mechanism is not very well-understood.

2.4.5 Polarity complex and cancer

Invasion and metastasis are the major cause of cancer related morbidity and mortality

(Hanahan and Weinberg 2011). Epithelial to mesenchyme transition (EMT) is particularly an

important hall-mark of carcinoma, where migrating tumor cells recapitulate developmental

process of EMT by altering the epithelium structure, disrupting the basal lamina and invading

the underlying tissues. Loss of integrity of tight junction and adherence junction and loss of

polarity are crucial features of EMT. As described in section 2.4.4, cell-cell contact is

maintained and regulated by several polarity proteins. These polarity proteins are often

targeted by EMT inducers leading to altered function of these complexes ultimately

facilitating process of carcinogenesis (Martin-Belmonte and Perez-Moreno 2012). EMT

related alterations include overexpression or deregulation of components of polarity complex,


mislocalization, deletion or production of altered protein through alternative splicing

(Guarino 2007; Etienne-Manneville 2008; Nolan, Aranda et al. 2008). Polarity complexes are

emerging as one of the key tumor suppressor complex in cancer progression. Past two

decades of research on polarity complex and carcinogenesis has provided a wealth of

knowledge for a better understanding of advanced stage carcinogenesis. Studies revealed

several components of polarity regulator complex (Crumbs, Scribble and Par) to be

deregulated in the process of carcinogenesis (Rothenberg, Mohapatra et al. 2010; Martin-

Belmonte and Perez-Moreno 2012)and see Table 2).


Table 2.2: List of alterations of polarity complex proteins in epithelial transformation

and human cancer

Polarity complex in cancer Polarity

complex

Component Alteration Cancer Origin

Scribble SCRIB Loss of expression

and mis-localization

Cervical cancer Primary tumor

Do Colon cancer Do

Do Breast cancer Do

Dlg1(Human

homologue of

Dlg)

Do Cervical cancer Do

Loss of expression Colon cancer Do

Lgl1 (Human

homologue of

Lgl)

Loss of expression Colon, breast, ovarian and

prostate cancers

Do

Splice alterations Hepatocellular carcinoma Primary tumor

and cell line

Lgl2 (Human

homologue of

Lgl)

Loss of expression

and mis-localization

Gastric cancer Primary tumor

Crumbs Crb3(Human

homologue of

Crb)

Loss of expression Mouse kidney epithelial

cells cells

Mouse model

Mis-localization Mouse mammary epithelial

cells

Do

Par aPKCzeta Overexpression Hepatocellular carcinoma Primary tumor

Do Bladder, head and neck

cancer, breast cancer and

PDAC

Primary tumor

and cell lines

aPKCiota Overexpression and

mis-localization

Ovarian cancer,

hepatocellular carcinoma,

NSLC and PDAC

Do

Par6A Overexpression Breast cancer and NSLC Do

Par6B Loss of expression Breast cancer Primary tumor

Par3

Deletion loss of

expression

OSCC Do

Papilloma formation Mouse skin cells Mouse model

Keratoacanthoma

formation

Do Do

PDAC- Pancreatic ductal adenocarcinoma, NSLC- Nonsmall cell lung carcinoma, OSCC-

Oesphageal squamous cell carcinoma


1) Role of scribble complex in cancer

Scribble complex is well studied polarity complex in cancer progression. This complex is

comprised of three components (DLG, LGL and SCRIB) first identified in Drosophila and

demonstrated to have tumor suppressor function (Gateff 1978; Bilder, Li et al. 2000). Further

studies strongly supported the tumor suppressor function of scribble complex components in

human cancers. For example, mis-localization and reduced expression of scribble complex

has been observed in invasive stage of cervical cancer (SCRIB component: (Nakagawa, Yano

et al. 2004), colon cancer (SCRIB and DLG components: (Gardiol, Zacchi et al. 2006),

prostate cancer (SCRIB component: (Pearson, Perez-Mancera et al. 2011) and colorectal

cancer (LGL component: (Schimanski, Schmitz et al. 2005). Loss of expression of scribble

complex proteins facilitates tumorigenesis in several ways. For example, a correlation of

SCRIB loss and elevation of MAPK signaling in the invasive stage of carcinoma has been

observed in prostate cancer mice model (Pearson, Perez-Mancera et al. 2011). Scribble

complex was also demonstrated to be deregulated by c-Myc mediated tumorigenesis in 3D

model of breast cancer (Zhan, Rosenberg et al. 2008). Interestingly, scribble complex

components were found to be reduced in cervical cancer and has been mechanistically linked

to HPV. HPV onco-protein E6 has been demonstrated to target SCRIB and DLG and induce

ubiquitin mediated degradation in several human cell lines (Nakagawa and Huibregtse 2000;

Thomas, Massimi et al. 2005).

2) Role of crumbs complex in cancer

Crumbs complex (components: CRB, PALS1 and PATJ) was first identified in Drosophila

and studies have shown that the complex is required for the maintenance and morphogenesis

of epithelial cell polarity (Bazellieres, Assemat et al. 2009). However, little is known about

the role of crumbs complex proteins in human carcinogenesis. Mouse models and human cell


line studies have shown a potential tumor suppressor role of crumbs complex in

tumorigenesis. For example, tumorigenic properties such as loss of contact inhibition, loss of

apico basal polarity and disassembly of tight junction in correlation with loss of expression of

CRB3 (human homologue of Crb) was identified in a study of Crb3 knockout mouse model

(Karp, Tan et al. 2008). In a breast cancer cell lines study, the loss of expression of Crb3

along with PALS1 was found to be mediated through ZEB1 and SNAIL (well-known

transcription repressor induced during EMT in epithelial cells) suggesting a possible link

between classical EMT and polarity regulators (Aigner, Dampier et al. 2007; Whiteman, Liu

et al. 2008). The role of PATJ and PALS1 (component of crumbs complex) in oncogenesis is

not well understood. An in vitro and human cell line based study revealed PATJ as a target

for ubiquitin mediated degradation through HPV protein E6 (Storrs and Silverstein 2007).

However, PALS1 was shown to be important for TJ maintenance and importantly, shown to

provide stability to Crb3 in crumbs complex through interaction with PATJ component

(Straight, Shin et al. 2004). Also, PALS1 has been shown to interact with aPKC which

suggested a link between crumbs complex and par complex in maintenance of TJ (Straight,

Shin et al. 2004).

3) Functional implication of “par” complex in cancer

“Par” complex is emerging as both oncogenic as well as tumor suppressive complex in a

context dependent manner during tumorigenesis (Aranda, Nolan et al. 2008). “Par” complex

comprises 3 components namely PAR3 (PAR3A and B in mammals), aPKC (aPKClambda,

aPKCzeta and aPKCiota in mammal) and Par6 (Par6A, B and G in human) (section 2.4.2). Par

complex plays a significant role for establishment and maintenance of tight junction in

epithelial cell polarity and hence has been shown to be deregulated and targeted during

carcinogenesis in several cancers (Aranda, Nolan et al. 2008). Multiple studies have shown


the link of aPKC deregulation in human cancers. For example, over-expression and/or mis-

localization of aPKC has been observed in non-small cell lung carcinoma (Regala, Weems et

al. 2005), ovarian carcinoma (Eder, Sui et al. 2005), hepatocellular carcinoma (Tsai, Hsieh et

al. 2000), breast carcinoma (Kojima, Akimoto et al. 2008) and pancreatic ductal

adenocarcinoma (Scotti, Bamlet et al. 2010) etc. The mechanism of deregulation of aPKC in

cancer is not clearly understood. However, a study of ovarian carcinoma identified a positive

correlation between over expression of Cyclin E and aPKC, which suggested a possible link

of aPKC with cell proliferation (Eder, Sui et al. 2005). Another component of “par” complex

is PAR3, which has been shown to have possible tumor suppressor function. For example,

significantly reduced expression of Par3 has been observed in esophageal squamous cell

carcinoma (ESCC) tumor tissues and shown to have a role in reestablishment of TJ in ESCC

derived cell line (Zen, Yasui et al. 2009). The mechanism of PAR3 mediated tumor

suppressor function is not clearly understood. However, studies in mouse model and cell lines

revealed possible link of PAR3 deregulation with classical oncogenic signaling pathways. For

example, in a study performed on rat epithelial cells, a dose dependent reduction of PAR3

transcript level has been identified upon TGFβ (well established inducer of EMT) treatment

and also demonstrated to have positive effect on TJ assembly in these cells (Wang, Nie et al.

2008).

Human PAR6 family consist of three isoforms namely PARD6A, PARD6B and PARD6G as

described above. PAR6 is a central scaffold molecule directly interacts with PAR3 and

CDC42 and activates the formation of TJ and its regulatory activity for TJ assembly is

determined by the interaction of aPKC (Aranda, Nolan et al. 2008). As mentioned earlier,

regulation of TJ is an important event to maintain epithelial cell polarity, PAR6 is therefore

expected to be deregulated in cancers and targeted by tumor signaling pathways. PARD6A

over-expression has been detected in breast cancer (Nolan, Aranda et al. 2008; Viloria-Petit,


David et al. 2009). Also, PARD6A has been shown to be over-expressed in stromal

compartment of non-small cell lung carcinoma tumor tissues and the high expression was

well correlated with good prognosis (Al-Saad, Al-Shibli et al. 2008). Moreover, recently, a

moderate reduction of PARD6B protein expression was identified in breast tumor tissue and

there was a positive correlation between PARD6B expression and maintenance of TJ

assembly (Cunliffe, Jiang et al. 2012). Except few studies, no information is available about

PAR6 isoforms expression level in cancers and hence the mechanism by which PAR6

isoforms are deregulated is not clearly understood. Mechanistically, a recent study of 3D

breast cancer model, ERBB2 (a well-known oncogene in breast cancer) was shown to replace

PAR3 in PAR6/PAR3/aPKC complex during in vitro acini formation (Aranda, Haire et al.

2006).

4) PARD6G as possible TSG in PaCa

Studies have suggested a possible oncogenic function for PARD6A and possible tumor

suppressor function for PARD6B as described in section 2.4.5. Interestingly, in a recent case

study of non-cirrhotic HFE hemochromatosis patient, hepatitis B virus (HBV) was shown to

integrate in 5’ upstream region of PARD6G and induce its over-expression (Pollicino, Vegetti

et al. 2013). But there is no direct evidence available for tumorigenic role for PARD6G

isoform in any cancer including PaCa. However, as described in section 2.4, the PARD6G

locus (18q23) is frequently lost in human cancers including PaCa. Therefore, given the

importance of PARD6A and PARD6B deregulation in cancers, PARD6G is expected to have

tumor suppressor function in PaCa.


2.5 Rationale and objectives of the present work

2.5.1 Characterization of ARID1B as novel TSG in PaCa

As discussed in section 2.3.4, ARID1A (isoform of ARID1B) was shown to be predominantly

inactivated and having low expression in several cancers such as ovarian carcinoma and

PaCa. But, the role of ARID1B in cancer progression (specifically PaCa) has not been well-

established. However, studies revealed frequent genomic loss at 6q25.3 in PaCa (Bashyam,

Bair et al. 2005; Ichimura, Mungall et al. 2006; Birnbaum, Adelaide et al. 2011; Shain,

Giacomini et al. 2012). Therefore, given the importance of 6q25.3 loss in PaCa, exploration

for a possible tumor suppressor function of ARID1B (located at 6q25.3) in PaCa is perhaps an

important step towards the discovery of novel therapeutic target.

2.5.2 Characterization of PARD6G as novel TSG in PaCa

As discussed in section 2.4, 18q harbors many TSGs and has been shown to be frequently lost

in cancers including PaCa. aCGH studies of PaCa cell lines and xenografts identified several

homozygous deletions including loss of 18q23 (Heidenblad, Schoenmakers et al. 2004;

Bashyam, Bair et al. 2005; Nowak, Gaile et al. 2005). In addition to PaCa cell lines and

xenografts, aCGH studies on primary PaCa tissues also documented loss of 18q23 region. For

instance, an aCGH study on 23 PDAC tissues detected homozygous deletion at 18q23

(Harada, Baril et al. 2007). An NCBI gene database search for 18q23 reveals an important

human gene partition defective 6 homolog gamma (PARD6G)

(http://www.ncbi.nlm.nih.gov/gene/?term=18q23). In spite of the fact that “par” complex is

important for various cellular functions, knowledge of its implication in cancer and

specifically PaCa remains to be elucidated. Importantly, as discussed in sections 2.4.2 &

2.4.5, PAR6 isoform PARD6A has been implicated to have an oncogenic function and one

study of PARD6B suggested a possible TSG function, but no information is available for the

http://www.ncbi.nlm.nih.gov/gene/?term=18q23


role of PARD6G in PaCa. Moreover, given the frequent loss of 18q23 in PaCa and

tumorigenic role of other isoforms of PAR6, PARD6G suspected to have tumor suppressor

function in PaCa. Therefore, current study on PARD6G may reveal a possible tumor

suppressor function of the gene in PaCa.

2.5.3 Objectives of present work

The aim of the present study is to analyze the tumor suppressor role of ARID1B and PARD6G

genes in PaCa using molecular genetics (mutation screening and epigenetics) and cell biology

approach.

Objective for molecular genetics approach include

1. Screening of ARID1B and PARD6G gene sequences for detection of point mutation (a

common mode TSG inactivation in cancers) in PaCa cell lines harboring suspected

single copy loss.

Assessment of tumor suppressor role of ARID1B and PARD6G through epigenetics approach

specifically included

1. Evaluation of ARID1B and PARD6G transcript level after Azacytidine and

Trichostatin A treatment in PaCa cell lines harboring reduced transcript level.

2. Prediction of CpG islands of both genes using bioinformatics approach.

3. Bisulfite genome sequencing (BGS) analysis of CpG island(s) of ARID1B and

PARD6G in PaCa.

To gain insight into the function of ARID1B gene in PaCa, cell based strategy was employed.

The specific objectives included

1. Generation of ARID1B stable expression clones of MiaPaCa2 cell line (harboring

homozygous deletion of ARID1B).

2. Analysis of rate of cell growth kinetics of ARID1B stable clones.


3. Assessment of colony formation potential of ARID1B clones in liquid and solid

media.

4. To evaluate the apoptosis properties of ARID1B stable clones.

5. To explore cell migration ability of ARID1B stable clones in an in vitro wound

healing assay.

TSGs usually have reduced level of expression in cancers. Down-regulation of TSGs occurs

at transcription level through promoter and other regulatory elements. The specific objectives

are to evaluate transcription regulation of PARD6G included

1. Experimental determination of transcription start point.

2. Prediction of promoter region and transcription factor binding elements using

bioinformatics approach.

3. Evaluation of transcription activity of putative promoter using Luciferase reporter

assay.

4. Prediction of transcription factor binding elements in transcription activator and

transcription repressor region.

Chapter 2shodhganga.inflibnet.ac.in/bitstream/10603/11669/11... · 2.3 Chromosomal deletion at...

Documents

Transcript of Chapter 2shodhganga.inflibnet.ac.in/bitstream/10603/11669/11... · 2.3 Chromosomal deletion at...