University of Southern California Dissertations and Theses

Investigating the effects of polycomb repressive complex inhibitors on cancer cell phenotypic plasticity

(USC Thesis Other)

Investigating the effects of polycomb repressive complex inhibitors on cancer cell phenotypic plasticity

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Investigating the Effects of Polycomb Repressive
Complex Inhibitors on Cancer Cell Phenotypic
Plasticity

By: Nofar Avihen Schahaf

A Thesis Presented to the
FACULTY OF THE USC KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
MOLECULAR MICROBIOLOGY AND IMMUNOLOGY

May 2020

ii

Acknowledgements

I would like to offer my gratitude to my advisor Dr. Amir Goldkorn for giving me the
opportunity to conduct this project. I am grateful for his mentorship, feedback and support
throughout this process. I would also like to thank my supervisor, Dr. Tong Xu for her guidance
and for teaching me the techniques required for this research. I am extremely grateful for the
assistance provided by Dr. Gareth Morrison and Emmanuelle Hodara, who helped to further
develop my project. I would also like to thank Dr. Yi-Tsung Lu and Alexander Cunha for their
insights and guidance. I would like to offer a special thanks to our collaborator, Dr. Oliver Bell,
for his extremely valuable cooperation throughout this research project, and to the guidance of
Daniel Bsteh who taught me some of the most important lab techniques. Lastly, I would also like
to thank Dr. Amir Goldkorn, Dr. Joseph Landolph and Dr. Keigo Machida for serving as members
of my committee.

None of this would have been possible without the support of my husband, Eidan Schahaf,
my parents and my sister, and I want to thank them for their encouragement, love, and especially
for believing in me.

iii

Table of Contents

Acknolwedgements..........................................................................................................................ii
Figures............................................................................................................................................iv
Abstract............................................................................................................................................v
Introduction......................................................................................................................................1
Hypothesis......................................................................................................................................14
Materials & Methods.....................................................................................................................16
Results............................................................................................................................................21
Discussion......................................................................................................................................32
References......................................................................................................................................37

iv

Figures

Figure 1 Schematic representation of histone modification regulation by PRC1 10
and PRC2 complexes
Figure 2 Chemical structure of UNC4976 12
Figure 3 Chemical structure of UNC7040 13
Figure 4 Chemical structure of UNC7263 14
Figure 5 Chemical structure of GSK126 15
Figure 6 Immunoblotting for analysis of H3K27me3 23
Figure 7 SP subpopulation size in J82 cells 6 days after GSK126 treatment 25
Figure 8 SP subpopulation size in J82 cells 6 days after UNC4976 treatment 26
Figure 9 SP subpopulation size in J82 cells 6 days after UNC7040 treatment 27
Figure 10 Confluence-dependent model with J82 cells 28
Figure 11 Confluence-dependent model with T24 cells 29
Figure 12 SP subpopulation size in J82 cells 1 day after UNC7040 treatment 30
Figure 13 SP subpopulation size in J82 cells 1 day after GSK126 treatment 31
Figure 14 Immunoblotting for analysis of H3K27me3 32
Figure 15 SP subpopulation size in J82 cells 6 days after lentivirus infection for 33
EZH2 overexpression.

v

Abstract

Stem cell-like phenotypes such as drug resistance and tumorigenicity have been
traditionally viewed as pre-existing characteristics of a specific subpopulation of cells within the
whole tumor population. This assumption has been challenged by recent studies, showing that the
stem-like drug-resistant phenotype can emerge de novo by conversion of a subpopulation of cells
initially exhibiting a well-differentiated drug-sensitive state. Previous studies from our laboratory
established the cancer stem-like plasticity model, by analyzing cancer cell lines over time using
flow cytometry with Hoechst 33342 exclusion. This method is commonly used to yield a side
population of cells with drug-resistant highly tumorigenic properties, and a non-side population
lacking these properties. This model showed that these two subpopulations are able to go back and
forth between the two phenotypes in a rapid and cyclical manner, making them a “moving target”.
This phenotypic plasticity allows cancer cells to spontaneously and without external selective
pressure switch their phenotype and acquire a survival advantage under fluctuating environmental
conditions. The rapid and dynamic phenotypic transition is assumed to be in part epigenetically
regulated, as alterations in chromatin accessibility and DNA methylation has been previously
observed between genetically-identical subpopulations by the Goldkorn lab and other groups. In
the present study, we investigated the potential involvement of the Polycomb Repressive
Complexes (PRC), inhibitory chromatin modifiers, in the epigenetic regulation of phenotypic
plasticity. To test this, we targeted EZH2 and CBX, components of PRC2 and PRC1, respectively,
that are involved in a cascade of events that leads to epigenetic transcriptional repression. EZH2
is a tri-methyl writer, responsible for targeting a specific genomic region and depositing the
H3K27me3 repressive mark subsequently recognized by CBX proteins. We used pharmacological
vi

inhibitors to target these interactions and performed side population analysis to identify the highly
tumorigenic stem-like cells and monitor for shifts in the emergence of this subpopulation post-
treatment.

1

Introduction

1.1. Cancer cell phenotypic plasticity
Most tumors exhibit intra-tumoral heterogeneity. Tumor cells undergo molecular and
phenotypic changes collectively referred to as cellular plasticity (Yuan et al., 2019). While it is
well known that genetic mutations are a driving force for tumor development and resistance,
phenotypic plasticity can also provide a critical adaptive advantage to the overall cancer cell
population by giving rise to drug-tolerant cells. Phenotypic plasticity is the development of
multiple phenotypes from a single genotype, leading to mutation-independent heterogeneity (Jolly
et al., 2018).
Tumor heterogeneity can manifest in two different ways: inter-tumoral, whereby genotypic
and phenotypic alterations are observed between patients with the same tumor type, or intra-
tumoral, leading to phenotypic and functional alterations between cells within the same tumor (da
Silva-Diz et al., 2018). Phenotypic diversity can result from microenvironmental cues, epigenetic
alterations or treatment-imposed selective pressures. These conditions lead to a heterogenous
tumor cell population and facilitate metastasis and therapeutic resistance (Yuan et al., 2019; Zheng
et al., 2015). Phenotypic plasticity allows cancer cells to change state (phenotype) and shift
dynamically between a well-differentiated state, with limited tumorigenic potential, and a tumor-
initiating stem cell-like state. Stem-like cells are associated with a more aggressive, drug-resistant
and tumorigenic phenotype relative to non-stem-like cells (Cabrera et al., 2015). Such phenotypic
plasticity can provide cells with a survival advantage in unpredictable fluctuating environments
(Jolly et al., 2018).
2

Epithelial–mesenchymal plasticity may be the best-known case of tumor cell plasticity,
where cells may switch between epithelial and mesenchymal states several times during
development (Zheng et al., 2015). A study in non-small cell lung cancer (NSCLC) demonstrated
that aggressive treatment of multiple NSCLC cell lines sensitive to epidermal growth factor
receptor (EGFR) tyrosine kinase inhibition led to the isolation of slow-cycling drug-tolerant
persisters. When propagated in drug-free media, persisters cells resume growth and regain
sensitivity to EGFR inhibition (Sharma et al., 2010). Similarly, glioblastoma stem cells showed a
drug-tolerant and persister-like phenotype following RTK inhibitor exposure, and regained
sensitivity after drug washout (Liau et al., 2017). Over recent decades, such dynamic and adaptive
phenotypic switching has presented an ongoing challenge to the design of cancer therapies. This
intra-tumor heterogeneity drives cancer progression and treatment failure, thereby hindering
cancer prognosis and promoting disease recurrence (da Silva-Diz et al., 2018).

1.2. The cancer stem-like plasticity model:
Cancer stem-like cells function as an aggressive subpopulation localized within cancer
niches and play critical roles in tumor initiation and metastasis. Traditionally, molecular
phenotypes such as high tumorigenic potential and drug resistance have been attributed to this
small pre-existing subpopulation of cancer cells (Tsuchida et al., 2008; Kelly et al., 2007).
Nonetheless, recent work in cancer cell lines suggest that drug-resistant tumor-initiating features
can emerge de novo, and that cancer cells initially lacking such phenotypes are capable of tumor
initiation under certain permissive conditions (Shmelkov et al., 2008; Sharma et al., 2010; He et
al., 2011). The established idea of a pre-existing cancer stem-like subpopulation was then called
into question. Remarkably, it has been shown that phenotypic plasticity exists between stem-like
3

and non-stem-like cells in spontaneously growing tumor cells without interventions, allowing cells
to go back and forth between the two states (He et al., 2011; Chaffer et al., 2011; 2013).
Stem-like cells are known to possess the ability to efflux certain fluorescent dyes such as
Hoechst 33342. This ability, granted by the Adenosine Triphosphate–Binding Cassette (ABC)
transporter, has long been exploited for the isolation of these cells by fluorescence-activated cell
sorting (FACS) (Lu et al., ,2013). This method requires the incubation of the cell population with
Hoechst 33342 and subsequent FACS analysis of dual-wavelength Hoechst fluorescence. The
stem-like cells, displaying low fluorescence, are gated together and are termed the side population
(SP). The rest of the cells, not exhibiting a stem cell-like phenotype, are referred to as the non-side
population (NSP) (Yano et al., 2005).
Side population analysis was used in the Goldkorn lab to identify and monitor the stem
cell-like side population over time. Using this method, it has been observed that cancer cells can
fluctuate in a cyclical manner between SP and NSP phenotypes. While SP cells can differentiate
into the NSP phenotype, NSP cells can lead to the re-expansion of the SP phenotype by conversion.
These results, observed in both culture and tumor xenografts, show evidence that drug-resistant
highly tumorigenic phenotype can be rapidly and simultaneously acquired by a large number of
cancer cells previously lacking such traits. Moreover, cancer promoting characteristics can not
only arise under external selective pressure, but as a spontaneous, highly regulated transition (He
et al., 2011).
The cyclic transition between SP and NSP subpopulations has been characterized by
following their day-by-day dynamics. Cells growing in their native state were monitored daily
through FACS analysis coupled with green fluorescent labeling, showing that the SP
subpopulation became rapidly depleted by conversion into NSP cells after 1 day of seeding, and
4

reconstituted by conversion of NSP cells into the SP phenotype between days 3-6. During this
cycle of SP depletion and reconstitution, the largest SP subpopulation size was observed at day 6,
as the cells reached confluence, and at that point the cells were subcultured into new plates and the
cycle was repeated. This 2-way state switching represents a significant demonstration of
phenotypic plasticity within an isogenic population, as this dynamic transition occurs within two
genetically identical subpopulations (He et al., 2011; 2014; Jolly et al., 2018).

1.3. Epigenetic mechanisms in cyclical cancer plasticity
Previous work by our group demonstrated that phenotypic reprogramming exists within
genetically homogeneous cancer cell populations. We showed that tumor cells are able to
spontaneously reacquire stem-like drug-resistant properties, and that this subpopulation fluctuate
in size cyclically and predictably through serial passages in culture (He et al., 2011). We
subsequently reported that this phenotypic transition is in part mediated by the PI3K/AKT
signaling pathway, leading to CBP/β-catenin transcriptional activation that is implicated in
survival and proliferation of tumor cells in many human cancers (He et al., 2014; Shang et al.,
2017). Other groups studying different cancer cell lines also showed the emergence of slow-
cycling drug-tolerant cells that were found to be dependent upon histone modifications and several
signaling pathways, suggesting a mutational-independent plasticity (Sharma et al., 2010; Roesch
et al., 2013; Liau et al., 2017). Collectively, these observations established that adaptive cancer
stem-like phenotypes can cyclically re-emerge in cancer cell populations, driven by transcriptional
and signaling networks. This dynamic and rapid transition between phenotypes does not conform
into a model of acquired genomic mutations, suggesting that the resistant phenotype is mediated
epigenetically (Knoechel et al., 2014; Xu et al., in press, 2020).
5

To establish the idea of epigenetic plasticity, our group used a new assay recently
developed by our collaborators in the Liang laboratory. This assay, termed AcceSssIble, involves
comparing methylation of nuclei treated with a CpG methyltransferase enzyme, M.SssI, which
specifically methylates unmethylated CpG sites, with a no-enzyme treated control using an
Illumina HumanMethylation450 BeadChip. The accessibility or the lack thereof to M.SssI is
determined by acquisition of methylation post-treatment compared with the untreated control, and
can be used to infer the positioning of nucleosomes and provide information on chromatin state
(Pandiyan et al., 2013). This method was used in order to test whether epigenetic alterations are
involved in the interconversion between SP and NSP cells and revealed the existence of differential
chromatin accessibility and global DNA methylation between the two subpopulations. This
epigenetic plasticity, leading to differential gene expression, was found to plays a key role in the
transition between the two phenotypes (Xu et al., in press, 2020).
Many epigenetic regulators have been implicated in chromatin remodeling. The Polycomb
Repressive Complex has been previously identified as a chromatin modifier involved in many
repressive activities that lead to transcriptional silencing (Müller and Verrijzer, 2009). The
Polycomb group (PcG) proteins, predominantly EZH2 and the CBX protein family, were found to
be involved in many biological processes, such as cell differentiation, epithelial–to-mesenchymal
transition and promotion of stem-like traits (Tang et al., 2019). These cancer-promoting activities,
as well as our findings of shifts in global chromatin accessibility, lead to the hypothesis that the
PcG proteins are involved in the phenotypic plasticity between SP and NSP cells.

6

1.4. The Polycomb Repressive Complexes:
The Polycomb group protein complexes are evolutionary-conserved epigenetic regulators
first discovered in Drosophila, that were found to be involved in chromatin remodeling and
transcriptional repression (Lewis, 1978). Polycomb Repressive Complexes 1 and 2 (PRC1 and
PRC2) are the best characterized forms of the PcG complexes and are recognized for their role in
many processes such as cell differentiation, stemness of embryonic stem cells, cell-cycle control
and X-inactivation (Müller and Verrijzer, 2009; Aloia et al., 2013; Richly et al., 2011).
Mammalian PRC2 complex is composed of Enhancer of zeste homolog 1 or 2 (EZH1,
EZH2), suppressor of Zeste 12 homolog (SUZ12) and embryonic ectoderm development (EED)
in combination with the accessory subunits AEBP2 and RbAp48 (Chittock et al., 2017). EZH2 is
the only PRC2 subunit with enzymatic activity and is responsible for the catalyzation of
trimethylation of Lysine 27 on histone H3 (H3K27me3) (Margueron et al., 2008). This
modification is recognized by the CBX protein family, which leads to PRC1 recruitment. The rest
of the core PRC2 components are responsible for complex assembly and for proper enzymatic
activity (Pasini et al., 2004; Margueron et al., 2009).
EZH2 alone is catalytically inactive, requiring interaction with both EED and SUZ12 in
order to form a minimal, active PRC2 complex. EZH2 is a SET-domain protein, capable of
transferring a methyl group from S-adenosyl-L-methionine to the amino group of a lysine residue
on histones, leaving a methylated lysine residue and the cofactor byproduct S-adenosyl-L-
homocysteine. Ezh2 gene knockout results in embryonic lethality in mice, highlighting the
fundamental role of EZH2 in development, and is highly expressed in embryonic and proliferating
cells. The precise mechanisms of PRC2 recruitment to target loci in mammals remain contentious
(Chittock et al., 2017).
7

Mammalian PRC1 complex is composed of RING1 protein, Polyhomeotic homolog
(PHC), Chromobox Protein (CBX) and polycomb group ring finger (PCGF) (Morey and Helin,
2010). The PRC1 complex has a diverse subunit composition, and further information is required
in order to decipher how these different compositions determine the functional output of the PRC1
complexes by tuning their genomic targeting. It is yet to be discovered how the enzymatic and
non-enzymatic activities of the PRC1 components regulate transcription in such a highly
coordinated way (Chittock et al., 2017).
There are at least eight members of the CBX proteins in both mouse and human genomes,
each of which contains a single N-terminal chromodomain (CD) (Wotton and Merrill, 2007). All
of the CBX protein family members are involved in the regulation of heterochromatin, gene
expression and development. The chromatin recruitment of canonical PRC1 complex is mainly
mediated by CBX proteins. The C-terminal polycomb repressor box of CBX proteins is involved
in transcriptional silencing and binding to other PRC1 components such as RING1B (Müller et al.,
1995; Bezsonova et al., 2009). Adjacent to the chromodomain, all of the CBX homologs have a
DNA binding motif (Senthilkumar et al., 2009).
It has been previously shown that CBX8 possess oncogenic properties, such as induction
of epithelial–to-mesenchymal transition (EMT), invasive migration and stem cell-like traits, which
are associated with increased tumor growth and metastasis (Tang et al., 2019). In addition, the
expression of CBX7 is lost in human malignant neoplasias and its downregulation has been
previously correlated with cancer aggressiveness and poor prognosis. CBX7 is able to
differentially regulate crucial genes involved in cancer progression and EMT (Pallante et al.,
2015). Consistently, the Cbx gene family has been associated with many types of human cancers
(Ma et al., 2014).
8

As mentioned above, the histone modification mark H3K27me3, previously deposited by
the PRC2 complex, serves as a signal for CBX protein-mediated PRC1 recruitment (Figure 1)
(Kaustov et al., 2011). The chromodomain of CBX proteins is responsible for recognizing
H3K27me3 modifications on the chromatin and allows for physical interaction with this marker.
This interaction leads to the recruitment of the PRC1 components to specific regions of the
chromatin and allows for stabilization of the complex as a whole (Bernstein et al., 2006; Buchwald
et al., 2006).
Upon its recruitment, PRC1 regulate chromatin structures and gene expression in many
different ways. It is able to compact nucleosome arrays to form knot-like structures, and function
by restricting template accessibility to factors needed for gene activity. This complex also modifies
histones with its RING subunit, catalyzing histone H2A lysine 119 ubiquitination (Figure 1).
PRC1-mediated chromatin compaction is highly conserved, as many PRC1-bound genomic sites
folded in compact structures have been identified. The chromatin remodeling activities of both
PRC1 and PRC2 complexes significantly affect DNA methylation status and chromatin
accessibility in a way that ultimately leads to transcriptional repression (Boettiger et al., 2016;
Rinaldi and Benitah, 2014; Vidal and Starowicz, 2017).

Figure 1 (Rinaldi and Benitah, 2014). Schematic representation of histone modification regulation by
PRC1 and PRC2 complexes. The H3K27me3 mark is deposited by the PRC2 complex and recognized by
9

PRC1 containing a CBX protein. PRC1 is shown to catalyze the ubiquitination of H2A lysine 119 and
induce chromatin compaction, resulting in transcriptional repression.

1.5. Polycomb Repressive Complex inhibitors:
1.5.1. UNC4976 - CBX7 inhibitor:
Previous assays have shown that canonical PRC1 targeting to a reporter locus could be
achieved by a chimeric fusion of the PRC1 core subunit CBX7 with a DNA binding domain
(Moussa et al., 2019). In an effort to assess potency and permeability for new compounds targeting
PRC1-associated CBX7 in the context of native chromatin, a cellular reporter assay was utilized.
A mouse embryonic stem cell (mESC) line was generated with a single integration of an array of
12 DNA binding sites for ZFHD1 transcription factor upstream of a CpG-free Green Fluorescent
Protein (GFP) gene. This has been done in order to direct CBX7-mediated specific recruitment of
canonical PRC1 and initiation of a polycomb repressive domain (Lamb et al., 2019).
Stable expression of the DNA binding domain ZFHD1 alone and ZFHD1-meditated CBX7
tethering was compared. ZFHD1 expression alone did not cause reduction of GFP expression or
lead to repressive chromatin modifications. In contrast, ZFHD1-meditated CBX7 tethering leads
to the formation of a PRC1 complex, transcriptional repression of the GFP reporter gene and
deposition of repressive histone modification. In this experiment, transcriptional repression
depended on ZFHD1-CBX7 tethering, but endogenous repression by PRC1 and PRC2 depends on
the interaction between CBX7 chromodomain and H3K27me3. For this reason, it was assumed
that inhibitors preventing the binding between CBX7 and H3K27me3 would inhibit complex
formation in ZFHD1 binding sites and result in reactivation of the GFP reporter gene (Ibid).
Using this Polycomb in vivo assay, a set of compounds whose structures varied at the
methyl-lysine mimetic position was screened. mES cells were treated with the different
compounds and GFP levels were examined 48 hours later via flow cytometry. Though most of the
10

tested compounds showed a dose-dependent reactivation of GFP expression, The N6-methyl-N6-
norbornyl compound, UNC4976 (Figure 2), was selected for further investigation. UNC4976 was
found to be most potent in the CBX7 reporter cell line. In addition to the polycomb in vivo assay,
mESC viability was evaluated after compound treatment. UNC4976 showed weak toxicity at 100
mM, which is outside of the concentration range utilized for this compound in the following SP
analysis (Ibid).

Figure 2 (Lamb et al., 2019). Chemical structure of UNC4976. UNC4976 is a cellular potent inhibitor
that blocks the chromodomain of CBX7. It efficiently reactivates PRC1 target genes in a highly selective
manner.

1.5.2. UNC7040 - CBX8 inhibitor:
In order to find an inhibitor for CBX8 protein, a very similar polycomb in vivo assay that
was used to design an inhibitor for CBX7 was performed. mESC cell line was engineered,
expressing a CBX8 protein fused to a Tet repressor domain (TetR). This cell line contains a landing
site with a cassette that harbors an artificial DNA binding array composed of 12xZFHD1, 4xGAL4
UAS, and 7xTetO sites upstream of a PGK promoter controlling the expression of Puro resistance
and a GFP gene. Similar to the CBX7 inhibitor assay, the fusion of TetR-CBX8 leads to the
recruitment and assembly of a functional canonical PRC1 complex at a TetR-DNA binding site,
11

which ultimately results in repression of the GFP gene as measured by a decrease in GFP signal
using flow cytometry (Suh, 2019).
These results lead to the assumption that adding CBX8 antagonists into this system would
inhibit the binding and catalytic events of endogenous PRC1 and PRC2, mediated by CBX8, that
will finally de-repress the GFP reporter gene. mESC cells were then treated with different CBX8
antagonists and GFP levels were examined 48 hours later via flow cytometry. The data was then
analyzed in order to compare the strongest GFP signals between the compounds. In general, it was
found that longer and bulkier groups at the lysine mimetic position displayed increased cellular
efficacy. The inhibitor UNC7040 was found to be the most potent and specific, with no
cytotoxicity up to 100 μM (Ibid).
UNC7040 (Figure 3) is a small-molecule peptidomimetic ligand that mimics methylated
histone tails. It binds to the chromodomain of CBX8, which is responsible for the recognition of
the H3K27me3 marker on histone H3. By blocking the chromodomain of CBX8, it interferes with
its recognition of the H3K27me3 marker, making it unable to recruit the rest of the PRC1
components and leading to the expression of its target genes (Ibid).

Figure 3 (Suh, 2019). Chemical structure of UNC7040. UNC7040 is a specific and potent inhibitor that
blocks the chromodomain of CBX8.

12

1.5.3. UNC7263 - inactive control for UNC7040:
Based on the CBX8 reporter assay, the same approach was used in order to create a
negative control for CBX8 inhibition. The resulting compound, UNC7263 (Figure 4), did not show
any cellular activity in the CBX8 reporter assay and showed no cytotoxicity up to 100 μM.
UNC7263 is an analogue of the active compound UNC7040, displaying a small difference in the
lysine mimetic position. This compound, which is still expected to bind to CBX8-CD in vitro and
different from typical negative control in this regard, possesses no cellular activity and does not
interfere with the recognition of H3K27me3 modification by CBX8 (Suh, 2019).

Figure 4 (Suh, 2019). Chemical structure of UNC7263. UNC7263 is an inactive control for UNC7040,
displaying no cellular activity.

1.5.4. GSK126 - EZH2 inhibitor:
GSK126 (Figure 5) is a potent and highly selective small molecule inhibitor of EZH2
(Huang et al., 2017). It is a S-adenosyl-methionine (SAM) competitive that interferes with the
methyltransferase activity of EZH2, causing global decrease of H3K27me3 modification levels.
This inhibition of EZH2 causes the reactivation of silenced PRC2 target genes. Previous analysis
13

of GSK126 demonstrated that inhibition of H3K27me3 began before 24 h and potency was
maximal after 3 days (Zeng et al., 2017).
The docking site for GSK126 is the SAM binding pocket of EZH2, which contains the
SET-domain responsible for its methyltransferase activity. It has been shown to be more than
1,000-fold selective for EZH2 over other histone methyltransferases, including EZH1, and is
associated with inhibition of migration and metastasis. Histones H3 and the PRC2 components,
excluding EZH2, were not affected by the actions of GSK126, indicating that the reduction of
H3K27me3 modification is not due to their degradation, but due to the inhibition of the
methyltransferase activity of EZH2 (McCabe et al., 2012).
Not surprisingly, the majority of transcriptional changes with GSK126 treatment involved
upregulation. Comparison of gene expression between GSK126 treated cells and EZH2
knockdown cells showed very similar results, indicating that these transcriptional alterations are
in fact due to the loss of EZH2 methyltransferase activity. Similar results were observed in both
cell culture and mouse xenograft models (Ibid).

Figure 5. Chemical structure of GSK126 is shown above. GSK126 is a potent, highly selective EZH2
methyltransferase inhibitor, >1000-fold selective for EZH2 over 20 other human methyltransferases.

14

Hypothesis

Over recent decades, phenotypic plasticity of cancer cells has presented an ongoing
challenge to the design of cancer therapies (da Silva-Diz et al., 2018). Phenotypic switching,
allowing cancer cells to switch states between a cancer stem-like state and a well-differentiated
state, represents a significant driving force for aggressive cancer progression and treatment failure,
thereby promoting disease recurrence. Such phenotypic plasticity can promote therapeutic
resistance and provide cells with a survival advantage under varying environmental conditions
(Jolly et al., 2018).
The cancer cell plasticity model states that cancer cells can fluctuate between a stem-like
state, termed the side population, and a non-stem-like state, termed the non-side population, in a
6-days cycle in vitro (He et al., 2011). This plasticity was shown to be attributed, at least in part,
to epigenetic mechanisms, involving DNA methylation and chromatin remodeling, as the two
populations of cells exhibit differential chromatin accessibility and DNA methylation status. These
findings suggest that epigenetic plasticity plays a key role in phenotypic switching between SP
and NSP cells (Xu et al., in press, 2020).
The Polycomb Repressive Complex is a well-known chromatin remodeling complex, with
an epigenetic activity that leads to transcriptional repression. EZH2, a trimethyl “writer” enzyme
that forms a part of the PRC2 complex, is responsible for catalyzation of the H3K27me3 repressive
mark. This mark is recognized by CBX proteins, the “reader” component of the PRC1 complex,
that is responsible for the recruitment of the rest of the PRC1 components. This recruitment
followed by complex assembly leads to transcriptional silencing by chromatin remodeling
activities (Boettiger et al., 2016; Müller and Verrijzer, 2009).
15

This data led to the hypothesis that both EZH2 and CBX8 are involved in chromatin
remodeling and transcriptional repression that contributes to the phenotypic plasticity between SP
and NSP cells. In order to test this hypothesis, we sought to inhibit the methyltransferase activity
of EZH2 and interfere with the interaction between CBX8 and H3K27me3, in order to investigate
the effect of these inhibitions on the transition between SP and NSP phenotypes.

16

Materials & Methods

3.1. Cell Culture:
Human bladder cancer cell lines J82 and T24 were purchased from ATCC (Manassas, VA)
and routinely used in our lab. Cell lines were authenticated prior to starting the experiments.
Human J82 bladder cancer cell line was maintained at 37 ℃, 5% CO2 in Dulbecco’s Modified
Eagle’s Medium (DMEM; Corning). DMEM was supplemented with 10% fetal bovine serum
(FBS; Omega), 1% penicillin (100 units/mL; Invitrogen), and 1% streptomycin (100 µg/mL;
Invitrogen). Human T24 bladder cancer cell line was maintained at 37 ℃, 5% CO2 in Roswell Park
Memorial Institute (RPMI; Corning) 1640 medium supplemented with 10% FBS, 1% penicillin
(100 units/mL) and 1% streptomycin (100 µg/mL).

3.2. Flow cytometry for side population analysis:
Adherent J82 cancer cells were trypsinized, counted, centrifuged at 1,000 rpm for 5
minutes and resuspended in prewarmed DMEM. Hoechst 33342 (Sigma-Aldrich) was added at a
concentration of 5 µg/mL, and cells were incubated for 2 hours at 37°C, gently inverted every 15
minutes during the course of incubation. Parallel sample aliquots were prepared in the presence of
50 µmol/L Verapamil (Sigma-Aldrich), an ATP-binding cassette transporter family inhibitor that
prevents Hoechst efflux, at room temperature for 10 minutes before adding the Hoechst 33342
dye. Cells were centrifuged at 1,000 rpm for 5 minutes after incubation and resuspended in
DMEM. Samples were incubated for at least 5 minutes on ice before FACS analysis (FACSAria
& FACSLSR-II; BD Biosciences; both equipped with UV lasers).

17

3.3. Inhibitor treatments:
Prior to SP analysis performed on day 6, J82 cells were seeded at day 0 at a concentration
of 50,000 cells/mL and kept on drug treatment for 6 days. For confluence-dependent SP analysis,
where analysis was performed at day 1, cells were seeded at a concentration of 800,000 cells/mL
and kept on drug treatment for 24 hours.
GSK126 was used at a concentration of 4 μM, UNC7040 was used at a concentration of
60 μM, UNC7263 was used at a concentration of 60 μM and UNC4976 was used at a concentration
of 20 μM. Inhibitors were received courtesy of the Bell laboratory at the University of Southern
California. Treatment with GSK126 and UNC4976 was compared to a vehicle control containing
cells with DMSO treatment.

3.4. Western Blot:
J82 cells were seeded in 10cm plates and harvested 72 hours later. For nuclear extraction
cells were washed with PBS and resuspended in buffer A (25mM Hepes, pH 7.6; 5mM MgCl2;
25mM KCl; 0.05mM EDTA; 10% Glycerol; 0.1% NP40). Cells were then incubated for 10
minutes and centrifuged at 1500 rpm for 5 minutes at 4°C. Pellet was washed with buffer A and
centrifuged at 1500 rpm for 5 minutes at 4°C, then resuspended in RIPA buffer (150mM NaCl;
5mM EDTA, pH 8.0; 50mM Tris, pH 8.0; 1% NP-40; 0.5% Sodium Deoxycholate; 0.1% SDS).
Lysate was transferred to Bioruptor pico micro tubes and sonicated with Bioruptor Pico (3-4 cycles
with 30sec ON/OFF intervals). Sonicated lysate was transferred to regular 1.5 ml safe lock tubes
and centrifuged at full speed for 10 minutes at 4°C. Supernatant was transferred to new tubes.
Mastermix of 4x Sample buffer [4x Laemli buffer Biorad, 10 mM DTT (1:100 1M stock), 0.5%
BME (1:200 of 100% stock)] was made, and 6 µL of this mastermix was added to each sample.
18

24 µL of RIPA lysis buffer was added and samples were boiled at 95°C for 5 minutes. Samples
were centrifuges and kept at room temperature until loading.
Gel chamber was assembled and filled with running buffer (Invitrogen). 10 µL Prestained
PAGERuler Plus (Thermoscientific) was used as Marker. Gel Electrophoresis was used for protein
separation at 120V for 1 hour. Immobilon-FL PVDF membrane was activated in 100% methanol
and blot at 300 mA for 2 hours. Membrane was shortly washed with water and 1xPBS-0,1% Tween
20, and subsequently blocked for 30 minutes with 3% nonfat dry milk powder in 1XPBS. Primary
antibodies were added (a-H3K27me3 Diagenode p069-050 1:1000 rabbit o/n; a-Lamin B1 Abcam
(ab16048) 1:20000 rabbit o/n), incubated overnight at -4°C, and later washed 3x for 5-10 minutes
with 1x PBS-0,1 % Tween 20. Secondary antibodies were then added (Goat anti-Rabbit IgG (H+L)
Secondary Antibody, DyLight 800 4X PEG Catalog # SA5-35571 & 35518 680nm for anti-mouse
1:10000), incubated for 1 hour, and subsequently washed 3x for 5-10 minutes with 1x PBS-0,1 %
Tween 20. Fluorescence based detection on the ODYSSEY CLx system was used for imaging.

3.5. Lentiviral Infection:
Lentivirus containing modified Ezh2 gene and Blasticidin resistance gene were received
courtesy of the Bell laboratory at the University of Southern California. Lentivirus were generated
in order to deliver a hyperactive EZH2 or a control catalytically dead EZH2. J82 cells were seeded
in 10 cm plates, infected with lentivirus and treated with Blasticidin until only Blasticidin-resistant
cells remained in culture.

19

3.6. RNA extraction and cDNA synthesis:
RNA extraction was performed using the Direct-zol™ RNA MiniPrep Plus kit (Zymo
Research). Cells were trypsinized, centrifuged at 1,000 rpm for 5 minutes and washed with
Phosphate-Buffered Saline (PBS; Corning). Cells were then collected into 1.7 mL tubes,
centrifuged and resuspended in 400 μL Trizol. 400 μL of 100% ethanol was added to the solution
and mixed well by pipetting. 700 μL of the mixed solution was transferred into an Easy-Spin
column and centrifuged at 12,000 rpm for 30 seconds. The flow through was discarded and 400
μL of RNA pre-wash buffer were added to the column and centrifuged again at 12,000 rpm for 30
seconds. Flow was discarded and this step was repeated. Then, 400 μL of RNA wash buffer was
added to the column and centrifuged for 2 minutes at 12,000 rpm. Flow through was discarded and
the column was transferred into a 1.7 mL RNase-free collection tube and 100 μL of DNase/RNase-
free water were added directly to the center of the spin column membrane for RNA elution. Tubes
were centrifuged at 12,000 rpm for 30 seconds. The RNA was obtained in 100 μl volume.
For cDNA synthesis, 16 μl of the RNA extracted above was added into the reaction tube.
4 μL of the cDNA SuperMix (QuantaBio) was then added to the mixture. The reaction conditions
for cDNA synthesis were as follows: 25°C for 5 minutes (1 cycle), 42°C for 30 minutes (1 cycle),
85°C for 5 minutes (1 cycle) and 4°C for ∞. After the synthesis, the cDNA was stored at -20°C.

3.7. Quantitative Real-Time PCR:
cDNA was used for real-time PCR amplification with specific gene primers and Quanta B-
R Sybr Green qPCR supermix (QuantaBio) using Bio-Rad MyiQ single color Real-Time PCR
Detection System (Bio-Rad) and Bio-Rad iQ5 (Bio-Rad). Primer sequences can be found below.

20

3.7.1. List of Primers (IDT):
Gene of Interest Forward Primer Reverse Primer
EZH2 5'- AGAAGGGACCGGTTTGTTGG -3' 5'- ACATTATAGGCACCGAGGCG -3'

3.8. Statistical analysis:
Experiments were conducted in biological triplicates and represented as means. Error bars
represent the standard deviation derived from biological replicates.

21

Results

4.1. Targeting EZH2 with GSK126 lead to a reduction in H3K27me3 marker:
GSK126 is a highly specific and potent inhibitor of EZH2 (McCabe et al., 2012), the
methyltransferase component of PRC2 complex that is responsible for deposition of the
H3K27me3 marker (Margueron et al., 2008). Inhibition of EZH2 protein with GSK126 leads to a
global decrease in H3K27me3 marker on chromatin. In order to test the efficiency of this inhibitor,
we seeded J82 cells in cell culture and treated them with GSK126 for 72 hours. Nuclear extraction
and immunoblotting with anti-H3K27me3 were subsequently performed.
GSK126 treatment resulted in a significant reduction of H3K27me3 levels. These results
suggest that EZH2 inhibition with 72-hours GSK126 treatment at a concentration of 4µM was
sufficient for a significant reduction of global H3K27me3 levels in J82 cells (Figure 6).

Figure 6. Immunoblotting for analysis of H3K27me3. Cells were treated with 4 μM GSK126 for 72 hours
prior to analysis. Lamin-B was used as loading control. J82 cells were seeded at a concentration of 300,000
cells/mL and harvested 3 days after treatment. Nuclear extracts were resolved by SDS page and
immunoblotted with anti-H3K27me3. These results show the efficiency of GSK126 inhibition of EZH2
that results in a significant global reduction in H3K27me3 marker.

4.2. Effects of EZH2 inhibition on side population size:
Having previously observed that NSP cells were capable of reconstituting the SP
subpopulation by conversion, we sought to investigate the potential involvement of the Polycomb
group proteins in this process. To accomplish this, we seeded J82 cells in cell culture and treated
22

them with GSK126, a pharmacological inhibitor of EZH2. Cells were counted at days 3 and 6 after
GSK126 treatment, and no significant effect on total cell number was observed (data not shown).
Side population analysis was performed 6 days following GSK126 treatment in order to measure
the size of the side population. Untreated J82 cells showed a high percentage of SP, as expected at
day 6, with cells treated with GSK126 initially showing a significant decrease in SP cells relative
to both untreated and vehicle control (Figure 7A).
While initial results showed a significant decrease in SP subpopulation size following
GSK126 treatment, subsequent experiments did not reproduce the same results. Although a second
set of experiments showed that GSK126 treated cells exhibit a significant increase in SP size
(Figure 7B), Neither the increase nor the decrease in SP size after GSK126 treatment was
reproducible in subsequent experiments. GSK126 treatment did not continue to show an effect on
the SP size, as no significant change relative to both untreated and vehicle controls was observed
(Figure 7C).
Inhibition of EZH2 with GSK126 showed an inconsistent effect on SP size, which may be
attributed to the length of the experiments. Side population analysis is an effective tool for
monitoring the SP subpopulation, but one major issue is that this protocol requires tracking the
side population long after drug treatment. This method was time consuming, as the entire protocol
required 6 days, which may lead to the observed variability in these results. After 6 days of culture,
as cells become overly confluent, it is generally believed that the signaling networks are altered
and that there may be a degree of variability in cell numbers that can lead to aberrant and
irreproducible results.

23

Figure 7. SP subpopulation size in J82 cells 6 days after GSK126 treatment. Verapamil was used as a
negative control. A total of 100,000 cells were seeded in triplicates into 6-well plates at day 0 and monitors
for SP size at day 6. GSK126 did not produce a consistent effect on SP size. (A) GSK126 treatment shows
a significant reduction in SP subpopulation size relative to both untreated J82 cells and vehicle controls.
(B) GSK126 treatment increased SP size relative to controls. (C) GSK126 treatment does not show a
significant change in SP subpopulation size relative to controls.

4.3. Effects of CBX7 inhibition on side population size:
J82 cells were seeded in cell culture and treated with UNC4976, an inhibitor of CBX7, in
order to determine the potential involvement of CBX7 in the plasticity between SP and NSP cells.
Cells were counted at days 3 and 6 after UNC4976 treatment, and no significant effect on total cell
number was observed (data not shown). Side population analysis was performed 6 days after
UNC4976 treatment in order to monitor the size of the side population. Cells treated with
UNC4976 showed no significant shift in SP size, relative to untreated and vehicle controls (Figure
8). These results cannot confirm our initial hypothesis, stating that the interaction between CBX7
and H3K27me3 leads to the reconstitution of the side population via conversion of NSP cells.
A B C
24

Figure 8. SP subpopulation size in J82 cells 6 days after UNC4976 treatment. Verapamil was used as a
negative control. A total of 100,000 cells were seeded in triplicates into 6-well plates at day 0 and monitors
for SP size at day 6. Inhibition of CBX7 with UNC4976 does not show a significant change in SP
subpopulation size relative to both untreated J82 cells and vehicle controls.

4.4. Effects of CBX8 inhibition on side population size:
J82 cells were seeded in cell culture and treated with UNC7040, an inhibitor of CBX8, in
order to investigate the potential involvement of CBX8 in the reconstitution of the SP
subpopulation by conversion of NSP cells. Cells were counted at days 3 and 6 after UNC7040
treatment, and no significant effect on total cell number was observed (data not shown). Side
population analysis was performed 6 days after UNC7040 treatment in order to monitor the size
of the side population. While untreated J82 cells showed a roughly high percentage of SP, as
expected at day 6, cells treated with UNC7040 initially showed a significant decrease in SP cells
(Figure 9A). Similar to the previous results showing the inconsistent effect of GSK126, subsequent
experiments observing the effect of UNC7040 treatment showed no significant effect on SP size
(Figure 9B). As suggested earlier, this experimental variability may also be attributed to the length
25

of the experiments, once again requiring a 6-days protocol. Shortening the time between drug
treatment and SP analysis could be a potential improvement for a more efficient protocol.

Figure 9. SP subpopulation size in J82 cells 6 days after UNC7040 treatment. Verapamil was used as a
negative control. A total of 100,000 cells were seeded in triplicates into 6-well plates at day 0 and monitors
for SP size at day 6. UNC7040 did not produce a consistent effect on SP size. (A) UNC7040 treatment
shows a significant reduction in SP subpopulation size relative to untreated J82 cells. (B) UNC7040
treatment does not show a significant change in SP subpopulation size relative to untreated J82 cells.

4.5. Confluence-dependent plasticity model:
We showed that both GSK126 and UNC7040 treatments did not show a consistent effect
of reducing or increasing the size of the side population. Hence, we decided to produce a more
efficient protocol, that will allow the performance of a significantly shorter analysis. So far, side
population analysis required a 6 days period from the moment of cell culture to the achievement
of SP reconstitution by NSP cell conversion. In this protocol, cells were seeded in cell culture at a
concentration of 50,000 cells/mL, which is equivalent to 10% confluence, and by the time the side
population is reconstituted at day 6, the cells are between 80%-100% confluence. We decided to
seed our cells at different levels of confluence, in order to examine whether high confluence is a
potential explanation for the high percentage of SP cells at day 6. We seeded J82 cells in cell
culture at increasing confluence, from 10% to 100%, and monitored the size of the side population
the following day.
A B
26

As expected, cells that were seeded at a higher level of confluence showed a higher number
of SP cells, relative to cells seeded at a lower confluence level (Figure 10). These results suggest
that it is the cells’ density (possibly proximity and interaction with one another) that can potentially
lead to a massive conversion of NSP cells into the SP phenotype. Similar results were observed
with T24 cells, a different human bladder cancer cell line (Figure 11). This new and rapid
confluence-dependent plasticity model allowed for an efficient and rapid set of experiments, that
may correct for the previous variability observed with both GSK126 and UNC7040 treatment.

Figure 10. Confluence-dependent model with J82 cells. Verapamil was used as a negative control. J82
cells were seeded in triplicates at a concentration range of 100,000-1,000,000 cells/mL (equivalent to 10%-
100% confluence). The SP subpopulation size increased as the level of cell confluence increased.

27

Figure 11. Confluence-dependent model with T24 cells. Verapamil was used as a negative control. T24
cells were seeded in triplicates at a concentration range of 100,000-1,000,000 cells/mL (equivalent to 10%-
100% confluence). The SP subpopulation size increased as the level of cell confluence increased.

4.6. Effect of CBX8 inhibition on SP subpopulation size at 80% cell confluence:
Having observed that a large side population size can be achieved 24 hours after seeding
cells at a high percentage of confluence, we sought to repeat UNC7040 treatment in order to
evaluate the effect of CBX8 inhibition on the SP size within 1 day of treatment. To accomplish
this, we seeded J82 cells in cell culture at a concentration of 800,000 cells/mL (equivalent to 80%
confluence) and treated them with UNC7040.
In order to better control this protocol, a new component was added in these sets of
experiments. UNC7263, as described earlier, is an analogue of the active compound UNC7040,
displaying a small difference in the lysine mimetic position. This compound does not interfere
with the recognition of the H3K27me3 modification by CBX8 (Suh, 2019).
After treating the cells with either UN7040 or its control UNC7263, we subsequently
performed side population analysis 24 hours later. Cells were counted and no significant effect on
28

total cell number was observed (data not shown). While the SP analysis showed that J82 cells
treated with UNC7040 displayed a significant decrease in SP subpopulation size, cells treated with
UNC7263 also displayed a decrease in SP (Figure 12). Repeated experiments confirmed this
finding (data not shown).

Figure 12. SP subpopulation size in J82 cells 1 day after UNC7040 treatment. Verapamil was used as a
negative control. A total of 800,000 cells were seeded in triplicates into 6-well plates at day 0 and monitors
for SP size at day 1. Inhibition of CBX8 with UNC7040 shows a significant reduction in SP subpopulation
size relative to untreated J82 cells. UNC7263, a cellular control compound analogue to UNC7040, also
show a significant reduction in SP size, relative to untreated control.

Previous studies showed that UNC7263 has no cellular activity and should not interfere
with the recognition of H3K27me3 marker by CBX8. These results suggest that UNC7263 may
have an off-target effect that is currently unknown and further investigation is required in order to
determine the effect of this compound on SP emergence. Since both UNC7040 and UNC7263
exhibit a similar effect on the side population, this reduction cannot be attributed to blocking the
interaction of CBX8 with H3K27me3.

29

4.7. Effect of EZH2 inhibition on SP subpopulation size at 80% cell confluence:
In order to evaluate the effect of EZH2 inhibition on the SP size within 1 day of GSK126
treatment, we seeded J82 cells in cell culture at a concentration of 800,000 cells/mL (equivalent to
80% confluence) and treated them with GSK126. We subsequently performed side population
analysis 24 hours after treatment.
Initial results showed that GSK126 treatment leads to a significant reduction in SP size
when compared with both untreated control and vehicle control (Figure 13A). Nonetheless,
subsequent experiments did not reproduce the same effect, as no shifts in side population size were
observed (Figure 13B). These results show that despite the reduction of a 6-days protocol to a 1-
day protocol, the effect of GSK126 on the side population size remained inconsistent.

Figure 13. SP subpopulation size in J82 cells 1 day after GSK126 treatment. Verapamil was used as a
negative control. A total of 800,000 cells were seeded in triplicates into 6-well plates at day 0 and monitors
for SP size at day 1. GSK126 did not show a consistent effect of SP size. (A) GSK126 treatment shows a
significant reduction in SP size relative to controls. (B) GSK126 treatment does not show a significant
change in SP subpopulation size relative to controls.

4.8. Six-day GSK126 treatment does not lead to a complete knock out of H3K27me3:
As part of our effort to investigate the reason behind the inconsistent results observed with
GSK126, we suggested that it is the insufficient dilution of the H3K27me3 modification that could
be a potential cause. We showed a partial dilution of H3K27me3 after 3 days (Figure 6), and next
we wanted to find out whether the dilution would be complete within 6 days, as required by our
A B
30

first protocol. To test this, we seeded J82 cells in cell culture and treated them with GSK126 for 6
days. Nuclear extraction and immunoblotting with anti-H3K27me3 were subsequently performed.
While 6 days of GSK126 treatment resulted in a significant reduction of H3K27me3 levels, the
modification was not completely diluted (Figure 14). Moreover, these results did not seem
significantly different from the results observed after 3 days of treatment. This suggests that the
currently used drug concentration may not be sufficient for a complete knock out of the
modification in question.

Figure 14. Immunoblotting for analysis of H3K27me3. Cells were treated with 4 μM GSK126 for 6 days
prior to analysis. Lamin-B was used as loading control. J82 cells were seeded at a concentration of 50,000
cells/mL and harvested 6 days after treatment. Nuclear extracts were resolved by SDS page and
immunoblotted with anti-H3K27me3. These results show the efficiency of GSK126 inhibition of EZH2
that results in a global reduction in H3K27me3 marker.

4.9. Effects of EZH2 overexpression on SP subpopulation size:
In our continuous investigation of the effect of EZH2 on the plasticity between SP and NSP
cells, we sought to overexpress EZH2 in order to evaluate the potential effect on the emergence of
the side population. To do so, we seeded J82 cells in cell culture at a concentration of 100,000
cells/mL and infected them with a lentivirus expressing either a hyperactive EZH2 protein, or a
catalytically dead EZH2 protein. Quantitative Real-Time PCR was successfully performed in order
to confirm the overexpression of EZH2, showing >1000-fold increase in EZH2 expression relative
to uninfected control (data not shown).
31

Infected cells were seeded and side population analysis was performed 6 days later.
Surprisingly, while cells expressing a hyperactive EZH2 protein showed a significant reduction in
SP size, cells expressing the inactive EZH2 protein showed a greater reduction in SP cells (Figure
15). Subsequent experiments consistently showed that cells overexpressing an inactive EZH2
protein led to a significant reduction in SP size (data not shown), but the mechanism by which it
is done is currently unknown. Nonetheless, any shift in side population size cannot be attributed
to the methyltransferase activity of EZH2, as a catalytically inactive EZH2 show similar results.

Figure 15. SP subpopulation size in J82 cells 6 days after lentivirus infection for EZH2 overexpression.
Verapamil was used as a negative control. A total of 100,000 cells were seeded in triplicates into 6-well
plates at day 0 and monitors for SP size at day 6. Cells expressing a hyperactive EZH2 proteins showed a
significant reduction in SP size. the cells expressing an inactive EZH2 protein show an even greater
reduction in SP cells.

32

Discussion

Intra-tumoral heterogeneity represents an ongoing challenge to the design of cancer
therapies, as it drives cancer progression and treatment resistance. Phenotypic plasticity, the
development of multiple phenotypes from a single genotype, can lead to mutation-independent
heterogeneity that facilitates therapeutic resistance, providing cells with a survival advantage
under varying conditions (Kelly et al., 2007; Tsuchida et al., 2008; Sharma et al., 2010; Shmelkov
et al., 2008; He et al., 2011; Jolly et al., 2018). Phenotypic switching is regulated at least in part
by several signaling pathways, such as the PI3K/AKT pathway or Notch signaling, that ultimately
involve transcriptional activation (Sharma et al., 2010; Liau et al., 2017; He et al., 2014). These
observations can be explained by the theory of “bet-hedging”, which provides a reason for the
degree of phenotypic diversity observed even in clonal populations. Bet-hedging – maintaining
two or more phenotypes – can be convenient in expending populations, optimizing fitness in
varying and unpredictable environments (Jolly et al., 2018; Villa et al., 2018).
The rapidity and reversibility of acquired resistance support an epigenetic rather than a
genetic tolerance mechanism. Epigenetic factors such as DNA methylation, accessibility,
nucleosome occupancy and histone modifications can drive phenotypic diversity, as demonstrated
in several cancer cell lines (Roesch et al., 2013; Knoechel et al., 2014; Xu et al., in press, 2020),
and represent potential mechanisms for the SP/NSP cyclical state transitions previously described
by our group (He et al., 2011). Recent studies associate altered expression of epigenetic modifiers,
such as the KDM (histone lysine demethylase) family of proteins, with induction of chromatin
modifications that regulate the two-way transition between stem-like and non-stem-like
phenotypes (Feinberg et al., 2016; Sharma et al., 2010; Liau et al., 2017).
33

In the present study, we explored whether the Polycomb Repressive Complex, subdivided
into the PRC1 and PRC2 complexes, plays a role in SP/NSP plasticity. This epigenetic modifier
has been associated with several malignancies (Morin et al., 2010; Ma et al., 2014), and is involved
in chromatin compaction, histone modifications and transcriptional repression (Boettiger et al.,
2016; Rinaldi and Benitah, 2014; Vidal and Starowicz, 2017). EZH2, the writer component of
PRC2, has a methyltransferase activity resulting in the deposition of the H3K27me3 modification,
subsequently recognized by CBX proteins, the reader component of PRC1. CBX can physically
interact with H3K27me3 with its chromodomain, resulting in the recruitment of the rest of the
PRC1 components. This cascade results in epigenetic silencing of genes (Müller and Verrijzer,
2009).
The epigenetic role of the Polycomb Repressive Complex set the foundation for our
hypothesis, suggesting that EZH2 and the CBX protein family are potentially involved in the
epigenetic plasticity previously observed between the SP and NSP subpopulations (He et al., 2011;
Xu et al., in press, 2020). To test this hypothesis, we sought to interfere with the activities of both
EZH2 and CBX and observe the effect of this inhibition on the emergence of the SP subpopulation.
To do so, we used three pharmacological inhibitors: GSK126, UNC7040 and UNC4976.
GSK126 is a highly selective inhibitor of EZH2, responsible for blocking its
methyltransferase activity (McCabe et al., 2012). When used to determine the effect on SP size 6
days following treatment, we were not able to produce a consistent effect, as repeated SP analysis
resulted in a varying SP subpopulation size. Similarly, CBX8 inhibition with UNC7040 also did
not produce a consistent effect 6 days after treatment. While initial results showed a significant
decrease in SP, subsequent experiments could not reproduce the same effect, and did not lead to
any shifts in SP size. Both GSK126 and UNC7040 exhibit a great variability in their effect on the
34

SP subpopulation, and as previously mentioned, this effect was assumed to be attributed to the
length of the experiments, requiring continuous drug treatment. Long 6-day experiments
unavoidably produced a certain degree of variability in the cell numbers, which in turn may have
an effect on SP size and lead to inconsistent effect of the inhibitors.
In an attempt to create a significantly more efficient set of experiments, we discovered a
way to improve our model by converting it into a confluence-based model. Traditionally, we
seeded our J82 cells at a low concentration, and waited 6 days until the cells reached confluence,
which is the point where we observe the highest number of SP cells. When seeding J82 cells at a
confluence level parallel to day 6, we discovered the following day that a similar size of SP is
produced. This experiment allowed the shift from a 6-days protocol to a 1-day protocol, composed
of the new confluence-dependent plasticity model.
Following the formation of a rapid plasticity model, requiring 1 day of drug treatment, we
investigated whether the previous variability can be fixed with a shorter set of experiments and
showed that despite the significant shortening of our protocol, GSK126 treatment continued to
show an inconsistent effect on SP size. The reason for this experimental variability remains
unclear, but one potential explanation could be an insufficient dilution of the H3K27me3 marker
after 1 day of GSK126 treatment. We showed by Western Blot analysis that H3K27me3 levels
were significantly reduced after 3 days of GSK126 treatment, but it might not be the case after 1
day of treatment. Dilution of the H3K27me3 marker require multiple cell divisions, as new DNA
is synthesized and EZH2 inhibition by GSK126 prevents it from depositing new modifications.
We showed that even after 6 days of treatment we did not achieve a complete knock out of
H3K27me3, suggesting that the concentration of drug and the length of treatment should be further
optimized. Pre-treatment of our cells may be necessary for a complete dilution of H3K27me3 and
35

could potentially explain the variability in these results, but since SP cells are able to rapidly
emerge within 24 hours, the slow dilution of the H3K27me3 marker might not be able to conform
to our model.
In comparison with the slow effect of GSK126, UNC7040 previously showed to have an
immediate inhibitory effect on CBX8 (Suh, 2019). Experiments investigating the effect of 1-day
inhibition of CBX8 resulted in a reduction in SP size, yet similar results were observed with the
inactive control compound UNC7263. These results do not allow us to demonstrate convincingly
that the reduction in SP size is attributed to the reader activity of CBX8 and to downstream
chromatin remodeling and suggest that both UNC7040 and UNC7263 may have off-target effects
that are currently unknown. Additional biochemical analysis of these compounds would be
necessary to determine these effects.
The over-expression of EZH2 was also used in order to test its effect on the SP size. While
cells over-expressing a hyperactive EZH2 showed a significant reduction in SP size, cells
expressing an inactive EZH2 protein showed a larger reduction in SP. The mechanism by which
this SP reduction occurs is currently unknown, but the methyltransferase activity of EZH2 may
not be correlated, due to a similar effect by an inactive EZH2 protein. Because over-expression of
a catalytically inactive version of this protein results in reduction in the SP phenotypes, the effect
cannot be due to increased activity of the protein in question, but instead must arise from an
alternative mechanism. Additional analysis of the potential off-target effects of these modified
EZH2 proteins is required to determine their activity. For example, biochemical analysis of binding
partners, investigation of any known biochemical activities or cellular localization under normal
and overexpressed conditions are all potential mechanisms to provide insights into interpreting the
reduction in SP phenotype observed in our cells.
36

Further investigation of the effects of GSK126, UNC7040, and UNC7263, as well as the
catalytically inactive EZH2 protein is required in order to determine the role of EZH2 and CBX8
in cancer cell phenotypic plasticity. Our current findings cannot confirm the involvement of neither
the methyltransferase activity of EZH2, nor the CBX-mediated recruitment of PRC1 in the
emergence of the side population. Although our initial hypothesis could not be confirmed by the
assays performed in this study, the many roles of EZH2 and CBX8 in chromatin remodeling and
gene silencing can be further tested in the context of cellular plasticity.

37

References
1. Aloia L, Di Stefano B, Di Croce L. Polycomb complexes in stem cells and embryonic
development. Development 2013; 140(12):2525-34.
2. Bernstein E, Duncan EM, Masui O, Gil J, Heard E, Allis CD. Mouse polycomb proteins bind
differentially to methylated histone H3 and RNA and are enriched in facultative
heterochromatin. Mol Cell Biol 2006; 26(7):2560–2569.
3. Bezsonova I, Walker JR, Bacik JP, Duan S, Dhe-Paganon S, Arrowsmith CH. Ring1B
contains a ubiquitin-like docking module for interaction with Cbx proteins. Biochemistry 2009;
48(44):10542-8.
4. Boettiger A.N, Bintu B, Moffitt J.R, et al. Super-resolution imaging reveals distinct chromatin
folding for different epigenetic states. Nature 2016; 529:418-422.
5. Buchwald G, van der Stoop P, Weichenrieder O, Perrakis A, van Lohuizen M, Sixma TK.
Structure and E3 ‐ligase activity of the Ring–Ring complex of Polycomb proteins Bmi1 and
Ring1b. The EMBO journal 2006; 25(11):2465-74.
6. Cabrera M.C, Hollingsworth Robert E, Hurt Elaine M. Cancer stem cell plasticity and tumor
hierarchy. World J Stem Cells 2015; 7(1):27–36.
7. Chaffer CL, Brueckmann I, Scheel C, Kaestli AJ, Wiggins PA, Rodrigues LO, et al. Normal
and neoplastic nonstem cells can spontaneously convert to a stem-like state. Proc Natl Acad Sci
U S A 2011; 108:7950–5.
8. Chaffer CL, Marjanovic ND, Lee T, Bell G, Kleer CG, Reinhardt F, D'Alessio AC, Young
RA, Weinberg RA. Poised chromatin at the ZEB1 promoter enables breast cancer cell plasticity
and enhances tumorigenicity. Cell 2013; 154(1):61-74.
38

9. Chittock E.C, Latwiel S, Miller Thomas C.R, Müller C.W. Molecular architecture of
polycomb repressive complexes. Biochemical Society Transactions 2017; 45:193–205.
10. da Silva-Diz V, Lorenzo-Sanz L, Bernat-Peguera A, Lopez-Cerda M, Muñoz P. Cancer cell
plasticity: Impact on tumor progression and therapy response. Seminars in Cancer Biology 2018;
53:48-58.
11. Endoh, M., Endo, T.A., Endoh, T., Isono, K.-i., Sharif, J., Ohara, O. et al. Histone H2A
mono-ubiquitination is a crucial step to mediate PRC1-dependent repression of developmental
genes to maintain ES cell identity. PLoS Genetics 2012; 8(7):e1002774.
12. Feinberg AP, Koldobskiy MA, Gondor A. Epigenetic modulators, modifiers and mediators in
cancer aetiology and progression. Nat Rev Genet 2016; 17: 284-99.
13. He K, Xu T, Goldkorn A. Cancer cells cyclically lose and regain drug-resistant highly
tumorigenic features characteristic of a cancer stem-like phenotype. Mol Cancer Ther 2011;
10:938-48.
14. He K, Xu T, Xu Y, Ring A, Kahn M, Goldkorn A. Cancer cells acquire a drug resistant,
highly tumorigenic, cancer stem-like phenotype through modulation of the PI3K/Akt/β-
catenin/CBP pathway. International Journal of Cancer 2014; 134(1):43-54
15. Huang T, Lin C, Zhong LL, et al. Targeting histone methylation for colorectal cancer.
Therap Adv Gastroenterol 2017; 10(1):114–131.
16. Jolly MK, Kulkarni P, Weninger K, Orban J, Levine H. Phenotypic plasticity, bet-hedging,
and androgen independence in prostate cancer: Role of non-genetic heterogeneity. Frontiers in
oncology 2018; 8:50.
39

17. Kaustov L, Ouyang H, Amaya M, Lemak A, Nady N, Duan S, Wasney GA, Li Z, Vedadi M,
Schapira M, Min J, Arrowsmith CH. Recognition and Specificity Determinants of the Human
Cbx Chromodomains. Journal of Biological Chemistry 2011; 286(1): 521–529.
18. Kelly PN, Dakic A, Adams JM, Nutt SL, Strasser A. Tumor growth need not be driven by
rare cancer stem cells. Science 2007; 317(5836):337.
19. Knoechel B, Roderick J.E, Williamson K.E, Zhu J, Lohr J.G, Cotton M.J, Gillespie S.M,
Fernandez D, Ku M, Wang H, Piccioni F, Silver S.J, Jain M, Pearson D, et al. An epigenetic
mechanism of resistance to targeted therapy in T-cell acute lymphoblastic leukemia. Nat Genet
2014; 46(4):364–370.
20. Lamb Kelsey N, Bsteh Daniel, Dishman Sarah N, Moussa Hagar F, Fan Huitao, Stuckey
Jacob I, Norris Jacqueline L, Cholensky Stephanie H, Li Dongxu, Wang Jingkui, Sagum Cari,
Stanton Benjamin Z, Bedford Mark T, Kenakin Terry P, Kireev Dmitri B, Wang Gang Greg,
James Lindsey I, Bell Oliver, Frye Stephen. Discovery and Characterization of a Cellularly
Potent Positive Allosteric Modulator of the Polycomb Repressive Complex 1 Chromodomain,
CBX7. Cell Chemical Biology 2019; 26:1-15
21. Lewis EB. A gene complex controlling segmentation in Drosophila. Nature 1978;
276(5688):565-70.
22. Liau BB, Sievers C, Donohue LK, Gillespie SM, Flavahan WA, Miller TE, Venteicher AS,
Hebert CH, Carey CD, Rodig SJ, Shareef SJ, Najm FJ, et al. Adaptive Chromatin Remodeling
Drives Glioblastoma Stem Cell Plasticity and Drug Tolerance. Cell Stem Cell 2017; 20:233-46.
23. Lu J, Cui Y, Zhu J, He J, Zhou G, Yue Z. Biological characteristics of Rh123high stem-like
cells in a side population of 786-O renal carcinoma cells. Oncol Lett 2013; 5(6):1903–1908.
40

24. Ma RG, Zhang Y, Sun TT, Cheng B. Epigenetic regulation by polycomb group complexes:
focus on roles of CBX proteins. J Zhejiang Univ Sci B 2014; 15(5):412–428.
25. Margueron R, Justin N, Ohno K, Sharpe ML, Son J, Drury WJ 3rd, Voigt P, Martin SR,
Taylor WR, De Marco V, Pirrotta V, Reinberg D, Gamblin SJ. Role of the polycomb protein
EED in the propagation of repressive histone marks. Nature 2009; 461(7265):762-7.
26. Margueron R, Li G, Sarma K, Blais A, Zavadil J, Woodcock CL, Dynlacht BD, Reinberg D.
Ezh1 and Ezh2 maintain repressive chromatin through different mechanisms. Mol Cell 2008;
32(4):503-18.
27. McCabe M.T, Ott H.M, Ganji G, et al. EZH2 inhibition as a therapeutic strategy for
lymphoma with EZH2-activating mutations. Nature 2012; 492(7427):108-112.
28. Morey L, Helin K. Polycomb group protein-mediated repression of transcription. Trends in
Biochemical Sciences 2010; 35(6):323-32.
29. Morin RD, Johnson NA, Severson TM, et al. Somatic mutations altering EZH2 (Tyr641) in
follicular and diffuse large B-cell lymphomas of germinal-center origin. Nat Genet 2010;
42(2):181–185.
30. Moussa H.F., Bsteh D., Yelagandula R., Pribitzer C., Stecher K., Bartalska K., Michetti L.,
Wang J., Zepeda-Martinez J.A., Elling U., et al. Canonical PRC1 controls sequence-independent
propagation of Polycomb-mediated gene silencing. Nature Communications 2019; 10:1931.
31. Müller J, Gaunt S, Lawrence PA. Function of the Polycomb protein is conserved in mice and
flies. Development 1995; 121(9):2847-52.
32. Müller J, Verrijzer P. Biochemical mechanisms of gene regulation by polycomb group
protein complexes. Current Opinions in Genetics & Development 2009; 19(2):150-158.
41

33. Pallante P, Forzati F, Federico A, Arra C, Fusco A. Polycomb protein family member CBX7
plays a critical role in cancer progression. Am J Cancer Res 2015; 5(5):1594–1601.
34. Pandiyan K, You J.S, Yang X, Dai C, Zhou X.J, Baylin S.B, Jones P.A, Liang G. Functional
DNA demethylation is accompanied by chromatin accessibility. Nucleic Acids Research 2013;
41(7):3973–3985.
35. Pasini D, Bracken AP, Jensen MR, Lazzerini Denchi E, Helin K. Suz12 is essential for
mouse development and for EZH2 histone methyltransferase activity. The EMBO Journal 2004;
23(20):4061-71.
36. Richly H, Aloia L, Di Croce L. Roles of the Polycomb group proteins in stem cells and
cancer. Cell Death & Disease 2011; 1(2): e204.
37. Rinaldi L, Benitah S.A. Epigenetic regulation of adult stem cell function. The FEBS Journal
2014; 282:1589–1604.
38. Roesch A, Vultur A, Bogeski I, Wang H, Zimmermann KM, Speicher D, Korbel C, Laschke
MW, Gimotty PA, Philipp SE, Krause E, Patzold S, et al. Overcoming intrinsic multidrug
resistance in melanoma by blocking the mitochondrial respiratory chain of slow-cycling
JARID1B(high) cells. Cancer Cell 2013;23: 811-25.
39. Senthilkumar R, Mishra RK. Novel motifs distinguish multiple homologues of Polycomb in
vertebrates: expansion and diversification of the epigenetic toolkit. BMC Genomics 2009;
10:549.
40. Shang S, Hua F, Hu ZW. The regulation of β-catenin activity and function in cancer:
therapeutic opportunities. Oncotarget 2017; 8(20):33972–33989.
41. Sharma SV, Lee DY, Li B, Quinlan MP, Takahashi F, Maheswaran S, et al. A chromatin-
mediated reversible drug-tolerant state in cancer cell subpopulations. Cell 2010; 141:69–80.
42

42. Shmelkov SV, Butler JM, Hooper AT, Hormigo A, Kushner J, Milde T, et al. CD133
expression is not restricted to stem cells, and both CD133þ and CD133 metastatic colon cancer
cells initiate tumors. J Clin Invest 2008; 118:2111–20.
43. Suh, J.L. Development of Small Molecules and Peptidomimetic Ligands Targeting
Epigenetic Reader Proteins. PhD dissertation, University of North Carolina at Chapel Hill 2019.
44. Taherbhoy A.M, Huang O.W, Cochran A.G. BMI1–RING1B is an autoinhibited RING E3
ubiquitin ligase. Nature Communications 2015; 6:7621.
45. Tang B, Tian Y, Liao Y, Li Z, Yu S, Su H, Zhong F, Yuan G, Wang Y, Yu H, Tomlinson S,
Qiu X, He S. CBX8 exhibits oncogenic properties and serves as a prognostic factor in
hepatocellular carcinoma. Cell Death & Disease 2019; 10(2):52.
46. Tsuchida R, Das B, Yeger H, Koren G, Shibuya M, Thorner PS, et al. Cisplatin treatment
increases survival and expansion of a highly tumorigenic side-population fraction by
upregulating VEGF/Flt1 autocrine signaling. Oncogene 2008; 27:3923–34.
47. Vidal M, Starowicz K. Polycomb complexes PRC1 and their function in hematopoiesis.
Experimental hematology 2017; 48:12-31.
48. Villa M.P, Muñoz M.A, Pigolotti S. Bet-hedging strategies in expanding populations. PLOS
Computational Biology 2018; doi: 10.1371/journal.pcbi.1006529
49. Wotton D, Merrill JC. Pc2 and SUMOylation. Biochem Soc Trans 2007; 35(Pt 6):1401-4.
50. Xu T, Li H.T, Wei J, Li M, Hsieh T.C, Lu Y.T, Lakshminarasimhan R, Xu R, Hodara E,
Morrison G, Gujar H, Rhie S.K, Siegmund K, Liang G, Goldkorn A. Epigenetic plasticity
potentiates a rapid cyclical shift to and from an aggressive cancer phenotype. International
journal of cancer 2020, in press.
43

51. Yano S, Ito Y, Fujimoto M, Hamazaki T.S, Tamaki K, Okochi H. Characterization and
Localization of Side Population Cells in Mouse Skin. Stem Cells 2005; 23:834–841
52. Yuan S, Norgard R.J, Stanger B.Z. Cellular Plasticity in Cancer. Cancer discovery, 2019;
9(7):837-851.
53. Zeng D, Liu M, Pan J. Blocking EZH2 methylation transferase activity by GSK126 decreases
stem cell-like myeloma cells. Oncotarget 2017; 8(2):3396–3411.
54. Zheng X, Carstens JL, Kim J, Scheible M, Kaye J, Sugimoto H, et al. Epithelial-to-
mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic
cancer. Nature 2015; 527:525–30.

Abstract (if available)

Abstract Stem cell-like phenotypes such as drug resistance and tumorigenicity have been traditionally viewed as pre-existing characteristics of a specific subpopulation of cells within the whole tumor population. This assumption has been challenged by recent studies, showing that the stem-like drug-resistant phenotype can emerge de novo by conversion of a subpopulation of cells initially exhibiting a well-differentiated drug-sensitive state. Previous studies from our laboratory established the cancer stem-like plasticity model, by analyzing cancer cell lines over time using flow cytometry with Hoechst 33342 exclusion. This method is commonly used to yield a side population of cells with drug-resistant highly tumorigenic properties, and a non-side population lacking these properties. This model showed that these two subpopulations are able to go back and forth between the two phenotypes in a rapid and cyclical manner, making them a “moving target”. This phenotypic plasticity allows cancer cells to spontaneously and without external selective pressure switch their phenotype and acquire a survival advantage under fluctuating environmental conditions. The rapid and dynamic phenotypic transition is assumed to be in part epigenetically regulated, as alterations in chromatin accessibility and DNA methylation has been previously observed between genetically-identical subpopulations by the Goldkorn lab and other groups. In the present study, we investigated the potential involvement of the Polycomb Repressive Complexes (PRC), inhibitory chromatin modifiers, in the epigenetic regulation of phenotypic plasticity. To test this, we targeted EZH2 and CBX, components of PRC2 and PRC1, respectively, that are involved in a cascade of events that leads to epigenetic transcriptional repression. EZH2 is a tri-methyl writer, responsible for targeting a specific genomic region and depositing the H3K27me3 repressive mark subsequently recognized by CBX proteins. We used pharmacological inhibitors to target these interactions and performed side population analysis to identify the highly tumorigenic stem-like cells and monitor for shifts in the emergence of this subpopulation post-treatment.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The cancer stem-like phenotype: therapeutics, phenotypic plasticity and mechanistic studies

PDF

RNA methylation in cancer plasticity and drug resistance

PDF

Polycomb repressive complex 2 subunit stabilizes NANOG to maintain self-renewal in hepatocellular carcinoma tumor-initiating stem-like cells

PDF

The noncanonical role of telomerase in prostate cancer cells: exploring a non-telomeric signaling role for telomerase protein (TERT) in a cancer cell line

PDF

Exploration of the roles of cancer stem cells and survivin in the pathogenesis and progression of prostate cancer

PDF

Interaction of epigenetics and SMAD signaling in stem cells and diseases

PDF

Exploring the effects of CXCR4 inhibition on circulating tumor cell populations in metastatic prostate cancer

PDF

Epigenetic plasticity of cultured female human embryonic stem cells and regulation of gene expression and chromatin by PR-SET7 mediated H4K20me1

PDF

The effect of tumor-mediated immune suppression on prostate cancer immunotherapy

PDF

Molecular signature of aggressive disease and clonal diversity revealed by single-cell copy number analysis of prostate cancer cells across multiple disease states

PDF

Targeting glioma cancer stem cells for the treatment of glioblastoma multiforme

PDF

Investigating the effects of T cell mediated anti-leukemia activity in FLT3-ITD positive acute myeloid leukemia

Asset Metadata

Creator Avihen Schahaf, Nofar (author)

Core Title Investigating the effects of polycomb repressive complex inhibitors on cancer cell phenotypic plasticity

School Keck School of Medicine

Degree Master of Science

Degree Program Molecular Microbiology and Immunology

Publication Date 05/03/2020

Defense Date 03/04/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag acquired resistance,bet-hedging,Bladder,bladder cancer,cancer,cancer cell,CBX7,CBX8,chromatin accessibility,confluence-dependent plasticity model,cyclical cancer plasticity,DNA,DNA methylation,DNA modifications,drug resistance,epigenetic regulation,epigenetics,EZH2,GSK126,H3K27me3,inhibitors,intra-tumoral heterogeneity,modifications,non-side population,OAI-PMH Harvest,phenotypic plasticity,plasticity,plasticity model,polycomb,polycomb repressive complex,polycomb repressive complex inhibitors,PRC1,PRC2,side population,stem cell-like,UNC4976,UNC7040,UNC7263

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Landolph, Joseph (committee chair), Goldkorn, Amir (committee member), Machida, Keigo (committee member)

Creator Email nofaravihen@gmail.com,nofarschahaf@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-293825

Unique identifier UC11663602

Identifier etd-AvihenScha-8399.pdf (filename),usctheses-c89-293825 (legacy record id)

Legacy Identifier etd-AvihenScha-8399.pdf

Dmrecord 293825

Document Type Thesis

Rights Avihen Schahaf, Nofar

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA